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
|
|---|---|---|---|---|---|
Central American countries face a major challenge in the control of Triatoma dimidiata , a widespread vector of Chagas disease that cannot be eliminated . The key to maintaining the risk of transmission of Trypanosoma cruzi at lowest levels is to sustain surveillance throughout endemic areas . Guatemala , El Salvador , and Honduras integrated community-based vector surveillance into local health systems . Community participation was effective in detection of the vector , but some health services had difficulty sustaining their response to reports of vectors from the population . To date , no research has investigated how best to maintain and reinforce health service responsiveness , especially in resource-limited settings . We reviewed surveillance and response records of 12 health centers in Guatemala , El Salvador , and Honduras from 2008 to 2012 and analyzed the data in relation to the volume of reports of vector infestation , local geography , demography , human resources , managerial approach , and results of interviews with health workers . Health service responsiveness was defined as the percentage of households that reported vector infestation for which the local health service provided indoor residual spraying of insecticide or educational advice . Eight potential determinants of responsiveness were evaluated by linear and mixed-effects multi-linear regression . Health service responsiveness ( overall 77 . 4% ) was significantly associated with quarterly monitoring by departmental health offices . Other potential determinants of responsiveness were not found to be significant , partly because of short- and long-term strategies , such as temporary adjustments in manpower and redistribution of tasks among local participants in the effort . Consistent monitoring within the local health system contributes to sustainability of health service responsiveness in community-based vector surveillance of Chagas disease . Even with limited resources , countries can improve health service responsiveness with thoughtful strategies and management practices in the local health systems .
The prevalence of Chagas disease in Central America decreased from 1 . 7 million in the 1990s to 0 . 4 million in 2010 as a result of successful vector control [1 , 2] . Of the two main vectors , Rhodnius prolixus is almost eliminated , but Triatoma dimidiata remains widespread in the region despite greatly reduced rates of household infestation [3–8] . To prevent transmission of Chagas disease resulting from re-infestation of houses by T . dimidiata in areas with limited resources , Guatemala , El Salvador , and Honduras implemented community-based surveillance , in which community members report the presence of bugs in houses to trigger a response by local health services of the Ministry of Health [5 , 9 , 10] . Community-based surveillance has been shown to be effective and cost-effective , but can be challenging to sustain [11–14] . Household infestation with vectors can be detected readily by inhabitants , and in turn , health services are expected to respond to every vector report by visiting houses to spray insecticide and provide educational advice [9 , 11] . However , little is known about the extent to which vector reports from the community are met with appropriate responses and the factors that determine responsiveness of health services to vector reports . Research on responsiveness of health services may provide insights that help sustain and strengthen vector surveillance throughout the region . We retrospectively analyzed health services’ response rates and underlying determinants in community-based vector surveillance of Chagas disease in Guatemala , El Salvador , and Honduras .
We selected 12 areas with community-based vector surveillance in Guatemala , El Salvador , and Honduras–four from each country ( Fig 1 ) . Each area was a conglomerate of villages and defined as being under the jurisdiction of a particular health center . For inclusion of an area in the study , the Ministry of Health had to have completed the attack phase in all villages by conducting multiple cycles of extensive insecticide spraying of at-risk houses to reduce household vector infestation , implemented community-based vector surveillance , and recorded data from 2008 to 2012 . To compare management styles in unevenly decentralized health systems , we included one area per Department . The selected study areas were rural and in the most endemic districts of the Departments . The study areas varied in population size ( 1 , 160 to 33 , 579 persons ) , geographic area ( 6 to 150km2 ) , entomological situation , and human resources ( Table 1 and Fig 2 ) . The main target vector for surveillance in the study areas was Triatoma dimidiata , although in six areas ( two in Guatemala and four in Honduras ) surveillance also focused on Rhodnius prolixus because of previous history of infestation . All 12 health centers had physicians , nurses , and operational technicians except San José de la Reunión in Honduras , which had no physicians , and Ojo de Agua in Guatemala , which had no nurses ( Table 1 ) . Operational technicians had different qualifications or responsibilities . Vector control was carried out by vector control specialists and occasionally assisted by unspecialized rural health technicians in Guatemala; and was jointly conducted by vector control specialists , health promoters , and environmental sanitation inspectors in El Salvador . In Honduras , environmental health technicians were responsible for food security , environmental sanitation , and zoonoses as well as vector control . Some technicians belonged to neighboring health centers or a departmental office , and covered the health centers through regular visits . Community health volunteers were present in all 12 health centers and insecticide sprayers were present in nine . We defined a health service’s response rate as the percentage of the number of households sprayed or visited for advice by the local health services divided by the number of households infested with Chagas disease vectors as reported by the community . The annual response rate was calculated for each study area between 2008 and 2012 , so that a total of 60 response rates ( 12 areas x 5 years ) potentially were available for analysis . If the response occurred during the year following notification , it was considered as an action of the year of notification . A household with consecutive notifications of vector infestation was counted as a single infested household until the health service responded , regardless of interval length between notification and response . Taking in account factors that might influence demand , supply , and work process in community-based vector surveillance , we selected for analysis the following eight variables as potential determinants of health service responsiveness: number of infested households as reported by the community; distance from health centers to departmental capitals; number of operational technicians per 1 , 000 households; numbers of community volunteers and sprayers per 1 , 000 households; interval between receipt of vector reports from the community and response by health services , i . e . <3 months , 3–12 months or >12 months; degree of decentralization of response to vector reports , i . e . by health center or departmental office; presence of consistent monitoring by departmental technicians; and presence of technical assistance by JICA . We collected data on surveillance activities , local demography , geography , and human resources during visits to the departmental health offices and health centers during 2013 . We interviewed personnel responsible for Chagas disease vector surveillance in each facility to identify any perceived factors or circumstances that might have influenced responsiveness during the five year period .
Communities reported a total of 2 , 630 households with T . dimidiata infestation in the 12 study areas between 2008 and 2012 . Of these , the Ministry of Health responded to 2 , 041 households ( response rate 77 . 6% , Table 2 and S1 Table ) . Of the 2 , 041 responses , 68 . 4% were by insecticide spraying and the reminder by providing education and advice . Values of the eight variables that potentially influenced health service’s response rates differed among the health centers , but remained relatively constant within health centers over the 5-year period , with the exception of number of infested households reported , consistent monitoring by departmental technical officials , and technical assistance by the JICA project ( Table 3 ) . Numbers of health workers fluctuated according to trainees’ temporary assignments , and the population size of areas grew over time , but we treated these data as constant over the five year period . Of the eight variables analyzed , two were found by linear regression and mixed-effects multi-linear regression to be significantly associated with health service responsiveness: consistent monitoring by departmental technicians and technical assistance by JICA ( Table 4 ) . In both regression analyses , consistent monitoring from the departmental level was correlated positively with health service responsiveness to a moderate degree ( r = 0 . 48–0 . 55 ) whereas the correlation of assistance from JICA was weak and negative ( r = -0 . 13 ) . Health centers in Dolores , Honduras and Comapa , Guatemala reported large numbers of infested households in 2009 and 2012 , following campaigns in schools to promote bug searches as explained during interviews with health center staff ( S1 Table ) . Response rates followed four general patterns over the 5-year period: 1 ) nearly 100% response for most of the period , 2 ) nearly 100% for years 1 to 3 but then falling , 3 ) fluctuating moderately ( between 50% and 100% ) , and 4 ) fluctuating substantially ( between 0% and 100% ) but with a tendency towards improvement . When mixed-effects multi-linear regression was clustered by response pattern , similar associations between health services’ response rates and regular monitoring by departmental technicians ( r = 0 . 71 , p<0 . 01 ) and assistance from JICA ( r = -0 . 15 , p<0 . 01 ) were seen as in earlier models ( S2 Table ) . Interviews with health center personnel offered insight into the reasons underlying the different patterns ( Table 5 ) . Centers with higher response rates appeared to be more prepared to react to reports of infested houses; had better trained and more engaged workers; had superior management skills for coordinating and solving problems; and had greater support from higher institutional levels and local stakeholders such as community health volunteers and municipalities . Interviews with the personnel of health centers and departmental health offices identified the persons responsible for different surveillance functions ( Table 6 ) . Bug detection was performed by the population in all study areas . Operational technicians or clinical staff of health centers were responsible for analysis , decision making and planning of response in 7 of the 12 study sites , while personnel at the Department level carried out this function in the other 5 areas ( Table 6 ) . Health promotion , bug reporting , and response to reports were conducted by distinct combinations of stakeholders in the different study areas . Overall Honduras recorded higher degrees of involvement by community personnel and clinical staff and lesser involvement by operational technicians than Guatemala and El Salvador ( Fig 4 ) .
We found that regular ( quarterly ) monitoring by departmental health offices was a significant determinant of health service responsiveness in community-based vector surveillance of Chagas disease in Guatemala , El Salvador , and Honduras . Perhaps surprisingly , response rates were significantly higher among health centers without presence of technical assistance by the donor ( JICA ) . However , this finding can be explained by the presence of JICA at early stages of planning and implementation of the surveillance program at each study area , during which time response rates were low or fluctuated but subsequently improved . Three-year bilateral projects to establish community-based vector surveillance began in 2008 in El Salvador and Honduras and in 2009 in Guatemala . Health service responsiveness was independent of the volume of bug reports from the community , distance between health centers and departmental offices , numbers of operational technicians in the local health service and community workers , intervals between vector report and institutional response , and degree of decentralization of response . Interviews with health center staff demonstrated the effectiveness of regular monitoring on responsiveness and a decline in response rates following the departure of departmental supervisors in two health centers . This finding confirms previous research on primary health care services in low-resource settings , which showed that work performance was not motivated by written guidelines but by monitoring [15] . Because monitoring in this study provided an opportunity for departmental technicians and health center staff to review surveillance data , check equipment and supplies , participate in meetings with community health volunteers , and exchange information and experiences , continuation of quarterly visits should maintain or improve work performance over time . On the contrary , the consequences of inadequate monitoring can be serious in the long run , as reported in Gran Chaco in Argentina , where failure to supervise community personnel caused dysfunction of vector surveillance and reemergence of Chagas disease transmission [12] . Interviews also shed light into the lack of association between the other potential determinants and health service responsiveness . Although greater numbers of vector reports , for example following campaigns at school , increased the workload of local health services , response rates did not decline because manpower was augmented to meet the demand and tasks were reassigned among local stakeholders . Departmental technicians temporarily increased response capacity by mobilizing operational staff from other districts ( as often occurs in reaction to dengue outbreaks ) and by organizing extensive spraying operations with health center staff and community sprayers from different villages in the jurisdiction . Stakeholder analysis showed that surveillance tasks normally carried out by health specialists were simplified by the National Chagas Program and shifted to less specialized personnel through training , as we and others have reported previously [9 , 16 , 17] . In short , such combinations of short-term and long-term strategies reinforced responsiveness of health services . Managerial responsibility for response at the departmental office rather than the health center did not appear to affect the response rate . Although the departmental response approach was more vertical and less integrated into primary health care services , interviews showed that both departmental health offices and health centers with high responsiveness were able to find solutions for difficult situations . For instance , departmental vector teams assigned a data collection technician to health centers , concentrated response efforts in time and space , and travelled by motorcycles to reduce transportation costs . A physician and a nurse at one health center posted a large map of the jurisdiction on a billboard in the waiting room and used thumbtacks to represent the number of households reporting vectors in each village and removed them following the appropriate response . Such strategies reinforced the management capacity of the local health services . Longer intervals between receipt of vector reports and health service response did not lead to either higher response rates because of greater efficiency from economies of scale , or to lower response rates due to increased demands to deal with greater number of bug notifications . However , longer intervals are worrisome because of extended time of exposure of the population to the vector and thus greater risk of transmission of infection . Another potential negative impact is that the community may become reluctant to participate in bug notification if the interval is perceived as too long . While portraying the reality of vector surveillance in Guatemala , El Salvador , and Honduras , this observational study has important limitations . The sensitivity of the analysis may have been affected by the limited number of infested households in certain areas and during specific years , and by lack of data at the individual household level , which would have detected repeatedly infested and responded households . Our resources were insufficient to measure outcomes such as household vector infestation rates and incidence and prevalence of Chagas disease . These data would enable analysis of the consequences of not achieving 100% response rate; the effect of spraying vs . educational advice; and the impact of variable quality of responses by specialized vs . lay workers . We were unable to conduct cost analyses that would allow us to compare the effectiveness of the different styles and approaches to integrated surveillance , which varied substantially among the 12 study areas [12] . Further research is needed to address these limitations as well as long-term effects of monitoring on community-based surveillance where stakeholders may be changing . The greatest challenges to control of Chagas disease in Central America are non-eliminable , widespread vectors and underfunded and irregularly decentralized health systems . Although the disease has been targeted for elimination [18] , a more realistic approach is to prepare for permanent control in the region [19] . The success of vector control efforts in reducing household infestation and disease prevalence have made vector bugs and patients less visible and made the interventions less likely to be prioritized for government budgets in the future . Prospects for external funding are not good , since international aid agencies are often attracted to health problems which are eliminable or reducible to a great extent in a short time . Thus , Chagas disease control strategies need to be extraordinarily cost-effective and sustainable , and intervention models should be simple enough to be readily integrated and monitored in local health systems at different stages of decentralization . Although in Guatemala , El Salvador , and Honduras community-based vector surveillance for Chagas disease is part of the local health systems and functions with existing human resources and minimum costs , reductions in budget could affect availability of transportation and insecticide , and consequently health service responsiveness . In the control of non-eliminable vectors , such as T . dimidiata , the roles of continued spraying of infested houses and alternative interventions must be determined . In our study , 33 . 5% of responses to infested households was by insecticide spraying in Guatemala , versus 95 . 8% in El Salvador and 84 . 0% in Honduras . This partly reflects periodic scarcity of insecticides in the Guatemalan Ministry of Health , but also a deliberate shift towards house improvement . Multiple cycles of insecticide spraying are effective in reducing household infestation [20] , but are costly and difficult to sustain in the long run . Moreover , continuous application of insecticide might promote emergence of resistance in vectors . On the other hand , risk factors such as cracked mud walls , dirt floors , thatched roofing , and improperly tiled roofing [21] can be mitigated using locally available materials [22 , 23] . The cost-effective approach for improving house structures and living conditions innovated by Guatemalan researchers was adapted by the country’s Ministry of Health [22 , 23] . Also , local operational technicians developed an effective community organization approach which promotes engagement by the population and local government , and efficient implementation and scale-up of the house improvement method [5] . Evaluation of these efforts should also be part of the future research agenda . This research found that consistent monitoring at the departmental level of the Ministry of Health makes a significant difference in health service responsiveness in community-based vector surveillance of Chagas disease . Other potential factors , such as the number of infested households , numbers of health personnel and community workers , distance from departmental health offices to health centers , and degree of decentralization of response seemed to have limited impact on health service responsiveness . Challenges related to these factors were met largely because of managerial efforts of the local health services in implementing short-term and long-term strategies . Basic management practices such as monitoring and supervision combined with thoughtful strategies can improve health service responsiveness in resource-limited settings .
|
Elimination of domiciliated vectors led to a decreased prevalence of Chagas disease in parts of Latin America . In Central America , where the domiciliated vector Rhodnius prolixus has been almost eliminated , Triatoma dimidiata , which cannot be eliminated , continues to threaten the population in vast areas . To maintain the risk of transmission of Trypanosoma cruzi at lowest levels despite limited resources , Guatemala , El Salvador , and Honduras integrated community-based vector surveillance into local health systems . One challenge to sustaining surveillance is to ensure continuous responsiveness to reports of household infestation from the community . Our research in 12 study areas in the three countries over a five-year period investigated eight potential determinants of health service responsiveness , including volume of vector notifications , local geography , demography , manpower , and managerial approach . We found that consistent ( quarterly ) monitoring by departmental personnel within the local health services was associated with high response rates . Results of interviews added additional insight .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Determinants of Health Service Responsiveness in Community-Based Vector Surveillance for Chagas Disease in Guatemala, El Salvador, and Honduras
|
Malaria in sub-Saharan Africa has historically been almost exclusively attributed to Plasmodium falciparum ( Pf ) . Current diagnostic and surveillance systems in much of sub-Saharan Africa are not designed to identify or report non-Pf human malaria infections accurately , resulting in a dearth of routine epidemiological data about their significance . The high prevalence of Duffy negativity provided a rationale for excluding the possibility of Plasmodium vivax ( Pv ) transmission . However , review of varied evidence sources including traveller infections , community prevalence surveys , local clinical case reports , entomological and serological studies contradicts this viewpoint . Here , these data reports are weighted in a unified framework to reflect the strength of evidence of indigenous Pv transmission in terms of diagnostic specificity , size of individual reports and corroboration between evidence sources . Direct evidence was reported from 21 of the 47 malaria-endemic countries studied , while 42 countries were attributed with infections of visiting travellers . Overall , moderate to conclusive evidence of transmission was available from 18 countries , distributed across all parts of the continent . Approximately 86 . 6 million Duffy positive hosts were at risk of infection in Africa in 2015 . Analysis of the mechanisms sustaining Pv transmission across this continent of low frequency of susceptible hosts found that reports of Pv prevalence were consistent with transmission being potentially limited to Duffy positive populations . Finally , reports of apparent Duffy-independent transmission are discussed . While Pv is evidently not a major malaria parasite across most of sub-Saharan Africa , the evidence presented here highlights its widespread low-level endemicity . An increased awareness of Pv as a potential malaria parasite , coupled with policy shifts towards species-specific diagnostics and reporting , will allow a robust assessment of the public health significance of Pv , as well as the other neglected non-Pf parasites , which are currently invisible to most public health authorities in Africa , but which can cause severe clinical illness and require specific control interventions .
Malaria in sub-Saharan Africa has historically been almost exclusively attributed to Plasmodium falciparum ( Pf ) . The identification of the Duffy antigen as the obligate trans-membrane receptor for P . vivax ( Pv ) infection of red blood cells by Miller et al . during the 1970s [1 , 2] stalled research into the epidemiology of Pv in Africa as indigenous populations on this continent were known to rarely express the Duffy antigen ( and therefore be resistant to infection ) and the dogma of “Pv absence from Africa” became entrenched [3 , 4] . However , against a backdrop of increasing appreciation for the clinical severity of Pv infection [5–7] , multiple sources of evidence suggest that Pv may be more prevalent on this continent than commonly perceived [3 , 8–10] . The absence of any thorough effort to unify these sporadic reports , however , precludes appraisal of their significance , and assessment of the public health significance of Pv in sub-Saharan Africa . Plasmodium vivax has certain biological and epidemiological characteristics distinguishing it from Pf . Lower peripheral parasitaemia of blood-stage infections and a perception of lower risk of clinical complications during the era of malariotherapy of neuro-syphilis patients [5 , 11] contributed to the classification of Pv as “benign” , despite the parasite causing the same spectrum of clinical symptoms as Pf [5 , 7 , 12 , 13] . A key difference is the Pv parasite’s ability to form dormant liver-stages ( “hypnozoites” ) which evade the human immune system and blood-stage therapy , and can trigger relapses of clinical episodes weeks to months following the initial infective mosquito bite [14 , 15] . Without adequate treatment , hypnozoites therefore impose on the host a cumulative burden of blood-stage infection , contributing to the disease’s severe malaria phenotype , notably as severe anaemia [16] . These ( and other ) life-stage features unique to Pv mean that several aspects of the control interventions designed against Pf , including preventive measures related to vector biting behaviour [17 , 18] , diagnosis and treatment [15 , 19 , 20] are not effective against Pv , therefore making this parasite a greater challenge than Pf in achieving malaria elimination [21 , 22] . Current WHO diagnosis and treatment guidelines for the WHO African Region ( AFR ) do not allow for the peculiarities of Pv infections , and are exclusively focused on Pf ( see Box 1 ) . Untreated , however , chronic relapses of Pv can cause severe morbidity and mortality [16] . Community surveys in Liberia published in 1949 reported 2 . 0% of all malaria infections being due to Pv [23] . However , the acceptance of P . ovale ( Po ) as a separate species [24 , 25] , first as a single species and later two [26] , and the demonstration of the dependency of Pv on the Duffy antigen meant that the idea of Pv absence from Africa became entrenched , and during the latter part of the twentieth century apparent diagnoses of Pv were routinely reclassified as Po [3] , a morphologically very similar parasite [24] . During this period of near-total reliance on microscopy diagnosis , Pv largely disappeared from sub-Saharan African epidemiological records . The rise of the molecular diagnostics era , however , provided the capacity to confidently differentiate human Plasmodium infections [27] , and with it came the first substantive evidence bringing into doubt both the universality of Miller’s observations and the absence of Pv from populations across sub-Saharan Africa . The other endemic Plasmodium species , Po and P . malariae ( Pm ) have lower parasitaemia than Pv [28] , so are also historically neglected by microscopy-based diagnostics , and therefore also rarely recorded without molecular diagnosis . While Pv is evidently not a major co-endemic parasite in sub-Saharan Africa , Mendis et al in 2001 nevertheless estimated an annual burden of 6–15 million cases across Africa [9] . Reports of Pv transmission from almost all countries across the continent [8] further justify closer investigation . A shifting attitude is discernible in the language of the 2014 World Malaria Report , which suggests that Pv can occur throughout Africa , but with a low risk of infection due to high prevalence of Duffy negativity [29]; and the 2015 WHO Treatment Guidelines [30] indicate that cases are rare outside the Horn of Africa ( with the exception of Mauritania and Mali ) [30] . Current WHO guidelines for diagnosis and treatment in AFR , however , do not reflect this risk , and for example , recommend rapid diagnostic tests that do not detect non-Pf species . Moreover , most sub-Saharan African countries have no country-specific guidelines for Pv treatment ( see Box 1 ) . The epidemiology of Pv in Africa remains poorly understood , despite being key to evaluating the need for adapting current control interventions , to estimating the burden of Pv disease globally , and to anticipating any changes that may result from reductions in Pf transmission [22] . Here we assess the existing evidence base of Pv transmission in sub-Saharan Africa and evaluate the hypothesised mechanisms which may be sustaining this .
A PubMed literature search starting 01/01/1985 was run using keywords “vivax” AND “[African country names] or [Africa]” ( last updated 19/12/2014 ) . Following abstract review for relevance , several categories of data emerged: locally-diagnosed clinical case reports , serological surveys , observations of infected vectors , cross-sectional prevalence surveys , and reports of imported malaria into non-endemic countries by infected travellers ( Fig 1 ) . The location , diagnostic details and numbers of Pv infections were recorded by data type . Pv-specific annual parasite incidence ( API ) data reported by health management information systems was also assembled ( see Gething et al 2012 for details [31] ) . A further literature search for references about African Anopheles vectors was run ( 15/12/2014 ) , using keywords “vivax” AND “[vector species names]” , with the seven dominant vector species across the Africa region [32]: Anopheles arabiensis , An . funestus , An . gambiae , An . melas , An . merus , An . moucheti and An . nili; An . pharoensis was also included as a WHO-defined “major Anopheles species” in certain African countries [29] . The numbers of Pv sporozoite-positive mosquitoes found in each survey location were recorded . The most commonly used epidemiological metric of malaria endemicity is the parasite rate ( PR ) , assessed by cross-sectional surveys of infection prevalence [33] . The Malaria Atlas Project ( www . map . ox . ac . uk ) PR surveys database , the product of nearly ten years of archiving [33] , was reviewed for evidence of Pv infections reported since 1985 ( last updated 17/2/15 ) . A dataset of imported malaria infections into malaria-free countries by travellers returning from sub-Saharan African countries was assembled from national and regional ( notably the European Centre for Disease Prevention and Control , ECDC ) surveillance programme reports ( search window November 2014 –January 2015 ) . A selection of malaria-free countries with robust surveillance systems were contacted to obtain summary surveillance data since 2000 on imported malaria cases into their country . Given that individuals from these non-endemic countries would have similar prior exposure ( i . e . be predominantly immunologically naïve ) , their country of origin was not relevant to the analysis , although varied countries were approached to increase the coverage of countries of infection ( for instance , countries from several continents [Europe , Americas , Asia , Oceania] and linguistic groups [Dutch , English , French , German , Portuguese , Spanish , etc] ) . Contributing countries and the reporting time intervals which could be assembled are listed in Fig 1 . Total numbers of infections were aggregated by probable country of infection . A framework was developed to assess the overall evidence of Pv transmission at the first sub-national administrative level ( Admin1 ) . First , the data available for each data type from each Admin1 were classified as being of category 1 ( strongest ) , 2 , or 3 ( weakest ) evidence ( Fig 2 ) , and second , the scores of the seven surfaces representing each evidence type were summed to provide an overall assessment of the relative strength of evidence of Pv transmission occurring in each Admin1 . The varied nature of the available data restricted this to being a qualitative analysis , without aiming at any quantitative estimate of endemicity/incidence or quantitative comparison with other Plasmodium parasites . The absolute number of positive reports was used as the metric of strength of evidence , rather than any proportional metric . The criteria for each evidence category varied according to the time window of exposure represented by the data type ( and thus the probability of local infection ) , the denominator size , and the diagnostic method sensitivity/specificity . For example , due to their limited range of movement , infected vector specimens were considered relatively strong evidence of local transmission . In contrast , the longer window of exposure detected by serological surveys meant sero-positivity could result from a transmission event in a different location prior to relocation to the study site . To reflect the reduced temporal specificity of the serological evidence , a relatively higher number of reports of sero-positive individuals was therefore necessary to provide the same strength of evidence of transmission as infections observed in real-time . The single time-point snapshot investigated during a community prevalence survey contrasts with a much larger time window and population denominator for clinical case reporting across a region . Therefore , evidence of parasitaemia from prevalence surveys was considered stronger evidence of ongoing transmission than routine reporting of symptomatic cases , and a larger number of positive symptomatic individuals was required to equate to the strength of evidence of asymptomatic infections from prevalence surveys . Annual Parasite Incidence data ( API ) indicating stable transmission ( ≥1 case/10 , 000 ) equated to the strongest category 1 , while unstable transmission ( <1 case/10 , 000 ) was category 3 . Clinical cases diagnosed by molecular diagnostic tools based on nucleic acid amplification techniques ( notably PCR ) were considered stronger evidence of transmission than by light microscopy or rapid diagnostic tests ( RDT ) due to the increased species-specific diagnostic accuracy . In the case of prevalence surveys , however , where infections were more likely to be of low parasitaemia , RDT and microscopy were expected to diagnose a smaller proportion of total infections than PCR-based diagnostics due to their differing limits of the detection [28] . It has been estimated that molecular diagnostics identify at least double the number of infections compared to conventional methods [34] . Therefore the numbers of infections diagnosed by the two methods correspond to differing proportions of the true parasite prevalence , with RDT and microscopy-based underestimating true prevalence . On the other hand , RDT and microscopy diagnostics were less specific . These two aspects therefore balanced out and it was considered that prevalence data from both diagnostics would correspond to similar strength of evidence of transmission . Reports of Pv infection among travellers returning from African countries could not be attributed the same strength of evidence as observations of infection among local residents . The relapsing nature of Pv parasites means that infections could have been acquired from travel prior to the reported journey . Furthermore , the nature of the imported infections dataset assembled here is highly opportunistic and incomplete . This data type was therefore down-weighted relative to reports of local infections . The data were mapped to the Admin1 level . Where multiple surveys within a data type were available from a single administrative region , the scores were based on the total number of reports . Returning traveller infections were only reported to the national level , so to allocate scores to Admin1 units , the overall number of reported Pv infections was divided by the total number of Admin1 units in the country . These infections were therefore strongly down-weighted relative to the other data . Finally , once the component data types were categorised and mapped , these were summed together in ArcMap 10 . 1 [35] . The weighting system accounted both for the strength of evidence in each category and for corroboration between data types: the more data types reporting Pv transmission , the higher the overall strength of evidence . The highest level , “conclusive evidence” , was only attained if two or more category 1 sources of evidence were available . The lowest level , “very weak evidence” corresponded to Admin1 units from which no direct evidence was available , only a relatively small number of infected traveller reports ( less than 1 case per number of Admin1 units , thus a low probability of local transmission ) . Using an approach similar to that previously published [8 , 31] , the geographic limits of Pv transmission were overlaid onto a 2015 population surface to estimate the PvPAR . Medical intelligence data and biological exclusions were applied to define the limits of potential transmission . First , API and routinely reported incidence data were reviewed [31] to identify areas which could be excluded for being risk-free of malaria; this exclusion was not species-specific due to the previously discussed imperfect diagnostic capacity of most countries , but instead only areas defined as “malaria free” were excluded . Second , areas where temperatures could not support Pv sporogony at any time in an average year were excluded by a modelled temperature suitability mask [36] . Aridity was not used as exclusion given that man-made conditions can facilitate vector breeding even in areas of extreme aridity; previous PvPAR estimates have “downgraded” risk from stable to unstable transmission levels based on aridity , but the lack of distinction between these levels here means no aridity exclusion was applied . Third , urban areas , as identified based on the Global Rural Urban Mapping Project ( GRUMP ) urban extents layer [37] , were excluded . While potentially overly-conservative , the exclusion of urban populations was consistent with previous PvPAR estimates [8 , 31] , and is based on the significantly lower infection risk in urban areas [38 , 39] . A population surface for 2015 was compiled from the WorldPop Project projections ( www . worldpop . org ) , supplemented by data from the Gridded Population of the World ( GPW ) v3 ( http://sedac . ciesin . columbia . edu/data/set/gpw-v3-population-count-future-estimates ) for the Comoros and São Tomé and Príncipe , which were not available from WorldPop . This was adjusted to the Duffy positive population using the previously published Duffy blood group frequency maps developed by geostatistical modelling of population surveys of blood group frequencies [4] . The median model prediction of the Duffy surface was used , together with the 25% and 75% quartiles ( 50% confidence interval ) of the Duffy model outputs . National-level PvPAR estimates were derived for the three Duffy thresholds within the defined limits of transmission . All spatial manipulations were run at 5 x 5 km resolution in ArcMap and ArcScene 10 . 1 [35] . The PR database was used to investigate aspects of Pv transmission across sub-Saharan Africa and compare them to the PvPAR and to Pf . Estimates of the prevalence of Duffy negativity at the PR survey locations were extracted from the modelled Duffy group frequency maps [4] . For the comparisons with Pf , only surveys with the diagnostic capacity to identify both species were included , ensuring that paired estimates of PvPR and PfPR were matched in space , time , and population sample characteristics . The analysis was also restricted to surveys of ≥50 individuals . A trio of spatially-matched estimates were therefore collated for PvPR , PfPR and Duffy negativity prevalence . This dataset was used to assess: ( i ) whether the observed PvPR values were consistent with infections being potentially limited exclusively to Duffy positive hosts , ( ii ) how the prevalence of infection and relative risk of infection differed between species among their specific subset of known susceptible hosts ( based on the assumption that the PvPAR was limited to Duffy positive hosts: PvPRFy+ = PvPR/Proportion of Duffy positive hosts ) , ( iii ) the relationship between infection prevalence of the two species , and whether prevalence of one could inform predictions of the other . Full methodological details are available in Text S1 . Reports of Pv infections in Duffy negative hosts were compiled from the different data types reviewed . Only studies that used molecular confirmation of both Pv infection and Duffy genotype were included . The reports were geopositioned and presented in relation to the local prevalence of Duffy blood group phenotypes estimated by the modelled geostatistical maps [4] .
The literature search identified 1 , 479 publications , of which 529 were considered relevant to studies of Pv in Africa . All reports of Pv infection were collated and geopositioned ( Fig 3 ) . A total of 15 surveys reporting infected Anopheles vectors were identified ( S1 Table ) , as well as six reports of Pv sero-positivity ( S2 Table ) , 109 sites with local Pv case reports ( S3 Table ) and 643 Pv-positive community prevalence surveys ( www . map . ox . ac . uk ) . Overall , reports of Pv were available from 21 of the 47 malaria-endemic countries examined . The maps in Fig 4 represent the aggregated scores mapped to the Admin1 level . Records of Plasmodium infections in travellers were accessed for 31 malaria-free countries , corresponding to a period of 182 surveillance-years between 2000 and 2014 ( Fig 1 ) . Overall , 1 , 339 Pv infections were reported from 42 suspected countries of infection spread across sub-Saharan African countries , corresponding to all but five countries considered here . Fewer than 20 cases of Pv were reported from 24 countries , while from nine countries there were 20–50 cases , and nine countries were associated with infection of 50 of more travellers over the reporting period . The number of cases that our searches assembled and the proportion of Pv infections relative to Pf are shown in Fig 5 , though these ratios were not included in the evidence-weighting framework which focussed exclusively on Pv with evidence strength classifications based on categorical classifications and not a quantitative analysis . Limitations to this dataset are discussed below . The composite map in Fig 6 reflects the strength of the overall available evidence in each Admin1 unit ( summarised by country in S5 Table ) . Of the 629 Admin1 units across malaria-endemic Africa , there were 28 units for which no evidence of transmission was identified ( island nations of Cape Verde and Mayotte , and the small states of Guinea-Bissau and Swaziland ) . In 488 Admin1 units , the only available evidence was from traveller infections , so evidence of transmission from these areas was considered weak ( n = 217 ) or very weak ( n = 271 ) . Of the remaining Admin1 units from which direct evidence was identified ( n = 113 ) , eight had conclusive evidence ( in Ethiopia , Madagascar and Mauritania ) , 20 had strong , 36 moderate and 49 weak evidence of local transmission . Overall , there was moderate to conclusive evidence of Pv from 18 countries , including four in West Africa , three in Central Africa , three in East Africa , two in Southern Africa , and all six countries of the “Horn of Africa+” ( HoA+ ) region ( which included Djibouti , Eritrea , Ethiopia , Somalia , Sudan and South Sudan ) . The weaker evidence categories represent areas of relatively higher uncertainty . An estimated 86 . 6 million individuals were at risk of Pv infection ( PvPAR ) across sub-Saharan Africa in 2015 ( interquartile range , IQR: 43 . 9–156 . 7 million; S6 Table ) . These individuals were Duffy positive hosts living outside urban areas across malaria-endemic Africa where temperatures were suitable for sporogony at some point during an average year . The confidence interval is based only on the IQR of the Duffy positive frequency surface [4] , without accounting for uncertainty in the population density dataset or the temperature suitability surface . The binary adjustment excluding risk of urban transmission results in a conservative PvPAR estimate which is also not represented in the confidence interval . Overall , 28% of the PvPAR was outside the HoA+ region , with 10 . 0 million in East Africa ( 4 . 7 million excluding Madagascar ) , 7 . 2 million in West Africa , 4 . 1 million in Southern Africa and 3 . 3 million in Central Africa ( S6 Table ) . In 12 countries distributed across all sub-regions , the national PvPAR was greater than one million; Ethiopia , Sudan and Madagascar carried the highest PvPARs . The spatial distribution of the PvPAR is illustrated in Fig 7 , showing clustered areas of relatively increased PvPAR across the continent . Transmission of Pv in sub-Saharan Africa has been considered unlikely given the high prevalence of Duffy negativity across the continent . However , the evidence indicates that Pv is present , so here we investigated ( i ) whether the observed infection rates could be consistent with transmission exclusively limited to Duffy positive hosts , and ( ii ) how Pv and Pf infection rates were related to one another in sub-Saharan Africa . Full discussion of the results of these analyses and additional figures are available in the Supplementary Information ( S1 Text and S1–S3 Figs ) . The data ( Fig 8A ) were first examined for consistency with transmission within the Duffy positive population subset . The scatter of the data in Fig 8B being almost exclusively concentrated below the x = y dashed line ( 1 , 540 of 1 , 546 surveys ) indicates that the proportion of individuals infected does not exceed the proportion of the population considered susceptible , thus fulfilling the necessary condition ( but without proving ) that Duffy negativity imposes an upper threshold on PvPR by infections being limited to Duffy positive hosts . Six PvPR values ( 1 . 7% of positive surveys ) were outliers to this , all from coastal areas of Western/Central regions , including 4 from São Tomé and Príncipe . The survey with the highest discrepancy was from Cameroon , where 13 of 269 individuals were PCR-Pv positive [40] . Fig 8B also indicates PvPR rising as the frequency of susceptible hosts increases , as would be expected . Next , PvPR estimates were re-scaled to prevalence of infection among the subset of Duffy positive hosts ( PvPRFy+ = PvPR/Proportion of Duffy positive hosts ) . S1 Fig plots PvPRFy+ against PfPR , revealing distributions of points across the graphs without any clear trends emerging between infection rates between species . Regional trends , however , were apparent , with higher relative risk of infection in regions of lower host availability . For instance , where PvPRFy+ was positive in the Western region ( n = 13 ) , all surveys showed higher likelihood of infection by Pv than Pf ( among the respective susceptible population sub-groups ) ( S1 and S2 Figs ) . In areas where Duffy positive hosts were rare ( across most of sub-Saharan Africa these are <5% ) , these individuals were at a higher risk of being infected by Pv than by Pf ( S2 Fig ) . As Duffy positive hosts become more common , the relative risk of infection becomes more evenly distributed around 0 . Finally , a logistic regression model was developed to test for an association at the population level between the parasite rates of the two species: could PR data from one species predict PR of the other ? No significant relationship could be identified from the paired PR dataset for surveys outside the HoA+ , and the association was not significantly different from a flat line ( S3 Fig ) . Within the HoA+ , PfPR was a highly significant predictor of PvPRFy+ ( p = 4 . 4x10-6 ) up to 15% PfPR , after which increases in PfPR did not result in further predicted increases in PvPRFy+; in contrast , PvPRFy+ within the HoA+ was a significant linear predictor of PfPR at all levels of endemicity . Substantial scatter and heterogeneity in the data , however , resulted in wide 95% confidence intervals around the predicted relationship even when a significant relationship was identified . The last few years have seen evidence emerging of Pv infections in Duffy negative hosts ( Fy-Pv+ infections ) , providing an additional potential mechanism sustaining Pv transmission across sub-Saharan Africa . These observations are from diverse geographic , demographic and host genetic landscapes . The PubMed literature review identified 19 sites across 6 countries of sub-Saharan Africa where Fy-Pv+ infections in 54 individuals have been confirmed by molecular diagnosis , all published 2010–2015 ( S4 Table and Fig 9 ) . Observations of Fy-Pv+ infections were widely distributed across the continent , including in populations with heterogeneous Duffy phenotypes ( with correspondingly higher PvPAR: Ethiopia and Madagascar ) as well as in populations where Duffy negativity was at near fixation ( where PvPAR was much lower: Angola , Cameroon , Equatorial Guinea and Mauritania ) . All Fy-Pv+ infections ( except one survey from Yaoundé , Cameroon ) were from areas classified as “rural” by GRUMP . The observations were from areas of varied population densities , ranging from <100 to >500 , 000 individuals in the surrounding 25 km2 area ( WorldPop Project data ) . Both symptomatic and asymptomatic Fy-Pv+ infections were reported . The reported repartition of Pv infections among Duffy negative and positive hosts ranged widely between surveys ( S4 Table ) , but the data did not permit any formal analysis of differential infection risk .
The PR data were found to be consistent with transmission being potentially sustained by Duffy positive hosts alone . However , an important limitation to the PR comparative analysis was the skewed distribution of positive surveys , with the majority coming from the HoA+ ( 73% ) , while the main area of interest was outside this region . The lack of any predictable association between PvPRFy+ and PfPR in regions outside the HoA+ suggests that change in the prevalence of one species is not reflected by the prevalence of the other species at the host population level . This outcome may indicate that the two infection rates are determined by different drivers of transmission linked to their differing biological characteristics , notably capacity to relapse . For instance in these areas of low Duffy positive host availability , the environmental drivers that are highly significant determinants of PfPR endemicity [50] , may be secondary to the impact of the local host dynamics for predicting PvPRFy+ . Alternatively , the noise from the heterogeneity of the small dataset ( exacerbated by the PvPR adjustment to PvPRFy+ ) may be masking a relationship similar to that in the more data-rich HoA+ region which would be possible to detect through a larger dataset . Bespoke transmission models accounting for relapse risks and the peculiarities of the predominantly Duffy negative landscape are required in order to estimate the critical community sizes and determine the transmission dynamics needed to sustain infection , to verify the plausibility of Pv in Africa being limited to Duffy positive individuals [51–56] . The quantitative PR data analysis presented here is heavily contingent on the Duffy negativity frequency map [4] . Of the 203 population surveys which informed that map , 41% were published pre-1990 . The increase in large-scale population movement since this time [57 , 58] means that an influx of Duffy positive alleles to the gene pool may be increasing the frequency of susceptible hosts with a strongly spatially clustered distribution , making the 86 . 6 million PvPAR estimate a potentially significant underestimate . Furthermore , the historical waves of migration from Duffy positive British , French , Indian , Lebanese , Portuguese etc , populations may also be underrepresented . The wide confidence interval of the PvPAR ( 43 . 9–156 . 7 million ) indicates the potential impact of changes to the Duffy maps . Furthermore , the GRUMP map which was used to identify urban populations for exclusion from the PvPAR , overestimate urban areas , so many high density rural populations may have been excluded , further contributing to a conservative PvPAR estimate . Plasmodium ovale represents an important potential confounder to the reliability of microscopic reports of Pv due to their morphological similarities [24] and the risk of Pv reports actually being misdiagnosed Po infections . Conversely , the opposite may also be true with Pv infections being misclassified as Po due to the belief of Pv absence [3] . Microscopy-based diagnoses were down-weighted relative to PCR-based data in the evidence framework to account for these uncertainties . There is also potential for cross-reactivity between Po- and Pv-specific antigens in serological screening . Little is known about the diversity of Po in particular , so there is a need for the development of additional species-specific reagents to reliably distinguish the two species . The finding of Pv being a common cause of traveller infections is consistently reported from varied sources ( Fig 5 ) . For instance , of 618 cases of Pv imported into Europe between 1999 and 2003 , 33 . 8% were reported to have been infected in Africa [59] . While molecular confirmation is not routine for returning traveller infections , this is increasingly common in China where thorough case investigations differentiate imported from indigenous cases [60] . Recent investigation of malaria infections among gold miners returning to China from Ghana after a median travel time of one year , found 42 of 874 infections were Pv and 1 was Po [61] . From a public health perspective , even if these several hundreds of infections are all misdiagnoses , given that the same treatment guidelines for radical cure apply to both species [30] , the potential misdiagnosis is not significant in terms of treatment policy and requires the same increased diagnostic capacity away from the perception of Pf exclusivity . A potential confounder to attributing all the assembled evidence of Pv to local transmission in Africa is the parasite’s ability to form dormant hypnozoites and present clinical symptoms only after a delayed period during which the patient may have travelled widely [14 , 62] . The evidence-weighting framework attempted to account for this possibility through the differing thresholds between categories . Again , though , even if a subset of the observed cases resulted from infection events elsewhere , the implications for diagnostic and treatment policy remain the same: appropriate capacity is required in the area of clinical presentation , irrespective of where transmission occurred . Liu et al [63] have published conclusive evidence of the origin of human Pv being from a genetically diverse parasite population whose natural hosts are gorillas and chimpanzees in Africa . Although the human parasite clade is distinct from the more diverse parasite strains that indiscriminately infect various ape species , there is evidence ( from one individual ) of the plausibility of cross-infection of parasite strains between ape and human hosts with similar clinical presentation [63 , 64] . Liu et al therefore argue that Pv infections reported from regions of high Duffy negativity frequency are zoonotic infections spilling over from ape reservoirs . Entomological evidence suggests that An . moucheti and An . vinckei may be potential vector species bridging transmission between apes and humans [64 , 65] , although the entomological evidence is very limited with only one reported infected specimen of each species and vector behaviour that is not conducive to frequent transmission events . The natural ranges of these ape reservoir species are restricted to the forests of Central Africa ( Fig 10 ) . So , while this infection mechanism may explain transmission in specific contexts , this does not represent a universal explanation for the evidence of Pv across predominantly Duffy negative regions of Africa . The reports of Duffy negative infections from diverse epidemiological settings ( 54 molecularly-confirmed symptomatic and asymptomatic infections identified across 19 sites; S4 Table ) represent a further potential mechanism sustaining Pv transmission across Africa . The public health burden of these infections is unclear . While the protective effect of Duffy negativity may not be as absolute as previously considered , this blood group’s high frequencies across Africa do nevertheless appear to represent a limiting force on potential Pv endemicity . In the absence of any limiting factor , and given the presence of competent vectors and a suitable climate , rates of Pv infection comparable to those in Asia or the Americas [31] would be expected . Fy-Pv+ infections demand further investigation , ideally across diverse epidemiological settings . The impact of reduced protection conferred by Duffy negativity ( currently assumed to be 100% ) would have important impacts on the resulting PvPAR ( see estimates in Zimmerman et al [67] ) . For the first time , this paper synthesises all available evidence of Pv transmission in Africa , and evaluates the varied mechanisms hypothesised to explain these observations across an area where transmission is largely unexpected . Here we argue that no single theory is sufficient to explain transmission across this varied continent , but instead , transmission pathways vary according to the complex ecology and epidemiology of each area and population . Evidence relating to Pv transmission across Africa appears inconsistent . Pv is clearly widely present , causing a substantial proportion of returning traveller malaria infections , clinical illness and asymptomatic infection among local residents , infecting vectors and presenting a history of exposure by local communities to infected bites . However , extensive surveys using high-sensitivity molecular methods have repeatedly failed to diagnose Pv [68–70] . For instance , PCR-based screening of 1 , 402 blood samples in southern Cameroon found no trace of Pv [69] , while a nearby PCR-based community survey of 269 individuals diagnosed a prevalence rate of 5% PvPR ( representing 13 . 8% of all Plasmodium infections ) [40] . The key difference between these neighbouring surveys in southern Cameroon was the demographic composition of the sampled populations: while no Pv was found in the remote , rural village communities [69] , the multi-ethnic and highly cosmopolitan population did have Pv [40] . While there is some evidence of an ape reservoir enabling zoonotic infections , this does not preclude the possibility of endemic transmission . Fig 10 makes evident the observations of Pv outside these ape natural ranges , and furthermore , the main group of travellers returning from Africa with Pv infections are visiting relatives [71 , 72] , and less likely to visit the tourist attractions which would bring them into contact with infected vectors [73] . Therefore , while there is evidence of the zoonotic reservoir being a contributing source of infection , other mechanisms of transmission ( and reservoirs of infection ) must also be involved . The multiple theories discussed so far for explaining the observations of Pv in Africa are not mutually exclusive . Instead , given the limitations to each hypothesis for explaining the overall diversity of transmission , it seems probable that all are involved to some degree . The dominant source of infection would vary between location: in remote forest areas , infections may be from the ape reservoir [64] , while in more cosmopolitan areas , transmission may be being sustained by an admixed Duffy positive population . The early production of gametocytes in infections [20] means that even where the host has developed immunity to symptomatic blood-stage parasitaemia from prior exposure , short-lived sub-clinical infections could still be helping to sustain transmission . It is likely that some reported cases of Pv are in fact Po misdiagnoses , but the sheer abundance of data from various diagnostic methods means this cannot be universal . Similarly , while there is convincing evidence of Fy-Pv+ infections , these are not reflected by the widespread Pv endemicity levels that would be expected if these infections were commonplace . The epidemiology of Pv in Africa is likely to be complex and multi-faceted , driven by different mechanisms in different regions , determined by the unique host genetic , vector , and reservoir characteristics of each area . These complimentary theories together may explain the observed evidence of Pv transmission . High sensitivity molecular methods could allow further insight into the population genetics of Pv infections diagnosed in Africa and therefore their epidemiology [74] , as well as helping to define true exposure and risk of exposure to infection . The non-specific clinical presentation of malaria means that diagnosis is only possible with a parasitological test [30] . The WHO T3 policy to “Test , Treat , Track” all cases of malaria is very logical in an era of resistance emergence and falling endemicity in many areas [75] . However , it becomes harmful if the “Test” component is not capable of detecting all potential parasites . In the many African countries exclusively reliant on HRP2-based RDT diagnosis ( see Box 1 ) , Pv , Po and Pm infected patients who present in clinics with fever symptoms will test negative and could leave untreated . The patients will persist in the community as sources of onward transmission , and , in the case of Pv and Po , be susceptible to the cumulative impact of clinical relapses . In turn , therapeutic options for Pv are inadequate . While Pf-targeted control will treat Pv clinical symptoms , it will not impact on the parasite reservoir sustaining the observed blood-stage clinical infections . Even in terms of blood-stage therapy , drug policy in Ethiopia ( ranked fourth largest contributor to Pv cases globally [29] ) maintains chloroquine as first-line treatment despite reports of resistance [76] . A further complication to delivering Pv-specific therapy in Africa is the high prevalence of glucose-6-phosphate dehydrogenase deficiency ( G6PDd ) [77] , causing a potentially dangerous intolerance to primaquine , the only available drug for treating the liver-stage parasites [78] . The same issues apply to Po radical cure . It has been hypothesised that the high prevalence of G6PDd may be the result of a selective advantage against malaria [79–81] , meaning that G6PDd might be under-represented in Pv patients and thus primaquine therapy a lesser concern . However , given Pv’s preference for invading reticulocytes–young red cells with highest levels of G6PD enzyme activity–it is unlikely that a relatively mild G6PD variant ( such as G6PDA- , the predominant African variant [82] ) , would provide any protective advantage against Pv infection . Finally , the evidence presented here has two main implications for Pv mapping . First , where PvPR surveys are available , the infection rates reported are consistent with transmission only among Duffy positive hosts . This therefore supports the approach previously followed of restricting the PvPAR to Duffy positive hosts [8 , 83] and of using the frequency of Duffy positive hosts as an upper threshold to the maximum potential prevalence of infection [31] . Second , the PfPR surface cannot be used to predict Pv infection prevalence in Duffy positive hosts , except potentially in transmission settings up to 15% PfPR in the HoA+ region . The implications of this paper should not be misinterpreted . This is not a “call to arms” requiring huge additional resources above and beyond the considerable efforts already ongoing into control of Pf . Current epidemiological data overwhelmingly indicates that Pf is the predominant malaria pathogen across most of sub-Saharan Africa [29 , 31 , 68] , and that where present , Pv prevalence remains low . Nevertheless , despite the estimated 86 . 6 million individuals at risk of infection , Pv is so drastically neglected across this vast and populous continent that without a commitment to add capacity for Pv in routine surveillance and reporting , this minor player will remain invisible and any quantitative burden estimates and evidence-based policy decisions impossible . Broadening the outlook beyond Pf would simultaneously also strengthen the evidence base relating to the other neglected endemic species , Po and Pm and increase access to diagnosis and treatment of these other neglected malaria species .
|
Plasmodium vivax ( Pv ) is the most widely distributed malaria parasite globally , but conspicuously “absent” from Africa . The majority of African populations do not express the Duffy blood group antigen , which is the only known receptor for Pv infection . Since this discovery in the 1970s , the low clinical incidence of Pv in Africa has resulted in a perception of Pv being completely absent and any apparent cases being misdiagnoses , and no public health allowances are made for this parasite in terms of diagnosis , treatment or surveillance reporting . As more sensitive diagnostics become available , Pv infection in Africa is increasingly reported from a variety of different survey types: entomological , serological , community prevalence surveys , as well as clinical infection data from local residents and travellers returning to malaria-free countries . A literature review was conducted to assemble these reports and assess the current status of evidence about Pv transmission in Africa . Moderate to conclusive evidence of transmission was available from 18 of the 47 malaria-endemic countries examined , distributed across all parts of the continent . Mechanisms explaining this reported transmission are evaluated , as well as alternative explanations for the observations . Combinations of complementary explanations are likely , varying according to regional ecology and population characteristics . The public health implications of these observations and recommendations for increased awareness of Pv transmission on this continent are discussed .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Plasmodium vivax Transmission in Africa
|
Genome-scale metabolic models have become a fundamental tool for examining metabolic principles . However , metabolism is not solely characterized by the underlying biochemical reactions and catalyzing enzymes , but also affected by regulatory events . Since the pioneering work of Covert and co-workers as well as Shlomi and co-workers it is debated , how regulation and metabolism synergistically characterize a coherent cellular state . The first approaches started from metabolic models , which were extended by the regulation of the encoding genes of the catalyzing enzymes . By now , bioinformatics databases in principle allow addressing the challenge of integrating regulation and metabolism on a system-wide level . Collecting information from several databases we provide a network representation of the integrated gene regulatory and metabolic system for Escherichia coli , including major cellular processes , from metabolic processes via protein modification to a variety of regulatory events . Besides transcriptional regulation , we also take into account regulation of translation , enzyme activities and reactions . Our network model provides novel topological characterizations of system components based on their positions in the network . We show that network characteristics suggest a representation of the integrated system as three network domains ( regulatory , metabolic and interface networks ) instead of two . This new three-domain representation reveals the structural centrality of components with known high functional relevance . This integrated network can serve as a platform for understanding coherent cellular states as active subnetworks and to elucidate crossover effects between metabolism and gene regulation .
So far , metabolic processes and gene regulatory events are typically considered individually in system-level investigations . However , ample evidence exists that the majority of cellular processes involves both , metabolism and gene regulation , and thus requires their joint examination [1] . One of the best-investigated individual examples in Escherichia coli ( E . coli ) is the phosphoenolpyruvate–carbohydrate phosphotransferase system ( PTS ) which is responsible for import and phosphorylation of sugars [2] . Additionally , the PTS is involved in the regulation of the import process depending on the available carbohydrate mixtures in the growth medium . By carbon catabolite repression and inducer exclusion , primarily the uptake of a preferred carbon source to be metabolized , such as glucose , is selected from other carbohydrates present in the growth medium . In order to understand the underlying principles , not only the effects of both ‘layers’ , metabolism and regulation , need to be taken into account , but also their interface [3] . On a more qualitative level , the importance of the interface of metabolism and gene regulation can be illustrated by having a closer look at their most prominent representatives , namely , enzymes and metabolic transcriptional regulators . Both examples are proteins and can be thought of as a component type organizing the interplay of genes and metabolic reactions ( Fig 1 ) . For enzymes the connection is straightforward: The majority of metabolic reactions can only take place if the corresponding genes of the catalyzing enzymes are expressed . These genes , in turn , are often involved in regulatory processes , especially if they are associated with central biochemical reactions . In contrast , metabolic transcriptional regulators can be illustrated by looking at transcription factors , the probably best-investigated transcriptional regulators . Some of them require the binding of a metabolite to be active and are therefore called metabolic transcriptional regulators . In the context of the integrative view discussed here , it is noteworthy that only the interaction with a metabolic component enables their functionality as gene expression regulators . Conventional reconstructions of E . coli’s metabolism as well as of its gene regulation thoroughly describe the process itself but usually lack information on interacting elements of the other biological system . While there are numerous genome-scale metabolic reconstructions available [4–9] , only a few large-scale transcriptional regulatory networks exist that are mainly based on the information from RegulonDB [10] . First attempts to integrate both cellular processes started from metabolic reconstructions which were expanded by regulatory genes and stimuli of the associated encoding metabolic genes [11 , 12] . Both studies started from the metabolic model of [5] and include 104 regulatory genes and 583 regulatory rules regulating approximately 50% of the metabolic genes . In this manner , the close proximity of regulatory events was captured but more far-reaching and global effects , e . g . , self-contained regulatory dynamics among genes , could not be considered . Further approaches examine the regulatory processes of the metabolic network based on the aforementioned pioneering attempts [13 , 14] . For this purpose , the information about regulatory events was assembled in terms of Boolean rules as a variant of Boolean network models . More recently , [15] introduced a method called probabilistic regulation of metabolism , a new variant of regulatory flux balance analysis , i . e . , the class of approaches behind some of the pioneering integrative models discussed above [11 , 12] . The state of the many variants of integrating regulatory information into flux-balance analysis models has been reviewed in [16] and [17] . The necessity of achieving such data integration , even on the network level , has recently been discussed in [18] . To a certain degree all these studies consider the regulation of metabolism but only cover the proximity rather than a genome scale . A very recent example of a successful topological characterization of biological networks , in order to understand the interplay of gene regulation and metabolism , is the analysis of gene regulatory—metabolic feedback loops [19] . In [19] joint representations of gene regulatory and metabolic networks have been compiled for two organisms , E . coli and B . subtilis . These network representations are then analyzed with a focus on the hierarchical structure of the resulting network and on feedback loops between gene regulation and metabolism . These feedback loops are then further characterized in terms of their potential impact ( defined as the number of genes downstream of the transcription factor receiving the feedback from metabolism ) , their possible function ( e . g . , in the processing of environmental information ) and other properties [19] . In this way , the authors obtain insight in the algorithmic features of the interface between gene regulation and metabolism . Understanding the interplay of metabolism and gene regulation will help to gain insight in cellular , system-wide responses such as to changing environmental conditions . Here , we present the database-assisted reconstruction of an integrative E . coli network capturing metabolic as well as regulatory processes . The attribution of network components ( in terms of individual vertices ) to the metabolic and regulatory domains , as well as the protein interface enables the further characterization of the network in terms of its modular organization , its path statistics and the vertex centrality . In particular , we formulate a new measure by evaluating domain-traversing paths , in order to quantitatively assess the role of components in the interface domain and thus identify cross-systemic key elements contributing to both regulatory and metabolic processes . In all cases , these topological assessments highlight system components and functional subsystems , which are well known for their biological relevance , thus emphasizing the predictive power of network topology . Employing observations on the topological ( structural , network-architectural ) level , in order to identify components in the system of particular functional relevance has a long tradition in network biology ( and in network science in general ) . The main results of our investigation are: We present an integrated network representation of gene regulation and metabolism of E . coli and illustrate how it is a promising starting point for the structural investigation of system-wide phenomena . In particular , the network perspective suggests the explicit consideration of a protein interface between the genetic and metabolic realms of the cell . Employing network metrics we ( 1 ) argue that a three-domain partitioning is architecturally and functionally plausible , and ( 2 ) show that prominent components of the network according to the structural investigation tend to be of evident biological importance . Especially , the evaluation of possible paths through the interface domain of the network reconstruction yields well-known functional subsystems . The overlap of structural and biological relevance here suggests that a careful analysis of such a structural model can guide biological investigations by focusing on a limited number of structurally outstanding components . This network model can also serve as a starting point for a range of topological analyses with methods developed in statistical physics ( see , e . g . , [20] for a recent review ) . Summarizing , in contrast to the separate analyses of ( e . g . , the metabolic or gene regulatory ) subsystems , we expect that the integrative network model shown here will draw the attention to system-wide feedback loops not contained in the individual subsystems and to different roles of individual components , which become only visible from the perspective of interdependent networks .
By now , the dramatic growth of bioinformatics databases [21] , both in content and in diversity , allows addressing the challenge of integrating regulation and metabolism on a system-wide level . We devised a semi-automated framework to integrate information from EcoCyc database [22] and RegulonDB [10] into a network for E . coli including major cellular processes , from metabolic processes via protein modifications to a variety of regulatory events ( see Methods ) . Networks are an efficient data structure for integrating this wealth of information [23–25] . In this way , the vast amount of data contained in the bioinformatics databases provide an ‘architectural embedding’ for metabolic-regulatory networks and guides subsequent steps of model refinement and validation . We augmented and validated the resulting network based on existing reconstructions of metabolic [6 , 8 , 26–28] as well as of gene regulatory processes [10] . The integrative E . coli network constructed here comprises the three major biological components , genes , proteins , and metabolites , as well as the metabolizing reactions summing up to more than 12 , 000 components . Represented as a graph , the network has seven types of vertices depicting the major biological components ( Fig 2 , Table A in S1 Text ) and seven different types of edges including two types of encoding associations , i . e . , transcription and translation processes , four types describing the associations within biochemical reactions , and one type summarizing regulatory relations ( Table B in S1 Text ) . Two small annotated biological examples are shown in Fig B in S1 Text . The graph representation facilitates the mapping of reactions and their catalyzing enzymes , as both are depicted as vertices . In contrast , metabolic systems are often represented as hypergraphs to illustrate the Boolean ‘AND’ association of reaction educts and the fixed stoichiometric ratio of the involved metabolites . Those aspects are assigned explicitly as edge properties in the graph representation . For the purpose of measuring the propagation of perturbations through the network , for example , the following logical assignments are helpful ( see [29] for details on these definitions ) : Besides the associations of reaction educts , the encoding relations of protein complexes are of Boolean ‘AND’ type , termed conjunct links . On the contrary , associations representing isoforms of protein subunits , isoenzymes as well as reaction products are implemented by Boolean ‘OR’ links , called disjunct . The third linkage type , regulation , covers approximately 7 , 300 regulatory associations , i . e . , transcriptional , translational as well as metabolic ones ( Table C in S1 Text ) . The comparison with existing models reveals that the presented integrative network is a comprehensive representation of the metabolic and regulatory processes in E . coli . The very first approach of embedding metabolic processes in the regulatory context of [11] , the iMC1010 model , started from a metabolic model which was extended by the regulation of the encoding genes of the catalyzing enzymes . For the purpose of determining the overlap of the integrative metabolic-regulatory network and the iMC1010 model , transport reactions as well as the artificial biomass reaction have been disregarded and , moreover , only unique metabolites ( neglecting compartmentation ) have been taken into account . Else , the different levels of details of the transport systems such as PTS as well as of the compound compartmentation would render a correct mapping impossible . Overall , the iMC1010 model is covered by our model to more than 89% ( Fig 2 , see Table D in S1 Text , column 3 ) . To assess the coverage of E . coli’s metabolic processes , the embedded metabolic processes of the integrative E . coli network have been associated to the ones of an established E . coli metabolic reconstruction , namely the iAF1260 model from [6] . About 67% of the involved biochemical reactions , compounds and genes could be mapped directly ( see Table D in S1 Text , column 4 ) . Particularly , these two thirds capture almost all biologically relevant components in terms of in silico viability . Using flux balance analysis for simulating the biomass production capacity of the iAF1260 model and taking the overlap with mapped components of the integrative E . coli network revealed that for the default medium setup approximately 75% of the essential reactions ( to yield 1% biomass ) are covered by the integrative E . coli network . Analogous to the metabolic processes , the coverage of E . coli’s gene regulation has been determined using the transcriptional regulatory network from RegulonDB [10] . This model has been assembled in a similar fashion but is accounting only for transcription factors and their regulated genes . With a coverage of more than 98% , the transcription-related regulatory processes are considered as completely recorded in the integrative E . coli network ( see Table D in S1 Text , column 5 ) . Apart from that , for this assessment of overlap a comparison of regulatory processes associated with RNA translation as well as metabolic regulatory events is not possible since the RegulonDB transcriptional regulatory network does not consider protein and metabolic interaction processes . The most conspicuous links between metabolic and gene regulatory processes are metabolic transcription factors , i . e . , gene expression regulators binding metabolites , and metabolic genes , i . e . , genes with significant and coordinated response on the metabolic level such as encoding enzymes . Intuitively , the interface is considered so far as the direct interactions of metabolic elements and gene regulatory elements , and the integrative E . coli network can be partitioned into metabolic and regulatory domain ( MD—RD ) . However , by examining those interactions in more detail the topological role of proteins becomes apparent . Regarding the metabolic transcription factors , the respective metabolite binds to a protein and this metabolite-protein complex then subsequently regulates the gene expression . In the case of metabolic genes , ultimately the respective gene encodes a protein which either by itself or as a complex serves as an enzyme . In line with this , the interface of metabolic and gene regulatory processes should be considered as the series of interactions of metabolites and genes , respectively , with proteins and subsequent protein modifications . Thus , the interface does not only comprise interactions ( edges ) but also components ( vertices ) , and the integrative E . coli network will in the following be divided into a metabolic domain , a protein interface and a regulatory domain ( MD—PI—RD ) . In the next section , the plausibility of the three-domain partition ( and the set of biologically motivated rules devised to create it ) will be assessed in comparison to the likewise proposed two-domain ( MD—RD ) representation . The integration of metabolic and regulatory events allows us to determine the key elements of E . coli , especially those beyond the individual processes . In particular , the functional three-domain partition facilitates to recover network components ( in terms of individual vertices ) of evident biological relevance , e . g . , by means of simple centrality measures . In the following , two different aspects of centrality have been examined [40]: degree centrality depicting the direct linkage of a vertex , and betweenness centrality which can be thought of as the participation of a vertex in the network flow [41] . Starting with the prominent local vertex structure , the so-called hubs ( here , vertices with a total degree larger than 50 ) , it is noticeable that they are primarily compounds and proteins , in particular protein complexes and appear in all three domains ( see Table H in S1 Text , columns 3–5 ) . In the metabolic domain , hubs include trivial compounds such as H+ and H2O and , so-called , currency metabolites , e . g . , ATP , NAD ( P ) H and coenzyme A , while hubs of regulatory processes are obviously global regulators which characteristically exhibit a remarkably strong asymmetry of in-degree and out-degree . Particularly , well-known transcriptions factors top this list such as FNR ( fumarate and nitrate reduction ) [42] , Fis ( factor for inversion stimulation ) and H-NS ( histone-like nucleoid structuring protein ) [43] . As stated above , hubs predominantly occur in metabolic and gene regulatory domain while only a few are affiliated to the protein interface . However , it was not to be expected to identify cross-systemic elements solely based on their degree . To assess/detect cross-systemic key elements an extended approach of degree centrality has been used that additionally accounts for the domain boundaries . The intra-domain degree fraction ξ , also termed embeddedness [44] , denotes the ratio of the internal degree of a vertex , within a domain , and the total degree in the network . This measure very clearly distinguishes between , on the one hand , metabolic and regulatory hubs which show intra-domain degree fractions ξ > 0 . 87 ( except one single compound with ξ = 0 . 185 ) and hubs in the interface which in contrast have ξ ≤ 0 . 06 ( see Table H in S1 Text , last column ) . Thus , while metabolic and regulatory hubs are embedded in their respective domains , hubs in the protein interface are mainly connected to vertices in the neighboring domains . In total , seven hubs show a significant low intra-domain degree fraction pointing to their prevalent interactions with the other two domains ( Fig A in S1 Text and Table K in S1 Text , column 5 ) . Six of them are affiliated to the protein interface exhibiting numerous interactions with the regulatory domain . Their linkages to the metabolic domain become visible when considering their composition , in case of the protein complexes , and their modes of action , respectively . The former involve the four protein-compound complexes Crp-cAMP ( cyclic-AMP receptor protein binding cyclic-AMP ) [31 , 45 , 46] , DksA-ppGpp ( dnaK suppressor binding guanosine 3’-diphosphate 5’-diphosphate ) [47–49] , NsrR-NO ( nitrite-sensitive repressor binding nitric oxide ) [50–52] and Lrp-Leu ( leucine-responsive regulatory protein binding leucine ) [53–55] whose naming schemes already indicate the metabolic link . The latter , namely , protein complex Cra ( catabolite repressor activator ) and protein monomer Lrp ( leucine-responsive regulatory protein ) form in the presence of appropriate metabolites , i . e . , fructose 1 , 6-bisphosphate/fructose 1-phosphate and leucine , complexes affecting their regulatory effect . The remaining hub is the metabolic-domain vertex representing guanosine 5’-diphosphate 3’-diphosphate ( ppGpp ) . Besides its special domain-affiliation among the low intra-domain degree hubs , ppGpp acts as an important regulator of both , metabolism and transcriptional processes . More precisely , it regulates several enzyme activities as well as numerous transcription initiations by allosterically binding to RNA polymerase . So far , we demonstrated that the protein interface of the E . coli network reconstruction acts as a bridging module between regulatory and metabolic domain enabling their interaction and communication . Therefore , we expect the betweenness centrality to directly highlight vertices from the interface . Indeed , ten out of the top-25-ranked ( still including currency metabolites ) vertices are from the interface ( see Table I in S1 Text , column 5 ) , while overall the interface only accounts for about 18% of the vertices of the network . Especially , the already mentioned protein-compound complexes Crp-cAMP and DksA-ppGpp are among these compounds . In general , currency metabolites and trivial compounds ( see above ) as well as global regulators are among the central components with respect to betweenness . Apart from that , biochemical reactions building up and/or breaking down these metabolites and proteins as well as the other involved reactants pertain to the most betweenness-central components . Component association to functional systems allows to assess the systemic feature and by considering the corresponding network affiliation to depict the candidates for cross-systemic key elements . In this manner the network analysis allows us to detect the central role of Crp-cAMP , Lrp-Leu and ppGpp on purely topological grounds , as each component is the focus of such a functional system with high betweenness . Additionally among the top-ranked vertices with respect to betweenness centrality are five further cross-systemic components which are assigned to the protein interface , namely , phosphorylated PhoB ( PhoB-P ) , Fur-Fe2+ , and three outer membrane proteins ( Omp ) , OmpC , OmpE and OmpF ( Table I in S1 Text ) . The former two components are transcription factors and therefore acting in the gene regulatory domain , while at the same time they are protein complexes binding a metabolic small molecule depicting the connection to the metabolic processes . The latter three , the outer membrane porins , form hydrophilic channels , enabling non-specific diffusion of small molecules across the outer membrane [56–58] . In this role these proteins represent the most obvious connections of gene regulatory and metabolic domain—their encoding genes are highly regulated while the porins enable numerous metabolic transport reactions . By focusing on the connecting domain of gene regulation and metabolism , the two centrality measures reinforce the key role of further cross-systemic elements . Considering the protein interface-induced subgraph both centralities point out the vertices that top the list of the above-discussed downwards traversing paths ( Table J in S1 Text ) . In more detail , both major systems contributing to the downwards traversing paths are represented each by three vertices , namely , PTS and RNR system ( Fig 7 , panels A and B ) . Having a look at the intra-domain degree fraction , which put the focus on protein interface vertices as described above , additionally highlights a representative of the upwards traversing path system NtrBC ( Fig 7 , panel C ) , as the second non-hub ( Table K in S1 Text ) . This corroborates the predictions from the traversing paths and , thus , shows that our new topological measure reveals cross-systemic elements which otherwise only stand out under detailed scrutiny of a large amount of biological information .
With their balance of structural detail and functional simplicity , network models are capable of revealing organizational principles , which are hard to recognize on a smaller systemic scale ( e . g . , by analyzing individual pathways ) or in functionally richer system representations ( e . g . , in dynamical models ) . One purpose of the network provided here is to enable work at the interface of statistical physics and systems biology , where the rich toolbox of complex network analysis is employed to identify functionally relevant non-random features of such biological networks . The recent work of [69] , for example , showed that network structure can reveal , whether an enzyme is susceptible rather to genetic knockdown or pharmacologic inhibition . While in the present study , the network measures do not distinguish between different kinds of vertices or links , the rich biological meta data concerning the different biological roles of the components could be translated into distinct vertex and edge classes . In our own investigation [29] we used this fact to study , in a further example of such an interdisciplinary effort , the balance of robustness and sensitivity in the interdependent network of gene regulation and metabolism , based on the reconstructed network provided here . In general , we expect that our network reconstruction can serve as a relevant data resource for the application of methods from the analysis of multiplex [70] and other multilayer networks [20 , 71] . Recently , there has been a growing interest in the properties of these systems , especially in the presence of explicit interdependencies between vertices [70 , 72] . In contrast to monoplex networks interdependent networks can show a qualitatively different robustness against failures , i . e . , cascading failures leading to a sudden system breakdown at a critical initial attack size [73 , 74] . The case of different vertex types ( as opposed to different edge types ) has been considered , for example , in the context of secure communication in a network where eavesdroppers control sets of vertices [75] . On a general level , analyzing statistics of paths with respect to the network’s large-scale structure , like the domain-traversing paths used here , might prove useful for the evaluation of other networks that show ( possibly more than one ) interface-like features . In summary , the analysis of network topology allows to determine key system components in the integrative E . coli network . In line with expectations , trivial compounds as well as currency metabolites showed up regardless of the measure that has been applied . In addition , further obvious components including several global regulators were identified . More striking is the detection of components and systems which solely emerge when analyzing specifically the interface . These hidden elements are associated to two of the biologically well-investigated functional subsystems , PTS and NtrBC . Both well-established and newly designed measures of the interface point out the same subsystems , and even the analysis of the entire network discloses components indirectly related to these hidden subsystems . Apart from trivial and currency metabolites , every detected key element of the entire network contributes to some extent to the downwards and/or upwards interface . This unlooked-for cross-systemic property is reflected either in the complex composition , the intra-domain degree fraction , the proximity to key systems , and/or the interplay with regulatory and metabolic processes . The biological relevance of these components supports their detection and reinforces the predictive power of the novel traversing path measure . In general , we believe that the presented integrative E . coli network allows further investigations of the interplay of metabolism and gene regulation which will provide insights into cellular , system-wide responses .
First , relevant information of the database has been extracted and arranged ( Algorithm 1 in Fig 9 ) . For each regulatory process , the respective source and target were specified and converted to match one of the vertex types ( ‘regulation . dat’ , file name of the EcoCyc-archive ) . To this end , the transcript units were separated into promoter , genes and terminator ( if applicable ) , and the regulatory processes were multiplied per comprising gene . Moreover , each regulating RNA has been translated into its encoding gene to meet the vertex types . In case of the metabolic processes , the reaction educts and products as well as the catalyzing enzymes have been assembled and converted to match one of the vertex groups , the respective educt and product stoichiometry have been assigned and the reaction compartmentation and reversibility have been assessed ( ‘reactions . dat’ ) . Thereby , as cell compartments the periplasmic space , the inner membrane , and the cytosol have been taken into account and reversible reactions have been split up . Second , vertex candidates have been validated ( ‘reactions . dat’ , ‘compounds . dat’ , ‘proteins . dat’ , ‘genes . dat’ , ‘rnas . dat’ ) and divided into reaction , compound , protein monomer , protein-protein complex , protein-compound complex , protein-RNA complex , and gene . In doing so , generic terms such as DIPEPTIDES have been substituted ( ‘classes . dat’ ) and double annotations , e . g . , CPD-15709 and FRUCTOSE-6P have been decoded . Thereupon , the compositions and the encoding genes of the assembled proteins have been gathered and matched to the vertex groups and the respective logical operation and stoichiometry have been annotated ( ‘protcplxs . col’ ) . Based on the validated vertex lists , the regulatory and metabolic processes have been updated whereby each process was removed with at least one unidentified vertex resulting in the final edge lists . Fig C in S1 Text provides a flowchart of this algorithmic procedure . With the validated vertex and edge lists the graph has been assembled and its largest weakly connected component has been extracted . The three domain partition MD—PI—RD ( Table 1 and A in S1 Text ) as well as the two-domain partition are implemented as vertex properties affiliation and metabolic . The initial categorization of both partitions is based on the vertex type ‘reaction’ which is denoted as purely metabolic and interface-related if all educts and products are compounds and proteins , respectively . Mixed educt and product types demand further clarification later on . Similarly , non-ambiguous vertices of type ‘compound’ , ‘protein’ and ‘gene’ are affiliated based on the affiliation of their neighbor vertices . This means that , if the influential adjacent vertices have the same affiliation , the vertex will be assigned to the same or its assignment needs a detailed consideration . In this way , genes and proteins that are not involved in any regulatory process can be assigned to the metabolic domain ( MD ) . Subsequently , deferred vertex affiliations are resolved iteratively based on their neighbor vertices affiliation until no further vertex affiliations can be assigned . The final ambiguous vertices , in total 13 of 12868 , are assigned as interface vertices since they cannot be uniquely assigned to metabolic or regulatory domain . Algorithms 2A and 2B in Figs 10 and 11 show the detailed domain affiliation process of a vertex . Moreover , the mapping to the E . coli model of [11] has been annotated which integrates the metabolic network iJR904 published by [5] and the transcription regulatory events related to the encoding genes of the catalyzing enzymes . To this end , genes , proteins , metabolites as well as biochemical reactions of the metabolic model have been mapped to the EcoCyc database ( release 20 . 0 ) , in a first step automatically based on their identifier and the resulting dictionaries have been manually curated . As the EcoCyc database does not account for compartmentation of compounds and reactions as well as for exchange reactions , unique metabolites and internal reactions have been considered resulting in a coverage of more than 93% . By additionally disregarding internal transport reactions a coverage of 96 . 5% can be achieved ( Table 1 ) . Integrating the manually curated Covert dictionaries , each vertex has attributed ( 1 ) a unique identifier , according to the EcoCyc identifier but also indicating the compartment , ( 2 ) a unique type reference , ( 3 ) a unique assignment of the model components from [11] , if applicable , and ( 4 ) the affiliations of the two- and three-domain partition . Furthermore , vertices of types gene and reaction have ( 5a ) a name assigned , the Blattner ID and the EC number , if applicable . The remaining vertices have additionally ( 5b ) a compartment assigned , where cytosol ( c ) , extracellular space ( e ) , periplasmic space ( p ) , inner membrane ( i ) , outer membrane ( o ) and membrane in general ( m ) were taken into account . Similarly , each edge of the network has the attribute ( 1 ) type , specifying the connected vertices , and the corresponding ( 2 ) stoichiometry , where zero is assigned if not applicable or ambiguous . For edges depicting regulatory processes the stoichiometry actually denotes the mode of regulation , namely activation ( + , 1 ) inhibition ( − , −1 ) or combined ( 0 ) . These edges additionally have assigned ( 3 ) an identifier , according to the EcoCyc identifier and ( 4 ) a name , specifying the regulation type . All other edge types can be classified as either representing conjunct or disjunct links in the sense that all or solely one incoming link is required for functionality ( Table B in S1 Text ) . The fully annotated integrative reconstruction of E . coli’s metabolic and regulatory processes is provided as a graph representation in S1 File . The following measures have been used in the assessment of the graph partitioning scheme . Inter-module edge fraction c . Given the set of vertices with the domain label D , edges connecting these vertices to a vertex of a different label are considered external , while edges between vertices of the same label are internal . We call c D = # ( external edges of D ) # ( external + internal edges of D ) the inter-module edge fraction of domain D . Network modularity M . denotes the degree to which a given partition divides the network in highly connected groups , modules , which are comparably sparsely connected among each other . Therefore , the intra-module links are counted against the total degree of the module vertices ( Eq 1 ) , M = ∑ j = 1 N M ( L ( v M j , w M j ) L G - ( deg ( v M j ) 2 L G ) 2 ) ( 1 ) with NM − # of modules , LG − # of links of graph G , L ( v M , w N ) = ∑ v ∈ M ∑ w ∈ N link ( v , w ) , deg ( v M ) = ∑ v ∈ M k v . Here kv is the degree of vertex v and link ( v , w ) denotes an undirected edge between vertices v and w . Note that this formulation of modularity ( taken from [30] ) coincides with the definition from [76] . A traversing path connects the regulatory and the metabolic domains via the protein interface , specifically , a traversing path of length k is of the form [ ( u , v 1 ) , ( v 1 , v 2 ) , … ( v k - 1 , w ) ] ( 2 ) where the vertices u and w are from the regulatory and the metabolic domain ( and vice versa ) and the vertices vi are distinct and part of the protein interface . Starting from the set of edges directly at the intersection of two domains iteratively the vertex successors of the interface domain as well as the final , first successor in the third domain have been determined ( Algorithm 3 in Fig 12 ) . The key elements of the integrative E . coli network have been determined based on two graph properties . Degree Centrality DC . is a local centrality measure and denotes the total number of in- and out-going edges of a vertex , ( Eq 3 ) , D C ( v ) = k v = k v in + k v out . ( 3 ) Here , the vertices with a total degree greater than 50 are termed hubs ( see the degree distribution in Fig D in S1 Text ) . By additionally accounting for the domain boundaries , the intra-domain degree fraction ξ ( also termed embeddedness [44] ) has been defined as ratio of internal degree , within domain D , and total degree of a vertex , ( Eq 4 ) , ξ D ( v ) = k v int k v = 1 k v ∑ w ∈ D ( A v w + A w v ) ( 4 ) where A denotes the adjacency matrix of the graph . Betweenness Centrality BC . describes the impact on the flux through the network , under the assumption that the transfer follows the shortest paths . In particular , it quantifies the fraction of shortest paths between all pairs of vertices which involve the designated vertex ( Eq 5 ) , B C ( v ) = ∑ s ≠ v ≠ t ∈ V σ s t ( v ) σ s t ( 5 ) where σst is the number of all shortest-paths between the vertices s and t while σst ( v ) yields the number of these paths that run through v [41] .
|
Networks—the compact representation of systems in terms of nodes and links—are an efficient data structure for biological information . They also allow us to establish relationships between network structure and dynamical function and thus hold the potential of implementing a systems-level view on biological processes . Using the formal language of networks and careful manual curation , we unite information from a range of publicly available databases , in order to provide a metabolic-regulatory network model for the gut bacterium Escherichia coli , which allows us to provide novel topological characterizations of system components based on their positions in the entire network . From the network representation we derive a new partition of the system into three network domains , one predominantly associated with gene regulation , a second , which covers all metabolic processes and a third domain , containing protein interactions and serving as an interface between the two other domains . This has consequences for the topological prediction of the biological relevance of the system components . We discuss specific examples , where this new three-domain representation reveals the structural centrality of components with known high functional relevance . This integrated network can serve as a platform for understanding biological phenomena jointly mediated by gene regulation and metabolism and thus provide insight relevant for genetic and metabolic engineering .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"protein",
"metabolism",
"gene",
"regulation",
"metabolic",
"processes",
"regulatory",
"proteins",
"metabolic",
"networks",
"enzymology",
"dna-binding",
"proteins",
"network",
"analysis",
"enzyme",
"metabolism",
"transcription",
"factors",
"enzyme",
"chemistry",
"computer",
"and",
"information",
"sciences",
"transcriptional",
"control",
"proteins",
"enzyme",
"regulation",
"gene",
"expression",
"biochemistry",
"genetics",
"biology",
"and",
"life",
"sciences",
"metabolism"
] |
2019
|
A system-wide network reconstruction of gene regulation and metabolism in Escherichia coli
|
Progress in science depends on the effective exchange of ideas among scientists . New ideas can be assessed and criticized in a meaningful manner only if they are formulated precisely . This applies to simulation studies as well as to experiments and theories . But after more than 50 years of neuronal network simulations , we still lack a clear and common understanding of the role of computational models in neuroscience as well as established practices for describing network models in publications . This hinders the critical evaluation of network models as well as their re-use . We analyze here 14 research papers proposing neuronal network models of different complexity and find widely varying approaches to model descriptions , with regard to both the means of description and the ordering and placement of material . We further observe great variation in the graphical representation of networks and the notation used in equations . Based on our observations , we propose a good model description practice , composed of guidelines for the organization of publications , a checklist for model descriptions , templates for tables presenting model structure , and guidelines for diagrams of networks . The main purpose of this good practice is to trigger a debate about the communication of neuronal network models in a manner comprehensible to humans , as opposed to machine-readable model description languages . We believe that the good model description practice proposed here , together with a number of other recent initiatives on data- , model- , and software-sharing , may lead to a deeper and more fruitful exchange of ideas among computational neuroscientists in years to come . We further hope that work on standardized ways of describing—and thinking about—complex neuronal networks will lead the scientific community to a clearer understanding of high-level concepts in network dynamics , and will thus lead to deeper insights into the function of the brain .
Philosophers of science have yet to develop a robust definition and interpretation of models and simulations [39]–[42] . Most of that debate focuses on models in physics , but Peck [31] gives an interesting review of models and simulations in ecology , while Aumann [32] thoroughly discusses requirements of successful modeling of ecological systems; Wooley and Lin [43] give an overview of modeling and simulation in biology . The only comparable assessment of the role of models and simulations in computational neuroscience is part of a book chapter by Clark and Eliasmith [44] . A recent appraisal of the role of models in neuroscience [45]–[47] , based on a general reappraisal of the role of computational models by Humphreys [48] , has mostly focused on connectionist models . We shall not attempt to provide a general analysis of models and simulations in computational neuroscience here . Our aim is more practical: to promote standards for the description of neuronal network models in the literature , to further sharing of knowledge and facilitate critique . Thus , our focus is narrower yet than that of Eliasmith and Anderson [49] , Ch . 1 . 5 , who proposed a “Methodology” of neural engineering . For our purposes , we adopt a quite restricted working definition of a model: Several aspects of this definition deserve note: The model is first of all a mental model formed in the brain of a researcher . It is her hypothesis about the function of a part of the brain . Heinrich Hertz expressed this idea first in his textbook “Prinzipien der Mechanik” in 1894: Scientific progress depends critically on the ability of neuroscientists to communicate models , i . e . , hypotheses , among each other: When Anna presents her model to Bob and Charlie—will both build the same mental model in their minds as Anna ? Or will some nuances be lost , some aspects interpreted differently , some parts misunderstood ? Only a precise , unambiguous notation for models will allow Anna , Bob and Charlie to discuss their individual understandings of the model and thus to truly share models . Efficient communication dictates that scientists should use a common notation to describe their models , as it is demanding to thoroughly acquaint ourselves with any advanced notation . It is tempting to consider implementations of neuronal network models in a specific simulator software as a sufficient model description , as it is explicit , specific and describes structure and dynamics . We believe this to be a fallacy . Implementations come most often in the form of scripts or computer programs , which tend to be difficult to reverse engineer: It is simply not possible to infer the overall network structure from the bits and pieces of a large script . Secondly , most simulation scripts rely on properties hidden in a simulator , which may even change as a simulator evolves over time . Translating a given implementation first to a mental model and then to a second simulator software for independent testing , opens for errors in both translation steps . We believe that while scientific productivity benefits from sharing simulation code through repositories such as ModelDB [50] and standard languages such as NeuroML [51] , implementations do not fill the need for precise human-readable model descriptions in the scientific literature . Based on experiences in systems biology , Wimalaratne et al . [36] stress that it is crucial to identify biophysical concepts as logical abstractions in order to create meaningful and re-usable model implementations . It is also worth mentioning that the translation of a mathematical model into a computer program is lossy and irreversible . The translation is lossy due to the finite precisions of computers . For example , most real numbers cannot be represented on a computer . This is obviously problematic in the analysis of chaotic systems where small errors have a big influence on the state trajectories of the system . The translation is generally not reversible , because the commonly used programming languages are not accessible to formal analysis . It is generally not even possible to prove that a function , implemented in a common language such as C++ , is correct . In some cases , one may even have to add equations to models in the computer implementation to preserve stability and obtain results in agreement with experimental observation [42] , [52] . While mathematical model descriptions can be treated with formal methods , their computer implementations generally cannot . This means that if we want to validate the claims about a model , we must start from the description in the scientific publication . If we start from the model implementation of the authors , we can never refute that the model may be faulty or doing something entirely different than what was claimed in the publication . Taking a given implementation of a model or hypothesis and simply executing it again does not constitute independent testing , nor does it fulfill the criterion of falsifiability: the same program run twice should yield identical results .
A complete model description must cover at least the following three components: ( i ) The network architecture , i . e . , the composition of the network from areas , layers , or neuronal sub-populations . ( ii ) The network connectivity , describing how neurons are connected among each other in the network . In most cases , connectivity will be given as a set of rules for generating the connections . ( iii ) The neuron and synapse models used in the network model , usually given by differential equations for the membrane potential and synaptic currents or conductances , rules for spike generation and post-spike reset . Model descriptions should also contain information about ( iv ) the input ( stimuli ) applied to the model and ( v ) the data recorded from the model , just as papers in experimental neuroscience do , since a reproduction of the simulations would otherwise become impossible . Neuronal network models are usually described by a combination of five means: prose ( text ) , equations , figures , tables and pseudocode . We shall discuss these in turn . Prose is a powerful means of communicating ideas , intentions and reasons . It is flexible and , if used carefully , precise . Unfortunately , prose can easily—often unintentionally—become ambiguous . Previous knowledge and ideas in the mind of the reader will shape the reader's understanding of a textual description of a model and may lead to misunderstandings . Prose that strives to be strictly unambiguous and provide all required detail , on the other hand , will often be difficult to read . Mathematical notation ( equations ) is compact and unambiguous . Suitably chosen notation compresses complex relationships in concise expressions , which allow for further manipulation in our mind , as illustrated by the matrix exponentiation in the Introduction . The now common mathematical notation emerged alongside the great scientific achievements of Newton , Leibniz and others between the 17th and 19th century [53] , [54] . Unfortunately , not all mathematical notation is understood easily , and variations in notation , as is common in computational neuroscience ( cf . Table 2 ) , can present serious obstacles to effective communication . Figures communicate the architecture and connectivity of network models well , since vision is the dominating sense in most humans . Most readers will first scan the figures in a paper to get an overview of what the paper is about , using figure captions as a guide , and read the full text of the paper only later . Thus , figures and captions will shape the initial idea a reader forms about a neuronal network model , and the ideas thus established may be difficult to correct through textual description . Specifying complex networks precisely in figures can be difficult , and disciplines depending strongly on exact diagrams , such as mechanical and electrical engineering , have developed precise standards for such diagrams ( see , e . g . , [55] ) . Systems biologists have yet to arrive at a definite standard for depicting their models , but they at least have an open debate about graphical representations [56]–[59] . Tables are a useful means of organizing data , especially model parameters . Data presented in table form is far more accessible than data dispersed throughout a text , facilitating , e . g . , comparisons of parameter choices between different papers and proof-reading of simulation scripts against papers . Pseudocode is often used to present algorithms in concise , human readable form , without resorting to a specific programming language . It will be an efficient means of communication only if the pseudocode notation is sufficiently well established to be unambiguous . The placement of model descriptions within a scientific publication depends on the focus of the paper and the journal it is published in . Traditionally , model descriptions were either given in the body text of a paper , or in an appendix . It has now become common to give only brief model overviews in the paper itself , and to relegate detailed model descriptions to supplementary material published online , or even to place simulation code online in community repositories such as ModelDB .
Figure 1 summarizes the placement of the description of architecture , connectivity and neuron and synapse models , respectively , across all papers; for details , see Tables S1 , S2 and S3 in the Supporting files . All papers present at least an overview of the model they investigate in the main body of the paper . Details are frequently provided in supplementary material available online , especially in more recent papers; appendices are used to a lesser degree . Model descriptions in some papers are incomplete in the sense that the authors refer to other publications for details of neuronal dynamics in particular . Within the body text of the paper , model descriptions were placed in the “Methods” sections in 10 of the 14 papers surveyed , even though the neuronal network model is in itself a product of significant scientific analysis and synthesis [32] . As such , it would rather belong in the “Results” section of a paper . Whether the placement of the model description in the “Methods” section genuinely reflects the way in which authors perceive their models , or rather is a consequence of editorial policies shaped by “wet” neuroscience , is not clear at present . It is interesting to note in this context that papers in theoretical physics generally do not follow the strict “methods-results-discussion” pattern . We would like to point out two interesting aspects of the placement of model descriptions . First , the text of a paper manuscript , including the appendix , undergoes thorough peer review and copy editing , ensuring high standards in content and presentation . It is not , at present , clear whether all material published as supplementary material receives the same scrutiny in the review process; it is often not copy-edited to the same standards as the paper proper . Second , source code published in community repositories represents an implementation of a model , not the model itself [52] . It can thus serve only as a service to the community to facilitate code-reuse , but not to communicate the content of the model proper . Incidentally , none of the 14 papers surveyed here describes re-use of neuronal models available in repositories , such as ModelDB [50] . Nor does any paper mention that the source code for the model implemented in the paper was made available to the community , even though models from several papers are at present available from ModelDB [4] , [5] , [15] . In recent years , though , there appears to be a slowly growing trend to explicitly reference and re-use existing models from ModelDB; see http://senselab . med . yale . edu/modeldb/prm . asp for an up-to-date list ( Michael Hines , personal communication ) . Figure 2 shows that equations are mostly used to describe the dynamics of model neurons , while connections are most often presented in a combination of prose and figures , occasionally in form of pseudocode . We will review the quality of these descriptions in detail below . Table 3 shows how parameters are presented in papers . It regrettably indicates that too few authors make parameters easily accessible in tables . Network model descriptions in the literature show no consistent order of description . Among the papers surveyed here , six begin with a description of the neuron models and then proceed to network architecture , seven papers use the opposite order , while one paper mixes the description of neurons and network . We find the latter option least useful to the reader . Authors differ greatly in their efforts to anchor their models in empirical data . Destexhe et al . [4] go to great lengths to justify the design of their neuron and synapse models with respect to the neurophysiological literature . They thus provide the synthesis document proposed by Aumann [32] as the basis of any modeling effort . Unfortunately for those readers who want to investigate the resulting model , though , model description and justification are tightly intertwined in the terse methods section , making it quite demanding to extract the model description as such . Among all papers surveyed here , only Destexhe et al . [4] and Izhikevich and Edelman [8] show responses of individual synaptic conductances and individual neurons to test stimuli , while all other authors only show responses of the entire network . This means that researchers who attempt to re-implement a model and find themselves unable to reproduce the results from a paper , will not be able to find out whether problems arise from neuron model implementations or from a wrong network setup . We will now analyze in detail which difficulties arise in describing a network model , considering in turn network architecture , connectivity , and neuron models , and point out examples of good descriptions . Descriptions of network architecture become challenging as network complexity increases . Networks with a small number of populations , random connectivity and no spatial structure are easily described in a few lines of prose , as in Brunel's paper [3] . A combination of prose and simple figures is usually sufficient to describe architecture of networks composed from a small number of one- or two-dimensional layers of individual neurons; examples are Destexhe et al . [4] and Kirkland and Gerstein [9] . Complex models spanning several brain areas with detailed spatial , layered , and functional substructure , such as Lumer et al . [10] and Izhikevich and Edelman [8] , are more challenging to describe . Authors generally adopt a top-down approach , giving first an overview of the brain areas involved , before detailing the structure of the individual areas . In models of systems with clearly defined signal flow , areas are often visited in the predominant order of signal flow [6] , [7] , [14] , while others present the more complex cortical structures before descending to subcortical structures [8] , [10] . The most detailed explicit model studied here is the thalamocortical model presented by Lumer et al . [10] . The description of the cortical areas in this model ( Vp and Vs ) , while complete , lacks in our opinion the clarity desirable of a good model description , and may thus help to identify rules for ideal model descriptions . For one , discussions on model design and properties are embedded in the model description , e . g . , the reduction of a total of 32 “combinations of response selectivities” to just two included in the model , and a comparison of the number of neurons in the model to that found in animals . We believe that design decisions and model review should be kept separate from the model description proper for the sake of clarity , since they are independent intellectual endeavours [32] . Second , Lumer et al . mix different views of their layer architecture without providing sufficient guidance to the reader . They begin by describing the Vp layer as a grid of 8×8 macro-units , with two “selectivities within a macro-unit” , each containing “a collection of 5×5 topographic elements , each of which corresponded to a contiguous location in retinal space” , before proceeding to state that “[t]opographic elements in Vp were organized in maps of 40×40 elements for each of the two modeled orientation selectivities . ” We find it difficult to interpret this description unambiguously . We are in particular in doubt about the localization of macro-units and topographic elements in retinal space . In our view , the most parsimonious interpretation is as follows: 5×5 topographic elements placed in each of 8×8 macro-units result in a grid of 40×40 topographic elements . ” This interpretation is sketched in Fig . 3 . Another interesting aspect is that model composition is often described from a perspective orthogonal to the description of connections . Lumer et al . [10] , e . g . , present the primary thalamus and cortex as grids of 40×40 topographical units , each containing an excitatory and an inhibitory neuron ( thalamus ) and a microcolumn composed of 10 neurons organized in three laminae ( cortex ) . Connections are then described by looking at this architecture from an orthogonal perspective: Thalamus is described as two layers , one of excitatory and one of inhibitory neurons , while cortex is split into six layers , one of excitatory and one of inhibitory neurons for each of the three laminae in the model . We believe that it may be more sensible to base the model description on the perspective used in defining connections , as connectivity is the central aspect of a network model . Izhikevich and Edelman [8] present a significantly more complex model , covering the entire human cortex and thalamus . Concerning the spatial placement , they only state that “[n]euronal bodies are allocated randomly on the cortical surface , whose coordinates were obtained from anatomical MRI . ” No further information is given on how MRI measurements were converted to neuron densities in space . Thus , even if one had access to MRI data of the human brain , it would be difficult to reproduce the neuron distribution investigated by Izhikevich and Edelman . In such cases it would be advantageous to either use datasets available from community databases or to make data available to others . Figures of network architecture vary widely between papers . We will discuss them in the following section together with connections . Describing the connections well is the most challenging task in presenting a neuronal network model . For networks with random connections and no spatial structure , connectivity is easily described in a few sentences [3] . Haeusler and Maass [5] additionally represent connection strengths and probabilities in a figure; this works well for their six-population model . If yet more populations were involved , such a figure would soon become cluttered , and it becomes more useful to present connection parameters in tables , cf . supplementary material in ref . [8] . Even in these simple networks , care must be taken to specify details: Few authors are explicit on all these points , although these choices may have significant consequences for network dynamics ( Tom Tetzlaff , personal communication; see also Kriener et al . [60] ) . Models incorporating spatial structure have more complex connection patterns , which we will call topographic connections , since they usually describe the spatial distribution of connection targets relative to the spatial location of the sending neuron , i . e . , connections are typically described as divergent connections . In most cases , connections have a random component: they are created with a certain probability . In simple cases , such as Kirkland and Gerstein [9] , connections are made to neurons in a rectangular mask with equal probability . In more complex models , connection probability depends on the relative locations of the neurons that are candidates for a connection , e . g . , [10] , [11] . Unfortunately , few authors provide the equations for these probability functions; Mariño et al . [11] is a laudable exception . It is somewhat paradoxical if papers present long tables of parameters for these connection probability functions , but do not provide the equation into which these parameters enter . Mariño et al . [11] are the only authors who explicitly discuss self-connections ( in their supplementary material ) , and as far as we can see , no authors have discussed whether multiple connections between any two neurons may be created . Another neglected issue is precisely how probabilistic connections are created . The following approach seems to be implied: For each pair of neurons from the sender and target population , a connection is created if a random number is smaller than the connection probability for the pair . But one might equally well determine the total number of connections to be made first , and then distribute the connections according to the spatial probability profile [61] . Such schemes offer significant performance gains [62] . A complete specification of the connection algorithm should thus be given . Among the papers surveyed , Izhikevich and Edelman [8] has by far the most complex connectivity and the authors go to great lengths to present gray-matter connectivity in figures , tables , and prose . Alas , some information appears to be missing: It is not clear from the text exactly how connections are distributed within the axonal spans , and how they are distributed across dendritic compartments of neurons with more than one compartment in a cortical layer . We have also been unable to find specific information on how synaptic weights and delays were assigned to connections . Finally , no details are provided about the white-matter ( long-range ) connections , which were based on diffusion-tensor imaging ( DTI ) data . Without access to the DTI data it is thus impossible to re-implement the model presented . Paper authors draw network diagrams in quite different ways , both in the overall style of their diagrams and in use of symbols . Figure 4 shows network diagrams of a model loosely based on Einevoll and Plesser [63] , Fig . 3 , drawn in the style of three of the papers surveyed here . The diagram in the style of Hayot and Tranchina [6] ( Fig . 4A ) gives a reduced but clear overview of the overall architecture of the model; it provides no details . The style of Haeusler and Maass [5] ( Fig . 4B ) carries most information , with weights and probabilities shown next to connection lines , and line widths proportional to the product of weight and probability . Figure 4C , which imitates the style of Lumer et al . [10] , is rather illustrative: it provides no quantitative information and the structure of the connectivity is less prominent than in the other two figures; on the other hand , it is the only figure hinting at the spatial structure of the network . Interestingly , all three diagram styles use different ways of marking excitatory and inhibitory connections: bars vs circles , black vs red , and arrows vs bars . Indeed , bars at the end of connection lines mark excitatory connections in Hayot and Tranchina's style , but inhibitory connections in the style of Lumer et al , nicely illustrating the lack of standards in the field . Izhikevich and Edelman [8] have illustrated their brain model using diagrams presenting significantly more detail than in the diagrams shown in our Fig . 4 . Unfortunately , we cannot reproduce Figures 2 and 8 from the supplementary material of the paper by Izhikevich and Edelman here due to copyright issues; the figures are available on the internet at http://www . pnas . org/content/105/9/3593 . figures-only and http://www . pnas . org/content/105/9/3593/suppl/DC1 , respectively . Their diagrams , though , provide so much detail of interest to the re-implementer , that the reader will have difficulty to form a clear conceptual model from the diagram . This is in many ways the curse of complex models as the following analogy may illustrate: when a physicist or electrical engineer sees a diagram of an RLC circuit , she will intuitively “see” the circuit oscillate . When presented with the complete wiring diagram for a modern analog radio receiver , though , it is hardly likely she will “hear the music” . The figure in the style of Haeusler and Maass [5] takes a middle ground . Since the individual populations are homogeneous , they can be represented by one circle each , with annotated lines providing information about connection structure and parameters . By marking connection strength through line width and differentiating excitation and inhibition by line color , the figure appeals quite directly to our intuition . It is clear , though , that any further populations would increase the complexity of the diagram to the point of illegibility . There is no established standard for the order in which connections within a network are described . Some authors proceed from local connectivity ( e . g . , intracortical intralaminar ) towards global connectivity [10] . Others rather follow the signal flow through the network , from retina via LGN to cortex , e . g . , Kirkland and Gerstein [9] , Hayot and Tranchina [6] , and Troyer et al . [14] . Neuron and synapse models are commonly described by a mixture of prose and equations , cf . Fig . 2; tables are used inconsistently to present parameters , see Table 3 . Some authors do not provide complete model specifications in their paper , but rely heavily [4] or even entirely [5] on references to earlier work . While the desire to avoid repetition is understandable , we believe that authors here walk a thin line toward incomprehensibility , especially if the models used are spread over three or more publications . Even though the re-use of neuron model implementations provided in repositories such as ModelDB may save effort and contribute to a standardization in the field , none of the papers we studied made use of available model implementations—or the authors failed to point out that they did . Table 2 shows the membrane potential equations found in several papers and demonstrates that there is a reasonable amount of variation in the way this central equation is written down . There is in particular no widespread agreement on whether to include the membrane capacitance explicitly in the equation or rather to subsume it in a membrane time constant . Some authors , such as Tao et al . [13] , even chose to normalize the membrane potential equation by defining . Yet greater variation is found in the representation of synaptic currents . This means that phrases such as “we use the standard equations for integrate-and-fire neurons” , which are not uncommon in the literature , are essentially meaningless , since there are no established “standard equations” for integrate-and-fire neurons . Spike generation and detection , including subsequent reset and refractory behavior , are usually described in prose , sometimes with interspersed equations . “ was reset to … , when it exceeded a threshold of … −51 mV … , at which point a spike was recorded , and relayed … , ” is a typical formulation [10] . Unfortunately , it does not state precisely how threshold crossings are detected , which times are assigned to spikes , or when exactly the reset is executed . All these issues can have significant consequences for network dynamics [64]–[66] . The previous sections have documented a wide variety of approaches to model descriptions in the literature . We believe that this variety is detrimental to the field , as it makes it difficult to communicate neuronal network models correctly and efficiently . At the same time , we believe that the field of computational neuroscience is too young to establish exacting standards for model descriptions . We will return to this problem and its various causes in the discussion . As a middle road , we propose to establish a good model description practice for the scientific literature . We will refer to it as “good practice” below for brevity . Some of our suggestions are motivated by a recent analysis of modeling techniques in ecology [32] , but see also [49] . We propose a practice with the following elements: We will discuss these elements in turn below , followed by more detailed discussions about how to render specific aspects of a network model . As an illustrative example , Figures 5 and 6 provide a concise description of the Brunel [3] model following the good practice format . A similar description of the Lumer et al . [10] model is given in Figures 7–9 . We would like to stress that we present the good practice here to stimulate the debate on model descriptions within the computational neuroscience community . If it is adopted widely throughout the community , it will provide numerous advantages: authors will have guidelines that will allow them to check their descriptions for completeness and unambiguousness; referees will more easily be able to assess the correctness and quality of a model; and readers will find it easier to comprehend and re-implement a model , and to compare different models .
The model survey presented here revealed a wide variety of approaches to describing the composition and connectivity of neuronal networks . We believe that this is , at least in part , due to a lack of common high-level concepts for composition and connectivity from a modeling perspective . Developing such high-level concepts describing , e . g . , certain types of randomized connectivity patterns , is thus an important task for the computational neuroscience community . The challenge at hand is perhaps best clarified when trying to draw diagrams representing neuronal network models . Such diagrams have two aims: To give the reader an intuitive understanding of model properties central to the dynamics of the model , and to unambiguously provide the necessary detail to allow a reconstruction of a model . In the absence of a mathematical formalism for model specification , diagrams often seem better suited than prose to present unambiguous detail . Simple models , such as that by Brunel [3] , can be depicted in a single diagram , as illustrated in Fig . 6 . The four panels in that figure , though , show that one may choose from a wide variety of styles for such diagrams , and it is not a priori clear which style is best . In panels A–C in the figure we propose three ways to differentiate between excitatory and inhibitory connections ( line styles and endings ) as well as to mark connectivity patterns ( line endings , styles , annotations ) . Panel D differs from the other three in the way the external input is represented . Brunel [3] states that “[each neuron] receives connections from excitatory neurons outside the network . External synapses are activated by independent Poisson processes with rate . ” This is rendered in detail in panel D , which shows Poisson generators per modeled neuron . In all other panels , these generators have been collapsed into an external excitatory population with the implicit assumption that this population contains the correct number of Poisson generators required by the model . Presenting complex models is even more challenging . In Fig . 9 , we present a set of three figures describing the model by Lumer et al . [10] at three levels of hierarchy: an overall view in panel A , details of the connectivity within the cortical populations tuned to vertical stimuli in panel B , and finally details of projection patterns into a single cortical population in panel C . All figures are simplifications of the full model , since we have left out the secondary thalamic and cortical areas . We are currently pursuing research to identify drawing styles and a hierarchy of diagrams that will be intuitive to a majority of computational neuroscientists and provide the necessary detail . Results will be presented elsewhere . Given the importance of comprehensible and precise model descriptions , it may seem surprising that no standards or good practices have emerged in computational neuroscience to date . Early proposals , such as the Neural Simulation Language [73] ( see also Eliasmith and Anderson [49] , Ch . 1 . 5 and Kumar [74] ) , have not been accepted widely in the community . At present , two developments appear promising . NetworkML , which is part of the NeuroML project [28] , [51] , provides a simulator-agnostic XML-based declarative standard for neuron network model descriptions . Simulation code for tools such as Neuron and Genesis can be generated from models defined in NeuroML . PyNN [29] , in contrast , is an imperative scripting language that can control a number of common neuronal network simulators , such as NEST , Neuron , and Brian . One reason why neither NetworkML nor PyNN has yet caught on as a means of widespread model exchange may be that neither of the two languages seems to aim at providing human-comprehensible model descriptions that might be included in publications . Another reason for the lack of model description standards may be that computational neuroscience has to a large degree been an ancillary science , an appendix of electrophysiology: The vast majority of publications in computational neuroscience compares its modeling results directly to specific sets of experimental data . And even though models have driven the development in some fields of neuroscience [75] , very few authors have compared the properties of different models with each other; Erwin et al . [76] is a notable exception . De Schutter [77] even argues that there currently is a trend away from the investigation of models as such , and back to a one-to-one matching of models to experiments . As long as computational neuroscientists focus on matching their models to specific experiments , rather than either to spar their models against each other , or build their models upon each other , the motivation to use a standard notation is obviously limited . We have no doubt that model sharing will increase in computational neuroscience in years to come . This raises the question of what model sharing precisely entails . At the simplest level , models may be shared as simulator code . While this seems convenient at first , it carries significant risk , as any code is likely to contain errors , in particular errors that may surface only once an existing model is used in a different context than the one in which it was originally developed . Indeed , in at least one case , high-profile publications ( outside neuroscience ) , had to be retracted after a subtle programming error was discovered in a widely shared scientific software [78] . Some scientists argue that everyone in a field should use the same , carefully maintained simulation software to avoid such problems , and to make computational science reliable [79] . We beg to differ: monoculture tends to create more problems than it solves . Establishing a new publication culture in computational neuroscience will require considerable effort within the community . We hope that the good model description practice that we have outlined in the previous section may be a good starting point . We believe in particular that a clear segregation of model derivation , model description , implementation , and model analysis , as proposed above , will make it easier for readers to discern the model as such , compare it to other models , and evaluate its relevance to their own research . The proposed Checklists for model descriptions will help to ensure that model descriptions themselves are reasonably complete and follow a common pattern , further improving the communication of models , while the Templates for tables invite a standardized presentation of details on various aspects of models; similarly , the Guidelines for diagrams should aid authors in illustrating their network models . Since all our proposals are informal , we hope that authors will find it straightforward to apply them when describing their network models , thus establishing a de facto standard for model descriptions . We are optimistic that we are beginning to see changes towards more cooperation within computational neuroscience , as witnessed by several collaborative reports on neuronal network simulations in the last two years [25]–[27] and the development of tools for the integration of various simulation software [29] , [30] , much helped by the establishment of the International Neuroinformatics Coordinating Facility ( INCF ) in 2005 . The Connection Set Algebra proposed by Djurfeldt [72] is an encouraging step towards establishing high-level concepts for neuronal network descriptions , i . e . , giving us a concise language to talk about our models . There is also much to be learned from model sharing and curation efforts in other communities , such as the IUPS Physiome and the European Virtual physiological human projects [37] , [80] . In closing , let us return to the power of notation , as exemplified by the matrix notation in the introduction . In July 1924 , Werner Heisenberg gave a manuscript full of complicated mathematics to his mentor Max Born , unsure whether it was worth publishing . Born worked through Heisenberg's ideas and realized that what Heisenberg had written down , actually amounted to the matrix mechanics of quantum theory . This insight of Born's unleashed the full power of Heisenberg's ideas and let Born discover the non-commutativity of quantum mechanics [81] , p . 125f . We are looking forward to the day when a good formalism will give us deeper insights into the secrets of signal processing in the brain .
|
Scientists make precise , testable statements about their observations and models of nature . Other scientists can then evaluate these statements and attempt to reproduce or extend them . Results that cannot be reproduced will be duly criticized to arrive at better interpretations of experimental results or better models . Over time , this discourse develops our joint scientific knowledge . A crucial condition for this process is that scientists can describe their own models in a manner that is precise and comprehensible to others . We analyze in this paper how well models of neuronal networks are described in the scientific literature and conclude that the wide variety of manners in which network models are described makes it difficult to communicate models successfully . We propose a good model description practice to improve the communication of neuronal network models .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"neuroscience/theoretical",
"neuroscience",
"computational",
"biology/computational",
"neuroscience"
] |
2009
|
Towards Reproducible Descriptions of Neuronal Network Models
|
Secondary bacterial infections from snakebites contribute to the high complication rates that can lead to permanent function loss and disabilities . Although common in endemic areas , routine empirical prophylactic use of antibiotics aiming to prevent secondary infection lacks a clearly defined policy . The aim of this work was to estimate the efficacy of amoxicillin clavulanate for reducing the secondary infection incidence in patients bitten by Bothrops snakes , and , secondarily , identify risk factors for secondary infections from snakebites in the Western Brazilian Amazon . This was an open-label , two-arm individually randomized superiority trial to prevent secondary infection from Bothrops snakebites . The antibiotic chosen for this clinical trial was oral amoxicillin clavulanate per seven days compared to no intervention . A total of 345 patients were assessed for eligibility in the study period . From this total , 187 accomplished the inclusion criteria and were randomized , 93 in the interventional group and 94 in the untreated control group . All randomized participants completed the 7 days follow-up period . Enzyme immunoassay confirmed Bothrops envenoming diagnosis in all participants . Primary outcome was defined as secondary infection ( abscess and/or cellulitis ) until day 7 after admission . Secondary infection incidence until 7 days after admission was 35 . 5% in the intervention group and 44 . 1% in the control group [RR = 0 . 80 ( 95%CI = 0 . 56 to 1 . 15; p = 0 . 235 ) ] . Survival analysis demonstrated that the time from patient admission to the onset of secondary infection was not different between amoxicillin clavulanate treated and control group ( Log-rank = 2 . 23; p = 0 . 789 ) . Secondary infections incidence in 7 days of follow-up was independently associated to fibrinogen >400 mg/dL [AOR = 4 . 78 ( 95%CI = 2 . 17 to 10 . 55; p<0 . 001 ) ] , alanine transaminase >44 IU/L [AOR = 2 . 52 ( 95%CI = 1 . 06 to 5 . 98; p = 0 . 037 ) ] , C-reactive protein >6 . 5 mg/L [AOR = 2 . 98 ( 95%CI = 1 . 40 to 6 . 35; p = 0 . 005 ) ] , moderate pain [AOR = 24 . 30 ( 95%CI = 4 . 69 to 125 . 84; p<0 . 001 ) ] and moderate snakebites [AOR = 2 . 43 ( 95%CI = 1 . 07 to 5 . 50; p = 0 . 034 ) ] . Preemptive amoxicillin clavulanate was not effective for preventing secondary infections from Bothrops snakebites . Laboratorial markers , such as high fibrinogen , alanine transaminase and C-reactive protein levels , and severity clinical grading of snakebites , may help to accurately diagnose secondary infections . Brazilian Clinical Trials Registry ( ReBec ) : RBR-3h33wy; UTN Number: U1111-1169-1005 .
Bothrops snakebites result in a subcutaneous and muscular lesion at the site of bite , which many times evolve to local complications [1–5] . In 2015 , the Brazilian Ministry of Health recorded 18 , 741 snakebites across the country [6] . In this country , a higher incidence is observed in the Brazilian region ( 37 cases/100 . 000 inhabitants ) . These values could be higher in remote areas of the Amazon because of the considerable case underreporting [7] . Snakebites are recorded mostly from remote rural or riverine areas , from where patients’ rescue to the health units is made exclusively by boat transportation , lasting several hours or even days [8–10] . Thus , late medical assistance is common and probably contributes to the high complication rates related to local necrosis [10–12] and secondary bacterial infections [10–15] , that can lead to permanent function loss and disabilities [3 , 10 , 16–19] . It was suggested that secondary bacterial infections from snakebites are related to the oral and fangs microbiota of the perpetrating snake [20–26] . Besides , traditional treatment also contributes for the emergence of secondary infections , such as tourniquet use , local alternative medicines , incision and suction of the bite site [8 , 18 , 27] . These factors increase the development of expressive forms of secondary infection , which has been primarily diagnosed with identification of cellulitis or abscesses [12 , 13 , 22] . In the Brazilian Amazon , Bothrops atrox is the most important venomous snake , causing 80–90% of the snake envenomings [28] , Despite the wide geographic distribution in the Amazon , B . atrox venoms share the same family of toxins , as PIII and PI snake venom metalloproteinase , phospholipase A2 , serine proteinase , cysteine-rich secretory protein , L-amino acid oxidase and C-type lectin-like [29 , 30] and are characterized by coagulant , hemorrhagic and proteolytic or acute inflammatory activities [31–33] . Spontaneous systemic bleeding and acute renal failure are common systemic complications from Bothrops envenomings [13 , 34] . Local envenomation range from a painless reddened injury to intense pain and swelling at the site of bite , starting minutes after the event . Enlargement of the regional lymph nodes draining the site of bite and bruising can also be observed some hours after bite , especially if patient delayed in reaching a health service [33 , 34][34][33][33][33][33] . In the first 24 hours , blistering and tissue necrosis may be evident . Cellulitis or abscess occurs mostly in the moderate or severe cases , generally as a polymicrobial infection . Gram-negative bacteria have been implicated in secondary bacterial infection , which frequency may vary according to region [35] . In Manaus , secondary bacterial infections were observed in around 40% of the Bothrops snakebites [13] . Several antimicrobial schemes were suggested for the treatment of secondary infections , but in general these recommendations were not based on good evidences from clinical trials [36 , 37] . For example , ampicillin/cephalosporin/cloxacillin [14 , 27] , ciprofloxacin [15 , 22] and clindamycin [36 , 38 , 39] were previously used for secondary bacterial infections resulted from snakebites , with variable effectiveness . In the Amazon , basic information about the bacterial agents responsible by the wound infection is still lacking since secondary infection diagnosis is mostly based only from clinical features without microbiological confirmation . Although common in endemic areas , routine empirical prophylactic use of antibiotics aiming to prevent secondary infection lacks a clearly defined policy , leading to wasteful inappropriate antibiotic use , which is costly and may promote bacterial antibiotic resistance [22 , 40] . Preemptive treatment efficacy of oral chloramphenicol monotherapy in Bothrops snakebites [41] and intravenous chloramphenicol plus gentamicin in Crotalus snakebites [42] showed no statistical difference between patients treated and untreated groups . Nowadays , however , the Infectious Diseases Society of America ( IDSA ) guidelines for diagnosis and management of skin and soft-tissue infections indicate amoxicillin clavulanate to reduce complications by prevention of secondary infection from animal bites [38 , 39] . Evidence supporting this recommendation came from a clinical trial carried out with patients bitten by dogs [43] , but efficacy of this regimen is still not available in snakebites . The aim of this work was to estimate the efficacy of amoxicillin clavulanate for reducing the secondary infection incidence in patients bitten by Bothrops snakes , and , secondarily , identify associated factors for secondary infections from snakebites in the Western Brazilian Amazon .
Ethical approval was obtained from the Fundação de Medicina Tropical Doutor Heitor Vieira Dourado ( FMT-HVD ) ( approval number 492 . 892/2013 ) . Written informed consent was obtained from all participants prior to randomization . This study was registered in the Brazilian Clinical Trials Registry ( ReBec ) : RBR-3h33wy and UTN Number: U1111-1169-1005 . This was an open-label , two-arm individually randomized superiority trial to estimate the efficacy of the preemptive amoxicillin clavulanate administration compared to no intervention for preventing secondary infection from Bothrops snakebites . Clinical trial was performed at the Fundação de Medicina Tropical Doutor Heitor Vieira Dourado ( FMT-HVD ) , in Manaus , Western Brazilian Amazon , from August 2014 to September 2016 . This tertiary hospital is the reference in the Amazonas state for snakebites treatment . In Manaus , FMT-HVD is the only hospital unit that performs the distribution and administration of snakebite antivenom . At admission , Bothrops snakebites were diagnosed with basis in clinico-epidemiological characteristics of the patient and , when the patient brought the snake responsible by the envenomation , by its identification made by a trained biologist . Sample size calculation was based on the mean of 240 snakebites/year attended at FMT-HVD , with an expected frequency of secondary infection of 40% [13] , a 50% risk reduction of infection , at an 80% power and 5% of significance level and an 1:1 randomization ratio . Adding 10% of losses in the follow-up , a sample size of 186 participants was obtained , with 93 patients in the intervention group and 93 in the untreated group . Patient was eligible if admitted to the hospital with less than 24 hours after the bite , without antivenom therapy in other hospital and without any sign of secondary infection at this time . Patients that used any antibiotic in the past 30 days , pregnant women or patients with previous history of allergic reactions to antibiotics were not included in this trial . After application of eligibility criteria , the study pharmacist was contacted to obtain the allocation group to the patient . Randomization sequences with an allocation ratio of 1:1 were computer-generated by a random table , to the intervention group ( preemptive amoxicillin clavulanate ) or to the control group ( no preemptive antibiotic prescription ) . All laboratory staff was blinded for treatment assignment . The antibiotic chosen for this trial was oral tablet amoxicillin clavulanate 875/125 mg to adults and 25 mg/kg/day to children twice per day for seven days , starting at the admission day . After patient inclusion , demographic and epidemiological information was collected using a standardized questionnaire , including gender , age ( in years ) , area of occurrence ( urban or rural ) , anatomical site of the bite , work-related bite ( yes or no ) , time elapsed from bite to medical assistance ( in hours ) , walking after bite ( in minutes ) , previous history of snakebite and pre-admission conduits ( use of topical or oral medicines , use of tourniquet and other procedures ) . A detailed clinical and laboratorial characterization was also made at this time . Pain assessment was made using the Numerical Rating Scale , with values rating from 1 to 10 [44]; pain was further classified as absent ( rate 0 ) , mild ( rated from 1 to 3 ) , moderate ( rated from 4 to 7 ) and severe ( rated from 8 to 10 ) . Edema was classified as absent , mild ( affecting 1–2 limb segments ) , moderate ( affecting 3–4 limb segments ) and severe ( affecting more than 5 limb segments ) [11] . Bite site temperature ( oC ) was measured using an infrared digital thermometer ( Color Check AC322 ) ; the difference between the bite site temperature and the contralateral limb site was calculated . Presence of local bleeding , lymphadenitis and necrosis was also assessed . Systemic signs and symptoms , such as systemic bleeding , signs of acute renal failure , headache , dizziness and vomiting were recorded from patients . Vital signs ( blood pressure , heart rate , respiratory rate and axillary temperature ) were also assessed . All the clinical information was collected through a standardized clinical registration form . Immediately after clinical examination , a 15 mL blood sample was taken for laboratorial analysis . Tests included leukocyte count ( cells/μL ) , fibrinogen ( mg/dL ) , platelet count ( number/μL ) , hemoglobin ( mg/dL ) , creatine phosphokinase ( IU/L ) , creatine phosphokinase-MB ( ng/mL ) , erythrocyte sedimentation rate ( mm/hour ) , lactate dehydrogenase ( IU/L ) , creatinine ( mg/dL ) , urea ( mg/dL ) , aspartate transaminase ( IU/L ) , alanine transaminase ( IU/L ) , clotting time ( in minutes ) , prothrombin time ( in seconds ) and C-reactive protein ( mg/dL ) . An aliquot was submitted to an enzyme immunoassay to confirm Bothrops envenoming diagnosis and to determine circulating venom levels in all patients [45] . All the laboratory results were transferred to a standardized registration form . According to clinical severity , patients were classified using to the Brazilian Health Ministry guidelines [7]: i ) mild cases: local pain , local swelling and bruising for Bothrops; ii ) moderate cases: local manifestations without necrosis and minor systemic signs ( coagulopathy and bleeding , no shock ) ; iii ) severe cases: life- threatening snakebite , with severe bleeding , hypotension/shock and/or acute renal failure . Both groups were submitted to the same local wound care with daily 0 . 9% saline cleaning . Thirty minutes before antivenom therapy , the intervention group of patients took amoxicillin clavulanate , supervised by a nurse . Twenty minutes after pre-medication with IV hydrocortisone ( 500 mg ) , IV cimetidine ( 300 mg ) and oral dexchlorpheniramine ( 5 mg ) , antivenom therapy was given to all patients from both arms in a dosage corresponding to the severity grading , according the Brazilian official guidelines [7] . Intervention and control groups were hospitalized for 3 days and returned to the hospital 7 days after admission . A full clinical and laboratory examination were performed 24 hours , 48 hours , 72 hours and 7 days after admission . The primary efficacy endpoint of this trial was the time free of secondary infection at snakebite site , defined as the presence of cellulitis and/or abscess [38 , 39] , until 7 days after hospital admission . The onset of secondary infection until 48 hours at admission was considered a secondary outcome . Cellulitis was defined by the presence of local inflammation signs ( erythema , edema , bruising and pain ) with association to fever , leukocytosis , lymphangitis and/or lymphadenitis [46] . An abscess was characterized by individual injuries , floating , presenting purulent secretion or serous-purulent secretion [46 , 47] . Two independent observers evaluated all the patients and came to a final agreement . Patients clinically diagnosed with secondary infection were additionally evaluated by ultrasonography and a sample collection for microbiology was obtained only in cases evolving to abscess . After secondary infection diagnosis , amoxicillin clavulanate was interrupted and patient was treated accordingly medical discretion . Before statistical analysis , two independent typists entered information using Epi Info 3 . 5 . 1 . A study researcher solved disagreements . The primary efficacy analysis was done on all randomized participants finishing the follow-up ( per protocol population ) . The primary efficacy endpoint , secondary infection-free at 7 days , was analyzed using Kaplan-Meier estimates . A two-sided log-rank test was done over the period using a 5% significance level . Patients that were secondary infection-free at day 7 were censored at this point . The effect of drug use ( amoxicillin clavulanate ) was assessed as a single block . Relative risk , relative risk reduction , absolute risk reduction and number needed to treat were assessed for primary and secondary outcomes . For the secondary infection risk analysis , at 48 hours and 7 days of follow-up , explanatory variables were grouped in hierarchical blocks [48] . Proximal block was composed by laboratory findings at admission , intermediate block by clinical findings at admission and distal block by demographic and epidemiological variables . Univariate regression analysis was carried out for each block individually . Variables with a significance level of p<0 . 2 were included in the multivariate analysis by block . All variables with a significance level of p<0 . 05 in the multivariate analysis by block were thus included in the overall model ( all blocks together ) . Crude odds ratios ( OR ) , adjusted odds ratios ( AOR ) with their respective confidence intervals were calculated for each hierarchical level and for the overall model . Accuracy of the final model was evaluated by Hosmer-Lemeshow goodness-of-fit test . Where zeros caused problems with computation of the OR and 95% CI , 0 . 5 was added to all cells [49 , 50] . Mann-Whitney tests were carried out to assess differences between median of treated and untreated groups . Statistical analyses were performed using the STATA statistical package version 13 ( Stata Corp . 2013 ) .
A total of 345 patients were assessed for eligibility in the study period . From this total , 187 accomplished the inclusion criteria and were randomized , with 93 in the interventional group and 94 in the untreated control group . One patient of the control group was lost to follow up ( Fig 1 ) . Enzyme immunoassay confirmed Bothrops envenoming diagnosis in all included patients . Epidemiological characterization showed predominance of males ( 82 . 3% ) , mostly occurring in rural areas ( 87 . 1% ) . The most affected age group was the 21–30 years old ( 22 . 6% ) . The most affected anatomical site was the foot ( 66 . 1% ) . A total of 40 . 3% of cases were classified as work-related bites and 65 . 6% of the patients walked after the snakebite . Time elapsed from bite to medical assistance was higher than 3 hours in 42 . 4% of the cases . Use of topical medicines was informed in 34 . 4% , oral medicines in 28 . 5% and tourniquet in 24 . 7% of the cases ( Table 1 ) . The most frequent manifestations observed at admission were severe pain ( 46 . 2% ) , mild edema ( 48 . 4% ) and local bleeding ( 46 . 8% ) . The difference of temperature between the bite site and the contralateral site was predominantly <1°C ( 51 . 9% ) . The most frequent systemic manifestations were headache ( 26 . 9% ) , dizziness ( 14 . 5% ) , gingival bleeding ( 8 . 6% ) , nausea ( 8 . 1% ) and vomiting ( 7 . 0% ) . Mean axillary temperature was 36 . 1 oC , mean heart rate was 79 . 3 bpm , mean respiratory rate was 20 bpm , mean systolic pressure ( mmHg ) was 129 . 8 mmHg and mean diastolic pressure ( mmHg ) was 82 . 6 mmHg . Snakebites were classified as moderate in 48 . 9% of the cases ( Table 2 ) . No compartmental syndrome , sepsis , gangrene , amputation or death was seen . Laboratorial characterization revealed mild leukocytosis , hypofibrinogenemia , increased creatine phosphokinase and creatine phosphokinase-MB activities , increased erythrocyte sedimentation rate and mildly increased lactate dehydrogenase activity . Clotting time presented incoagulable in 57 . 5% of patients . Prothrombin time presented incoagulable in 40 . 9% of patients . C-reactive protein was >6 . 5 mg/dL in 18 . 8% of the cases . Mean blood venom concentration was 50 . 9 ng/mL ( Table 3 ) . No statistical differences were observed between treated ( median = 12 ) and untreated ( median = 6 . 5 ) groups when CRP medians were compared ( Z = -0 . 161 , p-value = 0 . 872 ) . Of the 74 patients with secondary infection , cellulitis was diagnosed in 64 and abscess in 29 . Secondary infection rates observed in the intervention and control groups are shown in Table 4 . Survival analysis demonstrated that the time from patient admission to the onset of secondary infection was not different between amoxicillin clavulanate treated and control group ( Log-rank = 2 . 23; p = 0 . 789 ) ( Fig 2 ) . Secondary infection incidence until 7 days after admission was 35 . 5% in the intervention group and 44 . 1% in the control group [RR = 0 . 80 ( 95%CI = 0 . 56 to 1 . 15; p = 0 . 235 ) ] . Cellulitis rate was 30 . 1% in the intervention group and 38 . 7% in the control group [RR = 0 . 78 ( 95%CI = 0 . 52 to 1 . 16; p = 0 . 279 ) ] . The abscess rate was 15 . 1% in the intervention group and 16 . 1% in the control group [RR = 0 . 93 ( 95%CI = 0 . 48 to 1 . 82; p = 0 . 999 ) ] . Secondary infection incidence until 48 hours after admission was 22 . 6% in the intervention group and 36 . 6% in the control group [RR = 0 . 62 ( 95%CI = 0 . 38 to 0 . 98; p = 0 . 038 ) ] . Actually , survival analysis has shown a later onset of secondary infection in the treated group ( Fig 2 ) . No late secondary infection ( after 7 days of follow-up ) was observed . From the total of 74 patients presenting secondary infections , 88 . 2% were males and 88 . 2% occurred in the rural area . The age groups more affected by secondary infections after snakebite were 31–40 and 51–60 years old , with 20 . 6% each . Secondary infections were recorded mostly from bites in the foot ( 61 . 8% ) . Infections were secondary to work-related snakebites in 41 . 2% . Time to medical assistance was less than 3 hours after snakebite in 61 . 8% of the secondary infections cases . Secondary infections were mostly observed in moderate snakebites ( 58 . 8% ) . Use of local products was made in 35 . 5% of the secondary infections cases and of tourniquets in 26 . 5% . Samples from secondary infection injuries were collected for culture from 11 patients , with 6 positive cases . Microorganisms isolated were Morganella morganii ( five cases ) and Staphylococcus aureus ( one case ) . Considering proximal variables , secondary infections incidence in 7 days of follow-up was significantly associated to fibrinogen >400 mg/dL [AOR = 3 . 39 ( 95%CI = 1 . 72 to 6 . 66; p<0 . 001 ) ] , alanine transaminase >44 IU/L [AOR = 2 . 21 ( 95%CI = 1 . 03 to 4 . 75; p = 0 . 006 ) ] and C-reactive protein >6 . 5 mg/L [AOR = 3 . 90 ( 95%CI = 1 . 98 to 7 . 69; p = <0 . 001 ) ] . Regarding intermediate variables , moderate [AOR = 11 . 75 ( 95%CI = 2 . 47 to 55 . 86; p = 0 . 002 ) ] and severe pain [AOR = 16 . 97 ( 95%CI = 2 . 05 to 340 . 80; p = 0 . 001 ) ] , moderate [AOR = 3 . 46 ( 95%CI = 1 . 63 to 7 . 35; p = 0 . 001 ) ] and severe edema [AOR = 3 . 78 ( 95%CI = 1 . 20 to 11 . 90; p = 0 . 023 ) ] and moderate [AOR = 2 . 52 ( 95%CI = 1 . 32 to 4 . 82; p = 0 . 005 ) ] and severe snakebites [AOR = 2 . 80 ( 95%CI = 0 . 90 to 8 . 77; p = 0 . 076 ) ] . No distal variable was associated to secondary infections incidence ( Table 5 ) . In the final multivariate analysis model , secondary infections incidence in 7 days of follow-up remained significantly associated to fibrinogen >400 mg/dL [AOR = 4 . 78 ( 95%CI = 2 . 17 to10 . 55; p<0 . 001 ) ] , alanine transaminase >44 IU/L [AOR = 2 . 52 ( 95%CI = 1 . 06 to 5 . 98; p = 0 . 037 ) ] , C-reactive protein >6 . 5 mg/L [AOR = 2 . 98 ( 95%CI = 1 . 40 to 6 . 35; p = 0 . 005 ) ] , moderate pain [AOR = 24 . 30 ( 95%CI = 4 . 69 to 125 . 84; p<0 . 001 ) ] and moderate snakebites [AOR = 2 . 43 ( 95%CI = 1 . 07 to 5 . 50; p = 0 . 034 ) ] ( Table 6 ) . Secondary infections incidence in 48 hours of follow-up was significantly associated to C-reactive protein >6 . 5 mg/L [AOR = 4 . 28 ( 95%CI = 1 . 81 to 10 . 14; p = 0 . 001 ) ] , moderate [AOR = 5 . 87 ( 95%CI = 2 . 06 to 16 . 74; p = 0 . 001 ) ] and severe pain [AOR = 17 . 89 ( 95%CI = 1 . 71 to 186 . 97; p = 0 . 016 ) ] . Preemptive amoxicillin clavulanate was protective for secondary infections in 48 hours of follow-up [AOR = 0 . 42 ( 95%CI = 0 . 19 to 0 . 94; p = 0 . 034 ) ] ( S1 File ) .
Antimicrobial schemes for prevention or treatment of secondary infections from snakebites are not based on good evidences from randomized clinical trials [36 , 37] . Although the Infectious Diseases Society of America ( IDSA ) guidelines for diagnosis and management of skin and soft-tissue infections indicates amoxicillin clavulanate to reduce complications by prevention of secondary infection from animal bites [38 , 39] , to the best of our knowledge this is the first trial assessing the efficacy of this regimen for snakebites . In this trial , although lower rates of secondary infection was observed in the intervention group such differences between amoxicillin clavulanate treated and control groups over a follow-up of 7 days did not achieve statistical significance . Consistently , analysis of the subgroups of patients who had abscesses or cellulitis presented also not significantly difference in terms of intervention efficacy . A previous study with oral chloramphenicol showed a poor efficacy in preventing secondary infection from Bothrops snakebites [41] , although this drug was suggested as a good alternative for the treatment of local infections which may complicate bites by this snake genus [26 , 51 , 52] . Accordingly , intravenous chloramphenicol plus gentamicin showed no statistical difference between patients treated and untreated groups in Crotalus snakebites [42] . In a previous study amoxacillin was also used preemptively without any clinical benefit [53] . In general , these trials were not guided by the investigation of the bacterial agents responsible by the snakebite site infection in that area . Indeed , even for routine treatment purposes , secondary infection diagnosis is mostly based only from clinical features without microbiological confirmation followed of antimicrobial resistance profile . Oral microbiota of snakes comprises a wide range aerobial and anaerobial microorganisms , including Enterobacteriaceae ( namelly Morganella spp . and Escherichia coli ) , Streptococcus , Aeromonas spp . , Staphylococcus aureus and Clostridium spp . [20–22] . Few reports of microbiological confirmation of bacteria responsible for snakebites abscesses demonstrated a predominance of aerobics Enterobacteriaceae , mainly Morganella morganii [20 , 23 , 24] . The infection may not be necessarily associated to snake's mouth flora but also to local disorders induced by venom . The significant difference present at 48 hours but not at 7th day could be explained by the contamination between 3rd day and 7th day with origin other than oral microbiota of the snake , such as patient microbiota or iatrogeny . A limitation of this study was the absence of definitive identification of bacteria responsible by the infections for most of the participants . From six bacterial isolations , however , Morganella morganii was present in five snakebite site infections , however no antibiogram was routinely performed . Resistance to β-lactam antibiotics in Morganella species is very common and usually mediated by the presence of chromosomally encoded β-lactamases belonging to the AmpC β-lactamase family . These β-lactamases are typically inducible in the presence of β-lactam antibiotics [54] . As a result , agents such as ampicillin , amoxicillin , and first-generation and some second-generation cephalosporins may be ineffective [55] . This finding highlights the need of previous knowledge of the secondary infections epidemiology as a cornerstone in the preemptive antibiotics trials in snakebites . Secondary infection incidence until 48 hours after admission had a relative risk reduction of 38 . 3% for the group using preemptive amoxicillin clavulanate compared to control , showing that antibiotic regimen delayed the onset of secondary infection among treated patients . As the preventive effect did not extend until the end of the follow-up , this delay may represent a severe risk for patients using ineffective preventive antibiotics regimens , because in the absence of signs of local complications patients are usually discharged after 48 hours of hospitalization , and may develop secondary infection without proper medical attention . In the difficulty for riverine and indigenous populations living in remote areas returning to health centers for treatment of this complication , severe clinical conditions such as functional loss , amputation , sepsis and even deaths are possible [28] . Although common in endemic areas , our results point that routine empirical prophylactic use of antibiotics aiming to prevent secondary infection lacks a clearly defined policy , leading to a costly and wasteful inappropriate antibiotic use , and even a risk for the patient [22 , 40] . Unfortunately due to the limited funding , a double-blinded study was not performed , a possible limitation of the study . Another limitation was the enrollment of patients with less than 24 hours after the bite , what may have selected those less prone to develop secondary infection . In Bothrops snakebites , studies mostly describe factors associated to systemic complications , such as coagulopathy [11 , 20] , acute renal failure [11 , 56–59] and death [11 , 60 , 61] , with extreme age groups and time to medical assistance associated with these poor outcomes [61] . There is scarce information in relation to local complications , especially necrosis [59 , 62] and amputation [63] , associated to anatomical region bitten , systemic bleeding , renal failure , older age and use of tourniquet . Although secondary bacterial infections were observed in around 40% of the B . atrox snakebites in the Amazon [13] , as confirmed in this work , no epidemiological or clinical predictive marker is known for this complication . In this work , secondary infections incidence was significantly associated to higher levels fibrinogen , alanine transaminase and C-reactive protein , suggesting these laboratorial markers as auxiliary tools for the diagnosis of secondary infections allied to clinical signs of cellulitis and abscesses . Fibrinogen is an acute-phase protein and its serum concentration may be elevated in inflammatory and infectious conditions associated with vascular damage [64 , 65] . Even that B . atrox metalloproteinases cleave fibrinogen and induces a drop of fibrinogen levels with an increase in fibrin/fibrinogen degradation products ( FDP ) levels in vivo , with a foremost role in the pathogenesis of coagulopathy and intravascular hemolysis in the acute envenoming [66 , 67] , the intense inflamatory reaction linked to secondary infection persisting after the antivenom therapy may trigger an acute phase reactant response in affected patients . High C-reactive protein paralleling with high fibrinogen in patients with soft-tissue secondary infections [39] . Our study suggests that the proinflammatory profile present in secondary infections from snakebites may be responsible for hepatocellular dysfunction and further elevated alanine transaminases , as previously reported in patients with sepsis and urinary tract infections [68–72] . Cellulitis or abscesses occur mostly in the moderate or severe snakebite cases , as previously reported in southern Brazil [52] . The Infectious Diseases Society of America ( IDSA ) guidelines for diagnosis and management of skin and soft-tissue infections indicate amoxicillin clavulanate to prevent secondary infections from animal bites [38 , 39] . However , in this study , patients did not benefit from preemptive amoxicillin clavulanate in preventing secondary infection from Bothrops snakebites . As a perspective , antimicrobial selection to be used in future clinical trials should be pursued diligently in comprehensive snakebites infection management . Secondary infections incidence was significantly associated to higher levels fibrinogen , alanine transaminase and C-reactive protein , suggesting these laboratorial markers as auxiliary tools for a more accurate diagnosis of secondary infections allied to clinical evaluation .
|
Bothrops genus is responsible by 80–90% of the snakebites in the Brazilian Amazon , resulting in a subcutaneous and muscular lesion at the site of bite , which many times evolve to local complications , mostly secondary bacterial infections . In this region , late medical assistance is common and probably contributes to the high complication rates related to local necrosis and secondary bacterial infections , which can lead to permanent function loss and disabilities . Even with this high frequency , routine empirical use of antibiotics aiming to prevent secondary infection lacks a clearly defined protocol . In this work , we estimated the efficacy of amoxicillin clavulanate for reducing the secondary infection incidence in patients bitten by Bothrops snakes , and , identified factors related to secondary infections from snakebites . Amoxicillin clavulanate was not effective for preventing secondary infections from Bothrops snakebites , probably because of the resistance to β-lactam antibiotics in bacteria species commonly found infecting the snakebite site . This finding highlights the need of previous knowledge of the secondary infections epidemiology as a cornerstone in the preemptive antibiotics trials in snakebites . Laboratorial markers , such as high fibrinogen , alanine transaminase and C-reactive protein levels , and severity clinical grading of snakebites , may help to accurately diagnose secondary infections .
|
[
"Abstract",
"Introduction",
"Methods",
"and",
"materials",
"Results",
"Discussion"
] |
[
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"c-reactive",
"proteins",
"chemical",
"compounds",
"fibrinogen",
"aliphatic",
"amino",
"acids",
"pathology",
"and",
"laboratory",
"medicine",
"drugs",
"tropical",
"diseases",
"microbiology",
"creatine",
"vertebrates",
"organic",
"compounds",
"animals",
"signs",
"and",
"symptoms",
"antibiotics",
"reptiles",
"amino",
"acids",
"neglected",
"tropical",
"diseases",
"pharmacology",
"glycoproteins",
"snakebite",
"abscesses",
"alanine",
"proteins",
"chemistry",
"snakes",
"biochemistry",
"diagnostic",
"medicine",
"organic",
"chemistry",
"squamates",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"amniotes",
"glycobiology",
"organisms"
] |
2017
|
Poor efficacy of preemptive amoxicillin clavulanate for preventing secondary infection from Bothrops snakebites in the Brazilian Amazon: A randomized controlled clinical trial
|
When sequencing an ancient DNA sample from a hominin fossil , DNA from present-day humans involved in excavation and extraction will be sequenced along with the endogenous material . This type of contamination is problematic for downstream analyses as it will introduce a bias towards the population of the contaminating individual ( s ) . Quantifying the extent of contamination is a crucial step as it allows researchers to account for possible biases that may arise in downstream genetic analyses . Here , we present an MCMC algorithm to co-estimate the contamination rate , sequencing error rate and demographic parameters—including drift times and admixture rates—for an ancient nuclear genome obtained from human remains , when the putative contaminating DNA comes from present-day humans . We assume we have a large panel representing the putative contaminant population ( e . g . European , East Asian or African ) . The method is implemented in a C++ program called ‘Demographic Inference with Contamination and Error’ ( DICE ) . We applied it to simulations and genome data from ancient Neanderthals and modern humans . With reasonable levels of genome sequence coverage ( >3X ) , we find we can recover accurate estimates of all these parameters , even when the contamination rate is as high as 50% .
When sequencing a human genome using ancient DNA ( aDNA ) recovered from fossils , a common practice is to assess the amount of present-day human contamination in a sequencing library [1 , 2 , 3 , 4 , 5 , 6] . Several methods exist to obtain a contamination estimate . First , one can look at ‘diagnostic positions’ in the mitochondrial genome at which a particular archaic population may be known to differ from all present-day humans . Then , one counts how many aDNA fragments support the present-day human base at those positions . This is the most popular technique and has been routinely deployed in the sequencing of Neanderthal genomes [7 , 1] . However , contamination levels of the mitochondrial genome may sometimes differ drastically from those of the nuclear genome [8 , 9] . A second technique involves assessing whether the sample was male or female using the number of fragments that map to the X and the Y chromosomes . After determining the biological sex , the proportion of reads that are non-concordant with the sex of the archaic individual are used to estimate contamination from individuals of the opposite sex ( e . g . Y-chr reads in an archaic female genome are indicative of male contamination ) [8 , 1 , 10 , 4] . Another method uses a maximum-likelihood approach to estimate contamination , but is only applicable to single-copy chromosomes , like the X chromosome in individuals known a priori to be male [11 , 12] . Finally , one last technique involves using a maximum-likelihood approach to co-estimate the amount of contamination , sequencing error and heterozygosity in the entire autosomal nuclear genome [1 , 3] , using an optimization algorithm such as L-BFGS-B [13] . Afterwards , if the aDNA library shows low levels of present-day human contamination ( < ∼2% ) , demographic analyses are performed on the sequences while ignoring the contamination . If the library is highly contaminated , it is usually treated as unusable and discarded . Neither of these outcomes is optimal: contaminating fragments may affect downstream analyses , while discarding the library as a whole may waste precious genomic data that could provide important demographic insights . One way to address this problem was proposed by skoglund et al . [14] , who developed a statistical framework to separate contaminant from endogenous DNA fragments by using the patterns of chemical deamination characteristic of ancient DNA . The method produces a score which reflects the odds that a particular fragment is endogenous or not , based on these chemical patterns . This approach is effective at isolating truly endogenous fragments from contaminant fragments , but at the cost of potentially discarding some fragments that may not have chemical damage and still be endogenous . This becomes more problematic the younger the ancient DNA sample is , because younger samples will tend to have a higher proportion of non-deaminated ancient DNA , and so the method will lead users to discard a larger fraction of endogenous material . Instead of ( or in addition to ) attempting to separate the two type of fragments before performing a demographic analysis , one could incorporate the uncertainty stemming from the contaminant fragments into a probabilistic inference framework . Such an approach has already been implemented in the analysis of a haploid mtDNA archaic genome [15] . However , mtDNA represents a single gene genealogy , and , so far , no equivalent method has been developed for the analysis of the nuclear genome , which contains the richest amount of population genetic information . Here , we present a method to co-estimate the contamination rate , per-base error rate and a simple demography for an autosomal nuclear genome of an ancient hominin . We assume we have a large panel representing the putative contaminant population , for example , European , Asian or African 1000 Genomes data [16] . The method uses a Bayesian framework to obtain posterior estimates of all parameters of interest , including population-size-scaled divergence times and admixture rates .
We will first describe the probabilistic structure of our inference framework . We begin by defining the following parameters: We are interested in computing the probability of the data given the contamination rate , the error rate , the derived allele frequencies from the putative contaminant population ( w ) and a set of demographic parameters ( Ω ) . We will use only sites that are segregating in the contaminant panel and we will assume that we observe only ancestral or derived alleles at every site ( i . e . we ignore triallelic sites ) . In some of the analyses below , we will also assume that we have additional data ( O ) from present-day populations that may be related to the population to which the sample belongs . The nature of the data in O will be explained below , and will vary in each of the different cases we describe . The parameters contained in Ω may simply be the population-scaled times separating the contaminant population and the sample from their common ancestral population . However , Ω may include additional parameters , such as the admixture rate—if any—between the contaminant and the sample population . The number of parameters we can include in Ω will depend on the nature of the data in O . For all models we will describe , the probability of the data can be defined as: P [ a , d | r C , ϵ , w , Ω , O ] = ∏ j = 1 K P [ a j , d j | r C , ϵ , w j , Ω , O ] ( 1 ) where P [ a j , d j | r C , ϵ , w j , Ω , O ] = ∑ i = 0 2 P [ a j , d j | i , r C , ϵ , w j ] P [ i | Ω , O ] ( 2 ) Here , i is the true ( unknown ) genotype of the ancient sample , and P[i |Ω , O] is the probability of genotype i given the demographic parameters and the data . We focus now on computation on the likelihood for one site j in the genome . In the following , we abuse notation and drop the subscript j . Given the true genotype of the ancient individual , the number of derived and ancestral fragments at a particular site follows a binomial distribution that depends on the genotype , the error rate and the rate of contamination [1 , 3]: P [ a , d | i , r C , ϵ , w ] = a + d d q i d ( 1 - q i ) a ( 3 ) where q 2 = r C w ( 1 - ϵ ) + ( 1 - w ) ϵ + ( 1 - r C ) ( 1 - ϵ ) ( 4 ) q 1 = r C w ( 1 - ϵ ) + ( 1 - w ) ϵ + ( 1 - r C ) ( 1 - ϵ ) / 2 + ϵ / 2 ( 5 ) q 0 = r C w ( 1 - ϵ ) + ( 1 - w ) ϵ + ( 1 - r C ) ϵ ( 6 ) In the sections below , we will turn to the more complicated part of the model , which is obtaining the probability P[i|Ω , O] for a genotype in the ancient sample , given particular demographic parameters and additional data available . We will do this in different ways , depending on the kind of data we have at hand . First , we will work with the case in which O = y , where y is a vector of frequencies yj from an “anchor” population that may be closely related to the population of the ancient DNA sample . An example of this scenario would be the sequencing of a Neanderthal sample that is suspected to have contamination from present-day humans , from which many genomes are available . For all analyses below , we restrict to sites where 0 < yj < 1 . Note that it is entirely possible ( but not required ) that y = w , meaning that , aside from the ancient DNA sample , the only additional data we have are the frequencies of the derived allele in the putative contaminant population , which we can use as the anchor population too . However , it is also possible to use a contaminant panel that is different from the anchor population ( Fig 1A ) . We will assume we have sequenced a large number of individuals from a panel of the contaminant population ( for example , The 1000 Genomes Project panel ) and that the panel is large enough such that the sampling variance is approximately 0 . In other words , the frequency we observe in the contaminant panel will be assumed to be equal to the population frequency in the entire contaminant population . In this case , Ω = {τC , τA} , where τA and τC are defined as follows: We need to calculate the conditional probabilities P[i|Ω , O] = P[i|y , τC , τA] for all three possibilities for the genotype in the ancient individual: i = 0 , 1 or 2 . To obtain these expressions , we rely on Wright-Fisher diffusion theory ( reviewed in Ewens [17] ) , especially focusing on the two-population site-frequency spectrum ( SFS ) [18] . The full derivations can be found in the S1 Text , and lead to the following formulas: P [ i = 0 | y , τ C , τ A ] = 1 - y * e - τ C - 1 2 * y * e - τ A - τ C + y y - 1 2 e - τ A - 3 τ C ( 7 ) P [ i = 1 | y , τ C , τ A ] = y * e - τ A - τ C + y 1 - 2 y e - τ A - 3 τ C ( 8 ) P [ i = 2 | y , τ C , τ A ] = y * e - τ C - 1 2 * y * e - τ A - τ C + y y - 1 2 e - τ A - 3 τ C ( 9 ) We generated 10 , 000 neutral simulations using msms [19] for different choices of τC and τA ( with θ = 20 in each simulation ) to verify our analytic expressions were correct ( Fig 2 ) . The probability does not depend on θ , so the choice of this value is arbitrary . The above probabilities allows us to finally obtain P[i | yj , Ω , O] . Although the above method gives accurate results for a simple demographic scenario , it does not incorporate the possibility of admixture from the ancient sample to the contaminant population . This is important , as the signal of contamination may mimic the pattern of recent admixture . We will assume that , in addition to the ancient DNA sample , we also have the following data , which constitute O: We can then estimate the remaining drift parameters , the error and contamination rates and the admixture time ( β ) and rate ( α ) between the archaic population and modern population Y . The diffusion solution for this three-population scenario with admixture is very difficult to obtain analytically . Instead , we use a numerical approximation , implemented in the program ∂a∂i [20] . We incorporated the likelihood functions defined above into a Markov Chain Monte Carlo ( MCMC ) inference method , to obtain posterior probability distributions for the contamination rate , the sequencing error rate , the drift times and the admixture rate . Our program—which we called ‘DICE’—is coded in C++ and is freely available at: http://grenaud . github . io/dice/ . We assumed uniform prior distributions for all parameters , and the boundaries of these distributions can be modified by the user . For the starting chain at step 0 , an initial set of parameters X0 = {rC0 , ϵ0 , Ω0} is sampled randomly from their prior distributions . At step k , a new set of values for step k + 1 is proposed by drawing values for each of the parameters from normal distributions . The mean of each of those distributions is the value for each parameter at state Xk and the standard deviation is the difference between the upper and lower boundary of the prior , divided by a constant that can be increased or decreased to achieve a desired rate of acceptance of new states [21] . By default , this constant is equal to 1 , 000 for all parameters . The new state is accepted with probability: P [ a c c e p t ] = m i n 1 , P [ a , d | X k + 1 ] P [ a , d | X k ] ( 10 ) where P[a , d | Xk] is the likelihood defined in Eq 1 . Unless otherwise stated below , we ran the MCMC chain for 100 , 000 steps in all analyses , with a burn-in period of 40 , 000 and sampling every 100 steps . The sampled values were then used to construct posterior distributions for each parameter . Fu et al . [5] showed that , when estimating contamination , ancient DNA data can be better fit by a two-error model than a single-error model . In that study , the authors co-estimate the two genome-wide error rates along with the proportion of the data that is affected by each rate . Therefore , we also included this error model as an option that the user can choose to incorporate when running our program . Furthermore , we developed an alternative error estimation method that allows the user to flag transition polymorphisms , which are more likely to have occurred due to cytosine deamination in ancient DNA . These sites are therefore likely to be subject to different error rates than those common in present-day sequencing data [22 , 23] . Our program can then estimate two error rates separately: one for transitions and one for transversions . Finally , we incorporated an option to include an ancestral state misidentification ( ASM ) parameter , which should serve to correct for mispolarization of alleles [24] . The standard input for DICE is a file containing counts of particular ancestral/derived base combinations and SNP frequencies ( see README file online ) . As an additional feature , we also developed a module for the user to directly input a BAM file and a file containing population allele frequencies for the anchor and contaminant panels , rather than the standard input . The user can either choose to convert the BAM file to native DICE format using a program provided with the software package and then run the program , or run it directly on the BAM file . In the latter case , instead of calculating genome-wide error parameters , the program will calculate error parameters specific to each sequenced fragment , based on mapping qualities , base qualities and estimated deamination rates at each site ( see S2 Text ) .
We have developed a new method to jointly infer demographic parameters , along with contamination and error rates , when analyzing an ancient DNA sample . The method can be deployed using a C++ program ( DICE ) that is easy to use and freely downloadable . We therefore expect it to be highly applicable in the field of paleogenomics , allowing researchers to derive useful information from previously unusable ( highly contaminated ) samples , including archaic humans like Neanderthals , as well as ancient modern humans . Applications to simulations show that the error and contamination parameters are estimated with high accuracy , and that demographic parameters can also be estimated accurately so long as enough information ( e . g . a large panel of modern humans ) is available . The drift time estimates reflect how much genetic drift has acted to differentiate the archaic and modern populations since the split from their common ancestral population , and can be converted to divergence times in generations if an accurate history of population size changes is also available ( for example , via methods like PSMC , [35] ) . Although we cannot perform proper model testing , we found via extensive simulations that the posterior mode of an MCMC run was a robust heuristic statistic to help detect which panel was most likely to have contaminated the sample . We caution , however , that the fact that a particular panel yields a higher posterior mode than another is no guarantee that it is a better fit to the data for demographic scenarios that may be different from the ones we simulated . We also applied our method to empirical data , specifically to two Neanderthal genomes at high and low coverage , a present-day high-coverage Yoruba genome , and several ancient genome sequences of varying degrees of coverage , some obtained via shotgun-sequencing and some via SNP capture . For the high-coverage Yoruba genome , we infer no contamination , as would be expected from a modern-day sample , and drift times indicating the Yoruba sample indeed belongs to an African population . The contamination and sequencing error estimates we obtained for the Altai Neanderthal are roughly in accordance with previous estimates [4] . The drift times we obtain under the three-population model for the African population ( τC + τAfr ) are approximately 0 . 411 + 0 . 009 = 0 . 42 drift units . The geometric mean of the history of population sizes from the PSMC results in Prüfer et al . [4] give roughly that Ne ≈ 21 , 818 since the African population size history started differing from that of Neanderthals , assuming a mutation rate of 1 . 25 * 10−8 per bp per generation . If we assume a generation time of 29 years , and use our drift time in the equation relating divergence time in generations to drift time ( t/ ( 2Ne ) ≈τ ) , this gives an approximate human-Neanderthal population divergence time of 531 , 486 years . This number roughly agrees with the most recent estimates obtained via other methods [4] . Additionally , the Neanderthal-specific drift time is approximately 6 . 5 times as large as the modern human drift time , which is expected as Neanderthals had much smaller population sizes than modern humans [36 , 4] . The admixture rate from archaic to modern humans that we estimate is 1 . 72% , which is consistent with the rate estimate obtained via methods that do not jointly model contamination ( 1 . 5–2 . 1% ) [4] . In the case of the Altai Neanderthal , we observe that the sample was probably contaminated by one or more individuals with European ancestry . When testing modern human and Neanderthal ancient genomes of lower coverage than the Altai Neanderthal , we obtain reasonable parameter estimates for samples of medium to high-coverage . However , we run into problems in estimation when the samples are of low coverage . For these reasons , and from our simulation results , we recommend that our method should be used on nuclear genomes with >3X coverage . The method may converge under certain conditions at coverages as low as 0 . 5X ( for example , in the case of the Mezmaiskaya genome under the two-population model when using AFR as the anchor and contaminant panel ) , but , in such cases , we caution the user to check convergence is achieved before drawing any conclusions from the estimates . For SNP capture data , we obtain reliable estimates for samples with a minimum coverage of 500 , 000 sites that are polymorphic in the anchor panel . The demographic models used in our approach are simple , involving no more than three populations and a single admixture event . This is partly due to limitations of known theory about the diffusion-based likelihood of an arbitrarily complex demography for the 2-D site-frequency spectrum—in the case of the two-population method—and to the inability of ∂a∂i [20] to handle more than 3 populations at a time . In recent years , several studies have made advances in the development of methods to compute the likelihood of an SFS for larger numbers of populations using coalescent theory [37 , 38 , 39] , with multiple population size changes and admixture events . We hope that some of these techniques could be incorporated in future versions of our inference framework .
|
When extracting and sequencing ancient DNA from human remains , a recurrent problem is the presence of DNA from the paleontologists , archaeologists or geneticists that may have handled the fossil . If a DNA library is highly contaminated , this will introduce biases in downstream analyses , so it is important to determine the amount of extraneous DNA . Different methods exist for this purpose , but few are applicable to the nuclear genome , and none of them can extract reliable genomic information from highly contaminated samples . Thus , samples with high rates of contamination are usually discarded . Here , we present a method to jointly estimate contamination and error rates , along with demographic parameters , like drift times and admixture rates . Our method can serve to uncover important details about the evolutionary history of archaic and early modern humans from ancient DNA samples , even if those samples are highly contaminated .
|
[
"Abstract",
"Introduction",
"Methods",
"Discussion"
] |
[
"sequencing",
"techniques",
"demography",
"social",
"sciences",
"anthropology",
"ancient",
"dna",
"contaminants",
"neanderthals",
"genome",
"sequencing",
"materials",
"science",
"genome",
"analysis",
"paleontology",
"dna",
"paleogenetics",
"molecular",
"biology",
"techniques",
"paleoanthropology",
"research",
"and",
"analysis",
"methods",
"genomic",
"libraries",
"genomics",
"hominids",
"materials",
"by",
"attribute",
"molecular",
"biology",
"people",
"and",
"places",
"hominins",
"biochemistry",
"physical",
"anthropology",
"nucleic",
"acids",
"earth",
"sciences",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"dna",
"sequencing"
] |
2016
|
Joint Estimation of Contamination, Error and Demography for Nuclear DNA from Ancient Humans
|
The [PSI+] prion may enhance evolvability by revealing previously cryptic genetic variation , but it is unclear whether such evolvability properties could be favored by natural selection . Sex inhibits the evolution of other putative evolvability mechanisms , such as mutator alleles . This paper explores whether sex also prevents natural selection from favoring modifier alleles that facilitate [PSI+] formation . Sex may permit the spread of “cheater” alleles that acquire the benefits of [PSI+] through mating without incurring the cost of producing [PSI+] at times when it is not adaptive . Using recent quantitative estimates of the frequency of sex in Saccharomyces paradoxus , we calculate that natural selection for evolvability can drive the evolution of the [PSI+] system , so long as yeast populations occasionally require complex adaptations involving synergistic epistasis between two loci . If adaptations are always simple and require substitution at only a single locus , then the [PSI+] system is not favored by natural selection . Obligate sex might inhibit the evolution of [PSI+]-like systems in other species .
[PSI+] is the aggregated prion form of the protein Sup35 [1] . [PSI+] aggregates appear spontaneously at a low rate [2] . Once established , [PSI+] causes normal Sup35 proteins to misfold to form more [PSI+] [3] . This self-catalytic conversion allows for transgenerational inheritance [4] . The normal , non-prion form of Sup35 is involved in stop codon recognition during gene translation [5] , [6] . The depletion of normal Sup35 through its incorporation into prion aggregates leads to readthrough errors at stop codons [7] . This phenotypically reveals previously cryptic genetic variation beyond stop codons [8]–[11] . Revealed variation can sometimes lead to faster growth [10] and adaptation [12] under stressful lab conditions . These observations have controversially suggested a role for [PSI+] in promoting evolvability [13]–[17] . [PSI+] may tap into stocks of variation at times of stress when they are most likely to be needed [18] . [PSI+] induces only low levels of any given adaptive readthrough product . A simple point mutation at the stop codon will produce much higher levels . Let the ancestral allele at locus i with an intact stop codon be designated ( appended gene product - sensu Masel and Bergman [15] ) and the derived , adaptive allele with its stop codon destroyed by mutation be designated ( Figure 1 ) . [PSI+] promotes evolvability by acting as a stopgap mechanism . [PSI+] spontaneously appears far more often than stop codon mutations ( [2] , [19] , [20]; see parameter estimates below ) . This provides partial but nevertheless rapid and therefore easily accessible adaptation . [PSI+] buys time for the lineage to expand , providing more opportunities for more precise adaptation later through genetic assimilation via the appearance of the stop codon mutant agp+ [9] , [14] . [PSI+] may provide a convenient model system for the more general study of evolvability via evolutionary capacitors [21] . Evolutionary capacitors are molecular mechanisms that act as switches to control the storage and release of cryptic genetic variation . Cryptic stocks of variation are likely to be pre-enriched for potential adaptations , making this mechanism of evolvability more potent than a reliance on new mutation [22] . Despite these experimental and theoretical results , a role for [PSI+] in evolvability has remained controversial . In particular , although data directly show that [PSI+] can sometimes promote rapid growth and adaptation in novel environments [10] , [12] , this does not imply that indirect selection for evolvability explains the emergence and evolutionary conservation of the [PSI+] system . Theoretical results support the evolution of the evolvability properties of the [PSI+] system , but all such studies to date have neglected sex [15] , [21] , [23] , [24] . This is of concern , since the evolution of another putative evolvability mechanism , namely mutator alleles , is dramatically inhibited by sex [25]–[27] . This is because recombination rapidly breaks up linkage between mutator alleles and the adaptations they generate , severely limiting the mutator's ability to hitchhike on the success of those adaptations . This argument does not , however , apply in an identical form to the [PSI+] system , since revealed variants remain dependent on continued [PSI+] expression , preventing their adaptive separation by recombination until genetic assimilation has occurred [15] . When linkage equilibrium evolves only slowly , evolvability may be favored by natural selection [28] . Here we examine for the first time the effect of realistic rates of Saccharomyces sex on the evolution of the evolvability properties of the [PSI+] system . Consider a modifier locus prf ( prion-forming - sensu Masel and Bergman [15] ) that affects whether [PSI+] is formed . Examples of modifiers of [PSI+] formation in nature include the [PIN+] prion [29] , chaperone molecules [30]–[32] and changes in the Sup35 sequence [33] , [34] . In our analysis , prf is an abstract modifier in the tradition of theoretical population genetics , rather than a specific , empirically identified modifier . Let the prf0 allele completely suppress de novo [PSI+] formation and the prf+ allele allow for it . We track allele frequencies at the prf locus in order to infer whether the [PSI+] system is favored by natural selection . Both alleles allow propagation of [PSI+] , once present . Usually , [PSI+] is deleterious , and so the prf+ allele incurs small ongoing costs by generating [PSI+] lineages . But on rare occasions , [PSI+] and hence prf+ may be adaptive . The prf0 allele avoids the costs , but is still partially able to usurp the benefits by acquiring the cytoplasmically inherited [PSI+] element through sex with a [PSI+] strain . prf0 can therefore be thought of as a “cheater” allele . When outcrossed sex is rare , however , as it is in Saccharomyces [35] , prf0 will on average acquire [PSI+] only after a potentially significant delay , during which a prf+ lineage may have already hitchhiked to high frequency in association with [PSI+]-facilitated adaptation . Here we determine whether prf+ is able to outcompete prf0 , implying that the [PSI+] system is favored by natural selection on evolvability , given empirically estimated [35] rates of sex in Saccharomyces . An interesting aspect of evolutionary capacitors in general , and the [PSI+] system in particular , is the fact that variants at many loci are exposed simultaneously . It has long been speculated that certain adaptations might involve multiple simultaneous changes , and that a temporary period of relaxed selection would allow multiple mutations to accumulate , providing greater diversity as the raw material for evolution [10] , [36] . Of course , a potential problem with this idea is that cryptic genetic variation may also contain an accumulation of highly deleterious mutations . This may thwart adaptation , since revealing a stock of variation that includes both highly deleterious and mildly adaptive mutations will on balance likely be deleterious . However , capacitors such as [PSI+] tap into stocks of cryptic genetic variation that had remained subject to low levels of selection while in the cryptic state [22] . This low level of “pre-selection” is sufficient to weed out strongly deleterious alleles , while allowing mutations of small effect to accumulate [22] . One consequence of this pre-selection is that when variation is finally released through a capacitor , adaptations involving multiple simultaneous changes occur far more readily than they would without a capacitance mechanism [22] . Here we consider the evolution of the [PSI+] system via the prf modifier locus in the presence of sex , a fluctuating environment in which [PSI+] occasionally promotes adaptation , and both with and without complex adaptations involving multiple loci . We find that in the presence of realistic frequencies of Saccharomyces sex , complex adaptations are both necessary and sufficient for natural selection on evolvability to drive the evolution of the [PSI+] system .
The simulated diploid Saccharomyces population experiences a fluctuating environment . All environments where [PSI+] is deleterious we label “1” and the environments where [PSI+] generates an adaptation we label “2” . The probability of switching from environment 1 to 2 is Ω12 per generation , and the probability of switching from environment 2 to 1 is Ω21 per generation . We explore environmental switching rates between 10−7 and 10−3 per generation . The population starts in environment 1 , with prf+ and prf0 allele frequencies of 0 . 5 , and evolves for 5/Ω12 generations . This process is replicated to determine the proportion of runs for which the prf+ frequency increases . Model parameters are listed in Table 1 , including default values when not otherwise specified . In environment 2 , [PSI+] can mediate adaptation by expressing novel gene product ( s ) at either one locus ( simple adaptation; i = 1 ) or two loci ( complex adaptation; i∈{1 , 2} ) . Our “agp” notation and our parameter estimates are based on the assumption that adaptation comes from addition to a protein C-terminal through stop codon readthrough . The same formalism can , however , still be applied if the variation revealed by [PSI+] is instead mediated via nonstop mRNA decay [11] , via +1 frameshifting at shifty-stop sites [37] , or via variation in genes regulated downstream through any of these mechanisms [37] . Each switch to environment 2 is considered unique and involves a new set of agp loci , whose frequencies are initialized at this time . After switching back to environment 1 , this set of agp loci is no longer tracked . If only a single agp+ allele is required for adaptation , and it is already present or very soon appears in the population , then adaptation will proceed via this more direct route rather than via [PSI+] , yielding no benefit to a prf+ allele . If , however , two different readthrough products are simultaneously involved in a complex adaptation , then it becomes exceedingly unlikely that both alleles will initially be present in the same individual . In this case [PSI+] will have an advantage , since it will cause simultaneous readthrough at both loci , reaping synergistic benefits and promoting complex adaptations . Competing paths of direct vs . [PSI+] mediated adaptation are shown in Figure 2 for the two locus case . We track individual genotypes at the agp and prf loci . Haploids must have different alleles ( α vs . a ) at the mating-type mat locus in order to conjugate , and so we also track the mat locus for its potential effect on inbreeding . We do not model mutation at the mat and prf loci , except implicitly through the possibility of mother-daughter haplo-selfing ( see Reproduction below ) . There is free recombination between all loci . Individuals therefore have either three or four genetic loci , depending on whether we are modeling simple or complex adaptations with one or two agp loci , respectively . Each of the three to four loci has two alleles , plus there are also two possible cytoplasmic states ( [PSI+] versus [psi−] ) . The point mutation rate in Saccharomyces is around 5×10−10 per base pair per cell division [19] . We approximate the frequencies of the 3 stop codons TAA , TAG and TGA as equal and all mutational substitution types as equally likely . All point mutations at the first position destroy the stop codon . So do all but two at the second position ( namely those between TAA and TGA ) out of 3 possible substitutions at each of the 3 stop codons . Similarly , of the 9 possible substitutions at the third position , only those between TAA and TAG preserve the stop codon . The total rate of stop codon destruction by point mutation is therefore estimated as ( 1+7/9+7/9 ) ×5×10−10 = 1 . 3×10−9 per cell division . Although mutations that precisely reverse stop codon loss are rarer than this , compensatory mutations can also create alternative stop codons nearby , leading to a functionally equivalent gene product . We therefore assume symmetric mutation rates at the agp loci . The back mutation rate is primarily important only for setting agp allele frequencies at mutation-selection-drift equilibrium in environment 1 ( see Text S1 ) . We explore both the case where the prf+ allele is completely dominant ( h = 1 ) , and the case where it is completely recessive ( h = 0 ) relative to the prf0 allele . prf+ individuals form [PSI+] with probability m = 10−5 per generation [2 , Tuite MF pers . comm . ] . [PSI+] is lost with probability m′ per generation during cell division . Empirical work shows that m′<0 . 0002 [38]; here we follow the common assumption that m′≈m . Note that m increases by as much as 60-fold in response to environmental stress [18] . This responsiveness increases the ability of prf+ to promote evolvability . Here we make the conservative assumption that m does not depend on the degree of adaptation to the current environment . In both environments , readthrough products at any locus induced either by [PSI+] or by point mutations in stop codons are likely to incur a fitness cost . This cost could be related to gain or loss of function , and hence specific to the gene in question , or it could be a more general metabolic cost . Here , in order to develop a general , parameterized model , we assume a metabolic cost . In environment 2 , the metabolic cost of readthrough is ameliorated because of the adaptive effects of a substrate-dimer reaction involving readthrough products at one or two loci . Readthrough probabilities are δpsi− = 0 . 003 and δPSI+ = 0 . 01 in [psi−] and [PSI+] cells respectively [7] . Let Ei be the level of readthrough at locus i . Ei is equal to δj for genotypes , ( 1+δpsi− ) /2 for genotypes and 1 for genotypes , where j∈{psi− , PSI+} . The unit concentrations for all equations below is now given relative to a typical expression level of a gene defined as E = 1 . The fitness of an individual in environment 1 is ( 1 ) where L is either 1 or 2 , depending on whether simple or complex adaptation is assumed , and βd is a constant that weights the metabolic cost of readthrough at the potentially adaptive loci relative to the metabolic cost of readthrough across the whole genome . Since there are ∼5000 genes in Saccharomyces , we assume that βd = 1/5000 . Equation 1 yields a fitness of 1 in the absence of readthrough , decaying exponentially towards zero as levels of readthrough increase . The parameter α controls the strength of selection against readthrough . In environment 2 , an individual's fitness depends both on the metabolic cost above , and on a benefit accruing from readthrough at agp loci . We assume that the readthrough Agp+ gene product has adaptive function when in the form of a dimer . For simple adaptations , this is an Agp+ homodimer . For complex adaptations , this is an heterodimer . These dimeric scenarios allow us to capture synergistic epistasis in a realistic way that allows direct comparison between one-locus and two-locus models . Fitness in environment 2 is given by ( 2 ) where βb is a parameter controlling the magnitude of the adaptive effects and t1/2 is the half-life of a substrate acted on by a catalytic Agp+ dimer . The first term represents the metabolic cost of readthrough , and is identical to fitness in environment 1 . βb is set such that the relative fitness of [PSI+] homozygous agp+ individuals is 1+s2 where s2 = 0 . 001 , 0 . 01 or 0 . 1 . Fixing s2 in this way allows appropriate comparisons between the 1-locus and 2-locus models . t1/2 captures how the strength of adaptation depends on the extent of readthrough at each of the L loci . The biochemical model for calculating t1/2 depends on the Agp+ dimer concentration and is presented in the Text S2 and Figure S1 . Masel and Griswold [39] estimate the strength of selection against [PSI+] . This estimate depends on the frequency of [PSI+] as a rare polymorphism in wild , [PSI+]-competent Saccharomyces populations . Following expression of a Sup35-GFP fusion protein , a few cells from wild populations show aggregates almost immediately [40] . This suggests the pre-existence of [PSI+] cells containing Sup35 aggregates at a frequency of ε = 1% . If some of these aggregates are false positives , then the true value of ε could be lower . Assuming populations are in epimutation-selection balance , the strength of selection against [PSI+] is [39] ( 3 ) where psex is the probability an offspring is formed sexually . Given that an individual is formed sexually , pauto is the probability it is formed via automixis and pamphi is the probability it is formed via amphimixis ( see below ) . Although Equation 3 is complex , its inference of the strength of selection against [PSI+] depends largely just on the observed [PSI+] frequency ε and the rate of [PSI+] appearance m [39] . Given selection s1 against [PSI+] in environment 1 , α is given by ( 4a ) ( 4b ) Equation 4 is derived from Equation 1 by equating 1−s1 to the fitness of [PSI+] individuals relative to psi− individuals . We useε to calculate s1 and hence α , and α to run our simulations . Since there is uncertainty in the estimate of the equilibrium frequency ε of [PSI+] when deleterious , we explore cases when ε is 0 . 01% , 0 . 1% and 1% . We analyze evolutionary competition between prf+ and prf0 alleles by initializing a population in environment 1 with a 0 . 5 frequency of each allele , and simulating evolution for 5/Ω12 generations to determine how often prf+ increases in frequency . We use mutation-selection-drift balance theory to initialize [PSI+] frequencies , and also to initialize agp frequencies at the moment when the switch to environment 2 occurs ( see Text S1 for details ) . We assume initial linkage equilibrium between all loci . Although epimutation tends to associate [PSI+] with prf+ , we neglect this association during initialization since it is not tractable , and since in any case it establishes itself very rapidly on the timescale of our simulations . Reduced covariance between prf+ and [PSI+] prior to a change from environment 1 to 2 inhibits the maintenance of prf+ ( pers . obs . ) and so the approximation of linkage equilibrium is conservative relative to inferring whether prf+ can be maintained . Given genotype and epigenotype frequencies in one generation , we calculate the effects first of reproduction and epimutation ( described below ) , then of mutation at the agp locus , and finally of selection ( according to fitnesses described above ) to calculate expected genotype frequencies in the next generation . We then sample realized genotype frequencies from expected genotype frequencies using the multinomial distribution to capture genetic drift in a finite population of size Ne . The effective population size Ne in Saccharomyces can be estimated as θ ( 1+F ) / ( 4μ ) where θ is the pairwise sequence divergence estimated as 0 . 0032–0 . 0038 [35] , the inbreeding coefficient F = 0 . 98 [35] , and the per-base pair per replication point mutation rate μ is around 3 . 3×10−10 [20] to 5×10−10 [19] . This yields Ne≈3×106–6×106 . We use Ne = 5×106 . Saccharomyces is generally diploid , and normally reproduces asexually , with only around psex = 0 . 1% of offspring formed via sex [35] . We ignore the haploid stage of the life cycle in our calculations of both mutation and selection , thus assuming that there is no fitness cost to sex in terms of a delay in forming the next generation of diploid offspring . We calculate a combination of sexual and asexual diploid offspring produced instantaneously in each generation . Given sex , only around pamphi = 1% of offspring are generated through amphimictic random mating in the population [35] . pauto = 94% of sexual offspring are formed by automictic within-tetrad mating , while phaplo = 5% of sexual offspring are formed when the products of a haploid mother-daughter mitotic division mate with one another following mating-type switching [35] . In our simulations we explore the effect of varying the overall probability of sex psex , but hold the relative proportions of amphimixis pamphi , automixis pauto and haplo-selfing phaplo constant at the values estimated by Tsai et al . [35] . Amphimictic and automictic mating are only allowed to occur between cells of opposite mating type specified at the mat locus . All sexual reproduction involves independent segregation at each genetic locus . Propagation of [PSI+] state is slightly more complex . Both sexual and asexual reproduction consist of cell division followed by cell growth . During cell division , [PSI+] is lost with probability m′ . During subsequent cell growth , [PSI+] appears spontaneously in prf+ cells with probability m . When reproduction is sexual , both contributing individuals first have the opportunity to lose [PSI+] during meiosis with probability m′ . The new diploid individual is then [psi−] only if both parent cells are [psi−] . This allows prf0 lineages to capture the benefits of [PSI+] . During diploid cell growth following mating , [PSI+] has a single opportunity to appear with probability m in prf+ cells . Some simulations were initialized with only a single prf+ mutant rather than with a 50% allele frequency . This single mutant appears in a random genetic background , and in environment 1 rather than environment 2 with probability Ω21/ ( Ω12+Ω21 ) . When prf+ appeared in environment 1 , simulations were carried out in the same way as for an initial 0 . 5 frequency described above . For single mutants , simulations continued forward in time until prf+ went either extinct or fixed in the population , rather than observing whether its frequency was greater or less than 0 . 5 after a certain number of generations . Fixation probability was then compared to the neutral expectation of 1/N . When prf+ appeared in environment 2 , simulations began at the time of the previous environmental switch from 1 to 2 . Both the time tprf+ of the appearance of the prf+ allele by mutation and the time t21 of switching back to environment 1 were preset as follows . First , tprf+ and t21 were drawn from geometric distributions with mean 1/μprf+ and 1/Ω21 respectively where μprf+ is the probability a prf+ allele arises per generation and was set to an arbitrarily low value . Then while tprf+>t21 , we reset t21 to equal t21−tprf+ . It is important to note that it is possible for the population to adapt to environment 2 prior to the arrival of the prf+ allele . If the population adapts prior to the arrival of the prf+ allele , prf+ will be unconditionally deleterious . 95% confidence intervals in the figures are calculated using the approximate method suggested by Agresti and Coull [41] .
We see in Figure 4 that with psex = 10−3 , as estimated for S . paradoxus [35] , prf+ is favored given complex but not simple adaptations . This inference does not depend on the extreme rarity of yeast sex: with complex adaptations , prf+ would still be maintained even if the probability of sex were raised an order of magnitude . Once sex becomes as frequent as 0 . 1 , prf+ is maintained only if selection on [PSI+]-mediated adaptations is strong . From these results , it seems unlikely that a [PSI+]-like evolvability system could be favored by natural selection in an obligately sexual species under the conditions considered here . For most of our simulations , we assume prf+ is dominant ( h = 1 ) . When prf+ is completely recessive ( h = 0 ) , sex provides even less of a barrier to the evolution of evolvability ( Figure 4 ) . Inference of the strength of selection against [PSI+] in Equation 3 depends on the estimate ε = 1% of the mean frequency of [PSI+] in prf+ populations at mutation-selection-drift equilibrium ( see [39] for details ) . This estimate may contain false positives and instead be an upper bound . In Figure S2 , we see that uncertainty in ε is not important , since lower values of ε , implying stronger selection against [PSI+] in environment 1 , lead us to the same conclusions . If environment 2 is too short-lived for selective sweeps to be completed , then capacitance cannot evolve ( Figure 5 ) . This agrees with previous work using a different modeling approach [42] . Opportunities for adaptation must also arise at a minimum frequency for capacitance to evolve ( Figure 6 ) . Previous work in an asexual model found that a capacitor must be useful at a minimum frequency of Ω12>1/Ne per generation in order to be favored by natural selection [23] . With realistic levels of Saccharomyces sex ( i . e . , psex = 0 . 001 ) , we see in Figure 6 that prf+ increases in frequency when Ω12>2×10−6 , corresponding instead to Ω12Ne>10 . This still corresponds to an exceptionally mild and plausible absolute requirement on the rate of environmental change . A prf+ frequency of 0 . 5 is a very artificial starting condition , and was chosen for computational efficiency . To test the sensitivity of our results to this starting condition , we also did an invasion analysis starting with a single new prf+ mutant ( Figures 7 and 8 ) . The neutral expectation of fixation with probability 1/N is shown by a dashed line . In agreement with results using a 0 . 5 starting condition ( Figure 3 ) , we find that prf+ will fix with a probability greater than the neutral expectation , provided that sex is not too common and selection is not too weak ( Figure 7 ) . prf+ fixes more often than the neutral expectation when Ω12≥10−7 ( Figure 8 ) , favoring evolvability at even lower levels of Ω12 than with a 0 . 5 starting condition ( Figure 6 ) , in agreement with previous work in an asexual model that Ω12>1/Ne per generation is the necessary and sufficient condition for prf+ to be favored by natural selection [23] . Our more comprehensive calculations above that began with a prf+ frequency of 0 . 5 seem to be mildly conservative with respect to the evolution of evolvability .
When realistic levels of yeast sex are accounted for , indirect selection for evolvability can still favor the evolution of the [PSI+] system . This is only true , however , if adaptation involves at least two loci with synergistic epistatic effects on fitness . Otherwise , with an effective population size as large as that of yeast , all single-locus mutants are readily accessible through mutation . [PSI+] is a stopgap adaptation that incurs costs as well as benefits , and is never preferred to direct adaptation . However , simultaneous direct adaptation at multiple loci is extremely rare , and modifiers of [PSI+] hitchhike to high frequency by virtue of facilitating it . Evolutionary capacitors , by exposing multiple variants simultaneously , have long been believed to facilitate complex adaptations involving multiple sites [10] , [22] , [36] . Here we find that the converse is also true: complex adaptations facilitate the evolution of capacitors . This illustrates the intricate relationship between the two . Sex strongly inhibits the evolution of mutator genes , but here we find that its effect on modifiers of capacitance is much weaker . Nevertheless , were yeast to undergo obligate sex , this would be sufficient to disrupt the evolution of [PSI+] under a model of 2-locus adaptation . Our model is specific to the parameters of the [PSI+] system in Saccharomyces , and the evolution of other putative capacitors in the presence of sex still remains to be determined .
|
Can evolvability evolve ? One obvious way to evolve faster is via mutator alleles that increase the mutation rate . Unfortunately , recombination will rapidly separate a mutator allele from the advantageous alleles that it creates . Mutators , therefore , gain very little benefit from promoting adaptations and are thought not to evolve in sexual organisms . Here we find that the [PSI+] prion , unlike mutator alleles , will evolve to promote evolvability in sexual yeast species . Together with previous laboratory studies of [PSI+]–mediated adaptation , and with bioinformatic studies consistent with [PSI+]–mediated adaptation in the wild , our theoretical results firmly establish [PSI+] as a model system for the evolution of evolvability . We also shed light on the importance of complex adaptations involving multiple genes . Adaptations involving multiple simultaneous genes drive the evolution of evolvability in this system . This work is an important proof of principle , showing that evolvability can sometimes evolve under realistic conditions .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"computational",
"biology/evolutionary",
"modeling",
"evolutionary",
"biology/microbial",
"evolution",
"and",
"genomics"
] |
2009
|
Complex Adaptations Can Drive the Evolution of the Capacitor [PSI+], Even with Realistic Rates of Yeast Sex
|
Anthrax is a soil-borne disease caused by the bacterium Bacillus anthracis and is considered a neglected zoonosis . In the country of Georgia , recent reports have indicated an increase in the incidence of human anthrax . Identifying sub-national areas of increased risk may help direct appropriate public health control measures . The purpose of this study was to evaluate the spatial distribution of human anthrax and identify environmental/anthropogenic factors associated with persistent clusters . A database of human cutaneous anthrax in Georgia during the period 2000–2009 was constructed using a geographic information system ( GIS ) with case data recorded to the community location . The spatial scan statistic was used to identify persistence of human cutaneous anthrax . Risk factors related to clusters of persistence were modeled using a multivariate logistic regression . Areas of persistence were identified in the southeastern part of the country . Results indicated that the persistence of human cutaneous anthrax showed a strong positive association with soil pH and urban areas . Anthrax represents a persistent threat to public and veterinary health in Georgia . The findings here showed that the local level heterogeneity in the persistence of human cutaneous anthrax necessitates directed interventions to mitigate the disease . High risk areas identified in this study can be targeted for public health control measures such as farmer education and livestock vaccination campaigns .
Recently it has been suggested that the burden of anthrax has not been fully realized [1] . Interest in anthrax as a biological weapon has grown since the 2001 bioterrorist attacks [2] yet the disease is often undervalued as a public health threat . Anthrax is considered a neglected zoonosis , disproportionately afflicting rural areas in developing nations [3] . Although several countries have implemented successful control strategies , the disease continues to persist globally with an estimated 20 , 000 to 100 , 000 new cases yearly [4] . However , the true disease burden is likely unknown . Areas most heavily impacted by the disease include countries of sub-Saharan Africa , and former Soviet states in Southwestern and Central Asia [5] , [6] . The causative agent , Bacillus anthracis , is a soil-borne , Gram-positive bacterium , which primarily infects herbivores and secondarily afflicts humans . The bacterium is able to persist in the environment for years , possibly even decades , in alkaline soils with high calcium and organic matter [7] and has been shown to be limited geographically by other environmental factors such as temperature [8] , [9] , [10] . Human transmission is often a result of coming into contact with infected animals or contaminated animal materials during agricultural activities including the butchering of livestock or industrial exposures through the processing of hair and bone [8] , [9] , [10] . Manifestation of the disease in humans typically occurs in three acute clinical forms with cutaneous anthrax comprising ∼95% of all reported cases , followed by gastrointestinal and inhalation anthrax [8] , [9] , [10] . Although the most common form of the disease is treatable with antimicrobials , if left untreated , human cutaneous anthrax can be highly fatal with case fatality ratios in excess of 20% [11] , [12] . In neglected endemic regions anthrax can result in substantial economic losses from livestock mortality and lost worker productivity . In the country of Georgia , anthrax is classified as endemic [13] and has persisted for centuries with the first description of human disease thought to be anthrax documented in 1697 [4] . Following the collapse of the Soviet Union , public and veterinary health services in Georgia experienced tremendous setbacks due to deteriorating organization and finances [14] . During this period of transition to independence Georgia experienced an increase in the number of reported human anthrax cases with 118 reported cases during 1991 to 1996 compared to 36 cases reported during 1985 to 1990 [15] . Recent evidence indicates that the situation has continued to worsen and that incidence rates have surpassed that of neighboring countries , including hyperendemic Turkey [5] . Although human anthrax outbreaks are in some instances sporadic , as evident by the recent outbreak in Bangladesh after a more than 20 year absence [14] , [16] , the ability of the bacterium to survive in the environment [16] , [17] can give rise to areas of persistence and disease recurrence . Additionally , the heterogeneous nature of the disease [18] necessitates control efforts that target high risk areas rather than employing blanket control efforts . This is especially true in developing nations where resources are often limited . Therefore , identifying local areas of persistent may allow for the more efficient implementation of public health intervention strategies such as livestock vaccination , increased awareness of the disease , and placement of improved diagnostics for local veterinary and human health care facilities . This study had three objectives: 1 ) to analyze human anthrax at a local scale to better understand the distribution and burden of disease over the past decade , 2 ) identify areas of persistence for targeted public health control measures 3 ) identify environmental and anthropogenic factors associated with areas of persistence .
No human subjects work was undertaken in this study , human anthrax case data were extracted from government reports prepared and approved by the Georgian National Center for Disease Control and Public Health ( NCDC ) . These government reports provided summarized count data of patients diagnosed at health care facilities by category of disease and year . All data were anonymised . Anthrax is a nationally reportable infectious disease in Georgia . Surveillance and documentation within the country is undertaken by the NCDC , which is comprised of a reporting network of 11 regional and 66 district public health centers [7] , [19] . A spatial database of reported human cutaneous anthrax was constructed with a geographic information system ( GIS ) using case data from 2000 to 2009 ( Figure S1 ) . Latitude and longitude coordinate pairs were matched to the reported community using the National Geospatial Intelligence Agency GEOnet names server [20] , Index Mundi [21] , and the GeoNames database [22] . Population data for each community were derived from the Population Statistics of Eastern Europe database [23] for the year 2002 . Of the communities that were included in this human anthrax database we were unable to identify population estimates for seven of these locations . In these instances , the population of these seven communities were estimated using the Gridded Population of the World ( GPW3 ) population dataset for the year 2000 [24] . Yearly population estimates were extrapolated using the United Nations ( UN ) medium variant population growth rates for Georgia during the periods 2000 to 2005 ( −1 . 17 ) and 2005 to 2010 ( −0 . 57 ) ( http://data . un . org/ ) using the following formula: where a is the UN population growth rate for either the period 2000 to 2005 or 2005 to 2010 aforementioned , b is the number of years from the reference population and ( p ) is the population . Yearly national incidence rates per one million population were calculated with the total number of cases reported in Georgia for each year as the numerator and the yearly national population obtained from GeoStat [25] as the denominator . The study period was divided into two equal five-year periods to identify potential changes in the reporting of the disease over time and space . Cumulative incidence risk was calculated per community for each five-year period ( 2000 to 2004 ) and ( 2005 to 2009 ) with the total number of cases during each period as the numerator and the median year community population of each period as the denominator . Cumulative incidences per 10 , 000 population were mapped at the community level and at the district level using ArcGIS 9 . 3 . 1 [26] . The methodology for mapping at the district level is described in Text S1 . The average incidence per 10 , 000 population was calculated to investigate the global clustering among communities . Smoothed risk estimates were calculated for each five-year time period using the Empirical Bayes Smoothing ( EBS ) in the GeoDa software package [27] . The EBS technique can be used to adjust for instability in the risk estimates caused by heterogeneity in the distribution of cases and the population . Posterior risk is estimated from a weighted combination of the local risk and the risk in the surrounding areas ( the prior ) . It has been suggested that the EBS methodology can be implemented in several scenarios , such as when the numerator data total less than three cases , which was the situation in this analysis [28] . In order to maintain a standard comparison of rates between time periods , crude and EBS estimates of cumulative incidence rates ( per 10 , 000 ) were mapped using graduated symbols with the same data bins ( 0–3 , 4–9 , 10–18 , 19–30 , 31–50 , >50 ) . A chi-square analysis was used to test for a relationship between the number of communities reporting anthrax in the east and west of the country during each time period . Boxplots were used to illustrate differences between the crude and EBS rates for both of the five-year time periods . To test for spatial dependence in EBS average anthrax incidence rates between communities , the global Moran's I test was implemented using OpenGeoDa [29] . This test analyzes the degree of spatial autocorrelation present in the data in relation to a single variable of interest and is written following Moran [30]:where N is the number of communities , is the average EBS incidence at each community , Xi and Xj are the EBS incidence in community i and j , and Wij is the spatial weight matrix defining the spatial relationships between communities . In this instance Wij was defined as a series of increasing distances from 1 to 10 kilometers ( km ) to test a range of dependencies . The statistic produces values between −1 and +1 similar in interpretation to the Pearson's correlation with large positive values of I indicating the similarity of values between locations with the presence of either positive or negative autocorrelation and negative values indicating a dissimilarity among values . Significance of the test statistic was assessed with a pseudo p-value generated using 999 random permutations [29] . To identify the spatial and spatio-temporal clustering of anthrax cases during the 10-year period the Poisson clustering models using the space-only and space-time scan statistics were implemented in SaTScan v9 . 0 [31] . SaTScan uses a series of moving windows of varying diameter to detect spatial clusters across a study area and simultaneously implements a series of cylinders of varying height to detect temporal clustering as well [29] . The relative risk ( RR ) is calculated based on the number of observed and expected observations and the likelihood function for each window location and size depending on the assumed distribution [32] . For the SaTScan analyses , community latitude and longitude coordinates were used as the case locations . The model was run on yearly case data , using the total number of yearly cases per community while adjusting for the underlying population of each community . Space-only models were run for each five-year time period using 25% of the population at risk . This threshold for the population at risk was chosen to identify local clusters of anthrax rather than using the default of 50% . Maximum spatial extents of 25% , 20% , 15% , and 10% of the population at risk were chosen with no change in cluster size or location , therefore 25% was used in the analysis . All SaTScan cluster were chosen at the p≤0 . 05 level of significance . In this study we were particularly interested in identifying areas of human anthrax that experienced recurrent infections or persistence over space and time . To identify persistence we implemented the Space-Time Poisson model in SaTScan using 25% of population at risk with a maximum temporal window of 90% of the study period . The temporal window was extended to 90% to allow SaTScan to search across a greater portion of the study period and thereby identify communities that may have persisted as clusters in both space and time across a large temporal window . Additionally , a SaTScan model examining spatial variation in a temporal trend was used to identify individual communities that may have been contributing to an increasing trend in incidence over time . SaTScan fixes the temporal window during the analysis and compares the trend in risk inside and outside of the varying size spatial window [32] , [33] . The most likely spatial cluster location is chosen by the cylinder that maximizes the change in the trend over time . Environmental and anthropogenic variables used in this analysis have been suggested in the literature to be associated with the presence of human anthrax ( Table 1 ) [32] , [33] . Continuous environmental variables included elevation , average annual precipitation [32] , mean land surface temperature ( LST ) , mean mid infra-red reflectance ( MIR ) [3] , [6] , [7] , [8] , and minimum soil pH ( soil pH ) [34] . Continuous anthropogenic variables included population density [35] , travel time to cities with a population >50 , 000 ( travel time ) [36] and categorical variables for urban/rural classification ( UR ) [26] and cattle density dichotomized into equal groups <19 head per 1 km2 and >19 head per 1 km2 ( cattle density ) [37] . Logistic regression was used to identify differences in risk factors between communities that were identified as a persistent space-time cluster in SaTScan and those communities that were not . A total of nine environmental/anthropogenic variables ( Table 1 ) were used to construct univariate , and a final multivariate , logistic regression models using a binary response 0/1 with clustered communities = 1 ( n = 24 ) and non-clustered communities = 0 ( n = 80 ) as the dependent variable . The final model was constructed in SAS using a backward stepwise approach with entry p-values of 0 . 15 and stay p-values of 0 . 20 . Variables used in the stepwise model selection were limited to those that did not show significant collinearity ( ρs≤0 . 70 ) using a Spearman's rank correlation test . Goodness of fit was evaluated with the Hosmer and Lemeshow test and interaction terms were tested between categorical and continuous independent variables . To assess the discrimination ability of the final logistic regression model we used receiver operating ( ROC ) curves , based on the area under the curve ( AUC ) [38] . The AUC provides a measure of model accuracy with values ranging from 0 . 0 to 1 , where values of 1 indicates all locations were correctly classified by the model and 0 . 5 indicating no better than random .
During the study period there were 340 cases of human anthrax reported in Georgia with a median number of reported cases per year of approximately 33 . 5 ( CI95%: 22 . 5 , 42 . 0 ) . Yearly national incidences per one million displayed inter annual variability ranging from a low of 3 . 4 ( CI95%:1 . 9 , 5 . 7 ) in 2002 and to a high of 13 . 9 ( CI95%: 10 . 7 , 17 . 9 ) in 2008 ( Figure 1 ) . There were 143 cases [range: 15–45] during period one ( 2000 to 2004 ) with a median number of yearly cases of 27 ( CI95%: 9 , 39 ) and a total of 197 cases [range: 18–61] during period two ( 2005 to 2009 ) for a yearly median of 38 ( CI95%: 15 , 58 ) . Figure 2 shows the spatial distribution of crude ( insets 1A and 2A ) and smoothed ( insets 1B and 2B ) community level cumulative incidence rates in Georgia for each time period , respectively . There were a greater number of communities reporting human anthrax during period two ( n = 79 ) compared to period one ( n = 44 ) with an increased number occurring in the west of the country during period two ( n = 47 ) compared to period one ( n = 23 ) ; however this difference was not significant ( χ2 = 0 . 6 , p = 0 . 44 ) . . Community level cumulative incidence rates per 10 , 000 population during period one ranged from a low of 0 . 09 ( CI95%: 0 . 05 , 0 . 17 ) to a high of 136 . 4 ( CI95%: 83 . 3 , 210 . 07 ) and during period two ranged from a low of 0 . 18 ( CI95%: 0 . 11 , 0 . 29 ) to a high of 242 . 3 ( CI95%: 97 . 07 , 499 . 2 ) . Box plots illustrated that the cumulative incidence rates for each time period were unstable due to population heterogeneity with a greater than 30% reduction in the estimates applying an EBS rate adjustment ( Figure S2 ) . District level estimates of cumulative risk are displayed in Figure S3 . The Moran's I test identified the presence of significant positive spatial autocorrelation ( Figure 3 ) . A correlogram of Moran's I index values showed increases in positive values up to a distance of 3000 meters and then a subsequent sharp decline up to a distance of 8000 meters indicating spatial dependence at a relatively local scale . The space-only SaTScan analysis detected the presence of three significant clusters during period one ( Figure 4A ) and two clusters during the period two ( Figure 4B ) . Spatial clusters in period one were smaller , and contained fewer communities compared to period two . Clusters varied spatially between periods with a large cluster in period two located in the center of the country that was absent during period one . Persistent clusters in both periods were identified around the main urban center , the capital Tbilisi in the southeast . Results from the space-time model identifying persistence revealed the presence of four significant persistent clusters with two cluster located near the capital Tbilisi ( Figure 5A ) . Clustered communities had an average of 1 . 01 cases per year compared to 0 . 2 cases per community outside of clusters . Clusters were predominantly located in the central and eastern part of the country excluding a large number of communities in the west . The spatial variation in a temporal trend model indicated a single cluster of communities that were contributing to an increasing trend in human anthrax reporting ( Figure 5B ) . Communities within the central part of the country were shown to have incidence of anthrax increasing at an average of 38 . 1% per year as compared to 2 . 8% per year for the rest of the communities in the study area . Results from the logistic regression models are shown in Table 2 . The Hosmer and Lemeshow test ( p = 0 . 96 ) revealed the model was a good fit to the data . There were no statistically significant interaction terms at the p≤0 . 10 . Accuracy of the final main effects model evaluated by the AUC from the ROC analysis showed good discrimination ( AUC = 0 . 85 , p<0 . 001 ) . . Environmental and anthropogenic factors were both found to be associated with the presence of persistence . Strong positive associations shown in odds ratios ( OR ) were found between soil pH ( OR 4 . 58 , CI95%: 1 . 55 , 13 . 51 ) , and UR ( urban areas have greater odds OR , 4 . 67 CI95%: 1 . 11 , 19 . 64 ) . Cattle density ( cattle density above 19 head per 1 km2 have greater odds , OR 2 . 78 , CI95%: 0 . 91 , 9 . 55 ) was also shown to be positively associated although not at the 0 . 05 level of significance ( p = 0 . 10 ) .
Anthrax in the country of Georgia continues to represent a threat to public and veterinary health . The current national classification of endemicity does not allow for an efficient implementation of control measures given the limited resources available locally . Previous human studies have focused on national reporting [39] , investigations of a single outbreak , or outbreaks across a short period of time [40] . In contrast , this study represents a fine scale spatial investigation of naturally acquired human anthrax during a decade of reporting . . We identified persistent clusters of disease at the community level that were associated with both environmental and anthropogenic factors . Establishing these sub-national estimates of human anthrax risk provides a crucial first step in implementing targeted public health interventions . The high rates of anthrax observed during the study period are likely due to a combination of factors resulting from decreased funding for public and veterinary health management following the collapse of the Soviet Union as well as changes to agricultural production [17] . Since the transition to independence in the 1990s , reliance on agriculture has grown with an increase in agricultural employment from 25% in 1990 to ∼55% in 2009 [18] , [41] , [42] . During this period agricultural decollectivization has marked a switch to individual ownership [5] , [43] . This has resulted in a greater number of individual animal holdings and a subsequent increase in peridomestic arrangements [43] . Greater contact among livestock and humans living in close proximity ( peridomesticity ) has been shown elsewhere to facilitate disease transmission [44] . The increases in reporting observed here may have been due to increased diagnostic capacity and awareness , although the issue of anthrax as a neglected , endemic disease [43] , along with funding cuts to public health facilities , has likely led to an under reporting in both humans and animals [45] . In Georgia the number of documented livestock cases are few , with the number of human cases out of proportion suggesting an anthropocentric reporting system [3] . Additionally , the absence of gastrointestinal cases of anthrax during the study period , given the large number of cutaneous cases , further highlights the probable under reporting of the disease [5] . Results from this study point to a primarily localized transmission of human anthrax identified by the strong spatial dependence of the Moran's I statistic at a relatively short distance ( 3000 meters ) ( Figure 3 ) . This is consistent with research that has described the local dynamics of infection related to the communal handling , butchering , or sharing of infected meat across short distances [14] , although imported cases from contaminated animal products have been documented elsewhere [46] . In Georgia , anthrax was spatially heterogeneous at the community level as evident by the presence of clustering ( Figure 4 ) . These findings are similar to studies in livestock/wildlife , which are often the source of human infection , that found variation in cases related to environmental constraints on the pathogen [42] , [47] , [48] . Reports on the spatial distribution of human cases are largely absent in the literature although in Kazakhstan human reporting was shown to be located in specific areas that comprised historical foci of the disease [49] . By expanding the maximum temporal window to 90% in the SaTScan space-time analysis we identified areas that represented anthrax persistence over time . The possibility persistence is likely due to a combination of ecological factors that permit pathogen survival and anthropogenic factors that accommodate disease transmission [20] , [48] , [50] . Soil pH showed a strong positive association with anthrax persistence , as documented elsewhere [51] . Additional environmental factors MIR and LST were also significant in the model . Alkaline soils with adequate moisture ( MIR ) and an ambient temperature above 15°C ( LST ) have been suggested as factors related to incubator areas for bacterium [8] , [52] , [53] . These factors are similar to previous studies that have modeled the potential distribution B . anthracis in the United States and Kazakhstan [7] , [8] , [50] . Soil pH has long been considered an important factor for spore survival with the bacterium preferring alkaline soils with a relatively high pH [7] . In Tanzania , a higher proportion of anthrax seropositive wildlife and domestic animals were identifed in areas with alkaline soils [8] , [10] . Arable land is limited in Georgia ( ∼17% of all land ) , leaving pasture/grasslands for livestock grazing in more alkaline soils [7] , [54] , [55] . Cattle density had a positive association OR 2 . 78 ( CI95%: 0 . 91 , 9 . 55 ) with persistence , however , it was not significant . Infection ratios comparing humans to livestock can vary widely , ranging from 1∶3 to 1∶16 as seen in Zambia [50] and the density of cattle may not be as important as the presence or absence of livestock . Human anthrax has been described as a disease of rural communities [43] . However , our study suggests that urban areas had >4 . 5 times the odds of being a persistent cluster when compared to rural areas . Health seeking behaviors of urban and rural populations differ , which can bias reporting due to access to healthcare facilities . In the Caucasus region , rural populations were shown to under-utilize the national healthcare leading to an under reporting of illnesses [47] . Anthrax infections in livestock , as seen elsewhere may also prompt owners to slaughter animals and bring the meat to market more quickly [6] . Urban areas often function as the market centers , hence contaminated livestock brought to market may increase the risk of exposure . Also , occupational exposures may differ in and around urban areas with animal processing facilities close to central markets . This study had several limitations that warrant discussion . The study was an ecological analysis that examined aggregate level data and may not represent the true associations of risk . Deriving census population as denominator data for small populations is often difficult , particularly in low resource settings . In this study , population totals for most communities were obtained from census records although several estimates were also obtained from gridded population data , which may skew the cumulative risk estimates and SaTScan results by either overestimating or underestimating the denominator figures . Model accuracy of the logistic regression was not evaluated with training and testing data due to the sample size of the persistent clusters identified . Although the disease was likely under reported , the reporting may have been biased towards urban areas with more access to healthcare facilities resulting in an under estimation of risk in rural areas . Furthermore , the method of classifying urban areas , while well documented in the literature , may in some instances incorrectly classify communities . Epidemiological data were not available on individuals , therefore individual level risk factors such as age , gender , and the source of infection were not considered in this study . Although anthrax is associated with specific factors on the landscape identified here , cultural and socio-economic factors not available in this study may also influence the occurrence of the disease . Research has reiterated the need for continued surveillance in order to distinguish between an intentional release and natural occurring disease [56] . This is especially true in light of the 2001 anthrax bioterrorist attack in the United States , which has renewed fears of its use as a weaponized biological agent [18] . Public health officials may be able to implement directed intervention strategies targeting high risk areas identified in this study which can include: training on proper carcass disposal , targeted livestock vaccination programs , education about the disease , and limiting occupational exposures through personal protective equipment [26] , [57] , [58] . While changing behavior related to zoonotic disease transmission may be difficult [26] , [59] , initiating targeted livestock vaccination in areas with persistence risk can help to control future outbreaks . Future research should focus on collecting epidemiologic data on individual cases while also further exploring factors associated with an increasing rate of disease in communities identified in this study .
|
Anthrax is a zoonotic bacterial disease that occurs nearly worldwide . Despite a large number of countries reporting endemic anthrax , persistence of the disease appears to be associated with specific ecological factors related to soil composition and climatic conditions . Human cases are most often associated with handling infected livestock or contaminated meat and most cases are in cutaneous form ( skin infections ) . Following the collapse of the Soviet Union , the country of Georgia has undergone major restructuring in land management and livestock handling and anthrax remains a serious public health risk . Few studies have evaluated the local spatial patterns of human anthrax . Here we identify areas on the landscape where human cutaneous anthrax persisted over the last decade . Persistence was found to be associated with both anthropogenic and environmental factors including soil pH and livestock density . These findings aid in the establishment of spatial baseline estimates of the disease and allow public health officials to adopt targeted anthrax control strategies , such as livestock vaccination campaigns and farmer education .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"bacterial",
"diseases",
"infectious",
"diseases",
"anthrax",
"geography",
"veterinary",
"diseases",
"social",
"and",
"behavioral",
"sciences",
"environmental",
"epidemiology",
"epidemiology",
"infectious",
"disease",
"epidemiology",
"zoonotic",
"diseases",
"spatial",
"analysis",
"spatial",
"epidemiology",
"population",
"biology",
"biology",
"veterinary",
"science",
"human",
"geography"
] |
2013
|
Evidence of Local Persistence of Human Anthrax in the Country of Georgia Associated with Environmental and Anthropogenic Factors
|
Insertions of parasitic DNA within coding sequences are usually deleterious and are generally counter-selected during evolution . Thanks to nuclear dimorphism , ciliates provide unique models to study the fate of such insertions . Their germline genome undergoes extensive rearrangements during development of a new somatic macronucleus from the germline micronucleus following sexual events . In Paramecium , these rearrangements include precise excision of unique-copy Internal Eliminated Sequences ( IES ) from the somatic DNA , requiring the activity of a domesticated piggyBac transposase , PiggyMac . We have sequenced Paramecium tetraurelia germline DNA , establishing a genome-wide catalogue of ∼45 , 000 IESs , in order to gain insight into their evolutionary origin and excision mechanism . We obtained direct evidence that PiggyMac is required for excision of all IESs . Homology with known P . tetraurelia Tc1/mariner transposons , described here , indicates that at least a fraction of IESs derive from these elements . Most IES insertions occurred before a recent whole-genome duplication that preceded diversification of the P . aurelia species complex , but IES invasion of the Paramecium genome appears to be an ongoing process . Once inserted , IESs decay rapidly by accumulation of deletions and point substitutions . Over 90% of the IESs are shorter than 150 bp and present a remarkable size distribution with a ∼10 bp periodicity , corresponding to the helical repeat of double-stranded DNA and suggesting DNA loop formation during assembly of a transpososome-like excision complex . IESs are equally frequent within and between coding sequences; however , excision is not 100% efficient and there is selective pressure against IES insertions , in particular within highly expressed genes . We discuss the possibility that ancient domestication of a piggyBac transposase favored subsequent propagation of transposons throughout the germline by allowing insertions in coding sequences , a fraction of the genome in which parasitic DNA is not usually tolerated .
Paramecium belongs to the ciliate phylum , a deep radiation of highly diverse unicellular eukaryotes . The hallmark of ciliates is nuclear dimorphism: each unicellular organism harbors two kinds of nuclei with distinct organization and function . A diploid “germline” micronucleus ( MIC ) undergoes meiosis and transmits the genetic information to the next sexual generation but is not expressed . A polyploid “somatic” macronucleus ( MAC ) contains a version of the genome streamlined for gene expression and determines the phenotype . A new MAC is formed at each sexual generation by programmed rearrangements of the entire zygotic , germline-derived genome , and the maternal MAC is lost . The MAC genome of P . tetraurelia has been sequenced [1] revealing a series of whole genome duplications ( WGDs ) in the lineage that provide a unique tool for evolutionary analyses . Ciliate genome rearrangements and their epigenetic control by non-coding RNAs have been recently reviewed [2]–[4] . In Paramecium , genome rearrangements involve ( i ) endoreplication of the DNA to about 800 haploid copies , ( ii ) imprecise elimination of genomic regions that contain , in particular , transposons and other repeated sequences , usually leading to chromosome fragmentation and ( iii ) elimination of Internal Eliminated Sequences ( IES ) by a precise mechanism . The accuracy of this process is crucial for IESs located within coding regions , to correctly restore open reading frames . The characterization of fewer than 50 IESs identified by cloning MIC loci [5] showed that they are short ( 26–883 bp ) , unique copy elements that are located in both coding and non-coding regions of the genome . The IESs are invariably flanked by two TA dinucleotides whereas only one TA is found at the MAC chromosome junction after IES excision ( Figure 1 ) . IESs have also been discovered by cis-acting mendelian mutations that prevent their excision , conferring a mutant phenotype [6]–[10] . The mutations in almost all cases were found in one of the flanking TA dinucleotides , which seem to be an absolute sequence requirement for IES excision . Extrapolation of the number of IESs found mainly in surface antigen genes led to the estimation that there could be as many as 50 , 000 IESs in the Paramecium genome . Such massive presence of unique copy IESs inserted in genes is not a characteristic of all ciliates . The estimated 6 , 000 IESs of the related oligohymenophorean ciliate Tetrahymena [11] are excised by an imprecise mechanism [12] , are usually multi-copy including recognizable transposons [13]–[15] and are rarely found in coding sequences [16] , [17] . Klobutcher and Herrick [18] first reported a weak consensus at the ends of 20 IESs from Paramecium surface antigen genes ( 5′-TAYAGYNR-3′ ) that resembles the extremities of Tc1/mariner transposons . These authors hypothesized a “transposon link” to explain the origin of IESs , suggesting that they are the decayed relics of a Tc1/mariner transposon invasion and that they are excised from the MAC DNA by a Tc1/mariner transposase encoded by a gene that has become part of the cellular genome [19] . In this model , IES excision represents the exact reversal of Tc1/mariner transposon integration into its TA target site with duplication of the TA dinucleotide , an evolutionary novelty that may have appeared more than once in the ciliate phylum . One problem with the model is that transposition catalyzed by Tc1/mariner transposases usually leaves a 2 or 3 bp “footprint” at the donor site [20] while IES excision is precise . A decisive step towards understanding the mechanism of IES excision and validating a transposon link for the origin of the IES excision machinery was the identification of a domesticated piggyBac transposase in Paramecium [21] . Baptized PiggyMac ( Pgm ) , the protein is encoded by the PGM gene which is expressed only late in sexual processes , at the time of genome rearrangements . Pgm , localized in the developing new MAC , was found to be required for the excision of all IESs tested and for the imprecise elimination of several regions containing transposons or cellular genes [21] . A similar piggyBac-derived transposase is found in Tetrahymena and is required for heterochromatin-dependent DNA elimination [22] . Since the Paramecium and Tetrahymena proteins appear to be monophyletic , based on a broad phylogeny of piggyBac transposases ( L . Katz and F . Gao , personal communication ) , the domestication event may have preceded the divergence of these two ciliates , estimated at 500–700 Ma ( million years ago ) [23] . Most significantly , the in vivo geometry of IES excision , initiated by staggered double-strand breaks ( DSBs ) that generate 4-base 5′ overhangs centered on the TA at both ends of the IES [24] , is fully compatible with the in vitro reaction catalyzed by a piggyBac transposase isolated from an insect [25] , whose target site is a 5′-TTAA-3′ tetranucleotide . piggyBac elements leave behind no scar when they jump to a new location: only ligation is required to join the fully complementary 5′ overhangs . Limited processing of 5′ and 3′ ends is further required for precise closure of the Paramecium IES excision sites since only the TA dinucleotides at the center of the 4-base 5′ overhangs are always complementary [24] , [26] . We report here a genomic approach to exhaustively catalogue the IESs in the Paramecium tetraurelia germline genome in order to study their evolutionary dynamics and seek evidence for a transposon origin of these elements . We obtained DNA highly enriched in un-rearranged germline sequences , from cells depleted in Pgm by RNA interference . Deep-sequencing of this DNA ( hereafter called “PGM DNA” ) allowed us to identify a genome-wide set of nearly 45 , 000 IESs , by comparing contigs assembled using the PGM DNA ( hereafter called “PGM contigs” ) with the MAC reference genome [1] . The hypothesis that Pgm is required for excision of all IESs was tested by genome-scale sequencing of a source of DNA from purified MICs [27] , providing validation of the IES catalogue . The evolutionary dynamics of the IESs was studied by exploiting the series of WGDs that have been characterized in Paramecium [1] . The study provides , to our knowledge , the first genome-wide set of IESs , in Paramecium or any ciliate , and provides new evidence that IESs have deleterious effects on fitness and that at least a fraction of IESs do derive from Tc1/mariner transposons that have decayed over time . The IES sequences evolve rapidly . The constraints we could detect concern their size distribution , suggestive of the assembly of a transpososome-like excision complex and a weak consensus at their ends , which resembles the extremities of Tc1/mariner elements . We discuss the possibility that ancient domestication of the Pgm transposase favored subsequent propagation of transposons throughout the Paramecium germline genome , by providing a mechanism for their precise somatic excision , therefore allowing insertions in coding sequences .
An overview of the strategy for identification of a genome-wide set of IESs is presented in Figure S1 . The first step was next-generation deep sequencing of DNA enriched in un-rearranged sequences , isolated from strain 51 cells that had undergone the sexual process of autogamy after depletion of Pgm protein by RNAi ( Figure S2 ) . In the absence of Pgm , the zygotic DNA is amplified but rearrangements are impaired . The sample that was sequenced contained a mixture of 60–65% un-rearranged DNA from the developing new MACs and 35–40% rearranged DNA from the fragments of the maternal MAC still present in the cytoplasm , as judged by Southern blot quantification of MIC and MAC forms at one locus ( Figure S3 ) . The PGM sequence reads ( Table 1 ) were mapped to the MAC reference genome of strain 51 ( see Materials and Methods ) , and putative IES insertion sites were defined as sites with a local excess of ends of read alignments ( pipeline MIRAA for “Method of Identification by Read Alignment Anomalies” ) . This excess of ends of alignments arises when a read contains a MIC IES junction , since only part of such a read can align with a MAC chromosome , either starting or ending at the IES insertion site , expected to be a TA dinucleotide . Using MIRAA , we identified 45 , 739 potential IES insertion sites . Essentially all ( 99% ) of the insertion sites contained a TA dinucleotide , even though this was not assumed by the pipeline . In order to obtain the sequence of the IESs , the paired-end PGM DNA sequence reads were assembled into contigs ( cf . Table S1 for assembly statistics ) and compared to the MAC reference genome assembly ( pipeline MICA for “Method of Identification by Comparison of Assemblies” ) . We looked for insertions in the PGM contigs with respect to the MAC reference assembly . Any insertion bounded by TA dinucleotides after local realignment was considered to be an IES . Using this pipeline we identified 44 , 928 IESs . The fact that 96% ( n = 43 , 220 ) of the IESs identified by MICA correspond to an IES insertion site identified by MIRAA ( Figure S1 ) testifies to the overall reliability of the procedure . Experimental validation of 6 IESs identified only by MICA and 17 insertion sites identified only by MIRAA was carried out by PCR amplification of an independent preparation of PGM DNA . The results ( Table S2 ) show that the 6 IESs and at least 12 of 17 insertion sites tested do correspond to the presence of an IES . Interestingly , among the IES sites identified only by MIRAA , we found 8 examples of a pair of IESs separated by one or only a few nucleotides ( in 5/8 cases , these tandem IESs are located in exons , a proportion similar to that found for the genome-wide IES set , see below ) . This case is not handled by the MICA pipeline since the initial global alignment with BLAT would have detected a single large insertion that would have been rejected by the local realignment filter , which requires the insertion to be flanked by TA dinucleotides . This is the first report of such closely spaced IESs , although nested IESs ( Figure 1 ) have been previously documented [8] . In order to see whether the set of 44 , 928 IESs is likely to be exhaustive , we looked for the 53 previously characterized IESs identified directly by cloning MIC loci in P . tetraurelia strain 51 cells ( Table S3 ) . All 53 previously cloned IESs were found , with the exception of one IES that had been assembled into the MAC reference genome and one IES form that represents use of an alternative boundary . In addition , two small IESs , each of which is nested within a larger IES , were found in PGM DNA but were not identified by our pipeline as IESs . Indeed , nested IESs can only be identified by time-course experiments or if the outer IES is retained in the MAC e . g . as the result of a point mutation [8] . Since 49 of 51 non-nested IESs were identified by MICA , the IES identification procedure has a sensitivity of at least 96% . The entire IES identification approach is based on the assumption that the excision of all IESs in Paramecium requires the Pgm domesticated transposase activity . In order to test this assumption , we sequenced inserts from a lambda-phage library constructed some 20 years ago [28] , using DNA from MICs that had been separated from MACs by Percoll gradient centrifugation [27] . This library has been extensively used to clone MIC loci with specific probes . Although the contigs assembled from the phage DNA reads only partially covered the MAC reference genome ( Table 1 ) , 98 . 5% of the 13 , 377 IESs that could be identified using the phage DNA and the MICA pipeline had also been identified using the PGM DNA . The difference of 1 . 5% is within the estimated sensitivity of the MICA pipeline . We conclude that all Paramecium IESs very likely require Pgm for excision , and that our data set does represent a genome-wide set of P . tetraurelia IESs . The genome-wide set of IESs has an overall G+C content of 20% , significantly lower than the 28% G+C content of the MAC reference genome [29] but comparable to the G+C content of intergenic regions ( 21% ) . The IESs are found in exons ( 76 . 8% ) , introns ( 5 . 4% ) and intergenic regions ( 17 . 8% ) , suggesting a nearly random distribution of IESs with respect to genes , since the MAC reference genome is composed of 76% exons , 3 . 2% introns and 20 . 8% intergenic DNA [1] . However , IESs are not randomly distributed along the chromosomes . Intriguingly , as shown in Figure S4 for the 8 largest MAC chromosomes , IESs tend to be asymmetrically distributed along MAC chromosomes . The MAC assembly ( 188 scaffolds >45 Kb constitute 96% of the 72 Mb assembly ) contains 115 telomere-capped scaffolds , varying in size from ∼150 Kb to ∼1 Mb , that are considered to represent complete MAC chromosomes . For 70 of these telomere-capped scaffolds , IESs display non-uniform distributions ( p<0 . 002 , median scaffold size 417 Kb ) while for the remaining 45 telomere-capped scaffolds , the IES distribution is uniform ( median scaffold size 275 Kb ) . Thus the larger the MAC chromosome , the greater the chance of observing a non-uniform IES distribution . The distributions for all scaffolds are easily visualized using the ParameciumDB [30] Genome Browser . The significance of the asymmetry in IES distribution is not clear , but might be related to the global organization of MIC chromosomes , currently unknown ( discussed in [29] ) . The genome-wide set of IESs covers 3 . 55 Mb ( mean IES size 79 bp ) , compared to 72 Mb for the MAC reference genome assembly . The IESs thus add about 5% to the sequence complexity of the part of the MIC genome that is collinear with MAC chromosomes . The total complexity of the PGM contigs ( after elimination of contigs with low PGM read coverage and high G+C content , assumed to represent bacterial contamination as confirmed in many cases by BLASTN matches against bacterial genomes ) is ∼100 Mb , however the use of a single paired-end sequencing library with small inserts ( ∼500 bp ) may have perturbed assembly of repeated sequences , possibly leading to underestimation of repeated sequence content . We infer that ∼25 Mb of germline-specific DNA corresponds to the imprecisely eliminated regions located outside of the MAC-destined chromosomes i . e . the part of the MIC genome that is not collinear with MAC chromosomes . We have not further characterized this fraction of the PGM DNA . However , we did identify the first germline P . tetraurelia Tc1/mariner transposons ( Figure S5 ) , by using the phage-lambda library of MIC DNA [28] to walk past the end of MAC scaffold_51 , which bears the subtelomeric 51G surface antigen gene [31] . In all , 5 phage inserts and 4 cloned PCR products corresponding to part or all of different copies of the element downstream of the 51G surface antigen gene , named Sardine , were sequenced ( EMBL Nucleotide Sequence Database accession numbers HE774468–HE774475 ) and a consensus for the ∼6 . 7 Kb transposon was constructed ( Figure S5 and Text S1 ) . The ends of the Sardine copies contain intact or partially deleted 425 bp terminal inverted repeats ( TIRs ) which are themselves palindromic , containing a unique , oriented region nested within outer inverted repeats ( Figure S5 ) . Sardine contains up to 4 ORFs . One ORF is a putative DD35E transposase of the IS630-Tc1 family , like the DDE transposases of the TBE and Tec transposons found in stichotrich ciliates [32] . Another ORF , as in Tec transposons [33] , encodes a putative tyrosine recombinase . The other two ORFs are hypothetical , though ORF2 shows some similarity ( 31 . 7% identity and 55 . 4% similarity over 202 aa ) to the hypothetical ORF1 of the Tennessee element from P . primaurelia [34] . One of the Sardine copies ( copy S6 ) is interrupted by the insertion , within the putative tyrosine recombinase gene , of a different but similar element , named Thon ( French for “tuna” ) , which also contains a DD35E transposase , a tyrosine recombinase , possibly the two hypothetical ORFs , and palindromic TIRs of ∼700 bp ( Figure S5 ) . For a handful of IESs , it has been shown experimentally that they are single copy elements [5] . In order to see whether this is generally the case , we looked for all IESs present in more than 1 fully identical copy ( 100% sequence identity ) . We found 44 , 210 IESs to be unique copy ( 98 . 4% ) . We examined all IESs present in 2 or more identical copies and found 39 cases of duplicate IESs as a result of errors in assembly of the MAC reference genome that had led to small , partially redundant scaffolds ( 4% of the MAC assembly is contained in scaffolds <45 Kb and some of these are partially redundant with the chromosome-size scaffolds [1] ) . The rest of the 319 IESs found in 2 copies were inserted in homologous genomic sites and appeared to be the result of recent segmental duplication or gene conversion . The 23 cases of IESs found in 3 to 6 copies correspond to expansion or recombination of repeated sequences such as tetratricopeptide repeat ( TPR ) domains or WD40 repeats . We performed an all by all sequence comparison of the IESs and of their flanking sequences to see whether we could identify homologous IESs inserted at non-homologous sites in the genome . As shown in Table 2 , we were able to identify 8 clusters of 2 to 6 IESs that share significant homology ( BLASTN E-value <10−10 ) over at least 85% of their length , inserted in non-homologous sites ( cf . Text S2 for the alignments ) . Moreover , we found significant nucleotide identity ( E-value 9×10−57 for the best match; nucleotide identity between 68 and 78% for the HSPs ) between the IESs of cluster 5 and one of the Tc1/mariner-like transposons identified using the phage library ( Thon , Figure S5 ) . This is a strong indication that these IESs are derived from recently mobile elements . However , the IES sequences of this cluster correspond to a single palindromic TIR . This might reflect assembly problems given use of a single insert size for the paired-end sequencing , either because these IESs contain sequences repeated elsewhere in the genome or because the Thon TIRs are large ( ∼700 bp ) and palindromic so that the assembly might have jumped from one TIR to the other deleting the rest of Thon . We therefore used a long-range PCR strategy capable of amplifying large DNA fragments containing each of the IESs to verify their size and attempt to obtain sequences ( detailed in Text S3 ) . Amplification products of the expected sizes were obtained for all of the IESs from cluster 5 , making it unlikely that these IESs correspond to a complete Thon element that had failed to be assembled from the paired-end sequencing reads . Three IESs were chosen for sequencing , and the sequences of the corresponding PCR products confirmed the IESs , indicating that they had been correctly assembled . Identification of 6 IESs ( at non-homologous genomic sites ) that share sequence identity with a P . tetraurelia Tc1/mariner solo TIR argues that at least a fraction of IESs do originate from Tc1/mariner-like elements . We therefore adopted a complementary strategy , using the PFAM-A library of curated protein domains to search for domain signatures in the genome-wide set of IESs . Matches at a BLASTX E-value cutoff of 1 were inspected visually to filter out matches with PFAM-A protein domains from Paramecium and matches owing to compositional bias ( high A+T content ) . This left 6 IESs , ranging in size from 2416 to 4154 bp , with a DDE_3 ( PFAM accession number 13358 ) DDE superfamily endonuclease domain characteristic of IS630/Tc1 transposons . The peptides encoded by the IESs were subjected to an HMM search of the PFAM-A hmm profiles ( http://pfam . sanger . ac . uk/search ) for confirmation of the conserved residues and to validate the statistical significance of the match ( E-values of 0 . 02 to 2 . 1×10−15 for the 6 peptides ) . The IESs were aligned with MUSCLE and a neighbor-joining tree grouped 4 of them together with good bootstrap values ( not shown ) . The 4 IESs were used to search for sequence similarity with the genome-wide set of IESs and this allowed identification of 28 IESs ranging in size from 1251 to 4154 bp ( Table S4 ) . The IESs were aligned to provide the consensus sequence for 2 distinct Tc1/mariner-like 3 . 6 kb transposons from the same new family , baptized Anchois ( Anchovy ) . Manually adjusted alignments used to reconstruct the AnchoisA and AnchoisB elements , consensus sequences and annotation are provided in Text S1 . Alignment of the DDE domains of the reconstituted Anchois transposons with the DDE domains from bacterial IS630 elements , invertebrate Tc1 transposons and all known ciliate Tc1/mariner elements indicates that the Anchois transposase belongs to the IS630/Tc1 subfamily ( Figure 2A ) . Unlike Thon and Sardine but like the P . primaurelia Tennessee element , Anchois TIRs are short and lack internal palindromes , moreover Anchois does not contain a putative tyrosine recombinase . Anchois has 2 hypothetical ORFs in addition to the DDE transposase ( Figure 2B; Text S1 ) . The ORF2 of Anchois displays homology to ORF2 of Sardine ( 36 . 2% identity and 56 . 2% similarity over 210 aa ) and to ORF1 of Tennessee . Interestingly , for 6 of the 28 IESs that initially identified the copies of Anchois , the Anchois TIRs do not correspond to the extremities of the IES , raising the possibility of Anchois insertions within pre-existing IESs . The discovery of the Anchois elements and the fact that several IESs appear to be full-length copies , provides a strong , direct link between IESs and transposons . The size distribution of the genome-wide set of IESs is shown in Figure 3A , for the 93% of the IESs that are shorter than 150 bp . The most remarkable feature is a periodicity of ∼10 bp , which corresponds to the helical repeat of double-stranded DNA . The first peak of the size distribution has maximal amplitude at 26–28 bp and includes 35% of all identified IESs . The abrupt cutoff at 26 bp represents the minimum IES size . A second peak appears to be forbidden and contains only a few IESs . The following peaks are centered at approximately 45–46 , 55–56 , 65–66 bp etc . and the distance between these peaks is best fit by a 10 . 2 bp sine wave ( not shown ) . At the far end of the spectrum , 95 of the IESs are between 2 and 5 Kb in size . Similar periodic size distributions are found for IESs inserted in coding sequences and for IESs inserted in non-coding sequences ( Figure S6 ) . This indicates that the constraint on the distance between IES ends is an intrinsic property of the IESs and is not related to the locus in which they are inserted in the genome . Whatever their size , the IESs adhere to the weak , Tc1/mariner-like end consensus first reported for 20 IESs located in surface antigen genes [18] , as illustrated in Figure 3B for the whole set . Differently sized subsets of the IESs all display essentially the same end consensus ( data not shown ) . We further examined constraints on IES size and sequence by evaluating IES conservation with respect to the 3 WGDs in the Paramecium lineage . We used the large number of paralogs ( hereafter termed “ohnologs” ) of different ages ( Table 3 ) that could be identified for each of the WGD events [1] to ask whether IESs are present , at the same position relative to the gene coding sequences , in ohnologs of the different WGD events . This analysis makes the assumption that IES insertions are rare events so that if IESs are present at the same position in ohnologous genes , then they must have been acquired before the WGD and can be considered to be “ohnologous” IESs . As shown in Table 3 , we found 84 . 5% , 23 . 2% and 5 . 9% conservation of IESs with respect to the recent , intermediate and old WGDs respectively . For comparison , more than 99% intron conservation was found for 1 , 112 pairs of genes related by the recent WGD [35] . This indicates that the dynamics of IES insertion or loss over evolutionary time is relatively fast compared to that of introns . The only phylogenetic study of IESs , carried out for two loci in a few different stichotrich ( formerly called hypotrich ) ciliates , which are very distantly related to Paramecium , also concluded that the intragenic IESs in those species evolve very rapidly [36] . We found that the ohnologous IESs related by the recent WGD are highly divergent in sequence . In more than 90% of cases , the sequence identity was too low for detection by BLASTN ( E-value threshold of 10−5 ) . This high level of sequence divergence is consistent with the pattern expected for neutrally-evolving non-coding regions , since the average synonymous substitution rate measured between ohnologous genes derived from the recent WGD is about 1 substitution per site [1] . However , if we compare the lengths of IESs that are conserved with respect to the recent WGD ( Figure 4A ) , for ∼55% of the pairs , both IESs are found in the same peak of the IES size distribution . The honeycomb appearance of the plot ( Figure 4A ) , with off diagonal cells that result from ohnologous IESs in different peaks of the distribution , underscores the strong evolutionary constraint that is exerted on IES size . In order to investigate the rate of IES insertions and losses during the evolution of the Paramecium lineage , we examined gene families , which we call “quartets” , for which all 4 ohnologs issued from duplication of an ancestral gene at the intermediate and then the recent WGD are still found in the present day genome . Of the 1350 such quartets identified in the MAC genome [1] , 878 contain at least one IES in at least one of the 4 duplicated genes . We evaluated the conservation of IESs at the same position with respect to the coding sequence for all members of each quartet ( Figure S7 ) , and identified 2126 IES groups , each group containing an IES conserved either in all 4 genes ( N1111 = 190 ) , in 3 genes ( N1110 = 64 ) , in 2 genes on the same intermediate WGD branch ( N1100 = 1304 ) , in 2 genes on different branches ( N1010 = 10 ) or in only one of the 4 genes ( N1000 = 558 ) . Under the assumption that two IESs present at the same location in ohnologous genes derive from a single ancestral IES ( i . e . the probability of two insertion events occurring at the same site after a WGD is considered negligible ) , and that the rate of IES losses has remained constant , it is possible to estimate the rate of IES gain during the evolution of the Paramecium lineage ( the model is developed in Text S4 ) . The quartet analysis is fully consistent with a model whereby IES acquisition has been ongoing since before the intermediate WGD ( 15% of the IESs predating this WGD ) , with a peak in the period between the intermediate and the recent WGD events: 69% of IESs were acquired during the interval between these two WGDs , vs . 16% during the period since the recent WGD , which corresponds to about the same evolutionary time . Genome-wide IES data for other Paramecium species will be necessary in order to test the assumption of a constant rate of IES losses . However , even if we relax this assumption ( i . e . rates of IES losses are allowed to vary over time ) , the model still strongly rejects the hypothesis that all IESs were acquired before the intermediate WGD ( cf . Text S4 ) . Thus , with the presently available data and biologically reasonable assumptions , we conclude that IESs have been acquired in all 3 of the time periods delimited by the intermediate and recent WGD events . We compared the cumulative size distributions of the N1111 , N1100 and N1000 IESs ( Figure 4B ) . The N1111 IESs , which must have been acquired before the intermediate WGD , are much shorter than the IESs of the two other samples , with almost 80% of the IESs in the first peak , compared to 20% for N1100 IESs , which may mainly result from IES acquisition after the intermediate but before the recent WGD , and only 16% for N1000 IESs , at least some of which may have been acquired since the recent WGD . In addition , the curves are significantly shifted with respect to each other , in particular , 30% of the N1000 IESs are larger than 150 nt , compared to scarcely any IESs larger than 150 nt for the two other samples . This analysis shows that the older an IES , the shorter it is likely to be , consistent with a decay process involving progressive shortening of IESs by accumulation of small deletions , in addition to the accumulation of point mutations . Quartet analysis is restricted to IESs in genes that have been retained in 4 copies ( fewer than 10% of all IESs ) . Similar distributions of IES size are found if we consider all ohnologous IESs ( 45% of all IESs , cf . Table 3 ) . IESs conserved with respect to the intermediate WGD ( 76% of IESs in first peak ) are significantly shorter than IESs conserved only with respect to the recent WGD ( 30% of IESs in the first peak ) ( data not shown ) . The size distribution of IESs conserved with respect to the old WGD is poorly determined because of the small number of conserved IESs ( Table 3 ) , which are moreover often in genes that have undergone recent gene conversion judging from the nucleotide divergence of the ohnologs ( data not shown ) . It is therefore uncertain that IESs were present in the genome before the old WGD , consistent with the absence of TA-bounded IESs in Tetrahymena , which diverged from Paramecium after the old WGD event [1] . Since we found essentially no IESs shorter than 26 bp , it seems likely that some mechanism ( s ) other than decay of the sequence through internal mutations and deletions is responsible for the complete loss of an IES . In order to explore this question , we examined case by case , using both nucleotide and conceptual protein alignments , all of the N1110 quartet IES groups ( n = 64 ) , which are most parsimoniously explained by insertion of an IES before the intermediate WGD followed by loss of an IES after the recent WGD . We examined the raw read alignments and PGM and phage contigs in order to be sure that there was sufficient read coverage and no evidence suggesting presence of an IES at any site of putative IES loss . We found 4 different explanations for the quartet triplets: precise loss of the fourth IES ( n = 17 ) , gain of the third IES by gene conversion between intermediate WGD ohnologs ( n = 1 ) , recruitment of the fourth IES into the exon sequence ( n = 6 ) , and deletion of the region that encompasses the fourth IES ( n = 23 ) , often testifying to the formation of a pseudogene . In addition , we found 5 errors in IES detection ( the fourth IES probably exists as it can be found in the phage contigs or is predicted by the MIRAA pipeline ) . In the remaining cases ( n = 12 ) , annotation or alignment problems made it difficult to conclude . The observation of 17 cases of precise loss of an IES from the germline DNA raises the possibility that there is a mechanism for conversion of a MIC locus to the IES-free form using a MAC genome template . However , we cannot rule out the possibility that IESs can be precisely excised from the MIC DNA , and therefore lost , by the same Pgm-dependent mechanism as that involved in MAC genome assembly . The analysis of sequence variability in the polyploid ( 800n ) MAC genome , carried out by comparing the MAC assembly representing a “consensus” sequence with the 13× Sanger sequencing reads used to build the assembly , revealed nearly 2000 “TA-indels” that were presumed to be produced by the IES excision machinery and to reflect excision errors [29] . As shown schematically in Figure 5 , “residual” TA-indels ( n = 739 ) , that were suggested to represent occasional retention of IESs on some macronuclear copies , were absent from the assembly ( “major” form in Figure 5 ) , but present in at least one sequence read ( “minor” form ) . For 689 of the residual TA-indels ( 93% ) , we found an IES at the corresponding site in the genome . Interestingly , in 134 cases ( 19 . 4% ) , the TA-indel was shorter than the IES and case by case inspection indicated that most of these TA-indels may be products of IES excision that used an alternative IES boundary located within the IES ( Figure 5 ) . In this case , the TA-indel would only correspond to part of a larger IES . A few cases of use of an alternative IES boundary that may confer a mutant phenotype have been reported [7] , [37] . “Low frequency” TA-indels ( n = 1090 ) , previously suggested to represent excision of MAC-destined sequences [29] , were present in the assembly ( major form , Figure 5 ) , but absent from at least one sequence read ( minor form ) . We could not look for the “low-frequency” TA-indels directly among the genome-wide set of IESs , since they are part of the MAC genome assembly . However , we examined the ends of the low-frequency TA-indels and found 249 cases ( 23% ) where the TA dinucleotide at one of the ends corresponds to the insertion site of an IES in the genome-wide set ( Figure 5 ) , indicating that the TA-indel was generated by use of an alternative IES boundary located outside of the IES . The whole of the analysis supports the previous conclusion [29] that TA-indels are products of the IES excision machinery . The high incidence of alternative boundaries in both classes of TA-indels , revealed by comparing them with the genome-wide set of IESs , strengthens the previous conclusion [29] that TA-indels reflect IES excision errors ( see below ) . Thus TA-indels cannot be considered to be IESs in the absence of further experimental support . IESs are tolerated in coding sequences and evolve under a strong constraint on their size and end-consensus , properties that are presumably important for their precise and efficient excision . However , the excision machinery can commit errors , as revealed by TA-indels ( cf . above ) and by the use of alternative IES boundaries [7] , [37] . We therefore looked for evidence that the rate of excision errors is high enough to represent a fitness burden for the organism . First , only 47% of genes contain at least one IES , and the IESs are less represented in strongly expressed genes . Figure 6 shows the density of IESs in genes as a function of gene expression level determined by microarray experiments [38] , [39] . The density varies from about 0 . 7 IESs per Kb ( i . e . an IES on average every 1 . 4 Kb ) in genes with low expression to less than 0 . 3 IESs per Kb ( i . e . an IES on average every 3 . 3 Kb ) for the genes with the highest expression . The inverse correlation observed across all levels of expression indicates that IESs are less-well tolerated the more a gene is expressed . Second , IESs inserted in protein-coding exons display a characteristic bias in their size . There is a statistically significant deficit in IESs whose length is a multiple of 3 , compared to IESs found in non-coding regions . Furthermore , this bias is only found for 3n IESs that do not contain a stop codon in phase with the ORF of the upstream coding sequence ( Table 4; cf . Table S5 for a more detailed analysis ) . A similar 3n bias was reported for introns in eukaryotic genomes , and experiments in Paramecium showed that the Nonsense Mediated Decay ( NMD ) pathway destroys mRNAs containing unspliced introns , provided the intron retention leads to a premature stop codon [35] . Retention in mRNA of a 3n stopless intron would not be detected by NMD and therefore could lead to translation of potentially harmful proteins , explaining the deficit in 3n stopless introns . The fact that IESs display a similar deficit suggests that the rate of IES retention is high enough to represent a fitness cost , so that IESs in exons are under selective pressure to be detected by NMD in case they are retained in the MAC genome . We were able to test this hypothesis by looking at the size bias for IESs located in the exons of the 25% of Paramecium genes that are the most highly expressed hence subject to the strongest selective pressure . As shown by the last 2 lines of Table 4 ( samples Q1 and Q4 ) , the deficit in 3n IESs is the greatest for the IESs found in the most highly expressed genes ( 28 . 3% ) , where IES retention would be the most deleterious .
Previous studies of Paramecium IESs all relied on a small reference set of about 50 IESs . For the first time in any ciliate genome , in so far as we are aware , we have carried out an exhaustive identification of IESs . Since it is not yet possible to isolate Paramecium MICs in the quantity and of the purity required for genomic sequencing , we relied on nuclear DNA isolated from cells depleted in Pgm , the domesticated transposase required for introduction of the DSBs that initiate IES excision [21] . We fortunately were able to use the only genomic library ever made from purified MICs [28] – but heavily contaminated by bacterial DNA – to obtain genome-scale evidence that Pgm is required for excision of all Paramecium IESs and to estimate that our IES reference set includes ∼98 . 5% of all IESs . Although this IES reference set will prove useful for a variety of studies , it is important to keep two things in mind . First , the IES definition used here is necessarily a genomic definition involving comparison of MIC and MAC sequences . Our procedure does not allow identification of nested IESs ( unless the external IES is retained in the MAC ) , or of any IES located in part of the MIC genome that is not collinear with MAC chromosomes . The complexity of the assembled PGM DNA is almost 100 Mb , although we could not properly assemble repeated sequences . We thus estimate that at least 25% of the germline is not collinear with the MAC chromosomes , and might contain unique copy IESs or transposons , the excision of which could only have been detected if the flanking region were retained in the MAC . Second , this reference set does not provide information about the variability in IES excision patterns that might exist between different , though genetically identical , cell populations . Many IESs are under maternal , epigenetic control [40] , [31] , [41] . The genome scanning model [42] posits that every time Paramecium undergoes meiosis , the scnRNA pathway compares the maternal MIC , in the form of 25 nt scnRNAs [43] , with the maternal MAC , in the form of long non-coding transcripts [44] . The scnRNAs that cannot be subtracted by base pairing with the long maternal transcripts are licensed for transport into the new developing MAC [45] where they target homologous sequences for elimination , probably via deposition of epigenetic marks on the chromatin ( cf [3] , [4] for recent reviews of genome scanning in Paramecium and Tetrahymena ) . The scnRNA pathway in theory provides a powerful defense mechanism against transposons that invade the germline and can explain the molecular basis of alternative MAC rearrangement patterns that are maintained across sexual generations [31] , [40] , [41] , [46] , [47] . Hence the following caveat: any genome-wide set of IESs is identified with respect to a particular MAC reference genome sequence . There can be no “universal” IES reference set for the species . Since IESs can be a source of genetic variation as discussed in [48] , the IES catalogue we have established will make it possible to study this variation , for example by surveying IES retention in the MACs of geographic isolates and in stocks that have been experimentally subjected to different types of stress . The remarkable sinusoidal distribution of IES sizes retained by evolution reflects strong constraint on the distance between IES ends . We assume that the selection is exerted through the excision mechanism , since the retention of an IES in the MAC can impair gene function . An IES that cannot be efficiently excised is expected to be counter-selected . We propose an interpretation of the IES size distribution based on its similarity with data generated by “helical-twist” experiments , which have provided evidence of DNA looping between distant protein-binding sites in various , mainly prokaryotic , DNA transaction systems ( transposition , gene control , replication initiation , site-specific recombination , etc . reviewed in [49] ) . In these experiments , the distance between transposon ends [50] , [51] , repressor binding sites [52]–[55] or site-specific recombination sites [56] is varied , on plasmids or on the bacterial chromosome , and the activity of the system is measured in vivo . The observed periodicity in the length-dependence of the activity corresponds to the helical repeat of the DNA , since the same face of the double helix must interact with the protein at each end , and given the prohibitive energetic cost of twisting the double helix to fit the binding site to the protein . This is especially true for DNA fragments whose size is close to the persistence length of double stranded DNA ( ∼150 bp ) or shorter . The persistence length , a physical measure of the bending stiffness of a polymer in solution , is the length above which there is no longer a correlation between the orientation of the ends of the molecule . For DNA longer than its persistence length , it becomes possible for the 2 ends to encounter each other to form a loop , without any external intervention . Almost all ( 93% ) of the IESs in the genome are shorter than the persistence length of DNA . The size distribution , which appears as a series of regularly spaced peaks , can be decomposed into three parts . The largest peak is centered on 28 bp but displays an abrupt minimum size cutoff at 26 bp . A second peak seems to be of forbidden size . Finally , there follow a series of peaks that are best fit by a sine wave with a ∼10 . 2 bp periodicity . In the helical-twist experiments , the amplitude of the measured biological activity peaks tends to decrease with decreasing distance between interacting sites . However , for the IES size distribution , the decay of IESs over time imposes the opposite tendency: the peak heights increase as IES size decreases . Our working model for assembly of an active IES excision complex is shown in Figure 7 . We propose that , starting at the third peak ( 44–46 bp ) , the IESs assemble into the excision complex by forming a double-stranded DNA loop compatible with presentation of the same face of the double helix to the Pgm endonuclease at both IES ends . The near absence of the second peak , the minimum IES size of 26 nt and the 13 bp size of each piggyBac TIR [25] lead us to suggest that the IESs in the first peak are able to assemble an active excision complex without formation of a DNA loop . The IESs in the nearly absent second peak would not be efficiently excised , as they would be too short to form a DNA loop and too long to form an active excision complex without a DNA loop . Molecular analysis of the IES excision mechanism supports the involvement of such a transpososome-type excision complex . First , the domesticated Pgm transposase , which has retained the catalytic site of piggyBac transposases [21] , is very likely to be the endonuclease responsible for the cleavage reaction , involving the introduction of DSBs at each end of the IES [24] . Second , for IESs larger than 200 bp , covalently closed circular molecules containing the excised IES have been detected as transient intermediates during MAC development [57] . Third , if one end of an IES bears a mendelian mutation in the TA dinucleotide , no DSB occurs at either end of the IES . This indicates that the two IES ends must interact , directly or indirectly , before cleavage can occur [58] . It is worth noting that “canonical” TIRs of cut-and-paste transposons are often bipartite . They are composed of an internal sequence motif recognized and bound by the transposase , and of a few nucleotides at the termini that constitute the DNA cleavage site [59] . The obligatory conservation of a TA dinucleotide at IES ends is indicative of a requirement for DNA cleavage but is not sufficient for specific recognition , even if we take into account the weak consensus over the 6 internal nucleotides . The lack of a sufficiently long conserved motif in IESs makes it unlikely that Pgm recognizes IESs by binding to a specific sequence . For IESs under maternal control [31] , it is currently thought that Pgm is recruited to its substrate via epigenetic marks deposited on the chromatin by the scnRNA pathway [3] , [21] , [42] . The picture of an IES excision complex that emerges from these considerations , which must of course be tested biochemically , requires very short pieces of DNA to form loops ( Figure 7 ) . Proteins that bend DNA , such as HMG proteins [60] , could be involved . What is quite remarkable here , beyond the fact that evolution has performed such a nice “helical-twist” experiment , is that the DNA loops might be as short as ∼45 bp , shorter than almost any reported case of DNA looping . The minimal in vivo value reported for cut-and-paste bacterial transposons is 64–70 bp [50] , [51] and this is also the minimum size reported for HMG assisted DNA loop formation in vitro [60] . The only indication of shorter loops comes from detection of a minor peak of activity in vivo and in vitro for ∼50 bp DNA loops in the E . coli Hin invertasome , provided that invertasome assembly occurs in the presence of HU , a bacterial nucleoid protein that bends DNA [56] . Given the unusually high A+T content of IESs ( 80% ) , local melting might favor the deformations in the double helix required to make the very small looped structures of the postulated IES excision complex . Ciliate MICs have long been recognized as safe havens for transposons , since removal of the transposons from the somatic DNA during development would decrease the burden on host fitness , as discussed in [19] . Our study provides the first global vision of IESs in any ciliate germline and provides strong support for the “transposon link” hypothesis that present day IESs are remnants of transposons [18] , [19] . Although we do not yet have a complete picture of the transposon landscape of the P . tetraurelia germline genome , we have identified 3 families of Tc1/mariner elements , with 2 quite different structures . The Thon and Sardine transposons have long , palindromic TIRs , a tyrosine recombinase and a DDE transposase characteristic of the IS630/Tc1 subfamily , with a short spacer ( 32 aa ) between the 2nd and 3rd catalytic residues . This clearly distinguishes these transposons from the piggyBac family characterized by a long spacer and a DDD catalytic triad . The IESs related to these elements that we were able to identify appear as solo TIRs . Given the presence of repeated , palindromic subsequences in each TIR , we can speculate that the solo TIRs result from recombination between short direct repeats present within the complex TIRs , as proposed to explain the incidence of solo LTRs derived from LTR retrotransposons in the genomes of some organisms [61] , [62] . The other transposon family we have identified , Anchois , is characterized by much shorter TIRs which do not contain internal palindromes , a similar DDE transposase and the absence of a tyrosine recombinase . This structure is similar to that of the P . primaurelia Tennessee transposon [34] . In the case of Anchois , we could find a number of IESs that appear to correspond to the entire transposon or large portions of it , including IESs with a recognizable but degenerate DDE transposase ORF . It is possible that we have only scratched the tip of the iceberg since the germline genome is expected to contain other mobile elements . Indeed , we were able to identify 8 clusters of homologous IESs inserted at non-homologous genomic sites , suggesting recent mobility , and one of these clusters turned out to consist of IESs that are solo TIRs of the Thon element . The other clusters could be the remains of as yet unidentified elements . Both the Thon and the Anchois IES homologies were detected among the largest IESs in the genome-wide set ( i . e . the 380 IESs >500 bp ) , and for none of them could we detect ohnologous IESs from the recent WGD , an indication that these IESs were recently acquired . Since over 90% of present day IESs have decayed to very short sizes ( <150 bp ) it is not surprising that internal transposon motifs can no longer be recognized . These very short IESs nonetheless display the short degenerate Tc1/mariner end consensus . The existence of this consensus at IES ends may testify to their evolutionary transposon origin . This end consensus would eventually have become a requirement for efficient cleavage by the IES excision machinery . We can imagine two instances of such convergent evolution: i ) other families of mobile elements could be eliminated by the PiggyMac-dependent mechanism and ii ) genomic sequences that adhere to the end consensus could be excised just like IESs . We conclude that at least a fraction of IESs are decayed Tc1/mariner transposons , and we consider highly probable that some IESs are derived from other mobile elements . Since IES excision is not 100% efficient , IES insertions are in general deleterious , consistent with the different kinds of selective pressure we have observed: ( i ) a constrained IES size distribution likely reflecting assembly of the excision complex; ( ii ) a bias against IESs that do not lead to premature stop codons in case of IES retention in the MAC; ( iii ) an inverse correlation between IES insertions and gene expression level . IESs can in addition be considered to constitute a mutational burden , in the same way as introns are considered to constitute a mutational burden in intron-rich eukaryotic genomes [63] , since IESs are present in large number in Paramecium , and any mutation in a flanking TA dinucleotide abolishes IES excision . Nonetheless , the system can give rise to beneficial new functions , as attested by use of the IES excision machinery to provide a regulatory switch for mating type determination ( D . Singh , personal communication ) . Since IESs are in general deleterious and constitute a fitness burden for the organism , and since we have detected cases of probable clean IES loss from the germline DNA suggesting that a mechanism exists for precise IES excision in the MIC , we may ask why Paramecium has any IESs at all . This question can be easily answered if we consider that IESs arise from selfish genetic elements ( SGEs , defined as elements – typically transposable elements or viruses – that can enhance their own transmission relative to the rest of the genome , with deleterious or neutral effects for the host [64] ) . The number of IESs reflects the balance between the number of IES insertions ( e . g . invasion by SGEs that subsequently decayed to become unique-copy IESs ) and the strength of selection against these insertions , which either prevents fixation of new insertions in the population or favors loss of already fixed insertions . This genetic conflict is mediated by an “arms race” between SGEs and the host as discussed by Werren [64] . In all kingdoms of life , non-coding RNAs are used to defend host genomes against parasitic nucleic acids , as exemplified in eukaryotes by small RNA pathways involved in protection against viruses or in silencing transposons to ensure integrity of the germline genome [65]–[67] . In ciliates , nuclear dimorphism provides the potential for an additional layer of protection by physically separating the chromosomes that store the genetic information from the rearranged chromosomes that express the genetic information . Additional host defense machinery providing precise excision of transposons/IESs from somatic DNA , might have allowed the invasion of a fraction of the genome in which SGEs are not usually tolerated , namely the coding and regulatory sequences required for gene expression . In the case of Paramecium , Pgm domestication has provided the mechanism for precise excision of TA-bounded insertions from the somatic DNA , allowing transposons/IESs to be cleanly excised from genes in the MAC . Since this would reduce the fitness burden caused by transposition , we presume that it allowed transposons to spread throughout the MIC genome . Recognition of the IESs is however ensured by the scnRNA pathway [3] , itself an example of the more ancient mechanism of small RNA-based host immunity against foreign nucleic acids , and this epigenetic recognition may in part explain the less than 100% efficiency of IES excision . In Tetrahymena , which has both a scnRNA pathway and domesticated piggyBac-like transposases [4] , [22] , only excision of intergenic IESs has been studied for the moment and use of heterogeneous cleavage sites was found . This imprecise excision would not be compatible with insertion in genes since gene expression would be compromised . Tetrahymena has only about 6 , 000 IESs and indeed , they are not usually found within genes [17] . Why doesn't Tetrahymena have intragenic IESs ? We can only speculate that a Tc1/mariner invasion after the divergence of Paramecium and Tetrahymena was instrumental in the evolution of a precise excision mechanism in Paramecium , necessary for spread of these elements throughout the genome . In support of this hypothesis , a recent genome-scale identification of hundreds of Tetrahymena IESs [17] revealed a new class of TTAA-bound IESs that are precisely excised . They were found to contribute 3′ exons to genes that are expressed from the zygotic genome during genome rearrangements . These elements might be derived from piggyBac transposons , which have TTAA target sites , and perhaps testify to the ancient piggyBac invasion that led to domestication of the transposase . A contrasting situation is found in some stichotrich ciliates . The stichotrich ciliates are very distantly related to the oligohymenophorean ciliates and are characterized by highly fragmented somatic genomes consisting of nanochromosomes that usually bear a single gene . Intragenic IESs are more abundant in the germline genomes of Oxytricha and related strichotrichs than in Paramecium , with an estimate of at least 150 , 000 IESs per haploid genome [68] . Both single-copy IESs and transposons are precisely excised and the precise IES excision is assured by guide RNAs transcribed from the maternal MAC [69] , which are even capable of re-ordering the scrambled MAC-destined gene segments that occur frequently in Oxytricha and related stichotrichs [70] . There is also evidence that the endonuclease required for cleavage in Oxytricha is actually a transposase from germline TBE transposons [71] . However , there is currently no evidence for a scnRNA pathway specialized in the control of DNA elimination , although gene silencing by RNAi in Oxytricha testifies to the presence of small RNA machinery [69] . Thus the high precision and fidelity of the guide RNA mechanism for genome rearrangements in Oxytricha spp . seems to have tipped the balance even further in favor of intragenic IES insertions . The case of Euplotes , a stichotrich ciliate distantly related to Oxytricha and probably lacking scrambled genes , merits special attention . Beautiful work carried out by the Jahn and Klobutcher labs in the 1990s showed ( i ) the existence of high copy number Tc1/mariner elements , Tec1 ( 2 , 000 copies per haploid genome ) and Tec2 ( 5 , 000 copies ) , as well as lower copy number Tec3 elements ( 20–30 copies ) [33] , [72] , [73]; ( ii ) at least a fraction of these Tec elements are precisely excised between TA dinucleotides [74]; ( iii ) an estimated 20 , 000 short TA-bounded IESs [33] , bearing a Tc1/mariner end consensus just like the Paramecium IESs [18] , are excised precisely between TA dinucleotides leaving a single TA at each excision site on the MAC destined chromosomes [33] and ( iv ) molecular characterization of excised circular forms of both Tec elements and short IESs revealed an unusual junction consisting of 2 TA dinucleotides separated by 10 bp of partially heteroduplex DNA , showing that both the Tec transposons and the short IESs are excised by the same mechanism [74] , [75] . The mechanism is moreover different from that of precise IES excision in Paramecium [24] , [57] . Neither the endonuclease responsable for IES cleavage nor the repair pathway has currently been identified in Euplotes . It will be fascinating to see whether the same actors , i . e . a domesticated piggyBac transposase and the NHEJ ( non-homologous end-joining ) pathway , are responsible for a mechanism that in its details is not the same as that found in Paramecium , or whether completely different cellular machinery has been recruited to carry out the same function i . e . the precise excision from somatic DNA of the Tc1/mariner family Tec transposons and of short TA-bounded IESs presumed to be their relics [19] . In conclusion , different ciliates have evolved different host defenses in response to germline SGE insertions . In all cases that have been examined at the molecular level , maternal non-coding RNAs are involved in programming genome rearrangements . In Paramecium and some other lineages , the co-evolution of host defense machinery and SGEs has provided mechanisms for precise somatic excision , uniquely allowing the colonization of coding sequences by Tc1/mariner and likely other transposable elements . This phenomenon is so far only paralleled by the spread of introns into eukaryotic coding sequences , also thought to result from domestication of precise excision machinery , derived in this case from mobile self-splicing ribozymes [76] .
A lambda-phage library was provided by John Preer . This library had been made from DNA obtained after isolation of stock 51 wild type micronuclei [82] and further purified by cesium chloride density gradient centrifugation to eliminate G+C-rich DNA supposed to represent bacterial contaminants [28] . The library consisted of 70 , 000 recombinant phages ( lambdaGEM11 ) , expected to represent a 7× coverage of the MIC genome . We amplified the original library in 1995 and stored it at 4°C . Phage particles from 1 mL of the reamplified library ( approximately 105 particles ) were fully recovered by ultracentrifugation ( 42 min at 113898 g in a TLA-55 rotor; Slambda particle = 410 according to [84] ) and concentrated in ∼30 µL . Given the limited amount of material ( ∼18 pg of 40 Kb phage genomes corresponding to ∼4 . 5 pg of inserts ) , the cloned DNA was amplified by PCR using primers located next to the cloning sites ( LambdaL2 GGCCTAATACGACTCACTATAGG; LambdaR2 GCCATTTAGGTGACACTATAGAAGAG ) . Non-genomic sequences should only represent 0 . 6% of the total PCR-amplified DNA . As PCR inhibitors prevented direct amplification from the concentrated suspension of phage particles , 230 50 µL-PCR reactions were performed from 3 µL of a 30× dilution in SM . The Expand Long-Template PCR System ( Roche ) was used as recommended by the supplier with 23 amplification cycles , an annealing temperature of 60°C and 12 min for the extension time . PCR reactions were concentrated by ethanol precipitation and ∼35 µg of 9 to 13 Kb PCR products were obtained after purification from 0 . 6% low-melting-temperature agarose gels and treatment with β-agarase ( Sigma ) . DNA was sequenced by a paired-end strategy using Illumina GAII and HiSeq next-generation sequencers . The shotgun fragments were ∼500 bp and the paired-end reads 108 nt for DNA enriched in un-rearranged sequences ( PGM DNA ) . The fragments were ∼200 bp and the paired-end reads 101 nt for DNA prepared from the lambda-phage library . In the latter case , short reads that overlapped were merged . All Illumina short reads were mapped to the strain 51 reference genome ( see below ) using BWA [85] ( version 0 . 5 . 8 ) . Alignments were indexed using samtools [86] ( version 0 . 1 . 11 ) . The P . tetraurelia MAC genome [1] was assembled from 13× Sanger sequencing reads from different insert size librairies of strain d4-2 DNA . Strain d4-2 only differs from strain 51 at a few loci . We corrected sequencing errors in the scaffolds using Illumina deep sequencing in two stages , the first stage using the same strain d4-2 DNA sample that had been used for the original Sanger sequencing ( 84 million 75 nt paired-end reads ) , the second stage using two different samples of strain 51 MAC DNA ( 155 million 75 nt paired-end reads ) . The electronic polishing pipeline used for each stage consisted of the following steps . ( i ) Gap filling was achieved by assembling the Illumina reads into contigs using the Velvet [87] short read assembler ( Kmer = 55 -ins_length 400 -cov_cutoff 3 -scaffolding no ) . The contigs were mapped to the draft assembly using BLAT [88] and locally realigned with Muscle [89] . If the contigs spanned a sequencing gap , then it was filled . ( ii ) The Illumina reads were mapped to the draft genome using BWA [85] . ( iii ) Alignments were indexed using samtools [86] . ( iv ) Samtools mpileup program and homemade Perl scripts were used to identify all positions covered by at least 10 reads and where at least 80% of the reads did not confirm the reference sequence . ( v ) The reference sequence was corrected using the list of errors . Steps ( ii ) through ( v ) were repeated a few times at the second stage of correction using the strain 51 reads , since BWA mapping has low error tolerance , and more reads could be mapped as the correction progressed . At the end of the process 442 of 861 sequencing gaps were filled , 13 , 758 substitutions were corrected , 929 deletions of 1–2 nt were filled and 10 , 339 insertions of 1–2 nt were removed , to yield the strain 51 reference genome that was used for IES identification . The strain 51 reference genome is available via ParameciumDB [90] . Determination of IESs that are conserved in genes duplicated by a WGD event involved identification of the position of the IES with respect to the beginning of the alignment , either using a protein alignment of ohnologs , back translated into nucleotide sequence , or using nucleotide alignment of the 2 genes . In both cases , the alignments were carried out using Muscle [89] ( version 3 . 7 ) . If the relative positions of the IES is the same within a 2 nt tolerance , then the IESs are considered to be conserved . A phylogenetic tree was computed by concatenation of the alignment of 1350 protein families corresponding to quartets of ohnologs preserved after both the intermediate and recent WGD events . All gap-containing sites were excluded from the alignment , which is therefore robust with respect to possible annotation errors . The tree was constructed using BioNJ [91] with Poisson correction for multiple substitutions . The average length of the 2 branches between the intermediate and recent WGDs is 0 . 085 substitutions/site . The average length of the 4 branches between the recent WGD and the present is 0 . 0825 substitutions per site . Assuming a constant substitution rate , we can infer that the time between the intermediate and recent WGD events and between the recent WGD and the present are equivalent , although we cannot date the events since we do not know the substitution rate in Paramecium . The MAC reference genome used for this study ( strain 51 ) and the genome-wide set of IESs are available at http://paramecium . cgm . cnrs-gif . fr/download/ . The IESs have also been integrated into ParameciumDB BioMart complex query interface and the ParameciumDB Genome Browser [90] . The short read datasets have been deposited at the European Nucleotide Archive ( Accession numbers ERA137444 and ERA137420 ) . Oligonucleotides were designed to flank the IES insertion site at a distance of 150–200 nt to allow detection of amplification products with or without an IES . All PCR amplifications were performed with an Eppendorf personal mastercycler . Standard PCR amplifications were performed with 1 unit of DyNazyme II with reagent concentrations according to instructions provided by Finnzyme ( dNTP: 200 µM each , primers: 0 . 5 µM each ) with 50 ng of template DNA . The program used is 2 min at 95°C , 10 cycles of 45 sec at 95°C , 45 sec at annealing temperature , and 1 min at 72°C , 15 cycles of 20 sec at 95°C , 20 sec at annealing temperature and 1 min at 72°C , followed by a final incubation for 3 min at 72°C . Amplified products were analyzed on 3% Nusieve ( Lonza ) in TBE 1× . Long and AT-rich PCR amplifications were performed with 1 unit of Phusion ( Finnzymes ) using the following concentrations of reagents ( dATP and dTTP: 400 µM each , dCTP and dGTP: 200 µM each , primers: 0 . 5 µM each ) with 50 ng of template DNA . The program used was 1 min at 98°C , 25 cycles of 10 sec at 98°C , 30 sec at annealing temperature and 5 min at 72°C , followed by a final incubation of 2 min at 72°C . Amplified products were analyzed on 1% UltraPure agarose ( Invitrogen ) in TAE 1× . The template DNA for the amplification reactions was an aliquot of PGM DNA enriched in un-rearranged sequences , prepared as described above . Isolation of inserts from the MIC lambda-phage library [28] was carried out as previously described [31] . Phage inserts and long-range PCR products obtained by amplification of total DNA from vegetative cells were isolated and subjected to Sanger sequencing as in [34] . The lambda-phage inserts and the cloned long-range PCR products used to characterize the Sardine and Thon transposons have been deposited in the EMBL Nucleotide Sequence Database with accession numbers HE774468–HE774475 . Several IESs with homology to the PFAM DDE_3 domain were used to find other IESs sharing nucleotide identity , leading to a set of 28 IESs that were aligned with Muscle [89] to identify 2 Anchois transposons . In a second step , the alignment was refined and manually adjusted in order to reconstruct the AnchoisA and AnchoisB transposons . These second step alignments were built using IESs along with some PGM contigs that correspond to germline-restricted , imprecisely eliminated regions of the genome containing Anchois copies ( Text S1 ) . Statistical analyses and graphics were performed in the R environment for statistical computing [92] using standard packages , as well as the ape package [93] for phylogenetic analysis . Sequence logos were generated using weblogo software [94] .
|
Ciliates are unicellular eukaryotes that rearrange their genomes at every sexual generation when a new somatic macronucleus , responsible for gene expression , develops from a copy of the germline micronucleus . In Paramecium , assembly of a functional somatic genome requires precise excision of interstitial DNA segments , the Internal Eliminated Sequences ( IES ) , involving a domesticated piggyBac transposase , PiggyMac . To study IES origin and evolution , we sequenced germline DNA and identified 45 , 000 IESs . We found that at least some of these unique-copy elements are decayed Tc1/mariner transposons and that IES insertion is likely an ongoing process . After insertion , elements decay rapidly by accumulation of deletions and substitutions . The 93% of IESs shorter than 150 bp display a remarkable size distribution with a periodicity of 10 bp , the helical repeat of double-stranded DNA , consistent with the idea that evolution has only retained IESs that can form a double-stranded DNA loop during assembly of an excision complex . We propose that the ancient domestication of a piggyBac transposase , which provided a precise excision mechanism , enabled transposons to subsequently invade Paramecium coding sequences , a fraction of the genome that does not usually tolerate parasitic DNA .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome",
"evolution",
"evolutionary",
"biology",
"genome",
"sequencing",
"dna",
"recombination",
"molecular",
"genetics",
"dna",
"sequence",
"analysis",
"transposons",
"comparative",
"genomics",
"biology",
"molecular",
"biology",
"nucleic",
"acids",
"dna",
"transposons",
"dna",
"repair",
"genomics",
"molecular",
"cell",
"biology",
"genomic",
"evolution",
"computational",
"biology",
"genetics",
"and",
"genomics"
] |
2012
|
The Paramecium Germline Genome Provides a Niche for Intragenic Parasitic DNA: Evolutionary Dynamics of Internal Eliminated Sequences
|
CpG islands were originally identified by epigenetic and functional properties , namely , absence of DNA methylation and frequent promoter association . However , this concept was quickly replaced by simple DNA sequence criteria , which allowed for genome-wide annotation of CpG islands in the absence of large-scale epigenetic datasets . Although widely used , the current CpG island criteria incur significant disadvantages: ( 1 ) reliance on arbitrary threshold parameters that bear little biological justification , ( 2 ) failure to account for widespread heterogeneity among CpG islands , and ( 3 ) apparent lack of specificity when applied to the human genome . This study is driven by the idea that a quantitative score of “CpG island strength” that incorporates epigenetic and functional aspects can help resolve these issues . We construct an epigenome prediction pipeline that links the DNA sequence of CpG islands to their epigenetic states , including DNA methylation , histone modifications , and chromatin accessibility . By training support vector machines on epigenetic data for CpG islands on human Chromosomes 21 and 22 , we identify informative DNA attributes that correlate with open versus compact chromatin structures . These DNA attributes are used to predict the epigenetic states of all CpG islands genome-wide . Combining predictions for multiple epigenetic features , we estimate the inherent CpG island strength for each CpG island in the human genome , i . e . , its inherent tendency to exhibit an open and transcriptionally competent chromatin structure . We extensively validate our results on independent datasets , showing that the CpG island strength predictions are applicable and informative across different tissues and cell types , and we derive improved maps of predicted “bona fide” CpG islands . The mapping of CpG islands by epigenome prediction is conceptually superior to identifying CpG islands by widely used sequence criteria since it links CpG island detection to their characteristic epigenetic and functional states . And it is superior to purely experimental epigenome mapping for CpG island detection since it abstracts from specific properties that are limited to a single cell type or tissue . In addition , using computational epigenetics methods we could identify high correlation between the epigenome and characteristics of the DNA sequence , a finding which emphasizes the need for a better understanding of the mechanistic links between genome and epigenome .
CpG islands are genomic regions characterized by an exceptionally high CpG dinucleotide frequency [1–3] . In humans , they are among the most important regulatory regions , with functional roles in both normal and disease-related gene expression [4 , 5] . Originally , CpG islands were discovered by virtue of an epigenetic property , namely , the absence of DNA methylation: when the human genome was experimentally digested with methylation-sensitive restriction enzymes , some genomic regions were cut into small fragments , while the bulk of the genome remained uncut [6] . Since the restriction enzyme ( HpaII ) used cuts DNA only at unmethylated CpG dinucleotides , it was concluded that a small but significant fraction of the genome is reproducibly unmethylated . After a sample of these so-called HpaII tiny fragments had been sequenced , it became obvious that they were highly GC-rich and CpG-rich [3] . In an early computational analysis , this observation was utilized to define such regions as CpG islands , and a simple set of criteria was suggested to identify them based on their DNA sequence alone [7] . According to these so-called Gardiner-Garden sequence criteria , a genomic region has to fulfill three conditions to classify as a CpG island: ( 1 ) GC content above 50% , ( 2 ) ratio of observed-to-expected number of CpG dinucleotides above 0 . 6 , and ( 3 ) length greater than 200 basepairs ( bp ) . Because the amount of sequence data strongly outnumbered the experimental data available for DNA methylation , this definition quickly replaced the original methylation-based concept . Since their initial discovery , CpG islands have been subject to extensive research . Today it is known that CpG islands ( according to the DNA sequence criteria mentioned above or slightly modified variants ) associate with more than three-quarters of all known transcription start sites [8] and with 88% of active promoters identified in primary fibroblasts [9] , indicating that they bear important regulatory functions . Furthermore , they are hot spots of specific histone modifications [10 , 11] , they frequently bind ubiquitous transcription factors such as SP1 [12] , and they exhibit particularly accessible chromatin structures [13] . For these reasons , CpG islands have become indispensable for genome analysis and annotation . For example , they play a fundamental role for promoter prediction [14] , and their use as candidate regions for aberrant DNA methylation has contributed significantly to our understanding of the epigenetic causes of cancer [15] . However , the current sequence-based definitions of CpG islands [7 , 16] incur several disadvantages , which hamper both their theoretical and practical value . First , they are based on three threshold parameters that lack a clear biological justification . For example , it is unclear why 200 bp should be the most appropriate minimum length to define CpG islands , especially since even a random permutation of the genome sequence would give rise to a substantial number of CpG islands with this minimum length . A length threshold of 500 bp is also widely used , and its use was motivated by its ability to exclude most Alu-repeat-associated regions [16] , but again , no systematic analysis or parameter selection method has been applied to justify this particular value . Second , current definitions are purely binary , i . e . , a particular region either qualifies as a CpG island or not . This not only fails to account for the fact that CpG islands can differ considerably in terms of their sequence composition and epigenetic states [17] , it can also lead to unintuitive special cases . For example , even if a short CpG-rich region fails to fulfill CpG island criteria on its own , the same region may well fulfill the criteria after small and seemingly unrelated changes of a few neighboring nucleotides . Thus , the mapping of CpG islands is inherently unstable and depends not only on the definition used but also on the exact implementation of the mapping software . In contrast , the introduction of a numerical score for CpG island strength would allow distinguishing weak , intermediate , and strong CpG islands , without the necessity of a fixed all-or-nothing threshold . Third , and most critically , sequence-based CpG island criteria fail to distinguish between “bona fide” CpG islands—which are typically unmethylated , serve as transcription regulators , and exhibit an open and transcriptionally competent chromatin structure—and CpG-rich regions lacking these characteristics . More precisely , current CpG island criteria seem to be sufficiently sensitive in the sense that they detect most bona fide CpG islands in the human genome , but their specificity is low , i . e . , they give rise to a substantial number of false positive classifications . For example , Yamada et al . observed that almost a third of the putative CpG islands analyzed showed significant DNA methylation [18] , in contradiction to the original concept of CpG islands as unmethylated regions . To resolve the significant drawbacks of current sequence-based CpG island criteria , it was suggested to abandon the concept of CpG islands altogether and to replace it by direct counting of CpG dinucleotides [19] . In this study , we propose a less radical but arguably more viable strategy . Our approach maintains the high sensitivity of current CpG island criteria , but substantially improves their specificity , it introduces a more biologically meaningful way of selecting thresholds , and it accounts for the fact that CpG islands quantitatively differ in their strength . The fundamental concept of this study is to combine an initial , sequence-based mapping of CpG islands with subsequent prediction of CpG island strength . CpG island strength is expressed as a single quantitative score per CpG island , summarizing its inherent tendency—across different cell types and tissues—to exhibit an unmethylated , open , and transcriptionally competent chromatin structure . It is calculated as a combination of epigenome predictions and provides a measure for discrimination between bona fide CpG islands and regions that are just CpG-rich but show no evidence of the epigenetic and functional characteristics of bona fide CpG islands . We evaluate the predicted CpG island scores by comparison with large-scale experimental datasets on DNA methylation and transcription initiation sites , since absence of DNA methylation and presence of promoter activity are regarded as characteristic of bona fide CpG islands . Figure 1 provides a schematic overview of our approach , which is necessarily complex since we derive and benchmark four different scores of CpG island strength using combinations of large-scale epigenome datasets . From left to right , the first step comprises preparation of seven training datasets , based on pairwise overlaps between CpG island maps and epigenome datasets . In the second step , a prediction model is trained and its performance is estimated for each training dataset . The resulting prediction models are then used to score all CpG islands genome-wide . From these scores—in step three—four CpG island scores are calculated . In step four , a performance comparison on two large-scale evaluation datasets shows that the “combined epigenetic score” is the best indicator of CpG island strength and most predictive of bona fide CpG islands . All training and testing in this study is performed on Chromosomes 21 and 22 for reasons of data availability . Predictions are calculated and validated on the entire genome . The entire workflow as outlined in Figure 1 was repeated three times , for three widely used CpG island maps . By comparing the results , we show that CpG island strength predictions provide an improvement over each map , and we are able to select the most appropriate setup for the final maps of predicted bona fide CpG islands .
Our prediction of CpG island strength and mapping of bona fide CpG islands started from traditional CpG island maps , which we derived by means of widely used sequence-based CpG island criteria . This approach is unlikely to significantly reduce the completeness of our mapping since the original CpG island criteria [7] are regarded as highly sensitive and there is no evidence that they miss a substantial number of bona fide CpG islands . The application of traditional CpG island finder algorithms faces the problem of repetitive DNA in the genome . Some evolutionarily recent repeat insertions are CpG-rich ( e . g . , Alu elements ) and could erroneously be identified as CpG islands even though they most likely bear little regulatory function [16] . Several methods have been suggested to address this problem , but their efficacy has not been systematically investigated . We therefore applied and compared three widely used calculation methods: ( 1 ) repeat exclusion by using strict thresholds for GC content ( 55% ) , CpG observed-to-expected ratio ( 0 . 65 ) , and CpG island length ( 500 bp ) as suggested by Takai and Jones [16]; ( 2 ) repeat exclusion by combining the standard Gardiner-Garden thresholds [7] with subsequent removal of all CpG islands that comprise less than 200 bp of nonrepetitive DNA; and ( 3 ) repeat exclusion by applying the standard thresholds [7] to the repeat-masked genome . Using each of these methods , we derived a genome-wide map of CpG islands . Method 1 , which we refer to as TJU ( for “Takai Jones unmasked” ) , gave rise to 37 , 531 CpG islands genome-wide . Method 2 , which we refer to as GGF ( for “Gardiner-Garden filtered” ) , gave rise to 94 , 450 CpG islands genome-wide . And method 3 , which we refer to as GGM ( “Gardiner-Garden masked” ) , gave rise to 109 , 600 CpG islands genome-wide . All three maps were processed in parallel through most of this study . Absence of DNA methylation and presence of promoter activity are regarded as characteristic of bona fide CpG islands . Therefore , we hypothesized that computational predictions of DNA methylation and promoter activity might provide suitable scores of CpG island strength and thus indicators for the genome-wide mapping of bona fide CpG islands . In previous work focusing on human lymphocytes , we showed that prediction of CpG island methylation is possible with high accuracy based on the DNA sequence plus additional information such as the DNA helix structure and the distribution of repetitive DNA elements [20] . Our finding has recently been independently confirmed for brain tissue [21 , 22] and is expected to hold for a wide range of cell types and tissues . Computational promoter prediction is a well-studied topic and is also feasible with high accuracy across different cell types and tissues ( see Bajic et al . [14] and references therein ) . We therefore prepared training datasets for DNA methylation and promoter activity ( calculation step 1 in Figure 1 ) , to be processed with our epigenome prediction pipeline ( see next section ) . Each training dataset was constructed by identifying pairwise overlaps between the three CpG island maps ( Figure 1 , orange cylinder ) and experimental epigenome datasets on DNA methylation and promoter activity ( Figure 1 , brown cylinder ) , giving rise to a set of positives ( i . e . , regions that exhibit characteristics of bona fide CpG islands ) as well as a set of negatives ( i . e . , regions that do not ) for both DNA methylation and promoter activity ( Figure 1 , grey cylinders between calculation steps 1 and 2 ) . For the prediction of unmethylated versus methylated CpG islands , training datasets were constructed using DNA methylation data that Yamada et al . established for Chromosome 21q [18] . Similarly , for the prediction of CpG islands that show evidence of promoter activity versus those that do not , training datasets were constructed using the genome-wide list of polymerase II preinitiation complex binding sites that Kim et al . established for primary fibroblasts [9] ( for consistency with additional predictions that we report below , we restricted the latter dataset to Chromosomes 21 and 22 ) . Based on our experience with DNA methylation prediction [20] , we constructed a general epigenome prediction pipeline , which performs calculation step 2 in the overview diagram ( Figure 1 ) . Each prediction takes a training dataset as input and performs three subsequent steps: calculation of prediction attributes , performance estimation by cross-validation , and genome-wide prediction . The outputs of the pipeline are an overall performance estimate , a table of most predictive attributes , and a predicted score for each CpG island genome-wide . More specifically , these steps are performed as follows . ( 1 ) Prediction attributes are calculated: for each case in the respective training dataset , the pipeline calculates 847 potentially predictive attributes from genome data . These attributes belong to six groups: DNA sequence patterns , repeat distribution , predicted DNA helix structure [23 , 24] , predicted transcription factor binding sites , genetic variation , and CpG island attributes ( genes and gene-related information are deliberately omitted to ensure that the predictions are independent of manual curation expertise ) . ( 2 ) Performance is estimated by cross-validation: using the above attributes and the training data , the pipeline trains a linear support vector machine to predict whether a CpG island belongs to the set of positives or to the set of negatives . Prediction performance is evaluated by calculating the average correlation and accuracy over ten runs of a 10-fold cross-validation . Furthermore , to understand which attributes contribute most significantly to high prediction performance , two additional analyses are performed . First , the support vector machine is trained not only on the combination of all attributes but also on each of the six attribute groups separately . Second , Wilcoxon rank-sum tests are calculated to identify the most significant of all 847 attributes . On this basis , the optimal combination of attribute groups can be selected ( we use repeat distribution plus predicted DNA helix structure plus CpG island attributes throughout this study because these three attribute groups achieve high prediction performance and capture complementary aspects of the DNA ) . ( 3 ) CpG island scores are predicted genome-wide: the linear support vector machine is trained on all training data and is then applied to calculate a prediction score between zero and one for all CpG islands genome-wide . The resulting score describes the likelihood that a particular CpG island belongs to the set of positives ( i . e . , regions that exhibit characteristics of bona fide CpG islands ) and is therefore a potential indicator of CpG island strength . Processing the training data for DNA methylation and promoter activity through our epigenome prediction pipeline showed that the pipeline can distinguish with high accuracy between unmethylated and methylated CpG islands and , similarly , between CpG islands that exhibit evidence of promoter activity ( namely polymerase II preinitiation complex binding sites ) and those that do not ( Tables 1 , S1 , and S2 ) . A closer inspection of the most predictive attributes helps us to understand how this prediction performance is achieved ( Tables S3 and S4 ) . First , unmethylated CpG islands contain significantly fewer tandem repeats and segmental duplications than their methylated counterparts . Second , polymerase II preinitiation complex–bound CpG islands overlap more frequently with highly conserved regions than do unbound CpG islands . And third , both unmethylated and polymerase II preinitiation complex–bound CpG islands are highly enriched with CpG-rich sequence patterns and regions of low predicted DNA rise ( which is an important aspect of DNA helix structure [23 , 24] ) . These results support the hypothesis that the prediction score for DNA methylation at CpG islands as well as the prediction score for polymerase II preinitiation complex binding at CpG islands are both suitable indicators of CpG island strength . We denote their genome-wide prediction values derived by the epigenome prediction pipeline as the “predicted unmethylated score” and the “predicted promoter activity score , ” respectively , and evaluate their predictiveness for CpG island strength below . CpG island scores that focus exclusively on the absence of DNA methylation or on evidence of promoter activity may be insufficient for capturing all aspects of the complex epigenetic and functional states that characterizes bona fide CpG islands . To construct a more comprehensive epigenetic scoring of CpG island strength , we collected five additional large-scale epigenome datasets from the literature , each one describing a different aspect of an open and transcriptionally competent chromatin structure: histone H3K4 di- and trimethylation [10] , histone H3K9/14 acetylation [10] , DNase I hypersensitivity [13] and SP1 transcription factor binding [12] . All these datasets cover the nonrepetitive parts of human Chromosomes 21 and 22 , to which we confine our analysis . A genomic co-localization analysis that we performed showed that all five datasets of epigenetically modified regions indeed exhibit significant overlap with all three CpG island maps ( Figure 2 ) . Briefly , this analysis involved two steps . First , the absolute number of pairwise overlaps along Chromosomes 21 and 22 was counted for each pairwise combination of epigenetic modification map and CpG island map ( Figure 2A ) . Second , these numbers were normalized by the expected frequency of overlap under the assumption of CpG islands and epigenetically modified regions being uniformly distributed ( Figure 2B ) , to correct for length and frequency differences ( see Materials and Methods for details ) . Intriguingly , the enrichment observed in the genomic co-localization analysis was not independent between datasets but highly skewed towards a specific subset of CpG islands that frequently overlap with several epigenetic modifications simultaneously ( Table 2 ) . For example , CpG islands that show evidence of two out of five epigenetic modifications simultaneously are observed 10-fold to 20-fold more frequently than expected under a random model . We therefore concluded that all five epigenetic modifications do in fact capture different epigenetic indicators of a single concept , namely , whether or not a particular CpG island fosters an open and transcriptionally competent chromatin structure . To convert this observation into a method for scoring CpG island strength , we prepared training datasets and applied our prediction pipeline separately for each of the five epigenetic modifications ( calculation steps 1 and 2 in Figure 1 ) . In all cases , the support vector machine was able to distinguish with significant accuracy between CpG islands that overlap with the particular epigenetic modification and those that do not ( Tables 3 and S5 ) . A closer inspection of the most predictive attributes showed that CpG islands exhibiting overlap with the epigenetic modification are more likely to contain CpG-rich patterns , are more conserved , and exhibit a characteristic predicted helix structure ( see Table S6 for a list of most significant differences ) . Furthermore , we observed high correlations between the prediction scores for all five epigenetic modifications ( Table S7 ) , which provided additional support for the conclusion that they represent aspects of a single concept . Therefore , for each CpG island we calculated the average over all five predictions and thereby derived a single “open chromatin score” ( calculation step 3 in Figure 1 ) . Finally , since the predicted unmethylated score , the predicted promoter activity score , and the open chromatin score can be assumed to capture complementary aspects of a CpG island's epigenetic and functional state , we combined these three scores into an additional consensus score that we call the “combined epigenetic score” of CpG island strength . Up to this point , we carried out all analyses in parallel for the three CpG island maps that we derived using different repeat-exclusion strategies ( TJU , GGF , and GGM ) . To select the most appropriate setup for the final map of predicted bona fide CpG islands , we benchmarked these strategies on both evaluation datasets . Since ROC curves cannot easily account for the different number of CpG islands in each of the three maps , we constructed an alternative type of diagram for this purpose ( Figure 5 ) . This diagram plots the precision of the classification ( i . e . , the percentage of predicted bona fide CpG islands that are supported by the DNA methylation dataset [Figure 5A] or by the transcription start site dataset [Figure 5B] ) and the true positive rate ( i . e . , the percentage of unmethylated CpG islands [Figure 5A] or CpG islands harboring transcription start sites [Figure 5B] that are correctly predicted ) against the total number of CpG islands that are selected for any particular threshold . The results show that there is generally high agreement between the performance of the combined epigenetic score on each of the three CpG island maps ( Figure 5 ) , apart from the trivial fact that the overall sizes of the three maps differ . Nevertheless , the combined epigenetic score performs slightly better on the GGM map ( i . e . , repeat exclusion using RepeatMasker , with subsequent application of the Gardiner-Garden criteria for CpG island detection ) than on the two alternative maps , and this setup was therefore chosen . The GGM map has two additional advantages . First , in contrast to the GGF map , it does not require choosing a cutoff for the maximum repeat content that is permitted per CpG island . Second , in contrast to the TJU map , the DNA sequence parameters used to derive the GGM map are so permissive that virtually every nonrepetitive , CpG-rich region that exceeds 200 bp is selected and scored . Thus , scores are also calculated for regions that show little potential to be bona fide CpG islands but which may be of interest for comprehensive scans of particular genomic regions . At http://rd . plos . org/10 . 1371_journal . pcbi . 0030110_01 , we report the combined epigenetic score for all CpG islands that fulfill the Gardiner-Garden criteria on the repeat-masked genome ( GGM ) . Since our evaluations showed that the combined epigenetic score provides an accurate and robust estimate of CpG island strength ( i . e . , of a CpG island's inherent tendency to exhibit an open and transcriptionally competent chromatin structure ) , these scores are useful for a number of applications . For example , they add important quantitative information to support functional genome annotation as well as the interpretation of experimental epigenome data , and they can be used to prioritize candidate regions , e . g . , when selecting a fixed number of most promising regulatory CpG islands for experimental followup . Although our analysis emphasizes the importance of quantitative information on CpG island strength , to distinguish gradually between bona fide CpG islands and those CpG-rich regions that show no evidence of a regulatory role ( Figures 3 and 4 ) , we acknowledge that certain applications would benefit from a fixed threshold on the combined epigenetic score . For example , to derive a genome-wide list of predicted bona fide CpG islands or for selecting regions to be spotted on a CpG island microarray , it is necessary to make a tradeoff between thresholds that are low enough to achieve high sensitivity ( i . e . , most bona fide CpG islands are included ) and high enough to maintain high specificity ( i . e . , few CpG-rich regions that show no evidence of a regulatory role are selected ) . Fortunately , the way the combined epigenetic score is defined immediately suggests a threshold that balances sensitivity and specificity and carries a biologically meaningful interpretation . Since the combined epigenetic score is the average of the confidences with which a particular CpG island is predicted ( 1 ) to be unmethylated , ( 2 ) to exhibit promoter activity , and ( 3 ) to foster open chromatin structure , it assigns a value between zero and one to each CpG island that reflects its overall epigenetic and functional state . A value of zero thus corresponds to a completely silenced , inactive , and inaccessibly buried CpG island , while a value of one corresponds to an unmethylated , highly accessible CpG island with strong promoter activity . Between these two extremes , a value of 0 . 5 corresponds to CpG islands that are equally likely to be bona fide CpG islands or not . This value therefore provides a suitable threshold for CpG island mapping , as it balances sensitivity and specificity . We would recommend this threshold for most applications . Nevertheless , certain tasks ( e . g . , genome annotation ) may require increased sensitivity to annotate as many bona fide CpG islands as possible and would therefore profit from a less stringent threshold , such as 0 . 33 . Conversely , a highly conservative threshold of 0 . 67 is useful when selecting candidate regulatory regions for experimental followup , to minimize the risk of wasting resources on false positives . To support decision-making about the most appropriate map to use for a particular application , Table 4 provides quantitative data on true positive rates and false positive rates calculated for both evaluation criteria , DNA methylation and promoter activity . Using the GGM map as the basis ( 109 , 600 CpG islands for the entire human genome ) and the combined epigenetic score as the indicator of CpG island strength , we calculated maps of predicted bona fide CpG islands . Using the balanced 0 . 5 threshold , 21 , 631 genomic regions are predicted as bona fide CpG islands ( 19 . 7% ) ; for the highly sensitive 0 . 33 threshold , this value is 46 , 182 ( 42 . 1% ) ; and for the highly specific 0 . 67 threshold , we predict 10 , 281 bona fide CpG islands genome-wide ( 9 . 4% ) . All CpG island maps are available for download and inspection as UCSC Genome Browser [28] tracks at http://rd . plos . org/10 . 1371_journal . pcbi . 0030110_01 . The genomic distribution of bona fide CpG islands is summarized in Table S8 . Furthermore , we assessed how frequently bona fide CpG islands associate with genes , exons , annotated transcription start sites , and highly conserved regions ( Table S9 ) . As expected , predicted bona fide CpG islands are highly associated with annotated transcription start sites and evolutionarily conserved regions , and this effect is stronger for the specific threshold than for the balanced and the sensitive thresholds . However , even of the 10 , 281 strongest CpG islands in the human genome , i . e . , those whose scores exceed the highly specific 0 . 67 threshold , more than 40% do not overlap with an Ensembl-annotated transcription start site . Thus , we conclude that our prediction of CpG island strength identifies a significant number of regions with open and transcriptionally competent chromatin structure that are not known promoters of protein-coding genes . As outlined above , the combined epigenetic score has a conceptual advantage over more conventional ways of predicting CpG island strength because it directly links CpG island maps to the epigenetic and functional role that CpG islands are assumed to play in the human genome . However , it bears one significant disadvantage: the calculation of the combined epigenetic score is complex and computationally demanding . While we alleviate this issue by providing precalculated maps for the current assemblies of the human genome , it would be helpful to have a second estimate of CpG island strength available that is significantly simpler to calculate , even at the cost of a somewhat reduced performance . As suggested above and supported by Figure 3 , CpG island length can be used in this way . It is substantially , though not perfectly , correlated with the combined epigenetic score ( Pearson's r = 0 . 59 ) , and it gives rise to a ROC area under the curve [27] performance that is not dramatically lower than that of the combined epigenetic score ( Figure 3 ) . However , it is unclear what might be suitable thresholds to map bona fide CpG islands on the basis of their length , since—in contrast to the combined epigenetic score—CpG island length does not reflect any specific epigenetic concept . We propose that the most appropriate solution is to select thresholds such that the resulting maps resemble those calculated from the combined epigenetic score in terms of the false positive rate . That is , the length heuristic should not make more errors when detecting bona fide CpG islands than the combined epigenetic score , but it may well detect fewer ( worse ) or more ( better ) bona fide CpG islands , as measured by the true positive rate . Table 4 provides a performance comparison of bona fide CpG island maps derived from the combined epigenetic score versus maps derived using the CpG island length heuristic , with thresholds selected such that the false positive rate is as close as possible to that of the maps derived from the combined epigenetic score . Taking the results for both evaluation datasets into account and rounding to the closest hundred , we concluded that a minimum length of 700 bp is the most appropriate threshold for the balanced case . For sensitive mapping , the most appropriate minimum length is 300 bp , and for specific mapping , the most appropriate minimum length is 1 , 400 bp . Direct performance comparison with the maps derived from the combined epigenetic score shows that this length-based heuristic performs equally well for sensitive mapping ( slightly worse for DNA methylation , slightly better for promoter activity ) , but falls short for both the balanced and the specific maps ( Table 4 ) . Differences are particularly strong for the specific case , where the map based on the combined epigenetic score predicts 65% ( DNA methylation: true positive rate of 36 . 2% versus 22 . 0% ) and 56% ( promoter activity: 21 . 7% versus 13 . 9% ) more bona fide CpG islands than the heuristic when false positive rates are fixed to 1 . 2% for both maps . We conclude that the length-based heuristic can be used for a general mapping of bona fide CpG islands , preferably with a minimum length threshold of 300 bp . However , as soon as high specificity is desirable , we strongly recommend using the maps of predicted bona fide CpG islands that are based on the combined epigenetic score . This conclusion is consistent with the observation that exclusively sequence-based CpG island maps achieve high sensitivity but lack specificity , i . e . , they include many regions that fail to exhibit the epigenetic and functional characteristics of bona fide CpG islands .
The CpG island strength as a theoretical concept captures the inherent tendency of a particular CpG island to exhibit the characteristic epigenetic and functional state of bona fide CpG islands . This includes , but is not limited to , absence of DNA methylation as well as presence and strength of promoter activity . The concept of CpG island strength is abstracted from any tissue-specific or cell-type-specific variation of the epigenetic states . It should be viewed as a description of the default state that is encoded in the DNA sequence of a particular CpG island , and which the CpG island will assume in the absence of any strong influences towards variation ( such as imprinting-related differential methylation or cancer-related epigenetic silencing ) . Since we observed clear-cut quantitative differences among CpG islands ( Figure 4 ) and a highly significant clustering of epigenetic modifications in a subset of CpG islands ( Table 2 ) , we conclude that this concept adds important information to traditional CpG island maps . Furthermore , it provides a straightforward solution for the lack of specificity of these maps . To predict CpG island strength for each CpG island in the human genome , we initially predicted multiple epigenetic modifications independently . These genome-wide predictions were highly correlated with each other , hence we could combine them into a consensus prediction of CpG island strength . The predictive power of this combined epigenetic score ( and of several alternative CpG island scores ) was evaluated on large-scale experimental datasets of DNA methylation and promoter activity . We also selected and justified biologically plausible thresholds on the combined epigenetic score , leading to maps of predicted bona fide CpG islands that are more accurate than current sequence-based maps . For example , even the most restrictive definition [16] of CpG islands ( TJU ) gives rise to approximately one-third methylated CpG islands , i . e . , CpG-rich regions that fail to exhibit the characteristics of bona fide CpG islands according to our evaluation dataset . Using a sensitive threshold of 0 . 33 on the combined epigenetic score , this value can be reduced by two-thirds , while losing less than 8% of unmethylated , potentially bona fide CpG islands ( Figure 3A ) . Similar improvements were observed when evaluating promoter activity and for two additional CpG island maps ( GGF and GGM ) . We therefore conclude that a post-processing step utilizing bioinformatic predictions significantly increases the accuracy of CpG island mapping and can help overcome the weaknesses of current CpG island definitions . We also showed that a simple length-based mapping heuristic that selects only CpG islands with a minimum length of 300 bp on the repeat-masked genome is suitable for sensitive mapping of bona fide CpG islands but performs substantially worse than the combined epigenetic score when high specificity is desired . The fundamental advance of our analysis was to move beyond a purely sequence-based definition of CpG islands ( which many researchers have tried to optimize in the past [29–33] ) and to incorporate epigenome and chromatin data . This approach is consistent with the common notion of CpG islands being functionally and epigenetically exceptional regions , but gave rise to two conceptual difficulties . First , such data are tissue-specific and cell-type-specific . It is thus necessary to abstract the experimental data from these variations to derive a single CpG island map for the human genome ( instead of specific maps for all major tissues and cell types ) . Second , comprehensive epigenome data are currently available only for Chromosomes 21 and 22 , not for the entire genome . We addressed both issues by introducing epigenome prediction as the method for scoring CpG island strength , instead of using epigenome data directly . Our epigenome predictions utilize a strong link that connects the DNA characteristics of individual CpG islands with their epigenetic states . As illustrated schematically in Figure 6 , CpG islands differ in terms of their epigenetic states , in particular in their inherent tendency towards either open and transcriptionally competent or inaccessible and silenced chromatin structure . Similarly , CpG islands differ in terms of their DNA characteristics and genomic locations , including length and CpG frequency , preferred DNA helix structure , association with conserved regions , frequency of transcription start sites , and distribution of repetitive DNA elements . Intriguingly , epigenetic state and DNA characteristics are highly correlated , as indicated by the consistently high prediction accuracies that we observed throughout this study: CpG islands that are frequently unmethylated , exhibit promoter activity , and/or foster open chromatin structure also exhibit exceptional DNA characteristics , including high levels of CpG enrichment , high conservation , significant repeat depletion , and a specific predicted helix structure . On the other hand , methylated and transcriptionally inactive regions ( that still fulfill the traditional CpG island criteria ) exhibit converse DNA characteristics . This high degree of correlation between DNA characteristics and epigenetic state extends beyond CpG islands: our prediction pipeline also achieved high prediction performances for the distinction between regions that exhibit an open and transcriptionally competent chromatin structure and a set of randomly selected genomic regions ( unpublished data ) . We therefore conclude that the human genome and epigenome are significantly correlated . Potential limitations of this study arise from the epigenome datasets that were employed for training and evaluation . First , two out of the five ChIP-on-chip datasets that we used are based on ligation-mediated PCR amplification [9 , 12] , which creates an experimental bias towards GC-rich regions ( the other three are based on a more appropriate linear DNA amplification method [10] ) . Second , the lists of over-represented regions from the ChIP-on-chip studies that we used are most likely overly conservative [34] . However , in spite of these shortcomings of the underlying datasets , we observed consistent results across multiple datasets , which were obtained from different cell types , in different labs , and with different experimental protocols . Therefore , such error sources are highly unlikely to invalidate our main results . A second limitation concerns our ability to exhaustively evaluate the performance of the predictions: because the concepts of CpG island strength and of bona fide CpG islands describe inherent properties of CpG islands , which abstract from their epigenetic state in a particular tissue or cell type , they are difficult to measure experimentally . We therefore performed our evaluations on datasets that significantly deviate in their experimental and biological characteristics from all training data that was used , and we paid as much attention to deriving consistent and biologically plausible predictions of CpG island strength as to achieving the highest performance on the evaluation criteria . Finally , for reasons of data availability we focused on epigenetic modifications that are associated with open and transcriptionally competent chromatin . Future extensions of this work should include repressive epigenetic modifications as well , such as histone H3K9 methylation and H3K27 methylation . On this basis , combined with larger datasets , it may be possible to deconstruct the predicted CpG island strength into individual components for all major epigenetic modifications . The CpG island strength predictions and maps of predicted bona fide CpG islands described in this study are currently being used in several ongoing research projects , with topics ranging from imprinting regulation and epigenome profiling [35] to cancer-related hypermethylation , and have so far proved to be highly useful , both for guiding the selection of candidate regulatory regions and for supporting the interpretation of experimental results .
To calculate genome-wide CpG island maps according to the traditional sequence-based definition , we downloaded both the unmasked and the repeat-masked versions of the hg17/NCBI35 human genome assembly from the UCSC Genome Browser Web site [28] , and we ran a slightly modified version of the CpG Island Searcher script [16] with the following parameters . Calculation of the TJU map: GC content above 55% , CpG observed-to-expected ratio above 0 . 65 , length above 500 bp , based on the unmasked genome . Calculation of the GGF map: GC content above 50% , CpG observed-to-expected ratio above 0 . 6 , length above 200 bp , based on the unmasked genome . Calculation of the GGM map: GC content above 50% , CpG observed-to-expected ratio above 0 . 6 , length above 200 bp , based on the repeat-masked genome . Finally , for GGF we determined the number of nonrepetitive basepairs by comparison with the repeat-masked genome version and discarded all CpG islands for which this value was below 200 bp . For the prediction of DNA methylation , promoter activity , and the five components of the open chromatin score , we implemented a custom computer program . This epigenome prediction pipeline is based on our experience with the prediction of DNA methylation published previously [20] , and it implements several significant extensions . First , a 20-fold speedup of the program over the original version , achieved by optimization of the source code and of the database structure , now permits genome-wide analysis at acceptable speed . Second , a front end for Web-based analysis was implemented , which enables us to make the prediction pipeline available to interested researchers on a cooperation basis ( see http://rd . plos . org/10 . 1371_journal . pcbi . 0030110_02 for details ) . Unrestricted public access to this Web service is not yet feasible because of high computational demand of the prediction pipeline , but it is planned for the future . Briefly , the prediction pipeline works as follows . It takes a training set as input that consists of two lists of genomic positions ( i . e . , chromosome , start and end of the region relative to the hg17/NCBI35 genome assembly ) , the first one representing the positive cases and the second one the negative cases . Then , four consecutive steps are performed . First , to prepare the DNA-based prediction , a wide range of DNA attributes are calculated for all training cases and , in addition , for all CpG islands in the human genome . These attributes belong to six classes: ( 1 ) DNA sequence patterns and properties ( 426 attributes ) , ( 2 ) repeat attributes , frequency , and distribution ( 311 attributes ) , ( 3 ) predicted DNA helix structure ( 28 attributes ) , ( 4 ) predicted transcription factor binding sites ( 68 attributes ) , ( 5 ) evolutionary conservation and single nucleotide polymorphisms ( ten attributes ) , and ( 6 ) CpG island attributes ( four attributes ) . Most attributes take the form of frequencies or numerical scores , averaged over the CpG island and standardized to a default window size of one kilobase ( see Table S10 for the full list of attributes and for information on their calculation ) . The data for most of these attributes were collected from annotation tracks in the UCSC Genome Browser [28] ( as of September 2005 ) , with some exceptions: the attributes for classes 1 and 6 were calculated directly from the DNA sequence , and the attributes for class 3 were calculated from the DNA sequence by averaging over oligonucleotides with known structure [24] . Second , to estimate the prediction performance that a linear support vector machine can achieve for classification of positives and negatives , it is repeatedly trained and tested on partitions of the training dataset following a four-step procedure . ( 1 ) If the larger set ( either positives or negatives ) contains more than twice as many sites as the smaller set , it is randomly downsampled such that the class imbalance never exceeds 67% versus 33% ( this precaution limits potential bias towards predicting the majority class ) . ( 2 ) Using 10-fold cross-validation , a linear support vector machine [36] as implemented in the Weka package [37] is repeatedly trained on 90% of the cases and tested on the remaining 10% ( with default parameters ) . ( 3 ) Cross-validation is repeated ten times with random partition assignments . ( 4 ) The overall prediction performance is measured by the correlation coefficient between the predictions and the correct values on the test set of the cross-validations and by the percent accuracy of correctly predicted test set cases [38] , averaged over all cross-validation runs . Third , to understand which DNA attributes contribute to high prediction performances , the analysis described in the previous step is repeated for all six attribute groups separately ( Tables S1 , S2 , and S5 ) . In addition , single-attribute significance testing is performed on all 847 attributes ( Tables S3 , S4 , and S6 ) , using the nonparametric Wilcoxon rank-sum test with an overall significance threshold of 5% per statistical analysis . p-Values are adjusted for multiple testing alternatively by the highly conservative Bonferroni method ( which controls the family-wise error rate ) and by a more recent method that controls the false discovery rate [39] . Fourth , to derive a score for all CpG islands in the human genome , a linear support vector machine is trained as described above , but now on the full training dataset ( with downsampling if necessary , to enforce a maximum class imbalance of 67% versus 33% ) . The trained prediction model is then used to predict the likelihood of belonging to the set of positives for each CpG island genome-wide . Likelihoods are calculated as implemented by the Weka package [37] . The resulting quantitative predictions can assume values between zero and one , where a value of zero corresponds to a high-confidence negative prediction , a value of 0 . 5 to a borderline case , and a value of one to a high-confidence positive prediction . This quantitative prediction can then be used directly as a CpG island score or it can be subjected to further calculations as described below . The calculation of all four CpG island scores made use of the prediction pipeline , combined with appropriate training data . Calculations were performed on the hg17/NCBI35 genome assembly . Where necessary , data were remapped using the UCSC Genome Browser liftOver tool [28] . The predicted unmethylated score is based on training data from an experimental analysis of CpG island methylation in human lymphocytes [18] ( dataset obtained from the supplementary material of [18] ) . Using methylation-specific restriction enzyme and PCR , Yamada et al . measured DNA methylation states for 149 CpG-rich regions on Chromosome 21q , of which 132 cases showed an unambiguous methylation pattern and could be mapped to the current genome assembly . All CpG islands that overlap ( by at least 1 bp ) with one of the 103 unmethylated regions were combined into the positive training set , and all CpG islands that overlap with one of the 29 methylated cases were combined into the negative training set . The resulting training dataset was then processed by the prediction pipeline to derive predicted unmethylated scores for all CpG islands according to TJU , GGF , and GGM . The predicted promoter activity score is based on training data from an experimental analysis of polymerase II preinitiation complex binding in human fibroblasts [9] ( dataset obtained from the supplementary material of [9] ) . Using the ChIP-on-chip protocol and a highly conservative method for identifying regions of over-representation from the raw data , Kim et al . derived a genome-wide map of the most likely binding sites . All CpG islands on Chromosome 21 and 22 that overlap by at least 1 bp with one of these binding sites were combined into the positive training set . The negative training set was constructed from those CpG islands on Chromosome 21 and 22 that are at least 500 bp away from the nearest binding site . The resulting training dataset was then processed by the prediction pipeline to derive predicted promoter activity scores for all CpG islands according to TJU , GGF , and GGM . The open chromatin score is based on training data from several large-scale analyses . ( 1 ) Using the ChIP-on-chip protocol , Bernstein et al . [10] derived histone modification data for the HepG2 cell line , including H3K4 di- and trimethylation and H3K9/14 acetyla-tion ( dataset obtained from http://www . broad . mit . edu/cell/chromatin_study ) . Their analysis comprised the nonrepetitive parts of Chromosomes 21 and 22 , for which they calculated sites of significant over-representation . ( 2 ) Using DNase I digestion and subsequent massively parallel signature sequencing , Crawford et al . [13] derived a genome-wide profile of DNase I hypersensitive sites in CD4+ T cells ( dataset obtained from the UCSC Genome Browser [28] ) . ( 3 ) Using the ChIP-on-chip protocol , Cawley et al . [12] derived binding data for the ubiquitous transcription factor SP1 in the Jurkat cell line ( dataset obtained from http://transcriptome . affymetrix . com/publication/tfbs ) . Their data comprises the nonrepetitive parts of Chromosomes 21 and 22 , for which they calculated sites of significant over-representation . For each of the five epigenetic modifications , respectively , we constructed a training dataset as follows . All CpG islands on Chromosome 21 and 22 that overlap with the most significant sites for the respective epigenetic modification ( as reported by the original authors ) were included in the positive training set , and all CpG islands on Chromosome 21 and 22 that were at least 500 bp away from the nearest site were included in the negative training set . All five resulting training datasets were then processed by the prediction pipeline , and the five predictions for each CpG island were averaged , to derive open chromatin scores for all CpG islands according to TJU , GGF , and GGM . The combined epigenetic prediction score is calculated for each CpG island as the ( unweighted ) average of its predicted unmethylated score , its predicted promoter activity score , and its open chromatin score . Since all three components can assume values from zero to one , the same is true for their average . For the evaluation on DNA methylation , we used a dataset by Rollins et al . [25] , who identified 3 , 073 unmethylated and 2 , 565 methylated domains in human brain tissue ( dataset obtained from http://epigenomics . cu-genome . org/html/meth_landscape ) . Their data are based on paired-end sequencing from two DNA libraries that were constructed by digestion with methylation-sensitive restriction enzymes , such that one library is highly enriched with unmethylated regions while the other contains almost exclusively methylated regions . We regarded a CpG island as unmethylated if it overlapped by at least 25% with an unmethylated domain and as methylated if it overlapped by at least 25% with a methylated domain . No cases were observed where a single CpG island overlapped with an unmethylated and a methylated domain simultaneously . For the evaluation of promoter activity , we used a dataset from the FANTOM3 consortium [26] , who performed large-scale CAGE analysis ( i . e . , tag sequencing of 5′ ends of full-length mRNA ) on cDNA libraries derived from a wide range of tissues and cell types ( dataset obtained from http://gerg01 . gsc . riken . jp/cage_analysis/export/hg17prmtr ) . All CpG islands that contained at least three tags ( i . e . , experimental evidences of independent transcription initiation events ) were regarded as CpG islands with promoter activity , while all other cases were regarded as CpG islands that show either no or only spurious promoter activity . ROC curves were constructed in the usual way [27] , using the ROCR library [40] and the R statistical package ( http://www . r-project . org ) . The diagrams that compare the different repeat-exclusion strategies ( Figure 5 ) were constructed using the same tools , with some customizing to ensure that every unmethylated domain is counted only once for the true positive rate , even if it overlaps with several CpG islands simultaneously . All R scripts are available on request . To show that the five components of the open chromatin score exhibit significant overlap with each other and with the three CpG island maps ( TJU , GGF , and GGM ) , we performed a co-localization analysis of these eight datasets on Chromosomes 21 and 22 . To this end , a custom script was written that counts the number of sites of one type that overlap with a second type , for all pairs of site types ( i . e . , epigenetically modified regions and CpG islands ) . From these values , overlap percentages were calculated and plotted as a heat map ( Figure 2A ) . However , frequent and long regions are obviously more likely to overlap with other sites than are rare and short regions . We therefore normalized the observed frequency of overlap by the expected frequency for a uniform distribution , using the following procedure . ( 1 ) For each site type , we derived a random control set with similar set size , length distribution , and repeat overlap . Technically , for each record in the corresponding dataset , a random site of identical length was drawn from the entire length of Chromosomes 21 and 22 . If this random site was within five percentage points of its corresponding record in terms of repeat content , it was retained; otherwise , a new random site was drawn . ( 2 ) Pairwise frequencies of overlap between all control regions were counted . ( 3 ) Steps 1 and 2 were repeated 20 times , and frequencies of overlap were averaged . ( 4 ) The observed frequencies of overlap for the real data were divided by the averaged random overlap frequencies , giving rise to n-fold over- and under-representations . Figure 2B reports base-2 log scores of these over-representations ( under-representation relative to the expected overlap did not occur ) . Genome-wide maps of predicted bona fide CpG islands and CpG island strength scores can be accessed online and downloaded at http://rd . plos . org/10 . 1371_journal . pcbi . 0030110_01 . Furthermore , they are available as custom tracks on the UCSC Genome Browser Web site and as Distributed Annotation System tracks ( www . biodas . org ) for visualization within the Ensembl genome browser . The source code of the prediction pipeline is available on request from cbock@mpi-inf . mpg . de .
|
A key challenge for bioinformatic research is the identification of regulatory regions in the human genome . Regulatory regions are DNA elements that control gene expression and thereby contribute to the organism's phenotype . An important class of regulatory regions consists of so-called CpG islands , which are characterized by frequent occurrence of the CG sequence pattern . CpG islands are strongly associated with open and transcriptionally competent chromatin structure , they play a critical role in gene regulation , and they are involved in the epigenetic causes of cancer . In this article we make several conceptual improvements to the definition and mapping of CpG islands . First , we show that the traditional distinction between CpG islands and non-CpG islands is too harsh , and instead we propose a quantitative measure of CpG island strength to gradually distinguish between stronger and weaker regulatory regions . Second , by genome-wide comparison of multiple epigenome datasets we identify high correlation between features of the genome's DNA sequence and the epigenome , indicating strong functional interdependence . Third , we develop and apply a novel method for predicting the strength of all CpG islands in the human genome , giving rise to an improved and more accurate CpG island mapping .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"homo",
"(human)",
"mammals",
"computational",
"biology",
"molecular",
"biology",
"genetics",
"and",
"genomics"
] |
2007
|
CpG Island Mapping by Epigenome Prediction
|
As tourism is the mainstay of the Maldives’ economy , this country recognizes the importance of controlling mosquito-borne diseases in an environmentally responsible manner . This study sought to estimate the economic costs of dengue in this Small Island Developing State of 417 , 492 residents . The authors reviewed relevant available documents on dengue epidemiology and conducted site visits and interviews with public health offices , health centers , referral hospitals , health insurers , and drug distribution organizations . An average of 1 , 543 symptomatic dengue cases was reported annually from 2011 through 2016 . Intensive waste and water management on a resort island cost $1 . 60 per occupied room night . Local vector control programs on inhabited islands cost $35 . 93 for waste collection and $7 . 89 for household visits by community health workers per person per year . Ambulatory care for a dengue episode cost $49 . 87 at a health center , while inpatient episodes averaged $127 . 74 at a health center , $1 , 164 . 78 at a regional hospital , and $1 , 655 . 50 at a tertiary referral hospital . Overall , the cost of dengue illness in the Maldives in 2015 was $2 , 495 , 747 ( 0 . 06% of gross national income , GNI , or $6 . 10 per resident ) plus $1 , 338 , 141 ( 0 . 03% of GNI or $3 . 27 per resident ) for dengue surveillance . With tourism generating annual income of $898 and tax revenues of $119 per resident , results of an international analysis suggest that the risk of dengue lowers the country’s gross annual income by $110 per resident ( 95% confidence interval $50 to $160 ) and its annual tax receipts by $14 per resident ( 95% confidence interval $7 to $22 ) . Many innovative vector control efforts are affordable and could decrease future costs of dengue illness in the Maldives .
With approximately half of the world’s population at risk , dengue remains the most important mosquito-borne infection world-wide , [1 , 2] costing almost $9 billion globally per year for prevention and control . [3] The ecology of Small Island Developing States and territories ( SIDS ) , particularly with regards to temperature and precipitation , keeps dengue a continuing threat . [4] Outbreaks of dengue in SIDS can cause high burden , affecting the majority of island residents and overwhelming health systems . [5] SIDS are at risk of vector borne diseases as they are prone to natural disasters , often lack safe water supply , sanitation and waste management strategies , and their local governments have limited resources to implement effective vector control and manage outbreaks . [6] The Republic of Maldives ( the Maldives ) is a South Asian SIDS in the Indian Ocean made up of around 1 , 192 islands . Its 417 , 492 residents ( in 2015 ) lived on 187 inhabited islands ( island inhabited by national residents ) plus 126 resort islands , all grouped in 20 administrative atolls . [7] The capital city , Malé , is the most populous island . The Maldives ranks high in South East Asia in World Bank health indicators , [8] thanks to a well-developed public health system with a publicly funded health center on each inhabited island , a hospital for each atoll , and both public and private referral hospitals in Malé . Maldivians are enrolled in the country’s universal insurance system , Aasandha , which covers the cost of medical treatment , prescriptions , transfers , and if necessary , overseas care . Dengue was first reported in the Maldives in 1979 and became endemic in 2004 , when all atolls began reporting a high incidence of infections . [9] The Maldives’ Health Protection Agency ( HPA ) is responsible for dengue and vector surveillance and control , supporting personnel on each atoll and each inhabited island . As dengue is transmitted by Aedes mosquitoes , and in the absence of effective treatment or a public vaccination program , the main prevention strategy relies on controlling the vector population through integrated vector management . [10 , 11] Tourism is the major economic sector of the Maldives , with more than 1 . 2 million visitors arriving in 2014 , [12] mostly from Europe . To sustain this economy , the Maldives maintains its reputation as a “paradise island destination” by ensuring a clean environment with minimal risk of infectious diseases . The coefficient ( ± standard error ) of “dengue” in a regression on tourist arrivals was highly significant . These calculations suggest that the risk of dengue reduced the number of international tourist arrivals by 11 . 7% ( 95% confidence interval 5 . 3% to 18 . 0% ) compared to the expected number without such a risk . [13] As 9 out of 10 guests stay in self-contained resort islands , which generally practice comprehensive waste management and source reduction strategies , their risk of dengue infection is limited . Over the last decade , however , tourists increasingly visit and stay in guest houses on inhabited islands . Mosquito populations are abundant on these islands , increasing the risk of exposure to dengue virus . There are reports of tourists[14 , 15] or visiting workers[16] contracting dengue in the Maldives . Predominant vector breeding sites on inhabited islands stem from unmanaged waste and the presence of unprotected water storage containers . With support from the World Bank , the Maldives is expanding its national plan for waste management , [17] initiating a $17 million program to improve waste collection and management at regional and island levels . [18] In addition , health care workers on islands are trained to promote the safe storage of water and prevention of breeding sites to residents . Each island council is tasked with developing its own vector control strategies based on local needs to obtain national funding support . Waste management is important in the Maldives not only to mitigate the risk of Aedes-borne infections , but also to maintain the pristine environment underpinning the country’s tourist industry . While the literature documents several studies on preventing and controlling dengue , [3 , 19] economic studies on dengue specifically in SIDS or tourism-based economies are rare . To better understand the economic cost of dengue illness , control and preventive efforts and inform future control efforts , we undertook an economic analysis of dengue prevention and case management in this tourism-based economy .
The Maldives has a well-developed public health system with a publicly-funded health center on each inhabited island , a hospital for each atoll , and both public and private referral hospitals in Malé . Dengue can be treated at each of these health centers . Maldivians are also enrolled in the country’s universal insurance system , Aasandha . In addition , the State Trading Organization ( STO ) , a public company primarily owned by the Government of the Maldives , operates at least one public pharmacy in every inhabited island where Maldivians can obtain prescribed medicines . The country’s latest ( 2016 ) gross national income per capita was US $10 , 630 . [20] We visited health offices and facilities across the country’s health care spectrum in December 2016 . These comprised island health centers and local council offices ( on Haa Dhaalu [HDh . ] Hanimaadhoo , Kaafu [K . ] Dhiffushi and K . Maafushi ) , a resort island ( Thulhagiri ) , a regional hospital ( on HDh . Kulhudhuffushi ) and the Indira Gandhi Memorial Hospital ( IGMH ) , the country’s main referral hospital ( in Malé ) . The location of each of these sites is marked in Fig 1 . We further conferred with senior officials from the HPA , National Bureau of Statistics , Allied Health Insurance , Aasandha Health Insurance , the STO , the Maldives Association of Travel and Tour Operators , Ministry of Environment and Energy , and the Ministry of Tourism . We obtained aggregate statistical data on dengue cases from the Maldives HPA . Population data for atolls were obtained from the Maldives National Bureau of Statistics 2014 Census . The Maldives National Bureau of Statistics Housing and Household Characteristics Statistical Release 2014 provided data on waste disposal and water sources . To analyze these data , we classified atolls as “high risk” for waste disposal when they disposed of their waste “in open garbage sites” or “on the beach or in the bush” . We designated atolls as “high risk” for water storage when there was “rain water collection” or “presence of an open well . ” We analyzed the relationship between risk factors related to waste disposal and water sources on dengue by regressing the six-year average incidence rate on these risk factors . To ensure comparable observations , we excluded the outlier atolls South Ari and South Thiladhunmathi as well as the capital , Malé . This study adopted a bottom-up costing approach . First , all elements of the dengue control program were identified . Thereafter , data on resource utilization and unit costs of each resource were obtained or derived for 2016 . Total program costs were then derived from the sum of the product of resource utilization and unit costs for each element . Data collected included both capital and recurrent expenditures for dengue control activities . Buildings were assumed to have a 20-year useful lifespan as they were generally small , made of local materials , and faced constant sun and humidity . The one piece of equipment ( thermal fogger ) was assumed to have a 10-year useful life . We recorded data for resource use and costs at the district level in a matrix by line item and function . All items combine amortized capital costs ( e . g . , buildings and equipment ) and recurrent ( e . g . , utilities , fuel , and maintenance ) costs . Thereafter , costs for the line items were summed up to provide the total cost of dengue control activities for each district . Data on vector control activities and scenarios about hypothetical typical cases were collected through structured interviews with managers of health centers , hospitals , island councils , and a resort . Additional financial statements and database extracts were obtained from the STO , Aasandha and IGMH . The dengue fraction reflects the supervisors’ best estimate of the share of the relevant staff time devoted to dengue control activities . All data were entered into Microsoft Excel and the statistical software R for analysis . [21] Finally , we estimated the impact of dengue on the tourism sector using a regression coefficient from the international literature . [13] The study did not access or use any individual patient data .
The number of officially reported dengue cases by year from 2011 through 2016 was 2909 , 1083 , 680 , 775 , 1881 , and 1931 , respectively , with an annual average of 1 , 543 cases . Fig 2 depicts the incidence of dengue cases per 100 , 000 inhabitants of each atoll from 2011–2016 . Based on these numbers , the average number of dengue cases per month per island health center from 2011–2016 was 0 . 98 . The total resident population of the Maldives was 417 , 492 in 2016 . [20] Data from the Maldives provide suggestive evidence on the effectiveness of two source reduction strategies in controlling dengue: waste removal and water source management . Fig 3A and 3B depict dengue incidence nationally against means of waste disposal and water management , respectively . The trends suggest that for both higher percentage risky waste disposal ( p = 0 . 06 ) and higher percentage risky water management ( p = 0 . 08 ) are associated with elevated risk of dengue infection . Table 1 shows the estimated annual cost of dengue surveillance based on staffing needs at national , regional , atoll , and island levels . The surveillance costs measured relate to the staffing and information technology required to capture , analyze and disseminate epidemiological data . The total cost for dengue surveillance is $1 , 338 , 141 per year , or $3 . 27 per Maldivian resident per year . Local initiatives , like those on the island of Hdh . Hanimaadhoo ( inhabitants: 1 , 800 ) , demonstrated effective systems of source reduction–regular waste collection run by the local council . For a small fee , residents put out their waste and laborers gather it from public places . It is loaded into trucks and taken to a central place on the island for storage and , in the future , for recycling . The costs include laborers , supervision and vehicles and are described in Table 2 . This initiative costs US$35 . 93 per island resident per year . A source reduction initiative , addressing household water storage and related breeding sites , involved regular household visits ( approximately twice per year ) to show homeowners safe water storage practices and explaining other measures to reduce breeding sites . This includes application of larvicides; Bacillus thuringiensis israelensis or temephos ( from 2016 onwards ) provision of larvivorous fish for households with wells or fresh water holding bodies and instruction about sealing the pipe connections into water tanks to prevent mosquito entry . Table 2 details the cost of these initiatives , resulting in US$7 . 89 per island resident per year . The Maldives HPA currently recommends space spraying ( thermal fogging or using a mist blower ) as a measure only during an outbreak and in targeted priority locations . In a prior year , authorities on K . Dhiffushi conducted fogging to reduce adult mosquito populations . As shown in Table 2 , fogging cost US $0 . 15 per year per island resident . To reduce risks to guests and maintain the reputation of the resort as well as the country as a whole , resort islands engage in extensive vector control . The costs of these activities are described in Table 2 . Thulhagiri resort employed 2 full-time staff for waste reduction , breeding site identification and gutter cleaning for the sole purpose of mosquito control for 160 guests . Another 7 employees work full time to ensure clean facilities are maintained , of which 1/7th of their time is dedicated to cleaning potential vector breeding sources . With employees’ benefits including free accommodation and food , these vector control activities cost $1 . 60 per hotel-guest night . To project the cost of extending the local waste management and household visit programs nationally , we estimated that scaling could halve the per capita costs from $43 . 82 ( $35 . 93 for waste collection plus $7 . 89 for house visits ) to $21 . 91 . The resulting national cost would be $9 . 1 million annually . Similarly , the projected vector control costs for resort islands would be $11 . 2 million annually , based on 1 , 234 , 248 tourists staying for an average of 5 . 7 days in 2015 . [12] Table 3 first examines cost at the lowest level of the system , a health center , on the inhabited island of K . Dhiffushi . Two types of cases are considered: an ambulatory mild case of dengue , and a hospitalized case of dengue . An ambulatory case is estimated to consult the medical officer in the island health facility two times , be tested for NS1 antigen and receive a simple prescription , resulting in an overall economic cost of $49 . 87 per ambulatory dengue case . A hospitalized case is also estimated to consult the medical officer in the island health facility two times , be tested for NS1 antigen , receive prescriptions , and stay at the health facility for one night . This resulted in an overall economic cost of $127 . 74 per hospitalized dengue case . Aasandha payments for the product or service , where available , are shown as financial costs . Table 3 also analyzes costs at a regional hospital . The hospital had 50 beds with 60% occupancy 365 days per year generating 10 , 950 bed-day equivalents . For outpatient services , it had 76 , 026 annual outpatient visits ( about 255 per weekday ) , each generating 0 . 32 bed day equivalents , [22] or 25 , 480 bed day equivalents . In total , the hospital produced 36 , 430 bed day equivalents , of which 30% were generated by inpatient stays and 70% by outpatient visits . The estimated economic cost of one hospitalized dengue case is US$1 , 164 . 78 . Finally , Table 3 examines the economic cost of care at IGMH for 2015 . Assuming an average stay for a dengue patient of 4 days , the 2015 healthcare cost of a hospitalization within IGMH was $1 , 655 . 50 . The derivation of these costs is presented in Supporting Information file S1 Table . In addition to funding conventional services , Aasandha also covers certain specialized services needed because of the Maldives’ status as a SIDS with a small and dispersed population . Table 4 shows the quantities and costs of these services: sea evacuations , air evacuations , and overseas hospitalizations for 2015 . With only 140 evacuations in one year , the quantity of these specialized services is low , but their aggregate cost is substantial at $116 , 901 in 2015 . The most expensive cost was overseas hospitalization to Sri Lanka ( 89% ) and India ( 11% ) . The average cost of the sea evacuations ( $519 ) was almost as high as that for the air transfers ( $628 ) . Using epidemiological data , unit costs of hospitalization and ambulatory care from this study , and indirect costs of dengue cases infection derived from a systematic literature review , [3] we calculated the overall cost of dengue illness in the Maldives in 2015 , as shown in Table 5 . We estimated a total cost of $2 , 495 , 747 , of which 48% was direct costs , 5% for emergency evacuations , and 47% indirect costs . The Maldives’ largest economic sector , tourism , brought a gross income of US $375 million ( $898 per resident ) and tax receipts of US$50 million ( $119 per resident ) in 2014 . [12] Based on the aforementioned regression analysis , [13] the risk of dengue lowers the country’s gross annual income by $44 million ( $110 per resident , 95% confidence interval $50 to $160 ) and its annual tax receipts by $6 million ( $14 per resident , 95% confidence interval $7 to $22 ) .
To our knowledge , this is the first economic evaluation of dengue prevention and control in the Maldives , a SIDS actively working to reduce the burden of vector borne diseases . [23] Using data from 2015 , we found the cost of dengue in the Maldives to be $2 , 495 , 747 ( $6 . 10 per resident ) for dengue illness plus $1 , 338 , 141 ( $3 . 27 per resident ) for surveillance . The study results suggest that the economic costs through depressing tourism are substantially greater than the economic cost of illness . This paper describes four mosquito source reduction strategies seen in the Maldives: ( 1 ) waste collection by island councils on inhabited islands , ( 2 ) household visits by community health workers in inhabited islands , ( 3 ) thermal fogging of insecticides on inhabited islands and ( 4 ) intensive waste and water management on a resort island . The costs of these interventions varied based on the setting: inhabited islands spent from $0 . 15 to $35 . 93 per person per year , while a resort island invested $1 . 60 per hotel guest night . The projected cost of scaling the local waste management and household visit programs nationally was $9 . 1 million annually , similar to the estimated $11 . 2 million already spent protecting tourists on resort islands . Costs on resort islands are higher than those inhabited by regular residents due to the resort islands’ imperative to maintain a clean and safe environment for tourists , higher wages , and more comprehensive benefits for workers , while island councils on inhabited islands must allocate their resources across a range of services . Sharing best practices could likely reduce future costs compared to these projections . Locations such as the inhabited island of Hdh . Hanimaadhoo and the resort island of Thulhagiri operate model integrated vector management programs , providing tourists staying on these islands with the best possible protection from dengue . These tourists , however , frequently visit neighboring inhabited islands , thereby increasing the risk of infection . By sharing best practices between resort islands and other inhabited islands , both tourists and locals would enjoy greater dengue protection . Regular surveillance and monitoring would allow both resort islands and the HPA to evaluate and refine their efforts . As climate change likely increases the burden of vector borne diseases , [24] extending effective vector control to all islands will become more important in the future . Financial and economic costs of dengue case management varied depending on the location and case classification . For mild dengue cases presenting at island health centers , the cost was $49 . 87 per ambulatory case , and almost tripled to $127 . 74 for hospitalized cases . Case management of complicated dengue patients in regional or national hospitals significantly increased the cost to $1 , 164 . 78 and $1 , 655 . 50 per hospitalized case , respectively . This is more expensive than the closest neighboring island , Sri Lanka , where average costs per hospitalization are between US$196 to $866 for adult cases , depending on disease severity and treatment setting . [25] Medical fees for non-citizens are double those for citizens , and Aasandha does not cover non-citizens . For example , health staff on K . Maafushi island estimated that foreign workers faced an incidence of symptomatic dengue 30 times that of its Maldivian residents ( i . e . , 300 versus 10 per 100 , 000 population per year ) . Similarly in Singapore , migrant workers are more at risk to arbovirus infection . [26] Besides a higher risk of contracting dengue , foreign workers are often reported to delay treatment , so their risk of serious illness , and possibly death , is higher . Migrant workers in the Maldives often come from dengue endemic countries and may work on several islands during their stay , thereby increasing the risk of dengue transmission throughout the country . With its 187 inhabited islands , the Maldives necessarily relies on boats , seaplanes and airplanes for transportation . Although scheduled routes exist , medical transports frequently cannot wait and require ad-hoc emergency evacuations of patients from island health facilities to referral hospitals , including chartering a speed boat . In our study we found that despite a high number of evacuations , these contributed to only 7% of total costs . We found plane evacuations more economic than boats , likely due to difference in costs of purchasing a plane ticket on scheduled flights compared to the cost of chartering a speed boat . Hospitalizations of complicated dengue cases overseas in our study cost on average $1 , 225 . This is in line with the previously reported average cost of $1 , 470 to obtain treatment of any kind in overseas health facilities from the Maldives . [27] The aggregate cost of overseas hospitalizations for dengue in 2015 ( $15 , 926 ) , however , forms only a small part of the estimated $68 . 9 million spent on overseas hospitalization for all conditions for Maldivian citizens . [27] We estimated that each island health facility clinician sees on average of only one dengue case per month . Due to this low frequency and high turnover of clinicians on these islands , investing resources in strengthening clinical capacity for dengue would be relatively ineffective . Maldivian authorities have implemented a “dengue hotline” to allow inexperienced clinicians to seek advice from experts at IGMH . Similar health hotlines exist in the UK[28 , 29] and the USA , [30] which contribute to the triage and surveillance of cases . The current limitation of the hotline in the Maldives , however , is that calls are not screened , and island clinicians use the service for complicated dengue cases only . An expanded service could advise callers on interpretation of symptoms and test results to assess the likelihood of dengue , appropriate treatment to minimize risk , and decide appropriately on transfers . Based on international experience , we expect that an enhanced hotline would prove favorable in terms of better patient outcomes , less disruption to patients and families , and lower costs to the Maldivian health care system . We suggest that the Ministry of Health consider introducing an enhanced hotline phased in by atolls covered , monitoring hotline calls under both existing and enhanced phases to examine utilization , health outcomes , and costs of services and evacuations , and extending and refining the enhanced hotline based on lessons learned . This study is limited in that only a selected number of interventions were costed , and as such , may not reflect the costs seen on each island . Furthermore , this study did not address the burden of other important arboviruses , such as chikungunya and Zika , which also have been reported in the Maldives . [31 , 32] Including this in the analysis would likely increase the cost of case management , but not the cost of vector control activities , as the diseases share the same mosquito vector . In estimating the total cost of dengue in the Maldives , some of the data items were not available for the country and had to be derived from results in other countries . Using country-specific data for those items , if available , would strengthen the analysis . Based on the importance of the tourism sector the need to mitigate the potential impact of climate change , the Maldives introduced a “green tax , ” charging tourists on resort islands and guest houses $6 and $3 per person per night , respectively . While this tax revenue does not directly support the health sector , it is levied to strengthen solid waste management throughout the Maldives . By reducing plastic and other receptacles that are used by Aedes mosquitos to breed , this tax indirectly lowers the burden of dengue . Some additional transport or tourism tax could be established , if necessary , to fund the time , travel , and follow-up expenses needed to expand vector control services nationally . This study reinforces the economic rationale for investment in effective dengue control .
|
As tourism is the mainstay of the Maldives’ economy , this country recognizes the importance of controlling mosquito-borne diseases in an environmentally responsible manner . This study sought to estimate the economic costs of dengue in this Small Island Developing State of 417 , 492 residents with an annual average of 1 , 543 reported symptomatic dengue cases . Overall , the cost of dengue illness in the Maldives in 2015 was $3 million ( $6 . 10 per resident ) and surveillance cost an additional $1 million ( $3 . 27 per resident ) . The risk of dengue lowers the country’s gross annual income by $110 per resident and its annual tax receipts by $14 per resident . Rigorous elimination of debris on some resort islands demonstrates effective and environmentally sound vector control . Many innovative vector control efforts are affordable and could decrease future costs of dengue illness in the Maldives .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"geomorphology",
"medicine",
"and",
"health",
"sciences",
"landforms",
"topography",
"geographical",
"locations",
"social",
"sciences",
"biological",
"locomotion",
"health",
"care",
"infectious",
"disease",
"control",
"islands",
"infectious",
"diseases",
"health",
"economics",
"epidemiology",
"economics",
"walking",
"people",
"and",
"places",
"infectious",
"disease",
"surveillance",
"asia",
"earth",
"sciences",
"physiology",
"disease",
"surveillance",
"biology",
"and",
"life",
"sciences",
"atolls",
"maldives"
] |
2018
|
Economic analysis of dengue prevention and case management in the Maldives
|
Integrative and Conjugative Elements ( ICEs ) of the SXT/R391 family disseminate multidrug resistance among pathogenic Gammaproteobacteria such as Vibrio cholerae . SXT/R391 ICEs are mobile genetic elements that reside in the chromosome of their host and eventually self-transfer to other bacteria by conjugation . Conjugative transfer of SXT/R391 ICEs involves a transient extrachromosomal circular plasmid-like form that is thought to be the substrate for single-stranded DNA translocation to the recipient cell through the mating pore . This plasmid-like form is thought to be non-replicative and is consequently expected to be highly unstable . We report here that the ICE R391 of Providencia rettgeri is impervious to loss upon cell division . We have investigated the genetic determinants contributing to R391 stability . First , we found that a hipAB-like toxin/antitoxin system improves R391 stability as its deletion resulted in a tenfold increase of R391 loss . Because hipAB is not a conserved feature of SXT/R391 ICEs , we sought for alternative and conserved stabilization mechanisms . We found that conjugation itself does not stabilize R391 as deletion of traG , which abolishes conjugative transfer , did not influence the frequency of loss . However , deletion of either the relaxase-encoding gene traI or the origin of transfer ( oriT ) led to a dramatic increase of R391 loss correlated with a copy number decrease of its plasmid-like form . This observation suggests that replication initiated at oriT by TraI is essential not only for conjugative transfer but also for stabilization of SXT/R391 ICEs . Finally , we uncovered srpMRC , a conserved locus coding for two proteins distantly related to the type II ( actin-type ATPase ) parMRC partitioning system of plasmid R1 . R391 and plasmid stabilization assays demonstrate that srpMRC is active and contributes to reducing R391 loss . While partitioning systems usually stabilizes low-copy plasmids , srpMRC is the first to be reported that stabilizes a family of ICEs .
Integrative and conjugative elements ( ICEs ) are highly prevalent and widely distributed in bacterial genomes [1–3] . Their ability to self-transfer by conjugation between genetically unrelated bacteria contributes to the emergence of multidrug resistant pathogens in diverse taxonomic groups [4–6] . ICEs usually reside within and replicate with the host cell’s chromosome to be vertically inherited . ICEs eventually excise from the chromosome and form circular covalently closed molecules that serve as the substrate for the conjugative machinery that translocates the ICE DNA to recipient cells [6 , 7] . With a few exceptions reported only for ICEs of Actinobacteria , this conjugative machinery usually consists of a relaxase , a type IV coupling protein and a type IV secretion system [1–3 , 8] . The SXT/R391 family of ICEs encompasses one of the largest and most diverse set of ICEs studied , including elements that have been found over the past 40 years in clinical and environmental isolates of diverse species of Gammaproteobacteria [9 , 10] . ICEs of the SXT/R391 family largely contribute to the spread of antibiotic resistance genes in the seventh-pandemic lineage of Vibrio cholerae , the etiologic agent of cholera , which remains a major cause of mortality and morbidity on a global scale [11] . The ICE SXT is a prototypical member of the SXT/R391 family originally isolated from a 1992 Indian multidrug resistant clinical isolate of V . cholerae O139 [12] . SXT and several variants detected in V . cholerae O139 , O1 and non-O1 non-O139 isolates confer resistance to sulfamethoxazole , trimethoprim , streptomycin and chloramphenicol [9] . The second prototypical member of this family is R391 , which was originally detected in a 1967 South African isolate of Providencia rettgeri [13] . R391 confers resistance to kanamycin and mercury . Members of the SXT/R391 family all share a common integration site , the 5’ end of prfC , and a highly conserved core of genes and sequences that mediate their regulation , integration/excision and conjugative transfer [10] . Expression of the conjugative function of SXT/R391 ICEs is tightly regulated by SetR , which represses the expression of the master activator genes setC and setD . Their products activate transcription of int , xis and conjugation-associated operons [14] . Repression of setC and setD is alleviated by induction of the bacterial response to DNA damage , which promotes autoproteolysis of SetR [15] . SXT and R391 can exist co-integrated in a tandem fashion in prfC in the same host cell [16 , 17] . Such tandem arrays are suitable substrates for frequent homologous recombination events yielding hybrid ICEs that can be easily segregated in exconjugant cells [16 , 18] . Interestingly , R391 was reported to be found as a circular extrachromosomal replicative form in a recA recipient strain bearing an integrated copy of R997 , another SXT/R391 ICE found in Proteus mirabilis [19] . A similar behavior was also reported for R997 entering a recA recipient bearing an integrated R391 . However , no extrachromosomal form of R391 or R997 could be recovered from recA+ hosts . These observations suggest that , at least in specific circumstances , SXT/R391 ICEs are capable of autonomous replication . Autonomous replication was previously suspected for several ICEs and recently well characterized for ICEBs1 , an ICE of the Gram-positive bacterium Bacillus subtilis [20–25] . Plasmid-like replication was also shown to be essential for the stability of ICEBs1 [24] . However , whether autonomous replication is relevant to the biology and stability of SXT/R391 ICEs remains to be established . Breaking with old paradigms about ICEs , we report here that replication is a key step of the lifecycle of SXT/R391 ICEs by using R391 as a model . By monitoring the frequency of excision , the ICE copy number as well as the frequency of loss of a set of mutants , we show that the putative relaxase TraI and the origin of transfer ( oriT ) are essential for R391 replication and its stability in the progeny of host cells . Furthermore , we demonstrate that , besides diverse non-conserved toxin-antitoxin systems , all SXT/R391 ICEs also encode a conserved plasmid-like type II partitioning system that enhances their stability . Together , these results unravel an unforeseen similarity between the biology of ICEs and conjugative plasmids .
To have a better understanding of SXT/R391 ICEs biology , we evaluated five key factors of R391 lifecycle in Escherichia coli: ( i ) the dynamics of excision/integration , which will be reported as the frequency of excision in the rest of the manuscript , ( ii ) the frequency of transfer , ( iii ) the average copy number per cell in the whole cell population , ( iv ) the average number of extrachromosomal circular copies per cell , and ( v ) the ICE stability in the cell population . The frequency of R391 excision was assessed by quantifying by real-time quantitative PCR ( qPCR ) the relative amount of free integration site ( attB ) resulting from R391 excision per chromosome as measured by the amount of chromosomal prfC target ( Fig 1A ) . R391 excised at a frequency of 1 . 90×10-3 ± 0 . 38×10-3 ( Fig 2A ) , which is about tenfold lower than the excision frequency of SXT ( 1 . 76×10-2 ± 0 . 65×10-2 , P = 0 . 0140 , two-tailed Student t-test ) in similar conditions . Mating assays showed that R391 transfers at about 5 . 01×10-4 ± 0 . 31×10-4 exconjugant/donor ( Fig 2B ) , which is about 20 fold higher than SXT ( 2 . 71×10-5 ± 0 . 55×10-5 exconjugant/donor ) . Hence there is no correlation between the frequency of excision of these elements measured in the whole cell population and their respective frequency of transfer . This observation indicates that excision is not a factor that limits the rate of dissemination of these two ICEs . Using the same approach , we then measured the mean copy number per cell of R391 in the whole cell population as the ratio between the amount of R391-borne int target and the amount of chromosomal prfC target ( Fig 1A ) . This ratio was found to be 0 . 96 ± 0 . 04 as expect for a single copy of R391 integrated in the chromosome . We also measured the mean copy number of the extrachromosomal circular form of R391 per cell by establishing the ratio between the amount of attP recombination site resulting from R391 excision and the amount of unoccupied chromosomal attB sites . In theory , each event of R391 excision is expected to yield one unoccupied attB site on the chromosome and one attP site on the circular excised R391 ( attP/attB = 1 ) . We observed that this ratio reached 21 ± 2 ( Fig 2C ) , suggesting that R391 is capable of replicating in a small subset of the cell population in which it is excised from the chromosome . This observation is consistent with results previously reported for SXT , for which 4 attP sites on average exist for each unoccupied attB site [16 , 26] . New measurements carried out in this study to confirm these reports revealed 3 . 6 ± 0 . 2 attP sites per unoccupied attB site for SXT . We then assessed the stability of R391 by monitoring the number of cells lacking R391K in the cell population after 16 hours of growth ( about 20 generations ) in LB medium with or without selective pressure . R391K is tagged with the galK reporter gene under the control of the Plac promoter to enable high-level galactokinase activity in a lacI mutant strain such as E . coli VB38 [18] , a ΔgalK derivative of E . coli CAG18439 ( lacI42::Tn10 ) ( Fig 1B ) . The frequency of loss was determined as the percentage of white colonies ( galK- , devoid of R391K ) on MacConkey indicator agar supplemented with 1% galactose ( Fig 1C ) . R391K was found to be inherently stable because it was lost in only 0 . 0037% of the cell population in the absence of selective pressure ( Fig 2D , WT ) , whereas no detectable loss was observed when cells were grown with kanamycin in liquid culture ( detection limit of 0 . 0001% ) . Stability of many mobile genetic elements relies on a post-segregational killing mechanism , which induces a strong selective disadvantage or even death to cells that have lost them [27–31] . While previous studies have shown that two functional toxin-antitoxin ( TA ) systems , mosAT and s045-s044 , enhance the stability of SXT [32 , 33] , neither of these TA systems was found in R391 . Nevertheless , in silico analysis of the R391 sequence revealed that the two overlapping open reading frames ( ORFs ) orf02 and orf03 , which belong to the variable region I located upstream of xis , encode a putative hipAB-like TA system ( Fig 1B ) [34 , 35] . Indeed , orf03 ( hipA ) is predicted to encode a HipA-like toxin , while orf02 ( hipB ) likely codes for the HipB cognate antitoxin , which carries a DNA-binding HTH-XRE ( HTH_19 ) domain . To measure the impact of this putative hipAB-like TA system on R391 stability , we constructed a ΔhipA mutant of R391K . This mutation did not impair the transfer of R391K and had no effect on the excision or extrachromosomal copy number of the element ( Fig 2A , 2B and 2C , compare WT and ΔhipA ) . However , ICE stability was affected as R391K loss increased by 12 fold for the ΔhipA mutant compared to wild-type ( Fig 2D ) . No loss of R391K ΔhipA was detectable in the presence of kanamycin . These results revealed the functionality of the hipAB TA system and its involvement in the stability of R391 , as previously demonstrated for mosAT of SXT [32] . However , like the mosAT and s045-s044 loci of SXT , hipAB of R391 is not a conserved feature of SXT/R391 ICEs; therefore hipAB is likely not an inherent and ancestral mechanism used by SXT/R391 ICEs to enhance their stability in their respective hosts . Cell death or growth reduction associated with hipAB after R391 loss was likely to hinder our investigations on R391 stability . To circumvent this issue , the ΔhipA mutant provided us with a useful tool for additional investigations aimed at unraveling other stabilization mechanisms conserved among ICEs of the SXT/R391 family . Conjugation has been shown to be a powerful stabilization mechanism for conjugative plasmids that can reenter in cells having lost them by infectiously spreading in the cell population [28 , 36] . While Wozniak and Waldor [32] have shown that conjugation does not promote SXT loss , their assay does not allow to conclude whether conjugation is an efficient mechanism of stabilization of SXT/R391 ICEs . To answer this question , we looked at the frequency of loss of a ΔtraG mutant of R391K . traG codes for an inner-membrane component of the donor cell mating pair formation apparatus that is essential for SXT transfer [37] . As previously reported for SXT , deletion of traG abolished R391K transfer ( Fig 2B ) . However , the inability to transfer did not reduce the stability of R391K , as the frequencies of loss of wild-type R391K and its ΔtraG mutant were nearly identical ( Fig 2D ) . No detectable loss of R391K ΔtraG was observed when kanamycin was added during liquid culture . The frequencies of excision of the ΔtraG and ΔhipA ΔtraG mutants were reduced by ~2 . 3 fold , while their extrachromosomal copy numbers were more than twice as high as wild-type R391K and its ΔhipA mutant , reaching up to ~54 copies per cells ( Fig 2A and 2C ) . This observation suggests that , once excised from the chromosome , the circular form of R391K accumulates in the cell possibly because a defective mating apparatus caused by the ΔtraG mutation cannot mediate its transfer to a recipient cell . The plasmid-like replication of ICEBs1 was shown to be essential for its stability [24] . Rolling-circle replication of ICEBs1 requires the relaxase NicK and oriT , a cis-acting locus initiating the translocation of DNA through the mating pore . In a subset of a cell population , R391 seems to be in a multicopy plasmid-like state that may be important for preserving ICE stability when it is excised from the chromosome in actively dividing cells . To test this hypothesis , oriT and traI deletion mutants were constructed in R391K ( Fig 1B ) . traI codes for the putative relaxase of SXT/R391 ICEs that recognize the origin of transfer ( oriT ) [38] . As expected , the ΔtraI mutation abolished R391K conjugative transfer ( Fig 3B ) . We also observed a ~5-fold reduction of the extrachromosomal copy number of R391K when either traI or oriT were missing compared to wild-type ( Fig 3C ) . The ΔtraI mutation also led to a ~20-fold increase of the frequency of excision and to an 11-fold increase of R391K loss ( Fig 3A , 3B and 3D ) . Combined ΔtraI and ΔhipA mutations led to a 34-fold increase of R391K loss , thereby confirming that traI is important for R391 stability ( Fig 3D ) . Expression of traI in trans from the arabinose-inducible PBAD promoter in pTraI restored and even enhanced the transfer and the stability of both ΔtraI and ΔtraI ΔhipA mutants compared to wild-type ( Fig 3B and 3D ) . Although the copy number of the plasmid-like form of R391K ΔtraI doubled upon complementation with pTraI , it failed to reach the wild-type level ( Fig 3C ) . Interestingly , ΔtraI mutants were so unstable that selective pressure exerted by kanamycin in liquid culture did not , or only slightly , improved R391K stability ( Fig 3D ) . The high instability affecting ΔtraI mutants also led to the formation of sectored colonies on agar plates likely resulting from loss of R391K during colony development ( Fig 1C ) . Since conjugative transfer of SXT/R391 ICEs is known to be stimulated by DNA-damaging agents , we tested the effect of mitomycin C on the stability of the ΔhipA , ΔtraG and ΔtraI mutants of R391K . We observed that the drug did not induce high-frequency loss of the ΔhipA , ΔhipA ΔtraG or ΔtraI mutants of R391K ( Fig 3E ) . In striking contrast , deletion of both ΔhipA and ΔtraI led to a hypersensitivity of R391K to mitomycin C treatment as the ICE was lost in more than 90% of the cell population ( Fig 3E ) . We suspect that ΔtraI mutants are highly unstable; yet in the presence of hipAB , cells that have lost R391K ΔtraI likely have no progeny or strong growth reduction due to the persistence of the HipA toxin , thereby masking this high instability in conditions that strongly induce R391 excision . In silico analysis of R391 sequence using CD-search on the Conserved Domain Database v3 . 11 [39 , 40] and protein fold recognition server Phyre2 [41] revealed that orf07 , the first gene of an operon containing int , codes for a predicted actin-like NTPase structurally related to the ParM plasmid segregation proteins of plasmids R1 and pSK41 ( Fig 1B ) . ParM proteins are a key component of type II ParMRC partitioning systems that mediate plasmid DNA segregation during cell division via a pushing mechanism [42 , 43] . ParR adaptor protein connects parC , a cis-acting centromere-like locus , to the ParM filament . ParR proteins have low conservation and their genes are found downstream of the parM gene . The open reading frame orf06 , located downstream of orf07 , is predicted to code for a small basic protein ( pI 9 . 3 ) with no recognizable domain ( Fig 1B ) . Hence , orf06 may encode a ParR DNA-binding protein that binds the centromere-like region in partitioning systems . Based on these observations and results described below , orf06 and orf07 were renamed srpR and srpM for SXT/R391 ICEs partitioning proteins R and M , respectively ( Fig 1B ) . By functional analogy with the parMRC partitioning systems carried by the plasmid R1 of E . coli and the staphylococcal plasmid pSK41 [42] , the DNA fragment located upstream of srpM likely corresponds to the centromere-like region bound by SrpR and was annotated srpC ( Fig 1B ) . The srpMRC locus is strictly conserved in all SXT/R391 ICEs , suggesting that it may somehow play an important role in their biology . Deletion of srpR , srpM or both had no measurable effect on the frequency of excision or the extrachromosomal copy number of R391K ΔhipA , and had a slight inhibitory effect ( about 4- to 14-fold reduction ) on the frequency of transfer ( Fig 4A , 4B and 4C ) . The lack of impact on the excision frequency confirmed that neither deletion has a polar effect on the expression of the int gene located immediately downstream of srpR ( Figs 1B and 4A ) . Since the deletion of srpM had very little effect on R391 transfer , we used this mutation to further study the phenotype associated with a non-functional srpMRC locus . Cumulative mutations of traG , hipA and srpM did not affect the frequencies of excision or loss of R391K compared to the ΔtraG ΔhipA mutant ( Fig 4A and 4D ) . However , the extrachromosomal copy number of R391K dropped nearly 4 fold when comparing the same mutants ( Fig 4C ) . Furthermore , a ΔtraI ΔhipA ΔsrpM triple mutant exhibited a visible but statistically non-significant 35% reduction of stability compared to the ΔtraI ΔhipA mutant ( Fig 4D ) . However , deletion of srpM led to a 3-fold increase of the stability of R391K ΔtraI ΔhipA after 40 generations ( Fig 4D , dark grey bars ) . Finally , overexpression of traI from pTraI did not completely prevent the loss of R391 ΔtraI ΔhipA ΔsrpM , which was lost about 4 times more frequently than the ΔtraI ΔhipA mutant ( Fig 4D ) . These data revealed that srpM is important for R391 stability when the number of copy of the ICE is low and thus could be a functional active partition system . To test further whether srpMRC is a functional DNA partitioning system , plasmid stabilization assays were carried out using pBeloBAC11Δsop , an unstable derivative of the single-copy plasmid pBeloBAC11 that lacks its native sopABC partitioning system . In SXT/R391 ICEs , expression of srpM , srpR and int was shown to be driven from Ps003 , a promoter exclusively dependent upon activation by the transcriptional activator SetCD ( Fig 1B ) [14] . To bypass the need for SetCD , the srpMRC loci of R391 and SXT were cloned into pBeloBAC11Δsop downstream of the IPTG-inducible Plac promoter ( Fig 4E ) . Expression of srpMRC loci of SXT ( pSrpSXT ) or R391 ( pSrpR391 ) upon IPTG induction led to a respective ~1 . 8 and ~2 . 1-fold increase of pBeloBAC11Δsop stability , thereby confirming that srpMRC is a functional plasmid stabilization system ( Fig 4E and S1 Fig ) . The absence of srpR , srpM or srpC prevented plasmid stabilization , which was then comparable to the empty vector ( Fig 4E and S1 Fig ) . SrpR lacks homologies with known ParR proteins that have been shown to bind parC-centromere-like sequences upstream of parMR genes in plasmids such as R1 . To test whether SrpR is capable of binding the srpC locus , we carried out electrophoretic mobility shift essays ( EMSA ) experiments using purified C-terminally 6xHis-tagged SrpR protein ( predicted molecular weight of 10 . 4 kDa ) . EMSA assays revealed a specific binding of SrpR to the 615-bp fragment ig ( srpM-mobI ) which corresponds to the intergenic region between srpM and mobI and likely contains the srpC region ( Fig 5A and 5BI ) . Addition of high concentrations of sonicated salmon sperm DNA ( non-specific competitor ) did not destabilize SrpR binding to this probe . Further investigation confirmed that SrpR specifically binds the 251-bp srpC region as SrpR binding to srpC was resilient to the addition of the non-specific competitor DNA ( Fig 5A and 5BII ) . The presence of multiple specific shifts suggests that SrpR binds multiple sites or binds as different multimeric forms ( Fig 5BII ) . While SrpR was able to bind to the 298-bp fragment containing oriT , addition of the non-specific competitor DNA destabilized SrpR binding ( Fig 5BIII ) . Non-specific SrpR binding to oriT indicates that SrpR exhibits a significant non-specific affinity for DNA molecules . In silico analysis of the centromere-like srpC region of R391 and SXT using the Multiple Em for Motif Elicitation tool ( MEME ) [44] revealed four conserved 13-bp direct and inverted repeats that might be recognized by SrpR ( Fig 5C and 5D ) . EMSA results showed that a 40-bp fragment containing either R1-R3 or R2-R4 was bound by SrpR ( Fig 5E ) . Addition of competitor DNA strongly decreased SrpR binding but did not completely alleviate the interaction . Binding of SrpR to the sequences R2 or R4 used as probes was abolished by the addition of the competitor suggesting that half-sites do not produce stable complexes with SrpR ( Fig 5E ) . Since ParR-like DNA binding proteins have been shown to form multimeric complexes [45] , SrpR multimerization assays were carried out using glutaralehyde cross-linking . These assays suggests that , even without any DNA substrate , SrpR seems to be able to assemble as dimeric and tetrameric complexes in solution as shown by the apparition of large bands migrating at compatible molecular weights in a SDS page gel ( Fig 5F ) . Copious amounts of SrpR were also trapped in the well , thereby suggesting that SrpR could be able to assemble in complexes of higher order . To assess the relationship of SrpMRC with other type II partitioning systems , we carried out a phylogenetic analysis based on the ParM actin-like homologs found by BlastP . Since the ParR adaptor proteins and parC sequence usually retain low conservation , they were not included in the analysis . Our analyses revealed that as expected , SrpM of R391 clusters with close orthologs encoded by all SXT/R391 ( Fig 6A , green branch ) . SrpM is also closely related to ParM orthologs encoded by a putative type II partitioning system carried by conjugative plasmids of the IncA/C ( Fig 6A , red branch ) and pAQU groups [46–48] . SrpM and all of these orthologs cluster with more distantly related plasmids such as Rts1 ( IncT ) [49] and the catabolic plasmids pCAR1/pDK1 ( IncP7 ) [50 , 51] . Interestingly , with the exception of Rts1 , which seems to lack a gene coding for a ParR protein , the genetic contexts of the orthologous parMRC loci in all these mobile elements are strikingly similar , located between traG- and mobI-like genes , thereby supporting their common ancestry and divergent evolutionary pathways ( Fig 6B ) . This large group of related par loci is distantly related those carried by diverse plasmids broadly distributed among bacterial species of Delta- and Gammaproteobacteria , including the most distantly related parMRC systems carried by the conjugative plasmids R1 and R27 ( Fig 6A ) [52–54] .
In our modern world , antibiotics are widespread in most environments , subjecting microorganisms to a strong and constant selective pressure [57 , 58] . ICEs circulating among environmental and pathogenic bacteria can take advantage of this selective pressure by collecting and accumulating antibiotic resistance-conferring genes . The selective advantage conferred by antibiotic resistance enhances the stability of ICEs in their hosts as well as their odds to eventually spread into and invade a new bacterial population . However , ICEs likely predate the antibiotic era and have evolved other means to prevent their loss . Indeed , several ICEs are stably maintained despite the lack of genes coding for any obvious selective advantage for their host [59 , 60] . One strategy of stabilization consists in a tight control of the excision of the ICE from the chromosome . However , too tight a regulation could prevent its efficient dissemination . For ICEs of the SXT/R391 family , excision and transfer were shown to be coupled with the activation of the host’s SOS response [15] . In bacteria such as E . coli , spontaneous induction of the SOS response in the absence of DNA damaging agents has been shown to occur in 0 . 3 to 3% of the cell population [61] , thereby inherently leading to unscheduled excision that is detrimental to ICE stability . Indeed , cell division occurring after ICE excision can generate ICE-free cell lineages , which likely have a competitive advantage in the absence of selective pressure . Between attL and xis , R391 bears genes coding for a HipAB-like TA system that enhances the stability to the ICE as inactivation of hipA increased R391K loss by 12-fold ( Fig 2D ) . hipAB is also found at the same position in ICEVchMex1 , another member of the SXT/R391 family , which does not seem to confer any heavy metal or antibiotic resistance to its original host [10 , 60] . The toxin/antitoxin system mosAT has been shown to strongly improve the stability of SXT [32] . Interestingly , mosAT expression was found to be correlated with activation of SXT excision and conjugative transfer [32] . However , coupling of mosAT expression with SXT excision was later shown to be circumstantial to the activation by SetCD of the expression of the upstream traVA genes [14] . Furthermore , since neither mosAT nor hipAB are conserved in all SXT/R391 ICEs [10] , element-specific TA systems located in variable regions should only be considered as auxiliary determinants of stabilization for this family of ICEs . The same holds true for the tad-ata-type TA system s044-s045 carried by SXT in the variable region located between traIDJ and traLEKB [10 , 33] . In addition to diverse TA systems encoded by variable DNA , we have shown here that SXT/R391 ICEs rely on specific and conserved strategies to enhance their stability within their host genome . Besides the most obvious one , which is their integration within the host chromosome , our data support the notion that SXT/R391 ICEs are not only capable of replication , but they can also actively segregate the resulting plasmid-like forms . Conjugation has been shown to be a possible stabilization mechanism for IncP-1 conjugative plasmids in cell populations , allowing the recolonization of plasmid-free cells [36 , 62] . However , our results show that conjugation is not a key factor for the stability of SXT/R391 ICEs as a traG mutant that is unable to transfer was no less stable than wild-type R391K ( Fig 2D ) . Interestingly , we found that deletion of traG , which prevents translocation of the ICE DNA to the recipient cell , had unexpected side effects . In such mutants , the frequency of excision decreased while the extrachromosomal copy number increased ( Fig 2A and 2C ) . A plausible explanation for this phenotype is that plasmid-like molecules of R391K can somehow accumulate due to the blocked mating pore . This accumulation of R391K circles would then tend to displace the site-specific recombination reaction from excision toward reintegration , hence the lower excision frequency ( Figs 2A and 7 ) . Consistent with this hypothesis , deletion of traI produced the exact opposite effect . We observed that while deletion of traI drastically reduced the number of R391K circles , the frequency of excision was very much increased ( Fig 3A and 3C ) . Impaired replication of R391K seems to displace the site-specific recombination reaction equilibrium toward excision instead of integration ( Fig 7 ) . We observed that the frequency of excision does not correlate with the frequency of transfer when comparing SXT and R391 . R391 excises at a frequency that is ~10 fold lower than SXT whereas it transfers at a frequency that is ~20 fold higher . We previously reported for SXT that neither its excision from the chromosome of donor cells nor its integration in the chromosome of recipient cells was a step limiting the rate of transfer [14 , 26] . This suggests that assembly of the mating apparatus , initiation of transfer or DNA translocation across the cell membranes through the mating pore was the limiting factor . In fact , our data revealed that availability of TraI is a key regulatory element since ovexpression of traI in cells bearing R391K ΔtraI increased the frequency of transfer by 2 logs over wild-type R391K . This observation is supported by RNA-seq data that revealed the relatively low level of expression of traI compared to other tra genes in SXT , R391 and ICEVflInd1 , another member of the SXT/R391 family [14] . Therefore , initiation of transfer and/or replication , both depending on TraI and oriT , seem to determine the rate of transfer of SXT/R391 ICEs . Deletion of traI or oriT drastically reduced the copy number of R391 circles ( Fig 3C ) , which is consistent with a form of replication initiated at oriT by the relaxase TraI . The process of conjugation usually relies on an intercellular rolling-circle replication of conjugative elements , making their intracellular replication also virtually possible [8 , 63] . Although ICEs were initially defined as non-replicative elements [6] , several recent reports strongly support that single-stranded DNA transferring ICEs can replicate as extrachromosomal plasmid-like molecules , in both Gram-positive and Gram-negative bacteria [20–25] . This replication is initiated at oriT by the relaxase together with other ICE- and host-encoded auxiliary factors [22 , 24] . Notably , the transient replication associated with the conjugative transfer of ICEBs1 of B . subtilis , while not required for transfer , plays an important role in the stability [24] . It relies on oriT used as an origin of replication ( oriV ) and on the conjugative relaxase NicK used as the replication initiator protein . Therefore , the rolling-circle replication module being an intrinsic part of the conjugation module , many ICEs , if not all , might be able to transiently replicate as plasmid-like molecules . Our work revealed yet another intriguing feature of ICEs of the SXT/R391 family besides replication , which seems to blur the frontier between ICEs and plasmids even more . All SXT/R391 ICEs carry srpMRC , a locus coding for a functional active partition system . Contrary to low-copy plasmids such as F , which must actively segregate in the daughter cells following cell division , SXT/R391 ICEs usually remain quiescent , integrated into the chromosome of their host , and are passively passed on from one generation to another . Active partition of these ICEs would only be required in their transient excised state , even more so if their copy number per cell is low , such as in the traI mutant ( Fig 4D ) . In agreement with this observation , srpMRC is part of the same operon coding for the integrase that catalyzes both the integration and excision of SXT/R391 ICEs , all directly under control of the SetCD activator [14] . Therefore , srpMRC is expressed only prior to excision , replication and transfer of the ICE . We observed that a ΔhipA ΔtraG ΔsrpM R391K mutant has an extrachromosomal copy number similar to the wild-type . The apparent suppression of the effect of the ΔtraG mutation on the extrachromosomal copy number by the loss of srpM suggests a link between conjugation and partition that remains to be elucidated . Active partition of ICEs could be an overlooked feature that is in fact rather common among ICEs . The ICE PAPI-1 of Pseudomonas aeruginosa encodes the putative active partition system Soj . Deletion of soj leads to high-frequency loss of PAPI-1 [64] . Although the exact mechanism of action of Soj is not well understood , its expression was shown to be stimulated when PAPI-1 excises . ICEs of the pKLC102-ICEclc group , including PAPI-1 and ICEHin1056 , were shown to be able to replicate and code for putative partitioning systems [21 , 23 , 65–67] . Moreover , the core region of Tn4371-like ICEs and the ICE pNOB8 from Sulfolobus codes for ParA and ParB proteins , whose homologs are known to play a role in plasmid partition [68–72] . Finally , ICEA of Mycoplasma agalactiae encodes a ParA homolog that could be part of a partitioning system [73] . All these putative partitioning systems could also be involved in incompatibility with other ICEs and/or plasmids , as well as in transcriptional regulation of ICE- and/or host-borne loci [74–76] . Classification of mobile genetic element is extremely laborious mostly because of their modular structure . Our increasingly precise comprehension of their biology unravels some unexpected features that make them even harder to label [77] . On the one hand , ICEs exhibit phage-like behaviors , such as integration by site-specific recombination and , for some ICEs , regulation controlled by CI- or ImmR-like regulators [15 , 37 , 59 , 78 , 79] . On the other hand , ICEs also share several characteristics with plasmids , such as a single-strand DNA intermediate during transfer , their conjugative apparatus and entry exclusion systems ( traG/eex ) [80 , 81] . For instance , the conjugation modules and master activators SetCD and AcaCD of SXT/R391 ICEs and conjugative plasmids of the IncA/C group share a common ancestry [10 , 14 , 82] . As such SXT/R391 ICEs and IncA/C plasmids offer a dramatic example of divergent evolution from a common ancestor into two different lifestyles . Although SXT/R391 ICEs are capable of transient replication using the relaxase TraI and oriT , this lifestyle does not seem to be sustainable over multiple generations [14] . IncA/C plasmids lack the int and xis genes required for integration and excision , and instead carry a dedicated RepA/C replicon , allowing autonomous , stable and efficient replication in the cell . IncA/C plasmids code for a putative ParMRC-like partitioning system closely related to SrpMRC ( vcrx152/vcrx151 in pVCR94 ) ( Fig 6A and 6B ) . Interestingly , expression of parMRC-like locus of SXT/R391 ICEs and IncA/C plasmids is directly under the control of similar yet distantly related class II transcriptional activator complexes: SetCD for SXT/R391 ICEs and AcaCD for IncA/C plasmids [14 , 82 , 83] ( Fig 6B ) . Given the pleitropic role of these activators , this mode of regulation directly pairs the expression of DNA segregation functions to expression of conjugative transfer functions . However , although IncA/C plasmids retain a type II parMRC-like partitioning system ( actin-type ATPase ) , they also carry a type I parABC-like partitioning system ( Walker-type ATPase ) ( vcrx031/vcrx032 in pVCR94 ) , which does not seem to be regulated by AcaCD [48 , 82 , 84 , 85] . The exact function and eventual redundancy of each par locus remains to be investigated for IncA/C plasmids . The IncHI1 conjugative plasmid R27 also contains two independent partitioning loci , a type I partitioning system , and a type II partitioning system [54] . The type I partitioning system was shown to be the major stability determinant of R27 whereas type II is the minor stability determinant . Finally , our results put an end to a long standing question: Do SXT/R391 ICEs behave like plasmids and replicate ? R391 and related elements such as R705 , R748 , R997 , and pMERPH were initially reported as R factors belonging to the same J incompatibility group ( IncJ ) [9 , 86] . R391 and R997 were even isolated as circular molecules and physically mapped by restriction analysis [19] . Subsequent identification of SXT as an integrative element , and reports of the site-specific integration of R391 and R997 into the same chromosomal site as SXT highlighted seeming incongruities between otherwise extremely similar mobile genetic elements as revealed by sequence comparison [10 , 12 , 87–90] . In fact , our results indicate that replication , coupled with partition , is a normal yet transitory step of the lifecycle of SXT/R391 ICEs . The transitory nature of this replication does not allow stable maintenance and inheritance as a plasmid-like form . Therefore , integration into the chromosome remains the main mechanism ensuring stable vertical transmission of SXT/R391 ICEs over multiple generations . In the end , despite using similar strategies for their maintenance in the cell population and transfer between cell populations , ICEs and conjugative plasmids remain distinct entities regarding their respective maintenance by integration or replication .
The bacterial strains and plasmids used in this study are described in Table 1 . The strains were routinely grown in lysogeny broth ( LB-Miller , EMD ) at 37°C in an orbital shaker/incubator and were preserved at -80°C in LB broth containing 15% ( vol/vol ) glycerol . Antibiotics were used at the following concentrations: ampicillin ( Ap ) , 100 μg/ml; chloramphenicol ( Cm ) , 20 μg/ml; kanamycin ( Kn ) , 50 μg/ml; mitomycin C ( MC ) , 50 ng/ml; nalidixic acid ( Nx ) , 40 μg/ml; rifampicin ( Rf ) , 50 μg/ml; spectinomycin ( Sp ) , 50 μg/ml; sulfamethoxazole ( Su ) , 160 μg/ml; tetracycline ( Tc ) , 12 μg/ml; trimethoprim ( Tm ) , 32 μg/ml . When required , bacterial cultures were supplemented with 0 . 02 mM of isopropyl β-D-1-thiogalactopyranoside ( IPTG ) or 0 . 02% L-arabinose . Conjugation assays were performed by mixing equal volumes of each donor and recipient strains that were grown overnight at 37°C . The cells were harvested by centrifugation for 3 min at 1200g , washed in 1 volume of LB broth and resuspended in 1/20 volume of LB broth . Mating mixtures were then deposited on LB agar plates and incubated at 37°C for 6 hours . The cells were recovered from the plates in 1 ml of LB broth and serially diluted before plating . Donors , recipients and exconjugants were selected on LB agar plates containing appropriate antibiotics . Plasmid DNA was prepared using the EZ-10 Spin Column Plasmid DNA Minipreps Kit ( Biobasic ) according to manufacturer’s instructions . All the enzymes used in this study were purchased from New England BioLabs . PCR assays were performed with the primers described in S1 Table . The PCR conditions were as follows: ( i ) 3 min at 94°C; ( ii ) 30 cycles of 30 sec at 94°C , 30 sec at the appropriate annealing temperature , and 1 minute/kb at 68°C; and ( iii ) 5 min at 68°C . When necessary , PCR products were purified using an EZ-10 Spin Column PCR Products Purification Kit ( Biobasic ) according to manufacturer’s instructions . E . coli was transformed by electroporation as described by Dower et al . [94] in a BioRad GenePulser Xcell apparatus set at 25 μF , 200 V and 1 . 8 kV using 1-mm gap electroporation cuvettes . Sequencing reactions were performed by the Plateforme de Séquençage et de Génotypage du Centre de Recherche du CHUL ( Québec , QC , Canada ) . Plasmids and oligonucleotides used in this study are listed in Table 1 and S1 Table . pTraI and pTraG were constructed by cloning traI of SXT and traG of R391 into the TA cloning expression vector pBAD-TOPO ( Invitrogen ) according to the manufacturer’s instructions . traI was amplified by PCR with its native Shine-Dalgarno sequence using primers pBad-traI_Fw and pBad-traI_Rev and genomic DNA of E . coli HW220 as the template . traG was amplified using the primer pair traGEcoRI . for / traGEcoRI . for and genomic DNA of E . coli GG13 as the template . pVB15 was constructed by amplifying the origin of replication of pUC19 ( oriVpMB1 ) using the primer pair pUC_oriF/pUC_oriR and subsequent cloning into the 5 838-bp fragment of NheI/NotI-digested pAH56 to replace oriVR6K-attPλ and generate the high-copy number expression vector pVB15 . ps002-his was then obtained by cloning s002 ( srpR ) from SXT amplified with the primer pair s002F/s002-hisR into the 4 319-bp fragment of NdeI/BamHI-digested pVB15 . Plasmids used for plasmid stabilization assays were derived from pBeloBAC11Δsop , a pBeloBAC11 vector derivative from which the partitioning system sopABC was deleted by NdeI digestion and re-ligation . The srpMRC locus of SXT ( srpMRCSXT ) and R391 ( srpMRCR391 ) were amplified by PCR using genomic DNA of strains containing either SXT or R391 as the templates and primers pairs SXTpartHindIIIstop . for/SXTR391partHindIII . rev and R391partHindIIIstop . for/SXTR391partHindIII . rev , respectively . Amplicons were digested by HindIII and cloned into HindIII-digested pBeloBAC11Δsop to generate pSrpSXT and pSrpR391 . Subsequent deletions of segments of srpMRCR391 were obtained by high fidelity PCR amplification of the pSrpR391 vector using primer pairs pBeloDelSO02 . for/pBeloDelSO02 . rev , pBeloDelSO03 . for/pBeloDelSO03 . rev or pBeloDelparC . for/pBeloDelparC . rev , digestion by NheI and ligation using the T4 DNA ligase . The resulting plasmids were verified by restriction profiling and DNA sequencing . Deletion mutants of R391::galK ( R391K ) [18] were constructed using the one-step chromosomal gene inactivation [92] and P1vir transduction [95] techniques . Deletion of hipA , srpR , srpM , srpRM , traI and traG were constructed using primer pairs R391DhipAnoFRT . for/R391DhipAnoFRT . rev , 2SXTR391DSO02 . for/2SXTR391DSO02 . rev , R391DSO03 . for/2SXTR391DSO03 . rev , R391DSO03 . for/2SXTR391DSO02 . rev , R391DtraInoFRT . for/R391DtraInoFRT . rev , R391DtraGnoFRT . for/R391DtraGnoFRT . rev , respectively . Gene resistance cassettes were amplified using the pVI36 , pKD3 and pKD4 vectors . The λRed recombination system was expressed using pSIM5 or pSIM6 as described by Datta et al . [91] . If possible , the antibiotic resistance cassette was removed from the resulting construction by Flp-catalyzed excision using the pCP20 vector [96] . All deletions were verified by PCR and antibiotic resistance profiling . The stability of R391::galK and derivative mutants was monitored based on the methodology described by Wozniak and Waldor [32] . Cells were grown for 16 hours in 4 ml of LB medium supplemented or not with kanamycin . Serial dilutions were plated on MacConkey agar plates supplemented with 1% D-galactose . Loss of R391 resulted in the formation of white clones ( Fig 1C ) . For each experiment , at least 16 white clones were purified and tested on agar plate for their susceptibility to kanamycin . These clones were also tested by PCR amplification of an internal fragment of R391 using primer pair R391HipBM1 . for/R391HipB . rev . The stability of pBeloBAC11Δsop and derivatives containing the srpMRC locus of SXT or R391 was tested for 16 hours in M9 or LB medium using the approach described by Sanchez et al . [97] . Expression of the srp locus from Plac was induced by addition of 0 . 02 mM IPTG . Relative stability was calculated as the ratio of chloramphenicol resistant colonies in the population in the induced compared to the non-induced conditions . For both ICE and plasmid stability assays , each experiment was carried out at least in biological triplicate . The frequency of excision as well as total copy number in the population and copy number of the excised circular form of the ICE were assessed by real-time quantitative PCR as described elsewhere [20 , 26] . Genomic DNA was obtained from cell cultures of E . coli CAG18439 bearing SXT , R391K or its mutants grown for 16 h in LB medium . prfC , attB , attP and int were quantified using primer pairs prfC . qec . F1/prfC . qec . R1 , attB . qec . F2/attB . qec . R2 , attP . qec . F2/attP . qec . R2 and int . qec . F1/int . qec . R1 , respectively ( S1 Table ) . For frequency of excision and copy number determination , E . coli VI61 , which contains one chromosomal copy of attB , attP and prfC , was used to simulate 100% of excision and normalize the results . qPCR experiments were performed in triplicate on the RNomics platform of the Laboratoire de Geénomique Fonctionnelle de l’Universiteé de Sherbrooke ( http://lgfus . ca ) ( Sherbrooke , QC , Canada ) . Macroscopic observations were done using a SZX7 zoom stereomicroscope with a DF PLAPO1X-4 objective coupled to a SC30 digital camera via a U-TV1X-2 & U-CMAD3 adaptor ( Olympus ) . To express and purify SrpR tagged with a 6×His C-terminal epitope ( SrpR6×His ) , cultures of E . coli BL21 bearing ps002-his were grown overnight , diluted 1:500 in fresh 2×YTA broth and incubated at 37°C with agitation . At mid-exponential phase ( OD600 of 0 . 6 ) , protein expression was induced with 0 . 1 mM IPTG and cultures were incubated for 3 hours . Cells were then harvested by centrifugation at 1500×g for 10 min at 4°C and stored at -20°C . The cell pellet was weighted and re-suspended in Native Purification Buffer ( NPB ) ( 50 mM NaH2PO4 pH 8 . 0 , 2 . 5 M NaCl ) containing 0 . 1% Triton X-100 , 1 mM phenylmethanesulfonylfluoride ( PMSF ) , and protease inhibitors at 1 ml / 20g of cell pellet ( Protease Inhibitor Cocktail , Sigma ) . Purification of SrpR6×His was done by Ni-NTA affinity chromatography following the manufacturer’s instructions ( Qiagen ) . Cells were lysed by sonication , cell debris was pelleted by centrifugation , and the supernatant was incubated for 1 h at 4°C with 750 μl of Ni-NTA Agarose resin ( QIAGEN ) with agitation . The Ni-NTA Agarose resin was then transferred into a column and washed 4 times with 1 . 25 ml of native wash buffer ( NPB with 20 mM imidazole , pH 8 . 0 ) . SrpR6×His was eluted with native elution buffer ( NPB with 250 mM imidazole , pH 8 . 0 ) and stored at -20°C . Protein concentration was estimated using a Bradford protein assay ( BioRad ) and purity was determined by SDS-PAGE analysis . The linear double-stranded DNA probes srpC ( 251 bp ) , oriT ( 298 bp ) and ig ( srpM-mobI ) ( 615 bp ) used in the EMSA experiments were amplified by PCR using primer pairs RRintF/RRintR , oriT2F/oriT2R and MELR1/oriT1R , respectively , and E . coli HW220 as the template ( Table 1 and S1 Table ) . Probes were purified using an EZ-10 Spin Column PCR Products Purification Kit ( Bio Basic ) according to the manufacturer's instructions and their concentration was determined using a NanoDrop ND-1000 . Probes R3-R1 , R4-R2 , R2 and R4 were obtained by mixing equimolar concentrations ( 50 μM ) of primers , srpCSXTR3R1F and srpCSXTR3R1R , srpCSXTR4R2F and srpCSXTR4R2R , srpCSXTR2F and srpCSXTR2R , or srpCSXTR4F and srpCSXTR4R , respectively . The primer mixtures were heated at 95°C for 3 min , then annealed by slow cool down overnight . EMSA assays were carried out using the Electrophoretic Mobility-Shift Assay Kit with SYBR Green & SYPRO Ruby EMSA stains ( Life Technologies ) according to the manufacturer's instructions . Briefly , a total of 40 ng of DNA probe was used in each reaction . Quantities of SrpR6×His and of the non-specific competitor DNA ( sonicated salmon sperm DNA ) varied from 10 to 100 ng , and 10 to 200 pg , respectively . The non-specific competitor DNA was mixed with the probe before adding SrpR6×His to maximize competition . All binding reactions were done in a total volume of 10 μl for 15 min at room temperature followed by 10 min incubation on ice . Samples were then loaded on a pre-run ( 25 min at 100 V ) non-denaturing 4% acrylamide gel containing 1× TBE buffer and migration was carried out at 4°C during electrophoresis . SYBR Green staining was done according to the manufacturer's instructions and gel pictures were scanned using a Typhoon FLA 9500 ( GE Healthcare Life Sciences ) with a LPB filter for SYBR Green I at a 100 μm resolution . Dimerization assays were carried out using 2 μg of purified SrpR6×His . Samples containing srpC were carried out using 1 μg of DNA probe and were incubated prior to the dimerization assay in the same conditions as for the EMSA assays . Samples were incubated with or without 0 . 6% glutaraldehyde for 30 min at room temperature and 3% of β-mercaptoethanol was added to the samples prior to denaturation at 95°C for 3 min . Samples and ladder ( Precision Plus Protein Kaleidoscope Standards , BioRad ) were separated by electrophoresis on a 12% SDS-PAGE gel , later stained using Coomassie Brilliant Blue R-250 . The molecular phylogenetic analysis of SrpM was conducted in MEGA6 [98] The primary sequence of SrpM encoded by R391 was used to search for homologous sequences in the Genbank Non-redundant protein sequence ( nr ) database using blastP [99] . Phylogenetic analyses were computed using a protein alignment generated by MUSCLE [100] and poorly aligned regions were removed with the trimAl v1 . 3 software using the automated heuristic approach [101] prior to phylogenetic analyses . The evolutionary history was inferred by using the Maximum Likelihood method . Initial tree ( s ) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model , and then selecting the topology with superior log likelihood value . A discrete Gamma distribution was used to model evolutionary rate differences among sites ( 5 categories ( +G , parameter = 4 . 1321 ) ) . The rate variation model allowed for some sites to be evolutionarily invariable ( [+I] , 2 . 6007% sites ) . The tree is drawn to scale in iTOL v2 [102] , with branch lengths measured by the number of substitutions per site . The analysis involved 60 amino acid sequences with a total of 261 positions in the final dataset .
|
Integrative and conjugative elements ( ICEs ) constitute a class of mobile genetic elements defined by their ability to integrate into the chromosome of their host cell and to transfer by conjugation . Some of the most studied ICEs belong to the SXT/R391 family , which are major drivers of multidrug resistance dissemination among various pathogenic Gammaproteobacteria . Transfer of SXT/R391 ICEs to a new host first requires its excision from the chromosome as a circular molecule , which may be lost if the cell divides . In silico analyses revealed several putative stabilization systems carried by R391 , a prototypical member of the SXT/R391 ICEs family originally isolated from Providencia rettgeri . We discovered that , besides stabilization by integration into the chromosome , stability of SXT/R391 ICEs also depends on toxin/antitoxin systems and plasmid-like features including intracellular replication and active partition . Thus , although it has been known for a long time that ICEs and conjugative plasmids use similar strategies to transfer between bacterial populations , our work reveals additional unforeseen similarities in their mechanisms of maintenance in the host cell .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Replication and Active Partition of Integrative and Conjugative Elements (ICEs) of the SXT/R391 Family: The Line between ICEs and Conjugative Plasmids Is Getting Thinner
|
Certain cutaneous human papillomaviruses ( HPVs ) , which are ubiquitous and acquired early during childhood , can cause a variety of skin tumors and are likely involved in the development of non-melanoma skin cancer , especially in immunosuppressed patients . Hence , the burden of these clinical manifestations demands for a prophylactic approach . To evaluate whether protective efficacy of a vaccine is potentially translatable to patients , we used the rodent Mastomys coucha that is naturally infected with Mastomys natalensis papillomavirus ( MnPV ) . This skin type papillomavirus induces not only benign skin tumours , such as papillomas and keratoacanthomas , but also squamous cell carcinomas , thereby allowing a straightforward read-out for successful vaccination in a small immunocompetent laboratory animal . Here , we examined the efficacy of a virus-like particle ( VLP ) -based vaccine on either previously or newly established infections . VLPs raise a strong and long-lasting neutralizing antibody response that confers protection even under systemic long-term cyclosporine A treatment . Remarkably , the vaccine completely prevents the appearance of benign as well as malignant skin tumors . Protection involves the maintenance of a low viral load in the skin by an antibody-dependent prevention of virus spread . Our results provide first evidence that VLPs elicit an effective immune response in the skin under immunocompetent and immunosuppressed conditions in an outbred animal model , irrespective of the infection status at the time of vaccination . These findings provide the basis for the clinical development of potent vaccination strategies against cutaneous HPV infections and HPV-induced tumors , especially in patients awaiting organ transplantation .
Papillomaviruses ( PVs ) infect mucosal and cutaneous squamous epithelia , where they can cause hyperproliferative lesions . In the case of high-risk genital human papillomavirus ( HPV ) types , a causative link has been established between HPV infection and the development of malignant diseases , especially cervical carcinoma [1] . For cutaneous types , the association between HPV infection and skin cancer is still a matter of debate [2] , although there is increasing evidence that supports their role as a cofactor with UV radiation in the development of non-melanoma skin cancer ( NMSC ) [3] . Indeed , it has been shown that certain cutaneous HPVs display transforming properties and tumorigenic features , both in vitro and in vivo [4]–[7] . Furthermore , epidemiological data that support an association of HPV infection and NMSC have been found for two distinct populations: patients with the rare hereditary disease Epidermodysplasia verruciformis ( EV ) and immunosuppressed organ transplant recipients ( OTR ) . Compared to the general population , the incidence of NMSC is up to 250-fold higher in OTR [8] , [9] . Additionally , OTR suffer from benign and premalignant skin lesions , such as actinic keratosis , keratoacanthomas and cutaneous warts , which are indisputably caused by cutaneous HPVs [10] , [11] . Such lesions appear over a large area of the skin , persist for years and significantly reduce life quality . Hence , the high incidence of PV-induced warts and premalignant lesions in immunosuppressed OTR represents a great burden , which demands new prophylactic strategies to prevent such skin manifestations . We suggest that peri-transplant immunization with vaccine against cutaneous HPV types could reduce the incidence of virus-induced skin lesions that can progress to NMSC . Vaccination against genital HPV types is currently being used worldwide to prevent infection and , in turn , the development of PV-induced lesions in the mucosa , including cervical carcinoma . The two licensed vaccines are composed of HPV virus-like particles ( VLPs ) , which elicit high titers of neutralizing antibodies that protect from a subsequent infection by the targeted HPV types [12] . Both vaccines are very effective when applied in individuals with no previous exposure , but efficacy decreases when analyzed in patients with positive HPV serology [13] . A unique preclinical model to investigate PV-associated skin tumorigenesis is the African multimammate mouse Mastomys coucha , originally taxonomically designated as Mastomys natalensis [14] . This species belongs to the rodent family Muridae , as the laboratory mouse Mus musculus . The colony maintained at the German Cancer Research Center ( DKFZ ) is naturally and persistently infected with Mastomys natalensis papillomavirus ( MnPV ) and Mastomys coucha papillomavirus 2 ( McPV2 ) , which - like cutaneous and genital HPVs - infect epidermal and mucosal tissues , respectively [15] , [16] . Throughout their lifetime these animals spontaneously develop epithelial lesions of the skin ( mainly papillomas and keratoacanthomas ) as well as papillomas at the tongue and condylomata at the anus , vulva and penis . Both papillomaviruses persist episomally without any indication of integration , analogous to cutaneous HPVs [16] , [17] . Naturally MnPV-induced lesions rarely regress and can efficiently form squamous cell carcinomas ( SCC ) after a single topical application of a carcinogen followed by repeated challenge with a tumor promoter [18] . Additionally , the potential oncogenic capacity of MnPV could be also demonstrated in transgenic mice carrying the E6 oncoprotein , in which viral expression was targeted to the basal layer of the skin [19] . In contrast to the cottontail rabbit papillomavirus [20] , MnPV shares the trademark of cutaneous HPVs that lack a functional E5 open reading frame ( ORF ) [21] . This fact , together with the alignment of cis-responsive elements in its regulatory region and the phylogenetic assessment of parts of the E6 , E1 , and L1 ORFs , indicates that MnPV is related to HPV types found in lesions of cutaneous epithelia , in particular to those associated with EV [22] . Considering the natural infection mode , Mastomys acquire the virus very early after birth , as MnPV-DNA is found in the skin of four-week-old animals at the same time as seropositivity against viral antigens [23] . In fact , the infection status within our animal colony mimics the situation of some cutaneous HPVs , which are also acquired early during childhood [24] , [25] . As previously shown in a large serological study , a strong correlation between MnPV L1-specific antibodies and benign skin tumor formation is discernible . Notably , among the early expressed viral proteins , strong antibody responses against MnPV E2 can be measured at an age of one month . A prospective study additionally revealed that E2 seropositivity marks early and late stages of infection . Together with high anti-L1 reactivities at an age of 4 . 5 months , subsequent tumor development could be predicted [23] . In contrast to other animal models [26] , [27] , all progression stages can be monitored in an immunocompetent laboratory animal , starting from primary infection in newborns until lesion development in a natural host [28] . Since inbred strains may vary substantially from outbred counterparts in their immune response [29] , the outbred character of these animals does not only mimic the genetic heterogeneity in humans , but also allows to determine whether the protective effect of a VLP vaccine is representative and transferable to patients . Hence , the success of a VLP-based vaccination can be readily monitored by the absence of lesions as the ultimate read-out , thereby circumventing the necessity of an indirect challenge with luciferase-encoding pseudoviruses to detect neutralizing antibodies by in vivo imaging [30] . The preclinical animal model described here allows the investigation of virus-host interactions under defined in vivo conditions . Particularly , this study addresses important aspects which differentiate cutaneous from genital infections and might affect the efficacy of a vaccine against skin PV types . Firstly , will immunity conferred by VLP vaccination be effective against infections of the skin in a natural host ? Secondly , what will be the outcome of vaccination when applied to animals with previously infected skin ? In the case of a vaccine against cutaneous HPV types , this issue would be of the outmost importance because infections are acquired early in lifetime [24] . Thirdly , can a vaccine also protect against malignant skin tumors in its natural host ? And finally , what is the influence of immunosuppression on the long-term efficacy of the vaccine ? Answering these questions will be fundamental for the clinical development of a cutaneous HPV vaccine , given that the main targeted group of individuals would be OTR before undergoing systemic immunosuppression . In principle , the proof-of-concept of effective vaccination against a single skin PV type under the aforementioned conditions would represent the first step towards establishing a successful vaccination strategy in humans . In this regard , Mastomys provides an optimal model as the animals are latently infected with MnPV , exhibiting a similar tropism and pathogenicity as skin-associated HPV .
To investigate the effect of MnPV VLP vaccination under different conditions , we used for immunization young animals derived from naturally infected Mastomys and also a virus-free colony that can be infected under defined experimental conditions . Based on the fact that MnPV is not transmitted in utero [28] , virus-free Mastomys were obtained by hysterectomy [31] . Newborns were fostered by specific-pathogen-free ( SPF ) mice . The siblings were mated to establish a colony and their offspring analyzed at different ages for MnPV DNA in tissue biopsies , as well as for MnPV E2- and L1-specific seroconversion . Virus screening was performed by an established PCR protocol [23] . Samples were taken from potential infection sites [16] , [28] , namely the furred skin from the back , the eyelid , anogenital tissue and the tongue , which were all found to be negative . During the 3 years of existence of the virus-free colony , no MnPV-induced tumors were observed . After vaccination , subgroups from both cohorts were also immunosuppressed ( Fig . 1A , see also Table 1 for details ) . Since our study involved a large cohort of animals which were to be followed up for more than one year , it was important to first define the optimal vaccination parameters ( i . e , time , dosage , use of adjuvant , number of boosts ) . Groups of five 4-week-old Mastomys from the naturally infected colony were immunized subcutaneously with different vaccine formulations . The primary read-out for this pilot study was the generation of humoral immune responses against the MnPV major capsid protein . The induction of anti-L1 antibodies was measured by a newly set-up VLP-ELISA . All animals vaccinated with VLPs produced in baculovirus-infected insect cells ( Fig . 1B ) had a response above the calculated cut-off ( Fig . S1 ) . Since VLPs are highly immunogenic structures [32] , the use of low doses of antigen was sufficient to elicit a response . Here , the increase from 5 to 25 µg VLP per dose ( Fig . S1 , groups 1 and 5 ) did not have a substantial impact on the anti-L1 response . Conversely , the effect of adjuvant addition in the formulation was significant , as the use of Sigma Adjuvant System ( SAS ) proved to be very effective ( p<0 . 05 vs . no adjuvant; p<0 . 05 vs . aluminum hydroxide ) . Furthermore , two boosts after a first inoculation of VLPs+SAS proved to be optimal for inducing high antibody titers . Therefore , in all following experiments the vaccine was administered in a first dose of 10 µg VLPs formulated in PBS and SAS followed by two booster doses of 10 µg VLPs in PBS without adjuvant . To investigate the effectiveness of the vaccine in our model , the study was divided in two main branches ( Fig . 1A and Table 1 ) : VLPs were administered to Mastomys from the naturally infected colony as well as to virus-free animals which were subsequently infected . Seroconversion of the naturally infected animals at the time of vaccination was assessed by the appearance of antibodies directed against the MnPV E2 protein ( Fig . S2 ) , known to be the earliest marker of infection [23] . Two weeks after completion of the vaccination schedule , the sera showed high anti-L1 titers by VLP-ELISA , whereby all animals responded to the vaccine both in the naturally infected and in the virus-free colony ( geometric mean titer virus-bearing colony: 3 . 1×105; geometric mean titer virus-free colony: 3 . 5×105; range: 2 . 4×104–1 . 9×106 ) ( Fig . 2A ) . To test for the elicitation of neutralizing antibodies , we developed an in vitro neutralization assay which tests the ability of raised antibodies to prevent infection by MnPV pseudovirions that carry a reporter gene [33] . The system was first calibrated with two HPV L2-specific monoclonal antibodies ( K4L2 and K18L2 ) which cross-neutralize several papillomavirus types [34] . Both antibodies displayed a high neutralization titer , showing proper sensitivity in our experimental setup ( Fig . S3 ) . In a next step , we monitored the ability of sera from animals of all groups to prevent infection in the MnPV pseudovirion assay . As depicted in Fig . 2B , the degree of neutralizing activity of the sera displayed a high correlation with the anti-L1 titers as assayed by VLP-ELISA ( r2 = 0 . 8919 , n = 113 ) , regardless of the age of the animals as well as their infection and immune status ( Fig . S4 ) . We next analyzed the time course of L1-induced antibodies in the naturally and experimentally infected colonies both under normal and immunosuppressed conditions . Substantial responses could be discerned at around 5 months in non-vaccinated controls , independently of whether the sera were derived from naturally or experimentally infected immunocompetent animals ( Fig . 3A and C , pink bars ) . The natural immune response reached vaccination-like titers at an age of 5–7 months ( Table S1 ) . Considering the vaccinated cohorts ( Fig . 3B and D , light blue bars ) , antibodies reached a geometric mean titer of 3 . 1×105 and 3 . 5×105 in both colonies after the third immunization , declining to 3 . 6×104 and 1 . 9×104 after 14 months , respectively . To exclude that the observed seroresponses in both groups were due to continuous boosting of the immune system caused by viral infection , thereby overlapping the profiles depicted in Fig . 3A and C , we also vaccinated a group of animals which remained virus-free during the whole study ( Fig . 3D , grey bars ) . Anti-L1 antibody titers were stable over the same period , with a geometric mean titer of 8 . 5×105 immediately after vaccination that dropped to 7 . 2×104 after 14 months . The specific antiviral humoral immune response was also assessed in animals undergoing systemic immunosuppression . To mimic this situation , cyclosporine A ( CsA ) was incorporated into food pellets as previously described [35] , [36] . With this approach , immunosuppressive drugs can be given for an indefinite period of time without the animals suffering from the stress of daily intraperitoneal injections or gavage feeding . CsA concentration was measured in the heparinized blood both during the initial phase ( 574±132 ng/ml ) and the maintenance phase ( 235±81 ng/ml ) and was in accordance to the already reported immunosuppressive range for mice and humans [37] , [38] . In order to assure that biologically active CsA concentrations were present , we additionally monitored Mastomys-specific interferon-γ ( IFN-γ ) expression in spleen cells of individual animals either under immunocompetent or immunosuppressed conditions by quantitative real-time PCR . IFN-γ expression was down-regulated by 70% in CsA-treated animals , indicating biologically active levels ( Fig . S5 ) . Antibody titers after vaccination against L1 in the naturally infected or the virus-free colony , measured before immunosuppression ( at 3 . 2 months of age , see Fig . 1C ) , reached similar levels ( geometric means of 5 . 5×105 and 6 . 2×105 , respectively ) as in the immunocompetent groups at the same age ( Fig . 3B and D , compare dark and light blue bars ) . Likewise , the drop of approximately one log in antibody titer ( means of 1 . 6×104 and 4 . 3×104 , respectively ) after 14 months of immunosuppression showed that this treatment did not significantly affect the antibody response . Analyzing the course of L1-induced antibodies in the naturally and experimentally infected colonies after immunosuppression , antibody responses started at around 3–5 months and also increased in a time-dependent manner , as already revealed for immunocompetent animals ( Fig . 3A and C , red bars ) . As the ultimate read-out for the vaccination efficacy , we evaluated the ability of the VLPs to prevent the appearance of skin tumors . Animals were examined monthly starting at the age of 7 months . After 13 months of observation ( animal age: 20 months ) , none of the vaccinated animals had developed skin tumors in any of the vaccination groups . Conversely , the percentage of tumor-bearing animals at the end of the observation period was 28 . 0% for the naturally infected animals ( p<0 . 0001 vs . vaccinated ) and 17 . 5% for the experimentally infected colony ( p<0 . 01 vs . vaccinated ) ( Fig . 4A and B ) . Thus , vaccination with VLPs can effectively prevent skin tumor formation , even when already infected animals are vaccinated . In the case of the virus-free Mastomys , which had been experimentally infected with a MnPV-containing wart extract in the lower back , the observed skin tumors appeared only in this area ( Fig . 4E and F ) . In contrast , as described previously [23] , lesions could be detected over the entire body in the naturally infected colony ( Fig . 4G and H ) . Notably , tumors appeared at around 11 months in both groups of animals , independently of the mode of infection ( Fig . 4A and B ) . Also in immunosuppressed animals , a complete protection from tumor development was observed after VLP vaccination ( Fig . 4C and D ) , whereas 29 . 5% of the experimentally infected Mastomys ( p<0 . 05 vs . vaccinated ) and 14 . 9% of the naturally-infected ( ns vs . vaccinated; p = 0 . 0788 ) had tumors by the age of 20 months . This indicates that CsA treatment after vaccination does not affect the protection conferred by VLP vaccination , which is consistent with the presence of neutralizing antibodies even under immunosuppressive conditions ( see Fig . 2B and Fig . 3 ) . In a next step , we examined the histological features of the skin tumors obtained in non-vaccinated animals after different modes of infection . As previously reported , most lesions in Mastomys are characterized as papillomas and keratoacanthomas , but also squamous cell carcinomas can be discerned [39] , [40] . Indeed , the lesions found in our animals included not only benign , but also malignant tumors that can be classified as epidermal carcinomas . As depicted in Fig . 5A , keratoacanthomas show a thickened epidermis with principally maintained epidermal layers and a high proliferative index , as monitored by Ki-67 staining ( Fig . 5C ) . In general the lesion does not infiltrate , as demonstrated by an intact basement membrane uniformly stained by antibodies directed against laminin ( Fig . 5G ) . Also the keratin stain shows compact lobules of epidermal cells respecting the boundary to stroma ( Fig . 5E ) . In contrast , an epidermal carcinoma obtained by experimental infection depicts , similarly to naturally induced malignant tumors , highly proliferating cells distributed over the whole tissue section ( Fig . 5D ) . The basement membrane is no longer intact and keratin and Ki-67 positive singular cells can be visualized in the inflamed stroma ( Fig . 5F and H ) . An increased nucleocytoplasmic ratio , a moderate nuclear pleomorphism and nuclear polycromasia can be seen , indicative of malignancy of the lesion . These data show that vaccination not only protects against benign but also from malignant skin tumors . As reported previously , MnPV DNA can be found in the skin of virtually all animals older than 4 weeks in the naturally infected colony , suggesting an early transmission and long persistence [28] . Moreover , monitoring healthy skin or papillomas in older animals , the viral load increases in more differentiated regions ( such as the stratum spinosum and stratum granulosum ) to levels that could be detected by in situ hybridization ( Fig . 6A and B ) . To test whether the viral load of the skin can be used as a further surrogate marker for vaccination efficacy , quantitative PCR of skin samples from tumor-free animals reaching the study end-point was performed . As presented in Fig . 6C , vaccinated animals had a lower viral load than unvaccinated controls , suggesting that vaccination with VLPs at a young age effectively prevents the increase in viral load , which in turn predicts which animals are prone to subsequent tumor formation [23] . Since vaccinated animals displayed both a low MnPV load in the skin ( Fig . 6C ) and a complete protection from skin tumor formation ( Fig . 4 ) , we tested whether elicited antibodies were also protective under in vivo conditions . For this purpose , sera from vaccinated animals revealing a high antibody titer in the in vitro neutralization assay were pooled and administered intraperitoneally to virus-free Mastomys 24 hours prior to experimental infection . One week after passive immunization , RNA was extracted from infected skin areas and analyzed by RT-PCR for the appearance of the prevalent E1∧E4 transcript found in a recently performed Mastomys transcriptome analysis from a keratoacanthoma ( Vinzón et al . , manuscript in preparation ) . As shown in Fig . 7 , transfer of anti-VLP immune sera apparently prevented the virus from infecting the skin , as evidenced by the lack of the corresponding MnPV transcript at the site of infection . This result indicates that immune serum obtained from VLP vaccinated animals can protect from experimental infection by transmission of neutralizing antibodies .
The currently available HPV vaccines are a major breakthrough in the fight against HPV-associated genital cancers and proved the potential of prophylactic vaccines to prevent HPV-caused disease . These vaccines were developed on the basis of preclinical data obtained from animal models of PV infection [41] . Although valuable , these systems ( namely the bovine papillomavirus , the canine oral papillomavirus or the cottontail rabbit papillomavirus models ) present inherent difficulties that make their use in the laboratory technically challenging . Hence a small rodent model of papillomavirus infection would be beneficial for the testing of prophylactic vaccines as well as for the development of therapeutic approaches . In this regard , the multimammate rodent Mastomys coucha represents a unique outbred model , being naturally infected with both cutaneous and mucosal PV types which are etiologically linked to the formation of tumors [16] . We report here that a VLP-based vaccine against a cutaneous papillomavirus can effectively prevent the appearance of naturally and experimentally induced skin tumors , both under immunocompetent and immunosuppressive conditions , thus providing the basis for the implementation of vaccination strategies against cutaneous HPVs in OTR . For the vaccine formulation , VLPs were chosen on the basis of their previous success in eliciting a protective immune response against other PVs in animal models and in humans . VLPs have the advantage of generally inducing high titers of neutralizing antibodies , which represent the major effector mechanism of most preventive viral vaccines , particularly in those against genital high-risk HPV types [42] . We could show a strong correlation between the total IgG levels against the L1 capsid protein as measured by VLP-ELISA and the amount of antibodies which are able to effectively neutralize the virus ( Fig . 2B ) . Therefore , titers assessed by VLP-ELISA are reliable markers of a protective immune response . Using both serological methods [23] and direct detection of viral DNA [28] , previous studies have shown that natural infection with MnPV can occur earlier than four weeks after birth . This could hamper the evaluation of vaccination efficacy in preventing primary infection , since the actual time point of infection cannot be predicted for single cases . Therefore , a virus-free Mastomys population was established in order to assess the effect of vaccination on naïve animals . On the other hand , a particular advantage of the multimammate mouse is that the vaccine can be studied in a naturally infected colony , thus allowing the monitoring of immunization success in already virus-positive animals . This is of particular importance in the context of a human infection , in which some cutaneous HPVs are acquired early in childhood [24] . The currently licensed vaccines targeting HPV have not shown any evidence of accelerated clearance in patients who were DNA-positive at baseline for the type considered in the evaluation [13] . In our preclinical study , tumor formation was completely prevented even in previously infected animals , although VLP vaccination was not sufficient to completely clear PV infection in the skin ( Fig . 6C ) . Hence , administration of VLPs may have a “therapeutic” effect by preventing virus spread during early infection and therefore avert tumor development . Cutaneous immune surveillance is typically initiated by professional antigen-presenting cells which encounter the antigen in the skin and activate naïve T cells that finally orchestrate a proper adaptive immune response [43] . Although natural and experimentally infected control animals also raise a humoral response ( Fig . 3A and C ) resulting in the induction of neutralizing antibodies ( Fig . 2B ) , the appearance of tumors cannot be prevented ( Fig . 4A and B ) . Here , the immune system is apparently not able to handle early bursts of replication and high viral load accumulation that finally lead to a skin lesion . In fact , comparing the viral load in normal skin of vaccinated and control animals ( Fig . 6C ) revealed that the latter generally harbor more copies of the virus . This can also be visualized by in situ DNA hybridization where the amount of MnPV focally increases in cells of the strata spinosum and granulosum ( Fig . 6A ) . Conversely , considering a papilloma , a more even distribution of cells with highly amplified MnPV DNA could be discerned ( Fig . 6B ) . Consistent with earlier studies , viral persistence and high viral load correlated with the development of skin papillomas and keratoacanthomas [28] . Hence , high copy number may be a general feature of productive infections [44] . On the other hand , high viral load is also accompanied by enhanced oncogene expression that could finally predispose cells towards malignancy [45] . The novel finding that vaccination minimizes the viral load by at least 10–20-fold ( Fig . 6C ) clearly suggests that prevention of tumor development is due to the diminished efficiency of viral spread after an early infection . As it was already proposed for genital types , this effect is most likely conferred by neutralizing antibodies that in turn block de novo initiated productive viral cycles by preventing reinfection of cells in traumatized epithelium [46] . In fact , as demonstrated by passive transfer of sera from vaccinated to naïve animals , infection of the skin can be prevented ( Fig . 7 ) , showing that neutralizing antibodies directed against MnPV are sufficient for protection . The way by which neutralizing antibodies can reach the basal epidermal layer in order to control infection is still unclear . For IgG antibodies reaching the genital mucosa , two transport mechanisms have been described: transudation mediated by the neonatal Fc receptor [47] and exudation of systemic antibodies at the site of infection . In the case of the skin , transudation seems unlikely and an exudation mechanism would be favored . This is in agreement with the fact that , to enter basal cells , papillomaviruses first have to bind to the basement epithelial membrane at sites where the epithelium is traumatized [28] , [48] . Notably , it has been proposed that exudation might be responsible for the observed protection in the genital model , as clinical trials reported excellent protection from genital warts , which occur on cornified skin where protection by transuded antibodies is also unlikely to occur [12] . On the other hand , although it was previously thought that B cells and IgGs cannot enter the skin , there is accumulating evidence that exceptions exist , especially during chronic inflammatory processes or B cell malignancies . Firstly , although poorly understood , the skin can be infiltrated by a subset of B cells that connect innate with adaptive immunity [49] . Secondly , high-resolution two-dimensional immunoblotting assays revealed the presence of IgGs in the upper lesional epidermis of psoriasis plaques [50] , also known to be positive for cutaneous HPV types [51] . It remains to be determined how the exact spatial-temporal mechanism ( s ) that finally protects against a papillomavirus infection in the skin operates . Another important issue to address during the development of a vaccine against cutaneous PVs is whether protection can be still conferred to immunosuppressed patients . This matter is of paramount importance , when considering such a vaccine for OTR who undergo chronic systemic immunosuppression . Vaccination before transplantation is recommended for many other preventable diseases and has been reported to be effective , with a lower rate of success when patients are immunized after immunosuppression [52] , [53] . Chronic immunosuppression with CsA did not significantly affect the antibody response ( Fig . 3 ) , as it was previously reported for CsA-treated rabbits during rabbit oral papillomavirus infection [54] and in vaccination studies involving transplant patients [55] . In accordance to this and supporting a role for neutralizing antibodies in the protection from tumor formation , no lesions appeared in the vaccinated animals which underwent immunosuppression . Conversely , animals which were not vaccinated developed tumors ( Fig . 4 ) , as was seen in the immunocompetent animals . Although tumor incidence in immunosuppressed Mastomys was higher in experimentally infected and lower in naturally infected animals , neither of these differences were significant . The size of the immunosuppressed groups in this study was not designed to assess whether there is an effect of immunosuppression on tumor incidence and , therefore , we cannot draw any conclusions in this regard . Additionally , the administration regimen of CsA could have played a role in the lack of increased tumor formation , as it was recently shown in a UV-irradiated mouse model of skin cancer that CsA administered in a bolus rather than continuously is able to increase the number of tumors [56] . Nonetheless , we show that this vaccination strategy has credible potential in the difficult fight against skin lesions in immunosuppressed patients , especially those occurring in organ transplant recipients . Given the variety of cutaneous HPV types that could be involved in the development of non-melanoma skin cancer in humans ( such as HPV23 , 38 and the EV types HPV5 and HPV8 ) [57] , a broadly protective HPV vaccine would be ideal . Anti-L1 neutralizing antibodies are known to be highly HPV-type specific [12] . The papillomavirus minor capsid protein L2 , however , contains a major cross-neutralizing epitope that could be used to develop a second generation vaccine [30] . We are currently investigating whether administration of an L2-vaccine is as effective in preventing skin tumors in Mastomys coucha as its L1 counterpart . In summary , our data provide the first evidence that a vaccine targeting a cutaneous papillomavirus can effectively prevent the appearance of skin tumors even in animals which are already infected or will undergo systemic immunosuppression . Our results provide the basis for the clinical development of potent vaccination strategies against cutaneous HPV infections .
Mastomys coucha from the DKFZ breeding colony were maintained under conventional conditions ( 21–24°C , 55% relative humidity with 12–16 air changes per hour , mouse breeding diet and water ad libitum ) . Animals were checked monthly for the appearance of tumors . When tumors reached a size of 2 cm , the animals were sacrificed and blood was taken by cardiac puncture . Every two months , blood was taken from the submandibular vein . Animals were monitored throughout their whole lifetime until they had to be sacrificed for reasons of tumor development or decrepitude . Two weeks after the last vaccination dose ( age: 3 . 2 months ) , animals in the immunosuppressed groups were switched to the same mouse chow supplemented with cyclosporine A ( CsA , pharmaceutical grade powder; Fagron , Barsbuettel , Germany ) at a concentration of 250 mg/kg during the first 3 months of treatment and 125 mg/kg for the rest of the animal's life . Higher doses were used in the beginning to reflect the level of immunosuppression administered immediately after organ transplantation in OTR , which is typically reduced after a few months [58] . The drug–mouse chow premix was cold pelleted according to standard procedures by SNIFF Spezialdiaeten ( Soest , Germany ) . CsA concentration was measured in heparinized blood by HPLC ( University Hospital Regensburg Clinical Laboratories ) to confirm immunosuppressive levels were reached . M . coucha at the DKFZ are maintained in compliance with German and European statutes and all animal experiments were undertaken with the approval of the responsible Animal Ethics Committee ( Regional Council of Karlsruhe , Germany; 35-9185 . 81/G-124/08 ) . Hysterectomies were performed on pregnant Mastomys under sterile conditions [31] . The offspring ( four females and one male ) were nursed by foster specified pathogen-free ( SPF ) mice ( Mus musculus ) , kept in a specific pathogen free isolator unit at the DKFZ . From these animals , a virus-free colony was established . Testing for anti-MnPV and McPV2 antibodies as well as for viral DNA was regularly done on the progeny as described [23] . A wart extract containing infectious MnPV particles was obtained by grinding a frozen papilloma in a Mikro-Dismembrator S ( Sartorius ) and resuspending the homogenate in PBS . Before experimental infection , Mastomys from the virus-free colony were anaesthetized with 2 . 5% isoflurane . A 1 cm diameter patch on the back was shaved and gently scarified using a scalpel blade and 20 µL of the wart extract was applied to the patch . Animals were placed on their abdomens until the virus suspension dried before returning them to their cages . Animals were sacrificed by cervical dislocation and dissected tissue samples were frozen at −20°C . To avoid cross-contamination among different tissues , surgical instruments were changed after every sample dissection . The DNA was extracted as previously described [23] . Quantification of MnPV DNA was performed with the iTaq Universal SYBR Green Supermix ( BioRad ) using 50 ng of total DNA per reaction , following the manufacturer instructions . Detection was done with the CFX96 real time PCR detection system ( Bio-Rad ) . Binding sites of MnPV primers were located within the L1 genes ( forward primer: 5′-ACGGCAACTCATGCTTCTTC-3′ , reverse primer: 5′-CTCTGTGCCTGTCCATCCTT-3′ ) . To determine the number of input cell equivalents , the single-copy-number gene β-globin was validated and quantified ( forward primer: 5′-ACCATGGTGCACCTTACTGAC-3′ , reverse primer: 5′-TCCAGGCACCCAACTTCTAC-3′ ) . MnPV DNA copy numbers were determined in duplicate by using standard curves generated in the same PCR run with a standard containing MnPV and β-globin plasmids . Sensitivity was 5 copies of MnPV DNA per sample and quantification was linear from 5 to 5×108 copies MnPV . MnPV DNA load was defined as the number of MnPV DNA copies/2 β-globin copies [59] . ISH was performed as previously described [28] . Tissue sections were fixed in 4% buffered formaldehyde , embedded in paraffin , cut at 3 µm and stained by hematoxylin eosin or used for immunostaining . Immunohistochemistry was performed for keratin 14 ( K14 ) , laminin and the proliferation marker Ki-67 . K14 was labeled with a guinea pig antibody ( k14 . 2; Progen , Heidelberg , Germany ) at a dilution of 1∶100 and laminin was labeled with a 1∶50 dilution of rabbit polyclonal antibody donated by Dr . Stark , DKFZ , Heidelberg ( ln ( 537 ) ) ; Ki-67 was detected with a mouse monoclonal IgG1 ( NCL-L-Ki67-MM1; Novocastra ) used at a dilution of 1∶100 . Sections stained with k14 . 2 were pretreated by cooking for 20 minutes in citrate buffer , pH 6 . 0 and sections stained for Ki-67 were additionally pretreated with proteinase K for 15 minutes at 37°C . Staining was performed by streptavidin coupled to horseradish peroxidase or alkaline phosphatase , as described [60] . Sf9 cells ( Invitrogen ) were kept as described elsewhere [61] . Full-length MnPV L1 gene was cloned into the pVL1393 vector ( Invitrogen ) by PCR amplification . The construct was confirmed by DNA sequencing . For the production of recombinant MultiBac AcNPVs , 2 µg of the pVL1393 derived recombinant L1-encoding plasmid and 1 µg of MultiBac bacmid DNA were co-transfected by calcium phosphate precipitation using 1 ml Sf9 transfection buffer . After incubation for 5 h at 27°C and 5% CO2 , the cells were washed twice and maintained in Grace's medium under the same conditions for one week . The recombinant Ac virus was amplified before its employment for a productive infection of TN High Five cells and the titer determined by a plaque assay . For VLP production , 2×108 TN High Five cells were infected with wild-type or recombinant baculovirus at a MOI of 2 . Three days post-infection , cells were harvested and lysed by sonication . Subsequently , the lysate was cleared by centrifugation , layered onto a two-step gradient with 14 ml of 40% sucrose on top of 8 ml of a 57 . 5% CsCl solution and centrifuged for 3 h at 96 , 500× g at 10°C using a SW32 rotor ( Beckman ) . The interphase was collected and transferred into a Quick-seal tube ( Beckman ) . A CsCl gradient was produced by a 16 hour-centrifugation at 184 , 000× g at 20°C in a Sorval TFT 65 . 13 rotor and fractionated into 1 ml aliquots . Purity and L1 content of the collected fractions were assessed by SDS-PAGE and Coomassie staining . The peak fractions were pooled , dialyzed against 50 mM Hepes ( pH 7 . 4 , 0 . 3 M NaCl ) , and cleared from residual debris by centrifugation at 20 , 000× g for 10 min at 4°C . L1 protein concentrations were determined using image densitometry software ImageJ ( http://rsb . info . nih . gov/ij/ ) and bovine serum albumin as standard for the Coomassie-stained SDS-PAGE gel . The capsid quality was verified by electron microscopy . We produced a large batch of high quality VLPs sufficient for the whole vaccination study . Fractions were analyzed by electron microscopy four months and one year after storage , which revealed a robust stability of the particles . Electron microscopy ( EM ) was performed by Birgit Hub ( DKFZ , Heidelberg ) using a Zeiss EM-10 transmission electron microscope . The structure and quality of particles derived from VLP preparations were analyzed by negative staining . VLPs ( approximately 100 ng ) were applied on carbon coated grids , stained with 2% uranyl acetate and analyzed in the EM at 20 , 100 fold magnification . Animals were immunized at an age of 8 weeks with 10 µg dialyzed VLPs in the presence of the Sigma Adjuvant System ( SAS ) , containing monophosphoryl lipid A ( MPL ) and synthetic trehalose dicorynomycolate in squalene and Tween80 . The formulations were prepared as suggested by the manufacturer . The vaccine was delivered in VLP vaccine buffer ( 50 mM Hepes , pH 7 . 4 , 0 . 3 M NaCl ) and a volume of 200 µl was injected subcutaneously in a skin fold of the neck . Control animals were immunized with 200 µl of vaccine buffer alone . For MnPV VLP-ELISA , 96-well plates ( Polysorp , Nunc ) were coated overnight at 4°C with 1 µg/ml VLPs in 50 mM carbonate buffer pH 9 . 6 and blocked the next day with casein blocking buffer ( 0 . 2% casein in PBS , 0 . 05% Tween 20 ) . Plates were subsequently incubated 1 h at room temperature with three-fold dilutions of the sera in casein blocking buffer . Plates were washed and bound specific Mastomys IgG was detected with a HRP-conjugated goat anti-mouse IgG antibody ( heavy+light chain; Promega , Mannheim , Germany ) diluted 1∶10 , 000 in blocking buffer . Color development was performed by addition of 0 . 1 mg/ml tetramethylbenzidine ( Sigma , Steinheim , Germany ) and 0 . 006% H2O2 in 0 . 1 M sodium acetate pH 6 ( 100 µl/well ) . After 8 min , the enzyme reaction was stopped with 50 µl of 1 M sulfuric acid per well and the absorbance was measured in an automated microplate reader ( Labsystems Multiskan; Thermo Fisher Scientific , Waltham , USA ) at a wavelength of 450 nm . Antibody titer represents the last reciprocal serum dilution above blank . Pseudovirions were generated as previously described [62] with some modifications . 293TT cells were cotransfected with a plasmid encoding humanized MnPVL1 and L2 genes and a Gaussia luciferase reporter gene plasmid using TurboFect transfection reagent ( Fermentas ) according to the manufacturer's instructions . Cells were incubated at 37°C , 5% CO2 for 72 h , harvested and resuspended in the same volume of DBPS supplemented with 0 . 5% of Brij 58 ( Sigma ) and 1% RNAse A/T ( Fermentas ) . Cells were incubated overnight at 37°C under rotation to allow pseudovirion maturation . On the next day , pseudovirus were extracted in 0 . 8 M NaCl and incubated with 250 U of benzonase ( Merck ) for 1 h at 37°C . The pseudovirions were subsequently purified by an Optiprep ( Sigma ) gradient . Fractions with highest reporter gene expression were pooled and aliquots were stored in siliconized tubes at −80°C . The neutralization assays were performed as previously described [62] with some modifications . Briefly , 60 µl of diluted polyclonal or monoclonal antibodies were combined with 40 µl of diluted pseudovirus stocks and incubated at room temperature for 20 min . Next , 50 µl HeLaT cells ( HeLaT-clone 4 ) [63] ( 2 . 5×105 cells/ml ) were added to the pseudovirus-antibody mixture and incubated for 48 h at 37°C , 5% CO2 . The amount of secreted Gaussia luciferase in 10 µl of cell culture medium was determined using coelenterazine substrate and Gaussia glow juice ( PJK , Germany ) according to the manufacturer's instructions . A microplate luminometer ( Synergy 2 , BioTek ) was used to measure the samples 15 min after substrate addition . To assess the role of neutralizing antibodies on the skin infections , we performed a virion neutralization assay . Basically , a patch on the back of each anesthetized Mastomys was shaved with an electric razor and slightly scarified with a scalpel blade . Polyclonal serum from 5 vaccinated animals was pooled and diluted 1∶10 in PBS and 200 µL injected intraperitoneally one day before challenge with 20 µL of a wart extract containing MnPV infectious particles . One week after challenge , the animals were sacrificed and samples were taken from the treated areas of the skin . Total cellular RNA was extracted by using the RNeasy Kit ( QIAGEN ) , according to the manufacturer's instructions . To eliminate all traces of viral DNA to avoid false positive signals by RT-PCR , the RNA was additionally treated with DNase I ( QIAGEN ) . Reverse transcription was performed with the reverse transcriptase SuperScript II ( Invitrogen ) , according to the manual . To evaluate the infection with MnPV virions , the abundant E1∧E4 transcript was detected ( Vinzón et al . , unpublished ) after 32 PCR cycles with a forward primer spanning the splicing junction at nucleotides 808∧3144 and an appropriate reverse primer ( forward primer: 5′-TGAAGAAGCTCTACACCGCA-3′ , reverse primer: 5′-GTCTCCTCCTTTCGGGTGC-3′ ) . As a control , Mastomys GAPDH transcript was determined after 25 PCR cycles ( forward primer: 5′-CTTCATTGACCTCAACTACATGGTC-3′ , reverse primer: 5′- CACAGTCCATGCCATCACTGC-3′ ) . Prism 6 . 0 ( GraphPad Software ) was used for data analysis and graphic representation . Statistical analysis for comparisons of antibody titer of viral load was performed with the non-parametric Mann-Whitney U test . The distribution functions describing tumor incidence were determined using the Kaplan-Meier estimator . Differences in tumor incidence times between the different groups were evaluated using the log rank test .
|
Organ transplant recipients ( OTR ) frequently suffer from fulminant warts that are induced by cutaneous human papillomaviruses ( HPV ) . Moreover , some skin HPV types may also be involved in the development of non-melanoma skin cancer . Mimicking the situation of immunosuppressed OTR who acquire cutaneous HPV infections already in childhood , we explored the efficacy of a vaccine in infected animals that additionally underwent immunosuppression . We demonstrate for the first time the success of a vaccine against a skin papillomavirus in a natural outbred animal system , which completely prevents both benign and malignant skin tumor formation even under immunosuppressed conditions . Hence , our study provides the basis for clinical development of a vaccine against cutaneous HPV infections , which may be particularly useful in transplant recipients .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"vaccines",
"oncology",
"medicine",
"vaccination",
"dermatology",
"viral",
"vaccines",
"skin",
"neoplasms",
"clinical",
"immunology",
"immunity",
"virology",
"benign",
"skin",
"neoplasms",
"biology",
"microbiology",
"cancer",
"prevention",
"cancer",
"vaccines",
"malignant",
"skin",
"neoplasms"
] |
2014
|
Protective Vaccination against Papillomavirus-Induced Skin Tumors under Immunocompetent and Immunosuppressive Conditions: A Preclinical Study Using a Natural Outbred Animal Model
|
Human T-cell leukemia virus type 1 ( HTLV-1 ) infection is linked to the development of adult T-cell leukemia ( ATL ) and the neuroinflammatory disease HTLV-1 associated myelopathy/tropical spastic paraparesis ( HAM/TSP ) . The HTLV-1 Tax protein functions as a potent viral oncogene that constitutively activates the NF-κB transcription factor to transform T cells; however , the underlying mechanisms remain obscure . Here , using next-generation RNA sequencing we identified the IL-25 receptor subunit IL-17RB as an aberrantly overexpressed gene in HTLV-1 immortalized T cells . Tax induced the expression of IL-17RB in an IκB kinase ( IKK ) and NF-κB-dependent manner . Remarkably , Tax activation of the canonical NF-κB pathway in T cells was critically dependent on IL-17RB expression . IL-17RB and IL-25 were required for HTLV-1-induced immortalization of primary T cells , and the constitutive NF-κB activation and survival of HTLV-1 transformed T cells . IL-9 was identified as an important downstream target gene of the IL-17RB pathway that drives the proliferation of HTLV-1 transformed cells . Furthermore , IL-17RB was overexpressed in leukemic cells from a subset of ATL patients and also regulated NF-κB activation in some , but not all , Tax-negative ATL cell lines . Together , our results support a model whereby Tax instigates an IL-17RB-NF-κB feed-forward autocrine loop that is obligatory for HTLV-1 leukemogenesis .
The retrovirus human T-cell leukemia virus type 1 ( HTLV-1 ) infects between 10–20 million people worldwide [1] . HTLV-1 is the etiological agent of the neuroinflammatory disease HTLV-1-associated myelopathy ( HAM/TSP ) and adult T-cell leukemia ( ATL ) , a CD4+CD25+ T-cell malignancy [2] , [3] . ATL develops in about 5% of HTLV-1-infected individuals after a long latent period spanning 40–60 years [4] . The HTLV-1 genome encodes the Tax protein that exerts pleiotropic roles and is an essential regulator of viral replication and oncogenic cell transformation [5] . Tax modulates the activation of several key signaling pathways and cell cycle proteins to enhance T-cell proliferation and survival . One of the key cellular targets important for transformation by Tax is the NF-κB transcription factor [6] . NF-κB is composed of heterodimeric DNA binding proteins consisting of RelA , c-Rel , RelB , p50 and p52 [7] . In the canonical NF-κB pathway , NF-κB heterodimers are sequestered in the cytoplasm by ankyrin-repeat containing inhibitory proteins including IκBα [8] . A wide variety of stimuli including stress signals , proinflammatory cytokines or virus infection activate the IKK kinase complex consisting of the catalytic subunits IKKα and IKKβ and the regulatory subunit IKKγ ( also known as NEMO ) [9] . IKKβ phosphorylates IκB proteins to trigger their ubiquitin-dependent degradation thus allowing NF-κB to enter the nucleus and activate target genes [10] . In the noncanonical NF-κB pathway , tumor necrosis factor receptor ( TNFR ) superfamily members including BAFF , lymphotoxin-β and CD40 promote proteasomal processing of the p100 ( NF-κB2 ) precursor protein to yield p52 , which forms transcriptionally active heterodimers with RelB . The NF-κB inducing kinase ( NIK ) is a key regulator of this pathway by activating IKKα homodimers which in turn phosphorylate p100 leading to its processing . Tax constitutively activates both the canonical and noncanonical NF-κB pathways , in part by interacting directly with NEMO and IKK [11]–[14] . There is evidence that Tax may require upstream signaling molecules such as the kinase TAK1 to activate canonical NF-κB signaling [15] . Although the proximal signaling components of TNFR and interleukin-1 receptor ( IL-1R ) are dispensable for Tax to activate NF-κB [16] , whether Tax has usurped a distinct NF-κB pathway is unknown . Tax activation of the canonical and noncanonical NF-κB pathways fosters the aberrant expression of anti-apoptotic and pro-proliferative genes that leads to oncogenesis . Tax mutants defective in NF-κB activation expressed in an infectious HTLV-1 molecular clone are impaired in the immortalization of primary T cells [17] . NF-κB is also required for the survival of HTLV-1 transformed cell lines and patient-derived ATL cells [18] . Therefore , HTLV-1 transformed cell lines and primary ATL leukemic specimens exhibit a strict “addiction” to NF-κB for survival and proliferation , thus establishing the NF-κB pathway as an attractive target for novel ATL therapeutics . However , since Tax expressing cells are vigorously targeted by cytotoxic T cells and other arms of the host immune response , the majority of ATL tumors exhibit downregulated or lost Tax expression by mutations within Tax or deletion or methylation of the 5′ viral long terminal repeat region ( LTR ) [19] . Thus , Tax likely plays more important roles in the early events of transformation via persistent NF-κB activation , inactivation of p53 and other tumor suppressors and induction of genomic instability and aneuploidy [5] . However , canonical and noncanonical NF-κB signaling remains constitutive in ATL despite the loss of Tax . The interleukin 17 ( IL-17 ) cytokine family consists of six members ( IL-17A-F ) that play essential roles in host immunity and inflammatory diseases . IL-17A is the signature IL-17 cytokine and binds to an IL-17RA/IL-17RC receptor complex to orchestrate the host response against bacterial and fungal infections [20] . IL-17A controls the expression of cytokines and chemokines that enhance neutrophil recruitment . Dysregulation of this pathway has been implicated in numerous autoimmune and metabolic diseases and cancer [21] . IL-17E ( also known as IL-25 ) is essential for host defense against parasites by regulating expression of T helper 2 ( Th2 ) cytokines including IL-4 , IL-5 and IL-13 that promote eosinophil recruitment [22] . IL-25 has also been linked to allergic airway inflammation and asthma [23] . IL-25 is produced by diverse cell types such as epithelial cells , T cells , eosinophils , mast cells and basophils [24] , [25] . IL-25 binds to a heterodimeric receptor composed of IL-17RA and IL-17RB , of which IL-17RB is the specific receptor subunit for IL-25 [26] . IL-17RB is highly expressed in kidney , liver and other peripheral organs as well as memory and effector T lymphocytes [27] . IL-17RB expression can be regulated by IL-4 and TGF-β , however the precise transcriptional regulatory control of IL-17RB is unknown . Upon binding to IL-25 , IL-17RB recruits the Act1 ( also known as CIKS ) adaptor molecule via homotypic SEFIR ( similar expression to fibroblast growth factor genes and IL-17R ) domain interactions [28] , [29] . Act1 activates the ubiquitin ligase TRAF6 and the kinase TAK1 that in turn triggers NF-κB and MAP kinase activation to induce type 2 cytokines IL-4 , IL-5 and IL-13 as well as IL-9 [30] , [31] . IL-17B also serves as a ligand for IL-17RB , albeit with a lower affinity for IL-17RB compared to IL-25 [32] . In addition to IL-17RB regulation of host defense and allergic airway disease , this pathway can be oncogenic if dysregulated . The IL-17RB locus is a common site of retroviral integration in murine myeloid leukemias , resulting in the upregulation of IL-17RB expression [33] . IL-17RB is also overexpressed in a subset of breast tumors and is associated with poor prognosis [34] . In breast cancer , IL-17RB engagement by IL-17B triggers TRAF6 recruitment to IL-17RB , NF-κB activation and induction of the bcl-2 gene to inhibit apoptosis [34] . Although considerable progress has been made in our understanding of HTLV-1 oncogenesis , the precise mechanisms underlying HTLV-1-induced transformation remain unclear . Previous microarray studies have identified several anti-apoptotic , cell cycle and growth regulatory genes dysregulated by HTLV-1 [35]–[37] . However , due to experimental limitations of these studies and the advent of next-generation sequencing , RNA sequencing ( RNA-Seq ) has emerged as a powerful tool to evaluate gene expression , differential splicing , noncoding RNAs , RNA editing and gene fusions [38] . In this study , we used RNA-Seq to delineate the transcriptome of primary T lymphocytes immortalized by HTLV-1 . This work led to the identification of IL-17RB as an aberrantly overexpressed gene in HTLV-1 transformed cells that was induced by the HTLV-1 Tax protein . Surprisingly , the IL-17RB pathway was required for constitutive NF-κB activation by Tax and in HTLV-1 transformed cell lines . Furthermore , IL-17RB was overexpressed in leukemic cells from acute ATL patients and was essential for NF-κB activation in a subset of Tax-negative ATL cell lines .
To gain insight into the mechanisms of HTLV-1-induced T-cell immortalization , we used a well-established co-culture model [35] , [39] whereby primary human CD4+ T cells were purified by immunomagnetic beads from normal donor peripheral blood mononuclear cells ( PBMCs ) and co-cultured with lethally irradiated HTLV-1 transformed MT-2 cells ( to provide a source of HTLV-1 ) . Primary T cells were consistently immortalized in the presence of MT-2 cells between 6–8 weeks . Control T cells cultured in the absence of MT-2 did not proliferate after 4 weeks and were no longer viable at that time . The co-culture assay was performed with T cells from 4 independent blood donors . Of the 4 co-cultures , all produced immortalized T cell clones , however clone #1 ceased proliferation unexpectedly and was excluded from further studies . The immortalized T cell clones ( T-MT-2 ) #2-4 remained dependent on recombinant IL-2 for proliferation and expressed CD3 , CD4 and CD25 cell surface markers ( Figure 1A ) . Total RNA was harvested from T-MT-2 clone #2 ( week 12 after co-culture ) for RNA-Seq analysis as well as parental primary T cells ( week 0 ) , and T cells after 1 week of co-culture . A pure population of viable cells was obtained from the co-culture after removal of dead cells using magnetic labeling and separation . MT-2 RNA was also included as a control for RNA-Seq to confirm that the immortalized T cells expressed a unique genetic signature and were not simply MT-2 contaminants . RNA-Seq and bioinformatics analysis were performed with a total number of reads of 65 million ( week 0 ) , 73 million ( week 1 ) , 44 million ( week 12 ) and 52 million ( MT-2 ) . At 1 week after co-culture , the most abundant induced coding RNAs in the T cells were interferon-stimulated genes ( ISGs ) such as ISG15 , IFI27 , OAS1 and MX1 and these results were confirmed by real-time quantitative RT-PCR ( qRT-PCR ) ( Figure 1C and Table S1 ) . Conversely , the HTLV-1-immortalized T cells did not express ISGs , but rather expressed aberrant levels of genes regulating cell growth/cytokines ( IL-17RB , IL-5 , IL-9 , IL-13 , CADM1 ) , DNA damage ( DDIT4L ) , cell cycle ( CDC14B , CCNA1 ) , metabolism ( glycerol kinase 2 ) and migration/chemokines ( CCL1 , CXCR7 ) ( Table S2 ) . Also , these immortalized T cells had a distinct genetic signature compared to MT-2 cells ( Sequence read archive accession numbers SRS698576 and SRS698477 ) . Notably , many of the aberrantly expressed genes , including IL-17RB , have not previously been linked to transformation by HTLV-1 . IL-17RB was one of the most highly induced genes in HTLV-1 immortalized T cells ( Figures 1B and S1 and Table S2 ) . IL-17RB mRNA expression was sharply elevated in all 3 independent HTLV-1 immortalized T cell clones as shown by qRT-PCR ( Figure 1D ) . IL-25 , the high affinity ligand for IL-17RB , was expressed at variable levels in the three clones ( Figure 1E ) . Aberrant expression of CCL1 ( also known as I-309 ) , CXCR7 , DDIT4L , IL-9 and IL-13 in HTLV-1-immortalized clones was also confirmed by qRT-PCR ( Figure 1D and E ) . The chemokine CCL1 , shown previously to be overexpressed in ATL cells , functions in an anti-apoptotic autocrine loop [40] . Similarly , the chemokine receptor CXCR7 is induced by Tax and regulates the growth and survival of ATL cells [41] . Furthermore , Tax induction of both IL-9 and IL-13 may trigger the autocrine stimulation of HTLV-1 infected cells [42] , [43] . Taken together , our RNA-Seq studies have confirmed the dysregulation of known targets of HTLV-1 transformation and have also identified genes , such as IL-17RB , not previously demonstrated to be induced by HTLV-1 . Next , the cell surface expression of IL-17RB was examined in HTLV-1 transformed cell lines by flow cytometry . IL-17RB was highly expressed in the HTLV-1 immortalized T-cell clones and most HTLV-1-transformed cell lines , but not in Jurkat T cells ( Figure 1F ) . IL-17RB mRNA was also overexpressed in varying degrees in HTLV-1 transformed and ATL cell lines ( Figure 1F ) . Since IL-17RB forms heterodimers with IL-17RA , the expression of IL-17RA was examined in HTLV-1 transformed and ATL cell lines . IL-17RA and IL-25 mRNAs were also upregulated in a subset of HTLV-1 transformed and ATL cell lines ( Figure 1G and H ) . A previous study reported a role for TGF-β and IL-4 in the upregulation of IL-17RB expression [31] . Since NF-κB is important for the proliferation and survival of HTLV-1 transformed cells , we hypothesized that NF-κB may regulate IL-17RB induction . Thus , the HTLV-1 transformed T-cell lines C8166 and MT-2 were treated with sc-514 , a small molecule inhibitor of IKKβ , and qRT-PCR was performed for IL-17RB and the known NF-κB target gene CD25 . Treatment with sc-514 significantly diminished the expression of IL-17RB and CD25 mRNAs in these cells ( Figure 2A ) , thus supporting a role for IKKβ and NF-κB in the expression of IL-17RB in HTLV-1 transformed cells . However , sc-514 treatment had no effect on IL-17RB expression in Jurkat cells ( Figure 2A ) . To provide further evidence for a role of IKK in the regulation of IL-17RB , recombinant lentiviruses expressing either control scrambled short hairpin RNA ( shRNA ) or two distinct shRNAs specific for IKKα or IKKβ were transduced into C8166 cells . Both IKKα and IKKβ shRNAs strongly suppressed their respective mRNAs as shown by qRT-PCR , and these shRNAs significantly inhibited IL-17RB expression ( Figure 2B ) . Therefore , both IKKα and IKKβ regulate IL-17RB expression in C8166 cells . The HTLV-1 Tax oncoprotein dysregulates the expression of specific cellular genes as part of its oncogenic mechanism [44] . To determine if Tax was involved in the induction of IL-17RB expression we used a Jurkat cell line inducible for Tax expression ( Jurkat Tax Tet-On ) by doxycycline ( Dox ) [45] . Jurkat Tax Tet-On cells were treated with Dox for 1 , 2 and 3 days and mRNA was harvested for qRT-PCR for IL-17RB . Indeed , induction of Tax strongly upregulated IL-17RB mRNA ( Figure 2C ) . Conversely , shRNA-mediated knockdown of Tax in C8166 cells diminished the expression of IL-17RB mRNA ( Figure 2D ) . Knockdown of Tax also reduced the expression of IL-17RB protein in C8166 cells ( Figure 2E ) . Thus , both gain-of-function and loss-of-function studies support the hypothesis that Tax is the HTLV-1-encoded protein that promotes the aberrant overexpression of IL-17RB . Two commonly used Tax mutants M22 ( Thr130Leu131->Ala130Ser131 ) and M47 ( Leu319Leu320->Arg319Ser320 ) can be used to distinguish NF-κB or CREB-specific functions of Tax [46] . Tax M22 is defective for NF-κB and wild-type for CREB activation , whereas Tax M47 is defective for CREB and wild-type for NF-κB activation . Wild-type Tax , Tax M22 and Tax M47 were cloned into a lentiviral vector and recombinant Tax-expressing lentiviruses were used to transduce Jurkat T cells . Wild-type Tax and Tax M47 strongly upregulated IL-17RB mRNA expression as detected by qRT-PCR , however Tax M22 induction of IL-17RB was significantly diminished ( Figure 2F ) . These data further support the notion that Tax requires NF-κB to induce IL-17RB expression . Interestingly Tax induction of IL-17RB was not observed in 293 cells suggesting that this event may be specific for T cells ( Figure 2G ) . Tax activation of NF-κB was also independent of IL-17RB in 293 cells , since knockdown of IL-17RB in 293 cells had no effect on Tax activation of an NF-κB reporter ( Figure 2H ) . Finally , the expression of IL-25 was not regulated by Tax in T cells , therefore Tax induces the expression of IL-17RB but not its high affinity ligand ( Figure 2I ) . Since IL-17RB signals to NF-κB , we next asked if Tax required IL-17RB to trigger NF-κB signaling in T cells . Jurkat Tax Tet-On cells were transduced with lentiviruses expressing control or IL-17RB shRNA , yielding ∼60–70% knockdown efficiency ( Figure 3A ) . The cells were transiently transfected with NF-κB and HTLV-1 LTR reporters for dual-luciferase assays and also treated with Dox to activate Tax expression . Remarkably , Tax activation of NF-κB , but not the HTLV-1 LTR ( which is CREB-mediated ) , was dependent on IL-17RB ( Figure 3A ) . In agreement with these results , Tax induction of the NF-κB target genes , CD25 and cIAP2 , was impaired when IL-17RB expression was suppressed with shRNAs ( Figure 3B ) . Therefore , Tax induces IL-17RB expression to establish a positive feedback loop that is critical for Tax-induced NF-κB activation . Also , the requirement of IL-17RB for Tax-mediated NF-κB activation appears to be T-cell specific . To determine the role of the IL-17RB pathway in the early events of HTLV-1 transformation of primary human T cells , we conducted an in vitro T-cell immortalization assay with irradiated MT-2 cells and PBMCs from normal donors as described earlier . In this co-culture model , expression of both IL-17RB and IL-25 were significantly increased in primary T cells at early times ( 1–2 weeks ) after co-culture with MT-2 cells ( Figure 4A ) . PBMCs were transduced with lentiviruses expressing control shRNA or shRNAs for IL-17RB or IL-25 , co-cultured with irradiated MT-2 cells and puromycin was added to select for cells expressing shRNAs . Both IL-17RB and IL-25 were required for immortalization of primary T cells by HTLV-1 since cells expressing these shRNAs ceased to proliferate after 3 weeks of co-culture ( Figure 4B ) . However , as expected PBMCs expressing control shRNA yielded immortalized CD4+ T cells after 8 weeks ( Figure 4C ) . Taken together , these results suggest that both IL-25 and IL-17RB are required for the early events involved in the immortalization of primary T cells by HTLV-1 . Because Tax required IL-17RB for efficient NF-κB activation and NF-κB is critical for the survival of T cells transformed by HTLV-1 , we hypothesized that IL-17RB was essential for NF-κB activation and the viability of established HTLV-1 transformed cell lines . To address this notion , recombinant lentiviruses expressing control or IL-17RB shRNAs were transduced into three distinct Tax-expressing HTLV-1 transformed cell lines ( C8166 , MT-2 and HUT-102 ) . A total of three independent shRNAs to IL-17RB or scrambled control shRNA were expressed in these cell lines and selected with puromycin . Efficient knockdown of IL-17RB was confirmed by qRT-PCR in MT-2 and C8166 cells ( Figure 5B ) . The CellTiter-Glo Luminescent Cell Viability kit was used to quantify cellular ATP levels to determine cell viability and proliferation . Knockdown of IL-17RB significantly reduced the viability and proliferation of HTLV-1 transformed cell lines ( Figure 5A ) . Next , we examined expression of the NF-κB target genes CD25 , cIAP2 , IRF4 and IL-9 by qRT-PCR . The expression of each of these genes was significantly attenuated upon IL-17RB knockdown in C8166 and MT-2 cells ( Figure 5B ) . Importantly , Tax expression was unaffected by IL-17RB knockdown in these cell lines ( Figure 5B ) . An NF-κB DNA binding electrophoretic mobility shift assay ( EMSA ) was next performed with nuclear extracts from C8166 , MT-2 and HUT-102 cells expressing control or IL-17RB shRNA . NF-κB DNA binding was completely abrogated upon IL-17RB suppression in C8166 and MT-2 cells , but not HUT-102 likely due to inefficient lentiviral transduction ( Figure 5C ) . IL-25 was also suppressed by shRNAs in MT-2 cells and shRNA#3 was effective in reducing IL-25 expression ( Figure 5D ) . Knockdown of IL-25 with this shRNA also significantly attenuated NF-κB DNA binding and the expression of CD25 ( Figure 5C and D ) . NF-κB signaling can also be monitored with phospho-specific antibodies for IKK and p65 since these proteins are phosphorylated upon activation . Phosphorylation of IKK and p65 was constitutive in C8166 , MT-2 and MT-4 cells but reduced upon knockdown of IL-17RB or IL-25 ( Figure 5C and E ) . IκBα protein was increased upon suppression of IL-17RB ( Figure 5C ) , likely reflecting enhanced stability due to a loss of IKK-induced phosphorylation and proteolysis . IL-17RB knockdown also triggered an apoptotic response in HTLV-1 transformed cells as revealed by PARP and caspase 3 cleavage ( Figure 5C ) . The TNFR cell surface receptors CD40 and OX40 activate NF-κB , are strongly induced by Tax and are overexpressed in HTLV-1 transformed cell lines [47] , [48] . However , knockdown of either CD40 or OX40 had no effect on the proliferation or viability of C8166 cells ( Figure 5F ) . Thus , IL-17RB is a receptor that appears to be uniquely required for NF-κB signaling and the survival of HTLV-1 transformed cells . Our earlier results indicated that HTLV-1 immortalized T-cell clones expressed aberrant levels of IL-9 ( Figure 1E ) . A recent study has demonstrated that the IL-17RB pathway controls IL-9 expression in the context of allergic airway inflammation [31] . Furthermore , Tax has been shown to induce IL-9 expression and IL-9 can regulate the proliferation of primary ATL cells [42] . In light of these findings , we hypothesized that IL-9 may represent a key downstream gene of IL-17RB that governs the proliferation of HTLV-1 transformed T cells . First , to determine if IL-17RB regulated a soluble factor that was important for the proliferation of HTLV-1-transformed T cells , C8166 and MT-2 cells were transduced with lentiviruses expressing control or IL-17RB shRNAs and the media was then replaced with conditioned media from the corresponding cells . As expected , suppression of IL-17RB significantly reduced the proliferation of both C8166 and MT-2 cells ( Figure 6A ) . However , the conditioned media rescued the proliferative defects associated with loss of IL-17RB ( Figure 6A ) , suggesting that a soluble factor ( s ) is sufficient to restore the growth of these cells . As described above , IL-9 represented an attractive candidate as a pro-proliferative soluble factor in the conditioned media from HTLV-1 transformed T cells . To determine if IL-9 was necessary to restore the proliferation of IL-17RB knockdown cells , we collected conditioned media from cells transduced with control or IL-9 shRNAs for the culture of C8166 cells expressing control or IL-17RB shRNA . Our results revealed that the conditioned media from cells with suppressed IL-9 expression was unable to restore the proliferation of C8166 cells expressing IL-17RB shRNA ( Figure 6B ) . As expected , conditioned media from cells with control shRNA effectively restored C8166 cell growth ( Figure 6B ) . IL-17RB and IL-9 knockdown were confirmed by qRT-PCR ( Figure 6B ) . Next , to determine if IL-9 was sufficient to rescue the growth defect associated with suppressed IL-17RB expression we provided recombinant IL-9 to the media of HTLV-1 transformed cell lines expressing IL-17RB shRNA . The results indicated that provision of IL-9 was sufficient to restore cell proliferation of both C8166 and MT-2 cells ( Figure 6C ) . Therefore , IL-9 is a key cytokine downstream of IL-17RB that governs the proliferation of HTLV-1-transformed T cells . IL-17RB can form a heterodimeric receptor complex together with IL-17RA [26] , and upon binding to IL-25 , the active receptor recruits the Act1 adaptor protein [28] . In addition , the ubiquitin ligase TRAF6 can be directly recruited to IL-17RB via a TRAF6 binding motif and plays a critical role in IL-17RB-mediated NF-κB activation and gene expression [30] . Given the vital role of IL-17RB in NF-κB signaling and survival of HTLV-1 transformed T cells , we sought to determine the requirements of the upstream signaling molecules that constitute this pathway . To this end , knockdown experiments were conducted in HTLV-1 transformed cell lines using shRNAs specific for TRAF6 , IL-17RA and Act1 . Interestingly , knockdown of TRAF6 , but not IL-17RA or Act1 , attenuated NF-κB activation as determined by western blotting for phosphorylated forms of IKK , p65 and IκBα ( Figures 7A , S2C and S3C ) . Knockdown of TRAF6 also diminished the expression of NF-κB target genes CD25 and cIAP2 as shown by qRT-PCR in HTLV-1 transformed T-cell lines ( Figure 7B ) . Knockdown of IL-17RA , but not Act1 , modestly reduced the expression of CD25 and cIAP2 ( Figures S2B and S3B ) . However , the proliferation of HTLV-1 transformed cell lines was strongly dependent on the expression of both IL-17RA and Act1 ( Figures S2A and S3A ) . These results suggest that IL-17RA and Act1 regulate the proliferation of HTLV-1 transformed cells in an NF-κB independent manner . Since IL-17RA regulates chemokine mRNA stability independently of NF-κB [49] , this mode of regulation may explain how IL-17RA and Act1 contribute to the proliferation of HTLV-1 transformed T cells . Thus far our experiments support a model of a Tax-induced IL-17RB-NF-κB feed-forward autocrine loop that is essential for the in vitro immortalization of primary T cells by HTLV-1 and the proliferation and survival of established HTLV-1 transformed T cell lines . However , the majority of ATL tumors ( ∼60% ) have downregulated or lost Tax expression [19] , and these malignant cells have acquired mechanisms to activate NF-κB persistently despite loss of Tax expression [50] . We next asked if IL-17RB played a role in NF-κB activation in ATL cells lacking Tax expression . Tax-negative ATL cell lines ATL-43T , ED40515 ( - ) , TL-OM1 and MT-1 were transduced with control or IL-17RB shRNA lentiviruses . Interestingly , proliferation and viability were significantly diminished in ATL-43T and TL-OM1 , but not in ED40515 ( - ) , MT-1 and control Jurkat cells ( Figure 8A ) . Knockdown of IL-17RB was efficient in all cell lines except ED40515 ( - ) , likely due to poor lentiviral transduction ( Figure 8B ) . Thus , IL-17RB appears to be important for some , but not all Tax-negative ATL cell lines since MT-1 cells proliferated normally despite knockdown of IL-17RB ( Figure 8A and B ) . The NF-κB target genes CD25 and cIAP2 were suppressed in TL-OM1 and ATL-43T cells , but not in the other ATL cell lines ( Figure 8B ) . Phosphorylation of IKK and p65 was also inhibited by IL-17RB knockdown in ATL-43T and TL-OM1 cells ( Figure 8C ) . The loss of NF-κB activation also triggered an apoptotic response as shown by PARP and caspase 3 cleavage ( Figure 8C ) . We next examined the expression of IL-17RB , IL-17RA , IL-25 , IL-17B and Tax in eight primary acute ATL leukemic specimens . IL-17RB expression was significantly overexpressed in 3 out of 8 ATL samples compared to normal control PBMCs ( Figure 8D ) . IL-17RA was modestly elevated in all the ATL samples compared to controls ( Figure 8D ) . However , IL-25 expression was not detected in any of the ATL samples ( Figure 8D ) . Tax mRNA was found in a subset of the samples but did not correlate with IL-17RB expression ( Figure 8D ) , suggesting that ATL cells may regulate IL-17RB expression independently of Tax . Surprisingly , IL-17B was overexpressed in the majority of acute ATL samples ( Figure 8E ) , thus raising the possibility that IL-17B may serve as a ligand for IL-17RB in acute ATL samples .
The IL-25-IL-17RB pathway has been linked to allergic airway inflammation and host defense against parasites . Our study has established a novel connection of this pathway to HTLV-1-induced leukemogenesis and also reveal that IL-17RB overexpression can be oncogenic in T cells . Tax promotes the aberrant expression of IL-17RB via NF-κB signaling to establish an IL-17RB-NF-κB feed-forward autocrine loop that drives persistent NF-κB activation in T cells , the natural host cell of HTLV-1 . Therefore , Tax has hijacked the IL-17RB-NF-κB signaling axis to sustain high levels of NF-κB and coordinate the induction of a gene program consisting of inflammatory cytokines , chemokines and anti-apoptotic proteins that orchestrates pathogenic T-cell proliferation and survival . Together , these results provide a new framework for how Tax and HTLV-1 persistently activate NF-κB to promote the malignant transformation of T cells . HTLV-1-induced leukemogenesis is a multi-step process that commences with the IL-2-dependent polyclonal expansion of HTLV-1 infected T cells . The Tax oncoprotein is thought to play critical roles in driving T-cell proliferation and survival in the early events of transformation by HTLV-1 . However , at later stages Tax expression is largely dispensable , presumably due to genetic and epigenetic changes that may compensate for the loss of Tax . The HTLV-1-encoded HBZ protein may also exert oncogenic roles in ATL tumors in the absence of Tax expression [51] . Nevertheless , after loss of Tax expression , ATL tumors still exhibit constitutive canonical and noncanonical NF-κB signaling that sustains tumor cell proliferation and survival . However , the mechanisms of Tax-independent NF-κB activation in ATL tumors remain poorly understood . A recent study demonstrated that epigenetic downregulation of the microRNA miR-31 led to overexpression of NIK and activation of noncanonical NF-κB [52] . Our results reveal that IL-17RB drives canonical NF-κB signaling in a subset of ATL cell lines suggesting that the IL-17RB-NF-κB autocrine loop can be maintained in the absence of Tax , most likely by the acquisition of genetic and/or epigenetic changes . Comparative genomic hybridization ( CGH ) analysis has elucidated specific chromosomal imbalances associated with each of the clinical subtypes of ATL [53] . The highly aggressive acute ATL acquires more frequent chromosomal abnormalities , including characteristic gains at chromosomes 3p , 7q and 14q and losses at chromosomes 6q and 13q [54] . Interestingly , IL-17RB is encoded on chromosome 3p21 . 1 , one of the most frequently amplified regions in acute ATL [54] . We found that IL-17RB is significantly overexpressed in leukemic cells from 3/8 ATL patients ( 38% ) , comparable to the 37% of aggressive ATL cases with 3p21 gains [54] . Therefore , IL-17RB overexpression in a subset of acute ATL tumors may potentially regulate the constitutive canonical NF-κB activation in the absence of Tax expression ( Figure S4 ) . Nevertheless , additional studies with more ATL patient tumor specimens are warranted to further explore the mechanisms underlying IL-17RB overexpression . It will also be interesting to determine if somatic mutations occur in IL-17RB that render the receptor constitutively active in the absence of ligand . Finally , since IL-17B but not IL-25 , was expressed by acute ATL leukemic cells ( Figure 8 ) , future studies will need to examine if IL-17B plays a role in IL-17RB signaling , NF-κB activation and proliferation of primary ATL cells . IL-25 expression may potentially be suppressed by active mechanisms in ATL since it exerts pro-apoptotic roles in other tumor types [55] . IL-25 serves as the high affinity ligand for IL-17RB . IL-25 favors Th2 immune responses and orchestrates host defense against parasites by inducing the expression of IL-4 , IL-5 and IL-13 [56] . IL-25 signals through a heterodimeric receptor containing IL-17RA/IL-17RB , which in turn recruits Act1 and TRAF6 upon IL-25 stimulation to induce NF-κB and MAPK activation that regulate genes important for Th2 immunity , allergic responses and expulsion of helminths . Our results indicate that HTLV-1 transformed cells are critically dependent on the IL-17RB pathway for proliferation , however only IL-17RB and TRAF6 are essential for NF-κB activation . IL-17RB contains a TRAF6 interaction motif in its intracellular domain that propagates downstream NF-κB activation [30] . Furthermore , a previous study has shown that TAK1 , a kinase downstream of TRAF6 in the IL-17RB pathway , is also involved in Tax-mediated NF-κB activation [15] . Therefore , IL-17RB may signal through both TRAF6 and TAK1 to activate IKK in HTLV-1 transformed cells . We have recently identified a consensus TRAF6 interaction motif in the C-terminal region of Tax that mediates TRAF6 interaction and activation [57] , thus suggesting that Tax may activate TRAF6 to further enhance the Tax-IL-17RB-NF-κB positive feedback loop in T cells . A previous study claimed that TRAF6 was dispensable for Tax-induced NF-κB activation [58] , however they used a cell-free assay system using lysates from murine embryonic fibroblasts . Using intact T cells , we found that TRAF6 indeed plays a role in Tax-mediated NF-κB signaling . Our results also indicate that IL-17RB is dispensable for Tax to activate NF-κB in 293 cells , yet is critical for Tax-mediated NF-κB activation in T cells . Therefore , Tax activation of NF-κB appears to be distinct in T cells compared to other cell types and provides a strong rationale for Tax/NF-κB studies to be conducted in T cells . Our data has provided new insight into the transcriptional regulation of IL-17RB . Little is known regarding how IL-17RB expression is regulated , although a previous study demonstrated that TGF-β and/or IL-4 can induce IL-17RB expression in mouse T cells [31] . We have provided multiple lines of evidence supporting a role for NF-κB in Tax-induced expression of IL-17RB . First , an IKKβ inhibitor greatly reduced the expression of IL-17RB in HTLV-1 transformed T cell lines ( Figure 2A ) . Second , knockdown of IKKα or IKKβ with shRNAs diminished IL-17RB expression in C8166 cells ( Figure 2B ) . Finally , the Tax M22 mutant , defective for NEMO binding and NF-κB activation , was impaired in the induction of IL-17RB ( Figure 2F ) . Taken together , our data support a two-step model of Tax activation of NF-κB in T cells ( Figure S4 ) . First , Tax activation of canonical NF-κB commences through direct NEMO/IKK binding and IKK activation . The precise mechanisms remain poorly understood but may involve IKK oligomerization and inhibition of NEMO-associated phosphatase 2A [59] , [60] . Next , Tax and IKK-induced IL-17RB overexpression ( and engagement by IL-25 ) triggers downstream signaling to TRAF6 and further activates IKK to establish a positive feedback loop resulting in strong and sustained NF-κB signaling . It remains unclear whether NF-κB directly regulates the expression of IL-17RB , although we have identified a putative NF-κB site ( GGGAATTTCC ) ∼3380 base pairs upstream of the human IL-17RB transcriptional start site . Future studies will be necessary to identify important regulatory elements in the IL-17RB promoter . IL-17RB forms heterodimers with IL-17RA , and although IL-17RA does not directly engage IL-25 it appears to be essential for IL-17RB signaling in untransformed cells [26] . Since IL-17RA and Act1 were largely dispensable for NF-κB activation in HTLV-1 transformed cells ( Figures S2 and S3 ) , it is plausible that IL-17RA and Act1 regulate the proliferation of these cells by stabilizing chemokine mRNAs [49] . Because IL-17RB is overexpressed to a much greater degree than IL-17RA in HTLV-1 transformed T cells , it is likely that IL-17RB homodimers constitute the most abundant IL-17R complex that signals to NF-κB in these cells . Further studies are needed to examine the stoichiometry of IL-17R complexes and downstream signaling requirements in HTLV-1 transformed cells . Although IL-25/IL-17RB signaling has been previously linked to the induction of IL-9 and Th2 cytokines [24] , our study has identified additional genes regulated by this pathway that contribute to oncogenesis . We found that knockdown of IL-17RB in HTLV-1 transformed cell lines diminished the expression of cytokines ( IL-9 ) , cytokine receptors ( CD25 ) , anti-apoptotic genes ( cIAP2 ) and transcription factors ( IRF4 ) . Elevated expression of IRF4 in ATL tumors was shown to correlate with resistance to antiviral therapy with zidovudine ( AZT ) and interferon alpha [61] . Furthermore , cIAP2 was identified as a Tax regulated anti-apoptotic gene that was required for the survival of HTLV-1 transformed T cells [62] . IL-9 was also demonstrated to function as a key proliferative factor for ATL cells [42] . Consistently , we found that IL-9 was both necessary and sufficient to restore the cell proliferation of HTLV-1 transformed T cells with IL-17RB knockdown ( Figure 6B and C ) . These data support the notion that IL-9 is an important downstream target gene of IL-17RB that drives the proliferation of HTLV-1 transformed cells . Additional studies will be necessary to identify the full spectrum of genes regulated by IL-17RB in HTLV-1 transformed T cells that support oncogenic proliferation . Therapeutic blocking antibodies , such as those targeting HER2 and EGFR , have emerged as an important new treatment option in the clinic for carcinomas of the breast , lung and colon [63] , [64] . IL-17RB is overexpressed in a subset of breast tumors and is associated with poor prognosis [34] . Treatment with blocking IL-17RB therapeutic antibodies attenuated the tumorigenicity of breast cancer cells [34] . Given that IL-17RB overexpression can promote oncogenic NF-κB signaling in Tax-negative ATL tumors , this receptor may represent an attractive therapeutic target for ATL . IL-17RB may potentially serve as a biomarker to stratify ATL patients that could benefit from IL-17RB inhibition . Preclinical studies with IL-17RB ( or potentially IL-17B ) monoclonal blocking antibodies in both in vitro and in vivo ATL models will be required to establish the feasibility of this potential targeted therapy .
Blood from healthy donors was purchased from Biological Specialty Corporation ( Colmar , PA ) . PBMCs were collected from acute ATL patients ( n = 8 ) . This study was conducted according to the principles expressed in the Declaration of Helsinki . The study was approved by the Institutional Review Board of Kyoto University ( G204 ) . All patients provided written informed consent for the collection of samples and subsequent analysis . Human embryonic kidney cells ( HEK 293T ) and Jurkat T cells were purchased from ATCC . The HTLV-1-transformed cell lines MT-2 , HUT-102 , C8166 and MT-4 were described previously [65] , [66] . ED40515 ( - ) , MT-1 , and TL-OM1 cells are clones of leukemic cells derived from ATL patients , kindly provided by Dr . Michiyuki Maeda ( Kyoto University ) . ATL43T is a Tax-negative ATL cell line that was previously described [67] . Jurkat Tax Tet-On cells were kindly provided by Dr . Warner Greene [45] . 293T cells were cultured in Dulbecco's Modified Eagle's medium ( DMEM ) ; Jurkat , MT-2 , C8166 , MT-4 , ATL-43T , HUT-102 , ED40515 ( - ) , MT-1 and TL-OM1 cells were cultured in RPMI medium . Media was supplemented with fetal bovine serum ( FBS; 10% ) and penicillin-streptomycin ( 1× ) . MISSION shRNAs targeting human IL-17RB , IL-17RA , IL-25 , IL-9 , Act1 , CD40 , OX40 and control scrambled shRNA were purchased from Sigma . TRAF6 , Tax , IKKα and IKKβ shRNAs were cloned into pYNC352/puro . Target sequences for these shRNAs are listed in Table S3 . Tax WT , M22 and M47 were cloned in the pDUET lentiviral vector . Expression vectors encoding κB Luciferase ( Luc ) , pU3R-Luc , pRL-TK ( thymidine kinase ) have all been described previously [68] . Recombinant human IL-9 was purchased from R&D Systems . The IKKβ inhibitor SC-514 was from EMD Millipore . The following antibodies were used in this study: anti-hIL-17RB ( FAB1207P; R&D Systems ) , anti-hCD4 ( 555346; BD Pharmingen ) , anti-hCD3 ( 552851; BD Pharmingen ) , anti-hCD8 ( 555366; BD Pharmingen ) , anti-hCD25 ( 560989; BD Pharmingen ) , anti-β-actin ( AC15; Abcam ) , anti-IκBα ( SC-371; Santa Cruz Biotechnology ) , anti-phospho-IκBα ( 14D4; Cell Signaling ) , anti-p65 ( 8242S; Cell Signaling ) , anti-phospho-p65 ( 3031S; Cell Signaling ) , anti-IKKβ ( 2678; Cell Signaling ) , anti-phospho-IKKα/β ( 2697S; Cell Signaling ) , anti-IL-17RB ( SC-52925; Santa Cruz Biotechnology ) , anti-TRAF6 ( SC-7221; Santa Cruz Biotechnology ) , anti-PARP ( 9542S; Cell Signaling ) and anti-caspase-3 ( SC-7148; Santa Cruz Biotechnology ) . Human PBMCs from healthy donors were prepared from lymphocyte enriched human blood with a Ficoll-Hypaque gradient ( Pharmacia Biotech ) . Samples were tested and found to be negative for hepatitis B virus ( HBV ) , hepatitis C virus ( HCV ) and human immunodeficiency virus 1 ( HIV-1 ) . The cells were stimulated for 36 h with phytohemagglutinin ( PHA , 2 µg/ml ) and then cultured in RPMI medium supplemented with 20% FBS , 2 mM L-glutamine , penicillin-streptomycin , and 25 units/ml of human recombinant IL-2 ( Biological Resources Branch , NCI ) . Under these conditions , PBMCs continuously grew for up to 4 weeks in the presence of exogenous IL-2 . CD4+ T cells were isolated from PBMCs by negative selection using MACS MS Columns ( Miltenyi Biotec ) . The purity of the cells was confirmed by flow cytometry and was>95% . In vitro transformation of T cells with HTLV-I was performed as previously described [39] . Briefly , PHA-stimulated PBMCs were co-cultured with lethally γ-irradiated ( 50 Grays ( Gy ) ) HTLV-1 donor cells ( MT-2 ) in IL-2-containing RPMI medium . As expected , the virus-infected T cells became immortalized after about 6 weeks of co-cultivation . These cells proliferated vigorously when exogenous IL-2 was provided , a characteristic of T cells at an early stage of HTLV-1 infection . Under identical culture conditions , the uninfected control T cells or PBMCs typically ceased growth within 4 weeks , and the γ-irradiated MT-2 cells did not proliferate . The HTLV-1-immortalized T cells were maintained in RPMI medium supplemented with IL-2 and used as a bulk population . For shRNA knockdown studies , purified PHA-stimulated PBMCs were first infected with lentiviral particles expressing shRNAs to knockdown the indicated genes and subsequently co-cultured with lethally γ-irradiated ( 50 Gy ) HTLV-1 donor cells ( MT-2 ) in IL-2-containing RPMI medium . Puromycin was added after 3 weeks of co-culture to select for shRNA expressing cells . Total RNA was prepared from parental primary T cells , HTLV-1-infected cells after 1 week of co-culture , HTLV-1 immortalized T cell clones after 12 weeks of co-culture or MT-2 cells . Dead cells were removed from co-cultured cells after magnetic labeling and separation using the Dead Cell Removal Kit ( Miltenyi Biotec ) . RNA was isolated with RNeasy columns ( Qiagen ) . RNA-Seq and bioinformatics were conducted by the Johns Hopkins Sidney Kimmel Cancer Center next-generation sequencing core . Sequencing analysis was performed by aligning the paired end reads to hg19 using Bioscope . The differential expression analysis was performed using the DEseq R package and the GO enrichment was done with the topGO R package . Jurkat cells were transfected with TransIT-Jurkat ( Mirus ) according to the manufacturer's instructions . For lentivirus production , HEK293T cells were transfected with a lentiviral vector and gag/pol-encoding plasmids using GenJet ( SignaGen ) according to the manufacturer's instructions . Virus was harvested after 48 h by centrifugation at 49 , 000× g . Cells were transduced with lentivirus by the spinoculation protocol , cultured for 48 h and then selected with puromycin . For luciferase assays , cells were lysed 24 h after transfection using passive lysis buffer ( Promega ) . Luciferase activity was measured with the dual-luciferase assay system according to the manufacturer's instructions ( Promega ) . Firefly luciferase values were normalized based on the Renilla luciferase internal control values . Luciferase values are presented as “fold induction” relative to the shControl ( shCTR ) . Western blotting was performed essentially as described previously [69] . Whole cell lysates were resolved by SDS-PAGE , transferred to nitrocellulose membranes , blocked in 5% milk or bovine serum albumin ( BSA ) ( for phospho-specific antibodies ) , incubated with the indicated primary and secondary antibodies , and detected using Western Lightning enhanced chemiluminescence reagent ( Perkin Elmer ) . Quantitative real-time PCR ( qRT-PCR ) was performed as described previously [68] . Total RNA was isolated from cells using the RNeasy mini kit ( Qiagen ) . RNA was converted to cDNA using the First Strand cDNA synthesis kit for RT-PCR ( avian myeloblastosis virus [AMV]; Roche ) . Real-time PCR was performed using SYBR Green qPCR ( Sigma ) . Gene expression was normalized to the internal control 18S rRNA . PCR primers are listed in Table S4 . Cell viability and proliferation assay was determined using the CellTiter-Glo Luminescent Cell Viability Assay ( Promega ) . Cells were cultured in 96-well plates and the ATP content was quantified as an indicator of metabolically active cells . Small-scale nuclear extracts were prepared from cells as described previously [47] . The following sequence was used to generate double-stranded oligonucleotides for electrophoretic mobility shift assays ( EMSA ) : IL-2Rα NF-κB site: 5′-CAACGGCAGGGGAATCTCCCTCTCCTT . Nonradioactive EMSA was performed using LightShift Chemiluminescent EMSA Kit ( Thermo Scientific ) according to the manufacturer's instructions . Two-tailed unpaired T test was performed with Prism software . Error bars represent the standard deviation of triplicate samples . The level of significance was defined as: ***P<0 . 001 , **P<0 . 01 , *P<0 . 05 .
|
The retrovirus HTLV-1 is the causative agent of an aggressive lymphoproliferative disorder known as adult T-cell leukemia ( ATL ) . The HTLV-1 Tax regulatory protein constitutively activates the host NF-κB transcription factor to promote T-cell proliferation , survival and cell transformation . However , it remains unknown precisely how Tax persistently activates NF-κB in T cells . In this study , we used next-generation sequencing to identify genes that were differentially expressed upon HTLV-1 infection and immortalization of primary T cells . We found that IL-17RB , the receptor for the IL-25 cytokine , was highly induced in HTLV-1 transformed T cells and was required for NF-κB activation , cell proliferation and survival . Tax induced the expression of IL-17RB and established a positive feedback loop together with IL-25 that triggered persistent NF-κB activation and the upregulation of IL-9 and other genes critical for T-cell proliferation and survival . IL-17RB was also overexpressed in a subset of acute ATL patient specimens and therefore may potentially be targeted by monoclonal antibodies as a novel ATL therapy .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology",
"and",
"life",
"sciences",
"microbiology",
"medicine",
"and",
"health",
"sciences"
] |
2014
|
A Critical Role for IL-17RB Signaling in HTLV-1 Tax-Induced NF-κB Activation and T-Cell Transformation
|
Protein kinases use ATP as a phosphoryl donor for the posttranslational modification of signaling targets . It is generally thought that the binding of this nucleotide induces conformational changes leading to closed , more compact forms of the kinase domain that ideally orient active-site residues for efficient catalysis . The kinase domain is oftentimes flanked by additional ligand binding domains that up- or down-regulate catalytic function . C-terminal Src kinase ( Csk ) is a multidomain tyrosine kinase that is up-regulated by N-terminal SH2 and SH3 domains . Although the X-ray structure of Csk suggests the enzyme is compact , X-ray scattering studies indicate that the enzyme possesses both compact and open conformational forms in solution . Here , we investigated whether interactions with the ATP analog AMP-PNP and ADP can shift the conformational ensemble of Csk in solution using a combination of small angle x-ray scattering and molecular dynamics simulations . We find that binding of AMP-PNP shifts the ensemble towards more extended rather than more compact conformations . Binding of ADP further shifts the ensemble towards extended conformations , including highly extended conformations not adopted by the apo protein , nor by the AMP-PNP bound protein . These ensembles indicate that any compaction of the kinase domain induced by nucleotide binding does not extend to the overall multi-domain architecture . Instead , assembly of an ATP-bound kinase domain generates further extended forms of Csk that may have relevance for kinase scaffolding and Src regulation in the cell .
The Src family of tyrosine kinases ( SFKs ) is comprised of modular signaling enzymes involved in the control of cellular growth and differentiation . The members of this family contain three important structural domains: a C-terminal tyrosine kinase domain ( comprised of a small and large lobe ) , which is preceded in sequence by the non-catalytic regulatory SH2 and SH3 domains [1] , [2] , [3] , [4] . While phosphorylation of the activation loop is autocatalytic , phosphorylation of the C-terminal tail is inhibitory and requires Csk [5] . Csk contains the same structural domains as SFKs , but lacks an inhibitory C-terminal tail and an N-terminal sequence for membrane localization [6] . Additionally , Csk is not regulated through phosphorylation of its activation loop . Instead , Csk is constitutively active and increased activity is coupled to its association with membrane adaptor proteins . Each domain of Csk plays a role in interacting with various binding partners . The kinase domain binds and phosphorylates all nine members of the Src family [7] , a process that requires the binding of ATP and magnesium . The SH2 domain of Csk is responsible for binding a large number of scaffolding proteins . Since Csk is a cytosolic protein and the substrate Src is membrane localized , localization of Csk to Src requires interaction between the Csk SH2 domain and several scaffolding proteins . These scaffolding proteins include the ubiquitously expressed transmembrane protein Cbp , Caveolin-1 , Paxillin , the insulin receptor substrate IRS-1 , and Helicobacter pylori CagA [8] , [9] , [10] , [11] . Lastly , the SH3 domain has been shown to bind PKA and the phosphatase PEP [12] , [13] . The gamut of binding partners for Csk suggests it must be an adaptable protein in order to carry out all of its known functions . One aspect that may be central to this adaptability is nucleotide-derived conformational changes . Pre-steady-state kinetic studies suggest that a slow conformational change in Csk limits Src phosphorylation [14] , [15] . In addition , previous work from our lab examined the effects of nucleotide binding on Csk using hydrogen-deuterium exchange mass spectrometry ( DXMS ) and found binding of the ATP analog AMP-PNP as well as ADP led to changes in the protection of multiple peptides [16] , [17] , [18] . These data suggest that nucleotide binding to Csk has conformational effects on the protein that extend beyond the nucleotide pocket in the kinase domain and into the SH2 domain , the site of adaptor protein binding and catalytic regulation . While these studies establish the existence of long-range communication across Csk , they do not provide a structural framework in which to understand these observed inter-domain relationships . Small angle x-ray scattering ( SAXS ) and NMR studies combined with molecular dynamics simulations has become an increasingly powerful tool to understand the conformational ensemble properties of proteins in solution [19] , [20] , [21] , [22] , [23] , [24] , [25] , [26] . Using a combination of SAXS and structure-based models ( SBM ) , or SAXS-SBM , we showed that apo Csk is comprised of an ensemble of extended and compact conformations in solution , rather than adopting a single conformational form as depicted by the X-ray structure [27] . Here , we applied these methods to Csk to ask how nucleotide binding affects the conformational landscape of this flexible multi-domain kinase . We show that upon binding of the ATP analog AMP-PNP , there is a shift in the conformational ensemble of Csk towards more extended , open conformations . When bound to the reaction product ADP , the ensemble shifts even further towards these extended conformations . These data suggest that the orientation of regulatory domains in Csk is highly sensitive to the presence of nucleotides in the kinase domain . Most importantly , rather than inducing more compact forms , occupancy of the nucleotide pocket in the kinase domain induces more open forms of Csk .
We previously collected SAXS data on the apo form of Csk and identified an ensemble of compact and extended conformations that describe the protein in solution [27] . In the current study , we asked whether or not the binding of nucleotides to Csk might alter this ensemble . DXMS studies comparing apo Csk to nucleotide bound Csk ( AMP-PNP and ADP ) displayed changes in protection of many peptides , which suggests both the binding of ATP and delivery of the γ phosphoryl group of ATP induce global conformational changes in Csk [16] . In order to obtain a structural description of how binding of these nucleotides alters the conformation of Csk , we collected SAXS data on AMP-PNP and ADP bound Csk . To determine if there are significant global changes in the solution structure of these bound states , we initially compared the scattering profile of each protein using a Kratky plot ( Figure 1 ) . Kratky plots ( I ( q ) *q2 versus q ) provide a qualitative analysis of the “folded-ness” of a protein . Globular proteins display a bell shaped curve , whereas extended proteins lack a peak and will either plateau , or have elevated I ( q ) *q2 values for higher q values [28] . The apo protein plot indicates that Csk is well-folded with flexible linkers , which is consistent with our previous SAXS analysis , as well as crystallographic data [27] , [29] . However , both the AMP-PNP and ADP bound protein show signs of possible disorder and local unfolding , as indicated by the higher I ( q ) *q2 values at larger scattering angles ( higher q ) . While the AMP-PNP bound protein data shows moderate deviations from the apo , the ADP bound protein shows a dramatic shift towards more disorder . Additionally , the experimentally determined radius of gyration ( Rg ) values show the apo protein ( Rg = 39 Å ) to be more compact than the AMP-PNP bound ( Rg = 40 Å ) and the ADP bound protein ( Rg = 44 Å ) . Taken together , these data suggest significant shifts in the conformations of Csk when nucleotide is bound to the protein . To further investigate the nature of Csk's conformational change upon nucleotide binding , we modeled the experimental scattering data using SBM . In our previous studies using SAXS-SBM we showed the crystal structure does not accurately describe apo Csk in solution , but rather it is composed of an ensemble of compact and extended conformational states . The higher experimental Rg values of both the AMP-PNP and ADP bound forms of Csk suggest an ensemble of structures that are on average more extended than the apo protein . Using an all-atom SBM , we generated 40 , 000 candidate conformations for analysis . Theoretical scattering curves were generated for each individual conformation and then compared to the experimental scattering curve for all states of Csk using χ2 to measure the goodness of fit . Specifically , χ2 is a measure of the discrepancy between the theoretical and the experimental curves ( see Methods for more detail ) . For each experimental curve we selected candidate conformations from the simulation , using the 5% of conformations with the lowest χ2 values for each case . We then calculated P ( Rg ) ( Rg values calculated using the g_gyrate module of Gromacs ) , the probability in Rg space , for all low-χ2 conformations ( Figure 2 ) . This process effectively applies a low-χ2 filter to the candidate conformations , which range from Rg = 25 . 5 Å to Rg = 45 . 1 Å . In addition , we generated P ( Rg ) plots for the 2 . 5% and 7 . 5% of conformations with the lowest χ2 values and the data were consistent , ensuring the robustness of the 5% selection criteria . The low χ2 filtered sets each possess multiple distinct peaks , representing likely conformational populations of Csk in solution . These populations most likely contribute to the experimental signal , but the ensemble distribution in not extracted from P ( Rg ) directly ( see Methods ) . The apo protein prefers the most compact conformations , with peaks at ∼34 . 0 Å , ∼35 . 2 Å , and ∼37 . 5 Å , which is consistent with our previous analysis [27] . The AMP-PNP protein displays a shift towards more extended conformations with one major peak at ∼37 . 8 Å and two minor peaks at ∼39 . 5 Å and ∼41 . 6 Å . Lastly , the ADP bound protein displays two major peaks at ∼37 . 1 Å and ∼40 . 5 Å , as well as a minor but potentially significant population at ∼43 . 0 Å . Representative conformations from all peaks are depicted in Figure 3 . Common to all states of Csk studied here , we observe a similar major population ( Rg ∼37–38 Å ) that fits the experimental data well . However , each state of Csk displays conformations of differing Rg values that also describe the experimental data . The apo protein adopts more compact conformations than those observed in either of the nucleotide bound states . Upon binding of AMP-PNP , the predominant population ( peak 1 ) covers a broad Rg range centered around a peak at ∼37 Å , shifting away from the compact structures observed for the apo protein . When ADP is bound , the P ( Rg ) shifts even further towards higher Rg values , displaying two populations at high Rg that are not among the low χ2 conformations for the apo and AMP-PNP bound protein . Overall , the apo protein is the most compact form , binding of AMP-PNP shifts towards the more extended structures , and binding of ADP shifts to populations of even more extended conformations . While the P ( Rg ) analysis provides insight into the populations that are likely to be present in solution , it does not provide the probability of that population in solution . This is because each peak contains numerous conformations and possible bias by the simulation resulting in higher or lower sampling of a population . In order to quantify the probability of each population in solution we performed conformational ensemble analysis using conformations selected from the P ( Rg ) analysis . We selected four representative conformations from the three most prominent P ( Rg ) peaks for each state , where all the selected conformations are distinct ( RMSD>2 Å ) . Representative structures selected for ensemble analysis appear in Figure 3 . From this pool of candidate conformations we created all possible combinations that include one conformation from each population . Each combination was considered a candidate ensemble , with which we performed ensemble analysis ( see Methods ) . For each weighted ensemble , we generated a theoretical scattering curve , which was then compared to the experimental data ( Figure 4 ) . For each state of Csk , we find an ensemble containing conformations from the peaks in P ( Rg ) that accounts well for the experimental data . The apo protein is heavily weighted towards compact conformations , where the best-fitting ensemble ( χ2 = 2 . 6 ) arises from a mix of 90% peak 1 ( ∼34 Å ) and 10% peak 3 ( ∼37 Å ) . The absence of peak 2 ( ∼35 Å ) may be explained by a very small population or as a possible transient conformation between those represented at peaks 1 and 3 . Both the AMP-PNP and ADP bound Csk conformational ensembles are comprised of extended structures . For the AMP-PNP bound protein , 40% of the best-fit ensemble is comprised of a conformation from the population of peak 1 ( Rg ∼37 Å ) , with 30% each from peaks 2 ( Rg ∼40 Å ) and 3 ( Rg ∼41 Å ) . This ensemble fits the experimental data with a χ2 = 2 . 7 . When ADP is bound to Csk , there is a shift towards highly extended conformations with the best-fit ensemble consisting of 30% peak 1 ( Rg ∼38 Å ) , 30% peak 2 ( Rg ∼39 Å ) , and 40% peak 3 ( Rg ∼43 Å ) , with χ2 = 3 . 8 . The population analysis results in a unique conformational ensemble for each state of Csk . The apo protein adopts primarily compact conformations , whereas the AMP-PNP bound protein adopts more extended conformations and the ADP bound protein adopts highly extended conformations . To confirm that each ensemble best describes that state of Csk , we fit the theoretical curves for all of the best-fit ensembles to the experimental data of all other states of Csk ( Figure 4 , Table 1 ) . This comparison shows that the best-fit ensemble for each set of data describes the data better than any other ensemble does , reinforcing the notion that a unique solution is necessary for each state of Csk to describe the protein conformationally in solution .
Using SBMs to model SAXS data , we find Csk adopts a variety of conformations in solution that range from compact ( Rg = 34 Å ) to highly extended ( Rg = 43 Å ) . For every state of Csk examined here , the scattering data are best described by an ensemble of conformations , not a single structure . Interestingly , the P ( Rg ) distributions reveal a common population for the apo , AMP-PNP bound , and ADP bound protein , centered around 37 Å . For the apo protein , this population accounts for 10% of the best-fit ensemble and is the most extended conformation present in that ensemble . When AMP-PNP is bound to Csk , this population accounts for 40% of the best-fit ensemble , with the remainder of the ensemble composed of more extended conformations . Binding of ADP shifts this ensemble even further towards extended conformations , but still contains a contribution from the population common to all states ( 30% ) . Notably , 40% of the best-fit ensemble is comprised of a conformation from peak 3 ( 43 Å ) , a conformation that does not appear in the P ( Rg ) for either the apo protein or the AMP-PNP bound protein . This shift in the ensemble upon nucleotide binding , while surprising for a protein kinase , is consistent with observations for other proteins . Many recent studies highlight the role of dynamics and shifts in population during kinase catalysis . The energy landscape of PKA-C , as characterized using NMR and molecular dynamics , shows the apo enzyme can explore the landscape and access open and closed conformations , while ligand binding ( nucleotide , substrate , or inhibitor ) drives the enzyme to select alternate conformational states [30] . Combined crystallographic and NMR studies on arginine kinase display large substrate-induced domain motions and also show that the solution structure of the substrate-free protein is an equilibrium between substrate-bound and substrate-free forms [31] . The catalytic rates of two different forms of adenylate kinase are correlated with the timescales of motion in hinge and lid regions of the protein [32] . Similarly , allosteric effects upon ligand binding are observed in other nucleotide binding proteins . Solution NMR experiments on the nucleotide binding domain of the chaperone Hsp70 have revealed significant rotation in the subdomains when comparing the ATP and ADP bound states [33] . These changes impact the accessibility of a hydrophobic surface cleft distal from the nucleotide binding site , which is known to be essential for communication between the nucleotide binding domain and substrate binding domain . Allosteric effects are also observed upon cAMP binding to the exchange protein directly activated by cAMP ( EPAC ) , where long-range communication between the phosphate binding cassette and the N-terminal helical bundle is transmitted via two intramolecular pathways [34] . Interconversion between competing conformations is also critical in the function of kinase adaptor proteins , such as Crk where proline isomerization switches the protein between an autoinhibitory conformation and an uninhibited , active conformation [35] . In the SAXS studies for Csk , the extended conformations of the nucleotide-bound states arise primarily from rearrangement of the regulatory domains . The observed extension of the SH2-kinase linker and SH3-SH2 linker may allow the domains to more easily rearrange to both make catalytically functional interdomain contacts and interact with other binding partners ( discussed below ) . Communication between the SH2 and kinase domains is necessary for efficient catalysis as indicated by mutational studies of F183 , a critical hydrophobic residue in the SH2-kinase linker [17] . When F183 is mutated to a glycine this communication is broken and catalytic efficiency decreases nearly 1000-fold . Interdomain communication between the SH3 and kinase domains also aids in catalysis . Mutations in the SH3-SH2 linker decrease catalytic activity up to five-fold by disrupting the interaction surface between the domains characterized by NMR [36] . Kinetic experiments using a short peptide substrate revealed that Csk exists in both low and high activity forms [15] . Overall , the high activity form represents only a small fraction of the total Csk , based on ‘burst’ amplitudes in pre-steady-state kinetic experiments . Interestingly , the present study also indicates that much of the enzyme in solution adopts open forms that are not expected to reflect high-activity states . In particular , the small and large kinase lobes adopt open rotomers in most calculated species , which are not expected to position ATP for highly productive phosphoryl transfer . In contrast , a higher population of Csk transiently adopts the active form when using Src rather than a peptide substrate in pre-steady-state kinetic experiments . This implies that a physiological target with increased enzyme-substrate contacts may promote a more closed , enzymatically-competent kinase domain . However , regardless of these substrate-dependent changes in conformer distribution detected in kinetic experiments , turnover is limited by a slow conformational change in Csk when using Src as a substrate [14] . The large changes observed between AMP-PNP- and ADP-bound forms in the SAXS-SBM study may correspond to the rate-limiting changes observed in the pre-steady-state kinetic experiments for Src phosphorylation . If the AMP-PNP- and ADP-bound Csk complexes reflect population distributions before and after Src phosphorylation , then the SAXS studies may effectively capture important forms of the enzyme within the catalytic cycle . Thus , the conversion of the more open forms observed in the ADP-bound state into the more closed forms in the ATP-bound complex may be correlated with the observed slow conformational changes in Csk kinetic experiments . These finding are consistent with recent work on Abl kinase , where the ADP-bound kinase domain is more open and exhibits increased flexibility [37] . Csk is capable of interacting with a broad range of adaptor proteins through its SH2 domain [38] . These interactions likely require a high degree of adaptation and flexibility as Csk adopts a cytoplasmic location and then migrates to the plasma membrane in a phosphorylation-dependent manner [39] . This trafficking of Csk between the cytoplasm and membrane phospho-receptors may be facilitated by the regulatory SH2 and SH3 domains , which can transiently detach from the kinase core and become available for recognition at the plasma membrane . Diffusion of the SH2 domain is faster than the full protein , which may allow Csk to search for an SH2 domain binding partner more quickly . Such a scenario is reminiscent of the ‘fly-casting’ model for intrinsically disordered proteins , where unstructured regions of a protein may offer an increased capture radius for substrate recognition of specific targets [40] . Similarly , linker flexibility in Csk may allow the SH2 domain to effectively search for adaptors and then draw the kinase to the membrane for Src down-regulation . The SAXS-SBM results reported in this study provide new insights into these potential motions and how they may be coupled to both signaling and substrate processing in the cell .
The full-length Csk protein was expressed in Escherichia coli strain BL21 ( DE3 ) [41] , and purified by Ni2+ affinity chromatography [42] . The purified full-length enzyme was dialyzed against 50 mM Tris-HCl ( pH 8 . 0 ) , 150 mM NaCl , 5 mM DTT with 15% ( v/v ) glycerol and then concentrated to 107 µM and stored at −80°C . SAXS data were collected at 25°C at beamline 4-2 of the Stanford Synchrotron Radiation Laboratory . Scattering was independent of protein concentration ( 0 . 5 mg/mL–2 . 5 mg/mL ) indicating that interparticle ordering and aggregation are negligible . The AMP-PNP bound samples were generated through addition of AMP-PNP in a 50∶1 molar ratio to protein as well as the addition of 10 mM MgCl2 . Both AMP-PNP and MgCl2 were added from stocks prepared in Csk dialysis buffer . The ADP samples were prepared in the same manner . Data shown were obtained with 1 mg/mL protein for the apo and AMP-PNP bound proteins and 1 . 8 mg/mL for the ADP bound protein . A more detailed description of SAXS methodology for the study of proteins in solution may be found elsewhere [43] . The data were converted from a TIFF image to I ( q ) versus q using SasTool ( http://ssrl . slac . stanford . edu/~saxs/analysis/sastool . htm ) . q is the angular dependence of the scattering profile , which can be expressed as q = 4π ( sinθ/λ ) , where θ is half the scattering angle and λ is the wavelength of scattered X-rays . The radius of gyration , Rg , was determined by Guinier analysis . The goodness of the linear fit of the data in the low q range indicated no significant non-specific aggregation of the protein . We employed molecular dynamics simulations with an all-atom structure based model to generate an ensemble of candidate conformations ( i . e . structures that may explain the scattering profile ) . A detailed description of the all-atom model may be found elsewhere [44] . From this ensemble , theoretical scattering profiles were generated , and these profiles were compared to the experimentally measured data . We used an all-atom structure-based forcefield for the individual domains but only include steric interactions between domains . These models are based on the concepts of energy landscape theory [45] , [46] , [47] , have a low computational cost , and provide dynamic descriptions of proteins [48] , [49] , [50] , [51] and assemblies [52] that are in good agreement with experiments . The functional form of the all-atom structure-based forcefield isWherer0 , θ0 , ξ0 and φ0 are given the values found in the crystal structure ( PDB code: 1K9A ) [29] and σ = 2 . 5 Å , εr = 100/Å2 , εθ = 20/rad2 , εξ = 10/rad2 and εnc = 0 . 01 . Each native atom-atom contact interacts via a Leonard-Jones 12-6 interaction , where the energetic minimum corresponds to the native distance ( as defined by the crystal structure ) . A contact is defined as any atom pair that is ( a ) separated by less than 6 Å , is ( b ) separated by at least 4 residues in sequence , and ( c ) has no atom between them ( i . e . the “Shadow Algorithm”[53] ) . εnc and σnc define the excluded volume of each atom . Explicit representation of the atoms ensures that non-physical states , such as those with overlapping atoms or unrealistic bond angles , are disallowed . Contact and dihedral interactions were weighted as previously described [44] . In order to facilitate rapid sampling of all possible domain configurations , we removed all stabilizing inter-domain interactions and gave no configurational bias to the non-rigid dihedral angles ( i . e . dihedrals not restrained by orbital hybridization ) in the linkers . The forcefield files for Gromacs [54] were generated by an online resource ( http://smog . ucsd . edu ) [55] . A timestep of 0 . 0005 time units was used and the simulation was coupled to a temperature bath via Langevin dynamics . The total simulation time was 20 , 000 time units , which corresponds to approximately 100 µs [56] . Similar to previous studies [57] , [58] , [59] , we compare the candidate conformations from the simulation to the SAXS data by generating theoretical SAXS profiles for each conformation . Theoretical scattering curves were generated for all candidate conformations and ensembles using the CRYSOL software package [60] . Default parameters were used with the following exceptions: maximum order of harmonics = 50 , order of Fibonacci grid = 18 , number of points = 100 , and maximum s-value = 0 . 18 . Here , we measure the goodness of fit of each structure by calculating χ2 , which is defined as , where Nq is the number of data points in the scattering curve , I ( q ) is the SAXS intensity and σ ( q ) is the experimental error of Iexp ( q ) . Theoretical SAXS profiles were generated for linear combinations of conformations , where Ii ( q ) is the intensity profile of simulated conformation i and wi is the weight of conformation i . Icombined ( q ) was determined for all combinations of wi = 0 . 1n , where n is an integer of value 0 to 10 . χ2 was then calculated between each Icombined ( q ) and the experimental profile to determine the set of wi values that minimize χ2 .
|
The Src protein kinases are integral members of numerous signaling pathways involved in cellular growth and differentiation . The master regulator of the Src family is the protein kinase Csk , which adds a phosphate to the C-terminal tail , inhibiting Src Kinase function . Proper regulation of these signaling pathways by Csk is essential as unregulated activity in these pathways is correlated with the development of various cancers and autoimmune diseases . Understanding the nature of the mechanism and structure of Csk may lead to therapeutics and a better understanding of Src signaling pathways . Conformational changes associated with nucleotide binding and release have been shown to regulate the efficiency of Src down-regulation by Csk . To obtain insights into the nature of these nucleotide-induced structural changes , we examined the conformation of Csk in solution while bound to the ATP analog AMP-PNP and product ADP using a combination of small angle x-ray scattering and molecular dynamics . Surprisingly , both nucleotides induce extended conformations of Csk compared to the apo-enzyme , suggesting a novel mode of function . Further understanding of this mode of function may aid in the design of cancer therapeutics that act by regulating Src signaling pathways by modulating the function of Csk .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"biochemistry",
"enzyme",
"structure",
"protein",
"chemistry",
"cofactors",
"biochemistry",
"simulations",
"proteins",
"biophysic",
"al",
"simulations",
"enzymes",
"protein",
"structure",
"biology",
"computational",
"biology",
"biophysics",
"simulations",
"biophysics"
] |
2012
|
Substrate-Specific Reorganization of the Conformational Ensemble of CSK Implicates Novel Modes of Kinase Function
|
Malarial infection is associated with complex immune and erythropoietic responses in the host . A quantitative understanding of these processes is essential to help inform malaria therapy and for the design of effective vaccines . In this study , we use a statistical model-fitting approach to investigate the immune and erythropoietic responses in Plasmodium chabaudi infections of mice . Three mouse phenotypes ( wildtype , T-cell-deficient nude mice , and nude mice reconstituted with T-cells taken from wildtype mice ) were infected with one of two parasite clones ( AS or AJ ) . Under a Bayesian framework , we use an adaptive population-based Markov chain Monte Carlo method and fit a set of dynamical models to observed data on parasite and red blood cell ( RBC ) densities . Model fits are compared using Bayes' factors and parameter estimates obtained . We consider three independent immune mechanisms: clearance of parasitised RBCs ( pRBC ) , clearance of unparasitised RBCs ( uRBC ) , and clearance of parasites that burst from RBCs ( merozoites ) . Our results suggest that the immune response of wildtype mice is associated with less destruction of uRBCs , compared to the immune response of nude mice . There is a greater degree of synchronisation between pRBC and uRBC clearance than between either mechanism and merozoite clearance . In all three mouse phenotypes , control of the peak of parasite density is associated with pRBC clearance . In wildtype mice and AS-infected nude mice , control of the peak is also associated with uRBC clearance . Our results suggest that uRBC clearance , rather than RBC infection , is the major determinant of RBC dynamics from approximately day 12 post-innoculation . During the first 2–3 weeks of blood-stage infection , immune-mediated clearance of pRBCs and uRBCs appears to have a much stronger effect than immune-mediated merozoite clearance . Upregulation of erythropoiesis is dependent on mouse phenotype and is greater in wildtype and reconstitited mice . Our study highlights the informative power of statistically rigorous model-fitting techniques in elucidating biological systems .
Malarial infection of humans is a major cause of morbidity and mortality , continuing to cause around 250 million cases and close to a million deaths annually [1] . The vast majority of severe cases and deaths are due to Plasmodium falciparum , which is endemic in most of sub-Saharan Africa and other tropical areas [2] . Although there is no simple relationship between the pathogenic processes and clinical syndromes , disease only begins once the asexual parasite begins to multiply within the host's red blood cells ( RBCs ) [3] . The asexual dynamics depend on a complex interaction between the malaria parasite and the host's immune and erythropoetic responses [4] . Experimental methods have helped elucidate key aspects of this interaction . Such factors include the parasite's destruction of RBCs due to reproduction [5] , [6] , immune-mediated clearance of merozoites and parasitised RBCs ( pRBC ) [7] , [8] , and immune-mediated clearance of unparasitised RBCs ( uRBC ) . In particular , there is evidence that loss of uRBCs is responsible for the vast majority of the anaemia [9]–[11] . Suppression of RBC production ( dyserythropoiesis ) during the acute phase may also contribute to anaemia [12] , [13] , although recent modelling suggests that , overall , the level of erythropoiesis increases during malaria infection [8] , [9] , [14] . A full understanding of the infection dynamics requires quantitative analysis of the relative importance of the contributory factors [3] . Such an assessment is vital to help inform malaria treatment and intervention programmes [8] , [15] , [16] . In particular , the design of effective vaccines and immunotherapies depends largely on our understanding of the innate and adaptive immune responses [2] , [17] . In this context , rodent malaria models allow a highly replicable , highly controlled experiment . Although there are important differences between rodent and human malarias , a quantitative understanding of the rodent system , where we can control both host and parasite genetics , should help our understanding of the human case in which controlled experiments are unethical . As in other areas of science , mathematical models can be used to make inferences about complex dynamical systems by fitting them to data . This approach allows us to formally and quantitatively test and compare competing hypotheses , and to make quantitative predictions for future empirical testing . It is the most powerful and rapid way of culling possible , but incorrect , hypotheses . In the mathematical modelling literature on malaria , there are a number of studies that quantitatively fit models to data [6] , [8] , [9] , [14] , [15] , [18]–[22] . However , the poorly developed statistical , diagnostic and computational methodologies of fitting nonlinear dynamical models to noisy data ( see [23] and Discussion ) meant that these studies had to focus on particular aspects of the host-pathogen system in isolation . The method used was maximum likelihood . Its application to nonlinear systems is problematic because the nonlinearities create complex multi-dimensional likelihood surfaces . Search algorithms easily become trapped in local maxima , leading to false inferences [24] . Even if one is reasonably sure of having found the global maximum , evaluating parameter confidence intervals and covariances is computationally expensive and laborious , and computing predictive intervals practically impossible [25] . Recent developments in adaptive , population-based Markov chain Monte Carlo ( McMC ) methods overcome all of the problems associated with maximum likelihood [24] , [26]–[30] . The use of a Bayesian framework enables us to incorporate prior knowledge and uncertainty about the parameters . It allows us to quantify our relative belief in one model predicting the data over another , rather than accepting and rejecting models using conventional , but arbitrary , cut-offs . In order to use Bayesian statistics , we need to know the structure and variance of the measurement errors . Fortunately , these are known for our data sets . In this study we develop a set of models to test competing hypotheses describing the asexual stage of the malaria parasite . We fit the models to a set of data on Plasmodium chabaudi infections [31] using an adaptive McMC algorithm . We provide parameter estimates , examine differences between mouse and parasite strains , and make quantitative predictions about the immune and erythropoietic systems' dynamics , and their effects on the RBC population . In modelling the asexual dynamics , there are three general processes we need to consider: ( i ) the infection of RBCs , ( ii ) the immune response , and ( iii ) the response of the erythropoietic system to malaria-induced anaemia . The immune system's response to malaria is exceedingly complex and there is still much to learn about it qualitatively , let alone quantitatively [17] . Mathematical models have generally represented the immune response either as a single variable functionally linked to parasite density , or as separate innate and adaptive components [8] , [21] , [32]–[35] . The model of Recker et al . ( 2004 ) further discriminates , on the basis of human serologic data , between short-term , partially cross-reactive immune responses and long-term specific responses [36] . These models have given valuable insights into the immune dynamics , but it is important to acknowledge that the immune response consists of multiple arms , each targeting different aspects of the parasite [2] . Here we model the immune system as time-dependent immune-mediated clearance rates of merozoites , pRBCs and uRBCs . This allows us to bypass the debate about the highly interdependent innate and adaptive arms of the immune response , i . e . , when they are activated , what they target , and how they develop over time , and instead focus on the functional consequences in terms of the infection dynamics . We also draw attention to a key aspect of malaria asexual reproduction universally ignored in previous modelling studies . It is established that individual RBCs may be parasitised by more than one merozoite . Multiply-parasitised RBCs are often observed in experiments , but it is not known whether their subsequent behaviour is the same as that of singly-parasitised RBCs; previous models have generally assumed that their dynamics are identical . Here we test that assumption . In particular , we test whether multiply-parasitised RBCs have a greater death rate than other RBCs , and whether they produce a greater number of merozoites than singly-parasitised cells .
We used data obtained from a previous experiment [31] . Briefly , three different mice phenotypes were infected with either of two genetically distinct clones of Plasmodium chabaudi ( AS or AJ ) . Both clones were originally isolated from thicket rats ( Thamnomys rutilans ) in the Central African Republic [37] . The AS clone is associated with a lower peak density relative to AJ; it also has lower virulence , as measured by anaemia and weight loss [38] . Three different phenotypes of BALB/c mice were used: ( i ) wildtype mice; ( ii ) nu/nu mice ( “nude mice”; Harlan UK ) ; and ( iii ) nude mice reconstituted with T cells taken from wildtype mice . The mutation nu is a recessive mutation that blocks the development of the thymus . Nude mice therefore lack mature T cells , whereas heterozygotes ( nu/+ ) have a normal immune system [39] . Nude and reconstituted mice are smaller than the wildtype and are hairless . Mice of each phenotype ( wildtype , nude , reconstituted ) were innoculated with pRBCs of either AS or AJ . This gave six treatment groups . There were seven mice in both treatment groups for nude mice , and six mice in each treatment group for reconstituted and wildtype mice . Measurements of RBC and parasite density were taken on days 0 , 2 , 4 , and then daily until day 18 when the experiment was terminated . Parasite density was measured daily at 08:00 hrs using quantitative PCR , at which point the asexual merozoites have yet to replicate within pRBCs . Both RBC and parasite densities are expressed in terms of the number per microlitre ( ) of blood . Full details of the experimental methods are given in [31] . We removed a single data point from one of the reconstituted AJ-infected replicates . This mouse had much lower parasite density on day 14 than all the other mice . The data point was therefore considered to be an outlier . The averaged dynamics of each treatment are shown in Fig . 1 . The AJ clone ( solid line ) does not show the normal higher peak density compared to the AS clone ( dotted line ) ; however , the AJ clone exhibits a higher density during the exponential growth phase compared to AS . Parasite density tends to level off in reconstituted and nude mice from day 12 , but continues declining in the wildtype , presumably because of a stronger immune response in these mice . In reconstituted and wildtype mice , the AJ clone causes greater anaemia than the AS clone . All three mouse phenotypes show an earlier drop in RBC density from days 6–8 when infected with AJ compared to AS . The recovery of RBC density in nude mice is weaker than in reconstituted and wildtype mice . We discuss these observations below in relation to inferences from the model fitting . In Plasmodium chabaudi , pRBCs rupture synchronously every 24 hours , releasing on average 6–8 parasites ( merozoites ) into the bloodstream [40] . These newly released merozoites infect further RBCs and the cycle repeats . The rupture of pRBCS ( schizogony ) occurs at approximately midnight under normal lighting conditions [41] . We use a discrete-time formulation to model the dynamics , where each time step corresponds to a single day . The start of day is defined as the point immediately following rupture of pRBCs , before any infection has occurred ( i . e . , the point at which merozoites are released into the bloodstream ) . The densities of merozoites and uRBCs at the start of day are denoted and , respectively . We assume that the processes determining RBC density occur on two non-overlapping timescales . The first corresponds to the short infection phase during which merozoites infect RBCs , which occurs within a few minutes following schizogony . The second and subsequent timescale ( the remainder of the day ) corresponds to the RBC turnover phase: here the parasites replicate within pRBCs , new uRBCs are released into the bloodstream and , if active , the ( non-merozoite ) immune responses clear pRBCs and uRBCs . At the end of the RBC turnover phase , surviving pRBCs rupture and release new merozoites . Using the statistical algorithm described in the Supporting Information ( Text S1 , Figure S1 and Figure S2 ) , we fit to the observed data on RBC and parasite densities . The fitted parameters are given in Table 2 , along with their prior distributions . The prior distributions were based either on values taken from the literature , or approximate estimates obtained before the main model-fitting ( see Text S1 for details ) . Our aim is to find a set of minimal adequate models which explain the data well and contain as few parameters as possible . We take as our baseline the model described above . This is denoted and contains 20 fitted parameters . We developed a set of nested and non-nested models in which specific assumptions are made about the immune and erythropoietic responses ( outlined in Table 3 ) . All models were fitted to the data . Each model fit was then evaluated relative to that of the baseline . The model fits were compared using Bayes' factors , which naturally penalise overfitting [44] ( refer to Text S1 for further details on measurement error , the likelihood function , model fitting , assessment and comparison ) . We estimate parameters separately for each mouse , rather than across treatments . Even inbred mice are phenotypically different , and these differences result in variability in parasite and RBC dynamics . Immune responses are significant sources of variability in vivo; but we might also expect variation between mice in parameters such as infection rate , because of the multifactorial nature of such processes which involve the interaction of many host and parasite proteins . We therefore make no assumption about which parameters are invariant across mice and instead estimate parameters separately for each mouse . This method allows us to infer parameter ( and hence process ) variability within and between treatments from the posterior estimates .
We begin by analysing the baseline model , . The fits to all mice are shown in Figs . 2–4 . Note that the posterior predictive interval ( PPI ) of the dynamics widens from around day 15 for some mice because of the lack of data . The model fits appear to adequately explain the data . A more rigorous assessment of the model fits is attained by plotting the overlaid standardised residuals for parasite and RBC densities . Fig . 5 shows the standardised residuals ( the blue crosses ) for all mice across the six treatments . Poor fits are suggested by outlying and serially correlated residuals . The fits to parasite density are accurate but not perfect . There are no outliers , but the model tends to overestimate parasite density on day 4 . This is unexpected because we expect parasite density to be growing exponentially during this time , and indeed this is how the model behaves . This discrepancy suggests that parasite density may initially grow at a slower than exponential rate . Also , the model tends to underestimate parasite density on day 11 . Currently , we have no explanation for this . The fits to RBC density are accurate , except on day 8 where RBC density is marginally overestimated . Therefore , the model may not be correctly capturing the trough in RBC density . We calculated the Bayes' factors of models , for , relative to the baseline model . We adopt the scale of interpretation for Bayes' factors proposed by Jeffreys [45] and reproduced in Table 4 . For ease of interpretation , the Bayes' factors are converted to deciBans; i . e . , ( Bayes' factor ) . Bayes' factors were calculated for each mouse , giving a total of 38 values for each model comparison . The full list of Bayes' factors is given in Table S1 and Table S2 . For conciseness and clarity , we report for each model: ( i ) the sum of the deciBans for all mice within each treatment , and ( ii ) the sum of the deciBans for all mice across treatments ( Table 5 ) . We also report the standard error of the deciBans . Errors occur because deciBans are estimated from a finite sample of the posterior distribution . Our inferences are conservative; thus , we interpret a deciBan of as “barely worth mentioning” ( see Table 4 ) . Statistical comparison of parameter values between treatments was performed using Analysis of Variance ( ANOVA ) in the R statistical package [46] . The method was as follows . For a given parameter , we first took the mean of the posterior distribution , , for each individual mouse . The mean for a given treatment , , was then calculated as the average of the posterior mean values taken across all mice in the treatment .
The aim of this paper was to provide a quantitative assessment of the immune and erythropoietic responses in Plasmodium chabaudi infections . Hypotheses were drawn from experimental data and the mathematical modelling literature . These were fit to data on malaria infected mice using a Bayesian statistical framework . Crucially , by quantifying the experimental error , we were able to provide a rigorous assessment of the model fit . In particular , we were able to evaluate and compare the accuracy of different models in explaining the data . Models were compared using Bayes' factors , which impose a penalty for additional parameters . We interpreted our results with reference to the product of Bayes' factors ( sum of deciBans ) within and across treatments . Our results provide very strong evidence that the immune response to P . chabaudi involves clearance of both parasitised and unparasitised RBCs . Both effects were evident during the initial peak of parasite density , implying that control of the peak is regulated by the immune response . Previous modelling studies have suggested that innate or early specific immune responses regulate the initial dynamics of parasite density and anaemia [8] , [9] , [21] , [47] . Our study provides a statistically rigorous analysis in support of this hypothesis . Parasite-infected erythrocyte surface antigens ( PIESA ) may be an important immune target in both rodent and human malarias . In the case of P . falciparum , a longitudinal study of Kenyan children found that clinical malaria was caused by parasite isolates expressing PIESA variants that corresponded to gaps in the repertoire of antibodies carried by the children before they became ill [48] . We have shown that uRBC clearance by the immune system plays a key role in determining the infection dynamics of P . chabaudi . Experimental studies confirm that in the rodent malaria P . berghei [10] , and also in P . falciparum which infects humans [11] , the vast majority of anaemia is due to uRBC loss . Our results suggest that the level of anaemia following control of the initial peak ( from about day 12 ) , is mediated by the activity of the uRBC targeting response . This provides a mechanistic explanation for the variation in RBC dynamics between individual mice . If uRBC clearance decreases following control of the initial peak , then the increase in erythropoiesis that occurs from approximately day 10 allows the host to recover quickly from anaemia; in contrast , prolonged uRBC clearance is associated with a slow recovery from anaemia . Due to the lack of data to describe parasite density , the role of the pRBC targeting response during this later stage of infection is less certain . Our results suggest that the immune responses targeting pRBCs and uRBCs do not show a high degree of synchronisation . This implies they may be controlled by different effector mechanisms . Our model does not account for the antigenic variation that occurs in P . chabaudi infections [49] , [50] . Indeed , our model formulation only permits a single “switching-on” and “switching-off” of each immune component ( merozoite , pRBC , uRBC ) , and therefore does not distinguish between non-specific ( innate ) versus specific antibody responses to antigenically distinct variants . We modelled the infection up until day 18 , at which point antigenic variation may have a significant effect on the dynamics . In some mice , the data show a second drop in RBC density following the recovery from initial anaemia ( Fig . 2 ) . Although we only have data on parasite density up to day 14 , this second anaemia is commensurate with a recrudescent parasite density . However , it is significant that our model is able to explain the observed dynamics so well without including antigenic variation ( Fig . 5 ) . Extending the model up to , for example , day 30 post-infection would require explicit modelling of immune responses to the different antigenic variants . Such modelling would probably need to include both short-lived ( innate ) and long-lasting antibody responses , and may also need to consider cross-reactivity . The cascade of sequentially dominant antigenic variants seen in P . falciparum infections has recently been explained as the result of short-lived cross-reactive immune responses directed against shared epitopes [36] . Our results are consistent with the observation that T-cell-deficient ( “nude” ) mice have impaired immune responses , and are unable to resolve malaria infections [2] . T cells play an important role during the early stages of malarial infection , by amplifying the phagocytic and cell-mediated antiparasite responses; later in the infection , they help B cells to produce antibody , and assist in regulating the innate response [51]–[53] . Immune-mediated clearance of uRBCs was necessary to explain the dynamics in all three phenotypes , but nude mice had the higher maximum clearance rate of uRBCs . The infection rate of RBCs with merozoites , as reflected in the parameter , could also be higher in nude mice . As a simple proxy for the real biological system , these results indicate that nude mice are less able to limit the replication of the malaria parasite , and that their less specific immune response is associated with greater destruction of uninfected cells . At the individual mouse level , there was no evidence for immune-mediated clearance of merozoites . However , the cumulative evidence over all mice suggests that it is required to explain our data . Our interpretation of this result is that merozoite clearance is weak during the first few weeks of infection , and that pRBC and uRBC clearance are the major determinants of the dynamics during this time . Previous models have shown that , for a given level of immune activity , merozoite clearance is less effective at controlling parasite density compared to an equivalent response that clears pRBCs [7] , [8] . One explanation for this is that the duration of the merozoite infection phase ( estimated to be on the order of minutes ) is too short for the immune system to effectively target merozoites [7] . However , there is no a priori reason that one or several fast-acting , highly effective immune responses could not operate during this phase . There is considerable empirical evidence that merozoite surface protein one ( MSP1 ) is a target of immune mechanisms in malaria infections [54] . The presence of high levels of naturally acquired IgG antibodies to merozoite surface protein two ( MSP2 ) is also strongly associated with protection against clinical malaria [55] . Recent results suggest that this naturally acquired protection is not specific in relation to the major allelic dimorphisms of MSP2 [56] . We have shown that erythropoiesis upregulates during malarial infection , and that wildtype and reconstituted mice have higher upregulation than nude mice . This may reflect that the erythropoietic response only upregulates to the extent that the host is controlling the parasite . Recent theoretical results have shown that excessive upregulation of erythropoiesis facilitates the growth of the parasite , and may result in greater anaemia and a higher peak of parasite density [57] . We also investigated whether there is a time delay before the upregulation of erythropoiesis , and a lag in the feedback between RBC density and the level of erythropoiesis . The results for the AJ-infected wildtype mice suggest that upregulation of erythropoiesis occurs from day 10 . Both reconstituted and nude mice show no evidence of a time delay . Our results also suggest a time lag of 2–3 days in the feedback between RBC density and erythropoiesis in AJ-infected reconstituted mice; however there is no evidence for a time lag in the other treatments . The reasons for this are unclear . One possibility is that our putative time lag is compensating for another process not included in the model . Only analysis of other data sets may reveal what this may be . In summary , our results show that the immune system plays a key role in determining the RBC and parasite dynamics in malaria-infected mice . We have shown that immune-mediated clearance of both parasitised and unparasitised RBCs is necessary to explain the RBC and parasite dynamics . Previous models have examined the implications of RBC age structure for the infection dynamics [14] , [22] . In particular , recent work by Mideo et al . ( 2008 ) suggests that P . chabaudi may preferentially infect mature RBCs ( normocytes ) , but produce more merozoites in younger cells ( reticulocytes ) [14] . Future models may need to consider how RBC age structure and immune system dynamics can be combined to obtain a more complete picture of the asexual stage of malaria .
|
Malaria is a disease caused by a protozoan parasite of the genus Plasmodium . Every year there are around 250 million human cases of malaria , resulting in around a million deaths . Most of the severe cases and deaths are due to Plasmodium falciparum , which is endemic in much of sub-Saharan Africa and other tropical areas . The pathology of malaria is related to the asexual stage of the parasite . Understanding the infection dynamics during this stage is therefore essential to inform malaria treatment and vaccine design . Experimental infections of rodents represent an important first step towards understanding the more complicated human infections . We developed a series of models representing different hypotheses about the main processes regulating the infection dynamics during the asexual stage . Models were fit to data on Plasmodium chabaudi infections of mice , using a Bayesian statistical framework . The accuracy of different models in explaining the RBC and parasite densities was quantified . We identify the role of different types of immune-mediated mechanism , and show that RBC production ( erythropoiesis ) increases during infection . Differences between mouse phenotypes are explained . Our study highlights the informative power of model-fitting techniques in explaining biological systems .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"mathematics/statistics",
"public",
"health",
"and",
"epidemiology/infectious",
"diseases",
"immunology/immune",
"response",
"computational",
"biology"
] |
2010
|
Quantitative Analysis of Immune Response and Erythropoiesis during Rodent Malarial Infection
|
Homeostasis of most adult tissues is maintained by balancing stem cell self-renewal and differentiation , but whether post-transcriptional mechanisms can regulate this process is unknown . Here , we identify that an RNA methyltransferase ( Misu/Nsun2 ) is required to balance stem cell self-renewal and differentiation in skin . In the epidermis , this methyltransferase is found in a defined sub-population of hair follicle stem cells poised to undergo lineage commitment , and its depletion results in enhanced quiescence and aberrant stem cell differentiation . Our results reveal that post-transcriptional RNA methylation can play a previously unappreciated role in controlling stem cell fate .
Stem cells are defined by their ability to continuously maintain their population ( self-renewal ) while generating progeny ( differentiation ) . During self-renewal , stem cells have to avoid cell cycle exit and differentiation; however , when differentiating they have to evade uncontrolled proliferation . Thus , the question of how the balance between self-renewal and commitment is regulated is highly relevant to a fundamental understanding of cell differentiation and cancer . The hair follicle offers an excellent model to study stem cell fate , as it undergoes cyclic bouts of growth ( anagen ) , apoptosis-mediated regression ( catagen ) and rest ( telogen ) [1] . Multipotent hair follicle stem cells , located in a special microenvironment called the bulge , are slow cycling but exhibit long-term contribution to all hair compartments [2] , [3] . At the early onset of hair growth , single bulge cells migrate out of their niche in telogen and undergo proliferation as progenitors before they differentiate into hair [4] , [5] . Once a stem cell has left its niche , intrinsic and environmental cues converge to balance proliferation of progenitors with lineage-specific differentiation . For example , c-Myc is known to control the balance between stem cell expansion and differentiation [6] , [7] , [8] . When activated in epidermal stem cells , Myc triggers their exit from the stem cell compartment , induces proliferation of progenitors , and subsequently leads to lineage-specific differentiation [9] , [10] , [11] . Because the nucleolar protein Misu/NSun2 ( Myc-induced SUN-domain-containing protein ) is direct target gene of c-Myc [12] , we considered the possibility that its RNA methyltransferase activity may represent a novel mechanism to regulate stem cell fate . Misu catalyzes the formation of 5-methylcytidine ( m5C ) in tRNA and possibly other RNA species [12] , [13] , [14] . Whereas the function of ( cytosine-5 ) methylated DNA has been extensively analyzed [15] , the functional roles of methylated RNA are largely unknown [16] . To date , Misu ( NSun2 ) is one of only two identified m5C methylases with substrate specificity towards tRNA [13] , [17] , [18] , [19] . Here , we demonstrate that Misu is required for normal tissue homeostasis in vivo . Expression of Misu is up-regulated in the hair follicle bulge at the entry of anagen . Deletion of Misu prolongs stem cell quiescence leading to a delay in initiation of anagen . Thus , our data reveals a post-transcriptional modification as a novel mechanism used by stem cells to precise and temporally accurate balance self-renewal and differentiation .
To functionally analyse the role of Misu in vivo , we generated a reporter-tagged loss-of-function mouse model using an ES cell line carrying a Gene Trap in intron 8 of the NSUN2 gene ( GGTC-clone ID: D014D11 ) . The methyltransferase activity of Misu is mediated by the conserved SUN domain encoded by exon 2 to 12 ( Figure S1A; red box ) [19] . Insertion of the Gene Trap and fusion to the reporter β-galactosidase leads to disruption of the catalytic domain of Misu ( Figure S1A; blue box ) [20] . We confirmed deletion of Misu and presence of β-galactosidase by gene-specific PCR ( Figure S1B ) . Quantitative real-time PCR ( QPCR ) using a probe for NSun2 showed substantial reduction of full-length Misu in skin of heterozygous ( +/− ) and loss of expression in homozygous mice ( −/− ) ( Figure S1C ) . Western Blot analysis demonstrated that Misu −/− mice lacked Misu protein ( Figure S1D ) . Heterozygous mice for Misu deletion were viable and did not display an obvious phenotype ( Figure 1A ) . Homozygous mice were also viable , but were appreciably smaller than their wild type ( wt ) and heterozygous littermates ( Figure 1A ) . At three months of age , Misu −/− mice weighed around 30% less than controls ( Figure 1B ) . The smaller size of Misu −/− mice was confirmed by a second knock-out mouse model Nsun2tm1a ( EUCOMM ) Wtsi generated by the Wellcome Trust Sanger Institute ( http://www . sanger . ac . uk ) ( data not shown ) . Misu −/− males were sterile , and both genders had pelage that was comparable to controls after birth , but showed cyclic alopecia at around 10 months of age ( Figure 1C , 1D ) . Both , the yeast and human orthologue of Misu have previously been shown to catalyse the methylation of cytosine at position 34 ( C34 ) in intron-containing pre-tRNALeu ( CAA ) [13] , [19] . To investigate whether mouse Misu also methylated tRNALeu ( CAA ) , we used RNA bisulfite sequencing , which allows analysis of RNA methylation patterns in their native sequence context [21] . We find that intron-containing pre-tRNALeu ( CAA ) was methylated at C34 in wild-type mice but lacked the modification when Misu was deleted ( Figure 1E ) . We concluded that Misu was indispensable for methylation of tRNALeu ( CAA ) in skin . Based on our observation that older mice show signs of alopecia in the absence of Misu we speculated that Misu might have a functional role in maintaining skin homeostasis in the long-term . To determine the role of Misu in skin , we began by examining endogenous expression of Misu during embryogenesis and skin development by staining for LacZ ( Figure 2A–2Q ) . Misu-expression was detected from E3 . 5 in the inner cell mass of the blastocyst ( Figure 2A ) . After implantation and gastrulation , at E6 . 5 , Misu was observed throughout the extra-embryonic ectoderm ( Figure 2B , 2B′ ) , which gives rise to the nervous system and epidermis . Starting from E9 . 5 , expression of Misu became more restricted and at E13 . 5 and E14 . 5 Misu was enriched in developing whiskers ( arrow ) and eyes ( arrowhead ) ( Figure 2C–2F ) . From E15 . 5 , when the interfollicular epidermis ( IFE ) begins to stratify and follicular morphogenesis starts by forming hair placodes ( Figure 2G ) , highest expression of Misu was found in the suprabasal layer of IFE ( Figure 2G–2I; arrows ) . After birth , from postnatal day 1 ( P1 ) , expression of Misu in the IFE waned ( Figure 2J , 2K , 2N , arrows ) but increased in the matrix of growing ( anagen ) hair follicles ( Figure 2L , 2M , arrowhead ) . At the end of follicular morphogenesis , from around P14 to P19 , when hair follicles progress through the destructive ( catagen ) and resting ( telogen ) phase of the hair cycle , expression of Misu was absent ( data not shown and Figure 2N ) . At early anagen ( P23 ) , Misu was up-regulated in the hair germ ( HG; arrowhead ) and weakly expressed in the bulge ( Bu; arrow ) ( Figure 2O; insert ) . Throughout anagen , Misu was highly expressed in matrix but was also found in differentiated lineages of the hair follicles ( Figure 2P , 2Q , Figure S2 ) . We confirmed the dynamic expression pattern of Misu in skin by detecting endogenous RNA levels of Misu during morphogenesis ( M ) , catagen ( C ) , telogen ( T ) , and anagen ( A ) of the first postnatal hair cycle by QPCR ( Figure 2R ) . In conclusion , Misu was dynamically expressed during morphogenesis and in adult skin . In adult skin , expression of Misu was up-regulated in the bulge and hair germ as soon as the hair follicle entered its growing phase ( anagen ) . During anagen , cells of the hair germ give rise to the hair matrix , which showed highest expression of Misu . Matrix cells , which are often referred as to transit amplifying ( TA ) cells because they only survive through anagen [22] , are stem cell progenitors that divide a finite number of times until they become differentiated . Misu was originally identified as a transcriptional target of c-Myc in skin [12] . However , activation of c-Myc triggers epidermal stem cells to differentiate mainly into lineages of IFE and sebaceous glands [7] . Since a role for c-Myc in regulating bulge stem cells has yet not been identified and c-Myc expression levels remained unchanged during the hair cycle ( Figure 3A ) , we asked whether expression of Misu might also be regulated by hair-specific transcription factors . We examined the mouse Misu proximal promoter and detected a putative Lef1-binding motif ( Figure 3H ) . As described for Misu , also expression of Lef1 increased when hair follicles entered anagen ( Figure 2R; Figure 3B , 3C ) and expression of Misu and Lef1 overlapped in hair follicles at both early and late stages of anagen ( Figure 3F , 3G ) . To validate Misu as a target gene of Lef1 , we performed chromatin immunoprecipitation ( ChIP ) in mouse epidermis in anagen using an antibody for Lef1 ( Figure 3I ) . We detected Lef1-binding to the promoters of Misu and Msx2 , a known target gene of Lef1 [23] . We further confirmed Misu as a transcriptional target of the Lef1/β-catenin complex by luciferase assays using the Misu promoter ( pMisu ) and full-length Lef1 or ΔN63Lef1 ( ΔLef1 ) , a mutant construct lacking the β-catenin binding motif . The assays were performed in presence or as an additional negative control absence of β-Catenin ( Figure 3J ) . The Lef7 synthetic promoter ( pLef7 ) served as positive control ( Figure 3J ) . Luciferase activity using the Misu promoter increased around two-fold when the Lef1/β-catenin complex was present compared to the controls ( Figure 3J ) . Finally , we confirmed in vivo that Misu RNA levels decreased when ΔLef1 was over-expressed in the mouse epidermis ( K14ΔLef1 ) ( Figure 3K ) . We concluded that expression of Misu is up-regulated by Lef1 when hair follicles enter anagen . The complete lack of expression of Misu in adult skin in both the IFE and the bulge in either the catagen or telogen phase of the hair cycle excludes its expression in quiescent stem cells . However , Misu-expression , detected by LacZ staining , was up-regulated in the bulge region ( arrows ) and the hair germ ( HG ) as early as telogen to anagen transition ( Figure 4A , 4B; Figure S3A , S3B ) . We confirmed expression of Misu protein in the bulge ( arrows ) and the hair germ ( arrowheads ) in anagen by immunoflourescence using an antibody against mouse Misu ( Figure 4C–4E; Figure S3C ) . We next asked whether Misu-positive cells in the bulge were indeed stem cells . Hair follicle stem cells can be isolated by fluorescence activated cell sorting ( FACS ) , based on high expression of CD34 and α6 integrin ( Itgα6 ) ( Figure 4F–4H ) [24] . Progenitor cells of the hair germ are characterized by high expression of P-cadherin and low expression of Itgα6 ( Figure 4I–4K; Figure S4A–S4C ) [5] . Both , stem and progenitor populations were sorted from Misu +/− mice at the onset of anagen ( P21 ) and tested for expression of Misu based on β-galactosidase activity ( LacZ ) using fluorescein di-galactoside ( FDG ) ( Figure 4F–4K ) . Around 12% of bulge stem cells ( Itgα6high/CD34+ve ) and 17% of progenitor cells in the hair germ ( Itgα6low/P-cadherinhigh ) expressed Misu ( Figure 4F–4K ) . No signal for FDG was detected in keratinocytes from wild-type mice ( Figure 4G , 4J ) . To further confirm co-expression of Misu with stem cells markers , we isolated FDG+ve and FDG−ve keratinocytes from Misu +/− mice at P21 ( Figure S5A , S5B ) . QPCR analysis demonstrated that the stem cell markers CD34 , NFATc1 , and Lgr5 were enriched in Misu-expressing cells ( FDG+ve ) ( Figure 4L–4O ) . Both populations expressed Itgα6 at similar levels ( Figure S5C ) but FDG+ve cells were also enriched for the hair germ markers Lef1 , Wnt5a , and Sox6 ( Figure 4P; Figure S5D , S5E ) . Expression of differentiation markers was comparable or decreased compared to FDG−ve cells ( Figure S5F , S5G ) . We concluded that Misu was expressed in both bulge stem cells and cells of the hair germ at initiation of anagen . To test whether Misu might induce bulge stem cells to enter cell cycle at the transition of telogen to anagen , we FACS-sorted bulge stem cells , based on high expression of CD34 and Itgα6 ( Figure S6A ) and early progenitor cells of the hair germ , based on high expression of P-cadherin and low expression of Itgα6 ( Figure S6B ) at the onset of anagen at P21 in wild-type and Misu −/− mice . We then determined the cell cycle profile for all sorted populations ( Figure 5A–5I ) . We did not observe any difference in cell cycle profile when the whole epidermal population was analyzed , and as expected most of the cells were in G1 ( Figure 5B , 5C ) . However , the percentage of Misu −/− bulge stem cells ( Itgα6high/CD34+ve ) in S- and G2/M-phase of the cell cycle was significantly reduced compared to wild-type cells at anagen initiation ( Figure 5D , 5E ) . In contrast , the cell cycle profile of progenitor cells ( Itgα6low/P-cadherinhigh ) in Misu −/− and wild-type epidermis was comparable ( Figure 5F , 5G ) . At later stages , in anagen at P24 , the percentage of Misu −/− bulge stem cell population dividing was comparable to that of wild-type controls ( Figure 5H , 5I; Figure S6C ) . These data indicated that Misu was important to stimulate cell cycle entry of bulge stem cells at initiation of anagen and depletion of Misu might increase the quiescent phase of bulge stem cells . To test our hypothesis that lack of Misu resulted in increased number of quiescent bulge stem cells , we labelled Misu −/− and wild-type mice with BrdU , and detected label-retaining cells ( LRC ) after a chase period of four months ( Figure 6A–6D ) . A long chase period allows detecting differences not only in the number of LRC but also in the intensity of the BrdU-label , which correlates with number of divisions . The number of LRC in Misu-depleted tail hair follicles was significantly increased in the outer follicles of triplets ( Figure 6A–6C ) , which are known to go through the hair cycle concurrently , whereas the central follicle cycles asynchronously and usually has fewer LRC ( data not shown ) [25] , [26] , [27] . Additionally , the intensity of the BrdU-label was higher in Misu −/− skin suggesting that those cells divided slower than in their wt controls ( Figure 6D ) . Flow cytometry for CD34 and Itgα6 in anagen ( P30 ) , catagen ( P40 ) and telogen ( P49 ) confirmed that loss of Misu resulted in a two-fold increase of bulge stem cells only in telogen of the hair cycle ( Figure 6E , 6F; Figure S7A ) . The number of hair germ cells in telogen ( P49 ) was unchanged ( Figure S7B ) . Strikingly , we found that the increase of bulge stem cells in Misu −/− epidermis was due to an enrichment of a distinct cell population with lower Itgα6 expression compared to wild-type skin ( Figure 6E; red line ) , which might represent those bulge cells that failed to enter the cell cycle at the initiation of anagen ( Figure 5D ) . If our hypothesis was correct CD34+ve but Itgα6low cells from Misu −/− mice should be more quiescent and therefore express higher levels of bulge stem cell markers than the comparable wild-type population . To test our hypothesis , we sorted bulge stem cells into two populations: L ( CD34+ve/Itgα6low ) and H ( CD34+ve/Itgα6high ) ( Figure 6G ) . We then analysed both populations for expression of bulge markers in wild-type and Misu −/− epidermis ( Figure 6H ) . We found that population L obtained from Misu −/− mice showed consistently higher expression of the stem cell markers FGF-18 , Itgα6 , Sox9 , CD34 and NFATc1 relative to wild-type mice ( Figure 6H; Figure S8E ) . Importantly , population L did not show increased expression of the hair germ markers Wnt5a and Sox6 , or Lgr5 , a marker for cycling stem cells ( Figure 6H ) . In contrast , population H from Misu −/− mice expressed similar levels of stem cell markers than wild-type mice , but showed an increase in the expression of hair germ markers . Thus , our data indicated that depletion of Misu resulted in an accumulation of quiescent CD34+ve/Itgα6low stem cells in the bulge . To exclude the possibility that the increase in the quiescent stem cell population in Misu −/− mice was due to the general lack of Misu rather than being tissue-autonomous to skin , we generated a conditional knockout mouse model for Misu in the epidermis ( Methods ) ( Figure S8A–S8C ) . Like in Misu −/− animals , also in mice with conditionally deleted Misu in the basal , undifferentiated layers of the epidermis ( K14MisuΔ/Δ ) , a distinct cell population of stem cells with lower Itgα6 expression accumulated in the bulge ( Figure S8D; red line in right hand panel ) and expressed higher levels of stem cell markers than the Cre-negative Misuf/f controls ( Figure S8E ) . In vivo lineage tracing experiments suggest that a sub-population of bulge cells migrate into the hair germ in telogen to then undergo cell division at the onset of anagen [4] , [28] . We speculated that stem cells in the bulge of Misu −/− mice might fail to commit in the bulge and do not migrate into the hair germ . To determine whether lack of Misu affected stem cell migration into the hair germ , we labelled Misu −/− new born mice and control littermates with BrdU at late morphogenesis and chased them for one hair cycle until the second postnatal telogen ( P47 ) We then detected quiescent bulge stem cells using an antibody to BrdU and compared the number and location of label-retaining cells ( LRC ) in the upper ( high ) and lower bulge region and the hair germ ( Figure 6I–6M; Figure S9A–S9F ) [4] . Wild-type and Misu −/− epidermal cells incorporated BrdU at the same rates ( Figure S9A , S9B ) . After the chase period at P47 , we found a significant higher number of fully labelled LRC in the high bulge area of Misu −/− epidermis ( Figure 6J–6L ) . Moreover , the number of hair follicles without any LRC in the lower bulge or hair germ was three-fold higher in Misu −/− skin than wild-type littermates ( Figure 6M ) . Accordingly , we also found that the intensity of the BrdU-label increased in the higher bulge and decreased in the lower bulge and hair germ when Misu was depleted ( Figure S9C–S9F ) . Importantly , at this stage of the hair cycle the majority of bulge cells are yet not dividing , as shown by labelling for Ki67 ( Figure 6J , 6K ) , excluding the possibility that Misu −/− cells diluted the label by cell divisions . Our data showed that the increased quiescence of bulge stem cells in Misu −/− skin correlated with an accumulation of LRC in the upper part of the bulge , indicating that Misu −/− bulge stem cells are delayed in generating committed progenitor cells of the hair germ . To test whether Misu-deletion led to increased self-renewal capacity of stem cells in vitro , we measured the colony forming efficiency ( CFE ) of sorted Itgα6high/CD34+ve bulge stem cells ( Figure 7A ) . Out of all cell populations derived from skin , epidermal stem cells exhibit the highest CFE [29] . When seeded in clonal density , keratinocytes derived from Misu −/− mice formed more colonies than wild-type littermates ( Figure 7A ) , indicating that Misu −/− cells have a higher self-renewal capacity than wild-type cells . Similarly , unsorted keratinocytes obtained from Misu −/− epidermis showed higher CFE than control keratinocytes ( Figure 7B , 7C ) . Although expression of Misu was undetectable in adult IFE ( Figure 2N ) , cultured keratinocytes obtained from back skin expressed high levels of Misu RNA and protein ( Figure 7D; data not shown ) . Our data indicated so far , that loss of Misu increased accumulation of bulge stem cells in their niche at late telogen leading to an increased self-renewing but quiescent stem cell population . Therfore , we speculated that Misu −/− hair follicles should be delayed in entering anagen . Indeed , depletion of Misu led to a delay of entry into the first and second synchronized hair cycle in males ( Figure 7E–7L; Figure S10B , S10C ) . Compared to males , females exhibit a delayed hair cycle progression of around 2 days , yet even in Misu −/− females the percentage of hair follicles in the first anagen at P25 was lower than in their wild-type controls ( Figure S10A; Table S1 ) . Similarly , male mice with conditionally deleted Misu in the epidermis ( K14MisuΔ/Δ ) , displayed a delay in entering the first adult hair cycle compared to controls ( Misuf/f ) ( Figure 7M; Figure S10F–S10K ) . Later entry into anagen in Misu-depleted skin resulted in delayed differentiation of matrix cells ( Figure 7N–7U ) . The number of Lef1-postive cells , marking lineage committed hair follicle cells , was lower in matrix cells of Misu −/− mice ( Figure 7N , 7O; arrows ) [30] . Accordingly , expression of Dlx3 , Gata3 and BMP signalling , as determined by staining for phosphorylated Smad1/5/8 , were absent in Misu −/− anagen hair follicles ( Figure 7P–7U; arrows ) . Once Misu −/− mice entered anagen , the hair follicles were morphologically indistinguishable from those of wild-type mice ( Figure S10F , S10E , S10L–S10S ) . Our data suggested that Misu plays a role in accurately timing lineage commitment of hair follicle progenitor cells .
Here we show through generating general and skin-specific loss-of-function mouse models that the RNA methyltransferase Misu ( NSun2 ) defines expanding , committed progenitor populations in mammalian skin . Expression of Misu is absent in adult mouse interfollicular epidermis and the quiescent phases of the hair cycle ( catagen and telogen ) . As the hair follicle enters its growing phase ( anagen ) , Misu is expressed in the bulge and hair germ , both of which contain multipotent stem cells [3] , [31] , [32] . Cells in the hair germ give rise to transit amplifying ( TA ) cells in the hair matrix [4] , [5] . Matrix cells , which collectively show the highest expression of Misu of all cell types , subsequently differentiate into all hair lineages [3] . Uniquely at the transition of telogen to anagen , Misu is co-expressed with markers for both quiescent ( CD34+ve/Itgα6high ) and cycling ( Lgr5+ve ) stem cells from the bulge and the hair germ respectively . However , in contrast to those stem cell populations , Misu- expressing cells exhibit reduced self-renewal capacity . Thus , although Misu could reversibly commit stem cells to differentiate [33] , we clearly show that stem and committed progenitor cells co-exist within the bulge , the hair follicle stem cell niche . Our findings now raise the question of how only a few selected stem cells are activated during each hair cycle . Misu is required for cellular division of bulge cells only at the telogen-to-anagen transition , indicating that its function is temporarily and spatially controlled in a strict manner . One key pathway that drives epidermal stem cells from telogen into anagen and specifies hair follicle lineages is the canonical Wnt pathway . Activation of Wnt signaling by transient expression of N-terminally truncated β-catenin in the epidermis is sufficient to induce ectopic hair follicle formation [34] , [35] , [36] , [37] , [38] . Conversely , when the pathway is inhibited by β-catenin ablation or expressing of a ΔNLef1 mutant , hair follicle formation is impaired [30] , [39] , [40] , [41] . The identification of Misu as a direct downstream target of Lef1 further supports Misu as a key component during lineage commitment of bulge stem cells at the initiation of anagen . Thus , our data support a model in which stem and committed progenitors are distinct populations within the hair follicle that can be distinguished by their expression of Misu . Misu belongs to a large family of highly conserved methyltransferases , modifying cytosine-5 in RNA ( RNA:m5C-MTase ) [16] . Misu/NSun2 is the human orthologue of S . cerevisiae Trm4 , both of which have substrate specificity towards tRNA [12] , [13] , [19] . Like for human and yeast NSun2 , we confirm pre-tRNALeu ( CAA ) as a direct target substrate for mouse Misu in vivo . Although post-transcriptional methylation of tRNA at cytosine-5 is one of the most frequently encountered modifications , Dnmt2 and Misu are as yet the only identified tRNA:m5C MTases [13] , [17] , [18] , [42] . Although it has been recently shown that m5C methylation protects tRNA from cleavage and degradation , the biological function this may mediate remains unclear [17] , [43] . Similar to the depletion of Misu in mice , loss of Dnmt2 in zebrafish also results in reduced body size and impaired differentiation of specific tissues [44] . It remains , however , unclear why the RNA-methyltransferase activity of both proteins is critical in maintaining tissue homeostasis [44] . One possible mechanism of Misu's function is that methylated tRNA species could be directly involved in regulating stem cell differentiation . An alternative and intriguing possibility could be that methylation regulates the cleavage of tRNAs or intron-splincing of pre-tRNAs to generate products with microRNA-like features , which may offer an additional mechanism of post-transcriptional control [17] , [45] . Recent RNA deep-sequencing data identified a new set of small RNAs derived from tRNAs , including intron sequences , which are associated with Dicer and Argonaute proteins , strongly suggesting a role of these fragments in RNA silencing [46] , [47] . On the other hand , modified nucleotides in tRNA are well known to affect their structural and metabolic stability , and thus are likely to influence directly the rate or efficiency of protein translation [42] , [48] . Interestingly , protein translation is hierarchically controlled during stem cell self-renewal and differentiation; while parsimonious during self-renewal , it enhances during differentiation [49] . This increased protein synthesis capacity in stem cells could allow rapid elevation of translational rate in response to differentiation signals [49] . In summary , we demonstrate that the RNA methyltransferase Misu ( Nsun2 ) stimulates a sub-population of stem cells to leave the hair bulge and become committed progenitor cells in the hair germ . Thus we identify post-transcriptional RNA modifications as a novel mechanism by which stem cells control the balance between stem cell self-renewal and differentiation .
All mouse husbandry and experiments were carried out according to the local ethics committee under the terms of a UK Home Office license . Misu −/− mice were derived using 129S2 ES cell line carrying a Gene Trap in intron 8 of the NSUN2 gene ( GGTC-clone ID: D014D11 ) generated by the German Gene Trap Consortium , and then mated with C57Bl6/J CBA F1 mice . F1 progeny was subsequently inbred . A PCR-based strategy was developed to distinguish the wild-type and Misu gene trap alleles . The primers are as follows: SR2 , ( 5′GCC AAA CCT ACA GGT GGG GTC TTT ) and B34 ( 5′- TGT AAA ACG ACG GGA TCC GCC ) amplify a fragment 650 bp of the ß-geo cassette . The primers Misu-Int8-5′ ( 5′AGG TGG ACC TGA TCA TGG AG ) and Misu-Int8-3′ ( 5′-AGGG AGG GTC TGG AAA GATG ) amplify a fragment of 500 bp of the wild-type allele . Mice containing a floxed allele of the NSUN2 gene were obtained by first crossing Nsun2tm1a ( EUCOMM ) Wtsi mice , generated by the Wellcome Trust Sanger Institute , with transgenic mice expressing Flp recombinase to delete the LacZ-neo cassette [50] . The offspring , containing two LoxP sites flanking exon 6 , were then crossed with KRT14-cre mice ( The Jackson Laboratory ) . In KRT14-cre mice , Cre-recombinase is expressed under the control of the keratin14 ( KRT14 ) -promoter leading to deletion of Misu in all basal , undifferentiated cells of the epidermis ( K14MisuΔ/Δ ) . All mouse lines were bred to a mixed genetic background of CBA×C57BL/6J . Primers to identify Misuwt , Misuflox and MisuΔ alleles are F1 ( 5′CCC CCA CTG CTG CTC AAC G ) and R1 ( 5′CAA TGC CAC CAC AAC CTC CTT ) . Total of 97 Misu −/− mice with their wild-type controls and 29 K14MisuΔ/Δ with their Misuf/f control mice were genotyped and grouped by gender . Samples were taken from same regions from dorsal skin and process for H&E staining . Each mouse was classified into specific stages of the hair cycle based on established morphological guidelines [51] . The total numbers were represented as percentages . Bisulfite conversion of tRNA was carried out as previously described [21] . cDNA was PCR amplified using primers specific for the deaminated sequences of tRNA-998Leu ( CAA ) and tRNA-1911Leu ( CAA ) . tRNALeu ( CAA ) sequences were obtained from the Genomic tRNA Database ( http://gtrnadb . ucsc . edu/ ) . Primer sequences are as follows: Fw_Leu_De: 5′GAT GGT TGA GTG GTT TAA GGT GTT , Rv_Leu998_De: 5′CAC CTC CAA AAA AAA CCA AAA C and Rv_Leu1911_De: 5′CAC CTC CAT TCA AAA ACC AAA AC . DNA label-retaining cells ( LRC ) were generated by repeated BrdU ( Sigma ) injections of neonatal mice at P10 [25] . For LRC assays animals were chased for the times indicated . For tracing migration of bulge LRC into the hair germ , animals were sacrificed at P47 [4] . LRC were detected by BrdU immunostaining in tail skin whole mounts . Z-stack volumes of random areas of the slide were collected using a confocal microscope ( Leica SP5 ) . Maximum projected images were quantified using Volocity software ( PerkinElmer ) . Images were segmented to identify and measure intensity of BrdU-positive cells constrained to specific Regions Of Interest ( ROI ) defining the whole hair follicle ( bulge and hair germ ) , high and low bulge region and hair germ . Frequency distributions of BrdU intensity were calculated with Microsoft Office Excel 2007 software ( Microsoft ) . Immunostainings were performed on 10 µm paraffin sections . After citrate epitope retrieval , sections were permeabilized for 5 minutes with 0 . 2% Triton×100 at room temperature , blocked for 1 hour with 5% FCS and incubated overnight with the appropriate antibody dilution . Immunostaining on cryosections or cultured cells was performed as for paraffin after fixation for 10 minutes in 4% paraformaldehyde at room temperature . Tail epidermal whole mounts were prepared and immunolabelled as described previously [25] , [36] . For LacZ staining , whole mounts of embryos or freshly obtained skin samples were fixed for 30 minutes at room temperature in buffer containing 0 . 1 M phosphate buffer , 5 mM EGTA , 2 mM MgCl2 and 0 . 2% glutaraldehyde . Samples were then washed three times for 15 minutes each in wash buffer ( 2 mM MgCl2 and 0 . 1% Nonidet P40 in 0 . 1 M phosphate buffer ) and stained for 12 hours in a solution consisting of 1 mg/ml X-gal ( Melford ) , 5 mM K3Fe ( CN ) 6 and 5 mM K4Fe ( CN ) 6 in wash buffer . The skin samples were then embedded in paraffin , sectioned at 10 µm and stained with eosin or used for stainings . Primary antibodies were used at the following dilutions: rabbit monoclonal antibody to Ki67 ( 1∶100; SP6 , Vector Labs ) , rabbit polyclonal anti mouse keratin 14 ( 1∶2000; Covance ) , rabbit polyclonal anti mouse keratin 10 ( 1∶500; Covance ) , mouse monoclonal anti Gata3 ( 1∶50; HG3-31 , Santa Cruz Biotechnology ) , rabbit polyclonal anti Lef1 ( 1∶50; Cell Signaling Technology ) , mouse polyclonal anti Dlx3 ( 1∶200; Abnova ) , guinea pig polyclonal anti hair keratins 31 , 71 and 72 ( 1∶200; Progen ) , rabbit polyclonal anti keratin 6 ( 1∶5 . 000; Babco ) , rabbit polyclonal to phosphor-Smad1/5/8 ( 1∶50; Cell Signaling Technology ) , rabbit polyclonal CUK-1079-A antibody to mouse Misu ( 1∶1000; produced by Covalab ) , mouse monoclonal to keratin 15 ( 1∶1000 ) [25] ) , rat monoclonal anti BrdU ( 1∶100; Abcam ) , goat polyclonal anti P-cadherin ( 1∶100; R & D Systems ) , rat monoclonal anti α6 integrin ( 1∶500; GoH3 , AbD Serotec ) , Secondary antibodies ( Alexa Fluor 594- and 488-conjugated anti-rabbit , mouse , rat and guinea pig , Invitrogen ) were added at a dilution of 1∶500 for 1 hour at room temperature together with DAPI to label nuclei . White field images were acquired using an Olympus IX80 microscope and a DP50 camera . Confocal images were acquired on a Leica TCS SP5 confocal microscope . Z-stacks were acquired at 100 Hz with an optimal stack distance and 1024×1024 dpi resolutions . Z-stack projections were generated using the LAS AF software package ( Leica Microsystems ) . All the images were processed with Photoshop CS4 ( Adobe ) software . Proteins were extracted from cultured keratinocytes or total skin . 1 cm2 pieces of total back skin ( dermis and epidermis ) were snap-frozen in liquid N2 , transferred to lysis buffer ( 1% NP-40 , 200 mM NaCl , 25 mM Tris-HCl , pH 8 , 1 mM DTT ) including protease inhibitor cocktail ( Roche ) and homogenised for 30 seconds . Samples were incubated on ice for 20 minutes . Protein lysates were cleared by centrifugation at 13 , 000 rpm . Total protein concentration was quantified using Dc Protein Assay ( Bio-Rad ) . Equal amounts of protein were run in 7 . 5% polyacrylamide gels and blotted onto Hybond-P PVDF membranes ( GE Healthcare ) , which were incubated in TBST-blocking solution ( Tris-buffered saline , pH 8 . 8 , with 5% skimmed milk powder ) . Blots were incubated overnight at 4°C with primary antibodies , washed and incubated with the appropriate HRP-conjugated secondary antibodies ( GE Healthcare ) . α-Tubulin ( Sigma ) was used as a loading control . The chemiluminescent signal was detected using the ECL Plus Detection System ( GE Healthcare ) . Total RNA from mouse skin or cultured keratinocytes was prepared using Trizol reagent ( Invitrogen ) according to the manufacturer's instructions . Total RNA from flow-sorted cells was purified using Pure-Link RNA Micro Isolation Kit ( Invitrogen ) . Double-stranded cDNA was generated from 1 µg total RNA using Superscript III First-Strand Synthesis kit ( Invitrogen ) and random hexamer primers ( Promega ) . A minimum of two independent biological and three technical replicates was analysed . In case of Itgá6/CD34+ve-sorted cells one sample was pooled from four mice . For FDG-sorted cells one sample was pooled from one mouse . Real-time PCR amplification and analysis was conducted using the 7900HT Real-Time PCR System ( Applied Biosystems ) . The standard amplification protocol was used with pre-designed probe sets and TaqMan Fast Universal PCR Master Mix ( 2× ) ( Applied Biosystems ) . Probe set Mm00520224_m1 and Mm00487803_m1 were used to amplify mouse NSun2 ( Misu ) and c-Myc from total skin . The following probes were used to amplify selected genes from flow-sorted cells: α6 integrin ( Mm01333831_m1 ) , FGF-18 ( Mm00433286_m1 ) , CD34 ( Mm00519283_m1 ) , NFATc1 ( Mm00479445_m1 ) , Sox9 ( Mm00448840_m1 ) , Lgr5 ( Mm00438890_m1 ) , Sox6 ( Mm00488393_m1 ) , Wnt5a ( Mm00437347_m1 ) , Lef1 ( Mm00550265_m1 ) , Gata3 ( Mm01337569_m1 ) and Keratin 72 ( Mm00495207_m1 ) . GAPDH expression ( 4352932E ) was used to normalize samples using the ΔCt method . The −2068 to −48 bp DNA fragment of the mouse Misu promoter was cloned into the pGL3-Basic vector ( Promega ) . Plasmids for the reporter assays included pCDNA 3 . 1-hLef1-V5 ( Lef1 ) , pCDNA 3 . 1-ΔN63-hLef1-V5 ( ΔLef1 ) , and pCDNA 3 . 1-S33Y mCTNNB1 ( β-catenin ) . pLef7-fos-luc ( pLef7 ) was kindly provided by R . Grosschedl [52] . Hela cells were grown in DMEM ( Invitrogen ) supplemented with 10% fetal calf serum ( FCS ) in a humidified atmosphere at 37°C and 5% CO2 . Cells were transiently co-transfected with the promoter construct , pRL-TK renilla as an internal control and the indicated plasmids using Lipofectamine LTX transfection reagent ( Invitrogen ) . After recovery , cells were grown in media containing 0 . 2%FCS . Luciferase activity was measured 36–48 hours after the transfection using the Dual-Luciferase Reporter Assay System ( Promega ) on Glomax ( Promega ) . Each transfection was carried out in triplicate and the experiment was repeated twice . To isolate mouse keratinocytes from dorsal back skin we rinsed mouse back skin in 10% Betadine and 70% ethanol and washed it in PBS . The dermal side was thoroughly scraped to remove excess fat . The tissue was then floated on 0 . 25% Trypsin without EDTA ( Invitrogen ) for 2 hours at 37°C or overnight at 4°C . The epidermis was subsequently scraped from the dermis , minced using scalpels , disaggregated by gentle pipetting and filtered through a 70 ìm cell strainer . Trypsin was inactivated by addition of low-calcium medium with 10% FCS . The cells were pelleted and resuspended in the following antibodies: PE-conjugated Itgá6 ( clone GoH3 , eBiosciences ) , Alexa Fluor 647-conjugated CD34 ( RAM34 , eBiosciences ) and goat polyclonal anti-P-cadherin ( R & D Systems ) . After incubation for 45 minutes at 4°C , cells were washed twice in PBS . For detection of P-cadherin , cells were incubated for 10 minutes at 4°C with anti-goat Alexa Fluor 647-congugated secondary antibody ( Invitrogen ) . Cells were gated using forward versus side scatter to eliminate debris . Doublet discrimination was carried out using pulse width . The viable cells were then gated by their exclusion of DAPI using a 450/65 nm filter . Itgá6 PE stained cells were detected using a 580/30 nm filter and Alexa 647 cells were detected using a 670/30 nm filter . Cells were sorted with a MoFlo high-speed sorter ( Beckman Coulter ) . For cell cycle analysis of Itgá6high/CD34+ve or Itgá6low/P-cadherinhigh cell populations , cells were fixed with 1% paraformaldehyde for 5 minutes after immunolabelling , transferred to cold 70% ethanol and incubated for at least 1 hour before stained with propidium iodide ( PI ) or DAPI ( Sigma ) . After incubating the cells for 1 hour in RNase , analysis was carried out on a CyAN ADP analyzer ( Beckman Coulter ) . For detection of intracellular β-galactosidase activity , mouse keratinocytes from mice in early anagen ( P21 ) were loaded with fluorescein-di-β-D-galactopyranoside ( FDG ) ( Sigma ) using hypotonic shock . Briefly , equal volume of cells was mixed with warm 2× hypothonic shock solution ( 2 mM FDG in water ) and incubated for 30 seconds at 37°C , then cold media was added . Cells were washed and subsequently stained with the appropriate antibodies . Cells were then sorted based on FDG ( fluorescence detected using a 530/40 nm filter ) , after gating out dead cells ( based on DAPI staining ) . Mouse keratinocytes were isolated as described above and cultured on mitomycin-treated J2-3T3 feeder cells on collagen type I ( BD Biosciences ) coated plates ( BD Falcon ) . Mouse keratinocytes were grown in low-calcium FAD media ( one part Ham's F12 , three parts Dulbecco's modified Eagle's medium , 18 mM adenine and 0 . 05 mM calcium ) supplemented with 10% FCS and a cocktail of 0 . 5 µg/ml of hydrocortisone , 5 µg/ml insulin , 10−10 M cholera enterotoxin , and 10 ng/ml epidermal growth factor ( HICE cocktail ) and maintained in a humidified atmosphere at 32°C and 8% CO2 . Clonal growth was assayed by culturing 500 to 2500 Itgá6high/CD34+ve sorted cells or 5000 to 10000 viable epidermal cells per well in 6-well plates for 3 weeks . Cells were fixed and stained with 1% Rhodamine B . Three independent experiments were conducted . Mouse keratinocytes were isolated from mouse skin in anagen and processed according Chromatin immunoprecipitation Assay Kit ( Upstate ) . Chromatin was incubated with control anti-rabbit IgG and anti-Lef1 ( Cell Signaling ) antibody overnight at 4°C . The samples were eluted after washing . PCR reactions were performed by sets of specific primers: Misu TCT GTG CGG TCC TTT CTA CC ( forward ) and CGC GTC CTG CTA GCT ATG TT ( reverse ) ; Msx2 AAG GGA GAA AGG GTA GAG ( forward ) and CCC GCC TGA GAA TGT TGG ( reverse ) and GAPDH TAC TAG CGG TTT TAC GGG CG ( forward ) and TCG AAC AGG AGG AGC AGA GAG CGA ( reverse ) . The significance of quantitative data was tested using the unpaired , two-tailed Student's T test .
|
We demonstrate that the RNA methyltransferase activity of Misu/NSun2 is required for the proper maintenance of the epidermal differentiation program , and thus post-transcriptional mechanisms are involved in controlling the balance between stem cell self-renewal and differentiation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology",
"stem",
"cells",
"biology",
"adult",
"stem",
"cells"
] |
2011
|
The RNA–Methyltransferase Misu (NSun2) Poises Epidermal Stem Cells to Differentiate
|
A 30-node signed and directed network responsible for self-renewal and pluripotency of mouse embryonic stem cells ( mESCs ) was extracted from several ChIP-Seq and knockdown followed by expression prior studies . The underlying regulatory logic among network components was then learned using the initial network topology and single cell gene expression measurements from mESCs cultured in serum/LIF or serum-free 2i/LIF conditions . Comparing the learned network regulatory logic derived from cells cultured in serum/LIF vs . 2i/LIF revealed differential roles for Nanog , Oct4/Pou5f1 , Sox2 , Esrrb and Tcf3 . Overall , gene expression in the serum/LIF condition was more variable than in the 2i/LIF but mostly consistent across the two conditions . Expression levels for most genes in single cells were bimodal across the entire population and this motivated a Boolean modeling approach . In silico predictions derived from removal of nodes from the Boolean dynamical model were validated with experimental single and combinatorial RNA interference ( RNAi ) knockdowns of selected network components . Quantitative post-RNAi expression level measurements of remaining network components showed good agreement with the in silico predictions . Computational removal of nodes from the Boolean network model was also used to predict lineage specification outcomes . In summary , data integration , modeling , and targeted experiments were used to improve our understanding of the regulatory topology that controls mESC fate decisions as well as to develop robust directed lineage specification protocols .
mESCs are derived from the inner cell mass of a developing blastocyst and can be propagated indefinitely in culture . Cultured mESCs can contribute to all adult cell populations , including the germ-line . Human ESCs have similar in-vitro differentiation potential . It is now established that somatic cells can be reprogrammed into induced pluripotent stem cells ( iPSCs ) using simple combinations of transcription factors ( TFs ) [1]–[3] or other methods . Mouse and human iPSCs closely resemble ESCs , potentially removing ethical and tissue rejection barriers to applications in regenerative medicine . In order to harness the full potential of stem cell therapeutics there is a pressing need to further characterize the regulatory topology that controls pluripotency as well as commitment and differentiation to specific lineages . Pluripotency is maintained by a densely interconnected network of auto- and cross-regulatory TFs and other transcription regulators . These TFs and regulators promote the expression of other pluripotency genes and simultaneously suppress the expression of differentiation inducers [4] . To dissect the ESC regulatory topology , genome-wide high-throughput technologies such as cDNA microarrays , RNA-seq , ChIP-seq , immuno-precipitation followed by mass spectrometry ( IP-MS ) proteomics and phosphoproteomics , inhibitory RNA ( RNAi ) screens , as well as other emerging technologies have been applied . However , it remains a challenge to integrate multiple datasets , obtained from distinct sources and molecular regulatory layers into a systems level view of ESC regulation . Such data integration is necessary in order to build reliable predictive regulatory models that would provide a global view of the entire system . While static network diagrams can provide snapshot views of the information processing that controls cell fate decisions , it is necessary to develop regulatory models that capture the dynamical behavior of key regulatory components over time . In recent years , several stem cell-centered dynamical models have been developed from low-throughput functional studies . Most models employed ordinary differential equations ( ODEs ) and simulate interactions among a small number of well-studied TFs [5]–[7] . For example , a stochastic ODE model that linked Nanog , Oct4/Pou5f1 and Sox2 to an osteoblast differentiation circuit comprised of three additional TFs showed that cells can jump from one state to another if enough noise is added to the system [8] . In general , predictions made from computational dynamical models of embryonic stem cell have not been extensively experimentally validated until very recently . In the past year , few other comprehensive studies that integrated various datasets and constructed larger models of the ESC regulatory circuitry have emerged [9]–[12] . For example , Dowell et al . [11] integrated gene expression , ChIP-seq , protein interactions , RNAi screens and epigenetics markers to build a Bayesian network model of pluripotency genes . Their main focus was comparing human and mouse ESCs and their networks models are static . One of the advantages of their approach is that the network was not determined a priori which allowed the discovery of novel self-renewal and pluripotency components . Dowell et al . also present a database that is similar to our ESCAPE database [13] called StemSite . Lee and Zhou [10] combined ChIP-seq , gene expression and motif finding data to identify pairs of transcription factors that potentially work together within the pluripotency circuitry . They established 27 interactions between 14 factors . Many of the interactions they identified are consistent with our study . In another similar study Dunn et al . [9] implemented a data constrained Boolean model that connected 12 transcription factors through 16 interactions to suggest the minimal possible circuitry required to maintain pluripotency of mESC . In contrast with these studies , our study has primary gene expression data from single cell mESC , we also consider lineage markers and predict lineage propensity after perturbations , and provide extensive experimental validation with double and triple knockdowns followed by gene expression profiling of both pluripotency regulators and lineage markers . In order to address the need for multi-level data integration and broader experimentally validated dynamical models , we first extracted a signed and directed network from published ChIP-seq and knockdown followed by expression studies . All included 15 pluripotency network nodes and their interactions are supported by ChIP-seq data providing TF/target-gene binding evidence , as well as significant mRNA expression change following depletion or over-expression of a given TF . The ChIP-seq TF/target-gene binding evidence provides the directionality of the edges; whereas the knockdown or over-expression followed by genome-wide expression evidence establishes positive or negative regulatory edge sign . The underlying network regulatory logic was then learned using single-cell gene expression data collected with a microfluidic device . The learned model was validated by comparing predictions from in silico single or combinatorial node knockdowns to experimental single or combinatorial RNAi knockdowns of selected nodes followed by quantitative PCR expression level measurements of the remaining network components . Finally , lineage specification outcomes of single and combinatorial perturbations were predicted for all possible knockdown combinations .
We first extracted a signed and directed pluripotency network of mESCs consisting of 15 transcription regulators and 15 lineage markers from the ESCAPE database ( Figure 1 ) [14] . The ESCAPE database contains TF/target-gene interactions extracted from a collection of ChIP-seq studies applied to mESCs . In addition , the ESCAPE database also contains regulatory causal interactions connecting perturbed TFs to affected target genes based on mRNA expression . Such interactions were extracted from the loss-of-function ( LOF ) knockdown/knockout of TFs , or gain-of-function ( GOF ) over-expression followed by global transcriptional profiling measured by microarrays or RNA-seq database tables in ESCAPE . The 15 pluripotency regulators were selected only if both ChIP-seq and LOF/GOF evidence was available; whereas the selection of the 15 lineage markers was determined by expert curation . The network nodes represent 15 of the mostly well-studied pluripotency TFs and 15 of the most established early differentiation lineage markers . The selected factors also have phenotypic evidence that is important for sustaining normal mESC functions based on cell phenotype profiling after knockdown or over-expression . In addition most of the 15 selected pluripotency factors are well studied where some also were shown to play central role in iPSC reprogramming . The 15 lineage markers were selected to be the most established markers with having a relatively even representation for each lineage . The list of studies and the criteria used for inclusion of network nodes are described in Table S1 . The initial network created from the high-content studies had few links with conflicting signs . These conflicting signed links were resolved by citing specific publications from the literature ( Table S2 ) . The final network contains 30 nodes , 106 links , 10 positive auto-regulatory feedback loops , as well as 26 positive and 13 negative other feedback loops ( Figure S1 ) . The 30-node pluripotency network described above can suggest some novel regulatory mechanisms governing mESCs regulation . However , the directed and signed graph representation still lacks crucial regulatory information . Specifically , the transition functions ( logic gates ) that tie the activity of each node to the activities of its upstream regulators are not specified . In other words , the representation does not provide information on how the combinatorial state of upstream parent nodes determines the expression and activity of a given downstream node . The coding of transcriptional regulation to Boolean logic gates is a mathematical idealization and abstraction of the complex biochemical processes of transcriptional regulations . As such , a Boolean model can capture the essence of the regulatory relationships but may still lose important details . In addition , the network was created from data collected in different laboratories , under different experimental conditions and over different time scales . Moreover , all data were from bulk populations of mESCs and there is evidence for cell-to-cell heterogeneity in mESCs [15] , [16] . Hence , some links may be inaccurate , missing , or present in different contexts in individual cells , or in certain subpopulations . We therefore attempted to refine the network topology by learning the transition functions regulatory logic , and as a result enhancing the information about the network links based on gene expression measurements in single mESCs . Expression levels were measured using the Fluidigm microfluidic quantitative RT-PCR platform for single cell gene expression profiling [17] . In those experiments the expression of 96 genes was measured in 96 individual cells , including the 30 genes/nodes in the network assembled from ESCAPE ( Figures 2 , 3 , and S2 ) . The remaining 66 genes measured in single cells represent early differentiation markers and controls ( Table S3 ) . Measuring 96 individual genes in 96 single cells is a much more direct method to learn the regulatory logic of the pluripotency network . Analysis of data collected from a combined bulk of cells across a set of samples is less direct than single cell data because the regulatory relationships are masked by population averages . Two culture conditions were employed , +serum/LIF ( serum/LIF ) and –serum/+2i/LIF ( 2i/LIF ) . Serum can be a source of variability and with serum-free 2i media , two pharmacological inhibitors targeting the kinases MEK and GSK3β are sufficient to maintain pluripotency [18] . Both conditions benefit from the inclusion of LIF . However , it is still not completely clear how such variable conditions alter the connectivity of the core transcriptional network that maintains pluripotency . The various conditions when compared can give us clues about the required core circuitry that is needed to maintain pluripotency [19] , as well as the subtle differences that are expected to exist between conditions , particularly at the single cell level . Heterogeneous gene expression could arise from stochastic fluctuations or from natural global fluctuations in the pluripotent state [20] . As expected , a comparison of gene expression heterogeneity using indicator dispersion indices revealed less heterogeneity in the 2i/LIF condition as compared to serum/LIF ( p<10−5 , one-tailed paired t-test , Figure S2 ) . This can be seen in the generally narrower distributions for most mRNAs in the 2i/LIF condition ( Figure 2 ) . Our results show that Nanog and Zfp42/Rex1 expression is more homogeneous in the 2i/LIF condition compared with the serum/LIF condition , consistent with a previous study [21] . More globally , with the exception of Tbx3 , Fgf5 and Otx2 ( highlighted in bold and blue in Figure S2 ) , gene expression patterns are generally similar in the two conditions in terms of expression distributions ( Figure 2 ) . In agreement with previous observations , the 2i/LIF media stimulates the Wnt pathway , known to regulate the expression of Otx2 and Fgf5 [22] , [23] . Also , certain lineage-specific genes , namely Fgfr2 , Gata4 , T , Gata6 , Hand1 and Ncam1 display overall lower expression levels in 2i/LIF ( Table S5 ) . To address the possibility that these observations may be unique to the type of mESCs we used in our experiments , we compared the Esrrb expression level distribution to data from another study [24] ( Figure S2 and Text S1 ) . In both cases , Esrrb mRNA levels are bi-modal and similarly distributed in single cell populations . In addition , we demonstrated previously that Esrrb protein levels are comparable in the mESCs we used and wild-type cells [25] . Since the distribution of Esrrb mRNA expression is similar in the Esrrb removed ( Esrrb_R ) , CCE and Nanog removed ( Nanog_R ) mESCs using the same Fluidigm BioMark platform , we believe that Esrrb levels are well controlled in the Esrrb_R cell line and the results can be generalized to wild-type mESCs . Importantly , single cell expression profiles show bimodal distributions for numerous genes ( Figures 2 and S2 ) . Therefore , we reasoned that a Boolean framework for learning the underlying network regulatory logic by assigning values of 1 or 0 to high or low expression states , respectively , was a valid initial approach . Boolean modeling of gene-regulatory networks had first been proposed in the late 1960s [26] . Recently this approach regained popularity with the availability of more detailed systems level experimental data and networks [27] . In order to implement a Boolean representation , continuous Ct-value mRNA expression in single cells were converted to binary values ( 1 or 0 ) using a single K-means clustering step for all genes with K = 2 ( Figure 3A-B and Figure S2 ) . Delta Ct-values are used instead of the exponential of the delta Ct since this way it is more convenient for binarization . Hierarchical clustering of continuous ( Figure 3C–D ) and binarized Ct-values ( Figure S2 ) of the 30 network gene-products across the 96 cells show that there are no distinct single cell states but rather a level of variability across most cells . We actually expected to observe distinct subpopulations that represent few cell states . The bimodality in single cells has been observed before for single genes and the theory of having few distinct subpopulations is attractive from a modeling perspective . Yet the results confirm the bimodality of gene expression but not the idea of few distinct subpopulations within mESCs . Such variability may be associated with differentiation priming , perhaps to distinct lineages depending on the exact constellation of ON and OFF regulatory nodes in single cells . Priming toward different lineages is supported by the observation that lineage marker genes of the same lineage cluster together due to their high correlation coefficients of expression across the 96 single cells ( Figure 3E–F ) . For example , the ectoderm marker genes Rai1 , Ncam1 and Otx2 form a cluster and the trophectoderm marker genes Tead4 , Hand1 , Eomes and Cdx2 form another cluster . To learn the Boolean transition functions upstream of each network node , we applied an exhaustive symbolic search that is limited to the AND , OR and NOT logic operators and composition of these operators with the restriction of allowing each input to feed into only one of gate [28] . Based on the network topology extracted from the ESCAPE database and the single cell data measured with the microfluidic device , we attempted to fit all combinations of the three logic operators for each node given its parental inputs . The learning process used the single cell mRNA levels measured for each gene together with the original topology extracted from the ESCAPE database to derive a truth table for each Boolean transition function composed of AND , OR and NOT operators ( Figures 4 , S3 , and Table S4 ) . We assumed that single cell expression values reflect causal relationships between upstream regulators and downstream targets . Allowing only AND , OR and NOT gates , without nesting them , limits the search space to make the computation sufficiently efficient . However , such simplification may miss important biologically relevant Boolean functions . For example , exclusive OR ( XOR ) gates likely exist within mammalian cellular gene regulatory networks . In addition , threshold Boolean functions , which are Boolean functions that require at least several but not specific inputs to be active in order to turn on the target output are also missed by our symbolic search . The degeneracy of the possible functions listed in Table S4 suggests that such canalizing functions are likely present in the pluripotency circuits and finding them may reduce the degeneracy and complexity of the ensemble of dynamical Boolean models we obtained . Although there can be many Boolean functions that fit the same experimental observations , we attempted to find those that are most consistent with the direction and sign of edges of the network topology extracted from the ESCAPE database . For some input/output relationships many Boolean functions satisfied the input/output relationships ( Table S4 ) . For each gene , if no consistent function emerged with the initial topology , we performed a refinement process . The refinement process starts by testing all possible transition functions using a single regulator , sampling all possible regulators , then pairs and so on , until a defined threshold of input/output agreement was reached . For each gene , if no consistent function emerged , we performed the pruning refinement step by systematically removing all parental input links one-by-one and re-sampling all possible transition functions . This refinement and pruning procedure was executed recursively until a defined threshold of input/output agreement was reached . Specifically , if none of the combinations of plausible parents from the input network fulfilled the criteria , where is the number of single cell expression vectors , the algorithm exhaustively introduces single links from all the 15 pluripotency nodes . If still no single reassigned of a parent can explain the behavior of the downstream target by satisfying the criteria , we attempted all pairs of parents . For only four nodes from the serum/LIF learned network , and two nodes from the 2i/LIF learned network this process was necessary ( Table S4 ) . Auto-regulatory interactions are not considered in the dynamic Boolean models to keep the model dynamics simple . After learning , we obtained a large ensemble of Boolean network models where , in some cases , many distinct transition functions can satisfy the input/output relationships almost equally well . The average number of feedback loops after sampling 100 network models and without considering auto-regulatory loops , was 14 . 39 , including an average of 9 . 38 positive and 5 . 01 negative feedback loops ( Figure S3 ) . This number of feedback loops is significantly lower than the 39 feedback loops found in the initial network prior to learning the transition functions ( p-value <10−65 , t-test ) . The initial topology of the network created from the ESCAPE database has many more links than the pruned network , after fitting this network to be in concordance with the single cell data . The pruning and refinement step eliminates links iteratively , from the original topology of the initial network , with the goal of finding transition functions that are consistent with the original topology as well as with the single cell data . We next enumerated all feedback loops present in the sampled Boolean network models and ranked positive and negative loops based on their occurrence ( Figure S3 ) . The most common positive feedback loop in randomly selected models is the mutual activation of the Oct4/Pou5f1 gene by Nanog and the Nanog gene by the transcription factor Oct4/Pou5f1; whereas , the most common negative feedback loop is the activation of the gene Tcf3 by Oct4/Pou5f1 and the inhibition of the Oct4/Pou5f1 genes by the Tcf3 transcription factor ( Figure S3 ) . While it is established that 2i/LIF can replace serum/LIF to maintain the mESC ground pluripotency state [18] , detailed mechanisms responsible for culture-dependant similarities and differences remain elusive . Comparing the networks learned from both conditions , we found that the two networks re generally consistent with 21 out of 30 nodes having exactly the same connectivity . While the original topology of the serum/LIF and 2i/LIF networks is identical , during the pruning and refinement stage , some links can be removed , sign switched or added to obtain transition functions that are consistent with the single cell gene expression data . Therefore , the connectivity of the learned serum/LIF and 2i/LIF networks are slightly different from each other . We observed differences in the predicted regulation of the genes Klf4 , Tbx3 , Jarid2 , Fgf5 , Gata4 , Hand1 , Otx2 , Gli2 and Ptpn11 ( Figure 4 ) . In addition , the regulation by the key TF genes Oct4/Pou5f1 , Nanog , Sox2 , Esrrb and Tcf3 all showed some level of difference when comparing the two conditions ( Figure 4C ) . For example , Nanog appears to be a positive regulator of Klf4 and a negative regulator of Gata4 and Hand1 only in the 2i/LIF condition , while appearing as a negative regulator of Fgf5 under the serum/LIF condition . The family of learned Boolean functions confirmed known regulatory interactions and identified new ones . For example , Oct4/Pou5f1 is positively regulated by Sox2 and Nanog , which is known [29]–[31] . In turn , these links are reinforced by positive regulation of Sox2 and Nanog by Oct4/Pou5f1 . Furthermore , Esrrb was singled out as an activator of Klf4; whereas before learning , Nanog was a second potential Klf4 regulator . Direct activation of Klf4 by Esrrb may explain the ability of Esrrb to replace Klf4 for iPSC reprogramming [32] . Similarly , the repressive regulation of Nr0b1 ( Dax1 ) by Tcf3 and Sall4 was highlighted after refining the network topology with single cell data . Based on out-degree centrality ( direct targets per TF ) , Oct4/Pou5f1 emerges as the master regulator of the entire circuit ( Table S6 ) . This is not surprising since previous studies have shown that Oct4/Pou5f1 has the strongest effect on gene expression following its depletion in mESCs , and Oct4/Pou5f1 is the most critical factor for successful iPSC reprogramming [25] , [29] , [33] , [34] . Four new interactions were suggested from the learned model ( dark red links in Figure S3 ) . These are the inhibitory interactions from Oct4/Pou5f1 to Hand1 , Gata4 and Ptpn11 , and a negative link from Nr0b1 ( Dax1 ) to Jarid2 . Hand1 is associated with trophectodermal commitment . The inclusion of an inhibitory link from Oct4/Pou5f1 to Hand1 is consistent with the trophectodermal phenotype observed after Oct4/Pou5f1 depletion [35] . In addition , it has been reported that Oct4/Pou5f1 binds to Hand1 and Gata4 promoters in hESCs [36] . We confirmed the repressive effect of Oct4/Pou5f1 on Hand1 , Gata4 and Ptpn11 expression by depleting Oct4/Pou5f1 using two separate shRNAs followed by RT-PCR ( Figure S3 ) . In addition , motif analysis revealed potential Oct4/Pou5f1 binding sites within promoter regions of the three genes ( Figure S3 ) . While such data suggest direct regulation by Oct4/Pou5f1 for these genes , it is possible and still consistent with the model , that there are additional unidentified intermediates . Taken together , our results suggest that the learning framework can identify new regulatory relationships that can be experimentally validated . Although the network model topology does not consider protein-protein interactions , the learning process automatically captures functional relationships between multiple pluripotency TFs that are also likely to physically interact . The AND logic operator can indirectly suggest physical binding interactions . Among the 20 known protein-protein interactions that we have collected from prior studies that interconnect the 15 pluripotency regulators , 15 are supported by at least one AND operator ( Table S7A ) . This is a high over-representation compared to random assignments of logic gates ( p<0 . 001 , one-tailed Fisher exact test ) . Conversely , we rank the protein-protein pairs connected by an AND gate by their co-occurrence frequency in all learned Boolean functions , among the top 10 ranked protein-protein pairs , 6 were previously reported as interaction partners in large-scale interactome studies and 4 were reported in low-throughput studies ( Table S7B ) . Therefore , the optimized network with the learned logic functions is capable of predicting potential protein-protein interactions between TFs . Nevertheless , an AND gate does not necessitate a physical interaction between TFs . Since the top two ranked AND relationships were not supported by direct physical interaction studies , we decided to conduct a co-IP experiment to test one of these , namely the interaction between Nanog and Sox2 . Nuclear extract from a mESC line expressing an epitope-tagged Nanog was used to directly test this potential interaction ( see Text S1 and Figure S3 ) . The results were negative , suggesting that Nanog and Sox2 may not interact directly . This remains consistent with the model because an AND gate only requires that two or more TFs cooperate to regulate the same target genes by co-binding to promoter or enhancer regions but those factors do not have to physically interact . The refined 30-node network with the learned regulatory logic-gate relationships can be used for dynamical Boolean simulations . However , whether such simulations are predictive requires experimental validation . To this end , we first performed in silico and subsequently , experimental knockdowns of Esrrb , Oct4/Pou5f1 and Nanog individually and in all possible double and triple combinations ( Figure 5A ) . The learned ensemble of Boolean dynamical models was used to make predictions about the network response to perturbations . Many Boolean networks consisting of 30 genes/nodes and a set of Boolean functions were sampled randomly from all learned Boolean functions calibrated through the learning workflow . Computational simulations were achieved by forcing a node ( s ) into a stable OFF state . Starting with 100 random initial condition for each sampled network , step-wise simulations were performed and the resultant stable values for all network nodes ( steady-state attractors ) were obtained ( Figure 5B , Text S1 ) . Technically , simulations were performed using discrete Boolean dynamics with synchronous updating in 30 steps with 10 sampled networks , and 100 random initial conditions . In most cases , steady states were reached within 3 to 9 simulation steps . In the attractor space , on average 83% of the time there was a dominant attractor that is achieved from the 100 random initial conditions . We then performed the same knockdowns in mESCs and measured alkaline phosphatase activities ( Figure S4 ) as well as changes in mRNA and protein levels to verify knockdown efficiencies ( Figure S4 ) . Alkaline phosphatase ( AP ) is a pluripotency stem cell marker , whereas loss of AP activity , as determined by the AP assay , is used to access differentiation of mESCs . We also measured the expression levels of the 30 network genes/nodes using RT-PCR in bulk mESC population , with each experiment performed in duplicate ( Figure 5C ) . To compare the binary vectors from the in silico perturbations to continuous experimental Ct-values representing the mRNA levels in the cell population , we calculated post-knockdown gene expression changes relative to the unperturbed condition from both simulations and experiments ( Figure 5B–D ) . A logistic function ( see Text S1 ) was used to compute the consistency between the predicted and measured expression changes , including fold-change magnitudes ( Figure 5D , Table S8 ) . Overall , experimental measurements following single or combinatorial knockdowns showed significant agreement with the in silico predictions ( p-value <10−15 compared to a random predictor , see Text S1 for more details ) ( Figure 5D ) . All single and combinatorial in silico as well as experimental knockdowns repressed some but not all core pluripotency components and activated selective differentiation markers consistent with published experimental results [25] , [37] , [38] . In addition , in silico simulations showed that knocking down Oct4/Pou5f1 has the most significant effect , in agreement with previous experimental results [25] , [29] , [33] , [34] . However , our dynamical model is unsuccessful in predicting the correct activity values for Cdx2 , Fgf5 , Tcf3 and Zfp281 for various knockdown conditions . We adopted an alternative strategy to resolve these conflicts by utilizing all data sources to train the model: 1 ) the initial network topology; 2 ) the single cell gene expression measurements; and 3 ) the measurements of the network components after the various knockdowns . The re-learning process resulted in re-wired logic for Cdx2 and Fgf5 that resolved the conflicts for these two nodes ( Figure S4G–H ) . For Tcf3 and Zfp281 , the re-learning process did not improve the predictions . It is possible that additional upstream regulators are required to explain these discrepancies . For example , Tcf3 may be differentially regulated by the β-catenin/Wnt signaling pathway in mESCs . Alternatively , Tcf3 and Zfp281 could be differentially regulated in mESCs depleted of Oct4/Nanog/Esrrb . The inconsistent behavior of Tcf3 in the model versus the empirical observations in our experiments is intriguing . The model predicts that knockdowns of Oct4/Nanog/Esrrb would down-regulate Tcf3 . However , experimental knockdowns resulted in up-regulation of this factor . Increased Nanog and Oct4/Pou5f1 expression after Tcf3 depletion was previously reported [39] . Down-regulation of Tcf3 following depletion of Oct4/Pou5f1 or Nanog has also been demonstrated [40] . The latter result is consistent with our in-silico predictions while the former is not . Training the model with single cell data assumed concordance between mRNA levels and TF activities . For Tcf3 and Zfp281 this assumption may be incorrect . Indeed , in our previous studies after depletion of Nanog , we observed significant discordances between mRNA and encoded protein levels [41] . In addition to the Nanog , Oct4/Pou5f1 and Esrrb single and combinatorial knockdowns we also depleted Jarid2 and observed significant agreement between predicted and experimental results ( Figure S4 ) . Next , we asked if prediction accuracy is mostly the result of calibrating the initial network topology with the single cell data , or if it is already largely embedded in the topology extracted from the ESCAPE database , with the single cell data contributing only minor tuning . To address this question , we randomly and sequentially flipped binary single cell gene expression values or , similarly reassigned network links from ESCAPE ( see Text S1 ) . We demonstrate this methodology using a toy network with simulated data ( Figure S7 ) . Given any directed network and single cell gene expression data ( Figure S7 ) , the regulatory logic can be learned . Then , when in-silico knockdowns are performed , computational knockdowns can be compared with experimental knockdowns . Each entry in the comparison heatmap was calculated as 1 – discordance-score . Note that the score didn't reach 100 despite the ideal example , due to the logistic function used . Both shuffling single cell values , or links from the original topology of the network , reduced the predictive power of the model to approximately equal extents , demonstrating that both sources of data contribute to prediction accuracy ( Figure S4 ) . Because the predictions obtained from the dynamical model are generally reliable , we simulated all possible single or combinatorial knockdowns to predict mESC lineage commitment outcomes ( Figure 6 , Text S1 ) . The 15 lineage-specific markers in the network may be limited and biased because lineage specification involves more genes . Therefore , we linked the simulated knockdowns to a larger set of 40 lineage marker genes ( Text S1 ) . We compared predictions that are based on the resultant state of the network core components after simulated knockdowns , with predictions that are based solely on the additive effect of the direct targets of individual and combinations of TFs without simulations . Interestingly , predictions made through the Boolean modeling approach were more consistent with lineage commitment knowledge compared with the direct TF target-based predictions; for example , forcing Oct4/Pou5f1 into a stable OFF state in our dynamical model results in more pronounced predicted trophectodermal differentiation when compared with predictions that are based solely on Oct4/Pou5f1 target genes . Trophectoderm induction after Oct4/Pou5f1 depletion is well-established [42] . We also examined lineage marker gene levels after depleting Esrrb or Jarid2 . For Esrrb , simulation-based predictions were more consistent with experimental data showing greater effects on neuroectodermal than trophectodermal or mesodermal differentiation . In the case of Jarid2 , both simulation-based and direct targets lineage predictions were consistent with our experimental observations which point to primitive endoderm differentiation ( Figures 6 and S5 ) . Overall , the Boolean model appears to resolve indirect effects of TFs on lineage commitment . The model may be most useful for prioritizing combinations of knockdowns not easily tested in high-throughput . Finally , we constructed two additional dynamical models learned from two recently published single cell gene expression datasets collected from CCE and Nanog removed ( Nanog_R ) mESCs using the same Fluidigm platform in the same laboratory [43] . To compare the four dynamical models constructed from the various types of mESCs we counted unique and shared regulatory interactions ( Figure S6 ) . Interestingly , we identified 41 unique interactions present only in one model under one condition , and 13 consensus interactions shared by all four models and conditions . In addition , among all the learned interactions from the four models , there are 42 interactions that appear in 3 out of 4 models , and 62 interactions that are found in 2 out of 4 models . Hence , on average each model contains ∼10 unique interactions and ∼40 interactions shared by at least one another model . One of the consensus interactions is Nanog regulation of Zfp281 , reinforcing the importance of this interaction . Likewise , Oct4/Pou5f1 consistently positively regulates Zfp42/Rex1 and Sall4 . Intriguingly , the positive feedback loop between Oct4/Pou5f1 and Sox2 exists in all models in serum/LIF but not in 2i/LIF , suggestive of culture-specific regulatory interactions . Other differences can stem from cell-line differences , differences in the times the experiments were conducted , as well as instrument noise . Overall we consider the agreement among all 4 models to be relatively robust . For this study we first developed a directed and signed network consisting of 15 pluripotency and self-renewal regulators connected to 15 lineage marker genes . We then learned the underlying regulatory logic among network components utilizing single cell gene expression data . Gene expression in single cells uncovered some differences in mESCs cultured under serum/LIF versus 2i/LIF conditions . Characterizing such culture-dependent differences is important for understanding the enhanced iPSC reprogramming efficiency in the 2i/LIF condition [44] . Expression in single cells was found to be mostly bimodal and this fits well with a Boolean modeling framework . The utility of such Boolean modeling approach is the ability to perform predictions of network behavior after in silico perturbations , and the ability to control and constrain the free parameter space . The consistency such in silico perturbations were compared to experimental combinatorial shRNA perturbations . The good agreement between the model and the experimental validation suggest that the model does capture some of the real dynamics of the pluripotency and self-renewal circuit . Nevertheless , many challenges remain . Our model is binary , with genes and their products merged into single nodes assuming a direct correlation between TF activity and mRNA expression . We observed such correlations for Nanog , Oct4/Pou5f1 and Esrrb ( Figure S4 ) , but for other TFs this may not be the case . Furthermore , while there is ample evidence describing how pluripotency TFs regulate lineage-specific genes , little is known about lineage regulator-mediated suppression of the pluripotency circuit . In addition , gene expression is controlled by protein complexes and epigenetic modifications not explicitly incorporated into the Boolean model . Integration of TFs , histone modifications and DNA methylation may result in a more complex model but also such model will be more accurate and revealing [45] . Nevertheless , rapid progress in the field indicates that we will gradually be able to obtain more refined and dynamic view of pluripotency , self-renewal , lineage-specific commitment and differentiation , as well as better understand the process of iPSC reprogramming . Such views will enable the realization of pluripotent cell-based applications in regenerative medicine .
The Esrrb_R rescue cell line was derived from AINV-15 ESCs and cultured as previously described [25] in ESC media containing doxycycline ( 1 µg ml-1 Sigma ) . CCE mESCs are one of the first established stem cell lines [46] . They are derived from the 129/Sv mouse strain . All cells used in these experiments were under passage 80 . Serum-free ESC cultures were performed as previously described [47] , [48] . Briefly , cells were maintained without feeders in serum-free N2B27 media prepared as described [47] and supplemented with LIF ( Chemicon , Millipore ) and 2i inhibitors [18] . The two inhibitors ( Stemgent ) block GSK3β , ( CHIR99021; 3 µM ) and MEK1/2 ( PD0325901; 1 µM ) . All cultures were maintained at 37°C with 5% CO2 . Inventoried TaqMan assays ( 20× , Applied Biosystems ) were pooled to a final concentration of 0 . 2× for each of the 96 assays . Single Esrrb_R cells expressing both GFP and SSEA-1 were FACS-sorted directly into 10 µL RT-PreAmp Master Mix ( 5 . 0 µL CellsDirect 2× Reaction Mix , 2 . 5 µL 0 . 2× assay pool , 0 . 2 µL SuperScript III RT/Platinum Taq Mix from the ( CellsDirect One-Step qRT PCR Kits , Invitrogen ) and 1 . 3 µL TE buffer . Cell lysis and gene-specific reverse transcription were performed at 50°C for 20 min . Reverse transcriptase was heat-inactivated for 2 min at 95°C . Subsequently , single cell cDNA was pre-amplified using a multiplexed , target-specific amplification protocol ( denaturation at 95°C for 15 sec , and annealing and amplification at 60°C for 4 min for a total of 18 cycles ) . Pre-amplified products were diluted 5-fold prior to amplification using a Universal PCR Master Mix and inventoried TaqMan gene expression assays ( Applied Biosystems ) in 96×96 Dynamic Arrays on a BioMark System ( Fluidigm ) . Amplification included a 10 min , 95°C hot-start followed by 40 cycles of a two-step program consisting of 15 sec at 95°C and 60 sec at 60°C . Ct-values were calculated using BioMark Real-Time PCR Analysis Software v2 . 0 ( Fluidigm ) . Values greater than 35 were considered non-detectable and recorded as 35 . Gene-specific 19nt shRNAs were designed based on a previously described algorithm using an in-house Perl script [49] . All shRNA sequences were BLASTed to ensure specificity . Synthesized oligomers were annealed and ligated into the pSuper . puro vector ( Oligoengine ) . To make the Oct4/Nanog/Esrrb combinatorial shRNA constructs , ClaI-XhoI sites were used to insert H1-shRNA cassettes digested with BstBI-XhoI . The shRNA encoding sequences are: Oct4– GAAGGATGTGGTTCGAGTA ( shRNA_#1 ) and GCGAACTAGCATTGAGAAC ( shRNA_#2 ) , Nanog– GAACTATTCTTGCTTACAA , Esrrb– GATTCGATGTACATTGAGA and Jarid2 – TCACTGTCCTCCCAAATAA . Mouse CCE ESCs were cultured feeder-free on 0 . 1% gelatin-coated plates in ESC media ( Dulbecco's modified Eagle's medium ( DMEM; Hi-Glucose ) , 15% fetal bovine serum , non-essential amino acids , L-glutamine , β-mercaptoethanol , penicillin/streptomycin , sodium pyruvate and LIF ( Millipore ) . Serum was purchased from HyClone . This serum is embryonic stem cell qualified and therefore does not require heat inactivation . The specific lot of serum was rigorously tested to ensure robust self-renewal with little spontaneous differentiation as assessed by mESC morphologies and alkaline phosphatase staining . All cell cultures were maintained at 37°C with 5% CO2 . Gene-specific or scrambled shRNA constructs , the GFP-shRNA construct and empty vector ( all 3 ug ) were transfected using Lipofectamine 2000 ( Invitrogen ) according to the manufacturer's instructions . Transfected cells were selected for 48 hrs . in puromycin ( 1 . 5 ug/ml ) . Mock transfections resulted in no surviving cells after selection . Total RNA was Trizol-extracted ( Invitrogen ) , column-purified with RNeasy kits ( Qiagen ) , and reverse transcribed using the High Capacity reverse transcription kit ( Applied Biosystems ) . All quantitative PCR analyses were performed using the Fast SYBR Green Master Mix ( Applied Biosystems ) following the manufacturer's protocol on the LightCycler480 Real-Time PCR System ( Roche ) . Each PCR reaction generated a specific amplicon , as demonstrated by melting-temperature profiles ( Dissociation Curve analysis ) . No PCR products were observed in the absence of template . Data were normalized to Gapdh and represented relative to empty-vector transfected controls . Primer sequences are available in Table S9 . Cells were scraped/trypsinized , washed in PBS and incubated for 20 min in cold RIPA buffer without SDS . Protein concentrations were determined using Bradford Dye ( Bio-Rad ) . Total proteins ( ∼10 ug ) were separated on SDS–PAGE gels and transferred to PVDF membranes ( Millipore ) . Membranes were probed with specific primary antibodies followed by HRP-conjugated secondary antibodies and developed with ECL ( Amersham ) . Primary antibodies were: Oct4 ( sc9081 , Santa Cruz ) , Nanog ( A300-398A , Bethyl ) , Actin ( sc1615 , Santa Cruz ) , Sox2 ( sc17320 , Santa Cruz ) and Esrrb ( 300–748 , Novus Biologicals ) . Amido-black was used to detect core histones . Quantification of protein bands was performed using Adobe Photoshop . Relative protein level differences were calculated by normalization to actin levels and shown relative to empty-vector transfected sample . An alkaline phosphatase detection kit ( Stemgent ) was used to measure activity according to the manufacturer's instructions . From the ESCAPE database , a network containing 15 core pluripotency and 15 lineage-specific components was extracted . Arrows were established if there was evidence for binding of a specific TF to a target gene from mESC ChIP-chip/seq studies as well as a change in target gene expression level after loss-of-function ( LOF ) or gain-of-function ( GOF ) of the same TF in mESCs . We applied a majority-voting function giving more weight to LOF than to GOF evidence and binarized the output as either activation ( +1 ) or inhibition ( −1 ) using the sign function: ( 1 ) Where is 1 for activation or −1 for inhibition according to the LOF study where gene is depleted by RNAi or deleted by homologous recombination . is 1 for activation or otherwise −1 for inhibition according to the GOF study where gene is over-expressed . C with value 1 if there exists at least one protein-target gene promoter binding interaction connecting transcription factor to target gene from ChIP-chip/seq studies and 0 otherwise . A few links ( 8 out of 450 potential interactions ) were manually refined in cases with contradictory evidence using information from small-scale functional studies ( Table S2 ) . Experimental data from the Fluidigm platform measure transcript abundance of up to 96 genes analyzed in 96 single cells . Ct-values were normalized to the housekeeping gene Gapdh levels in each cell and converted to binary values ( 1 for high expression and 0 for low expression ) using K-means clustering with K = 2 ( MATLAB , Bioinformatics Toolbox ) . Histogram curves for normalized Ct-values were smoothened using the Kernel smoothing algorithm ( MATLAB , Statistics Toolbox ) . Hierarchical clustering for binarized and continuous expression levels of the 30 network genes in 96 individual cells were hierarchically clustered using the average-linkage clustering algorithm ( MATLAB , Bioinformatics Toolbox ) . The calibrated network inferred from the Boolean-function-learning-process is simulated using discrete Boolean dynamics with synchronous updating and the learned Boolean functions . An expression pattern is defined as a state vector where is determined by with the underlying Boolean functions as follows: . One Boolean network consisting of 30 genes and a set of Boolean functions is sampled randomly from all learned Boolean functions calibrated through the learning workflow . For each in silico simulation setting , 10 networks were sampled and the gene status ( 0 or 1 ) is recorded in for each gene in condition as follows: for each sampled network , 100 random initial conditions are evolved for 30 steps in synchronous mode . We set the step of 30 since all networks can reach steady states within a step of 30 in our case . In silico knockdowns are achieved by forcing gene ( s ) to the ‘OFF’ state all the time while recording the final state of the Boolean network . Each final state , as denoted by an attractor is recorded with weight based on the size of its basin ( defined as the set of initial states that lead to an attractor ) . Since a network can reach multiple steady states with certain probabilities , we weighted each stable state by the probability of being in that particular state . Thus the final binary state of each gene in each condition for a sampled network is determined by , where denotes the state of gene in attractor and is the weight of attractor . Then for 10 sampled networks is calculated as follows: The formula for calculating the discordance score for gene in condition between simulation and experiment is as follows: Where is absolute value function; reflects the expression change for gene in condition relative to unperturbed condition under in silico simulations , with values from representing ‘down-regulation’ , ’no-change’ and ‘up-regulation’ , respectively; reflects fold-change of gene expression in knockdown mESCs relative to empty-vector transfected controls measured by RT-PCR , with negative values representing ‘up-regulation’ and positive values representing ‘down-regulation’ . As the logistic function is monotonically increasing , larger value of would result in larger discordance score . Single cell expression data are gradually randomized by flipping data points with a certain percentage . The topology of the network is randomized by rewiring percent links represented as entries in the adjacency matrix underlying the 30 node network extracted from the ESCAPE database . An objective function is defined to quantify the error between in silico knockdown simulations and the experimental results . , where is the same discordance score defined above . Relative accuracy is defined as . For each , permutation level and learning process is repeated 10 times to obtain means and standard deviations . Mean discordance scores across all genes and perturbation conditions ( 30×7 = 210 ) from the model were compared to prediction results with a random predictor . This is a simple predictor with outputs randomly chosen from −1 , 0 or +1 representing ‘down-regulation’ , ‘no change’ or ‘up-regulation’ , respectively . A total of 500 random M30×7 matrices were generated using the random predictor . Rows represent genes and columns represent perturbation conditions . Individual mean discordance scores were calculated for each of the 500 random matrices . A one-sample t-test was performed to test the null hypothesis that the random sample mean is equal to mean discordance scores from the model .
|
For this study we first constructed a directed and signed network consisting of 15 pluripotency regulators and 15 lineage commitment markers that extensively interact to regulate mouse embryonic stem cells fate decisions from data available in the public domain . Given the connectivity structure of this network , the underlying regulatory logic was learned using single cell gene expression measurements of mESCs cultured in two different conditions . With connectivity and logic learned , the network was then simulated using a dynamic Boolean logic framework . Such simulations enabled prediction of knockdown effects on the overall activity of the network . Such predictions were validated by single and combinatorial RNA interference experiments followed by expression measurements . Finally , lineage specification outcomes upon single and combinatorial gene knockdowns were predicted for all possible knockdown combinations .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] |
[
"computer",
"and",
"information",
"sciences",
"systems",
"science",
"mathematics",
"stem",
"cells",
"animal",
"cells",
"cell",
"biology",
"embryonic",
"stem",
"cells",
"single",
"cell",
"primary",
"cell",
"culture",
"cell",
"culturing",
"techniques",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"biological",
"cultures",
"dynamical",
"systems",
"research",
"and",
"analysis",
"methods"
] |
2014
|
Construction and Validation of a Regulatory Network for Pluripotency and Self-Renewal of Mouse Embryonic Stem Cells
|
Genome-wide association studies have identified hundreds of risk loci for autoimmune disease , yet only a minority ( ~25% ) share genetic effects with changes to gene expression ( eQTLs ) in immune cells . RNA-Seq based quantification at whole-gene resolution , where abundance is estimated by culminating expression of all transcripts or exons of the same gene , is likely to account for this observed lack of colocalisation as subtle isoform switches and expression variation in independent exons can be concealed . We performed integrative cis-eQTL analysis using association statistics from twenty autoimmune diseases ( 560 independent loci ) and RNA-Seq data from 373 individuals of the Geuvadis cohort profiled at gene- , isoform- , exon- , junction- , and intron-level resolution in lymphoblastoid cell lines . After stringently testing for a shared causal variant using both the Joint Likelihood Mapping and Regulatory Trait Concordance frameworks , we found that gene-level quantification significantly underestimated the number of causal cis-eQTLs . Only 5 . 0–5 . 3% of loci were found to share a causal cis-eQTL at gene-level compared to 12 . 9–18 . 4% at exon-level and 9 . 6–10 . 5% at junction-level . More than a fifth of autoimmune loci shared an underlying causal variant in a single cell type by combining all five quantification types; a marked increase over current estimates of steady-state causal cis-eQTLs . Causal cis-eQTLs detected at different quantification types localised to discrete epigenetic annotations . We applied a linear mixed-effects model to distinguish cis-eQTLs modulating all expression elements of a gene from those where the signal is only evident in a subset of elements . Exon-level analysis detected disease-associated cis-eQTLs that subtly altered transcription globally across the target gene . We dissected in detail the genetic associations of systemic lupus erythematosus and functionally annotated the candidate genes . Many of the known and novel genes were concealed at gene-level ( e . g . IKZF2 , TYK2 , LYST ) . Our findings are provided as a web resource .
The autoimmune diseases are a family of heritable , often debilitating , complex disorders in which immune dysfunction leads to loss of tolerance to self-antigens and chronic inflammation [1] . Genome-wide association studies ( GWAS ) have now detected hundreds of susceptibility loci contributing to risk of autoimmunity [2] yet their biological interpretation still remains challenging [3] . Mapping single nucleotide polymorphisms ( SNPs ) that influence gene expression ( eQTLs ) can provide meaningful insight into the potential candidate genes and etiological pathways connected to discrete disease phenotypes [4] . For example , such analyses have implicated dysregulation of autophagy in Crohn’s disease [5] , the pathogenic role of CD4+ effector memory T-cells in rheumatoid arthritis [6] , and an overrepresentation of transcription factors in systemic lupus erythematosus [7] . Expression profiling in appropriate cell types and physiological conditions is necessary to capture the pathologically relevant regulatory changes driving disease risk [8] . Lack of such expression data is thought to explain the observed disparity of shared genetic architecture between disease association and gene expression at certain autoimmune loci [9] . A much overlooked cause of this disconnect however , is not only the use of microarrays to profile gene expression , but also the resolution to which expression is quantified using RNA-Sequencing ( RNA-Seq ) [10] . Expression estimates of whole-genes , individual isoforms and exons , splice-junctions , and introns are obtainable with RNA-Seq [11–18] . The SNPs that affect these discrete units of expression vary strikingly in their proximity to the target gene , localisation to specific epigenetic marks , and effect on translated isoforms [18] . For example , in over 57% of genes with both an eQTL influencing overall gene expression and a transcript ratio QTL ( trQTL ) affecting the ratio of each transcript to the gene total , the causal variants for each effect are independent and reside in distinct regulatory elements of the genome [18] . RNA-Seq based eQTL investigations that solely rely on whole-gene expression estimates are likely to mask the allelic effects on independent exons and alternatively-spliced isoforms [16–19] . This is in part due to subtle isoform switches and expression variation in exons that cannot be captured at gene-level [20] . A large proportion of trait associated variants are thought to act via direct effects on pre-mRNA splicing that do not change total mRNA levels [21] . Recent evidence also suggests that exon-level based strategies are more sensitive than conventional gene-level approaches , and allow for detection of moderate but systematic changes in gene expression that are not necessarily derived from alternative-splicing events [15 , 22] . Furthermore , gene-level summary counts can be biased in the direction of extreme exon outliers [22] . Use of isoform- , exon- , and junction-level quantification in eQTL analysis also support the potential to not only point to the candidate genes involved , but also the specific transcripts or functional domains affected [10 , 18] . This of course facilitates the design of targeted functional studies and better illuminates the causative relationship between regulatory genetic variation and disease . Lastly , though intron-level quantification is not often used in conventional eQTL analysis , it can still provide valuable insight into the role of unannotated exons in reference gene annotations , retained introns , and even intronic enhancers [23 , 24] . Low-resolution expression profiling with RNA-Seq will impede the subsequent identification of causal eQTLs when applying genetic and epigenetic fine-mapping approaches [25] . In this investigation , we aim to increase our knowledge of the regulatory mechanisms and candidate genes of human autoimmune disease through integration of GWAS and RNA-Seq expression data profiled at gene- , isoform- , exon- , junction- , and intron-level in lymphoblastoid cell lines ( LCLs ) . This is firstly performed in detail using association data from a GWAS in systemic lupus erythematosus , and is then scaled up to a total of twenty autoimmune diseases . Our findings are provided as a web resource to interrogate the functional effects of autoimmune associated SNPs ( www . insidegen . com ) , and will serve as the basis for targeted follow-up investigations .
Using densely imputed genetic association data from a large European GWAS in systemic lupus erythematosus ( SLE ) [7] , we performed integrative cis-eQTL analysis with RNA-Seq expression data profiled at five resolutions: gene- , transcript- , exon- , junction- , and intron-level . Expression data were derived from 373 healthy European donors of the Geuvadis project profiled in lymphoblastoid cell lines ( LCLs ) [18] . See S1 Fig for a summary of how expression at the five resolutions was quantified . A total of 38 genome-wide significant SLE loci ( S1 Table ) were put forward for analysis . To test for evidence of a single shared causal variant between disease and gene expression at each locus , we employed the Joint Likelihood Mapping ( JLIM ) framework [9] using summary-level statistics for SLE association and full genotype-level data for gene expression . Using JLIM , cis-eQTLs were defined if a nominal association ( P<0 . 01 ) with at least one SNP existed within 100kb of the SNP most associated with disease and the transcription start site of the gene was located within +/-500kb of that SNP ( as defined by authors of JLIM ) . JLIM P-values were corrected for multiple testing by a false discovery rate ( FDR ) of 5% per RNA-Seq quantification type ( i . e . at exon-level , JLIM P-values were adjusted for total number of exons tested in cis to the 38 SNPs ) . Causal associations of the integrative cis-eQTL SLE GWAS analysis across the five RNA-Seq quantification types are available in S2 Table and the full output ( including non-causal associations ) are available in S3 Table . The distribution of JLIM P-values across the five RNA-Seq quantification types are depicted in S2 Fig . We found the number of causal cis-eQTLs was markedly underrepresented when considering conventional gene-level quantification ( Table 1 ) . Only two of the 38 SLE susceptibility loci ( 5 . 3% ) were deemed to be causal cis-eQTLs at gene-level for three candidate genes . This is a similar proportion observed by Chun et al [9] who found that 16 of the 272 ( 5 . 9% ) autoimmune susceptibility loci tested were cis-eQTLs driven by a shared causal variant in the Geuvadis RNA-Seq dataset using gene-level quantification ( based upon the seven autoimmune diseases interrogated—not including SLE ) . Of note , transcript-level quantification did not increase the number of causal cis-eQTLs ( Table 1 ) . Transcript-level analysis did , however , yield a greater number of candidate genes ( seven individual transcripts derived from a total of four genes ) . Both junction- and intron-level quantification increased the number of causal cis-eQTLs to four ( 10 . 5% of the 38 total SLE loci ) . Using exon-level quantification , we were able to detect seven significant cis-eQTLs driven by a single shared causal variant ( 18 . 4% ) . Exon-level analysis also produced the greatest number of candidate gene targets: nine unique genes derived from 24 individual SNP-exon pairs ( Table 1 ) . Therefore , even with the severe multiple testing burden , we firstly conclude that exon- , junction- , and intron-level analysis detects more causal cis-eQTLs than gene-level . By combining all five types of RNA-Seq quantification ( gene , transcript , exon , junction , and intron ) we classified nine of the 38 SLE susceptibility loci ( 24% ) as being driven by the same causal variant as the cis-eQTL in LCLs ( Table 1 ) . This value , derived from interrogating only a single cell type , is almost equal to the total number of causal autoimmune cis-eQTLs detected by Chun et al [9] ( ~25% ) across three different cell types ( CD4+ T-cells–measured by microarray , CD14+ monocytes–microarray , and LCLs–RNA-Seq gene-level ) . We found that when considering the specificity of cis-eQTLs and target genes across the five RNA-Seq quantification types , both gene- and transcript-level quantification were redundant with respect to exon-level data; i . e . there were no causal cis-eQTLs or target genes detected at gene- or transcript-level that were not captured by exon-level analysis ( S3 Fig ) . Both junction- and intron-level quantification captured a single causal cis-eQTL each that was not captured by exon-level . We conclude that profiling at all resolutions of RNA-Seq is required to capture the full set of potentially causal cis-eQTLs . We compared the detection of cis-eQTLs using a pairwise comparison between the five RNA-Seq quantification types for matched SNP-gene cis-eQTL pairs ( Fig 1 ) . We only considered matched SNP-gene cis-eQTL association pairs that had a nominal cis-eQTL association P-value < 0 . 01 in both quantification types , and to be conservative , when multiple transcripts , exons , junctions , and introns were annotated with the same gene symbol , we selected the associations that minimized the difference in JLIM P-value between matched SNP-gene cis-eQTLs across RNA-Seq quantification types . There were over 250 matched SNP-gene cis-eQTL pairs per comparison . We firstly observed that the correlation of both cis-eQTL association P-values from regression and JLIM P-values across RNA-Seq quantification types reflected the methods in which expression quantification was obtained ( Fig 1A ) . Both cis-eQTL and JLIM P-values between matched SNP-gene pairs at gene- and transcript-level were highly correlated as gene-level estimates are obtained from the sum of all transcript-level estimates for the same gene . Exon-level and junction-level associations were also highly correlated due to split-reads being incorporated into the exon-level estimate . As expected , intron-level cis-eQTL and JLIM P-values for matched SNP-gene pairs were only weakly correlated against other quantification types—as reads mapping to introns are not included in the other quantification models . Interestingly , although cis-eQTL association P-values for matched SNP-gene pairs between transcript-level and junction-level were found to be relatively high ( r2 = 0 . 70 ) , we found the JLIM P-values for the matched pairs to be comparatively low ( r2 = 0 . 29 ) ; suggesting that whilst the statistical significance of matched cis-eQTLs maybe similar between these quantification types , the underlying causal variants driving the disease and cis-eQTL association are likely to be independent . By plotting the JLIM P-values for matched SNP-gene pairs between different quantification types , we found many instances of P-values distributed along the axes rather than on the diagonal ( Fig 1B ) . Our findings therefore suggest that often , one quantification type is more likely to explain the observed disease association than the other . When we compared conventional gene-level cis-eQTL analysis against exon-level results ( Fig 1C ) , we found that of the 296 matched SNP-gene cis-eQTL associations ( P<0 . 01 ) , eleven ( 4% ) shared the same causal variant at both gene- and exon-level using a nominal JLIM P-value threshold <0 . 01 . Only three of the 296 matched SNP-gene cis-eQTL associations ( 1% ) were captured by gene-level only—in contrast to the 26 ( 9% of total associations ) captured uniquely at exon-level . As expected , the overwhelming majority of cis-eQTL associations ( 86% ) did not possess a single shared causal variant at either gene- or exon-level . We performed this analysis for all possible combinations of quantification types ( Table 2 ) . In each instance , gene-level analysis detected only the minority of nominally causal associations for matched SNP-gene association pairs ( JLIM P<0 . 01 ) . Exon-level and junction-level analysis consistently detected more causal cis-eQTL associations than gene- , transcript- , and intron-level . In fact , when combined , exon- and junction-level analysis explained the most nominally causal associations for all significant SNP-gene cis-eQTL association pairs ( 24% ) . We functionally dissected the 12 candidate genes taken from the nine SLE associated loci that showed strong evidence of a shared causal variant with a cis-eQTL in LCLs ( Table 3 ) . We systematically annotated these genes using a combination of cell/tissue expression patterns , mouse models , known molecular phenotypes , molecular interactions , and associations with other autoimmune diseases ( S4 Table ) . We found the majority of novel SLE candidate genes detected by RNA-Seq were predominately expressed in immune-related tissues such as whole blood as well as the spleen and thymus . Based on our annotation and what is already documented at certain loci , we were sceptical on the pathogenic involvement of three candidate genes ( PHTF1 , ARHGAP30 , and RABEP1 ) . Although the cis-eQTL effect for these genes is evidently driven by the shared causal variant as the disease association , it is possible that these effects of expression modulation are merely passengers that are carried on the same functional haplotype as the true causal gene ( s ) and do not contribute themselves to the breakdown of self-tolerance ( detailed in S4 Table ) . We show the regional association plots and the candidate genes detected from cis-eQTL analysis in S4 Fig . The causal cis-eQTL rs2736340 for genes BLK and FAM167A was detected at all RNA-Seq profiling types . It is well established that the risk allele of this SNP reduces proximal promoter activity of BLK; a member of the Src family kinases that functions in intracellular signalling and the regulation of B-cell proliferation , differentiation , and tolerance [26] . The allelic consequence of FAM167A expression modulation is unknown . We found multiple instances of known SLE susceptibility genes that were concealed when using gene-level quantification . For example , we defined rs7444 as a causal cis-eQTL for UBE2L3 at transcript- and exon-level—but not at gene-level ( Table 3 ) . The risk allele of rs7444 has been associated with increased expression of UBE3L3 ( Ubiquitin conjugating enzyme E2 L3 ) in ex vivo B-cells and monocytes and correlates with NF-κB activation along with increased circulating plasmablast and plasma cell numbers [27] . Similarly , the rs10028805 SNP is a known splicing cis-eQTL for BANK1 ( B-cell scaffold protein with ankyrin repeats 1 ) . We replicated at exon- , and junction-level this splicing effect which has been proposed to alter the B-cell activation threshold [28] . Again , this mechanism was not detected using gene-level quantification . IKZF2 ( detected at the exon-level only ) is a transcription factor thought to play a key role in T-reg stabilisation in the presence of inflammatory responses [29] . IKZF2 deficient mice acquire an auto-inflammatory phenotype in later life similar to rheumatoid arthritis , with increased numbers of activated CD4+ and CD8+ T-cells , T-follicular helper cells , and germinal centre B-cells , which culminates in autoantibody production [30] . Of note , other members of this gene family , IKZF1 and IKZF3 , are also associated with SLE and can hetero-dimerize ( S4 Table ) [7] . We also believe LYST , ATG4D , and TYK2 to also be intriguing candidate genes . LYST encodes a lysosomal trafficking regulator [31] whilst ATG4D is a cysteine peptidase involved in autophagy and this locus is associated with multiple sclerosis , psoriasis , and rheumatoid arthritis [32] . TYK2 is discussed in greater detail in the following section . Interestingly , for the three causal SNP-gene pairs detected at gene-level ( rs2736340 –BLK , rs2736340 –FAM167A , and rs7444 –CCDC116 ) , we found that at exon-level , all expressed exons possessed causal cis-eQTLs . For example , rs2736340 is a causal cis-eQTL for all thirteen exons of BLK and for all three exons of FAM167A ( S5 Table ) . These data suggest that gene-level analysis is capturing associations where all—or the majority of exons—are modulated by the cis-eQTL . We found that within the SLE associated loci that showed evidence of a shared causal variant with a cis-eQTL ( Table 3 ) , there were many instances in which the proposed causal cis-eQTL modulated expression of only a single expression element . This enabled us to resolve the potential regulatory effect of the causal cis-eQTL to a particular transcript , exon , junction , or intron ( S5 Table ) . We were able to resolve to a single expression element in nine of the twelve candidate SNP-gene pairs . For example , rs9782955 is a causal cis-eQTL for LYST at junction-level for only a single junction ( chr1:235915471–235916344; cis-eQTL P = 1 . 3x10-03; JLIM P = 2 . 0x10-04 ) . We provide depicted examples of this isolation analysis for candidate genes IKZF2 ( S5 Fig ) , UBE2L3 ( S6 Fig ) , and LYST ( S7 Fig ) . We provide a worked example of resolving the causal mechanism ( s ) using RNA-Seq for the novel association rs2304256 with TYK2 ( Fig 2 ) . The top panel of Fig 2A shows the genetic association to SLE at the 19p13 . 2 susceptibility locus tagged by lead SNP rs2304256 ( P = 1 . 54x10-12 ) . Multiple tightly correlated SNPs span the gene body and the 3′ region of TYK2 –which encodes Tyrosine Kinase 2—thought to be involved in the initiation of type I IFN signalling [33] . In the panel below , we plot the gene-level association of all SNPs in cis to TYK2 and show no significant association of rs3204256 with TYK2 expression ( P = 0 . 18 ) . At exon- , and intron-level , we were able to classify rs2304256 as a causal cis-eQTL for a single exon ( chr19: 10475527–10475724; cis-eQTL P = 2 . 58x10-09; JLIM P<10−04 ) and a single intron ( chr19: 10473333–10475290; P = 2 . 20x10-08; JLIM P = 2x10-04 ) of TYK2 respectively as shown in the bottom two panels of Fig 2A . We show the exon and intron labelling of TYK2 in further detail in S8 Fig . We found strong correlation of association P-values of the SLE GWAS and the P-values of TYK2 cis-eQTLs against at exon-level and intron-level , but not at gene-level ( Fig 2B ) . The risk allele rs2304256 [C] was found to be associated with decreased expression of the TYK2 exon and increased expression of the TYK2 intron ( Fig 2C ) . By plotting the cis-eQTL P-values alongside the JLIM P-values for all exons and introns of TYK2 against rs2304256 ( Fig 2D ) , we clearly show that only a single exon and a single intron of TYK2 colocalize with the SLE association signal–marked by an asterisk ( note that rs2304256 is a strong cis-eQTL for many introns of TYK2 but only shares a causal variant with one intron ) . We show the genomic location of the affected exon and intron of TYK2 in Fig 2E ( exon 8 and the intron between exons 9 and 10 ) . Intron 9–10 of TYK2 is clearly expressed in LCLs according to transcription levels assayed by RNA-Seq on LCLs ( GM12878 ) from ENCODE ( Fig 2E ) . Interestingly , rs2304256 ( marked by an asterisk in Fig 2E ) is a missense variant ( V362F ) within exon 8 of TYK2 . The PolyPhen prediction of this substitution is predicted to be benign and , to the best of our knowledge , no investigation has isolated the functional effect of this particular amino acid change . We do not believe the cis-eQTL at exon 8 to be a result of variation at rs3204256 and mapping biases , as the alignability of 75mers by GEM from ENCODE is predicted to be robust around exon 8 ( Fig 2E ) . In fact , rs3204256 [C] is the reference allele yet is associated with decreased expression of exon 8 . In conclusion , we have found an interesting and novel mechanism that would have been concealed by gene-level analysis that involves the risk allele of a missense SNP associated with decreased expression of a single exon of TYK2 but increased expression of the neighbouring intron . Whether the cis-eQTL effect and missense variation act in a combinatorial manner and whether the intron is truly retained or if it is derived from an unannotated transcript of TYK2 is an interesting line of investigation . We re-performed our integrative cis-eQTL analysis with the Geuvadis RNA-Seq dataset in LCLs using association data from twenty autoimmune diseases . This was to firstly reiterate the importance of leveraging RNA-Seq in GWAS interpretation and to secondly demonstrate that our findings in SLE persisted across other immunological traits . As the raw genetic association data were not available for all twenty diseases , we were unable to implement the JLIM pipeline which requires densely typed or imputed GWAS summary-level statistics . We therefore opted to use the Regulatory Trait Concordance ( RTC ) method , which requires full genotype-level data for the expression trait , but only the marker identifier for the lead SNP of the disease association trait ( see methods for a description of the RTC method ) . We stringently controlled our integrative cis-eQTL analysis for multiple testing to limit potential false positive findings of overlapping association signals . To do this , we applied a Bonferroni correction to nominal cis-eQTL P-values separately per disease and per RNA-Seq quantification type . We rigorously defined causal cis-eQTLs , as associations with PBF < 0 . 05 and RTC ≥ 0 . 95 . An overview of the analysis pipeline is depicted in S9 Fig and S10 Fig . Using an r2 cut-off of 0 . 8 and a 100kb limit , we pruned the 752 associated SNPs from the twenty human autoimmune diseases from the Immunobase resource ( S6 Table ) to obtain 560 independent susceptibility loci . Our findings confirmed our previous results from the SLE investigation , and again support the gene-level study using the JLIM package . As before , we found that only 5% ( 28 of the 560 loci ) of autoimmune susceptibility loci were deemed to share causal variants with cis-eQTLs using either gene- or transcript-level analysis ( Fig 3A ) . Exon-level analysis more than doubled the yield to 13% ( 72 of the 560 loci ) with junction- , and intron-level analysis also outperforming gene-level ( 10% and 8% respectively ) . When combining all RNA-Seq quantification types , we could define 20% of autoimmune associated loci ( 110 of the 560 loci ) as being candidate causal cis-eQTLs—which corroborates our previous estimate in SLE using JLIM ( 24% ) . By separating causal cis-eQTL associations out by quantification type , we found over half ( 65% ) were detected at exon-level , and considerable overlap of cis-eQTL associations existed between both types ( Fig 3B ) . Unlike in our SLE analysis , gene- and isoform-level analysis did capture a small fraction of causal cis-eQTLs that were not captured at exon-level . Our data therefore suggest that although exon- and junction-level , and to a lesser extent intron-level analysis , capture most candidate-causal cis-eQTLs . It is necessary to prolife gene-expression at all quantification types to avoid misinterpretation of the functional impact of disease associated SNPs . We mapped the causal cis-eQTLs detected by all RNA-Seq quantification types back to the diseases to which they are associated ( Fig 3C ) . Interestingly , we observed the diseases that fell below the 20% average comprised autoimmune disorders related to the gut: celiac disease ( 7% ) , inflammatory bowel disease ( 14% ) , Crohn’s disease ( 16% ) , and ulcerative colitis ( 18% ) . We attribute this observation as a result of the cellular expression specificity of associated genes in colonic tissue and in T-cells [34] . Correspondingly , we observed an above-average frequency of causal cis-eQTLs detected in SLE ( 22% ) and primary biliary cirrhosis ( 37% ) ; diseases in which the pathogenic role of B-lymphocytes and autoantibody production is well documented [34] . Note that there are 60 SLE GWAS associations in this analysis as these originate from three independent GWA studies ( S6 Table ) . We further broke down our results per disease by RNA-Seq quantification type ( Fig 3D ) and in all cases , the greatest frequency of causal cis-eQTLs and candidate genes were captured by exon- and junction-level analyses . We provide the results from our analysis as a web resource ( found at www . insidegen . com ) for researchers to lookup causal cis-eQTLs and candidate genes from the twenty autoimmune diseases detected across the five RNA-Seq quantification types . The data are sub-settable and exportable by SNP ID , gene , RNA-Seq resolution , genomic position , and association to specific autoimmune diseases . See methods for a walkthrough of how to access results . By implementing a mixed model test of heterogeneity that accounts for the dependency structure arising from within-individual and within-gene expression correlations , we attempted to distinguish causal cis-eQTLs at transcript- , exon- , junction- , and intron-level that fitted either a systematic gene-model ( characterized by a similar effect on expression across all elements within a gene ) or a heterogeneous gene-model ( where the cis-eQTL signal is only evident in a subset of expression elements ) . The full results of this analysis are found in S7 Table . We found that across each RNA-Seq profiling type , the majority of causal cis-eQTLs exhibited heterogeneous effects on gene expression; indicative of alternative isoform usage ( Fig 4A ) . Junction-level causal cis-eQTLs had the greatest proportion of heterogeneous associations ( 49 of 65 causal cis-eQTLs were heterogeneous—75% ) . Both systematic and heterogeneous causal cis-eQTLs were then stratified by whether or not they were also causal at gene-level . As expected , we observed that causal cis-eQTLs that were also detected at gene-level ( Fig 4B ) showed a greater proportion of systematic effects on gene expression than associations not detected at gene-level ( Fig 4C ) . In both cases however , the heterogeneous model was more apposite . Interestingly , we found that the greatest frequency of systematic associations , which were not captured at gene-level , were observed at exon-level ( 42 of 76: 55% ) . This implies that exon-level analysis captures a near equal proportion of both systematic and heterogeneous effects that are not detected by gene-level analysis . We show four examples of systematic and heterogeneous causal cis-eQTLs stratified by their detection at gene-level quantification in Fig 5 . A previous investigation has suggested that causal variants of gene-level and transcript-level cis-eQTLs reside in discrete functional elements of the genome [18] . We therefore investigated whether this notion held true across the five RNA-Seq quantification types tested in this study . To accomplish this , we selected the causal cis-eQTLs from the twenty autoimmune diseases interrogated , and per quantification type , tested for enrichment of these SNPs across various chromatin regulatory elements taken from the Roadmap Epigenomics Project in LCLs ( using both the Roadmap chromatin state model and the positions of histone modifications ) . We implemented the permutation-based GoShifter algorithm to test for enrichment of causal cis-eQTLs and tightly correlated variants ( r2>0 . 8 ) in genomic functional annotations in LCLs ( see methods ) [25] . Results of this analysis are depicted in Fig 6 . We found the 28 gene-level cis-eQTLs were enriched in two chromatin marks: strong enhancers ( P = 0 . 036 ) and H3K27ac occupancy sites–a marker of active enhancers ( P = 0 . 002 ) . Transcript-level cis-eQTLs were also enriched in H3K27ac occupancy sites ( P = 0 . 039 ) but were not enriched in any other marks . The 72 exon-level cis-eQTLs were additionally enriched in active promoters ( P = 0 . 017 ) . Interestingly , the 54 causal cis-eQTLs detected at junction-level were found to be enriched in weak enhancers only ( P = 0 . 002 ) ; whilst the 43 intron-level cis-eQTLs were enriched in chromatin states predicted to be involved in transcriptional elongation ( P = 0 . 001; 83% of intron-level cis-eQTLs ) . Disease relevant cis-eQTLs detected at different expression phenotypes using RNA-Seq clearly localise to largely discrete functional elements of the genome . We quantified the number of causal cis-eQTLs and tightly correlated variants ( r2>0 . 8 ) per quantification type that were predicted to be alter splice site consensus sequences of the target genes ( assessed by Sequence Ontology for the hg19 GENCODE v12 reference annotation ) . We found only two of the 28 ( 7% ) gene-level cis-eQTLs disrupted consensus splice-sites for their target genes compared to the 14% and 13% detected at exon- and junction-level respectively ( Fig 6C ) . Our data suggest that although exon- and junction- level analysis leads to the greatest frequency of causal cis-eQTLs , the majority at this resolution cannot be explained directly by variation in annotated splice site consensus sequences ( splice region/donor/acceptor/ variants ) . We extended our investigation and performed genome-wide cis-eQTL analysis for all SNPs against gene- , transcript- , exon- , junction- , and intron-level quantifications . As with our integrative analysis of autoimmune risk loci , we found the greatest number of genome-wide significant cis-eQTLs and target genes ( at a genome-wide FDR threshold of 5% ) were detected at exon-level , followed by junction- and intron-level; with gene- and transcript-level being thoroughly outperformed ( S8 Table and S11 Fig ) . We confirmed that all of the causal cis-eQTL associations detected in our integrative analysis with autoimmune risk loci reached genome-wide significance—owing to the stringent Bonferroni multiple testing correction adopted ( S9 Table ) .
Elucidation of the functional consequences of non-coding genetic variation in human disease is a major objective of medical genomics [35] . Integrative studies that map disease-associated eQTLs in relevant cell types and physiological conditions are proving essential in progression towards this goal through identification of causal SNPs , candidate-genes , and illumination of molecular mechanisms [36] . In autoimmune disease , where there is considerable overlap of immunopathology , integrative eQTL investigations have been able to connect discrete aetiological pathways , cell types , and epigenetic modifications , to particular clinical manifestations [2 , 34 , 36 , 37] . Emerging evidence however has suggested that only a minority ( ~25% ) of autoimmune associated SNPs share casual variants with basal-level cis-eQTLs in primary immune cell-types [9] . Genetic variation can influence expression at every stage of the gene regulatory cascade—from chromatin dynamics , to RNA folding , stability , and splicing , and protein translation [21] . It is now well documented that SNPs affecting these units of expression vary strikingly in their genomic positions and localisation to specific epigenetic marks [18] . The eQTLs that affect pre-transcriptional regulation—affecting all isoforms of a gene—differ in the proximity to the target gene and effect on translated isoforms than their co-transcriptional trQTL ( transcript ratio QTL ) counterparts . Where the effect size of eQTLs generally increases in relation to transcription start site proximity , trQTLs are distributed across the transcript body and generally localise to intronic binding sites of splicing factors [18 , 21] . In over 57% of genes with both an eQTL influencing overall gene expression and an trQTL affecting the ratio of each transcript to the gene total , the causal variants for each effect are independent and reside in distinct regulatory elements of the genome [18] . In fact , three primary molecular mechanisms are thought to link common genetic variants to complex traits . A large proportion of trait associated SNPs act via direct effects on pre-mRNA splicing that do not change total mRNA levels [21] . Common variants also act via alteration of pre-mRNA splicing indirectly through effects on chromatin dynamics and accessibility . Such chromatin accessibility QTLs are however more likely to alter total mRNA levels than splicing ratios . Lastly , it is thought that only a minority of trait associated variants have direct effects on total gene expression that cannot be explained by changes in chromatin . As RNA-Seq becomes the convention for genome-wide transcriptomics , it is essential to maximise its ability to resolve and quantify discrete transcriptomic features so to expose the genetic variants that contribute to changes in expression and isoform usage . The reasoning for our investigation therefore was to delineate the limits of microarray and RNA-Seq based eQTL cohorts in the functional annotation of autoimmune disease association signals . To map autoimmune disease associated cis-eQTLs , we interrogated RNA-Seq expression data profiled at gene- , isoform , exon- , junction- , and intron-level , and tested for a shared genetic effect at each significant association . As we had densely imputed summary statistics from our SLE GWAS , we opted to use the Joint Likelihood Mapping ( JLIM ) framework [9] to test for a shared causal variant between the disease and cis-eQTL signals . This framework has been rigorously benchmarked against other colocalisation procedures . Summary statistics were not available for the remaining autoimmune diseases and therefore we implemented the Regulatory Trait Concordance ( RTC ) method for these diseases and set a stringent multiple testing threshold to define causal cis-eQTLs . We found the estimates of causal cis-eQTLs were near identical between the two methods used ( Table 1 and Fig 3A ) . Exon- and junction-level quantification led to the greatest frequency of causal cis-eQTLs and candidate genes ( exon-level: 13–18% , junction-level: 10–11% ) . We conclusively found that associated variants were in fact more likely to colocalize with exon- and junction-level cis-eQTLs when applying a nominal JLIM P-value threshold of <0 . 01 ( Fig 1B and Table 2 ) . Gene-level analysis was thoroughly outperformed in all cases ( 5% ) . Our findings that gene-level analysis explain only 5% of causal cis-eQTLs corroborate the findings from Chun et al [9] who composed and used the JLIM framework to annotate variants associated with seven autoimmune diseases ( multiple sclerosis , IBD , Crohn’s disease , ulcerative colitis , T1D , rheumatoid arthritis , and celiac disease ) . They found that only 16 of the 272 autoimmune associated loci ( 6% ) shared causal variants with cis-eQTLs using gene-level RNA-Seq ( with the same Geuvadis European cohort in LCLs as used herein ) . In our investigation , we argue that it is necessary to profile expression at all possible resolutions to diminish the likelihood of overlooking potentially causal cis-eQTLs . In fact , by combining our results across all resolutions , we found that 20–24% of autoimmune loci were candidate-causal cis-eQTLs for at least one target gene . Our study therefore increases the number of autoimmune loci with shared genetic effects with cis-eQTLs in a single cell type by over four-fold . Interestingly , using microarray data from CD4+ T-cells Chun et al classified 37 of the 272 autoimmune loci ( 14% ) as causal cis-eQTLs [9]—strengthening the hypothesis that autoimmune loci ( especially those associated with inflammatory diseases of the gut ) are enriched in CD4+ T-cell subsets and the cells themselves are likely to be pathogenic [25 , 34] . Microarray data are known to underestimate the number of true causal cis-eQTLs [10] . If we assume that by leveraging RNA-Seq we can increase the number of steady-state causal cis-eQTLs four-fold , we hypothesise that as many as ~54% of autoimmune loci may share causal cis-eQTLs with gene expression at multiple resolutions in CD4+ T-cell populations . A large RNA-Seq based eQTL cohort profiled across multiple CD4+ T-cell subsets will therefore be of great use when annotating autoimmune-related traits . Immune activation conditions further increase the number of causal cis-eQTLs detected in autoimmune disease [38] . We reason that although using relevant cell types and context-specific conditions will undoubtedly increase our understanding of how associated variants alter cell physiology and ultimately contribute to disease risk; it is clearly shown herein that we are only picking the low hanging fruit in current eQTL analyses . We argue it necessary to reanalyse existing RNA-Seq based eQTL cohorts at multiple resolutions and ensure new datasets are similarly dissected . Despite the severe multiple testing burden , we also argue that expression profiling at multiple resolutions using RNA-Seq may be advantageous even when looking for trans-eQTL effects . As trans-eQTLs are generally more cell-type specific and have a weaker effect size , we decided not to perform such analyses using the Geuvadis LCL data . Large RNA-Seq based eQTL cohorts in whole-blood will be more suitable for such analysis [19] . As well as biological reasons for using multiple expression phenotypes for integrative eQTL analysis , there are also technical factors to consider . Gene-level expression estimates can generally be obtained in two ways–union-exon based approaches [14 , 17] and transcript-based approaches [11 , 12] . In the former , all overlapping exons of the same gene are merged into union exons , and intersecting exon and junction reads ( including split-reads ) are counted to these pseudo-gene boundaries . Using this counting-based approach , it is also possible to quantify meta-exons and junctions easily and with high confidence by preparing the reference annotation appropriately [13 , 15 , 39] . Introns can be quantified in a similar manner by inverting the reference annotation between exons and introns [18] . Of note , we found intron-level quantification generated more candidate-causal cis-eQTLs than gene-level ( Fig 3A ) . As the library was synthesised from poly-A selection , these associations are unlikely due to differences in pre-mRNA abundance . Rather , they are likely derived from either true retained introns in the mature RNA or from coding exons that are not documented in the reference annotation used . Transcript-based approaches make use of statistical models and expectation maximization algorithms to distribute reads among gene isoforms—resulting in isoform expression estimates [11 , 12] . These estimates can then be summed to obtain the entire expression estimate of the gene . Greater biological insight is gained from isoform-level analysis; however , disambiguation of specific transcripts is not trivial due to substantial sequence commonality of exons and junctions . In fact , we found only 5% of autoimmune loci shared a causal variant at transcript-level . The different approaches used to estimate expression can also lead to significant differences in the reported counts . Union-based approaches , whilst computationally less expensive , can underestimate expression levels relative to transcript-based , and this difference becomes more pronounced when the number of isoforms of a gene increases , and when expression is primarily derived from shorter isoforms [20] . The Geuvadis study implemented a transcript-based approach to obtain whole-gene expression estimates . Clearly therefore , a gold standard of reference annotation and eQTL mapping using RNA-Seq is essential for comparative analysis across datasets . Our findings support recent evidence that suggests exon-level based strategies are more sensitive and specific than conventional gene-level approaches [22] . Subtle isoform variation and expression of less abundant isoforms are likely to be masked by gene-level analysis . Exon-level allows for detection of moderate but systematic changes in gene expression that are not captured at gene-level , and also , gene-level summary counts can be shifted in the direction of extreme exon outliers [22] . It is therefore important to note that a positive exon-level eQTL association does not necessarily mean a differential exon-usage or splicing mechanism is involved; rather a systematic expression effect across the whole gene may exist that is only captured by the increased sensitivity . By implementing a mixed model test of heterogeneity that accounts for the dependency structure arising from within-individual and within-gene expression correlations we found that causal cis-eQTLs captured by exon-level analysis that are not detected at gene-level , are derived from both systematic and heterogeneous effects on gene expression in almost equal proportions ( Fig 4 ) . Additionally , by combining exon-level with other RNA-Seq quantification types , inferences can be made on the particular isoforms and functional domains affected by the eQTL which can later aid biological interpretation and targeted follow-up investigations [10] . We clearly show this from our analysis of SLE candidate genes IKZF2 ( S5 Fig ) , UBE2L3 ( S6 Fig ) , LYST ( S7 Fig ) and TYK2 ( Fig 2 ) . For TYK2 we reveal a novel mechanism whereby the associated variant rs2304256 [C] leads to decreased expression of a single exon and increased expression of a neighbouring intron ( Fig 2 ) . By isolating particular exons , junctions , and introns , one can design more refined follow-up investigations to study the functional impact of non-coding disease associated variants . We show how our findings can be leveraged to comprehensively examine GWAS results of autoimmune diseases . We found nine of the 38 SLE susceptibility loci were causal cis-eQTLs ( Table 3 ) for 12 candidate genes which we later functionally annotated in detail ( S4 Table ) . Taken together , we have provided a deeper mechanistic understanding of the genetic regulation of gene expression in autoimmune disease by profiling the transcriptome at multiple resolutions using RNA-Seq . Similar analyses leveraging RNA-Seq in new and existing datasets using relevant cell types and context-specific conditions ( such as response eQTLs as shown in [38] ) will undoubtedly increase our understanding of how associated variants alter cell physiology and ultimately contribute to disease risk .
RNA-Sequencing ( RNA-Seq ) expression data from 373 lymphoblastoid cell lines ( LCLs ) derived from four European sub-populations ( Utah Residents with Northern and Western European Ancestry , British in England and Scotland , Finnish in Finland , and Toscani in Italia ) of the Geuvadis project [18] were obtained from the EBI ArrayExpress website under accession: E-GEUV-1 . The 89 individuals of the Geuvadis project from the Yoruba in Ibadan , Nigeria were excluded from this analysis . All individuals were included as part of the 1000Genomes Project . Expression was profiled using RNA-Seq at five quantification types: gene- , transcript- , exon- , junction- , and intron-level ( the files downloaded and used in this analysis have the suffix: ‘QuantCount . 45N . 50FN . samplename . resk10 . txt . gz’ ) . Full methods of expression quantification can be found in the original publication and on the Geuvadis wiki page: http://geuvadiswiki . crg . es/ ) ) . We have also provided a breakdown of the quantification methods in S1 Fig . Expression data downloaded represent quantifications that are corrected for sequencing depth and gene/exon etc length ( RPKM ) . Only expression elements quantified in >50% of individuals were kept and Probabilistic Estimation of Expression Residuals ( PEER ) had been used to remove technical variation [40] . We transformed all expression data to a standard normal distribution . In summary , transcripts , splice-junctions , and introns were quantified using Flux Capacitor against the GENCODE v12 basic reference annotation [16] . Reads belonging to single transcripts were predicted by deconvolution per observations of paired-reads mapping across all exonic segments of a locus . Gene-level expression was calculated as the sum of all transcripts per gene . Annotated splice junctions were quantified using split read information , counting the number of reads supporting a given junction . Intronic regions that are not retained in any mature annotated transcript , and reported mapped reads in different bins across the intron to distinguish reads stemming from retained introns from those produced by not yet annotated exons . Meta-exons were quantified by merging all overlapping exonic portions of a gene into non-redundant units and counting reads within these bins . Reads were excluded when the read pairs map to two different genes . SNPs genetically associated to systemic lupus erythematosus ( SLE ) were taken from the Bentham and Morris et al 2015 GWAS in persons of European descent [7] . The study comprised a primary GWAS , with validation through meta-analysis and replication study in an external cohort ( 7 , 219 cases , 15 , 991 controls in total ) . Independently associated susceptibility loci taken forward for this investigation were those that passed either genome-wide significance ( P<5x10-08 ) in the primary GWAS or meta-analysis and/or those that reached significance in the replication study ( q<0 . 01 ) . We defined the lead SNP at each locus as either being the SNP with the lowest P-value post meta-analysis or the SNP with the greatest evidence of a missense effect as defined by a Bayes Factor ( see original publication ) . We omitted non-autosomal associations and those within the Major Histocompatibility Complex ( MHC ) , and SNPs with a minor allele frequency ( MAF ) < 0 . 05 . In total , 38 independently associated SLE associated GWAS SNPs were taken forward for investigation ( S1 Table ) . Each susceptibility locus had previously been imputed to the level of 1000 Genomes Phase3 using a combination of pre-phasing by the SHAPEIT algorithm and imputation by IMPUTE ( see original publication for full details ) [7] . Primary trait summary statistics file . A JLIM index file for each of the 38 SLE associated SNPs was firstly generated by taking the position of each SNP ( hg19 ) and a creating a 100kb interval in both directions . Summary-level association statistics were obtained form the Bentham and Morris et al 2015 European SLE GWAS ( imputed to 1000Genomes Phase 3 ) . We downloaded summary-level association data ( chromosome , position , SNP , P-value ) for all directly typed or imputed SNPs with an IMPUTE info score ≥0 . 7 within each of the 38 intervals . The two-sided P-value was transformed into a Z-statistic as described by JLIM . Reference LD file . Genotype files in VCF format for all 373 European individuals of the Geuvadis RNA-Seq project were obtained from the EBI ArrayExpress under accession: E-GEUV-1 . The 41 individuals genotyped on the Omni 2 . 5M SNP array had been previously imputed to the Phase 1 v3 release as described [18]; the remaining had been sequenced as part of the 1000 Genomes Phase1 v3 release ( low-coverage whole genome and high-coverage exome sequencing data ) . Using VCFtools , we created PLINK binary ped/map files for each of the 38 intervals and kept only biallelic SNPs with a MAF >0 . 05 , imputation call-rates ≥ 0 . 7 , Hardy–Weinberg equilibrium P-value >1x10−04 and SNPs with no missing genotypes , we also only included SNPs that we had primary trait association summary statistics for . These are referred to as the secondary trait genotype files . We then used the JLIM Perl script fetch . refld0 . EUR . pl to generate the 38 reference LD files from the 373 individuals ( the script had been edited to include the extra 95 Finnish individuals ) . Cis-eQTL analysis . We created a separate PLINK phenotype file ( sample ID , normalized expression residual ) for each individual gene , transcript , exon , junction , and intron in cis ( within +/-500kb ) to the 38 lead SLE GWAS SNPs . We only included protein-coding , lincRNA , and antisense genes in our analysis as classified by Ensembl BioMart . Using the chromosome 20 genotype VCF file of the 373 European individuals ( E-GEUV-1 ) , we conducted principle component analysis ( PCA ) and generated an identity-by-state matrix using the Bioconductor package SNPRelate ( S9 Fig ) [41] . Based on these results , we decided to include the first three principle components and the binary imputation status ( as 41 individuals had been genotyped on the Omni 2 . 5M SNP array were imputed to the Phase 1 v3 release ) of the European individuals ( derived from Phase1 and Phase2 1000Genomes releases ) in the cis-eQTL analysis so to minimize biases derived from population structure and imputation status . We used PLINK to perform cis-eQTL analysis using the ‘—linear’ function , including the above covariates , for each expression unit ( phenotype file ) in cis to the 38 loci ( secondary trait genotype files ) . We performed 10 , 000 permutations per regression and saved the output of each permutation procedure . In cis to the 38 SLE SNPs were: 439 genes , 1 , 448 transcripts ( originating from 456 genes ) , 3 , 045 exons ( 400 genes ) , 2 , 886 junctions ( 332 genes ) , and 1 , 855 introns ( 443 genes ) . Joint likelihood mapping ( JLIM ) and multiple testing correction . Per RNA-Seq quantification type , a JLIM configuration file was created using the jlim_gencfg . sh script and JLIM then run using run_jlim . sh–setting the r2 resolution limit to 0 . 8 . We merged the configuration files and output files to create the final results table which included the primary and secondary trait association P-value , the JLIM statistic , and the JLIM P-value by permutation . Multiple testing was corrected for on the JLIM P-values per RNA-Seq quantification type using a false discovery rate ( FDR ) as applied by the authors of JLIM . A JLIM P-value <10−04 means that the JLIM statistic is more extreme than the permutation ( 10 , 000 ) . We classified causal cis-eQTLs as SLE associated variants that share a single causal variant with a cis-eQTL based on the following: if there existed a nominal cis-eQTL ( P<0 . 01 ) with at least one SNP within 100kb of the SNP most associated with disease , the transcription start site of the expression target was located within +/-500kb of that SNP , and the FDR adjusted JLIM P-value of the association passed the 5% threshold . Candidate genes modulated by the causal cis-eQTL . Using publically available resources , we systematically annotated the twelve SLE associated genes that were classified as being modulated by causal cis-eQTLs . The expression profiles at RNA-level across multiple cell and tissue types were interrogated in GTEx [42] and the Human Protein Atlas [43]—with the top three cell/tissue types documented per gene . We noted using Online Mendelian Inheritance in Man [44] any gene-phenotype relationships by caused by allelic variants and any immune-related phenotypes of animal models . Protein-protein interactions of candidate genes were taken from the BioPlex v2 . 0 interaction network ( conducted in HEK293T cells ) [45] . Using the ImmunoBase resource ( https://www . immunobase . org/ ) , we looked up each gene and noted if the gene had been prioritized as the ‘candidate gene’ within the susceptibility locus per publication . Finally , we counted the number publications from PubMed found using the keywords ‘gene name AND SLE’ . Autoimmune associated SNPs were taken from the ImmunoBase resource ( www . immunobase . org ) . This resource comprises summary case-control association statistics from twenty diseases: twelve originally targeted by the ImmunoChip consortium ( ankylosing spondylitis , autoimmune thyroid disease , celiac disease , Crohn's disease , juvenile idiopathic arthritis , multiple sclerosis , primary biliary cirrhosis , psoriasis , rheumatoid arthritis , systemic lupus erythematosus , type 1 diabetes , ulcerative colitis ) , and eight others ( alopecia areata , inflammatory bowel disease , IgE and allergic sensitization , narcolepsy , primary sclerosing cholangitis , Sjogren syndrome , systemic scleroderma , vitiligo ) . The curated studies and their corresponding references used in this analysis are presented in S6 Table . For each disease , we took the lead SNPs which were defined as a genome-wide significant SNP with the lowest reported P-value in a locus . Associations on the X-chromosome and within the MHC and SNPs with minor allele frequency < 5% were omitted from analysis , leaving 752 associated SNPs . We pruned these loci using the ‘—indep-pairwise’ function of PLINK 1 . 9 with a window size of 100kb and an r2 threshold of 0 . 8 , to create an independent subset of 560 loci . An overview of the integration pipeline using the twenty autoimmune diseases against the Geuvadis RNA-Seq cohort in 373 European LCLs is depicted in S10 Fig . Genotype data of the 373 individuals were transformed and quality controlled as previously described in the above methods sections ( biallelic SNPs kept with a MAF >0 . 05 , imputation call-rates ≥ 0 . 7 , Hardy–Weinberg equilibrium P-value >1x10−04 ) . We opted to use the Regulatory Trait Concordance ( RTC ) method to assess the likelihood of a shared causal variant between the disease association and the cis-eQTL signal [46] . This method requires full genotype-level data for the expression trait but only the marker identifier for the lead SNP of the disease association trait . SNPs within the 560 associated loci for the expression trait were firstly classified according to their position in relation to recombination hotspots ( based on genome-wide estimates of hotspot intervals ) [47] . Normalized gene expression residuals ( PEER factor normalized RPKM ) for each quantification type were transformed to standard normal and the first three principle components used as covariates in the cis-eQTL model as well as the binary imputation status ( as previously described above ) . All cis-eQTL association testing was performed using a liner regression model in R . Cis-eQTL mapping was performed for the lead SNP and all SNPs within the hotspot recombination interval against protein-coding , lincRNA , and antisense expression elements ( genes , transcripts , exons etc . ) within +/-500kb of the lead SNP . In cis to the 560 loci were: 7 , 633 genes , 27 , 257 transcripts ( originating from 7 , 310 genes ) , 52 , 651 exons ( 5 , 435 genes ) , 48 , 627 junctions ( 4 , 237 genes ) , 34 , 946 introns ( 6 , 233 genes ) . For each cis-eQTL association , the residuals from the linear-regression of the best cis-asQTL ( lowest association P-value within the hotspot interval ) were extracted . Linear regression was then performed using all SNPs within the defined hotspot interval against these residuals . The RTC score was then calculated as ( NSNPs—RankGWAS SNP / NSNPs ) . Where NSNPs is the total number of SNPs in the recombination hotspot interval , and RankGWAS SNP is the rank of the GWAS SNP association P-value against all other SNPs in the interval from the liner association against the residuals of the best cis-eQTL . We rigorously adjusted for multiple testing of cis-eQTL P-values using a Bonferroni correction per quantification type ( corrected for number of genes , isoforms , exons , junctions , and introns tested ) and per disease–as we wanted to keep our analysis as close to the authors of JLIM who themselves also adjusted per cell type and per disease . We stringently defined causal cis-eQTLs as associations with expression PBF < 0 . 05 and an RTC score ≥ 0 . 95 . Candidate genes are modulated by the cis-eQTL . Expression of gene elements ( for example exons ) within a gene are naturally correlated , as are expression data from the same individual . We therefore applied a linear mixed-effects model approach within each RNA-Seq quantification type to test for heterogeneity in cis-eQTL signal strength of causal associations . We firstly fitted a systematic gene-model containing a SNP allele dosage main effect ( encoded 0 , 1 , 2 ) and two random effects terms indexing each individual ( 1|Sample ) and each expression element found within the same gene ( 1|Target ) . We then fitted a heterogeneous gene-model containing the same terms plus a set of fixed-effect SNP dosage * expression element interaction terms . Both models were fitted via restricted maximum likelihood ( REML = FALSE ) using the lmer ( ) function of the lme4 R package . A likelihood ratio test was used to determine significance ( anova ) . P-values were corrected for multiple testing using a Bonferroni correction , correcting for all tests ( n = 230 ) across all quantification types . PBF < 0 . 05 was deemed significant for the heterogeneous model . To test for enrichment of causal cis-eQTL associations in chromatin regulatory elements we implemented the Genomic Annotation Shifter ( GoShifter ) package [25] . Chromatin regulatory elements were divided into two categories: chromatin state segmentation and histone marks . The genomic coordinates of the fifteen predicted chromatin state segmentations ( active promoter , strong enhancer , insulator etc . ) for LCLs ( in the GM12878 cell-line ) were downloaded from the UCSC Table browser ( track name: wgEncodeBroadHmmGm12878HMM ) . Histone marks and DNase hypersensitivity sites were obtained from the NIH Roadmap Epigenomics Project for LCLs ( GM12878 ) in NarrowPeak format . Sites were filtered for genome-wide significance using an FDR threshold of 0 . 01 and peak widths harmonised to 200bp in length centred on the peak summit ( as used in the GoShifter publication ) . We obtained all SNPs in strong LD ( r2 > 0 . 8 ) with the causal cis-eQTLs by using the getLD . sh script from GoShifter ( interrogating the 1000Genomes Project for Phase3 Europeans ) . Per quantification type , we then calculated the proportion of loci in which at least one SNP in LD overlapped a chromatin regulatory element ( conducted one at a time per chromatin mark ) . The coordinates of the chromatin marks were then randomly shifted , whilst retaining the positions of the SNPs , and frequency of overlap re-calculated . This was carried out over 1 , 000 permutations to draw the null distribution . The P-value was calculated as the proportion of iterations for which the number of overlapping loci was equal to or greater than that for the tested SNPs ( P < 0 . 05 used as significance threshold ) . Genome-wide cis-eQTL analysis was performed using the normalized expression residuals for each quantification type , four population principle components , and quality controlled SNP genotype data of the 373 European individuals as already described . Cis-eQTL association analysis was performed using the MatrixeQTL R package fitting the linear-model function for all SNPs within +/-500kb of protein-coding expression targets [48] . The total number of SNPs , genes , targets , and SNP-gene targets tested are documented in S8 Table and S11 Fig . The issue of multiple testing was addressed by calculating a False Discovery Rate for each SNP-target pair per quantification type and thresholding associations below 5% . R version 3 . 3 . 1 and ggplot2 was used to create heatmaps , box-plots , and correlation plots . Genes were plotted in UCSC Genome Browser [49] and regional association plots in LocusZoom [50] . To access the online results table , visit www . insidegen . com and follow the link ‘Lupus’ then ‘data for scientists’ . The table is found under title ‘Expression data associated with different autoimmune diseases’ .
|
It is well acknowledged that non-coding genetic variants contribute to disease susceptibility through alteration of gene expression levels ( known as eQTLs ) . Identifying the variants that are causal to both disease risk and changes to expression levels has not been easy and we believe this is in part due to how expression is quantified using RNA-Sequencing ( RNA-Seq ) . Whole-gene expression , where abundance is estimated by culminating expression of all transcripts or exons of the same gene , is conventionally used in eQTL analysis . This low resolution may conceal subtle isoform switches and expression variation in independent exons . Using isoform- , exon- , and junction-level quantification can not only point to the candidate genes involved , but also the specific transcripts implicated . We make use of existing RNA-Seq expression data profiled at gene- , isoform- , exon- , junction- , and intron-level , and perform eQTL analysis using association data from twenty autoimmune diseases . We find exon- , and junction-level thoroughly outperform gene-level analysis , and by leveraging all five quantification types , we find >20% of autoimmune loci share a single genetic effect with gene expression . We highlight that existing and new eQTL cohorts using RNA-Seq should profile expression at multiple resolutions to maximise the ability to detect causal eQTLs and candidate genes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"genome-wide",
"association",
"studies",
"rheumatology",
"medicine",
"and",
"health",
"sciences",
"immunology",
"clinical",
"medicine",
"genome",
"analysis",
"molecular",
"genetics",
"epigenetics",
"chromatin",
"systemic",
"lupus",
"erythematosus",
"genome",
"complexity",
"chromosome",
"biology",
"lupus",
"erythematosus",
"gene",
"expression",
"molecular",
"biology",
"genetic",
"loci",
"cell",
"biology",
"clinical",
"immunology",
"genetics",
"autoimmune",
"diseases",
"biology",
"and",
"life",
"sciences",
"genomics",
"computational",
"biology",
"introns",
"human",
"genetics"
] |
2017
|
Profiling RNA-Seq at multiple resolutions markedly increases the number of causal eQTLs in autoimmune disease
|
LINC complexes are evolutionarily conserved nuclear envelope bridges , composed of SUN ( Sad-1/UNC-84 ) and KASH ( Klarsicht/ANC-1/Syne/homology ) domain proteins . They are crucial for nuclear positioning and nuclear shape determination , and also mediate nuclear envelope ( NE ) attachment of meiotic telomeres , essential for driving homolog synapsis and recombination . In mice , SUN1 and SUN2 are the only SUN domain proteins expressed during meiosis , sharing their localization with meiosis-specific KASH5 . Recent studies have shown that loss of SUN1 severely interferes with meiotic processes . Absence of SUN1 provokes defective telomere attachment and causes infertility . Here , we report that meiotic telomere attachment is not entirely lost in mice deficient for SUN1 , but numerous telomeres are still attached to the NE through SUN2/KASH5-LINC complexes . In Sun1−/− meiocytes attached telomeres retained the capacity to form bouquet-like clusters . Furthermore , we could detect significant numbers of late meiotic recombination events in Sun1−/− mice . Together , this indicates that even in the absence of SUN1 telomere attachment and their movement within the nuclear envelope per se can be functional .
Nuclear anchorage and movement , including the directed repositioning of components within the nucleus , are essential for coordinated cell division , proliferation and development [1] . As these processes are largely dependent on cytoskeletal components , the cytoskeleton needs to interact with both the nuclear envelope ( NE ) and the nuclear content [2] . In this context , the so-called LINC ( linker of nucleoskeleton and cytoskeleton ) complexes emerged as the key players in that they represent the central connectors of the nucleus and its content to diverse elements of the cytoskeleton [2]–[4] . LINC complexes are widely conserved in evolution regarding their composition and function . They are composed of SUN ( Sad-1/UNC-81 ) domain proteins that reside in the inner nuclear membrane ( INM ) which bind to KASH ( Klarsicht/ANC-1/Syne/homology ) domain proteins of the outer nuclear membrane ( ONM ) [4] , [5] . Through specific interactions of SUN domain proteins with nuclear components , such as lamins , and the interactions of KASH domain proteins with the cytoskeleton , the SUN-KASH complexes are able to transfer mechanical forces of the cytoskeleton directly to the NE and into the nucleus [6] , [7] . During meiosis , telomeres are tethered to and actively repositioned within the NE . The characteristic telomere-led chromosome movements are an evolutionarily highly conserved hallmark of meiotic prophase I; they are a prerequisite for ordered pairing and synapsis of homologous chromosomes [8] , [9] . Directed chromosome movement , pairing and recombination are closely interdependent processes and their correct progression is essential for the faithful segregation of homologous chromosomes into fertile gametes . Failure in any of these processes leads to massive meiotic defects and , consistent with this , mutant mice showing defects in meiotic telomere attachment , chromosome dynamics or synapsis formation are mostly infertile due to apoptosis during prophase I [10]–[13] . The attachment of meiotic telomeres to the NE is mediated by SUN-KASH protein complexes [11] , [14]–[21] . Of the five SUN-domain proteins known in mammals , SUN1 and SUN2 have been shown to be the only ones that are also expressed in meiotic cells [11] , [22] . Recently , a novel meiosis-specific KASH domain protein , KASH5 , has been identified as a constituent of the meiotic telomere attachment complex [23] , [24] . With this , the first fully functional and complete mammalian meiotic LINC complex comprised of SUN1 and/or SUN2 within the INM and KASH5 as the ONM partner has been characterized . Nonetheless , many aspects of mammalian meiotic telomere attachment and movement , including its regulation , are not yet fully understood . To date , SUN1 and SUN1/SUN2 deficient mice have been studied to investigate both somatic and meiotic functions of SUN1 and SUN2 [11] , [21] , [25] , [26] . These studies have provided clear evidence that in somatic cells SUN1 and SUN2 play partially redundant roles . However , it also turned out that mice deficient in SUN1 are infertile due to serious problems in attaching meiotic telomeres to the nuclear envelope [11] , [21] , demonstrating the importance of SUN1 for meiotic cell division . Although SUN2 was found to be present at the sites of telomere attachment during meiotic prophase I , the SUN1 deficient phenotype demonstrated that SUN2 apparently is not able to effectively compensate for the loss of SUN1 in meiosis [11] , [21] , [22] . To learn more about the distinctive roles of SUN1 and SUN2 in meiotic telomere function and behavior we started a detailed re-evaluation of the meiotic phenotype caused by SUN1 deficiency . In our current study we now show that in the absence of SUN1 meiotic telomere attachment actually is not entirely lost , pointing to the existence of a SUN1-independent , partially redundant attachment mechanism . Consistent with this , we could find that in Sun1−/− mice NE-attached telomeres co-localize with SUN2 and KASH5 , suggesting that telomere attachment is mediated by SUN2/KASH5-LINC complexes in SUN1 deficient meiocytes . Furthermore , Sun1−/− meiocytes showed clustering patterns of the NE-attached telomeres that resembled typical bouquet-like configurations , indicating that SUN2 is not only sufficient to connect a significant portion of telomeres to the NE , but rather is part of a functional LINC complex capable of transferring cytoplasmic forces required to move telomeres .
In recent years , it has been established by several groups that meiotic telomere attachment in mammals involves SUN1 and SUN2 as part of the NE spanning LINC complex connecting the meiotic telomeres to the cytoskeleton [11] , [21] , [22] . To analyze SUN1 function , two independent SUN1 deficient mouse models have been generated so far ( here referred to as Sun1 ( Δex10-13 ) [11] and Sun1 ( Δex10-11 ) [21] ) , which both revealed a virtually identical , exclusively meiotic phenotype: both male and female SUN1 deficient mice showed severe meiotic defects , which were ascribed to massive problems in meiotic telomere attachment [11] , [21] . Although SUN2 compensates for the loss of SUN1 in somatic cells , SUN2 overtly does not have the competence to counterbalance loss of SUN1 in meiocytes , and hence it was described that telomere attachment is prevented in Sun1−/− ( Δex10-13 ) mice [11] . Since we have previously found SUN2 expressed in meiocytes , where it localizes to the sites of telomere attachment [22] , this raises the question of the real function of SUN2 in meiosis . To investigate the actual role of SUN2 during meiosis , we therefore initiated a detailed analysis of telomere attachment in SUN1 deficient meiocytes and started off with spermatocytes and oocytes from Sun1−/− ( Δex10-11 ) mice , which were previously demonstrated to be SUN1 deficient [21] . Worth mentioning , using antibodies recognizing an epitope encoded by exons 13 to 14 [27] we could confim that these mice in fact do not express a functional SUN1 protein ( data not shown ) . To study telomere behavior in SUN1 deficient mice , we combined telomere fluorescence in-situ hybridization with immunocytochemical labeling of the lamina and the synaptonemal complexes in spermatocytes and oocytes of SUN1 knockout and wildtype littermate mice ( Figure 1A ) . As expected , in wildtype spermatocytes and oocytes all telomere signals that are clearly associated with the ends of synaptonemal complexes , are embedded within the lamina ( Figure 1A and A″ ) . Consistent with the previously published results [11] , [21] , we found that telomere attachment to the nuclear envelope is significantly disturbed in SUN1 deficient meiocytes ( Figure 1A′ and A′″ ) . This is evident from telomere signals located in the nuclear interior , in significant distance to the NE . However , within the same meiocytes , we found that numerous telomere signals were still embedded within the lamina ( arrowheads in A′ and A′″ ) , indicating that in the absence of SUN1 telomere attachment may not be entirely lost , but only reduced . The unexpected high numbers of peripheral , nuclear envelope associated telomere signals that were observed in both spermatocytes and oocytes of Sun1−/− ( Δex10-11 ) mice ( see below ) gave the impression that at least a portion of the peripheral telomeres might be structurally anchored at the nuclear envelope , which would clearly contradict the previous notion that loss of SUN1 completely prevents telomere attachment [11] . To clarify whether these telomeres are truly attached or merely located in close vicinity to the NE , we therefore prepared testis tissue and ovary samples for electron microscopy , as both synaptonemal complexes and sites of telomere attachment can easily be detected in electron micrographs ( Figure 1B–B′ , C–C′ ) . To affirm that putative attachment does not depend on the knockout genotype , we analyzed samples from both currently available SUN1 deficient strains , Sun1 ( Δex10-13 ) [11] and Sun1 ( Δex10-11 ) [21] . As anticipated , fully synapsed stretches of synaptonemal complexes attached to the nuclear envelope were clearly evident in all control samples of pachytene spermatocytes and oocytes . Remarkably , oocytes and spermatocytes from both SUN1 deficient mouse strains revealed similar telomere attachment sites to the ones observed in the wildtype ( Figure 1B″–B′″ , C″–C′″ ) . Although many homologous chromosomes in both Sun1−/− mice strains fail to pair and synapse during pachynema [11] , [21] , partially completed synaptonemal complexes are still present in pachytene-like staged meiocytes . When these are tethered to the NE , wildtype-like attachment sites seem to be able to form . Together , the immunocytochemical ( see below ) and electron micrograph data show that telomere attachment is not completely abolished during meiosis in mice lacking SUN1 , irrespective of the genetic targeting strategy used to create the SUN1 deficient mouse strain . Together , our findings presented here in fact proved that even in the absence of SUN1 a subset of meiotic telomeres is still able to attach to the NE , and thus our results refute the previous assumption regarding the lack of telomere attachment in SUN1 deficient mice [11] . Particularly the use of electron microscopic analysis on SUN1 deficient meiocytes has revealed some of the phenotypic features , which have been overtly overlooked before . To define the percentage of attached telomeres in Sun1−/− ( Δex10-11 ) spermatocytes we quantified the number of attached and non-attached telomeres in 3 dimensionally preserved nuclei of cells , simultaneously labeled for the nuclear lamina , the synaptonemal complexes and telomeres . To evaluate further whether the absence of SUN1 impacts telomere attachment in a stage dependent manner during meiotic progression , we additionally quantified and compared telomere attachment in spermatocytes at early leptonema , zygonema and at pachynema . For this we prepared tissue samples of wildtype and knockout littermates aged 12 and 14 days post partum ( dpp ) . As in the first wave of spermatogenesis development of spermatocytes within the seminiferous tubules is nearly synchronized [28] , at 12 dpp most spermatocytes within the tubules could be found at early leptonema to early zygonema . In tubules where early leptotene spermatocytes predominated , telomere attachment was not complete in both wildtype and knockout spermatocytes , probably due to the very early meiotic stage ( 77 . 7% and 64 . 5% attached telomeres in wildtype and knockout , respectively; Figure 1 D; Figure S1 ) . In tubules where early zygotene spermatocytes were accumulated all wildtype spermatocytes showed complete telomere attachment , whereas in knockout zygotene spermatocytes not more than 71 . 2% of all telomeres appeared to be NE-attached ( Figure 1 D′; Figure S1 ) . We observed similar rates of telomere attachment in spermatocytes of 14 dpp mice , where pachytene stages predominated . Here , wildtype spermatocytes again showed complete attachment of all telomeres , whereas Sun1−/− ( Δex10-11 ) males only showed 69 . 8% of telomeres attached to the NE ( Figure 1 D″ ) . These results implicate that the process of telomere attachment is induced despite SUN1 deficiency , yet full telomere attachment is never reached . Almost equivalent rates of attachment could be detected in zygotene and pachytene spermatocytes of Sun1−/− ( Δex10-11 ) mice , suggesting that once telomeres succeed to attach they maintain their association with the NE throughout prophase I , even in the absence of SUN1 . This indicates that attachment of telomeres to the NE without SUN1 is stable enough to withstand potential mechanical forces generated by the chromatin or cytoskeleton . The unexpected , relatively large proportion of telomeres that , without SUN1 , are still capable of stably attaching to the NE clearly points towards the existence of a partially redundant and SUN1-independent attachment mechanism . Very recently , it has been described that meiotic tethering of telomeres to the cytoskeleton is mediated by the novel meiosis-specific KASH-protein KASH5 [23] , [24] . To clarify whether KASH5 is also involved in the attachment of telomeres in SUN1 deficient meiocytes , we conducted immunofluorescence experiments labeling KASH5 and SYCP3 , a major component of the lateral elements of synaptonemal complexes [29] , in wildtype and SUN1 knockout spermatocytes . Consistent with earlier reports [23] , [24] , strong KASH5 foci at the ends of synaptonemal complexes were detected in all wildtype pachytene spermatocytes ( Figure 2 A ) , labeling telomeres attached to the NE . However , in contradiction to earlier reports [23] , [24] , in our hands KASH5 foci were also consistently present in Sun1−/− ( Δex10-11 ) spermatocytes in several independent experiments and different animals tested ( Figure 2 A′ , A″ ) . Although significantly weaker than in the wildtype tissue , the KASH5 signals in SUN1 deficient meiocytes nevertheless showed a wildtype-like distribution . KASH5 in the SUN1 deficient spermatocytes was found to be localized just at those ends of synaptonemal complexes that are in close contact with the NE . These experiments again corroborate that in the absence of SUN1 the remaining NE-associated telomeres are indeed attached to the NE . Beyond this , the attached telomeres are connected to the cytoskeleton through a linkage that involves KASH5 . In an earlier publication [22] we were able to demonstrate that SUN2 is expressed throughout meiotic prophase I , where it co-localizes with attached telomeres in wildtype mice . Therefore , it is tempting to speculate that telomere attachment in the absence of SUN1 is mediated by SUN2 . To follow up on this , we generated SUN2 specific antibodies and used these in co-immunolocalisation experiments together with antibodies against SYCP3 . Consistent with our previous results , our newly generated antibodies produced the already reported SUN2 foci at the end of synaptonemal complex axes in both wildtype spermatocytes and oocytes ( Figure 3 A , B; [22] ) . Similar to the wildtype situation , SUN2 foci of comparable intensities were also present in spermatocytes and oocytes of different meiotic prophase stages from Sun1−/− ( Δex10-11 ) mice ( Figure 3 A′ , A″ , B′ , B″ ) . This again demonstrates that SUN2 is indeed located at meiotic telomeres . As SUN2 is the only SUN domain protein expressed in Sun1−/− meiocytes , it appears likely that it is in fact SUN2 that mediates the observed telomere attachment in the SUN1 deficient mice . To further investigate attachment of telomeres in the Sun1−/− ( Δex10-11 ) mice , in particular with regard to possible KASH protein partners , we conducted co-immunostaining experiments using KASH5 and SUN2 antibodies on paraffin testis sections from mice of different ages ( 12 dpp and adult ) ( Figure 3 C , C′ ) . Clearly , as anticipated for a functional meiotic LINC-complex , the KASH5 and SUN2 foci in the Sun1−/− ( Δex10-11 ) spermatocytes co-localized , labeling those telomeres that are attached to the NE in the absence of SUN1 . In summary , these results indicate that the SUN2 localization to meiotic telomeres can occur independently of SUN1 , which is in accordance with the previous reports of unchanged SUN2 localization in somatic nuclei of Sun1−/− mice [26] . Furthermore , by means of the results presented here , SUN2 appears to be , at least to some extent , sufficient for meiotic telomere attachment to the NE . Regarding its possible interaction with KASH5 , yeast-two-hybrid studies have previously shown that the KASH domain of KASH5 in effect is able to interact with both the C-terminal domain of SUN1 as well as of SUN2 [23] . This , in combination with our results , leads us to the conclusion that SUN2 may also form functional meiotic LINC complexes with KASH5 in vivo , which , at least in the absence of SUN1 , is able to tether meiotic telomeres to the NE . In a recent crystallography study investigating LINC complex structure , SUN and KASH domains were shown to interact as two sets of trimeric protein complexes [30] . Furthermore , several groups have proposed SUN1 and SUN2 to form hetero-multimeric complexes [31] , [32] . Taking into account that SUN2 is expressed during meiosis ( present study , [22] ) , sharing its localization with SUN1 and KASH5 , it is tempting to speculate that during wildtype meiosis SUN1 and SUN2 assembly heterotrimeric complexes that interact with KASH5 to form meiotic LINC complexes required for efficiently tethering telomeres to the NE . In the absence of SUN1 , such LINC complexes may only be composed of SUN2 and KASH5 , still tethering telomeres to the NE , yet in a less effective manner than a complete heterotrimeric SUN1/SUN2- KASH5 complex . This could then explain the only partially disturbed telomere attachment observed in both SUN1 deficient mouse models . In addition , our results presented here suggest at least partial redundancy between SUN1 and SUN2 in meiotic telomere attachment , consistent with what has been reported for nuclear anchorage in somatic cells [25] , [26] . Prophase I of meiosis is not only characterized by the stable association of telomeres with the NE , but also by directed telomere-led chromatin movements leading to the formation and release of the bouquet stage [8] . Because SUN1 seems to be , at least partially , dispensable for the formation of a meiotic LINC complex per se , we asked whether those telomeres , which attach to the NE despite the absence of SUN1 , are still able to move along and to cluster within the NE . To analyze the distribution of the attached telomeres in the Sun1−/− ( Δex10-11 ) mice , we used KASH5 and SYCP3 antibodies for labeling attached telomeres in relation to synaptonemal complexes in spermatocytes of wildtype and knockout siblings at 12 dpp ( Figure 4 ) . At this age , leptotene/zygotene stages showing clustered telomere patterns normally predominate within the synchronously maturing tubules . To define KASH5 distribution within the NE , we performed 3D reconstructions of single spermatocyte nuclei of wildtype ( n = 50 cells ) and knockout ( n = 64 cells ) mice . Spermatocytes showing typically clustered KASH5 patterns resembling bouquet-like conformations of the attached telomeres could be detected in both wildtype and SUN1 knockout siblings ( Figure 4 and Supplementary Video S1 ) . Further quantifications with respect to the appearance of clustered versus non-clustered KASH5 patterns revealed that at 12 dpp bouquet frequencies were similar and statistically indifferent between wildtype and Sun1−/− ( Δex10-11 ) siblings ( 70% and 79 . 6% , respectively; p-value 0 . 23 Pearson's chi square test ) . These analyses demonstrated that the remaining attached telomeres in SUN1 deficient males in fact are able to form bouquet-like clustered telomere patterns and that this is not a rare event but occurs at similar rates as in the wildtype siblings . It is noteworthy , that we never observed a real clustering of the internal non-attached telomeres in Sun1 deficient spermatocytes . Taken together , we conclude from this that telomeres need to be attached to the NE , likely connected to the cytoskeleton , to form bouquet-like clusters . In Smc1ß−/− mice [33] , another knockout mouse model where telomere attachment is partially disrupted , bouquet formation of attached telomeres was observed in knockout spermatocytes as well , although at reduced levels compared to the wildtype . Regarding this study and our results , it seems conceivable that completed telomere attachment per se is not an essential prerequisite for telomere clustering . Rather , any telomere which is attached to the NE by a LINC complex has the competence to move within the NE and to proceed to cluster formation . To investigate the impact of the residual telomere attachment and movement on progression of meiotic recombination events , we started a next series of experiments to analyze oocytes of wildtype and Sun1−/− ( Δex10-11 ) female mice aged 19 . 5 dpf ( days post fertilization ) for the appearance of late recombination events . Using antibodies against MLH1 , SYCP1 and SYCP3 together on chromosome spreads allowed us to simultaneously investigate late recombination events and the state of synapsis formation . As expected , we observed the expected one to two MLH1 foci per each synapsed chromosome pair on chromosome spreads of the heterozygous control oocytes ( Figure 5 A ) . Consistent with previous reports [11] , [21] , oocyte spreads from littermate Sun1−/− ( Δex10-11 ) mice ( Figure 5B , C ) showed large numbers of unpaired or incorrectly paired chromosome axes stained by SYCP3 , but not by SYCP1 . Despite these severe synapsis defects , MLH1 foci were not completely absent from Sun1−/− ( Δex10-11 ) oocyte spreads . Instead , a small number of homologous chromosomes in Sun1−/− ( Δex10-11 ) oocytes were apparently able to achieve intact synapsis as shown by the complete co-localization of SYCP1 and SYCP3 . Distinct MLH1 foci on these fully paired homologs show that they in effect were able to recruit MLH1 to their axis , thus forming cross-over sites . These results indicate that in the absence of SUN1 , the remaining attached telomeres and their directed movements within the NE are sufficient to allow at least partial pairing , synapsis and cross-over formation during later meiosis in females . Therefore , when attachment is effectually reached , this attachment per se and the following movement of the attached telomeres appear to be functional , at least to some extent , even without SUN1 . In conclusion , from our current study it has become evident , that although SUN1 is essential for the efficient attachment of telomeres to the NE , SUN2 also appears to be involved in the tethering of meiotic telomeres to the NE . In the absence of SUN1 , an unexpectedly large proportion of telomeres are still able to attach to the NE and , beyond this , are also able to move within the NE , forming bouquet-like clustered telomere patterns . This suggests that in the SUN1 deficient background some of the telomeres not only succeed to establish a tight connection to the NE , but even become linked to the cytoskeletal motor system . Consistent with this , in the SUN1 deficient meiocytes we found KASH5 , which interacts with cytoplasmic dynein–dynactin [23] , [24] , co-localizing with SUN2 at sites where telomeres are in contact with the NE . In a very recent study , Horn and colleagues [24] have shown that in mice deficient for KASH5 , homolog pairing , synapsis and recombination is severely disturbed . In addition , they never observed clustering of SUN1 foci in KASH5 deficient cells , indicating that KASH5 as the ONM component of meiotic LINC complexes is required for transferring forces to move the INM located SUN proteins and therewith the attached telomeres [24] . Remarkably , the meiotic phenotype observed in the Kash5-null mice appeared much more dramatic than the phenotype induced by SUN1 deficiency . As shown by Horn and colleagues Kash5-null spermatocytes overtly never reach full synapsis not even of single pairs of homologous chromosomes , while in a considerable proportion of Sun1-null spermatocytes full synaptic pairing of at least a subset of homologs could be observed [11] , [24] . This is consistent with our results demonstrating that attached telomeres in SUN1 deficient mice in effect are able to cluster , most likely mediated by a restricted LINC complex formed by KASH5 and SUN2 , hence supporting synapsis and recombination . To date , no mammalian model has been described where meiotic telomere attachment is completely lost . Instead there are a number of phenotypes with more or less severe partial telomere attachment defects , similar to the Sun1−/− phenotype described here [33] , [34] . This is unlike the situation in yeast , for example , where bqt4 has been identified as a key player without which no meiotic telomeres attach to the NE at all [15] . The meiotic telomere attachment in mammals , however , seems to be regulated by a more complex , partially redundant network of factors , of which some of the central players await identification in the near future .
All animal care and experiments were conducted in accordance with the guidelines provided by the German Animal Welfare Act ( German Ministry of Agriculture , Health and Economic Cooperation ) . Animal housing and breeding at the University of Würzburg was approved by the regulatory agency of the city of Würzburg ( Reference ABD/OA/Tr; according to §11/1 No . 1 of the German Animal Welfare Act ) . All aspects of the mouse work were carried out following strict guidelines to ensure careful , consistent and ethical handling of mice . Tissues used in this study were derived from wildtype , heterozygous and knockout littermates of either of the two currently existing SUN1 deficient mouse strains , Sun1 ( Δex10-13 ) and Sun1 ( Δex10-11 ) [11] , [21] . For immunofluorescence studies testes and ovaries from wildtype , heterozygous and SUN1 knockout progeny of the Sun1 ( Δex10-11 ) strain were fixed for 3 hrs in 1% PBS-buffered formaldehyde ( pH 7 . 4 ) . Tissues were then dehydrated in an increasing ethanol series , infiltrated with paraffin wax at 58°C overnight and embedded in fresh paraffin wax as described in Link et al . [13] . For EM analysis we prepared tissue material from wildtype , heterozygous and SUN1 deficient mice from both SUN1 deficient mouse strains , the Sun1 ( Δex10-13 ) and Sun1 ( Δex10-11 ) strain , according to the protocol described below . For the generation of SUN2 specific antibodies , a His-tagged SUN2 fusion construct ( amino acids 248–469 of the SUN2 protein ) was expressed in E . coli RosettaBlue ( Novagen , Darmstadt , Germany ) and purified through Ni-NTA agarose columns ( Qiagen , Düsseldorf , Germany ) . This peptide was used for immunization of a guinea pig ( Seqlab , Göttingen , Germany ) . The serum obtained was affinity purified against the SUN2 antigen coupled to a HiTrap NHS-activated HP column ( GE Healthcare , Munich , Germany ) . Similarly , for the generation of a KASH5 specific antibody , a His-tagged KASH5-fusion construct ( amino acids 421–612 ) was expressed and purified as described above . This peptide was used for immunization of a rabbit and the serum obtained was purified using a KASH5 antigen coupled HiTrap NHS-activated HP column . Further primary antibodies used in this study were: goat anti-Lamin B antibody ( Santa Cruz Biotechnology , Heidelberg , Germany ) , rabbit anti-SYCP3 antibody ( anti-Scp3; Novus Biologicals , Littleton , CO ) , guinea pig anti-SUN1 antibody [27] and mouse anti-KASH5 [23] . For TeloFISH analyses we further used monoclonal mouse anti-digoxigenin antibodies ( Roche , Mannheim , Germany ) . Corresponding secondary antibodies used for this study were: Cy2 anti-mouse , texas red anti-mouse , alexa647 anti-rabbit , texas red anti-rabbit , Cy2 anti-guinea pig and texas red anti-goat; all obtained from Dianova ( Hamburg , Germany ) and used as suggested by the manufacturer . Double-label immunofluorescence analyses were carried out on paraffin sections of testis or ovary tissue ( 3–7 µm ) as described in [13] , [27] . Paraffin sections were prepared for immunofluorescence by first removing the paraffin by two consecutive incubations of 10 min each in Roti-Histol ( Carl Roth , Karlsruhe , Germany ) . Then the tissue sections were rehydrated in a decreasing ethanol series . Subsequently , antigen retrieval was conducted by incubating the slides in antigen unmasking solution ( Vector laboratories , Burlingame , CA ) at 125°C and 1 . 5 bar for 7–20 min . After permeabilization of the tissue in PBS containing 0 . 1% Triton X-100 for 10 min and washing in PBS , slides were blocked for 30 min in blocking solution ( 5% milk , 5% FCS , 1 mM PMSF; pH 7 . 4 in PBS ) . After incubation with the first primary antibody either for 2 hrs at room temperature or overnight at 4°C , slides were washed in PBS and again blocked in blocking solution before incubating the samples with the second primary antibody for another 2 hrs at room temperature . Following two washing steps ( 10 min each ) in PBS and reblocking for 30 min in blocking solution slides were incubated with the appropriate secondary antibodies . DNA was counterstained using Hoechst 33258 ( Sigma-Aldrich , Munich , Germany ) . To label telomeres and selected proteins simultaneously , we combined telomere fluorescence in situ hybridization ( TeloFISH ) with immunofluorescence protocols on paraffin sections as described previously [13] . Paraffin sections were rehydrated and antigen retrieval was conducted as described above . Prior to TeloFISH , cells were permeabilized with PBS/0 . 1% Triton X-100 for 10 min . After rinsing in 2× SSC ( 0 . 3M NaCl , 0 . 03M Na-citrate; pH 7 . 4 ) cells were denatured at 95°C for 20 min in 40 µl of hybridization solution ( 30% formamide , 10% dextrane sulphate , 250 µg/ml E . coli DNA in 2× SSC ) supplemented with 10 pmol digoxigenin-labeled ( TTAGGG ) 7/ ( CCCTAA ) 7 oligomeres . Hybridization was performed at 37°C overnight in a humid chamber . Slides were washed two times in 2× SSC at 37°C for 10 min each and blocked with 0 . 5% blocking-reagent ( Roche , Mannheim , Germany ) in TBS ( 150 mM NaCl , 10 mM Tris/HCl; pH 7 . 4 ) . Samples were incubated with mouse anti-digoxigenin antibodies ( Roche , Mannheim , Germany ) according to the manufacturer's protocol and bound antibodies detected with Cy2-conjugated anti-mouse secondary antibodies . Following the TeloFISH procedure , samples were prepared for immunofluorescence by blocking with PBT ( 0 . 15% BSA , 0 . 1% Tween 20 in PBS , pH 7 . 4 ) . Slides were incubated with the first primary antibody overnight , washed two times in PBS for 10 min each and incubated with the corresponding secondary antibody as described above . Finally , slides were again washed in PBS before incubating with the second primary antibody . After repeated washing in PBS samples once again were exposed to the corresponding secondary antibodies . DNA was counterstained using Hoechst 33258 ( Sigma-Aldrich , Munich , Germany ) . For electron microscopy , fresh tissue from testis and ovary was prepared as described in [22] . The tissues were fixed in 2 . 5% buffered glutaraldehyde solution ( 2 . 5% glutaraldehyde , 50 mM KCl , 2 . 5 mM MgCl , 50 mM cacodylate; pH 7 . 2 ) for 45 min and washed in cacodylate buffer ( 50 mM cacodylate , pH 7 . 2 ) . This was followed by incubation in 2% osmium tetroxide in 50 mM cacodylate at 0°C . The samples were then washed several times in water at 4°C and contrasted using 0 . 5% uranyl acetate in water at 4°C overnight . Subsequently , the tissues were dehydrated in an increasing ethanol series and incubated three times in propylene oxide for 30 min . Finally , the samples were embedded in epon for ultrathin sectioning . Fluorescence images were acquired using a confocal laser scanning microscope ( Leica TCS-SP2; Leica , Mannheim , Germany ) equipped with a 63x/1 . 40 HCX PL APO lbd . BL oil-immersion objective . Images shown are pseudo colored by the Leica TCS-SP2 confocal software and are calculated maximum projections of sequential single sections . These were processed using Adobe Photoshop ( Adobe Systems ) . 3D reconstructions , as well as analysis and quantification of telomere attachment and clustering were conducted using the ImageJ software ( version 1 . 42q; http://rsbweb . nih . gov/ij ) .
|
Correct genome haploidization during meiosis requires tightly regulated chromosome movements that follow a highly conserved choreography during prophase I . Errors in these movements cause subsequent meiotic defects , which typically lead to infertility . At the beginning of meiotic prophase , chromosome ends are tethered to the nuclear envelope ( NE ) . This attachment of telomeres appears to be mediated by well-conserved membrane spanning protein complexes within the NE ( LINC complexes ) . In mouse meiosis , the two main LINC components SUN1 and SUN2 were independently described to localize at the sites of telomere attachment . While SUN1 has been demonstrated to be critical for meiotic telomere attachment , the precise role of SUN2 in this context , however , has been discussed controversially in the field . Our current study was targeted to determine the factual capacity of SUN2 in telomere attachment and chromosome movements in SUN1 deficient mice . Remarkably , although telomere attachment is impaired in the absence of SUN1 , we could find a yet undescribed SUN1-independent telomere attachment , which presumably is mediated by SUN2 and KASH5 . This SUN2 mediated telomere attachment is stable throughout prophase I and functional in moving telomeres within the NE . Thus , our results clearly indicate that SUN1 and SUN2 , at least partially , fulfill redundant meiotic functions .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"meiosis",
"telomeres",
"cell",
"division",
"transmembrane",
"proteins",
"chromosome",
"biology",
"proteins",
"biology",
"molecular",
"cell",
"biology"
] |
2014
|
Analysis of Meiosis in SUN1 Deficient Mice Reveals a Distinct Role of SUN2 in Mammalian Meiotic LINC Complex Formation and Function
|
Wolbachia pipientis is an intracellular endosymbiont known to confer host resistance against RNA viruses in insects . However , the causal mechanism underlying this antiviral defense remains poorly understood . To this end , we have established a robust arthropod model system to study the tripartite interaction involving Sindbis virus and Wolbachia strain wMel within its native host , Drosophila melanogaster . By leveraging the power of Drosophila genetics and a parallel , highly tractable D . melanogaster derived JW18 cell culture system , we determined that in addition to reducing infectious virus production , Wolbachia negatively influences Sindbis virus particle infectivity . This is further accompanied by reductions in viral transcript and protein levels . Interestingly , unchanged ratio of proteins to viral RNA copies suggest that Wolbachia likely does not influence the translational efficiency of viral transcripts . Additionally , expression analyses of candidate host genes revealed D . melanogaster methyltransferase gene Mt2 as an induced host factor in the presence of Wolbachia . Further characterization of viral resistance in Wolbachia–infected flies lacking functional Mt2 revealed partial recovery of virus titer relative to wild-type , accompanied by complete restoration of viral RNA and protein levels , suggesting that Mt2 acts at the stage of viral genome replication . Finally , knockdown of Mt2 in Wolbachia uninfected JW18 cells resulted in increased virus infectivity , thus demonstrating its previously unknown role as an antiviral factor against Sindbis virus . In conclusion , our findings provide evidence supporting the role of Wolbachia–modulated host factors towards RNA virus resistance in arthropods , alongside establishing Mt2’s novel antiviral function against Sindbis virus in D . melanogaster .
Heritable symbioses are pervasive in nature and exceedingly common in the insect world , where many endosymbiotic associations have been described [1] . Wolbachia pipientis is an alpha-proteobacterial , maternally transmitted endosymbiont that invades insect host populations by manipulating host reproduction , favoring infected females [2] . It is present in approximately 40% of insects , including several species of Drosophila and important disease vectors such as Aedes albopictus and Culex species of common house mosquitoes [3] . In D . melanogaster , Wolbachia strain wMel exhibits weak reproductive manipulation [4] , but recent work has shown that this Wolbachia strain can also protect D . melanogaster from common viral pathogens such as Drosophila C virus ( DCV ) , cricket paralysis virus ( CrPV ) , and Flock House virus ( FHV ) , as evidenced by increased survival and delay in virus accumulation [5 , 6] . Not surprisingly , Wolbachia–mediated antiviral protection , so-called “pathogen blocking” is considered an exciting phenomenon that could be leveraged to reduce arthropod-transmitted diseases [7] . While A . albopictus infected with its native Wolbachia strain ( wAlbB ) has little to no effect on RNA virus replication ( such as Dengue ) , transinfection of non-native Wolbachia strains like wMel into this mosquito species , as well as the naturally uninfected Aedes aegypti , have been shown to induce pronounced antiviral resistance [8 , 9] . Moreover , field trials conducted as a part of the global Eliminate Dengue project have demonstrated that these transinfected A . aegypti mosquitos can invade native A . aegypti mosquito populations , persisting over the seasons [10 , 11] . However , although Wolbachia’s pathogen-blocking ability is currently being implemented in vector control , we know little to nothing about the mechanism underlying Wolbachia-induced antiviral resistance . Mosquitos as a model suffer from a lack of widely accessible genetic tools , coupled with the added complexity of introducing Wolbachia transinfections . Conversely , the antiviral phenotype was originally characterized in Drosophila and the genetic tractability of this model organism makes it an ideal candidate platform for a mechanistic dissection of the tripartite interaction [5 , 6] . To this end , we have leveraged the power of Drosophila genetics , combined with the molecular tractability of a parallel tissue culture system , to probe previously uncharacterized aspects of Wolbachia-mediated resistance against the prototype alphavirus , Sindbis virus ( SINV ) . In this study , we show that presence of Wolbachia in D . melanogaster results in reduced SINV infectivity , viral RNA replication and protein synthesis . Furthermore , we have identified the DNA/RNA methyltransferase gene , Mt2 as a potential host factor responsible for Wolbachia-mediated antiviral resistance . Based on these results , we conclude that resistance towards SINV occurs at an early stage of virus replication that further affects subsequent stages of the virus life cycle and show evidence supporting the hypothesis that Wolbachia modulates the expression of a host methyltransferase gene ( Mt2 ) to target virus RNA synthesis .
We investigated the effect of Wolbachia on SINV titer in wild-type flies infected or uninfected with Wolbachia strain wMel , which we refer to as Wolb+ and Wolb- respectively from this point onwards ( Fig 1A ) . We did not observe death as a consequence of virus infection in Wolb+ or Wolb- flies . This was expected due to the non-pathogenic nature of SINV infection in Drosophila . Flies were collected 48 hours post-injection and virus titer was determined using end-point dilution assay on BHK cells . We found that infectious virus titer was significantly reduced in the Wolb+ individuals compared to their Wolb- counterparts . We and others have previously shown that SINV infectivity , measured by the ratio of total virus particles produced to infectious units , is influenced by the host cell environment in which the virus is cultivated [12] . To determine whether the presence of Wolbachia inside the host cell changes SINV infectivity , we used a cell culture based system comprised of Drosophila melanogaster-derived JW18 cell line infected with Wolbachia strain wMel and an antibiotic-treated Wolbachia free control cell line , which we refer to as JW18-dox . Cells were infected with SINV at an MOI of 100 and samples of growth medium were taken at 48 , 72 and 96 hours post infection ( Fig 1B ) . We performed conventional end point dilution assay to quantify infectious particles released during infection and a previously described qRT-PCR based method to quantify the total copies of virus genome present in the growth medium over the course of the infection , which is equivalent to the total number of released virus particles [12] . We found that virus derived from Wolb+ cells had a significantly high particle: infectious unit ratio relative to virus derived from Wolb- cells , indicating that on a particle basis this virus was less infectious than that derived from the Wolb- cells ( Fig 1B ) . We have previously shown that SINV infectivity changes over the course of infection in both mammalian and mosquito cells , and also varies in a host cell dependent manner , i . e . virus grown in one cell line may be more infectious on a per particle basis than virus grown in a different cell line [12] . In the current study we found that infectivity of particles produced deteriorated over time during infection in the Wolb+ cells . These data show that not only does the presence of Wolbachia lead to a decrease in virus produced , but also a decrease the infectivity of the SINV particles produced during infection . During SINV infection , the incoming 49S genomic RNA functions as a mRNA to encode the viral replication complex and later acts as a template for the synthesis of minus-strand RNA . The minus strand RNA itself then serves as a template for the synthesis of nascent full length 49S genomic RNA and the transcription of smaller 26S sub genomic RNA , which encodes the viral structural proteins . Synthesis of these different viral RNA species and their subsequent translation is required for the formation of virus particles . Given that we found that the presence of Wolbachia resulted in fewer infectious virus progeny , we hypothesized that viral RNA synthesis is inhibited when Wolbachia is present at the time of SINV infection . To test this , wild-type Wolb+ and Wolb- flies were challenged with SINV . Infection was allowed to progress for 48 hours before the flies were collected and snap-frozen . Following tissue homogenization and RNA extraction , both relative , as well as absolute quantities of the viral RNAs were determined by qRT-PCR ( Fig 2 ) . We found that presence of Wolbachia is accompanied by 2 . 5-fold reductions in both viral genome ( quantified using primers to nsP1 coding sequence ) and subgenome ( quantified using primers to E1 coding sequence ) RNA levels , suggesting that viral RNA synthesis is inhibited in the presence of the endosymbiont ( Fig 2A ) . We subsequently repeated this experiment in our D . melanogaster derived cell culture system to further determine the absolute copies of plus strand and minus strand RNA produced during virus infection . JW18 and JW18-dox cells were infected with SINV at an MOI of 100 . Infection was allowed to last for 96 hours before cells were collected for lysis and subsequent RNA extraction and absolute quantities of viral RNAs were determined using qRT-PCR ( Fig 2B ) . In line with our previous observation in flies , we found an average 10-fold reduction in total copies of virus plus and minus strand RNA . Given our observed reduction in virus titer and RNA levels , we expected a concomitant reduction in virus protein synthesis . During infection , SINV non-structural proteins are translated as a polyprotein from the first open reading frame present in the 49S genomic RNA . Sometime later during infection , the 26S subgenomic RNA is translated as a polyprotein that subsequently gives rise to the virus structural proteins [13] . Therefore , expression of a luciferase reporter present within any non-structural or structural protein can be used as a proxy for the net translational activity from the genomic and subgenomic RNAs respectively . To this end , we used a luciferase based viral translation assay to quantitatively determine translation of SINV non-structural and structural proteins . We used SIN- nsP3-nLuc virus , in which nanoluciferase ( nLuc ) has been translationally fused to the hypervariable domain of SINV nsP3 protein , and SIN-cap-nLuc virus , which has nLuc translationally fused to the C-terminus of SINV capsid protein . Wild-type Wolb+ and Wolb- flies were infected independently with each of the aforementioned viruses . Flies were collected 48hr post-infection and snap frozen and nLuc activity was measured post-homogenization ( Fig 3A ) . Our results indicated an average 10-fold reduction in nLuc activity in tissue derived from Wolb+ individuals challenged with SIN-nsP3-nLuc virus , suggesting inhibition of SINV non-structural protein synthesis ( Fig 3A ) . In contrast , quantification of nLuc activity in tissues derived from individuals challenged with SIN-cap-nLuc virus revealed a greater ( ~50-fold ) Wolbachia–mediated reduction in nLuc activity ( Fig 3A ) . These data indicate that Wolbachia causes a reduction in the synthesis of SINV non-structural and structural proteins and that expression of viral proteins synthesized off the subgenome are affected to a greater extent . However , it was not clear whether this was a consequence of decreased translational efficiency or simply a consequence of reduced quantities viral RNA available for translation ( Fig 2 ) . To determine the cause of reduced viral protein synthesis , we extracted total RNA from the tissue homogenates of individuals infected with SIN-nsP3-nLuc or SIN-cap-nLuc viruses and determined absolute quantities of viral genomic and subgenomic RNAs using qRT-PCR as before . Following the calculation of viral protein , expressed in terms of luciferase activity ( RLU ) , to viral RNA ( copies/ug ) , we failed to find any significant difference in the protein-to-RNA ratios between Wolbachia–infected and—uninfected individuals , suggesting that viral protein translation is reduced due to scarcity of viral transcripts and that the smaller number of viral transcripts produced in the presence of the bacterium can be translated with regular efficiency ( Fig 3B ) . Taken together , our data strongly indicates that Wolbachia-mediated inhibition of SINV infection occurs at an early stage of viral RNA replication that subsequently results in reduced viral protein synthesis and virus titer , which is consistent with previous reports regarding SFV infection in JW18 cells [14] . Based on evidence from other systems , we investigated whether the antiviral resistance mediated by Wolbachia could be explained by modulation of host immune gene expression by the bacterium [15–19] . We tested this hypothesis by profiling the transcriptional activity of several candidate genes spanning different pathways that have been previously implicated as being a part of the host antiviral defense . The initial examination of the effect of Wolbachia on host gene expression was performed in the absence of a viral infection . We reasoned that it is important to consider the cellular environment that the virus is initially exposed to upon entry into the cell , not necessarily one that it encounters later during infection . In addition to canonical innate immune pathway components we examined the expression of Mt2 , a gene encoding a nucleic acid methyl transferase previously shown to be required for DCV resistance in Drosophila [20] . Host gene expression was determined by qRT-PCR in Drosophila in the absence or presence of Wolbachia . We observed a general increase in expression of genes associated with canonical innate immune pathways such as Imd and Toll in the presence of Wolbachia ( S1 Fig ) . Interestingly , we also observed a significant increase in the expression of the host methyl-transferase gene Mt2 , with an average of 7-8-fold increase in transcript levels in the presence of Wolbachia ( S1 Fig , Fig 4A ) . While a number of previous studies have discounted the role of canonical immune pathways in Wolbachia-induced pathogen blocking , the role of Mt2 in this process is less clear . Interestingly , this elevated expression of Mt2 in Wolb+ flies decreased following SINV infection ( Fig 4A ) . To investigate the role of Mt2 in Wolbachia–mediated inhibition of SINV infection , we looked at the effect of Wolbachia on virus infection in a previously characterized , homozygous loss-of-function mutant of Mt2 ( Mt2 -/- ) [21] . SINV infection was established as before in Wolbachia infected and uninfected flies that were either wild-type or Mt2 -/- and the infection progressed for 48 hours before samples were collected , snap-frozen and virus titer was determined using end-point dilution assay on BHK cells ( Fig 4B ) . SINV titer was , on average , 10-fold higher in Wolb+ Mt2 -/- mutants compared to Wolb+ wild-type flies . Previous studies have established a direct correlation between Wolbachia density and the degree of antiviral resistance [22–24] . In light of this we examined whether the loss of function in the Mt2 gene in the Mt2 -/- mutants is accompanied by a reduction in Wolbachia titer , which could potentially explain the observed loss in virus inhibition . Quantification of DNA collected from samples using qPCR showed no reduction in Wolbachia titer in the Mt2 -/- mutants compared to the wild-type , indicating that Wolbachia titer is not responsible for the observed reduction in virus inhibition ( S2 Fig ) . Similar results were obtained following shRNA-targeted knockdown of Mt2 expression in two genetically distinct Wolbachia–infected fly lines , with each of the two shRNAs targeting a different area of the Mt2 gene ( Fig 5A and 5B , S3 Fig ) . Interestingly , the degree to which virus resistance was lost seemed to correlate roughly with the extent to which Mt2 expression was reduced ( Fig 5A–5C , S3 Fig ) . Virus titer from Wolbachia uninfected wild-type flies were still found to be higher compared to the Wolbachia infected Mt2 -/- mutants ( Fig 4B ) although the data failed to meet the threshold for statistical significance . Additionally , comparison of virus titer from Mt2-/- flies with and without Wolbachia indicated that pathogen blocking is mediated by factors in addition to Mt2 . However , once again the difference observed failed to reach statistical significance . Given the fact that our initial characterization of Wolbachia–mediated SINV resistance showed viral inhibition occurred at an early stage of RNA synthesis and consequently resulted in reduced viral protein levels , we next sought to determine the effect of Mt2 at these particular stages of virus replication . Quantification of SINV transcripts was performed using qRT-PCR as described above for absolute viral genome quantification ( Fig 4C ) . Compared to Wolbachia–infected wild-type individuals , loss of Mt2 resulted in an average of 2-3-fold increase in viral transcript levels . In contrast , we did not see any significant difference in the transcript levels between Wolbachia infected Mt2 -/- mutants compared to uninfected wild-type and Mt2 -/- mutants , showing that SINV RNA synthesis can be almost fully restored following the loss of Mt2 function in the presence of Wolbachia . While the data presented above indicated that inhibition at the stage of SINV RNA synthesis previously observed in the presence of Wolbachia is significantly reduced in Mt2 -/- mutant , it did not answer whether or not viral protein levels were restored in these mutants . We utilized our previously described SIN-cap-nLuc virus and luciferase based viral translation assay to quantify levels of SINV structural proteins in wild-type and Mt2 -/- mutants either in the presence or absence of Wolbachia . We found significantly higher expression of SINV structural proteins in the Mt2 -/- background ( Wolb+ and Wolb- ) , relative to that observed in their wild-type counterparts ( Fig 4D ) . This result implies that the antiviral effect of Mt2 targets a stage of viral RNA synthesis , with the loss of Mt2 restoring the RNA and consequent protein levels to wild-type Wolb- levels . We next examined whether the antiviral effect of Mt2 against SINV could be independent of Wolbachia , as it has been reported before to be antiviral in the context of native DCV infection in D . melanogaster [20] . To test whether overexpression of the Mt2 gene by itself resulted in SINV resistance , we used Gal4-UAS expression system to overexpress Mt2 in Wolbachia uninfected flies . While we were only able to achieve modest levels of overexpression , following infection with SINV , Mt2 overexpressing individuals were found to accumulate virus at an average of 2-fold lower titer than their wild-type sibling controls ( S4A Fig ) . Levels of viral RNA in Mt2 overexpressing flies were , on average , 2 . 5 fold lower compared to wild-type sibling controls ( S4B Fig ) . To further investigate the antiviral role of Mt2 against SINV , we asked whether Mt2 is responsible for our initial observation regarding the reduction of virus particle infectivity in the presence of Wolbachia . To this end , we utilized our Wolbachia uninfected JW18-dox cell culture model to knock down expression of Mt2 using targeted dsRNA against the methyltransferase gene . Cells were transfected with either Mt2-specific , or non-targeting dsRNA . SINV infection was established 48-hours post-transfection at an MOI of 100 and infection was allowed to last for 96 hours before cells and media were harvested separately . Following RNA extraction from harvested cells , qRT-PCR based quantification of Mt2 gene expression revealed an average 50% knockdown relative to non-targeting controls ( Fig 6A ) . Probing the effect of Mt2 knockdown on SINV RNA synthesis revealed viral RNA levels to be around 14-fold higher in cells transfected with Mt2 dsRNA ( Fig 6B ) , further confirming the results obtained in our animal model ( Fig 4C ) . Effect of Mt2 knockdown on SINV particle infectivity in JW18-dox cells was performed as described previously by determining the ratio of total particles to the total number of infectious units . Virus titer was calculated using standard end-point dilution assay on BHK-21 cells . Relative fold change in virus titer from cells transfected with Mt2 dsRNA was found to be on average , 200-fold higher compared to cells treated with non-specific dsRNA ( Fig 6C ) . Consequently , virus particle infectivity was found to increase for virus derived from cells in which Mt2 was knocked-down as evidenced by the 70 percent reduction in the particle-to-TCID50 ratio , indicating that Mt2 plays a role in regulating virus particle infectivity ( Fig 6D ) . Taken together , these results show Mt2 to possess antiviral activity against SINV .
Continued use of Wolbachia as an emerging vector control agent requires better understanding of the molecular mechanism underlying its antiviral resistance in the context of a tractable arthropod system . We provide evidence that the presence of Wolbachia reduces SINV particle infectivity and results in reduced virus titer . Furthermore , our data indicate a reduction in viral RNA synthesis , accompanied by decreased viral protein synthesis . Further characterization of early infection events is required to determine the exact nature of Wolbachia–mediated inhibition of SINV genome replication . Importantly , our data clearly demonstrate antiviral activity of the host RNA methyltransferase gene Mt2 in the presence and absence of Wolbachia . Given the data provided in this study , it is likely that the effect of Mt2 lies at the stage of viral RNA synthesis/stability . Future work will therefore focus on characterizing the nature of interaction between Mt2 and SINV RNA . The fact that both Wolbachia mediated antiviral resistance and the antiviral effect of Mt2 extends to members of different single stranded RNA virus families is significant , suggesting that existence of a common mechanism [5 , 6 , 9–11 , 14 , 20 , 22] .
Drosophila fly stock 6326 ( of a W1118 genetic background ) , carrying Wolbachia was obtained from the Bloomington Stock Center ( BDSC ) , and used as the infected wild-type background . A corresponding Wolbachia–uninfected line was created through tetracycline treatment ( approx . 20 ug/mL fed in the fly media for 3 generations ) . Wolbachia infection status was subsequently confirmed through quantitative PCR using published primer sets ( S1 Table ) [58] . The flies were repopulated with a wild-type microbiota post tetracycline treatment through culture in bottles previously occupied by untreated male flies of the same background ( stock 6326 ) . UAS-Mt2 and Mt2 loss of function flies ( provided by S . Bordenstein ) [54] were used to examine the role of the Drosophila DNA methyltransferase gene on the pathogen blocking phenotype . The Mt2 loss of function mutation is in a W1118 background as described by LePage et al . [54] . Wolbachia-infected TRiP mutant stocks 38224 ( y1 sc* v1; P {TRiP . HMS01667} attP40 ) and 42906 ( y1 sc* v1; P {TRiP . HMS02599} attP40 ) were used for shRNA-targeted knock-down of Mt2 gene expression by driving Mt2 shRNA expression using previously described Act5C-Gal4 driver males ( provided by Brian Calvi ) y1 w*; P{Act5C-GAL4}25FO1/CyO , y+ ( Fig 5 ) or y1 w*; P{w[Act5C-GAL4}17bFO1/TM6B , Tb1 ( S3 Fig ) . For overexpression of Mt2 , crosses were performed between UAS-Mt2 males and virgin Act5C-Gal4 driver females . Among the resulting progeny , straight-winged flies were considered to be overexpressing while siblings exhibiting Cyo phenotype were used as the wild-type control . All fly stocks were and maintained on standard cornmeal-agar medium supplemented with P/S at 25°C on a 24-hour light/dark cycle . In order to establish a systemic virus infection in vivo , flies were anesthetized with CO2 and injected in the thorax with 50nL of approximately 1010 PFU/mL of pelleted virus or control PBS using a glass capillary needle . Flies were collected two days post-infection , snap-frozen in liquid N2 and stored at -80°C for downstream processing . In all stocks harboring Wolbachia infection , the Wolbachia strain was wMel2 , confirmed by genotyping as shown in S5 Fig using primers described in Riegler et . al . 2005 [59] ( S1 Table ) . Wolbachia-infected Drosophila melanogaster JW18 cells and corresponding doxycycline-treated Wolbachia-uninfected JW18_dox cells ( a generous gift from W . Sullivan ) were maintained at 24°C in Shields and Sang media ( Sigma ) , supplemented with 10% heat-inactivated fetal bovine serum ( FBS ) and 1% Antibiotic-Antimycotic ( Gibco ) . BHK-21 cells ( American Type Culture Collection ) were grown at 37°C under 5% CO2 in MEM ( CellGro ) supplemented with 1% L-Gln , 1% Antibiotic-Antimycotic ( Gibco ) , 1% non-essential amino acids and 10% heat inactivated Fetal Bovine Serum ( FBS ) . JW18 cells were seeded in a 6-well plate at a density of approximately 2 . 26x106 cells/well 24h prior to infection . Infection was carried out using virus derived from BHK-21 cells , titered using a standard plaque assay . Virus was diluted in Shields and Sang media ( Sigma ) and the infection was established at an MOI = 100 . For doxycycline cleared JW18-dox cells , around 50–55% infection was observed at 96 hpi . Mock infections were carried out by treating the cells equivalently , without the addition of virus . Quantification of viral genome and subgenome translation was performed by introducing SINV luciferase reporter viruses nsP3-nLuc and cap-nLuc , respectively into 2-day old virgin female flies as described . The samples were homogenized in 1X Cell Culture Lysis Reagent ( Promega ) after they were collected 2-days post infection and clarified via centrifugation at 16 , 000 × g for 2 min . The samples were then mixed with luciferase reagent ( Promega ) , and luminescence was recorded using a Synergy H1 microplate reader ( BioTech instruments ) . In all cases , the luciferase readings were normalized to the levels of SINV genomic and sub-genomic RNA present in the assayed samples , determined by using qRT-PCR methods with ΔΔCT calculation of transcript levels . Wolbachia density from fly homogenates and tissue culture cells were determined via qPCR on whole DNA using an Applied Biosystems StepOne Real-time PCR system ( S1 Table ) and SYBRGreen Chemistry ( Applied Biosystems ) , previously described in [58] . Wolbachia density ( and infection status ) in JW18 cells was determined using DAPI staining , where Wolbachia were visualized in the form of cytoplasmic foci within the cells . Knockdown of Mt2 expression was achieved in doxycycline treated JW18 cells using target dsRNA against the Mt2 gene . Mt2 dsRNA was synthesized from corresponding dsDNA generated using self-annealing primer sets . First , custom oligonucleotide was designed with a 5’ T7-Polymerase binding site ( GAATTAATACGACTCACTATAG ) followed by a 3’ target sequence specific to the Mt2 coding region . Similarly , another oligonucleotide was designed with a similar 5’ T7 polymerase binding site followed by a 3’ end complementary to the 3’ end of the previous oligo ( S1 Table ) . Polymerase chain reaction was carried out using 100uM of primers and Q5 High-fidelity Master Mix [NEB] to produce dsDNA . dsRNA was synthesized via in-vitro transcription of this dsDNA , using T7-RNA Polymerase in the presence of the 5'cap analog 7'G5'ppp5'G [New England Biolabs] , followed by transfection of 500 ng dsRNA into JW18-dox cells using Lipofectamine LTX [Thermo Fisher Scientific] . Maximum knockdown was achieved at 48 hours post transfection . Whole flies were homogenized in TRIzol reagent , followed by RNA extraction . cDNA was synthesized using MMulV Reverse Transcriptase ( New England Biolab ) with random hexamer primers ( Integrated DNA Technologies ) . Negative ( no RT ) controls were performed for each target . Quantitative RT-PCR analyses were performed using Brilliant III SYBR green QPCR master mix ( Agilent ) with gene-specific primers according to the manufacturer's protocol and with the Applied Bioscience StepOnePlus qRT-PCR machine ( Life Technologies ) . The expression levels were normalized to the endogenous 18S rRNA expression using the delta-delta comparative threshold method ( ΔΔCT ) . Fold changes were determined using the comparative threshold cycle ( CT ) method ( S1 Table ) . P-values were calculated as described in the individual Fig legends . The average fold change ( FC ) in each experiment was calculated using the variable bootstrapping method , measuring the fold change between each potential pair of flies to determine the variability of the mean [12] . 95% confidence intervals ( CI ) were calculated using one sample t-test of log2FC values to determine the significance of distribution of the mean relative to the null using IBM SPSS Statistics Software 24 [60] .
|
Effective vector control is critically important to reduce the incidence of diseases caused by arthropod transmitted viruses . One proposed strategy involves the use of endosymbiotic bacteria Wolbachia pipientis as a novel biocontrol agent to prevent RNA virus transmission in mosquitoes . Previous work in the field suggests that the presence of this bacterium induces virus resistance within the host . However , the underlying mechanism of this antiviral phenotype is poorly understood , impeding its widespread use . Using the alphavirus , Sindbis as our model , we explored the tripartite interaction between the virus and the endosymbiont within its natural host , the fruit fly Drosophila melanogaster . In this study , we show that Wolbachia negatively influences multiple important aspects of the virus life cycle , extending our current understanding of the molecular nature of this interaction . We also provide evidence highlighting the role of a host gene , Mt2 , in Wolbachia–mediated antiviral resistance , while uncovering its previously unknown role as an antiviral host factor against Sindbis virus .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"nucleic",
"acid",
"synthesis",
"rna",
"extraction",
"microbiology",
"animals",
"wolbachia",
"animal",
"models",
"viruses",
"virus",
"effects",
"on",
"host",
"gene",
"expression",
"drosophila",
"melanogaster",
"model",
"organisms",
"rna",
"viruses",
"experimental",
"organism",
"systems",
"pharmacology",
"rna",
"synthesis",
"bacteria",
"drosophila",
"extraction",
"techniques",
"chemical",
"synthesis",
"research",
"and",
"analysis",
"methods",
"antimicrobial",
"resistance",
"viral",
"replication",
"insects",
"biosynthetic",
"techniques",
"arthropoda",
"biochemistry",
"rna",
"nucleic",
"acids",
"virology",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2017
|
Wolbachia elevates host methyltransferase expression to block an RNA virus early during infection
|
Identifying the genomic regions bound by sequence-specific regulatory factors is central both to deciphering the complex DNA cis-regulatory code that controls transcription in metazoans and to determining the range of genes that shape animal morphogenesis . We used whole-genome tiling arrays to map sequences bound in Drosophila melanogaster embryos by the six maternal and gap transcription factors that initiate anterior–posterior patterning . We find that these sequence-specific DNA binding proteins bind with quantitatively different specificities to highly overlapping sets of several thousand genomic regions in blastoderm embryos . Specific high- and moderate-affinity in vitro recognition sequences for each factor are enriched in bound regions . This enrichment , however , is not sufficient to explain the pattern of binding in vivo and varies in a context-dependent manner , demonstrating that higher-order rules must govern targeting of transcription factors . The more highly bound regions include all of the over 40 well-characterized enhancers known to respond to these factors as well as several hundred putative new cis-regulatory modules clustered near developmental regulators and other genes with patterned expression at this stage of embryogenesis . The new targets include most of the microRNAs ( miRNAs ) transcribed in the blastoderm , as well as all major zygotically transcribed dorsal–ventral patterning genes , whose expression we show to be quantitatively modulated by anterior–posterior factors . In addition to these highly bound regions , there are several thousand regions that are reproducibly bound at lower levels . However , these poorly bound regions are , collectively , far more distant from genes transcribed in the blastoderm than highly bound regions; are preferentially found in protein-coding sequences; and are less conserved than highly bound regions . Together these observations suggest that many of these poorly bound regions are not involved in early-embryonic transcriptional regulation , and a significant proportion may be nonfunctional . Surprisingly , for five of the six factors , their recognition sites are not unambiguously more constrained evolutionarily than the immediate flanking DNA , even in more highly bound and presumably functional regions , indicating that comparative DNA sequence analysis is limited in its ability to identify functional transcription factor targets .
Deciphering the transcriptional information contained in the extensive cis-acting sequences that direct intricate patterns of gene expression in animals is a major challenge in biology . Animal genomes encode several hundred ( e . g . , Drosophila ) to several thousand ( e . g . , human ) transcription factors [1–3] . These proteins mediate transcription by binding in a sequence-specific manner to cis-regulatory modules ( CRMs ) found throughout the nonprotein coding portions of animal genomes ( reviewed in [4–7] ) . The cells in which a CRM will activate or repress transcription of its target gene ( s ) are determined by the number , affinities , and arrangements of the transcription factor recognition sequences contained in the CRM , the expression patterns of the regulators that bind these sequences , and how the various factors interact . Animal sequence-specific regulators , however , generally recognize short , degenerate DNA sequences that occur frequently throughout the genome , and only a small subset of these predicted recognition sequences are thought to be functional targets of the transcription factor in vivo [4] . Because we do not yet understand the rules governing transcription factor binding or combinatorial interactions between factors , it is a major challenge to identify animal CRMs de novo or to predict how genes will be regulated based on their flanking DNA sequences alone . Closely linked to this challenge is the problem of understanding the control of morphogenesis by developmental regulatory networks . Animals are composed of complex three-dimensional arrays of cells whose movements , shape changes , divisions , and patterns of determination and differentiation are coordinated by master regulatory genes , many of which encode transcription factors . If we knew the range of genes directly controlled by these regulators , it would greatly aid studies of how they coordinate the complex processes of morphogenesis . To address these twin challenges , we have initiated an interdisciplinary analysis of the regulatory network controlling spatial patterning in the Drosophila melanogaster blastoderm embryo [8–11] . This network has been studied extensively , and we can be fairly confident that most of the major regulators have been identified [12 , 13] . Approximately 50 transcription factors are known to play a role in patterning the pregastrula embryo , forming a series of transcriptional cascades that regulate the formation of the anterior–posterior ( A-P ) and dorsal–ventral ( D-V ) axes . To decipher the combinatorial code by which transcription factors interact , it will be essential to have data for the great majority of factors in a system , and it should be possible to derive such comprehensive data for the early Drosophila network . In this system , A-P patterning is initially established by maternally controlled activity gradients of two transcription factors: Bicoid ( BCD ) , which has its highest activity in the anterior portion of the embryo and decays more posteriorly , and Caudal ( CAD ) , which has its highest activity in the posterior portion of the embryo and decays anteriorly ( Figure 1 ) . Amongst the earliest zygotically transcribed genes are four targets of BCD and CAD—hunchback ( hb ) , Krüppel ( Kr ) , knirps ( kni ) , and giant ( gt ) —the “gap” genes ( Figure 1 ) . These six genes encode transcription factors that work together to segment the A-P axis of the embryonic trunk ( a collection of additional regulatory factors are involved in patterning the head and tail ) [14–16] . For example , the second stripe of the pair-rule gene even-skipped ( eve ) is produced by the action of BCD , HB , KR , and GT on a CRM located approximately 1 . 5 kb upstream of the coding gene . In this case , BCD and HB act coordinately as activators in the same cells , whereas KR and GT each acts to repress expression in different parts of the embryo , restricting CRM output to a narrow stripe lying between the single band of KR expression and the anterior expression domain of GT [17 , 18] . This same collection of factors when bound to other CRMs produces different patterns of gene expression . The different combinations of recognition sequences in each CRM dictate the binding of factors in CRM-specific numbers and orientations . This binding , in turn , modulates the activity of each factor ( in some cases changing activators into repressors , in others leading to binding-site competition or cooperative interactions ) and produces a distinct transcriptional response [19–22] . Thus it is essential to study and model the action of these proteins in their native context . In this paper , we use chromatin immunoprecipitation ( ChIP ) and Affymetrix whole-genome tiling arrays to map the genomic DNA regions bound by these six factors in D . melanogaster embryos . Our results provide the most comprehensive in vivo DNA binding data for a set of cooperating transregulators specifying complex spatial patterns of expression in an animal . They provide a framework for ongoing efforts to decode transcriptional information in the genome and model developmental regulatory networks .
To identify the genomic regions bound in vivo by the gap and maternal factors controlling trunk segmentation , we adapted chromatin immunoprecipitation and microarray ( ChIP/chip ) methods [23 , 24] . Briefly , intact blastoderm embryos ( late stage 4 through stage 5 ) were treated with formaldehyde to crosslink proteins and DNA , after which chromatin was isolated , fragmented to an average length of 600 bp , and immunoprecipitated with antibodies recognizing the target protein [25] . The recovered material was amplified and hybridized to an Affymetrix whole-genome tiling array that contains over three million features representing 25-bp sequences spaced on average 35 bp apart across the unique portion of the D . melanogaster genome [26] . Our ChIP and DNA amplification protocols were optimized to maximize the signal-to-noise ratio , something that is especially critical in this system because these factors are only expressed at high levels in approximately 20% to 30% of cells ( Figure 1 ) . We also developed and optimized computational and statistical methods to provide an extensive , and accurate , high-resolution map of regions bound by each factor . Data were obtained using affinity-purified antibodies to KNI , KR , HB , GT , BCD , and CAD . In addition , to detect genes that are transcribed at this stage of development , further immunoprecipitations were performed using a monoclonal antibody recognizing the phosphorylated form of the C-terminal heptapeptide repeat of RNA polymerase II [27] . To reduce the possibility that the antibodies against gap and maternal factors might cross-react with proteins other than the one against which they were raised , we affinity purified all antisera against recombinant proteins engineered to remove amino acid sequences found in any other Drosophila proteins . For BCD , HB , KR , and KNI , we used two different antibody preparations that were independently purified against nonoverlapping epitopes; for CAD and GT , we were only able to obtain one set of purified antibodies per protein . For each purified antisera , two independent replicates of three different sample types were analyzed on separate arrays: ( 1 ) “Factor immunoprecipitates ( IPs ) ” obtained by immunoprecipitation using a factor-specific antibody; ( 2 ) “immunoglobulin G ( IgG ) control IPs” obtained by immunoprecipitation using a normal IgG antibody; and ( 3 ) “input DNA” obtained from the chromatin prior to immunoprecipitation , for a total of six arrays per antibody ( Figure 2A and 2B ) . To correct for the nonuniform hybridization response of the 25-bp oligonucleotides [28 , 29] , we divided the mean hybridization signal for each array element in the Factor IPs and IgG control IPs by the mean hybridization signal for the same feature in the input DNA ( Figure 2C ) . To further reduce noise , logarithms of these calculated oligonucleotide ratio scores were averaged ( throwing out the highest and lowest values to produce a “trimmed mean” ) in 675-bp windows ( approximately equal to the size of the immunoprecipitated fragments ) centered around each array element to give a hybridization window score ( Figure 2D ) . To determine which window scores represent significant enrichment in the Factor IP samples , we estimated false-discovery rates ( FDR; the fraction of windows with equal or greater scores that are not detectable enriched in the IP ) using two separate methods ( Figure 2F and 2G ) . One method determined empirically the distribution of scores for unenriched windows using the distribution of window scores in the IgG control IP ( Figure 2G; Materials and Methods ) ; the other method estimated this distribution directly from the Factor IP using a “symmetric null distribution” method [30] ( Figure 2F ) . For each FDR estimation method , all overlapping windows with mean hybridization scores whose corresponding FDRs were less than either 0 . 01 or 0 . 25 were collapsed into contiguous bound regions . Each bound region was assigned a hybridization score and FDR level equal to those of its highest scoring window , and the location of the maximum array hybridization within each bound region was determined and defined as its “primary peak window” ( Figure 2E ) . Our subsequent analyses focus on bound regions with FDRs below 0 . 01 and 0 . 25 ( the 1% and 25% sets , respectively ) . On average , the 25% FDR sets contained three to six times the number of genomic regions as their respective 1% FDR sets . Table 1 summarizes the number of bound regions identified for each factor; Tables S1 and S2 provide lists of the regions bound by each factor , and the locations of primary peak windows , as well as information on genes proximal to the regions for the 1% FDR and 25% FDR sets , respectively . There is excellent agreement between technical replicates as well as data from immunoprecipitation experiments using antibodies recognizing distinct epitopes on the same transcription factor ( see Figure 3 ) . There is also good correspondence of the bound regions identified for the different antibodies ( Table 1 ) . On average , 96% of the 1% FDR bound regions detected with one antibody overlap regions that score above the 25% FDR threshold for the second ( Table 1 ) ( 1% FDR regions were compared to 25% FDR regions to avoid not counting regions that lay just above the 1% FDR threshold for one antibody and just below for the other , which results in a somewhat lower overlap between the 1% FDR regions for both antibodies [Table 1] ) . Scatter plot analyses , presented later , also show a strong quantitative correlation between data from experiments using different antibodies to the same factor . The two methods for estimating FDRs also broadly agree , especially at the 1% FDR level ( Table 1 and Materials and Methods ) . ( For simplicity , in the remainder of the paper , we use only the symmetric null FDR estimates . ) To confirm the accuracy of our FDR estimates , we randomly selected 33 regions from the KR 1% FDR set and 23 regions from the BCD 1% FDR set , and amplified approximately 100-bp fragments close to the hybridization intensity peaks in each region by quantitative polymerase chain reaction ( Q-PCR ) from immunoprecipitated DNA . All of the regions tested were enriched in immunoprecipitations using both of the independently affinity-purified antibodies used for BCD and KR ( Figure S1 ) . In addition , 11 out of 16 KR bound regions selected from the bottom half of the 25% FDR list were enriched by Q-PCR from immunoprecipitated DNA ( Figure S1 ) , consistent with there being a significant fraction of bona fide bound regions between the 1% and 25% FDR thresholds . In vitro experiments have previously shown that UV or formaldehyde crosslinking correlates with levels of transcription factor occupancy on DNA [25 , 31 , 32] . To determine whether our experimental processing of immunoprecipitated material was distorting the levels of crosslinking , we carried out a series of control experiments . First , we applied the same series of amplification , labeling , and hybridization steps used for immunoprecipitated DNA to a sample of genomic DNA to which D . melanogaster bacterial artificial chromosomes ( BACs ) were added at known concentrations , and compared the data to unspiked genomic DNA . The ratio of signal intensities at oligos found in the BACs are lower than expected from the concentration of spiked BACs , but this compression is essentially monotonic and preserves the relative ranking of bound regions ( Figure S2B ) . Second , control Q-PCR experiments using immunoprecipitated chromatin samples also support the view that amplification and array hybridization do not profoundly distort relative DNA concentrations ( Figure S2A ) . Third , the relative primary peak window scores for BCD on a small collection of highly , moderately , and poorly bound regions are in line with the relative levels of in vivo UV crosslinking determined by direct Southern blot analysis of immunoprecipitated DNA [31] . In addition , the BAC experiments suggest that the great majority of regions significantly enriched after immunoprecipitation of chromatin will be detected by our array assay . We produced a set of 1% FDR regions for the spiked BAC DNA , and found that 100% of the regions present at 10-fold excess , 94% of the regions at 3-fold excess , and 51% of regions at 2-fold excess were present in the 1% FDR set . Although these model experiments do not precisely replicate the situation in the ChIP/chip experiments , they suggest that our amplification , hybridization , and array analysis methods are quite sensitive and should result in very few false negatives among moderately and highly enriched regions . The high concurrence between independent data from separate antibodies to the same factor also argue strongly against a significant false-negative rate among the 1% FDR regions . The most striking feature of the genome-wide data is the large number of bound regions identified for each factor ( see Table 1 ) , with most factors having thousands of bound regions . A total of 10 . 8 Mb are covered by 1% FDR bound regions identified for one or more of these factors , and a total of 6 . 6 Mb are within 500 bp of a primary peak for at least one of the regions bound above the 1% FDR threshold . These numbers represent 9 . 1% and 5 . 6% , respectively , of the 118 . 4 Mb euchromatic portion of the Release 4 genome sequence . A total of 40 . 38 Mb are covered by 25% FDR bound regions , and 31 . 2 Mb are within 500 bp of a primary peak of a 25% FDR bound region . These bound regions include all of the 43 CRMs known to be targets of one or more of these gap and maternal factors [9 , 33] ( see Figures 3 and 4 ) , arguing again that the false-negative rate in our ChIP/chip assay is very low . The known targets are , on average , more highly crosslinked than most of the identified bound regions , although some known targets are only poorly crosslinked ( Figures 5 and S3 ) , suggesting that lower levels of signal on the array may correspond to functional binding . In addition , the array hybridization signal frequently extends at low , but significant , levels many kilobases beyond the mapped edges of these known CRMs ( e . g . , Figure 3 ) . Although part of this flanking signal close to the CRMs is due to hybridization of immunoprecipitated DNA fragments that include the CRMs , because the signal extends well beyond the length of the DNAs immunoprecipitated , much of it must be due to crosslinking of the factors to sequences outside of the mapped CRMs ( for separate in vivo UV crosslinking data supporting this , see [32 , 34] ) . There is considerable overlap of the regions bound by each factor . Collectively , 82% of 1% FDR bound regions overlap a 25% FDR bound region by at least 500 bp for at least one of the other five factors ( Table 2 ) . Despite this extensive overlap among the regions bound by each factor in vivo , each factor binds with quantitatively different preferences . For example , whereas there is strong correlation of hybridization intensities between different antibodies for KR , there is a lower correlation between the hybridization intensity of KR and the other factors in multiply bound regions ( Figure 6 ) . Similar results are seen for the other factors ( Figure S4 ) . The variation in hybridization intensities for factors on the same target likely represents differences in the number of molecules of each factor occupying the shared target regions , either through differences in the number of recognition sequences bound and/or levels of occupancy at these sequences . Since many co-bound regions represent CRMs that direct different patterns of transcription , the quantitative differences in degree of binding of each factor must play a significant role in determining the unique output of these CRMs . Given the large number of in vivo binding regions identified in our analysis , the six gap and maternal factors may regulate a much broader array of genes and CRMs than the small collection of known target elements . To investigate to what extent the observed binding is associated with transcriptional regulation , we mapped each bound region to the gene transcribed in the blastoderm ( based on our RNA polymerase II ChIP/chip data ) whose 5′ end was closest to the center of bound region ( see Tables S1 and S2 ) . This mapping was imperfect due to the close packing of genes in the genome , the incomplete annotation of transcription units , and the ability of CRMs to act over large distances that sometimes skip intermediate genes; nonetheless , we still expected these associations to be broadly accurate . The most highly bound regions for each factor were preferentially found near genes that are transcribed in the blastoderm ( Figures 7 and S5 ) . For example , 49% of the 100 regions most highly bound by KR are within 10 kb of a transcribed gene , but at the 1% FDR threshold , a point at which there are insufficient false positives to significantly affect the analysis , only 22% of regions are within that distance of a transcribed gene ( Figure 7B ) . There is also a strongly nonrandom association of highly bound regions with genes that show patterned expression in the blastoderm ( Figures 7A , 7B , and S5 ) [35] . To further dissect these associations , we evaluated the enrichment of gene ontology ( GO ) terms of the putative targets of bound regions for each factor as a function of their position in the corresponding rank list ( Figures 8 and S6 ) . There is a consistent enrichment in the most highly bound regions for all factors of GO terms associated with A-P patterning , developmental control , and the regulation of RNA polymerase transcription . We also examined the percent of primary peak windows located in intergenic , intronic , 5′ and 3′ untranslated mRNA sequences , and protein coding regions ( Figure 9 ) . The great majority of the 500 most highly bound regions are found in intronic and intergenic sequences , as expected for transcription factor binding associated with gene regulation , but only at marginally higher frequencies than they would be found in a random selection of genomic regions ( Figure 9 ) . Surprisingly , for all factors except KR , many of the more poorly bound regions between the 1% and the 25% FDR thresholds are preferentially found in protein coding regions , which are not generally thought to code for CRMs ( Figure 9 ) . Thus it appears that many of the most highly bound regions are involved in patterning nearby genes and the set of highly bound regions likely includes many new blastoderm CRMs . In contrast , many of the thousands of poorly bound regions seem unlikely to be acting as classical CRMs directing transcription in the early embryo . Some may instead be active as CRMs later in development , when they may be more highly bound by these same factors; others may have some as yet undetermined function distinct from transcriptional regulation . But it is quite possible that a substantial proportion have no function at all . Among the most highly bound regions are four interesting classes of putative novel CRMs . ( 1 ) The noncoding regions flanking genes already known to be controlled by the early A-P regulators contain sequences bound highly by several of these factors that are distinct from their known CRMs ( e . g . , Figure 10A ) . ( 2 ) Many genes not previously known to be targets of the A-P regulators , but which are transcribed in spatial patterns in the blastoderm [35] , are associated with regions bound by these factors ( e . g . , Figure 10B ) . ( 3 ) A large proportion of transcribed microRNA ( miRNA ) genes are flanked by regions bound by A-P regulators ( Figure 10C , Table S3 ) , consistent with the increasing evidence that miRNAs are spatially patterned and play important roles in developmental processes [36 , 37] . ( 4 ) Noncoding sequences near the principle zygotic regulators of the D-V axis are highly or moderately bound by many of the A-P regulators ( e . g . , Figure 10D ) . Whereas the first three classes of putative target sequences are consistent with previous gene expression analysis , the observation that D-V regulators are bound was surprising as it has long been thought that they are not controlled by early A-P regulators . To test whether this binding might be functional , and whether previous regulation of D-V genes by A-P factors has been overlooked , we measured the mean levels of mRNA expression of four D-V regulators along the A-P axis using high-resolution imaging methods developed by the Berkeley Drosophila Transcription Network Project ( BDTNP ) [10 , 11] . Figure 11 shows that mRNA levels of the D/V regulators rhomboid ( rho ) , zerknult ( zen ) , twist ( twi ) , and snail ( sna ) change significantly along the A-P axis in wild-type embryos . The larger changes in expression can be readily seen in images of individual stained embryos , but smaller or more gradual changes—that can be as large as 40%—are only reliably detected by quantitative analysis ( Figure 11 ) . In mutant embryos that lack BCD , the expression of sna mRNA along the A-P axis changes in the manner expected in these mutants because the posterior half of the pattern is duplicated as a mirror image in the anterior . The BCD mutant data and our binding data suggest that the early A-P regulators control expression of D-V genes . This result parallels similar recent observations of the binding and regulation of A-P genes by D-V factors . The expression of many A-P genes is modulated up to 2-fold along the D-V axis [10 , 11] and analysis of genome-wide binding of the D-V factors Dorsal , Snail , Twist , Medea , and Zerknult show that these proteins bind to many A-P genes [38 , 39] ( unpublished data ) . One of our chief motivations for determining the sequences bound by transcription factors in vivo is to understand the molecular mechanisms that target factors to DNA . To begin this analysis , we examined the distribution of predicted recognition sequences for each factor in its bound regions and in regions where we do not detect it binding . We derived position weight matrices ( PWMs ) for each of the six factors either from DNaseI footprint data of recognition sequences found in known enhancers [40] or from in vitro selection ( SELEX ) experiments ( BDTNP , unpublished data ) . Two , BCD and KR , are shown in Figure 12A . Although such PWMs do not provide a complete description of a factor's binding specificity , they provide a first-order approximation [41] . We used these PWMs to identify all sequences across the genome that match each factor's in vitro binding specificity , and found that recognition sequences for each factor are enriched in their respective bound regions , the enrichment being greatest at the peak of array intensity hybridization ( Figures 12B and S7B ) . Such enrichment demonstrates that a significant fraction of the binding arises from the direct , sequence-specific interaction of each factor with its recognition sequences . However , bound regions , especially the most highly bound regions , show a marked G-C bias relative to their flanking sequences ( Figure 13A ) . This bias could lead to the spurious observation of recognition sequence enrichment , especially because many transcription factors recognize G-C–rich sequences . To ensure that the observed enrichment of recognition sequences for the gap and maternal factors in 1% FDR bound regions is not an artifact of general G-C bias , we repeated the enrichment analysis with PWMs generated by randomly permuting the order of columns within the real PWMs . Matches to these scrambled matrices are not enriched in bound regions , except in the case of HB whose homogenous PWM is not significantly altered by the permutation ( Figure S9 ) . However , a G-C bias would be expected to lead to a deficit of A/T-rich HB recognition sequences , and thus the enrichment of HB recognition sequences cannot be a result of G-C bias . As a separate control , we examined the enrichment of recognition sequences for each factor in regions not bound by the factor , but bound by at least one of the other five . These regions are G-C rich , but again , only very modest or no enrichment is observed ( Figure S10 ) . Thus , the enrichment of recognition sequences is largely specific to regions bound by the factor and to the factor's correct PWM . There is a strong positive correlation between the predicted affinity of a recognition sequence ( estimated here by its score against a factor's PWM ) and its enrichment ( Figure 12B and S7B ) . For example , the eight highest-affinity variants of the 8-bp BCD recognition sequence ( defined here as sequences that have log-likelihood scores against the BCD PWM of less than 0 . 0001; smaller log-likelihood values represent better matches to the PWM ) are enriched over 8-fold in BCD bound regions relative to flanking noncoding DNA . In contrast , the 184 medium-affinity variants ( those with log-likelihood scores between 0 . 001 and 0 . 003 ) are enriched only about 1 . 5 times over background . However , the total excess ( compared to noncoding sequences in which we did not detect binding ) of medium-affinity BCD recognition sequences in BCD bound regions , 1 . 3 recognition sequences per 1 , 000 bp of bound region , is higher than the excess of high-affinity recognition sequences , 1 . 0 per 1 , 000 bp ( unpublished data ) , suggesting that medium-affinity sites likely play a significant role in targeting factors to DNA . The enrichment of recognition sequences is greatest for the most highly bound regions , and declines with decreasing levels of in vivo binding ( Figures 12C and S7C ) . Despite this decrease in enrichment , the high-affinity recognition sequences are enriched far down the rank lists . Although recognition sequences are enriched on average in bound regions , consistent with previous data [32 , 34 , 42] , the enrichment is modest and is not uniform among bound regions . A significant number of bound regions contain fewer recognition sequences for the bound factor than are found in many unbound regions ( Figures 12D and S7D ) . For example , 80% of BCD 1% FDR primary peaks contain no predicted high-affinity BCD recognition sequences ( p < 0 . 0001 ) , and 20% do not contain any intermediate-affinity sequences ( p < 0 . 001 ) . Furthermore , there are numerous unbound regions that contain intermediate- and high-affinity recognition sequences . The excess of G-C bias in bound regions noted above warranted further analysis . We were concerned that the correlation between strength of binding and G-C content might reflect a bias for G-C rich sequence to hybridize more strongly to the array . We used the BAC data described above to investigate the effect of G-C content on the hybridization score in the 675-bp windows we used in our analyses . There is a tendency for windows with ( on average ) low G-C content to have lower mean window scores ( Figure S8 ) , which could bias the selection of peak hybridization windows within bound regions towards those with higher G-C content . However , this effect would likely be somewhat countered by the tendency for windows with high G-C content to have lower mean window scores ( Figure S8 ) . In addition , peak window scores correlate with enrichment measured by Q-PCR , which is not subject to G-C bias ( Figures S1 and S2 ) . Finally , a similar GC bias has been observed in a large collection of enhancers not identified by array hybridization [43] . We therefore conclude that bound regions are G-C rich relative to other noncoding DNA . Many lines of evidence suggest that animal transcription factors act in a context-dependant , combinatorial manner in which the action of one factor influences the behavior of another [44–49] . As a result , it is widely believed that a key to understanding how specific CRMs are constructed so that they are correctly bound by a defined set of transcription factors and produce specific patterns of expression lies in understanding a code that integrates information from multiple recognition sequences . For example , the ability to predict the locations of functional CRMs for the six early regulators is greatly improved when binding of multiple factors is considered at the same time , rather than when binding for factors is considered in isolation [8 , 9] . These observations suggest that the binding of one factor may influence the DNA binding or activity of other factors on an element . To begin to search for evidence of such effects in our dataset , we examined how the binding of additional factors influenced the frequencies of predicted recognition sequences for each factor in its bound regions ( Figure 14 ) . For some factors , such as BCD , little difference is seen , but for others . substantial , and in some cases counterintuitive , changes are seen . Predicted HB recognition sequences are enriched in regions bound exclusively by HB and by HB in combination with one or more of GT , KR , KNI , and CAD . However , predicted HB recognition sequences are not enriched in regions bound by HB and BCD ( Figure 14D ) , even though these regions are crosslinked 1 . 3-fold more highly by HB than regions not bound by BCD ( unpublished data ) . Thus there appears to be a complex influence of other factors on the manner in which at least HB recognizes its target sequences . A critical question unanswered by the above analyses is what fraction of the regions bound in vivo are biologically functional , be they involved in transcriptional regulation or some other cellular function . Many of the most highly bound regions overlap CRMs known to regulate important developmental processes , and it is likely that many of the remaining highly bound regions , especially those near important developmental control genes , will have regulatory activity in the blastoderm . However , these regions represent just a small fraction of the regions bound by these six factors . To begin to address the function of the remaining bound regions , we examined the evolutionary constraints on predicted recognition sequences in bound regions . We expect purifying selection to constrain substitutions at bases involved in protein–DNA interactions that mediate important regulatory events . Functional recognition sequences have consistently been observed to evolve more slowly than expected under neutral models [50] , and evolutionary constraint on noncoding DNA is often used as a proxy for regulatory function ( e . g . , [39] ) . The measurement and interpretation of constraint on recognition sequences , however , is not straightforward . First , Drosophila noncoding DNA is , in general , highly constrained . It has been estimated that over half of the bases in intergenic and intronic DNA have evolved under purifying selection [51 , 52] , compared to roughly 5% in mammals [53] . Thus , when compared to the presumptive neutral rate of substitution ( estimated in Drosophila from short introns ) , virtually any collection of recognition sites in Drosophila noncoding DNA will appear to be under evolutionary constraint , whether the sites are functionally bound or not . Second , despite this generally high level of noncoding constraint , functional recognition sequences are not always conserved . For example , it has been shown that several functional recognition sequences from the D . melanogaster eve stripe 2 enhancer are absent in other Drosophila species even though the enhancers themselves maintain their function [54–56] . Presumably , the loss of these sites is compensated for by the gain of sites elsewhere in the enhancer . Nonetheless , in most CRMs examined to date , at least a subset of recognition sequences are constrained over significant evolutionary distance , and thus it seemed reasonable that an analysis of the patterns of binding site constraint within bound regions might provide insight into their function . We began with two measures of sequence constraint: ( 1 ) rates of pairwise substitution between D . melanogaster and its sister species D . simulans , and ( 2 ) PhastCons scores measuring constraint across 12 sequenced , closely and distantly related Drosophila species [57 , 58] . Because D . melanogaster and D . simulans are so closely related , there is essentially no ambiguity in alignments of their genomes . However , the small number of changes also limits our ability to detect differences in rates of evolution between classes of sequences by rates of pairwise substitution . PhastCons scores that employ a wider diversity of species , in contrast , have a much greater statistical power , but can be confounded by alignment error in those species that are more distantly related [59] . For each transcription factor , we examined constraint on recognition sequences in 500-bp primary peaks from 1% FDR regions , unbound regions , and short introns ( see Table 3 ) . We also examined both measures of constraint down the rank lists of bound regions for each factor ( see Figures 15 and S12 ) . In both analyses , we focused exclusively on recognition sequences in noncoding DNA ( introns and intergenic sequences , with RNA coding genes excluded ) so as to avoid the confounding effects of conservation of coding capacity . As shown in Table 3 and Figures 15 and S12 , for each of the six factors , there is a general trend for recognition sequences in highly bound regions to be under the strongest constraint , followed by recognition sequences in poorly bound regions , recognition sequences in unbound regions , and recognition sequences in short introns . For example , for DNA sequences matching the KR PWM , their PhastCons scores are highest in the several hundred most highly bound regions ( Figure 15A ) , and their mean PhastCons scores in 1% FDR regions are somewhat higher than those in unbound noncoding DNA and much higher than those in short introns ( Table 3 , PhastCons Scores ) . A similar trend , however , is observed for the remaining parts of the bound regions , outside of the specific factor recognition sequences , so the observed effect is at least in part due to overall higher levels of constraint in highly bound regions ( “non-sites” data in Figures 15 and S12 , Table 3 ) . To evaluate the extent to which these patterns of constraint were specific to recognition sequences for the factor , we therefore examined constraint on recognition sequences predicted after randomly permuting the order of columns of the specificity matrix for each factor . For regions highly bound by BCD , CAD , and GT , none of the scrambled permutations produced recognition sequences that were as highly constrained as those from the real specificity matrix , and the average derived using many permutated PWMs was significantly lower than from the real PWM ( Figure 15 , Table 3 ) . In fact , the average of the scrambled PWMs was similar to that of the remaining parts of the bound regions , outside of the specific factor recognition sequences ( Figures 15 and S12 , Table 3 ) . Only in the case of BCD , however , did the patterns of constraint correlate with the in vivo DNA binding data because the additional constraint on recognition sequences , above the local background , dropped down the rank list , disappearing at around the 1% FDR threshold . For CAD and GT , the difference in conservation was consistent across the rank list all the way to the 25% FDR threshold , suggesting that the excess conservation may not be specific for bound recognition sequences . For HB , KNI , and KR , there is even less evidence for specific conservation of recognition sequences because many of the permutations produced recognition sequences that were more conserved than the real sites , and the average score of these permutated PWMs were not significantly different from those of the real matrix at either highly or poorly bound regions ( Table 3; Figure 15 ) . Overall , the comparative analysis adds further evidence that highly bound regions differ in character from poorly bound regions , but , with the exception of BCD , does not provide compelling evidence that the binding we observe contributes to organismal fitness . Although the six factors studied here are the initiating regulators of A-P expression in the embryonic trunk , it is likely that other factors are involved in activating or otherwise regulating their targets . For example , several known targets of maternal and gap factors are also regulated by genes in the terminal system that controls expression in the head and tail [60–62] . To investigate whether other factors may be binding to the regions bound by the maternal and gap transcription factors , we systematically searched for sequences enriched in the regions surrounding the primary peaks for each of the six factors . As shown above , the recognition sequences of each factor are enriched in their respective bound regions , and these sequences are routinely recovered in de novo searches for enriched sequences in bound regions . However , for each of the six factors , the most strongly enriched sequence was the heptamer CAGGTAG/CTACCTG ( Table S4 ) . This “TAGteam” sequence has been previously reported to control the timing of preblastoderm transcription [63 , 64] . Although its precise role in activating transcription is only beginning to be understood [63 , 65] , it is found in roughly 30% of bound regions and is concentrated in the most highly bound regions , emphasizing the broad role that it plays in early embryonic transcription . Furthermore , of all heptamers , CAGGTAG shows the greatest increase in interspecies conservation in bound regions relative to non-bound intergenic sequences .
Given that our ChIP/chip data identify several orders of magnitude more bound regions than the number of previously identified targets of these factors , the most immediate question is whether these bound regions are all functional . Several lines of evidence suggest that the bulk of the several hundred most highly bound regions are directly involved in regulating the transcription of neighboring genes . In particular , these most highly bound regions are preferentially found near genes transcribed in the blastoderm , these putative targets are enriched for patterned genes and genes with known roles in patterning and early development , and the most highly bound regions are preferentially conserved relative to other noncoding sequences ( Figures 3–5 , 7–8 , 10–11 , 15 , S12 , and Table 3 ) . The highly bound regions also tend to be located within intergenic and intronic sequences ( Figure 9 ) , as expected for regulatory sequences . All of these associations , however , dissipate down the rank lists for each factor , with an increasing percentage of more poorly bound regions mapping to genes that are not transcribed at this early stage of development and/or to protein coding regions or to noncoding regions that are less well conserved ( Figures 3–5 , 7–11 , 15 , S11 , S12 , and Table 3 ) . This suggests that the poorly bound regions have different , or perhaps even no , function . One possibility is that poorly bound regions regulate the transcription of adjacent genes , but more subtly than highly bound regions as it has been shown that many genes not directly involved in A-P patterning ( e . g . , housekeeping genes ) show weak A-P patterns at stage 5 [35 , 66 , 67] . Another possibility is that the low levels of binding seen at stage 5 may presage stronger binding and transcriptional regulation of adjacent genes later in development . In support of this , binding of HB increases in the neuroectoderm of stage 9 embryos at a subset of regions bound at low levels at stage 5 ( unpublished data ) , which , as genes become transcribed , likely results at least in part from a change in chromatin structure increasing access of factors to their recognition sequences [68–73] . A third possibility is that the observed binding is not involved in transcriptional regulation , but instead plays a role in regulating processes such as chromosome structure , DNA replication , or DNA repair . However , these six transcription factors have not been implicated in other cellular functions to date . Finally , and in many ways most tantalizingly , some lower-level binding may be truly nonfunctional and simply result from transcription factors binding to randomly occurring target sequences that , precisely because they do not significantly affect gene expression , are not selected against . Indeed , it has long been proposed on thermodynamic grounds that transcription factors would bind at low , nonfunctional levels throughout the genome either via sequence-independent [74–76] or sequence-specific DNA binding [32] . However , even with these factors bound poorly to many thousands of regions across the genome , at any instant they could only bind to a small fraction of their recognition sequences within the genome , and they would still inevitably have an indirect function in the system by buffering the molecules available for binding within CRMs . Determining which regions bound in vivo are functional and in what way ( s ) they function will be challenging . Our most reliable assay for sequences that regulate transcription—the construction of transgenic D . melanogaster embryos in which the sequence to be assayed is juxtaposed with a basal promoter and reporter gene—has several limitations . The assay only detects sequences that act independently of other sequences , whereas many bound regions are likely to augment the activity of other sequences or act redundantly [77 , 78] . Subtle or redundant regulatory activity is often difficult to detect in transgenes that use nonnative promoters and reporter genes . Finally , repressor , insulator , and other transcriptional regulatory activities require separate assays . Comparative sequence analysis also has the potential to contribute to the dissection of the function of bound regions and the recognition sites within them . These analyses , however , can be extremely complex and occasionally misleading . It is common in published analyses of regulatory sequence conservation to assume that recognition sequences occur in a homogenous background of nonconserved sequences . But neither the assumption of neutrality nor that of homogeneity is appropriate . A substantial fraction of Drosophila noncoding DNA is under selective constraint—and presumably involved in some function [51] ( Table 3 ) . Thus simply observing that a collection of recognition sequences is conserved ( i . e . , evolves slower than the presumptive neutral rate ) , as has frequently been done in the literature , does not reliably establish that transcription factor binding to these sequences contributes to fitness . It is necessary instead to use methods that attempt to detect conservation of binding potential of particular recognition sequences [50 , 79] . Even this is complicated , however , by variation in rates of constraint that are correlated with genomic features that are in turn related to transcription . For example , noncoding sequences flanking genes transcribed in the embryo are more conserved than randomly selected noncoding sequences . Since highly bound sequences are also associated with genes transcribed in the embryo , it appears , often incorrectly , that recognition sequences in highly bound regions are preferentially conserved . We have used several methods to control for these effects , but they may still be susceptible to other confounding factors . Our analysis to date has only been able to establish for one of the six factors ( BCD ) that its recognition sequences are constrained above the background in the flanking DNA , even within the most highly bound regions ( Figure 15 , Table 3 ) . For the other five factors , the results are ambiguous or not apparent ( Figures 15 and S12 , Table 3 ) . This raises the unpleasant possibility that evolutionary constraint may not be as useful as generally believed for distinguishing functional targets of transcription factors . Our data suggest that the rules governing factor targeting in vivo are likely to be subtle and complex . Consistent with in vivo crosslinking analyses of other animal transcription factors [39 , 42 , 80–84] , the more highly crosslinked regions in vivo do , on average , show greater enrichment of factor recognition sequences than poorly bound or unbound regions ( Figures 12B , 12C , and S7 ) , indicating that these factors' intrinsic DNA binding specificities play a role in determining the pattern of binding in vivo . Not only high-affinity recognition sequences , but also low-affinity sequences are enriched , suggesting that weaker sites help mediate binding ( Figure 12B and S7 ) . As shown previously [32 , 34 , 42] , however , in vitro DNA specificity alone cannot fully account for the distribution in vivo because many nonbound genomic regions contain higher densities of high-affinity recognition sequences than bound regions ( Figures 12D and S7 ) . Additional analyses indicate that factor targeting also depends on a local context established by other transcription factors: the degree of in vivo binding of a factor per number of specific recognition sequences at a given location is dependent on which other factors also bind the region in vivo ( Figure 14 ) . We do not know which factors are mechanistically responsible for establishing this context . The context could either be established by some of the A-P regulators themselves , or by one or more of the other factors whose recognition sequences are enriched in the bound regions ( Table S4 ) . In either case , the context establishing factors could act by cooperative interactions to increase A-P regulator DNA occupancy [85] , or by increasing accessibility to factor recognition sequences locally via an effect on chromatin structure [70 , 86] . The idea that chromatin structure plays a dominant role is appealing as it provides a ready explanation for why we observe multiple factors being targeted to the same highly overlapping set of regions ( Figures 6 and S4 , Table 2 ) . Each open chromatin region is expected to contain recognition sequences for many transcription factors because of the high frequency of such sequences throughout the genome . The factors would then be forced to bind the recognition sequences in open regions because regions elsewhere would not be available [70 , 86] . Some of our conclusions agree with those of other recent studies of in vivo DNA binding by sequence-specific transcription factors in animals . For example , some of these other proteins are also observed to bind extensively to a large number of genomic regions [38 , 42 , 80 , 83 , 87 , 88] . But in some important regards , our analyses and conclusions differ . Given that the earliest in vivo crosslinking studies of sequence-specific transcription factors established that , at least for some factors , genes are bound over a quantitative range that correlates with gene type , degree of gene regulation , and transcriptional state [34 , 78] , it is very surprising that all recent analyses have ignored this quantitative information and instead classified genomic regions as either bound or not bound [38 , 39 , 42 , 80 , 82–84 , 87 , 88] . A range of experiments suggest that crosslinking and ChIP/chip signals broadly correlate with different levels of transcription factor occupancy [32 , 34] ( Figures S1 and S2 ) . The analyses presented in this paper clearly reinforce how useful it is to consider the relative level of transcription factor occupancy in studying the complex range of genomic regions bound by animal transcription factors in vivo . Most analyses have either assumed that all regions bound in vivo must be functional targets or not actively considered whether a substantial fraction of bound regions may be nonfunctional [38 , 39 , 42 , 80 , 82–84 , 87 , 88] . Only a recent paper from the ENCODE Consortium [81] has seriously considered the possibility that a significant percent of in vivo binding may be nonfunctional , based on the lack of evolutionary constraint in bound sequences . However , the absence of constraint does not establish the absence of function , as it is well established that regulatory sequences can maintain their function in the absence of primary sequence conservation . We have shown that poorly bound regions lack many of the hallmarks of regulatory sequences . Another recent study of in vivo binding by sequence-specific transcription factors in Drosophila measured DNA methylation patterns of transcription factor/DNA adenine methyltransferase ( Dam ) fusion proteins . Binding was assayed over a 2 . 9-Mb region of the genome in tissue culture cells for seven ectopically expressed factors , including BCD [89] . These fusion proteins were strongly targeted to a common set of “hot spots . ” Because the factors have unrelated functions , it was proposed that hot spots are not classical cis-regulatory elements , but instead act either as sinks to sequester molecules , as mediators of interactions between distant genomic loci , or as unconventional enhancers at which many factors play only a minor role . The pattern of binding of endogenous BCD we observe in embryos , however , differs dramatically from that predicted by the methylation patterns in tissue culture cells . Within the 2 . 9-Mb region , only 16 1% FDR BCD bound regions are present . Of these , only five overlap the top 50 regions detected in the methylation assay , suggesting that the distributions mapped in tissue culture cells do not reflect binding by regulators normally expressed in other cells . In addition , we have mapped the binding of an additional 12 endogenous sequence-specific transcription factors in the early embryo that control D-V axis patterning and pair rule segmentation and which represent a broad range of transcription factor families ( unpublished data ) . Together with the six maternal and gap A-P regulators studied in this paper , these endogenous factors do in fact frequently target the same short genomic regions , though they bind these regions at very different relative levels . These commonly bound regions , however , include most of the well-known CRMs active in the early embryo . We suspect that many animal CRMs are bound by a much larger number of factors than currently realized , though it remains to be determined what fraction of this binding is functional .
Rabbit antisera were kindly provided by Sean Carroll ( HB ) , Gary Struhl ( BCD and CAD ) , Herbert Jackle , Ralf Pflanz , and Pilar Carrera ( KR ) . Rabbit antisera for KNI was raised from a 6xHIS-tagged full-length KNI protein expressed in Escherichia coli . For each of the six transcription factors , two sets of antibodies that recognize nonoverlapping portions of the protein were affinity purified from rabbit antisera . The Gateway cloning and expression system ( Invitrogen ) was used to generate parts of each protein for affinity purification . Each affinity reagent consists of at least 100 amino acids that do not contain any significant similarity with any other protein in the D . melanogaster genome ( no segment was used if it had greater than 20% identity to any other D . melanogaster protein , or perfect identity of ten amino acids or greater , as assessed by BLASTP [90] ) . To generate recombinant proteins , we cloned PCR-amplified fragments corresponding to the selected amino acid regions ( below ) using BP Clonase and the pDONR221 vector ( Invitrogen ) . After sequence verification of the entire amplified product , the fragments were transferred to the 6xHis-tagged bacterial expression vector pDest17 using the LR-Clonase and subsequently verified by PCR . Anti-HB ( HB1 ) and anti-HB ( HB2 ) were purified against HB amino acids 1–305 and amino acids 306–758 , respectively; anti-GT was purified using GT amino acids 182–353; anti-KR ( KR1 ) and anti-KR ( KR2 ) were purified against KR amino acids 1–230 and amino acids 351–502 , respectively; anti-CAD was purified against CAD amino acids 1–240; and anti-KNI ( KNI1 ) and anti-KNI ( KNI2 ) were purified against KNI amino acids 130–280 and amino acids 281–425 , respectively . The two affinity-purified sets of anti-BCD antibodies ( BCD1 and BCD2 ) were purified against BCD amino acids 56–330 and 330–439 , respectively , as described previously [34] . The monoclonal antibody H14 , which recognizes RNA polymerase II CTD repeats phosphorylated at Ser 5 , was obtained from CRP Inc . Further details on BDTNP antibodies and protein expression vectors are at http://bdtnp . lbl . gov/Fly-Net/ . The detailed protocol ( Protocol S1 ) was used . Briefly , 2–3-h-old embryos ( late stage 4 and early stage 5 ) were formaldehyde crosslinked , and chromatin was isolation by CsCl gradient ultracentrifugation as described previously [25 , 31 , 78 , 91] . The isolated chromatin was sonicated to an average size of about 600 bp prior to ChIP and dialyzed . Protein A–Sephacryl 1000 beads were prepared based on a method described previously [92] . The larger pore size was found to give at least a 3-fold higher yield of crosslinked chromatin after immunoprecipitation . The chromatin solution was precleared by incubating with normal rabbit IgG and the protein A–Sephacryl 1000 beads . Factor IP reactions were carried out in duplicate by incubating 100 μg of chromatin with 0 . 5–3 μg of the appropriate antibody for 3 h or overnight at 4 °C; parallel control IgG IP reactions , also in duplicate , were carried out with normal rabbit IgG . The immunoprotein–chromatin complexes were captured by incubating with protein A–Sephacryl 1000 beads , followed by consecutive washes and then eluted with a buffer containing 1% SDS and 0 . 1 M NaHCO3 ( pH 10 . 0 ) . In each ChIP experiment , a portion of the chromatin solution corresponding to 1% of that used in the ChIP reaction was used as input DNA control . The DNA from this sample , along with the Factor IP and IgG control IP samples , was purified by phenol/chloroform extraction and ethanol precipitation after the protein/DNA crosslinks had been reversed by incubation at 65 °C . Duplicate Factor IP , IgG control IP , and input DNA samples were amplified using a modified random primer-based DNA amplification protocol that gives significantly improved amplification consistency , particularly when the small quantities of genomic DNA recovered in our ChIP reactions ( <0 . 5 ng ) are amplified . After amplification , each DNA sample was fragmented with DNase I , biotinylated , and hybridized to Affymetrix Drosophila genomic tiling arrays [26] . Each array contains over 3 million oligo probes that cover the euchromatic portion of the genome at a resolution of about one per 36 bp on average . ChIP/chip array data were processed using TiMAT , a Java- and R-based open-source software package developed by the BDTNP ( http://bdtnp . lbl . gov/TiMAT/TiMAT2/ ) . Array images were visually inspected for blemishes and artifactual bright spots , which were then masked to the array's median probe intensity value in the few cases necessary . Only oligonucleotides present exactly once in the D . melanogaster genome ( release 4 . 0 ) were used for subsequence analysis . The complete set of six arrays from an experiment—Factor IP replicates , IgG control-IP replicates , and input DNA replicates ( Figure 2A and 2B ) —were scaled to a common median value and then quantile normalized against each other [93] . Replicates were averaged and log ( base 2 ) ratio scores were calculated for Factor IP and IgG control IP arrays: log ( mean Factor IP/mean input DNA ) and log ( mean IgG control IP/mean input DNA ) ( Figure 2C ) [94] . These scores were then smoothed using a sliding window of trimmed means 675 bp in length ( Figure 2D ) . Next , FDR estimates were calculated by two methods . The first used an assumption of a symmetric window scores null distribution to compute p-values , then applied a multiple testing correction [95] to control the FDR . More specifically , the symmetric null method constructed a null distribution estimate by using window scores to the left of the mode of the full distribution and then reflecting these scores over the mode ( Figure 2F ) . This approach is justified by the presumed relative scarcity of genomic regions enriched in the immunoprecipitation , and is a nonparametric variant on the approach of [30] . The second method simply computed the ratio of the number of windows scoring above a given cutoff in the IgG control IP data to the number in the Factor IP data—the assumption being that there are no true positives in the negative control data ( Figure 2G ) . Bound regions ( called “intervals” in TiMAT ) were then defined by first filtering out all windows with scores above the given FDR threshold , then collecting these into contiguous stretches of windows containing a minimum of ten windows , with a maximum allowable gap of 200 bp between any two adjacent windows ( Figure 2E ) . Both 1% and 25% FDR thresholds were considered . The location of the maximum array hybridization within each bound region was determined and defined as its “peak window , ” with the oligo having maximum intensity within each bound region defined as the “primary peak . ” Local peaks of array intensity were identified within each bound region using a recursive algorithm that considers peak shape , height , and period . In this paper , for simplicity , only the largest ( primary ) peak in each bound region has been used in subsequent analyses , which in some cases removed secondary and tertiary peaks that show characteristics of CRMs . Results , including oligonucleotide probe intensities , trimmed-mean window scores , bound region locations , peak magnitudes and locations , and nearby genes are reported in . sgr and . gff file formats as well as TiMAT's own text-based report files . The bound regions analyzed by Q-PCR were selected arbitrarily throughout the symmetric null 1% FDR rank list for BCD and KR . KR bound regions between the symmetric null FDR score 1% and 25% thresholds were chosen with a bias towards the more poorly bound regions . The oligonucleotide primers and probes were designed to be as close as possible to the peak of each binding region , the majority falling within 200 bp of the peak . Q-PCR reactions were carried either using the random-prime–amplified input DNA and Factor IP DNA samples or the original Factor IP and IgG control IP samples , i . e . , without random-prime amplification . Eight BAC plasmids , each containing about 170 kb of Drosophila genomic sequence were used . The BAC DNAs were mixed together at a relative molar concentration of one , four , ten , and 20 , with two BACS at each concentration . The BAC DNA cocktail was then mixed with Drosophila whole-genomic DNA to generate two samples , one containing BACs at one , four , ten , and 20 times the molar concentration of genomic DNA , and the other at two , eight , 20 , and 40 times . A total of 20 ng of each genomic DNA/BAC cocktail were random-prime amplified , and the resulting DNA samples were fragmented , biotinylated , and hybridized to chips following our protocol ( above ) . Bound regions were associated with the gene ( from release 4 . 3 of the D . melanogaster genome ) whose 5′ end was closest to the primary peak in the bound region . To identify the closest transcribed gene , the subset of release 4 . 3 annotations that completely overlap regions bound by RNA PolII in our ChIP/chip experiments was used . For BCD , HB , GT , CAD , and KR , PWMs were constructed from unpublished SELEX data using MEME [96] . KNI was constructed from DNaseI footprint data contained in [40] . Binding site positions were predicted using PATSER [97] with the indicated p-value cutoffs . Binding site enrichment was measured by dividing the density of sites in the bound regions by the density of sites in a control set of unbound sequences . The control set consisted of randomly selected noncoding sequences that did not overlap with 1% FDR regions for any factor . Evolutionary constraint on recognition sequences and bound regions was computed using 15 species ( 12 sequenced Drosophila species plus Anopheles gambiae , Apis mellifera , and Tribolium castaneum ) PhastCons scores obtained from the Univeristy of California Santa Cruz Genome Browser ( http://genome-test . cse . ucsc . edu/goldenPath/dm2/multiz15way/ ) and the pairwise D . melanogaster–D . simulans divergence was computed from LAGAN alignment of orthologous noncoding regions identified by a combination of BLAST and synteny . Mean constraints of recognition sequences in bound regions was compared to the same for recognition sequences in short ( less than 100 bp ) introns and in unbound noncoding regions . In addition , mean constraint was computed for recognition sequences predicted using randomly permuted PWMs for each factor in which the order of the matrix columns had been scrambled ( permutations whose recognition sequences were enriched in bound regions were excluded to avoid permuted matrixes that were too similar to the unpermuted matrix ) . For the pairwise D . melanogaster–D . simulans comparisons , only substitutions , and not insertions or deletions , were considered .
|
One of the largest classes of regulatory proteins in animals , sequence-specific DNA binding transcription factors determine in which cells genes will be expressed and so control the development of an animal from a single cell to a morphologically complex adult . Understanding how this process is coordinated depends on knowing the number and types of genes that each transcription factor binds and regulates . Using immunoprecipitation of in vivo crosslinked chromatin coupled with DNA microarray hybridization ( ChIP/chip ) , we have determined the genomic binding sites in early embryos of six transcription factors that play a crucial role in early development of the fruit fly Drosophila melanogaster . We find that these proteins bind to several thousand genomic regions that lie close to approximately half the protein coding genes . Although this is a much larger number of genes than these factors are generally thought to regulate , we go on to show that whereas the more highly bound genes generally look to be functional targets , many of the genes bound at lower levels do not appear to be regulated by these factors . Our conclusions differ from those of other groups who have not distinguished between different levels of DNA binding in vivo using similar assays and who have generally assumed that all detected binding is functional .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"developmental",
"biology",
"genetics",
"and",
"genomics"
] |
2008
|
Transcription Factors Bind Thousands of Active and Inactive Regions in the Drosophila Blastoderm
|
In Borrelia burgdorferi ( Bb ) , the Lyme disease spirochete , the alternative σ factor σ54 ( RpoN ) directly activates transcription of another alternative σ factor , σS ( RpoS ) which , in turn , controls the expression of virulence-associated membrane lipoproteins . As is customary in σ54-dependent gene control , a putative NtrC-like enhancer-binding protein , Rrp2 , is required to activate the RpoN-RpoS pathway . However , recently it was found that rpoS transcription in Bb also requires another regulator , BosR , which was previously designated as a Fur or PerR homolog . Given this unexpected requirement for a second activator to promote σ54-dependent gene transcription , and the fact that regulatory mechanisms among similar species of pathogenic bacteria can be strain-specific , we sought to confirm the regulatory role of BosR in a second virulent strain ( strain 297 ) of Bb . Indeed , BosR displayed the same influence over lipoprotein expression and mammalian infectivity for strain Bb 297 that were previously noted for Bb strain B31 . We subsequently found that recombinant BosR ( rBosR ) bound to the rpoS gene at three distinct sites , and that binding occurred despite the absence of consensus Fur or Per boxes . This led to the identification of a novel direct repeat sequence ( TAAATTAAAT ) critical for rBosR binding in vitro . Mutations in the repeat sequence markedly inhibited or abolished rBosR binding . Taken together , our studies provide new mechanistic insights into how BosR likely acts directly on rpoS as a positive transcriptional activator . Additional novelty is engendered by the facts that , although BosR is a Fur or PerR homolog and it contains zinc ( like Fur and PerR ) , it has other unique features that clearly set it apart from these other regulators . Our findings also have broader implications regarding a previously unappreciated layer of control that can be involved in σ54–dependent gene regulation in bacteria .
Bacterial gene expression is primarily controlled at the transcriptional level , which requires a central DNA-dependent RNA polymerase ( RNAP ) consisting of catalytic core ( α2ββ'ω; E ) and a dissociable σ factor [1] . Gene transcription occurs when the Eσ-promoter closed complex ( CC ) is converted to the open complex ( OC ) . Among various σ factors , the alternative σ factor σ54 ( σN , RpoN ) is employed by many bacteria to transcribe genes involved in a wide variety of cellular functions such as virulence , nitrogen metabolism , and stress responses [1] . Unlike other σ factors , the Eσ54 holoenzyme alone cannot melt the promoter . The Eσ54-CC ( held by the interaction of Eσ54 with the unique -24/-12 promoter ) remains in this conformation until the activator ATPase interacts with RNAP , which hydrolyzes ATP for promoter melting [2]–[4] . The activator ATPase , also known as the enhancer-binding protein ( EBP ) , usually binds to an enhancer site located ∼100–150 bp upstream of the promoter . Typically , the EBP interacts with the RNAP via a DNA looping mechanism that is modulated by a DNA-bending protein such as integration host factor ( IHF ) . Borrelia burgdorferi ( Bb ) , the Lyme disease spirochete [5]–[6] , encodes three σ factors: the housekeeping σ70 ( RpoD , BB0712 ) , and two alternative σ factors , σ54 ( RpoN , BB0450 ) and σS ( RpoS , BB0771 ) [7] . Abundant evidence [8]–[11] has revealed that Bb RpoN directly binds to the -24/-12 site in rpoS promoter and thus activates rpoS which , in turn , modulates the differential expression of more than 100 genes involved in Bb virulence , stress adaptation , and many other functions . Studies have indicated that the Bb RpoS regulon is triggered by various environmental stimuli including temperature , pH , cell density , and unknown mammalian host factors [12]–[15] . The RpoN-RpoS pathway ( or the Rrp2-RpoN-RpoS pathway ) [8]–[11] , [16]–[18] plays a central role in modulating the differential expression of Bb outer surface lipoproteins such as outer surface protein C ( OspC ) [14] , [19]–[20] and decorin-binding protein A ( DbpA ) [21]–[23] , which are critical for Bb to transmit from the arthropod tick vector to mammalian hosts and to maintain its natural life cycle . Activation of the RpoN-dependent rpoS gene requires the activation of Rrp2 ( BB0763 ) , a putative EBP [17] . Although Rrp2 was presumed to be the sole NtrC-like EBP in Bb , Rrp2 seems to be unconventional as an EBP in that it apparently does not bind specifically to the RpoN-dependent promoter of rpoS [8] , [24] . The efficient translation of rpoS mRNA also requires the small RNA DsrA and an atypical RNA chaperone Hfq [25]–[26] . Emanating from our general interest in virulence expression in the Lyme disease spirochete , we previously showed that a manganese transporter , BB0219 ( BmtA ) , is required for full virulence by Bb [27] . The implication that metal transport , and perhaps metal control over borrelial gene regulation , could influence Bb's virulence prompted us to expand our study to examine other molecules of Bb implicated in metal-sensing . To this end , BosR ( BB0647 ) is a ferric uptake regulator ( Fur ) -like homologue in Bb [7] , [28] that has been postulated to contribute to the regulation of oxidative stress responses [29] . In many bacteria , Fur globally regulates iron homeostasis and other functions [30] . Given the provocative finding that Bb seems not to accumulate iron [31] , it remained tenuous as to whether BosR is involved in iron uptake by Bb . Nonetheless , BosR may influence cellular functions other than iron acquisition . Recently , we and others surprisingly found that expression of the RpoS regulon was significantly impaired in a mutant deficient in bosR , leading to the additional unexpected finding that BosR somehow functions as a second activator to promote σ54-dependent rpoS transcription , and that such control by BosR ultimately governs the expression of virulence-associated membrane lipoproteins and mammalian infectivity by Bb [32]–[35] . Although this discovery has represented a major advance in further understanding the regulatory control of virulence expression by the Lyme disease spirochete , the observation has engendered many new unanswered questions . Among them include whether BosR is indeed a global regulator common to more than the one virulent strain of Bb ( examined in previous studies ) [32]–[34] , and how BosR may act mechanistically to exert its positive control over the RpoN-RpoS regulatory pathway in Bb . In this report , we provide further evidence for the direct involvement of BosR in the activation of rpoS , and thus the RpoS regulon , in a second virulent strain of Bb . We also present evidence that BosR functions as a DNA-binding protein , but it has many features that markedly distinguish it from either of its Fur or PerR homologs . Defining this novel involvement of BosR relative to its control over the RpoN-RpoS pathway is important for elucidating Bb's host adaptation and pathogenesis , and could lead to innovative strategies for thwarting Lyme disease . This study also expands our understanding of bacterial sigma factor regulatory networks , and establishes a new paradigm of an additional transcriptional activator that is absolutely required for σ54–dependent gene regulation in a bacterial pathogen .
Previously , we [34] and others [32] found that the mutation of bosR in Bb strain B31 abolished RpoS , OspC and DbpA expression . However , there are some notable discrepancies between these two studies . Hyde et al . [32] found that the mutant exhibited defects in growth in vitro and in the expression of NapA ( or Dps , implicated in protecting DNA from damage during starvation or oxidative stress ) , whereas we [34] showed that the bosR mutant had normal growth and NapA expression comparable to WT Bb . It is also well-documented that transcriptional regulators and gene control mechanisms can differ widely among bacterial pathogens of the same species [36]–[38] , and variations in strain-specific genetic contents , gene expression profiling , and pathogenicity have been observed , in particular , for different strains of Bb [39]–[40] . Thus , to more broadly investigate the role of BosR in Bb pathogenesis and gene regulation , we generated bosR mutants in another virulent WT Bb strain ( strain 297 ) via homologous recombination . As in strain B31 [34] , bosR in strain 297 was predicted to be cotranscribed and form an operon with bb0646 and bb0648 ( Fig . S1A ) . To verify this , RT-PCR using specific primers and Bb cDNA was performed . As shown in Fig . S1B , amplicons spanning the junction of bb0646/bosR ( lane 3 ) , bosR/bb0648 ( lane 5 ) , or bb0646- bb0648 ( lane 6 ) were generated , indicating the operonic nature of bb0646 , bosR , and bb0648 . Of note , the sequences of this operon and its flanking genes ( bb0645 and bb0649 ) are identical between both strains 297 and B31 ( data not shown ) . Thus , the same strategies used in the creation of the bosR mutant in B31 [34] were employed to inactivate bosR in Bb 297 . When the suicide vector pOY24 was transformed into Bb 297 , two kanamycin-resistant bosR mutant clones ( OY08/A11 and OY08/F4 ) were obtained . To cis-complement the bosR mutation , the suicide plasmid pOY83 [34] containing the bb0649-bb0648-bosR-PflgB-aadA cassette was introduced into the bosR mutants . As a result , two streptomycin-resistant clones ( OY33/A6 and OY33/F7 ) were isolated . The inactivation and complementation of bosR in these strains were confirmed using PCR ( Fig . S1C ) . Moreover , RT-PCR and immunoblot analyses revealed that BosR expression was detected in both WT and the complemented strains , but not in the mutants ( Fig . S1D and E ) . To ensure that all mutants and complements retained the plasmids cp26 , lp25 and lp28-1 that are essential for Bb virulence [41]–[42] , PCR-based plasmid profiling was performed . As shown in Fig . S2A , the WT and bosR mutant OY08/A11 contained the same plasmid profiles . In addition , OY08/F4 and the complemented strains ( OY33/A6 and OY33/F7 ) contained the same plasmid profiles as that of the WT strain ( data not shown ) . The bosR mutant exhibited spirochetal morphology and movement identical to that of WT under dark-field microscopy . No discernable growth defect was observed when bosR was inactivated , and the mutants displayed similar growth patterns to that of WT ( Fig . S2B ) . The role of BosR in Bb strain 297 mammalian infectivity was assessed using the murine needle-challenge model of Lyme borreliosis [43]–[44] . All mice inoculated with WT or the complemented strains at a 104 inoculum became infected , and motile spirochetes were isolated from all tissues from these mice ( Table 1 ) . In contrast , the bosR mutants were not recovered from any mice inoculated with either 104 or 107 bacteria . These data establish that previous findings implicating BosR in Bb's infectivity and virulence [32] , [34] were not unique to strain B31 , and thus BosR appears to be essential for conferring virulence to other pathogenic strains of Bb . To determine whether the loss of Bb strain 297 virulence in the bosR mutant correlated with a loss in the expression of rpoS , ospC , and dbpA , as has been noted for strain B31 [32] , [34] , we further assessed the effect of the bosR mutation on gene expression . WT 297 , the bosR mutants , and the cis-complemented strains were cultured at 37°C in BSK-H medium at pH 6 . 8 , conditions under which the RpoS regulon is highly induced [10] , [14]–[15] , [25] . Cells were harvested at late-log phase and subjected to immunoblot and RT-PCR analyses . As shown in Fig . 1 , when bosR was inactivated , the expression of rpoS , ospC and dbpA , was essentially abolished at both the protein ( Fig . 1A ) and mRNA ( Fig . 1B ) levels . When the bosR mutation was cis-complemented , gene expression was fully restored . To further investigate the influence of BosR on gene expression , a bosR expression construct ( pOY112 ) was created using a newly-developed lac-based inducible expression system [45] . In pOY112 , bosR transcription was placed under the direct control of the IPTG-inducible PpQE30 promoter . Bb 297 transformed with pOY112 were cultivated in the presence of various amounts of IPTG . Late log-phase cells were harvested and analyzed by immunoblot . Relative to protein levels in cells grown in medium without IPTG , 50 µM of IPTG induced the production of BosR , as well as increased the levels of RpoS , OspC , and DbpA ( Fig . 1C ) , suggesting that BosR activates expression of these genes . Gene expression in WT 297 containing the empty vector ( grown under various concentrations of IPTG ) was not altered ( data not shown ) . We also complemented the bosR mutation in trans using the IPTG-inducible bosR expression construct pOY112 . As shown in Fig . 1D , when BosR was expressed from pOY112 by IPTG , expression of RpoS , OspC and DbpA was consequently restored . These data , consistent with previous studies [32]–[34] , further corroborate that BosR activates the expression of rpoS , ospC and dbpA . Of note , expression of rrp2 , rpoN or bb0646 ( the gene downstream from bosR in the bb0648-bosR-bb0646 operon ) was not affected by the bosR mutation ( Fig . 1B ) , implying that the phenotypes observed were not due to impairment in the expression of rrp2 , rpoN , or bb0646 . In addition , NapA levels were found to be similar in WT , the bosR mutant , and the complemented strains ( Fig . 1A ) . Although our data suggested that the expression of both ospC and dbpA was activated by BosR , it remained unknown how BosR controlled the expression of these two lipoproteins . Given our finding that rpoS transcription was abolished in the bosR mutant , and the observations that expression of ospC and dbpA are directly regulated by RpoS through RpoS-specific promoters [46]–[48] , we hypothesized that BosR likely regulated the expression of rpoS which , in turn , influences ospC and dbpA . To test this hypothesis , an IPTG-inducible rpoS expression construct was generated to render RpoS synthesis independent of BosR . This vector , pOY110 , was then introduced into the bosR mutant OY08/A11 . When RpoS was induced from pOY110 by IPTG , expression of OspC and DbpA was consequently rescued , although expression of BosR was absent ( Fig . 2 ) , suggesting that the controlled induction of RpoS could overcome the BosR deficiency and that BosR indirectly , rather than directly , controls OspC and DbpA expression . Recombinant BosR ( rBosR ) was hyper-expressed in E . coli and purified to apparent homogeneity . SDS-PAGE analysis indicated that BosR has an apparent molecular mass of ∼18 . 7 kDa ( Fig . 3A ) , which is in agreement with the apparent mass of native BosR in Bb ( Fig . 3B ) . Furthermore , when analyzed by size-exclusion chromatography , purified BosR eluted predominantly as a dimer with a molecular mass of ∼38 kDa ( Fig . 3C ) . Although recombinant BosR has been obtained previously and Zn2+ was found to affect BosR's in vitro binding to DNA [28]–[29] , it remained unclear whether BosR contains bound metal . Therefore , metal content analysis was carried out by inductively coupled plasma atomic emission spectrometry ( ICP-AES ) [27] , [49] . rBosR did not contain detectable levels ( <0 . 001 ppm ) of metal ions such as Fe or Mn ( Fig . 3D ) . Rather , it contained 1 . 4 mol of zinc per mol of protein . Moreover , in order to remove bound metal ( s ) from BosR , we also dialyzed the protein in the presence of 10 mM EDTA . However , 0 . 3 mol of zinc/mol of proteins remained in the demetallated BosR ( Fig . 3D ) , suggesting that the recombinant protein bound zinc avidly . Of note , these properties are typical of the dimeric bacterial Fur protein [50] or the Bacillus subtilis H2O2 stress response regulator PerR [51] . In silico analysis predicted that Bb BosR contains an N-terminal winged helix-turn-helix DNA binding domain and a C-terminal dimerization domain . Three-dimensional ( 3D ) protein modeling using the Swiss-model program ( http://swissmodel . expasy . org/ ) indicated that the structure of the DNA-binding domain of BosR is quite similar to the Vibrio cholerae Fur protein [52] and the B . subtilis PerR protein [51] ( Fig . S3 ) , suggesting that , consistent with previous reports [28]–[29] , BosR may be a DNA-binding protein . Moreover , our aforementioned data revealed that BosR impacted rpoS expression at the transcription level . Thus , EMSAs were performed to examine potential interactions between BosR and the rpoS promoter . Consistent with previous studies [28]–[29] , BosR bound to the promoter of Bb napA ( from −336 to +48 , relative to the ATG start codon ) ( Fig . 4A ) . However , BosR did not bind to the ospC or dbpBA promoters under our tested conditions ( Fig . 4B and C ) , providing support that BosR likely does not impact ospC and dbpA directly . Although BosR did not bind to the probe ZM126 that encompasses the rpoS promoter from −67 to −8 ( Fig . 4D ) , BosR , in a dose-dependent manner , bound to the rpoS promoter ( PrpoS ) encompassing 277 bp of the rpoS upstream region and 245 bp of the rpoS encoding region ( Fig . 5A ) . Of note , binding of rBosR generated multiple shifted bands , suggesting the possible existence of multiple BosR binding sites ( BSs ) in the probe . As an initial approach to identify the BosR binding sequence , DNase I footprinting assays were performed . As shown in Fig . 5B , three BosR BSs were recognized in the PrpoS DNA . Specifically , BosR BS1 , BS2 , or BS3 spanned regions of −193 to −137 , −120 to −46 , or −29 to +43 ( relative to the ATG start codon , where A is +1 ) , respectively ( Fig . 5C ) . To corroborate the DNase I-footprinting data , EMSA employing synthesized double-stranded ( ds ) DNA oligonucleotides representing different BosR BSs were performed . As shown in Fig . 6B and C , BosR bound strongly to both ZM132 ( representing BS1 ) and ZM127 ( representing BS2 ) . Moreover , the binding of BosR to labeled BS1 ( ZM132 ) or BS2 ( ZM127 ) was inhibited by the addition of a 200-fold excess of unlabeled DNA , but not inhibited by the addition of a 200-fold excess of non-specific competitor ZM126 DNA ( Fig . 6D and E ) , suggesting that BosR binds to BS1 ( ZM132 ) or BS2 ( ZM127 ) specifically . Binding to both probes also was abrogated by the addition of α-BosR antibody , but not by the addition of control rat serum ( Fig . 6D and E ) , indicating that the DNA shift was indeed caused by BosR . Of note , BosR , like Fur or PerR , putatively comprises an N-terminal DNA binding motif domain and a C-terminal domain involved in protein dimerization . Both domains are essential for Fur/PerR recognizing and binding to its target DNA as a homodimer [51]–[52] . Therefore , interaction with either domain of BosR by antibody may also interrupt protein binding to DNA and prevent DNA-BosR complex formation . Similarly , we also examined the binding of BosR to BS3 using EMSA . As shown in Fig . 7 , although BosR did not bind to probe ZM160 that corresponds to the 5′ sequence of BS3 , the protein bound to probe ZM161 ( encompassing rpoS from +4 to +63 ) . EMSAs employing target DNA sequences ( representing the three BosR binding sites ) exposed to increasing concentrations of rBosR were used as means of inferring BosR binding affinities for the three binding sites . As shown in Fig . 6 and 7 , concentrations of 50 , 20 , or 200 nM of rBosR induced shifts by ZM132 ( BS1 ) , ZM127 ( BS2 ) , or ZM161 ( BS3 ) , respectively , suggesting that BosR has an affinity for these DNA targets in the order of BS2>BS1>BS3 . In addition , when 200 nM of rBosR was used , only a slight proportion ( <10% ) of ZM161 ( BS3 ) was shifted ( Fig . 7 ) , and probe ZM161 could not be saturated even by 10 , 000 nM of BosR ( data not shown ) . To more precisely assess the affinity of BosR for BS1 and BS2 , we measured the amount of bound DNA as a function of BosR concentration in EMSA assays ( Fig . 8A ) . The dissociation binding constants ( Kd ) for BS1 ( ZM132 ) or BS2 ( ZM127 ) were 210 . 2 or 36 . 6 nM , respectively . The relative affinities of these two DNA elements for BosR were also assessed by competition EMSA analysis ( Fig . 8 , B and C ) . Binding of labeled BS1 or BS2 was not inhibited by the non-specific competitor ZM126 ( NS ) , but was inhibited by unlabeled ( cold competitor ) BS1 or BS2 , respectively . Moreover , binding of labeled BS1 was inhibited approximately 90% by the addition of 200-fold unlabeled BS1 , but was completely competed out by 50-fold unlabeled BS2 ( Fig . 8B ) . The addition of 200-fold of unlabeled BS1 competed out only 15% of BS2 binding ( Fig . 8C ) . These data indicate that BosR has a higher affinity for BS2 than for BS1 . In silico analysis indicated that BosR contained two potential CX2C Zn2+ binding motifs in its C-terminus ( located at residues of 114–117 and 153–156 ) . These types of Zn2+ structural sites are crucial for Fur dimerization and binding DNA as a homodimer [30] , [53] . Accordingly , we found that BosR bound Zn tightly ( Fig . 3D ) . Moreover , consistent with previous observations [28] , our purified BosR appeared to exist principally as a dimer in solution ( Fig . 3 ) . Therefore , BosR may bind to the rpoS promoter as a homodimer , suggesting that the BosR binding sequence ( s ) may be a direct repeat ( DR ) sequence . In agreement with this assumption , close inspection of the three BosR BSs revealed a DR sequence ( TAAATTAAAT ) ( Fig . 5C ) . Of note , this sequence also consists of two contiguous pentamer direct repeats ( TAAAT ) . More specifically , BS1 contains one perfect DR at its 5′ sequence and one imperfect DR at its 3′ sequence; BS2 contains one perfect DR and one imperfect DR in the sequence upstream of the -24/-12 RpoN binding site; and BS3 contains one imperfect DR sequence at its 3′ . Of note , in both BS1 and BS2 , the DR1 and DR2 are located on opposite DNA strands . We hypothesized that if the DR is essential for binding to BosR , mutational changes in the sequence should abolish or inhibit BosR binding . Along these lines , we initially synthesized two DNA fragments , ZM155 and ZM156 , representing the 5′ or 3′ of BS1 , respectively ( Fig . 9A ) . Both ZM155 and ZM156 contain one DR sequence . After labeling these DNA fragments with digoxigenin , each DNA fragment was mixed with BosR and EMSAs were performed . As shown in Fig . 9B , BosR bound to both fragments . However , when the DR was mutated , the binding of BosR to either DNA fragment was abolished ( Fig . 9B ) , strongly supporting the notion that the DR sequence is critical for BosR binding . Using this same strategy , we also examined the two DRs in BS2 . As shown in Fig . 10 , although BosR still bound to probe ZM149 having sequences downstream of the -24/-12 site scrambled , BosR binding to BS2 was abolished when sequences upstream of the -24/-12 site were scrambled ( ZM147 ) , suggesting that the functional BosR binding sites are located in the sequence upstream of the -24/-12 site . Because sequences flanking the binding motif often play important roles in protein-DNA interactions , we synthesized another dsDNA ( ZM212 ) to represent the 5′ of BS2 , allowing added flanking sequences to the predicted DR sequences . EMSAs indicated that BosR still bound to probe ( ZM213 ) with the DR1 mutated ( Fig . 10B ) . When a mutation was introduced into DR2 ( ZM214 ) , BosR binding was dramatically reduced ( Fig . 10B ) . Moreover , when both DR sequences were mutated , protein binding was completely abolished ( Fig . 10B ) . These data suggest that DR1 and DR2 in BS2 are required for BosR binding . To identify other Bb genes potentially regulated by BosR , Katona et al . [28] performed a BlastN analysis based on the Per box consensus sequence . However , the DR identified in our current study is disparate from the Per box . To uncover additional BosR-regulated genes , we searched the Bb genome by using the Regulatory Sequence Analysis Tools ( http://rsat . ulb . ac . be/rsat ) , and queried for genes containing a perfect DR sequence in putative promoter regions . Gene promoter regions were defined as sequences from −400 to +50 bp ( relative to the ATG start codon ) . The results are shown in Table 2 . A total of 60 Bb genes were found to harbor one or multiple perfect DR sequence in their promoter regions . Thirty-one genes were located on the main chromosome , and 29 genes were on linear or circular plasmids . Of the 31 chromosomal genes , 16 genes encode proteins with assigned functions and15 genes encode hypothetical proteins . More importantly , 13 of these genes were found to be regulated by BosR in our recent microarray analysis [34] , further supporting that the DR sequence is important for BosR binding . Previously , BosR was reported to bind to the Bb napA promoter ( PnapA ) . Using footprinting assays , Boylan et al . [29] found that BosR protected a 50-bp region located at -222 to -173 ( relative to the ATG start codon ) in PnapA . Katona et al . [28] , using EMSAs , reported that BosR also bound to a DNA fragment encompassing PnapA from −152 to +3 . In addition , the latter researchers also reported that BosR bound to upstream regions of bosR . Interestingly , these two genes were not identified in our search ( Table 2 ) . However , when scrutinizing the upstream regions of bosR and napA , multiple imperfect DR sequences were detected ( Fig . 11A , 12A ) . Therefore , EMSAs using synthesized dsDNA were employed to examine the roles of these imperfect DR sequences in BosR binding . Specifically , two dsDNA fragments , ZM215 and ZM217 , were used to represent the BosR binding region in PnapA identified in previous studies [28]–[29] . As shown in Fig . 11B , BosR bound to both DNA fragments . When a mutation was introduced into the DR , binding of BosR to each probe was abolished . Similar data were also obtained for the probe ZM219 representing the bosR upstream region; BosR binding to the probe was abolished when the predicted DR was mutated ( ZM220 ) ( Fig . 12B ) . These data further substantiate the critical role of the DR in BosR binding . Bb RpoN activates rpoS directly through a canonical -24/-12 promoter , and mutation of G at -24 to T in the rpoS promoter resulted in a significant diminishment of in vitro RpoN binding and a dramatic decrease in rpoS expression [8] , [11] , [24] . Our analysis of BS2 , the site exhibiting the highest affinity for BosR among the three BosR binding sites in the rpoS promoter , revealed that a perfect DR sequence is located just upstream and adjacent to the -24/-12 σ54 promoter ( Fig . 13A ) . Given the importance of this locus for rpoS transcription , we further examined this site more closely using EMSA . As shown in Fig . 13B , when nucleotides in the -24/-12 site were mutated ( ZM157 , mutations of G-24T , G-25T , and C-12A ) , the binding of BosR was not altered . In contrast , when a mutation was introduced into the DR upstream of the -24/-12 site ( ZM166 , AATT replaced by GGCC ) , BosR binding was abolished ( Fig . 13 ) . These data strongly suggest that the key nucleotides for the binding of BosR and σ54 to the rpoS promoter are different .
The essential role of the RpoN-RpoS pathway for virulence expression by the Lyme disease bacterium is now well documented [8]–[11] , [13] , [16]–[17] , [24]–[25] . However , a number of the molecular details involved in the activation of this pathway have remained obscure . Whether Rrp2 , as an EBP , is present as a dimer and assembles into a functional hexamer or heptamer when activated is unknown . It is also unclear how Rrp2 interacts with Eσ54 to modulate RpoN-dependent rpoS expression in the absence of demonstrable specific binding to the rpoS promoter or upstream region [8] , [24] , but there is precedence for this anomaly in another bacterial pathogen , Campylobacter jejuni ( its EBP FlgR activates the σ54-dependent flagellar genes independent of DNA-binding ) [54] . Moreover , the mammalian host factors that influence rpoS expression [12]–[13] have not been identified . Recently it was shown in Bb strain B31 that a mutation in bosR led to a loss of mouse infectivity , as well as a block in the expression of the virulence-associated factors OspC and DbpA [32] , [34]–[35] . Given that both ospC and dbpA are under the control of RpoS [46]–[48] , we proposed that BosR somehow was involved in the activation of the RpoN-RpoS regulatory pathway [34] . However , these data and the resultant hypothesis emanated from studies involving only one virulent strain ( strain B31 ) of Bb . It has long been established that genetic regulators and control mechanisms can vary widely among strains of the same virulent species of pathogenic bacteria [36]–[38] . Hence , the question remained whether the novel RpoN-RpoS regulatory pathway and the important role of BosR were common to other virulent strains of Bb . The results of our study herein now confirm that the inactivation of bosR , which prevents activation of the RpoN-RpoS pathway by blocking the expression of rpoS , and , in turn , prevents expression of the virulence-associated genes ospC and dbpA , is not unique to a single virulent strain of Bb . BosR and its control over RpoN-RpoS activation thus appears to be an important global regulatory pathway essential for virulence expression by the Lyme disease spirochete . Our findings also provide new insights into the function of BosR . BB0647 ( BosR ) was originally predicted to be a Fur homologue [7] , [28] . Given the lack of precedence for interplay between a Fur homologue and an alternative sigma factor , such as RpoN or RpoS , in other bacteria , it was entirely unexpected that BosR would play a role in the induction of the RpoN-RpoS pathway [32] , [34] . Furthermore , the nomenclature of “BosR” , as Borrelia oxidative stress regulator , was derived from the sequence similarity of BosR to the B . subtilis PerR , and an observation [29] that , in the heterologous E . coli , BosR activated the expression of Bb napA implicated in the oxidative stress response [55] . It was previously reported that NapA production was inhibited in a bosR mutant of Bb , and that the mutant displayed an in vitro growth defect [32] . Our former B31 mutant deficient in bosR exhibited no such growth defect , and the expression of napA was not significantly altered in the bosR mutant [34]; these same wild-type-like phenotypes were also observed in our current bosR mutant derived from strain 297 ( Fig . 1B , Fig . S2B ) . Given these more recent findings , and the current paucity of compelling data that directly link BosR to an oxidative stress response in Bb , it is thus still premature to conclude that BosR is involved in modulating oxidative stress in Bb . Finally , despite the characterization of BosR as a Fur homologue , it also remains unclear whether BosR plays a role in regulating transition metal homeostasis in Bb . Our EMSA data clearly indicated that BosR binds to the rpoS promoter , suggesting that BosR directly influences rpoS expression . Moreover , DNA footprinting assays revealed three BosR-protected regions in rpoS . The occupation of multiple , rather than one , binding sites might stabilize and secure BosR binding to the rpoS promoter region . BosR exhibited binding affinity for these three sites in the order of BS2>BS1>BS3 . Among these three sites , BS2 was found juxtaposed to and partially overlapping with the -24/-12 RpoN binding site , whereas BS1 is located upstream of BS2 . BS3 is located in sequences downstream of BS2 , or more specifically , in the RpoS-encoding region . A previous study revealed that , when Bb was grown in vitro , a minimal PrpoS ( starting from the -24/-12 RpoN binding site ) , which contains only one intact BosR BS ( BS3 ) , is able to express RpoS at the same level as PrpoS containing all three BosR BSs [24] . Furthermore , RpoS expressed from the minimal PrpoS restored mouse infectivity to the rpoS mutant [24] . These data imply that PrpoS with only BS3 is sufficient to drive rpoS transcription , suggesting that our data of BosR binding to BS3 may be physiologically relevant . However , BS1 and BS2 may be required to coordinate rpoS expression under different in vivo conditions of the two diverse niches of Bb's complex life cycle . Our unanticipated finding that BS3 for BosR is located within the RpoS-encoding region is rare but not unprecedented for transcriptional activators; binding sites for other bacterial regulatory proteins have been noted to occur in the coding regions of their target genes [56]–[57] . It is thus possible that BS3 located in the rpoS encoding region may somehow strengthen the binding of BosR to PrpoS , and then cooperate in opening the RpoN-RNAP closed complex . Or , BosR binding at BS3 might allow rpoS expression to be controlled more tightly , especially if rpoS transcription requires transient modulation . A major finding of this study is the identification of a novel DNA binding sequence for BosR . As a Fur homologue , BosR dimers were reported [28]–[29] to bind in vitro to the Bb napA promoter , the upstream regions of bosR and bb0646 , and DNA containing a Fur box ( GATAATGATAATCATTATC ) or Per box ( TTATAAT-ATTATAA ) . In general , the Fur box is interpreted as two 9-bp inverted repeats ( GATAATGAT ) , or two heptamer inverted repeats ( TGATAAT ) , or three hexamer repeats ( GATAAT ) , whereas the Per box is recognized as two inverted repeat ( TTATAAT ) [53] , [58] . Despite the facts that the rpoS promoter contains neither a Fur ( or Per ) box , nor has significant similarity with the Bb napA promoter or the upstream sequences of bosR and bb0646 , BosR binds to the rpoS promoter . More importantly , we identified a DR sequence ( TAAATTAAAT ) that is critical for BosR binding . This assertion is strongly supported by several lines of evidence . First , the DR sequence is present in all three BosR BSs . Second , the DR sequence was identified in the promoter regions of 13 genes already known to be influenced by BosR . Third , imperfect DR sequences are present in previously-established BosR-binding DNA fragments , such as PnapA and a bosR upstream region . Finally , mutations in the DR severely reduced or completely abolished DNA binding by BosR . Of note , the DR sequence is markedly different from the direct or inverted repeats present in Fur or Per boxes . Thus , BosR appears to be able to recognize different DNA sequences , including the Fur box consensus , the Per box consensus , and the rpoS promoter element ( containing the DR sequence ) . Such promiscuous DNA recognition activity has been observed previously for the Bradyrhizobium japonicum Fur protein [59]; in addition to binding to the Fur box consensus , B . japonicum Fur also binds in vitro to the irr promoter ( with similar affinity ) , but , the irr promoter does not contain a Fur box . Rather , it contains three essential direct repeat sequences of TGCATC that differ markedly from the direct repeats ( GATAAT ) or inverted repeats ( GATAATGAT ) in the Fur box [59] . The mechanistic details of this anomaly remain unknown . One possibility for BosR is that its binding properties in vitro may depend largely on DNA conformation . However , when analyzing the conformation of the dsDNA ( including DNA containing mutated DRs ) used in our EMSAs by PREDICTOR ( http://www . farwer . staff . shef . ac . uk/PREDICTOR ) , which is a program calculating the three-dimensional atomic structure of dsDNA , no obvious differences in DNA conformation were revealed . Moreover , although BosR is predicted to share a similar three dimensional structure with Fur and the PerR protein ( Fig . S3 ) , BosR may harbor some subtle , unique structural feature ( s ) ( undetected by protein modeling ) that confer its DNA binding traits . Alternatively , under different in vivo conditions ( tick vector or mammalian hosts ) , BosR may display alternative structural conformations that differentially regulate gene expression . As such , differing conformations may prompt BosR to bind to different DNA sequences ( or with varying affinities ) . It is perhaps noteworthy that , after decades of intensive work , there is still much controversy over the molecular mechanisms and biochemistry of how Fur operates as a transcriptional repressor [60] . Thus , it is not surprising that there is yet much to learn regarding the molecular mechanism ( s ) that allow BosR to act as a regulator . Nonetheless , our finding that BosR , as a Fur or PerR homolog , recognizes disparate DNA sequences not only hints that the well-established model for Fur ( or PerR ) -DNA interaction may warrant further refinements , but also suggests that BosR employs mechanisms different from Fur or PerR to regulate gene expression . In addition , the recognition that BosR depends on a novel direct repeat for its binding to the rpoS promoter serves as a strong foundation for further mechanistic studies . Several aspects of our study engender a number of provocative possibilities surrounding the function of BosR as a transcriptional enhancer for rpoS . First , our mutagenesis experiments revealed that different key nucleotides are required for BosR or RpoN binding to the rpoS promoter , implying that BosR and RpoN may bind to different faces of the DNA helix ( comprising the rpoS promoter ) . Second , that BS2 is immediately adjacent to the -24/-12 RpoN binding site tempts speculation that BosR and RpoN ( and possibly Rrp2 ) may interact with one another at the -24/-12 site to initiate rpoS transcription . However , results from E . coli-based two-hybrid assays thus far have failed to show interactions between BosR and RpoN ( or Rrp2 ) ( unpublished data ) . These results , however , are not unexpected , because there is no precedence for interactions between Fur/PerR proteins and RpoN or an EBP ( e . g . Rrp2 ) in any bacterial system . Despite that , it is not impossible that BosR may transiently interact with RpoN or Rrp2 in vivo . It also remains possible that BosR may act as a critical molecule to recruit Rrp2 and/or RpoN to the rpoS promoter and the Eσ54-CC . The binding of BosR dimers to one face of the rpoS promoter at multiple sites may lead to DNA bending or other conformational changes that may facilitate the binding of Eσ54 to the -24/-12 site on the other strand of the promoter . BosR , Fur , and PerR share structural similarity ( Fig . S3 ) and all three are zinc- ( or other metal- ) containing regulatory proteins . However , there are other key features of BosR that markedly distinguish it from its putative Fur or PerR homologs . With Fur , metal-dependent DNA binding acts primarily as a repressor to avoid cellular iron toxicity [30] , [53] , [58] , [60] . Its positive regulatory role is often indirect , via the Fur-regulated anti-sense regulatory small RNA , RhyB [30] , [60] . In the case of PerR , transcriptional activation of its target genes ( involved in protecting the bacterial cell against oxidative stress ) occurs when the metal-bound PerR dissociates from the promoter [51] , [60] . Of the three regulators , BosR is the only one that works in concert with an alternative sigma factor ( RpoN ) and an EBP ( Rrp2 ) , and the only one that appears to activate gene transcription directly by DNA binding . However , at this time we cannot rule out the less likely possibility that BosR may prevent the binding of a repressor that blocks rpoS transcription . Nonetheless , from the metal ( zinc ) content of rBosR , it is tempting to speculate that BosR's DNA-binding activity may be metal-dependent . Although zinc was found in rBosR , the question remains whether zinc is the physiologically relevant metal that confers normal activity to native BosR . It is thus not out of the realm of possibility that other metal ( s ) may be physiological relevant during Bb's existence in ticks or mammalian hosts . Further studies are warranted to investigate these possibilities , although many of the potential experimental approaches have substantial obstacles . Our findings also reveal a new aspect of bacterial σ54-dependent gene activation and expands our understanding of transcriptional regulation by alternative sigma factors in general . Traditionally , for all known bacterial σ54-dependent genes , transcriptional activation requires only the cognate activator EBP ( ATPase ) [3]–[4] . For some σ54-dependent promoters , maximal induction relies on one or several auxiliary factors . For example , IHF , as a DNA-bending protein , can promote the interaction between Eσ54 and an EBP via DNA looping . In this case , however , IHF acts only as an architectural element to facilitate formation of the loop . It is the EBP , rather than the IHF protein , that modulates σ54-dependent gene expression [61] . In E . coli and Salmonella typhimurium , in the presence of arginine , the arginine repressor ArgR can induce the expression of the σ54-dependent astCADBE operon [62]–[63] . Nonetheless , ArgR plays only an accessory , rather than essential , role in the expression of astCADBE . In the absence of ArgR , genes are still expressed , although at more moderate levels . In response to flavonoids , the Azorhizobium caulinodans transcriptional activator NodD induces the transcription of NifA-RpoN-controlled NodA operon at an early stage [64] . In mature nitrogen-fixing nodules , the nodA gene is still transcribed in the nodD mutant in response to nitrogen-oxygen availability . In addition to the putative EBP ( Rrp2 ) , BosR is directly involved in the transcriptional activation of σ54-dependent rpoS in Bb . Unlike ArgR , NodD , and other accessory factors involved in maximizing the induction of σ54-dependent genes , BosR is essential for rpoS transcription in Bb . To our knowledge , this is the only demonstration , in any bacterial σ54-dependent transcriptional system , that transcription of a σ54-dependent gene requires an additional transcriptional activator . Bioinformatics indicate that homologues of BosR and σ54 are not only conserved in other Borrelia species ( such as B . garinii and B . afzelii ) , but may be encoded in numerous other bacterial species including Bordetella , Burkholderia , Shewanella , Campylobacter , Clostridium , Bacillus , Listeria , and others . Given this wide distribution of BosR homologs , our study may have broader significance in understanding the regulatory control over RpoN- or RpoS-dependent genes in other pathogenic bacteria .
Strains and plasmid used in this study are listed in Table 3 . Infectious Bb strain 297 [65] was used as the WT strain throughout this study . Bb was routinely cultured at 37°C and 5% CO2 in either BSK-II medium [66] or BSK-H medium ( Sigma ) supplemented with 6% rabbit serum ( Pel-Freeze ) . When appropriate , supplements were added to media at following concentrations: kanamycin , 160 µg/ml; streptomycin , 150 µg/ml . Spirochetes were enumerated by dark-field microscopy . This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . All animal procedures were approved by Institutional Animal Care and Use Committee at UT Southwestern Medical Center ( Animal Protocol Number 0165-07-14-1 ) . The bosR mutant OY08 was created by allelic exchange in Bb 297 using a suicide vector pOY24 [34] . The mutation in bosR was cis-complemented by transforming a suicide vector , pOY83 [34] , into the bosR mutant , generating OY33 . Transformation of Bb was performed as previously described [67] . Plasmid contents of all Bb strains were determined by PCR using specific primers . To artificially control BosR or RpoS expression in Bb , gene expression constructs were generated using a newly-developed lac-based inducible expression system [45] . First the PflaB-BblacI-PpQE30 cassette from pJSB252 [45] was ligated into pJD54 digested with BglII and BamHI , which generated pOY99 . 2 . Then rpoS or bosR was amplified from Bb 297 and cloned into pOY99 . 2 digested with NdeI and BglII , generating pOY110 or pOY112 , respectively . In these constructs , rpoS or bosR transcription was directly controlled by the IPTG-inducible T5 promoter in pQE30 ( PpQE30 ) . The infectivity of Bb clones was assessed using the murine needle-challenge model of Lyme borreliosis [43]–[44] . C3H/HeN mice ( Charles River Laboratories ) were infected via intradermal injection with various concentrations of Bb . At 4 weeks post inoculation , mice were sacrificed and skin , heart , and joint tissues were collected and cultured in BSK supplemented with 1× Borrelia antibiotic mixture ( Sigma ) . The outgrowth of spirochetes in these cultures was assessed using darkfield microscopy . Recombinant BosR was produced in E . coli using the Champion pET SUMO protein expression system ( Invitrogen ) . Briefly , bosR was amplified using primers ZM69F and ZM69R , and ligated into pET SUMO vector through TA cloning such that the resultant construct pOY73 encoded a fusion protein with a His6 –SUMO tag at its N terminus . Constructs were confirmed using PCR amplification , restriction digestion , and sequence analysis . The resulting construct , pOY73 , was then transformed into E . coli strain BL21-DE3 . After induction with 1 mM IPTG ( Sigma ) , recombinant His6 –SUMO-tagged BosR was purified using a Ni-NTA spin column under native conditions according to the manufacturer's instruction ( Qiagen ) . The N-terminal His6 –SUMO tag was removed via cleavage with SUMO protease ( Invitrogen ) at 30°C for 4 h in the buffer A containing 20 mM Tris , 20 mM NaCl , 100 mM L-arginine , pH 7 . 5 . The protease digestion mixture was concentrated and buffer exchanged with buffer A using an Amicon ultracentrifuge filter device ( Millipore ) with a 10 , 000 molecular weight exclusion limit . The concentrated protein was applied to a HiLoad 16/60 Superdex 200 prep grade column and purified on an Äkta fast performance liquid chromatography system ( GE Healthcare ) using buffer A . Subsequent to elution , peak fractions were analyzed by SDS-PAGE and Western Blot . At this stage , the protein was pure to apparent homogeneity and predominantly present as dimer . Fractions containing pure BosR with a homogeneity >95% were pooled . Protein concentration was determined using the BCA protein assay kit ( Thermo Scientific ) . Rat polyclonal antibody against BosR , Ab-BosR , was generated as previously described [34] . Metal contents in protein or buffer solutions ( as references ) were measured using inductively coupled plasma atomic emission spectrometry ( ICP-AES ) , by the Research Analytical Laboratory , University of Minnesota . Proteins were demetallated by dialyzing samples with 10 mM ETDA for 24 h , as described previously [49] . Three independent tests were performed , and average metal concentrations with standard deviations were presented . SDS-PAGE and immunoblot analysis were carried out as previously described [34] . Briefly , purified protein samples or a volume of whole cell lysate equivalent to 4×107 spirochetes were loaded per lane on a 12 . 5% acrylamide gel . Resolved proteins were either stained with Coomassie brilliant blue or transferred to nitrocellulose membrane for immunoblot analysis . BosR was detected using the anti-BosR rat polyclonal antibody , Ab-BosR . Rrp2 , RpoS , OspC , and DbpA were detected using anti-Rrp2 monoclonal antibody 5B8-100-A1 , anti-RpoS monoclonal antibody 6A7-101 , anti-OspC monoclonal antibody 1B2-105A , or anti-DbpA monoclonal antibody 6B3 , respectively . To confirm equal loading of bacteria in each lane , immunoblotting for the flagellar core protein ( FlaB ) was performed using a chicken IgY anti-FlaB antibody . Immunoblots were developed colorimetrically using 4-chloro-1-napthol as the substrate or by chemiluminescence using ECL Plus Western Blotting Detection system ( Amersham Biosciences ) . Primers used in EMSA are listed in Table S1 . PCR-amplified or synthesized DNA probes were end-labeled with digoxigenin using recombinant terminal transferase ( Roche Applied Science ) . The labeled probe ( 30 fmol ) was mixed with various amounts of purified BosR in 20 µl of the gel shift binding buffer containing 20 mM Hepes ( pH 7 . 5 ) , 50 µg/ml poly[d ( A-T ) ] , 5% ( w/v ) glycerol , 1 mM DTT , 100 µg/ml BSA , 1 mM MgCl2 , and 50 mM KCl . After being incubated at 37°C for 30 min , the samples were analyzed by 5% non-denaturing polyacrylamide gel electrophoresis at 80V for 1–3 h . Then DNA was transferred onto a positively charged Nylon membrane ( Roche Applied Science , USA ) by electroblotting . The digoxigenin-labeled probes were subsequently detected by an enzyme immunoassay using an antibody ( anti-digoxigenin-AP , Fab fragments ) and the chemiluminescent substrate disodium 3- ( 4-methoxyspiro {l , 2-dioxetane-3 , 2′- ( 5′-chloro ) tricyclo[3 . 3 . 1 . 13 , 7]decan}-4-yl ) phenyl phosphate ( CSPD ) ( Roche Applied Science , USA ) . For DNase I footprinting , DNA probe ( PrpoS ) containing the rpoS promoter was PCR-amplified using primers 88F and 88R . The resultant DNA was labeled with T4 polynucleotide kinase ( NEB ) and γ-32P-ATP ( PerkinElmer ) . A 50-µl reaction containing the radiolabeled probe ( 300 fmol ) and various amounts of BosR was incubated in the gel shift binding buffer at 37°C for 30 min . Then incubated DNA was digested with 0 . 01 unit of DNase I ( Invitrogen ) at room temperature for 2 min . The reaction was terminated by adding 100 µl of DNase I stop solution containing 200 mM Tris-HCl ( pH 7 . 5 ) , 50 mM EDTA , 2% SDS , 200 µg/ml of proteinase K , and 250 µg/ml of glycogen , followed by phenol/chloroform extraction . Then DNA was precipitated with 20 µl of 3M ammonium acetate , and 600 µl of 100% ethanol at −20°C overnight . The precipitated DNA was washed with 70% ethanol and air-dried . The pellet was resuspended in 10 µl of formamide dye ( 90% formamide , 1× TBE , and 0 . 02% bromophenol blue/xylene cyanol ) and analyzed in a 6% polyacrylamide/7 M urea gel at 75 W for 2 h . The gel was transferred onto a chromatography paper ( Fisher ) , dried , and exposed in a PhosphorImager screen . The signals were detected by Typhoon 9200 PhosphorImager ( GE Healthcare ) . Spirochetes were grown in BSK at 37°C under 5% CO2 , and harvested when bacterial growth reached a density of 5×107 cells per ml . Total RNA was isolated using Trizol ( Invitrogen ) according to the instructions . After genomic DNA was digested using RNase-free DNase I ( GenHunter Corporation ) , RNA was further purified using RNeasy Mini Kit ( Qiagen ) . cDNA was generated from 1 µg of RNA using the SuperScript III Platinum Two-step qRT-PCR kit according to the manufacturer's protocol ( Invitrogen ) .
|
Lyme disease , caused by the bacterium Borrelia burgdorferi ( Bb ) , remains the most common arthropod-borne illness in the United States . A critical strategy for Bb to maintain its presence in nature is adaptation to its diverse tick and mammalian ( mouse ) hosts . To accomplish this , Bb encodes a potential gene regulator , BB0647 ( BosR ) . Herein , we confirmed that BosR is essential for Bb to establish mammalian infection . We then found that purified recombinant BosR bound to the promoter DNA ( regulatory region ) of rpoS , suggesting that BosR directly controls the expression of the rpoS gene . This study has revealed a new mechanism of bacterial gene control . The discovery that BosR governs Bb's virulence may lead to new strategies to interrupt the bacterium's complex life cycle .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases/bacterial",
"infections",
"microbiology/medical",
"microbiology"
] |
2011
|
BosR (BB0647) Controls the RpoN-RpoS Regulatory Pathway and Virulence Expression in Borrelia burgdorferi by a Novel DNA-Binding Mechanism
|
Ascariasis remains the most common helminth infection in humans . As an alternative or complementary approach to global deworming , a pan-anthelminthic vaccine is under development targeting Ascaris , hookworm , and Trichuris infections . As16 and As14 have previously been described as two genetically related proteins from Ascaris suum that induced protective immunity in mice when formulated with cholera toxin B subunit ( CTB ) as an adjuvant , but the exact protective mechanism was not well understood . As16 and As14 were highly expressed as soluble recombinant proteins ( rAs16 and rAs14 ) in Pichia pastoris . The yeast-expressed rAs16 was highly recognized by immune sera from mice infected with A . suum eggs and elicited 99 . 6% protection against A . suum re-infection . Mice immunized with rAs16 formulated with ISA720 displayed significant larva reduction ( 36 . 7% ) and stunted larval development against A . suum eggs challenge . The protective immunity was associated with a predominant Th2-type response characterized by high titers of serological IgG1 ( IgG1/IgG2a > 2000 ) and high levels of IL-4 and IL-5 produced by restimulated splenocytes . A similar level of protection was observed in mice immunized with rAs16 formulated with alum ( Alhydrogel ) , known to induce mainly a Th2-type immune response , whereas mice immunized with rAs16 formulated with MPLA or AddaVax , both known to induce a Th1-type biased response , were not significantly protected against A . suum infection . The rAs14 protein was not recognized by A . suum infected mouse sera and mice immunized with rAs14 formulated with ISA720 did not show significant protection against challenge infection , possibly due to the protein’s inaccessibility to the host immune system or a Th1-type response was induced which would counter a protective Th2-type response . Yeast-expressed rAs16 formulated with ISA720 or alum induced significant protection in mice against A . suum egg challenge that associates with a Th2-skewed immune response , suggesting that rAS16 could be a feasible vaccine candidate against ascariasis .
Ascaris lumbricoides , Trichuris trichiura and the hookworm Necator americanus are the three major soil-transmitted helminths ( STH ) that infect more than one billion poor people in the world and are the leading neglected tropical diseases ( NTDs ) in terms of disability-adjusted life years ( DALYs ) [1] . New estimates from the Global Burden of Disease Study 2015 indicate that approximately 761 million people are chronically infected with A . lumbricoides , resulting in 2 , 700 annual deaths from ascariasis [2 , 3] . Global control of STH infections depends on the mass drug administration of anthelminthics such as albendazole or mebendazole targeting children between the ages of 1–14 years [4] . However , the rapid rates of post-treatment re-infection [5] , potential drug resistance [6] , low treatment coverage for children [7] , and low access to clean water [8] compromise the effect of anthelminthics alone as a suitable means to control or eliminate STH infections . Indeed , two systematic reviews have largely failed to confirm the beneficial effects of periodic deworming [9 , 10] . Thus , the development of a multivalent pan-anthelminthic vaccine targeting all three major STH infections would be a desirable biotechnology to prevent parasite reinfection and advance efforts for the control and elimination of these diseases [11] . To advance such a strategy , two major N . americanus hookworm vaccine antigens are undergoing clinical vaccine tests , but there is a need to simultaneously develop A . lumbricoides and T . trichiura candidate antigens as suitable vaccines to be integrated within the human hookworm vaccine development program [12] . With regard to the development of A . lumbricoides vaccine antigens , the genetically highly homologous pig parasite , Ascaris suum , is commonly used as a model to identify and evaluate vaccine candidates . A . suum and A . lumbricoides are morphologically , immunologically , and genetically very similar [13 , 14] and might even be subspecies variants . Indeed , A . suum has been shown to be an important cause of human ascariasis [15] . Similar to its natural host , the pig , mice can be infected with A . suum eggs and larvae will be released into the mouse intestine from which they will migrate to the lungs . However , in mice , these larvae cannot return to the intestine to develop into adult Ascaris worms [16–18] . Nonetheless , the mouse model has proven to be a valuable tool in the identification and evaluation of vaccine candidates against Ascaris infections [19–21] . Indeed , mice infected with A . suum eggs produced significant protection against A . suum egg challenge as judged by the significant reduction in the number of larvae migrating to the lungs or livers [19 , 20 , 22] and also by reduced lung pathology [23] . Using serum from infected mice or rabbit , several immunodominant antigens have been identified including As16 [20] , As14 [19] , As24 [24] , As37 [21] and As-Enol ( enolase ) [25] , and protective immunity has been induced by immunization with recombinant proteins [19 , 20 , 26 , 27] and with DNA [28] . As14 and As16 were the first two antigens previously identified by the Tsuji lab through immunoscreening of an A . suum cDNA library with serum from Ascaris infected rabbit [19 , 20] . They share 47% amino acid sequence identity and similar localization ( in larva and adult stages , as well as in excretory/excretory products ) [11] . Intranasal immunization with Escherichia coli expressed recombinant As16 and As14 conjugated with the cholera toxin B subunit ( CTB ) produced significant protection against A . suum infective egg challenge in mice [19 , 20] . In addition , rAs16 induced protection in a pig animal model [29] , and mice fed with As16-transgenic rice mixed with CTB were also protected against A . suum infection [30] . As14 fused with CTB was also successfully expressed in transgenic rice , but there was no oral immunization and protection reported [31] . Notably though , without CTB as the adjuvant , neither As16 nor As14 were able to induce protective immunity in any model . Here , we report the production of recombinant As16 and As14 in the yeast Pichia pastoris , a eukaryotic expression system with scalability and without the concern for endotoxin contamination as in E . coli-expressed proteins [32] . Mice immunized with yeast-expressed rAs16 mounted a Th2-biased immune response and showed significant protection in terms of lung larval reduction . However , the same immunization regime for rAs14 did not induce any protection in mice . The immunological mechanisms underlying rAs16-induced protection were evaluated compared to rAs14 that did not induce protection . The results in this study provide a feasible approach to developing a vaccine against ascariasis on the basis of the yeast-expressed rAs16 that can be produced at low cost and formulated with alum , an FDA-approved adjuvant for human use that induces the Th2 immune response necessary to achieve anti-Ascaris immunity .
All animal procedures were conducted in accordance with Baylor College of Medicine Institutional Animal Care and Use Committee ( IACUC ) approved protocol AN-6297 in compliance with the Animal Welfare Act , PHS Policy , and other Federal statutes and regulations relating to animals and experiments involving animals . A . suum eggs were originally obtained from an adult female worm collected from an infected pig at a pig slaughter house near Belo Horizonte , Brazil , and maintained in 0 . 2 N H2SO4 until most of them had developed into the embryonated infective stage ( 50–250 days ) . The infective embryonated eggs were shipped to our lab in Houston and used to orally challenge BALB/c mice as previously described [17] . The A . suum larvae hatch in the mouse intestine , and then migrate to the liver and lungs . The number of larvae recovered from mouse lung tissue eight days post infection were used as a biomarker to evaluate vaccine efficacy [17] . Crude extracts of A . suum eggs and lung-stage larvae were prepared by homogenization and sonication , and the insoluble pellet was removed by centrifugation as previously described [33] . Amino acid sequences were aligned using CLUSTAL W and prepared for display using BOXSHADE . The phylogenetic trees were generated for As16 and its homologues from different nematodes using Phylogeny . fr [34] ( http://www . phylogeny . fr/index . cgi ) . DNA coding for As16 without its signal peptide was codon optimized for expression in yeast and synthesized by GenScript . The DNA coding for As14 without its signal peptide was PCR amplified with As14 specific primers from A . suum larvae cDNA reverse-transcribed from total larval RNA . As16 and As14 coding DNAs were subcloned into the yeast expression vector pPICZαA ( ThermoFisher Scientific , Carlsbad ) . The correct sequences and reading frames of the recombinant plasmids were confirmed by double-stranded DNA sequencing using vector flanking primers , α-factor and 3’-AOX1 . The recombinant As16 and As14 ( rAs16 and rAs14 ) with a hexahistidine tag at its C-terminus were expressed in yeast stain P . pastoris X-33 under induction with 0 . 5% methanol for 48–72 hours and then purified by immobilized metal ion affinity chromatography ( IMAC ) , as described previously [35] . The purity of the recombinant proteins was determined by SDS–PAGE . The protein concentration was measured using BCA ( ThermoFisher Scientific , Waltham ) and Endotoxin clearance was confirmed using the Charles River Endosafe-PTS system ( Charles River , Houston ) . Six-week old female BALB/c mice were purchased from Taconic and divided into four groups of 20 animals each . Two vaccine groups were immunized subcutaneously with 50 μg of rAs16 or rAs14 emulsified with the adjuvant Montanide ISA720 ( Seppic , Paris , France ) in a total volume of 100 μl ( antigen/ISA720 = 30/70 v/v ) . Mice were boosted twice with the same dose on days 21 and 35 . The control groups were injected with PBS or PBS+ISA720 using the same regimen . Two weeks after the final vaccination , 5 mice from each group were sacrificed and blood and splenocytes were harvested for immunological tests . The remaining 15 mice from each group were challenged with 2 , 500 A . suum embryonated eggs in a total volume of 100 μl , administered by oral gavage . Eight days after infection , all infected mice were sacrificed , lungs were harvested , and A . suum lung-stage larvae were collected using a Baermann apparatus , as previously described [17] . Reduction in larval burden was calculated in all groups and the results were compared between the vaccine groups and the PBS and adjuvant control groups . To improve the protection induced by rAs16 and interpret the immunological mechanism underlying the As16-induced protective immunity , another vaccine trial was performed by formulating 25 μg of rAs16 with either 200 μg of Alhydrogel ( Brenntag , Mülheim , Germany ) , 20 μg of MPLA ( InvivoGen , San Diego ) , or 50 μl of AddaVax ( 50/50 , v/v ) ( InvivoGen , San Diego ) , each administered subcutaneously in a total volume of 100 μl per mouse given using the immunization regimen described above . Control groups were given adjuvant only . As a positive control , one group of 20 mice was orally infected three times with 1 , 000 A . suum embryonated eggs . After these immunizations , all mice were challenged with 2 , 500 A . suum embryonated eggs . Sera from all blood samples were isolated and frozen at -20°C . The sera samples were assayed for antigen-specific IgG isotypes ( IgG1 , IgG2a ) by a modified indirect enzyme-linked immunosorbent assay ( ELISA ) . Briefly , individual wells of Nunc-Immuno Maxisorp plates ( Thermo Scientific , Waltham ) were each coated with 100 μl of rAs16 ( 3 . 1 μg/ml ) or rAs14 ( 0 . 39 μg/ml ) in coating buffer ( KPL , Milford ) overnight at 4°C based on the pretested optimal signal/noise ratio . The coated plates were blocked overnight with 0 . 1% BSA in PBST ( PBS +0 . 05% Tween-20 ) , then incubated with diluted serum samples , starting at 1:200 in 0 . 1% BSA in PBST for 2 hours . Horseradish peroxidase ( HRP ) -conjugated goat anti-mouse IgG1 and IgG2a ( Lifespan Biosciences , Seattle ) were used as secondary antibodies ( 1:4 , 000 in PBST ) . Sure Blue TMB ( KPL , Milford ) was added as the substrate . The reaction was stopped by adding 100 μL of 1 M HCl . The absorbance was measured at 450 nm using a spectrophotometer ( BioTek , Winooski ) . Samples including crude extracts of A . suum lung-stage larvae and eggs and the recombinant proteins were separated by SDS–PAGE , then transferred onto PVDF membrane ( ThermoFisher , Waltham ) . After blocking with 5% ( w/v ) skim milk powder in PBST , the membrane was incubated with sera from mouse immunized with recombinant proteins or A . suum eggs . HRP-conjugated goat anti-mouse IgG ( Invitrogen , Carlsbad ) was used as a secondary antibody . The antibody recognized bands were developed by ECL ( GE Healthcare , Chicago ) . Recombinant Tc24 protein , expressed in yeast [36] , was used as a negative control . Spleens were obtained from mice two weeks after the third immunization and the splenocytes were disassociated using a 100 μm cell strainer . The cells were then suspended in complete RMPI medium containing 10% heat inactivated FBS and 1x pen/strep solution . After being centrifuged at 300 x g for 5 min , the cells were resuspended in 2 mL ACK lysis buffer ( Thermo Scientific , Waltham ) for 5 min . After centrifugation the splenocytes were resuspended in complete RPMI media containing 10% DMSO and stored in liquid nitrogen until use . For the cytokine stimulation assay , splenocytes were thawed in a 37°C water bath and transferred to 5 mL pre-warmed complete RPMI . Cells were washed once to remove residual DMSO . Splenocytes were seeded in a 96-well U-bottom culture plate ( Falcon , Corning ) at 1x106 cells per well in 250 μl medium and re-stimulated with either 25 μg/mL rAs14 or rAs16 at 37°C , 5% CO2 for 48 hours . Positive controls were stimulated with 20 ng/mL PMA and 1 μg/mL Ionomycin , and unstimulated negative control cultures were performed concurrently . After 48 hours the cells were pelleted by centrifugation at 300 x g for 5 min and the supernatants were collected for measuring cytokine production . The supernatant samples were tested for levels of IL-2 , IL-4 , IL-5 , IL-10 , IL-12 ( p70 ) , GM-CSF , IFN-γ and TNF-α using a Bio-Plex Pro Mouse Th1/Th2 8-plex kit ( Bio-Rad , Hercules ) . To save material and costs , and to increase the sensitivity of the experiment , the kit was used in combination with DA-Bead plates ( Curiox Biosystems , Singapore ) as previously described [37] . Samples were run on a Bio-Plex Magpix multiplex reader according to manufacturer's recommendations ( Luminex , Austin ) . Raw Luminex data were analyzed using the Bio-Plex Manager 6 . 0 software and plotted in GraphPad Prism 6 . 0 . The cutoffs of the cytokine standards were dependent on the lot number of the Bio-Rad kit . To remove individual baseline cytokine values , cytokine values from non-restimulated samples were subtracted from those associated with antigen restimulated samples . Statistical significance of differences between groups was determined using a Mann-Whitney test using Prism 6 . In Fig 5c the groups receiving the antigen + adjuvant were compared to the associated adjuvant alone group , the A . suum egg group , and to the PBS group using a Fisher’s LSD test . Data was presented as means ± standard deviation . For the statistical analysis , p < 0 . 05 was considered to be statistically significant .
The genes encoding As16 ( yeast codon optimized ) and As14 ( native sequence ) without their signal peptides were cloned into the Pichia expression vector pPICZαA . The hexahistidine-tagged rAs16 and rAs14 proteins were expressed in P . pastoris X-33 through induction with 0 . 5% methanol over 72 hours and then purified by immobilized metal ion affinity chromatography ( IMAC ) . The purified rAs14 and rAs16 proteins were analyzed by SDS–PAGE ( Fig 1A ) . The apparent molecular weight for both proteins was approximately 15 . 0 kDa , which corresponds well to the sizes of the predicted gene products ( 14 . 8 kDa for rAs14 and 15 . 4 kDa for rAs16 ) . To determine whether rAs14 or rAs16 were recognized by protective immune sera from mice repeatedly infected with A . suum eggs described below ( Protective immunity induced by immunization with rAs16 ) , the infected mouse sera were used for Western blot . We observed that only rAs16 was recognized by the infected mouse sera , while rAs14 was not recognized by the same sera ( Fig 1B ) . Even though the two proteins share 47% identity and 66% similarity in sequence ( Fig 1E ) , there was no obvious immunological cross reaction between them when using rAs16 or rAs14 immunized mouse sera individually ( Fig 1C and 1D ) . The results suggest that As16 antigen was exposed to the immune system during A . suum egg infection and larval migration , whereas As14 antigen may not have been immunologically accessible . As16 is a 16 kDa nematode specific protein found among several different nematode species ( Fig 2 ) . It is present in different developmental stages of A . suum , including larvae and adult worms , but its function remains unknown [20] . As16 is highly conserved in Ascaris spp . ; it shares 94% sequence identity with its counterpart in the human parasite A . lumbricoides ( Al-Ag2 ) , suggesting the possibility of achieving cross-protection for both Ascaris species if As16 were to be used as a vaccine antigen . Its homologues in filarial worms are ranked among the leading vaccine candidates against human onchocerciasis ( Ov-RAL-2 ) [38] and Brugia malayi filarial infections ( Bm-RAL-2 ) [39] . The As16 homologue from the canine hookworm ( Ancylostoma caninum , Ac-16 ) also protected dogs from blood loss and reduced worm fecundity [40] , and the Baylisascaris schroederi homologue , Bs-Ag2 , protected giant pandas from infection with that parasite [41] . Immunization with rAs16 and rAs14 formulated with the ISA720 adjuvant elicited significant titers of antigen-specific IgG1 and IgG2a antibodies in mice , with IgG1 as the predominant subclass ( Fig 3A ) . The IgG1/IgG2a ratio after As16 immunization ( 2662:1 ) was more than 100-fold higher than the ratio for As14 ( 206:1 ) ( Fig 3A ) , suggesting that a predominant Th2-type immune response occurs to both antigens , but in particular to rAs16 . Mice given PBS+ISA720 did not show any IgG isotype responses specific to rAs16 or rAs14 . To evaluate the cytokine profiles induced by immunization with rAs16 and rAs14 , mice were sacrificed two weeks after the final immunization and their splenocytes were isolated . Cytokine profiles were determined by measuring levels of IL-2 , IL-4 , IL-5 , IL-10 and IFN-γ in supernatants of splenocytes re-stimulated with 25 μg/ml rAs16 or rAs14 for 48 hours . Signal background in blank media was subtracted from re-stimulated samples . Statistically significantly increased levels of the cytokines IL-2 , IL-4 , IL-5 , IL-10 were detected in the supernatants of re-stimulated splenocytes from mice immunized with rAs16 and rAs14; however , IFN-γ was significantly increased only in mice immunized with rAs14 , not in the mice immunized with rAs16 ( Fig 3B ) . Splenocytes from the PBS +ISA720 control group did not show any detectable cytokine expression . All mice were orally challenged with 2 , 500 A . suum embryonated eggs two weeks after the final immunization . A . suum larvae were collected from the lungs of immunized mice eight days after egg challenge . The lung larva count showed that mice immunized with 50 μg of rAs16 formulated with ISA720 adjuvant showed a 36 . 7% larva reduction compared to the adjuvant-only control groups , constituting a statistically significant difference ( p<0 . 001 ) ( Fig 4A ) . In addition , the size of the larvae collected from rAs16 immunized mice was much smaller than those collected from adjuvant control mice , suggesting developmental stunting due to the vaccination ( Fig 4B ) . However , mice immunized with rAs14 did not show any protection in terms of reducing the number of larvae found in the lungs , or affecting the size of the larvae . To further understand the immunological mechanism underlying the protective immunity induced by rAs16 , and to select an adjuvant that performs best in protecting mice from infection , mice were immunized with only half the amount of As16 ( 25 μg , compared to 50 μg of rAs16 used for formulation with ISA720 ) , formulated with three different adjuvants ( Alhydrogel , MPLA and AddaVax ) . As a positive control , another group of mice was immunized through three trickle infections with 1 , 000 A . suum eggs . After three immunizations , all mice were challenged with 2 , 500 A . suum embryonated eggs . Mice immunized with 25 μg of rAs16 formulated with Alhydrogel , an alum adjuvant inducing a predominant Th2-type response , experienced a 38 . 9% lung larval reduction , which is statistically significant compared to the PBS and adjuvant-only control groups . However , rAs16 formulated with MPLA , a TLR4 agonist inducing a Th1/Th2-mixed type immune response , induced only a 26 . 1% lung larval reduction that was not statistically different from the control groups . Mice immunized with rAs16 formulated with AddaVax , an oil-in-water based adjuvant similar to the MF59 adjuvant that is licensed for flu vaccines in Europe and known to also induce a Th1/Th2 mixed immune response , were not statistically significantly protected against the A . suum egg challenge . Strikingly , mice immunized with three trickle infections with 1 , 000 A . suum eggs produced almost sterile immunity ( 99 . 6% lung larval reduction ) against A . suum challenge ( Fig 5A ) . Serological antibody measurement revealed that mice immunized with rAs16 formulated with three different adjuvants all produced high titers of IgG1 and IgG2a , with a bias towards IgG1 ( IgG1/Ig2a = 120–202 ) ( Fig 5B ) . Interestingly , mice repeatedly infected with a low-dose of A . suum eggs produced a 99 . 6% lung larva reduction and also showed a significant increase of anti-As16 specific IgG1 antibodies , but without any accompanying IgG2a response . This suggests native As16 is released and exposed to the host immune system during repeated low-dose A . suum infections and may be involved in the induction of the observed Th2 protective immunity . Cytokines released by splenocytes upon re-stimulation of rAs16 showed that immunization with rAs16 formulated with Alhydrogel induced the release of IL-5 , a major cytokine linked to Th2 responses in mice , and some level of IL-12 and GM-CSF . There were no detectable levels of IL-2 , IL-4 , IL-10 , IFN-γ or TNF-α observed in mice immunized with rAs16 + Alhydrogel . Conversely , mice immunized with rAs16 formulated with MPLA and AddaVax showed a significant release of Th1 associated cytokines ( IL-12 , IFN-γ and TNF-α ) , Th2 type cytokines ( IL-4 , IL-5 ) , and IL-10 , IL-2 and GM-CSF as well ( Fig 5C ) , even though there was no significant protective immunity observed in these Th1/Th2 adjuvant groups . Western blot analysis with sera from mice immunized with rAs16 and rAs14 demonstrated that anti-As16 mouse sera recognized a band at ~14 kDa in the soluble extracts of lung larvae of A . suum , but not in the extracts of A . suum eggs . For As14 , mouse anti-As14 sera specifically recognized a band of about 13 kDa in lung larval extracts , but not in A . suum eggs , indicating native As16 and As14 are expressed only at the larval stage , not in the eggs of A . suum ( Fig 6 ) . As16 and As14 are likely expressed in A . suum larvae after hatching and during the migration to the lungs . However , since infected mice only produced antibodies to As16 but not to As14 ( Fig 1 ) , it appears that only As16 is exposed to the immune system during larval migration .
As16 ( A . suum , BAC66614 . 1 ) ; As14 ( A . suum , BAB67769 . 1 ) ; Ce16 ( Caenorhabditis elegans , NP_495640 . 1 ) ; Dv16 ( Dictyocaulus viviparus , KJH51207 . 1 ) ; Acey16 ( Ancylostoma ceylanicum , EPB72254 . 1 ) ; Ac16 ( A . caninum , ABD98404 . 1 ) ; Ad16 ( A . duodenale , KIH68079 . 1 ) ; Sv-SXP ( Strongylus vulgaris , AGF90534 . 1 ) ; Na-SAA2 ( Necator americanus , XP_013290850 . 1 ) ; Al-Ag1 ( A . lumbricoides , ACJ03764 . 1 ) ; Bs-Ag1 ( Baylisascaris schroederi , ACJ03761 . 1 ) ; Av-RAL2 ( Acanthocheilonema viteae , AAB53809 . 1 ) ; Ll-SXP ( Loa loa , XP_003142836 . 1 ) ; Ov-RAL2 ( Onchocerca volvulus , P36991 . 1 ) ; WB14 ( Wuchereria bancrofti , AAC17637 . 1 ) ; Bm-RAL-2 ( Brugia malayi , XP_001900036 . 1 ) ; Bs-Ag2 ( B . schroederi , ACJ03762 . 1 ) ; Al-Ag2 ( A . lumbricoides , ADB45852 . 1 ) ; Hc16 ( Haemonchus contortus , CDJ91573 . 1 ) ; Asim16 ( Anisakis simplex , BAF43534 ) , Tc16 ( Toxocara canis , KHN84076 . 1 ) and As-RAL-2 ( Anisakis simplex , BAF75709 . 1 ) .
|
Roundworms ( Ascaris ) infect more than 700 million people living in poverty worldwide and cause malnutrition and physical and mental developmental delays in children . As an alternative or complementary approach to global deworming , a pan-anthelminthic vaccine is under development that targets ascariasis in addition to other human intestinal nematode infections . Towards this goal , two Ascaris suum antigens , As16 and As14 , were expressed in Pichia pastoris as recombinant proteins . Mice immunized with rAs16 formulated with ISA720 adjuvant produced significant larva reduction ( 36 . 7% ) and stunted larval development against A . suum egg challenge . The protection was associated with predominant Th2-type responses characterized by high levels of serological IgG1 ( IgG1/IgG2a > 2 , 000 ) and Th2 cytokines , IL-4 and IL-5 . A similar level of protection was observed in mice immunized with rAs16 formulated with alum that induces mainly a Th2-type immune response , whereas mice immunized with rAs16 formulated with MPLA or AddaVax , both inducing major Th1-type responses , were not significantly protected against A . suum infection . High-yield expression of rAs16 in yeast will allow for large-scale manufacture , and its protective efficacy when formulated with alum suggests its suitability as a vaccine candidate .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"cytokines",
"immunology",
"parasitic",
"diseases",
"vaccines",
"preventive",
"medicine",
"developmental",
"biology",
"immunologic",
"adjuvants",
"molecular",
"development",
"infectious",
"disease",
"control",
"vaccination",
"and",
"immunization",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"proteins",
"life",
"cycles",
"recombinant",
"proteins",
"immune",
"response",
"immune",
"system",
"biochemistry",
"helminth",
"infections",
"immunity",
"physiology",
"biology",
"and",
"life",
"sciences",
"larvae"
] |
2017
|
Yeast-expressed recombinant As16 protects mice against Ascaris suum infection through induction of a Th2-skewed immune response
|
To improve schistosomiasis control programs in Uganda , where intestinal schistosomiasis is a widespread public health problem , a country-wide assessment of the disease prevalence among all age ranges is needed . Few studies have aimed to quantify the relationships between disease prevalence and water and sanitation characteristics across Uganda to understand the potential to interrupt disease transmission with an integrated package of interventions . A nationally representative survey was undertaken that included a household and individual questionnaire followed by disease testing based on detection of worm antigens ( circulating cathodic antigen–CCA ) , diagnosis and treatment . A comprehensive set of questions was asked of randomly sampled individuals , two years of age and above , to understand their water and sanitation infrastructure , open defecation behaviors , exposure to surface water bodies , and knowledge of schistosomiasis . From a set of 170 randomly sampled , geographically diverse enumeration areas , a total of 9 , 183 study participants were included . After adjustment with sample weights , the national prevalence of schistosomiasis was 25 . 6% ( 95% confidence interval ( CI ) : 22 . 3 , 29 . 0 ) with children ages two to four most at risk for the disease with 36 . 1% infected ( 95% CI: 30 . 1 , 42 . 2 ) . The defecation behaviors of an individual were more strongly associated with infection status than the household water and sanitation infrastructure , indicating the importance of incorporating behavior change into community-led total sanitation coverage . Our results highlight the importance of incorporating monitoring and evaluation data into control programs in Uganda to understand the geographic distribution of schistosomiasis prevalence outside of communities where endemicity is known to be high . The high prevalence of schistosomiasis among the youngest age group , ineligible to receive drug treatment , shows the imperative to develop a child-appropriate drug protocol that can be safely administered to preschool-aged children . Water and sanitation interventions should be considered an essential investment for elimination alongside drug treatment .
Schistosomiasis is a neglected water-based , vector-borne disease transmitted indirectly through freshwater snails , estimated to affect more than 240 million individuals worldwide with 700–800 million people living at risk of infection [1–3] . In sub-Saharan Africa , approximately 280 , 000 deaths per year have been attributed to schistosome infections and the onset of complications caused by the disease [4] . The transmission cycle of the disease is perpetuated when infected individuals defecate ( predominately Schistosoma mansoni ) or urinate ( predominately Schistosoma haematobium ) in open waters and the miracidia , released from hatched schistosome eggs , infect suitable snail hosts [5] . The cercariae are then subsequently expelled from the snails and infect individuals in contact with cercariae-contaminated waters via penetration through the skin . The lack of sanitation infrastructure to adequately collect and treat human waste , along with the common behaviors and activities that involve frequent contact with open waters , has made schistosomiasis a challenging disease to control in sub-Saharan Africa . Approaches to estimate the scale of the disease have primarily focused on measuring disease prevalence in areas where endemic status is known to be high , particularly in shoreline communities where water contact is frequent [6–13] . In Uganda , where the disease is considered endemic , the control program has focused mainly on chemotherapeutic intervention using praziquantel with less emphasis on interrupting environmental transmission [14] . In 2018 , coverage of either annual or biennial mass drug treatment , coupled with health education , was achieved across all 82 districts where schistosomiasis is common . Despite this progress , there is doubt that the targets set by the World Health Organization will be met to control morbidity by 2020 and achieve elimination by 2025 [15] . To accurately assess progress toward schistosomiasis elimination in Uganda a comprehensive assessment of disease prevalence and intensity across the population is necessary . To understand the prevalence of schistosomiasis in Uganda and the associated risk factors , we undertook the first nationally representative survey , which included testing with detection of worm antigens ( circulating cathodic antigen–CCA ) in urine samples , diagnosis and treatment of schistosomiasis-positive participants within sampled households . The aims of this study were to assess the prevalence of schistosomiasis across Uganda and characterize the relationship between infection with S . mansoni ( the predominant circulating strain in Uganda [16–18] ) and its associated risk factors to help develop sustainable strategies to control schistosomiasis in Uganda .
This manuscript has been developed according to the consolidated standards of observational studies ( see S1 STROBE checklist ) . The study received ethical clearance from the Institutional Review Boards at Makerere University School of Public Health ( Kampala , Uganda; reference no . 424 ) , Johns Hopkins University ( Baltimore , Maryland reference no . IRB00007024 ) , the Uganda National Council for Science and Technology ( UNCST; Kampala , Uganda; reference no . HS 2069 ) and with permission from the Uganda Office of the President and district authorities . The following enrollment procedures were approved by the ethical committees: all participants or caregivers were informed about the purpose and procedures of the study and they were invited to sign a written consent form . In case of illiteracy , participants were asked to record their thumbprint alongside a literate witness . Those with informed consent were assigned a unique identifier . Diagnostic results were communicated to participants and those found positive for infection with schistosomiasis , and who met criteria for treatment ( e . g . , a single 40 mg/kg oral dose of praziquantel using height measured against a tablet pole as a proxy ) , were treated according to national guidelines [19] . We conducted rapid data collection using mobile technology under the Performance Monitoring & Accountability 2020 ( PMA2020 ) platform that utilizes a multi-stage cluster design with fixed , census-derived enumeration areas ( EAs ) and random selection of households [20] . To determine an accurate national prevalence rate for schistosomiasis in Uganda a three-tiered , stratified , clustered , random sampling was employed . The first tier of sampling randomly selected EAs according to their classification of either nearby or far from water bodies . The second tier of random sampling selected households within the EAs by first mapping the borders and occupied households within the selected EA . Once all households within the EA boundary were mapped and listed , a random sample of 30 households was made using a random number generator application . The third tier of sampling randomly selected individuals within each sampled household and tested them for schistosomiasis . All household members were listed during the household interview and were either usual members of the household or slept in the household the night before . A random number generator was used to randomly select an individual from the roster as a study participant . Therefore , no adjustment is made for the probability of selecting an individual from within the household . At all tiers , the sampling probabilities were available because the sample frame for all households and eligible individuals within each household was known . Monitoring of schistosomiasis in Uganda has historically focused on school-based sampling within districts , which results in a large variance and high design effect due to the strong intra-cluster correlation [21] . The sampling strategy for this study incorporated community distances from water bodies , such as wetlands and surface water sources to categorize an EA as near ( “endemic” ) or far ( “non-endemic” ) . In 2004 , Kabatereine et al . found S . mansoni prevalence rates , using egg counts in stool with the Kato-Katz method , were higher in communities that were 5km or less from the shores of Lake Victoria or Lake Albert and lower in communities more than 5 km [22] . In collaboration with the Uganda Bureau of Statistics ( UBOS ) , a nationally representative sample of EAs was randomly selected , which are geographic units of a size determined by UBOS that average 100 households . EAs were sampled based on their proximity to water bodies based on if their boundaries fall within ( near ) or outside of ( far ) 5 km of wetlands and water bodies . To determine an EA’s classification , UBOS established the distance to a water body based on detailed nationwide GIS data that included the boundaries of all seasonal wetlands and surface water bodies as small as 10m in one dimension . Rice paddies and similar wetland crops were not included because water intensive crops are expected to be in proximity or co-located with natural wetlands and surface water bodies . To determine the sample size a conservative assumption was used that the prevalence rate among individuals living in households in the near strata is 50% , leading to the largest possible sample size for that strata . For EA clusters over 5 km from water bodies in the far strata , 15% prevalence was assumed . We used 80% power and 5% type-I error and applied an estimated design effect of four . We adjusted the final sample size to account for non-response to reach a target of 170 EAs . The final proportion of EAs was chosen according to UBOS’s cartography unit estimate of approximately 70% of the Ugandan population near ( within 5 km of ) a water body for a final number of 120 EAs in the near strata and 50 in the far strata . With 30 households in each EA selected over two consecutive years , we calculated a sample size of 7 , 200 in the near strata and 3 , 000 in the far strata for a sample size target of 10 , 200 . To collect the sample a pooled cross-sectional design over the two years of 2016 and 2017 was necessary to measure the effect of praziquantel drug treatment on the subset of study participants with a positive diagnosis for schistosomiasis in the first survey round . The sample was pooled to enable more precise estimates in sub-analyses where sample sizes would be limited if only one survey round was used . Data were collected from 170 EAs between October to December in 2016 and from the same set of EAs in the same season in 2017 . Two different random samplings of households were selected for each year to allow for a pooled cross-sectional analysis where data are combined from 2016 and 2017 . The surveys were scheduled to precede the mass drug administration ( MDA ) planned by Ministry of Health Vector Control Division ( MOH/VCD ) in order to ensure that measurement of the national prevalence rate was not affected by the implementation of community-wide drug treatments . Community leaders in each EA were asked if MDA had ever been previously implemented in their community . For each household randomly selected for the survey a household questionnaire was administered by a trained enumerator fluent in the local language . The head of the household or a competent household member was interviewed for questions relevant to the household . The household questionnaire collected information on socio-demographic factors related to wealth , education , construction quality of the dwelling , and water , sanitation and hygiene ( WASH ) characteristics . The PMA2020 platform is designed for collection of WASH indicators [23 , 24] and was adapted for collection of these variables in the context of schistosomiasis . One household member , two years of age and older , was randomly selected to participate in the study . The individual questionnaire included a series of questions on WASH practices and open defecation behaviors , knowledge of schistosomiasis and uses of nearby water bodies that may expose the individual to schistosomiasis . At the end of the questionnaire the individual was tested for schistosomiasis . If the randomly selected individual in the household was a child under 8 years old then the primary caregiver of the child was interviewed . For children ages 8–17 the caregiver was present to assist the child with questions regarding the child’s knowledge , attitudes and practices with regards to schistosomiasis and the caregiver was interviewed regarding the child’s schistosomiasis infection history and demographics . This ensured that the most appropriate person was interviewed to obtain the most accurate information possible regarding schistosomiasis and the child . After the individual questionnaire was completed , respondents were asked to provide a urine sample in a sterile cup . If the individual was not able to provide a sample at the conclusion of the questionnaire , the enumerator came back at a time when the sample could be provided . Enumerators were trained by personnel from MOH/VCD to administer a rapid , commercially available antigen test , the Point-of-Care Circulating Cathodic Antigen test ( POC-CCA , Rapid Medical Diagnostics , Pretoria , South Africa ) , which measures the CCA of juvenile and adult S . mansoni that are released into the circulation . The test has been recommended by the WHO for screening of intestinal schistosomiasis and its protocol , with interpretation by color change , was well suited for training the enumerators [25] . The quantity of this antigen in urine correlates with the worm burden to indicate an active infection [26] . The POC-CCA test has been shown to have higher sensitivity than the standard Kato-Katz method when it was evaluated in endemic settings in the Africa region [27–31] . The POC-CCA test was conducted according to the manufacturer’s guidelines and was repeated a second time if the first test was considered invalid . Test results were scored as negative if the CCA band was absent and positive if present . Cases with trace results for CCA were considered as positive to most accurately estimate infection prevalence [32] . The tests were interpreted independently by enumerators and recorded by photo on the mobile devices for validation . Participants with a positive S . mansoni diagnosis by POC-CCA were treated for schistosomiasis with oral doses of praziquantel if they met the criteria for treatment as advised by the Uganda Ministry of Health . Children under five years of age found to be positive were given albendazole ( 400 mg ) but were not treated with praziquantel and breastfeeding or pregnant women were also not treated , per MoH/VCD guidance . Water body locations were recorded by field workers at each EA to understand the relationship between prevalence rates using CCA and proximity to water bodies , which may harbor the intermediate snail hosts . To measure individual distances between households and water bodies located within one kilometer of an EA boundary , field workers utilized the global positioning system ( GPS ) features on their mobile phones and recorded a GPS point for a water body based on the most likely place of entry into the surface water . GPS points were also recorded for the household during the household questionnaire and the difference between these two locations was calculated . If more than one water body was recorded for an EA then the median distance calculated for that household was analyzed as the distance between the household and a water body . All administered questionnaires had their responses directly entered into an Android smartphone using Open Data Kit ( ODK ) software similar to protocols from the PMA2020 program [20] . Following the interview , testing , diagnosis and treatment , household and individual questionnaires were submitted to a secure cloud server , where they were aggregated and monitored on a daily basis by data managers and quality assurance teams at Makerere University School of Public Health in Kampala . Technical assistance was provided by the PMA2020 team at Johns Hopkins University . To ensure data quality , the following procedures were put in place: i ) the POC-CCA tests were randomly replicated in spot checks by supervisors to validate the diagnosis of infection status; ii ) enumerators recorded their start and stop times that could later be cross-checked by the ODK data to verify that the diagnosis from the POC-CCA cassette was within the specified time window; iii ) photos of the POC-CCA test were recorded and verified with the diagnosis entered into ODK by quality assurance teams; and iv ) a set of questions were selected for monitoring of data quality during data collection by the quality control team . The national S . mansoni prevalence estimate was determined as the number of individuals testing CCA positive for schistosomiasis divided by the total number of participants with test results available . The national estimate was weighted using the adjusted sample weights . Pearson’s χ2 was used to test for differences of prevalence estimates across categories . Population wealth was measured using quintiles that were calculated using a pre-determined set of asset variables collected from the household questionnaire . An open defecation/urination ( OD/U ) index was constructed from the individual questionnaire where “low” represented no self-reported OD/U , “medium” indicated OD/U in the bush only and “high” indicated OD/U in open water bodies . The independent effect of factors associated with the CCA prevalence of S . mansoni infection was determined using robust “modified” Poisson regression and the results are reported as unadjusted and adjusted prevalence ratios ( PRs ) and 95% confidence intervals ( CI ) , along with the test for significance . Given the common occurrence of schistosomiasis prevalence ( greater than 10% ) prevalence ratios are reported to be “conservative , consistent , and interpretable” [33–35] . To account for clustering at the EA level , unadjusted and adjusted multilevel mixed-effects generalized linear models were applied using a "modified Poisson" approach to estimate the prevalence ratios for schistosomiasis risk-factors [36] . The adjusted model was built by first including risk-factor variables that are well established including sex , age group and the interaction between sex and age group [37–39] . Risk factors were then entered together with sex , age group and the interaction between the two and if significant ( using likelihood ratio test ) were added to the adjusted model [18] . The final adjusted model fit was assessed using both the deviance statistic and the Pearson statistic generated from the goodness-of-fit test [40] . P-values less than 0 . 05 were considered significant . Statistical analyses were done using STATA version 14 . 0 ( Stata Corporation; College Station , United States of America ) . Maps were created in QGIS Geographic Information System 3 . 0 . 2 Girona Open Source Geospatial Foundation Project ( http://qgis . osgeo . org ) . EA locations were geolocated by creating a centroid using the GPS coordinate data collected at each household listed in the EA and for confidentiality , the centroid GPS location was randomly displaced .
Enumerators reached all of the 170 randomly sampled EAs pre-selected for inclusion in the two-year serial cross-sectional survey . Household questionnaires were completed for 9 , 516 households ( completion rate 93 . 3% ) with the majority of non-responses due to not being at home ( 2 . 6% , n = 265 ) and refusal ( 1 . 4% , n = 146 ) . Individual questionnaires were completed for 9 , 183 randomly selected household members ( completion rate 96 . 5% ) with the majority of non-responses due to not being at home ( 2 . 2% , n = 208 ) and refusal ( 0 . 7% , n = 63 ) . The diagnostic urine test was successfully completed by 99% of the individuals who accepted to be interviewed for a total of 9 , 097 individuals included in the sample . The national prevalence of schistosomiasis using CCA was 25 . 6% ( 95% confidence interval ( CI ) : 22 . 3 , 29 . 0 ) after probability sample weighting . Disaggregated by year the 2016 national prevalence was 22 . 1% ( 95% CI: 18 . 7 , 25 . 6 ) and 2017 national prevalence was 29 . 0% ( 95% CI: 24 . 4 , 33 . 6 ) . Fig 1 presents the percent of participants who tested positive for schistosomiasis in each of the 170 randomly sampled EAs in Uganda . Comparisons of disease prevalence using CCA by socio-demographic categories show the greatest differences for age , religion , household size , sex , region and whether the EA received mass drug administration in years prior to the survey ( yes/no ) . Table 1 presents the percent of probability sample weighted estimates for those infected with schistosomiasis by the socio-demographic categories . The age group with the highest prevalence for schistosomiasis was the 2–4 year old children with 36 . 1% infected ( 95% CI: 30 . 1 , 42 . 2 ) . For subsequent age groups , 27 . 7% of children age 5–10 years old were infected ( 95% CI: 23 . 4 , 32 . 0 ) and 30 . 9% of 11–15 year olds were infected ( 95% CI: 25 . 8 , 36 . 1 ) . The prevalence of schistosomiasis steadily declined with each consecutive older age group . There were no differences in infection status by education levels and marginal differences by wealth quintiles . Age prevalence CCA curves in Fig 2 show the different burdens of disease by ages for high prevalence EAs , considered to be 10% or greater , and low prevalence EAs , considered to be less than 10% . The age groups were divided into 2–4 years old to represent the preschool aged children ineligible to receive praziquantel drug treatment . Subsequent ages were grouped in five year increments to analyze the impact of school-based drug administration: the 5–10 and 11–15 year olds most likely to receive drug treatment through the school-based drug administrations , and the 16–20 year olds who most likely no longer have access to school-based drug treatment programs . The older age groups , less likely to receive yearly access to mass drug administration , were grouped in ten year increments . The 2–4 years old group for both curves were the most at-risk group for schistosomiasis with 39 . 8% infected in the high prevalence EAs and 8 . 9% infected in the low prevalence EAs . Household water and sanitation characteristics were recorded from all 9 , 097 of the study participants . The prevalence of schistosomiasis using CCA did not significantly vary by household water and sanitation categories ( S1 Table ) including main drinking water source classified by the Joint Monitoring Programme [41] as “improved” or “unimproved” , use of surface water for any daily purpose ( e . g . drinking , washing , cooking , bathing ) and classification as near or far from a water body . The prevalence of schistosomiasis using CCA did vary based on individual water and sanitation practices ( S2 Table ) . Schistosomiasis was significantly higher among those who self-reported to defecate in surface water ( 31 . 2% ( 95% CI: 25 . 2 , 37 . 2 ) versus those that did not self-report the practice ( 24 . 1% infected ( 95% CI: 20 . 8 , 27 . 5 ) ( p = 0 . 0063 ) ) . For individuals self-reporting to submerge in surface water , defined as putting hands , feet , or any other part of the body into a water body , within the last year , 26 . 6% were infected ( 95% CI: 22 . 7 , 30 . 6 ) versus those that reported not to submerge themselves with 21 . 5% infected ( 95% CI: 17 . 7 , 25 . 4 ) ( p = 0 . 0133 ) . The prevalence of schistosomiasis significantly increased with each level of the OD/U index: low likelihood of OD/U had 18 . 4% infected ( 95% CI: 13 . 9 , 23 . 0 ) , medium likelihood 24 . 8% infected ( 95% CI: 21 . 6 , 28 . 0 ) and high likelihood 29 . 3% infected ( 95% CI: 24 . 3 , 34 . 4 ) ) ( p = 0 . 0005 for Chi-square for trend test ) . Fig 3 shows the community open defecation rates ( percent of individuals in a given enumeration area stating that they openly defecate in the bush or in water bodies ) versus the prevalence of schistosomiasis using CCA in that same enumeration area and shows an increasing trend in disease prevalence as the percent of open defecation in the community increases ( n = 170 ) . The percentage of the sampled population that had heard of schistosomiasis prior to the survey was 61 . 8% ( n = 5570 ) ( S3 Table ) . Those that had previously heard of schistosomiasis had a significantly higher prevalence of schistosomiasis infection using CCA ( 28 . 4% ( 95% CI: 24 . 3% , 32 . 5% ) versus those who had not heard of the disease ( 21 . 1% ( 95% CI: 17 . 9% , 24 . 2%; p = 0 . 0002 ) . The activity that respondents most frequently identified as being a risk factor for schistosomiasis infection was wading into water to wash clothes , fetch water or bathe oneself ( n = 2075 respondents ) ; there was no difference in disease prevalence between those who knew this was an activity that risked infection and those who did not know . Table 2 shows the prevalence ratios for risk factors of schistosomiasis infection ( unadjusted and adjusted ) among study participants in Uganda . The variables included in the final adjusted model , in addition to sex , age group , and the interaction between sex and age group , were wealth status , region , OD/U index , distance from the household to a water body , and whether the EA had received mass drug administration prior . The prevalence ratios of infection were significantly lower at all ages when compared to children 2–4 years . For children in the 5–10 year old age group , there was a reduction of approximately 48% in the prevalence of schistosomiasis ( 0 . 52 aPR , 95% CI: ( 0 . 36 , 0 . 77 ) ) compared to the 2–4 year olds . The household water and sanitation characteristics were not significant risk factors for infection with schistosomiasis while the individual water and sanitation related behaviors were found to be risk factors . For individuals with the highest likelihood to openly defecate and or urinate in a water body ( High OD/U index ) , compared to those with the lowest likelihood ( Low OD/U index ) , they were 1 . 36 times more likely to be infected with schistosomiasis ( 1 . 36 aPR , 95% CI: ( 1 . 15 , 1 . 61 ) . For every increase in a kilometer of distance there was a reduction in approximately 5% of schistosomiasis prevalence ( aPR = 0 . 95; 95% CI: ( 0 . 92 , 0 . 99 ) ) . Residence in an EA that received mass drug administration prior to the survey was associated with higher risk of schistosomiasis infection ( aPR 1 . 19 , 95% CI: ( 1 . 01 , 1 . 41 ) ) . A total of 433 water bodies were identified from field work that were recorded to be within 10 kilometers of a household . There was a decreasing trend in the prevalence of schistosomiasis as the household distance from a water body increased ( S1 Fig ) . The decreasing trend is also at the community level where prevalence of schistosomiasis decreases by 1 . 1% for every increase in a kilometer of mean distance between water bodies and the households in an EA ( S2 Fig ) .
This study found a national prevalence of schistosomiasis across Uganda at 25 . 6% ( 95% CI: 22 . 3 , 29 . 0 ) with CCA testing by using a large , representative sample of EAs and probability sample weights . The study had complete adherence to the random sample reaching all 170 pre-selected EAs by UBOS and found that intestinal schistosomiasis is prevalent and widespread across Uganda , even outside of areas known to be endemic . This is in line with a similar finding from western Tanzania where distribution of S . mansoni was more widespread than was previously understood [42] . This was the first national survey in Uganda to investigate schistosomiasis beyond known endemic areas and adds to work on the epidemiology of S . mansoni in Uganda by Kabatereine et al . that purposively selected communities in areas near large water bodies , where prevalence was expected to be high [22] . The known areas of high prevalence were included at random into the study sample and resulted in a national prevalence with CCA testing that averaged these areas of high transmission with areas of lower transmission . The finding that the Northern and Eastern regions of Uganda had higher prevalence is consistent with previous work that focused on the more endemic areas [12 , 22 , 43] . Prior estimates suggested that in Uganda 4 million individuals were infected with schistosomiasis caused by S . mansoni [44] and this study increases that estimate to over 10 million . This has implications for treatment programs to increase coverage across Uganda . The higher prevalence in the communities that had received mass drug administration prior to the survey ( 37 . 8% ( 95% CI: 32 . 0% , 43 . 5% ) ) indicates that the communities already receiving treatment may need more intense and focused efforts to bring the disease under control . Children of preschool age , from 2 to 4 years old , were identified as the most at-risk age group with a schistosomiasis prevalence of 36 . 1% ( 95% CI: ( 30 . 1 , 42 . 2 ) , n = 980 ) using CCA testing . This improves our understanding of the disease in this age group , which is often ignored in schistosomiasis control work given their ineligibility to receive drug treatment . The high prevalence found in preschool aged children supports prior studies that identified this age group as particularly important to treat to reduce heavy infections and early cumulative morbidity as they progress through childhood [45 , 46] . Early treatment of this at-risk age group could also mitigate the educational , learning , and memory deficits that Schistosoma infection and non-treatment have been found to be associated with in school aged children [47] . The high prevalence rate among children aged 2 to 4 years old also provides insight into how national control campaigns , that focus on drug delivery to school aged children , may have shifted the peak of age-based infection profiles to the preschool age groups , which previously peaked among school-aged children [1 , 22 , 48] . Preschool aged children may be spreading schistosomiasis as they engage in high risk activities for contaminating water bodies such as bathing and playing in surface water while their caregivers wash clothes [49] . The practical implication of this is that school-based treatments may be insufficient to reduce the disease burden in children [50] and there is a need to formulate a child-appropriate praziquantel tablet that can be safely administered to preschool aged children in control programs [51] . This high disease burden highlights the need for equitable treatment of young children to close the “praziquantel treatment gap” and reduce Schistosoma transmission in the general population [52] . As demonstrated in this study , preschool aged children should be treated with a praziquantel formulation in areas found to be highly endemic with CCA-based surveys [53] . The primary risk factors for S . mansoni that had strong associations with infection were an individual’s open defecation behaviors . Improved water and sanitation infrastructure in a household was not associated with reduced infection , indicating that the presence of adequate sanitation does not necessarily guarantee its use [54] . This is particularly important in the context of schistosomiasis where the bulk of the eggs reaching the freshwater snails are thought to stem from direct urination or defecation , largely by children that are bathing and swimming in the open water [55] . To sustain schistosomiasis transmission only a few eggs are required to enter freshwater and therefore a small number of individuals can continue the transmission cycle . For S . mansoni eggs that can survive out of water for up to eight days [56] , feces left in the bush or in latrines near water bodies may be washed in from rains or trodden into water bodies by people or animals [57] . The finding from this study that lower rates of open defecation in a community are associated with lower schistosomiasis disease prevalence is consistent with previous work that found increased latrine coverage along with a decrease in community wide open defecation reduced schistosomiasis prevalence [58] . In Uganda past surveying and mapping efforts have identified hotspots for schistosomiasis and other soil-transmitted helminths that pose a particular risk in the context of low sanitation coverage [59] . The protective relationship found between a household’s increasing distance from a water body and an individual’s infection with schistosomiasis confirms that human surface water use and surface water contact behaviors are contributing factors of disease transmission [60] . Notably , this protective relationship was found for individuals living in households up to 10 kilometer distances from water bodies , which is more geographically widespread than national control programs typically administer mass drug treatment . There was no significant difference in disease prevalence using CCA testing found between the near and far strata used for sampling EAs . The implication for this is that future sampling strategies should focus more on continuous distances from water bodies rather than dichotomous categories . Much of lakeside Uganda is characterized by largely mobile , itinerant fishing communities , whose frequent movements spread intestinal schistosomiasis outside of these hotspot areas and may also result in their children missing school-based annual mass drug administration [61] . The freshwater habitats for the intermediate snail hosts have also been found outside of targeted areas for control where natural transmission is thought unlikely [62 , 63] . For Uganda this requires control program implementation on a much larger scale than has been previously executed given that the majority of communities in Uganda are accessible to freshwater . A strength of the study is that the sampling frame allows the results to be interpreted as a national prevalence estimate for the Ugandan population . This was made possible by the sampling design and probability weights used from the collaboration with UBOS . Similar national prevalence surveys using CCA testing could be undertaken in other countries in the African Great Lakes Region plagued with high endemicity of schistosomiasis . Bureaus of statistics should be consulted to use a multi-stage cluster design to randomly sample EAs based on distances from water bodies and provide probability sample weights . Another strength was the use of POC-CCA , a commercialized urine-based antigen test that is used specifically for the detection of active S . mansoni infection in humans [64] . The use of POC-CCA enabled the large , nationwide scale of the study with its ease of use under field conditions and minimal training requirements for its application [65] . The use of urine samples instead of stool also increased its acceptability by respondents which make it ideal for epidemiological surveys in large populations [66] . The POC-CCA test has been suggested to be more sensitive than the Kato-Katz stool smears , which have historically been used for disease diagnosis in previous studies [67 , 68] . Related to this , the main limitation of the study is that the POC-CCA detection method may over-estimate prevalence in low prevalence areas due to cross-reactions with other helminth infections to produce false-positives [68] . The inclusion of lower intensity infections may have also over-estimated prevalence in low endemic areas where individuals without active schistosomiasis show residual reactivity ( trace ) POC-CCA [69] . This is also relevant in the context of control programs where the intensity of infection as measured by Kato-Katz is an important measure and one that is not available using POC-CCA . In a recent study that compared diagnostic methods to detect S . mansoni , the POC-CCA cassette test was recommended for moderate and high prevalence areas and used for screening and geographical mapping of S . mansoni infections [70] . This has also been validated in school-aged children in endemic settings around Lake Victoria to show that it is an appropriate and effective means of rapidly testing for intestinal schistosomiasis [29] . The current targets for eliminating schistosomiasis in Uganda by 2020 are most likely not going to be met as indicated by the national prevalence rate of 25 . 6% found in this study and stated by others in the literature [71] . There is a clear imperative for more accurate data to determine the treatment strategies for all age groups and incorporate the scale up of water and sanitation interventions [60 , 72 , 73] . To accelerate progress toward elimination of schistosomiasis in Uganda there is a need to strengthen evidence on how to deliver effective WASH interventions for neglected tropical diseases and break down existing programming silos built around single interventions [74 , 75] .
|
Schistosomiasis is a neglected tropical disease in sub-Saharan Africa that has remained intractable despite efforts to eliminate it through mass drug administration . The transmission cycle is perpetuated when sanitation infrastructure does not adequately capture infected urine or feces and local water bodies , with snail vectors , are contaminated . Schistosomiasis has been linked with stunting and cognitive deficits and there is particular concern for the most vulnerable age group under five years old who are undergoing critical intestinal development but are ineligible to receive drug treatment . Efforts to reduce the disease have focused on children and young adolescents in endemic areas , near water bodies where transmission is known to be high . In Uganda , where fresh water bodies are abundant and intestinal schistosomiasis is endemic , very little is understood about the disease prevalence at a national level . We conducted a large , nationally representative survey and found a national prevalence of 25 . 6% where the 2–4 year old children had the highest prevalence for schistosomiasis with 36 . 1% infected . The most significant risk-factor for the disease was an individual’s open defecation behaviors in surface waters . This emphasizes the need to include water and sanitation investments alongside drug treatment and behavior change to control schistosomiasis in Uganda .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"schistosoma",
"invertebrates",
"schistosoma",
"mansoni",
"medicine",
"and",
"health",
"sciences",
"surface",
"water",
"helminths",
"tropical",
"diseases",
"geographical",
"locations",
"uganda",
"parasitic",
"diseases",
"animals",
"health",
"care",
"age",
"groups",
"pharmaceutics",
"drug",
"administration",
"neglected",
"tropical",
"diseases",
"sanitation",
"africa",
"hydrology",
"public",
"and",
"occupational",
"health",
"people",
"and",
"places",
"helminth",
"infections",
"schistosomiasis",
"environmental",
"health",
"eukaryota",
"earth",
"sciences",
"population",
"groupings",
"biology",
"and",
"life",
"sciences",
"drug",
"therapy",
"organisms"
] |
2019
|
The prevalence of schistosomiasis in Uganda: A nationally representative population estimate to inform control programs and water and sanitation interventions
|
Soil-transmitted helminths ( STH ) infect more than 2 billion humans worldwide , causing significant morbidity in children . There are few data on the epidemiology and risk factors for infection in pre-school children . To investigate risk factors for infection in early childhood , we analysed data prospectively collected in the ECUAVIDA birth cohort in Ecuador . Children were recruited at birth and followed up to 3 years of age with periodic collection of stool samples that were examined microscopically for STH parasites . Data on social , demographic , and environmental risk factors were collected from the mother at time of enrolment . Associations between exposures and detection of STH infections were analysed by multivariable logistic regression . Data were analysed from 1 , 697 children for whom a stool sample was obtained at 3 years . 42 . 3% had at least one STH infection in the first 3 years of life and the most common infections were caused by A . lumbricoides ( 33 . 2% of children ) and T . trichiura ( 21 . 2% ) . Hookworm infection was detected in 0 . 9% of children . Risk of STH infection was associated with factors indicative of poverty in our study population such as Afro-Ecuadorian ethnicity and low maternal educational level . Maternal STH infections during pregnancy were strong risk factors for any childhood STH infection , infections with either A . lumbricoides or T . trichiura , and early age of first STH infection . Children of mothers with moderate to high infections intensities with A . lumbricoides were most at risk . Our data show high rates of infection with STH parasites during the first 3 years of life in an Ecuadorian birth cohort , an observation that was strongly associated with maternal STH infections during pregnancy . The targeted treatment of women of childbearing age , in particular before pregnancy , with anthelmintic drugs could offer a novel approach to the prevention of STH infections in pre-school children .
Soil-transmitted helminths ( STH ) , including A . lumbricoides lumbricoides , T . trichiura trichiura , and hookworm , are estimated to infect more than 2 billion humans worldwide [1] of which 51 million children are considered to be at risk of morbidity [2] . An estimated 35 million or more disability-adjusted life years [3] have been attributed to STH infections . Morbidity due to STH infections has primarily been associated with anaemia , malnutrition [4] , stunting [5] , and cognitive impairment [6] . Effects on childhood growth have been attributed to changes in appetite , digestion , nutrient absorption and iron loss [7] . Current strategies for the control of STH infections are primarily based upon periodic treatment of schoolchildren with anthelmintic drugs , and secondarily on education and improvements in sanitation . Treatment-based control strategies aim to control morbidity through reductions in the community transmission of STH infections [8] . Previous studies have shown that the main risk factors for STH infection are rural residency , low socioeconomic status and poor sanitation [2] , [3] . The use of pit latrines and improved drinking water have been associated with a reduced prevalence of STH infections [4] while maternal STH infections with increased risk [9] , indicating the importance of identifying risk factors amenable to targeted interventions . There are few data on the epidemiology of STH infections and risk factors for infection in pre-school children . Such data are relevant to the control of STH infections because pre-school children constitute an important reservoir of infection and are at risk of morbidity . To investigate the epidemiology of STH infections in early childhood and to identify risk factors for infection , we analysed data collected prospectively during the first 3 years of life in the ECUAVIDA birth cohort in tropical Ecuador .
Details of the design and methodology for the ECUAVIDA birth cohort study , an investigator-driven study , are provided elsewhere [10] . Briefly , 2 , 404 newborns in the district of Quinindé in Esmeraldas Province , Ecuador , were recruited at the District hospital between November 2005 and December 2009 . Inclusion criteria included being a healthy baby , the collection of a maternal stool sample , and planned residence in the District for at least 3 years . The study area is tropical at an elevation of up to 200 m , average annual temperature of 30°C and relative humidity of 75% , and with a population of ∼150 , 000 living in three towns and six rural parishes . The District is poor with limited access to clean water , sanitation , and basic services even within the towns . All children recruited to the ECUAVIDA birth cohort study were eligible for inclusion in this analysis . Children were actively followed from birth to 3 years of age with collection of stool samples at 3 , 7 , 13 , 18 , 24 , 30 and 36 months . Sample collection took place between November 2005 and December 2012 covering the period of cohort recruitment and the period to 3 years of age of the whole cohort . Stools at 3 , 18 , and 30 months were collected passively ( mothers were asked to provide a sample at the next time point during the previous follow-up ) , while mothers were actively requested for stool samples at the respective ages for the remaining time points . Household members were also asked to provide one stool sample around the time that the mother was enrolled into the study . Data on risk factors and potential confounders were collected by a questionnaire that was administered to the child's mother by a trained member of the study team around the time of birth of the child . Categories included in the questionnaire were; maternal and paternal data ( age , ethnicity , education , occupation , and number of live children [mother only] ) , urban versus rural location , socio-economic data ( monthly income , number of material goods [household electrical appliances including refrigerator , television , Hi-Fi , and radio] , a household electrical connection , household construction materials , sources of drinking water , and type of bathroom [for disposal of faeces] ) , number of sleeping rooms and number of people living in the household , and data on exposure to household pets , farming and farm animals . Household overcrowding was defined as number of people living in the household per bedroom . Data on number of anthelmintic treatments during the first 3 years of life was obtained from a questionnaire administered to the mother when the child was 7 , 13 , 24 , and 36 months . Single stool samples were collected and analysed for STH eggs and larvae by direct saline wet mounts ( for detection of all STH eggs including A . lumbricoides , T . trichiura , S . stercoralis , hookworm , and tapeworms ) , Kato-Katz ( for quantification of A . lumbricoides and T . trichiura ) and formol-ether concentration ( for detection of eggs/larvae of all STH and tapeworm parasites ) methods [11] . A . lumbricoides and T . trichiura infection intensities were expressed as eggs per gram ( epg ) of faeces using the results of Kato-Katz . The intensities of hookworm and S . stercoralis were not evaluated because of low prevalence and intensities in children of this age . A positive sample was defined by the presence of at least one egg or larva from any of the three detection methods . A sample size of 1 , 697 individuals included in this analysis was estimated to provide over 80% power at P<0 . 05 to detect exposure effects on risk of any STH prevalence with effect sizes of or less than 0 . 56 ( 10% exposure prevalence ) , 0 . 68 ( 20% ) , 0 . 72 ( 30% , and 0 . 74 ( 40% ) . Any STH infection was defined by the presence of at least one STH ovum or larva in any stool sample . Potential risk factors evaluated included parental factors , child factors ( gender , gestational age , and birth order ) , socioeconomic status and factors relating to the environment in which the child was living . A socio-economic status ( SES ) index was created using principal components analysis for categorical data by combining the socioeconomic variables . The first component that accounted for 30 . 0% of variation was divided into tertiles to represent low , middle , and high SES . The primary outcome was infection with any STH infection during the first 3 years of life . Secondary outcomes were age of first STH infection and any infection with A . lumbricoides or T . trichiura during the first 3 years of life . Estimates of effect for any STH infection and any infections with A . lumbricoides or T . trichiura during the first 3 years of life were calculated using multivariable logistic regression controlling for potential confounders using a backwards-stepwise procedure in which variables considered for inclusion were those with P<0 . 2 in univariate analyses . The variables , number of samples collected and number of anthelmintic treatments received and gender , were included in all analyses as potential confounding variables , the first because a greater number of samples collected increase the chances of detection of STH . Age of first STH infection was analysed by multinomial logistic regression , a technique that allows a single comparison group ( uninfected ) for more than one mutually exclusive outcome . This analysis allowed us to evaluate simultaneously , associations between risk factors and first STH infections acquired during the first , second and third years of life , respectively , compared to children without any documented STH infection . For multivariable multinomial logistic regression analysis , we used a backwards-stepwise method for selection of covariates – covariates considered in this analysis were those with P<0 . 005 in univariate analyses . The population-attributable fraction ( PAFs ) was calculated by: Pew X ( OR-1 ) /OR where Pew is the prevalence of maternal STH infection among children with any STH infection during the first 3 years . Statistical significance was inferred by P<0 . 05 . Analyses were done using SPSS ( Version 16 ) and STATA version 10 ( Statacorp , TX ) The study protocol was approved by the ethics committee of the Hospital Pedro Vicente Maldonado , Universidad San Francisco de Quito , and Pontificia Universidad Catolica del Ecuador , Ecuador . The study is registered as an observational study ( ISRCTN41239086 ) . Informed written consent was obtained from the child's mother and from household members for the collection of stool samples . Children with positive stools for STH infections were treated with a single dose of 400 mg albendazole if aged 2 years or greater and with pyrantel pamoate ( 11 mg/kg ) if aged less than 2 years , according to Ecuadorian Ministry of Public Health recommendations [12] , [13] .
Of 2 , 404 newborns recruited , 1 , 697 ( 70 . 6% ) of children had a stool sample examined at 3 years of age and these were the children included in the present analysis . Follow-up to 3 years of age for stool sampling for the whole cohort is illustrated in Figure 1 . There were no significant differences between excluded and included children except for non-maternal STH infections among household members that were more frequent among excluded children ( data provided in Table S1 ) . Of the 1 , 697 children analysed here , 718 ( 42 . 3% ) were infected with at least one STH infection during the first 3 years of life . The most frequent STH infection was A . lumbricoides ( 33 . 2% of children had at least one documented infection during the first 3 years ) , followed by T . trichiura ( 21 . 2% ) , Strongyloides stercoralis ( 1 . 4% ) and hookworm ( 0 . 9% ) . Other enteric parasites observed during the first 3 years of life included Hymenolepis spp . ( 2 . 2% of children ) , Giardia lamblia ( 44 . 8% ) and Entamoeba histolytica/dispar ( 28 . 5% ) . The prevalence of STH infections by age is shown in Figure 2 . STH infections appeared after 3 months of age and prevalence increased with age . The diagnostic methods used were not optimal for the detection of S . stercoralis: the infection was first detected at 13 months and prevalence did not vary substantially between 13 months and 3 years ( 0 . 4% at 13 months , 0 . 1% at 18 months , 0 . 6% at 24 months , 0 . 8% at 30 months , and 0 . 5% at 36 months . Mean age at first infection with an STH parasite among the study children was 23 months ( SD 9 months , range 7–36 months ) . For comparisons with other studies , we stratified the children's infection intensities according to WHO recommendations [14] . The majority of infections were of low intensity ( Figure 3 ) . Geometric mean infection intensities among children with A . lumbricoides or T . trichiura infections , respectively , were: 13 months - 931 epg ( range 35-182 , 910 ) or 244 epg ( range 35-6 , 860 ) ; 24 months - 1 , 859 epg ( range 35-176 , 750 ) or 248 epg ( range 35-21 , 875 ) ; 36 months - 1 , 973 epg ( range 35-283 , 930 ) or 243 epg ( range 35-47 , 425 ) . Anthelmintic treatments were widely available and were provided by the study team with a positive stool examination for STH infections and were also obtained directly from pharmacies by the mothers: 75 . 6% of children were treated at least once during the first 3 years of life and the proportions receiving anthelmintic treatments at 0–7 , 8–13 , 14–24 , and 25–36 months were 0 . 8% , 16 . 5% , 44 . 5% , and 47 . 0% , respectively . No differences were observed in the number of anthelmintic treatments received between infected and uninfected children ( Table 1 ) . At enrolment , STH infections were present in 45 . 7% of mothers , 31 . 0% of fathers and 53 . 3% of other household members . The estimated prevalence of S . stercoralis among mothers was 4 . 0% . The distributions of covariates between children with any documented STH infection and those without a documented infection are shown in Table 1 . Factors that were significantly more frequent among infected children were; being lower in the birth order , having a younger mother or Afro-Ecuadorian mother or illiterate mother , being of lower socioeconomic status , urban residence , living in a more crowded household , having a mother infected with an STH parasite during pregnancy particularly having a mother with moderate to high parasite burdens with A . lumbricoides or T . trichiura , and having a father or other household member infected with an STH parasite after the child's birth . The risk of infection was greater among children who provided 5 or more stool samples during the 3 years of observation . The number of anthelmintic treatments received by the child did not affect the risk of having any STH infection . There was some evidence that the prevalence of S . stercoralis was higher among mothers of children with STH infections compared to those without ( 6 . 0% vs . 2 . 3% , P<0 . 001 ) . The results of univariate and multivariable analyses are shown in Table 2 . In multivariable analyses , being lower in the birth order , maternal Afro-Ecuadorian ethnicity and younger age , being of low socioeconomic status , urban residence , and household overcrowding were significant independent predictors of STH infection during the first 3 years of life . Intensity of maternal A . lumbricoides infection produced the highest odds ratios; children were 11 . 6 times more likely to have STH if their mothers had moderate to heavy infection intensities with A . lumbricoides . This estimate was imprecise with high confidence intervals but is consistent with an analysis in which maternal A . lumbricoides infection intensities were stratified as tertiles and the highest tertile was compared with negatives ( adj . OR 4 . 46 ( 95% CI 2 . 68–7 . 41 ) , O<0 . 001 ) ( data not shown ) . Having household members with STH infections also increased the child's chance of having an STH infection as did the number of stool samples collected for each child . In a separate analysis , we estimated the association between any maternal STH infection and any STH infection in children , excluding the other maternal STH variables shown in Table 2 , giving an adjusted OR of 1 . 88 ( 95% CI 1 . 46–2 . 41 , P<0 . 001 ) . This estimate corresponds to a fraction of child STH infections attributable to maternal STH infections ( PAF% ) of 27 . 9% . To distinguish the effects of increased risk of child infection associated with STH during pregnancy from that of having a mother with STH infection during the first year of life , we stratified the data according to whether the mother was the child's primary carer during the first year of life or not: data were available for 1 , 639 children of whom mothers were the primary carer for 95 . 1% . The association between maternal STH infection and any child infections was greater among children with primary mother carers ( adj . OR 1 . 84 , 95% CI 1 . 40–2 . 41 , P<0 . 001 ) compared to those with non-mother carers ( adj . OR 0 . 35 , 95% CI 0 . 05–2 . 58 , P = 0 . 305 ) although the latter group consisted of only 81 children ( data not shown ) . Most STH infections during the first 3 years of life were caused by A . lumbricoides . Risk factors for any A . lumbricoides or any T . trichiura infection during the first 3 years of life in multivariable analyses were similar to those documented for any STH infection and showed strong associations with maternal STH infections: 1 ) Any A . lumbricoides infections was associated with greater maternal infection intensities with A . lumbricoides ( Moderate/heavy vs . uninfected , adj . OR 3 . 88 , 95% CI 2 . 12–7 . 08 , P<0 . 001 ) and maternal T . trichiura infections ( adj . OR 1 . 38 , 95% CI 1 . 05–1 . 82 , P = 0 . 021 ) ; 2 ) Any T . trichiura infection was associated with having a mother with greater maternal infection intensities with A . lumbricoides ( Moderate/high vs . uninfected , adj . OR 5 . 85 , 95% CI 3 . 29–10 . 40 , P<0 . 001 ) or infection with T . trichiura ( adj . OR 1 . 71 , 95% CI 1 . 36–2 . 56 , P<0 . 001 ) during pregnancy . ( Complete results are provided in the Table S2 ) . Multivariable analyses showed that maternal infections with either A . lumbricoides or T . trichiura were consistently strong predictors of age at first infection with STH parasites across the first 3 years of life ( Table 3 ) . Maternal infection with A . lumbricoides during pregnancy doubled the odds of a child being infected during the first year of life ( adj . OR 2 . 34 , 95% CI 1 . 61–3 . 40 ) - a significant effect was seen also for first infections acquired in the second year of life . Similar effects were observed for age of first infection among children born to mothers with T . trichiura infections . Maternal Afro-Ecuadorian ethnicity and low educational level were also associated with an increased risk across the first 3 years of life , although the association was not significant for low maternal educational level in the second year of life . Being lower in the birth order and having a household member with an STH infection were associated only with early infections with STH in children ( i . e . first year of life ) . Household overcrowding was associated with first infections after the 1st year of life .
In the present analysis , we investigated the epidemiology of and risk factors for STH infection during the first 3 years of life in a birth cohort from a largely rural District in tropical Ecuador . Over 40% of children had at least one STH infection documented during the first 3 years of life . Almost all infections ( 96 . 9% ) were caused by A . lumbricoides and T . trichiura , although few of these pre-school children harboured heavy parasite burdens . Markers of poverty were independent risk factors for any STH infections or infections with individual parasites . STH infection risk in the cohort children was strongly associated with maternal STH infections during pregnancy , particularly children with mothers with moderate to high infection intensities with A . lumbricoides who had an approximately 12-fold increased risk of infection . Potential limitations to the present study include losses to follow-up – we collected a stool sample from 70% of the original cohort at 3 years of age . Such losses could lead to selection bias . However , baseline variables were generally similar between those included and excluded from the analysis indicating that selection bias is probably not an important issue . Although we attempted to control for potential confounders , we cannot exclude confounding by uncontrolled factors or by highly correlated exposures as an alternative explanation for our findings . Approximately 76% of children were reported by mothers to have received at least one anthelmintic treatment during the first 3 years of life . This proportion did not vary significantly between infected and uninfected children indicating that mothers of children who did not receive anthelmintic treatment for their child for a positive stool examination were extremely likely to obtain anthelmintic drugs through other sources irrespective of a negative stool examination and concurs with our own experience that a lot of illness in children is attributed by mothers to the presence of ‘parasites’ and that self-medication is extremely common . Although 30% of all children had received 2 or more doses of anthelmintic drugs , number of treatments was not associated with risk of STH infections , an observation that might be explained by misclassification of this variable ( number of anthelmintic treatments ) or by high rates of reinfection over the year following treatment . The use of questionnaires to collect data on exposures is subject to reporting biases although these are unlikely to be systematic . Because data on risk factors was collected around the time of birth and before the measurement of outcomes , observation biases would seem unlikely to be important . Our findings are likely to be relevant to young children living in poor rural Districts of tropical Latin America and other similar regions elsewhere . Strengths of the study are the longitudinal nature of the study allowing repeated sampling of the same children over time and a large sample size . The study was originally designed and powered to examine the effects of maternal and early childhood infections with STH parasites on the development of atopy and allergic diseases [10] . These estimates that allowed losses to follow-up of 25% at 3 years for less common outcomes than childhood STH infections , nevertheless , had high power . Most STH infections are unable to replicate within the human host and the acquisition of increasing parasite burdens is time and exposure-dependent . Unsurprisingly STH infection prevalence increased with age in the cohort with the highest prevalence of 24 . 9% observed at 36 months . This is within the 19 . 6–35 . 5% estimate by PAHO-WHO of STH prevalence in Ecuadorian pre-school children [14] . Age-specific estimates of prevalence and intensity may have been lower than expected in the present study because of the ethical requirement to provide treatment whenever a positive sample was detected . Schoolchildren would be expected to have a higher prevalence of STH infections: a previous study of schoolchildren living in two other rural Districts in Esmeraldas Province estimated a prevalence of 74 . 9% using the same diagnostic methods and a single stool sample [15] . The use of several diagnostic methods as done in the present study is useful to maximise the sensitivity for STH detection [16] . However , the methods we used have limited sensitivity for the detection of S . stercoralis and could be improved by using more sensitive molecular diagnostics such as PCR . The strongest and most consistent risk factor for infection with STH in the first 3 years of life was maternal STH infections , particularly among children whose mothers harboured moderate to high parasite burdens with A . lumbricoides during pregnancy . A previous case-control analysis of a single stool sample collected from 1 , 004 children aged 7 months to 3 years , nested within the same cohort , showed that children of mothers infected with STH parasites during pregnancy had a higher risk of infection compared to children of uninfected mothers [9] . We now have extended these analyses to look prospectively at the acquisition of infection during the first 3 years of life in the whole cohort . Our data suggest that maternal STH infections , particularly moderate to heavy parasite burdens with A . lumbricoides , are an important independent determinant of risk of STH infections during early childhood . The association between STH infections in mothers and infections of children has two possible explanations . 1 ) STH infections of the mother during pregnancy may increase susceptibility to infection in offspring through tolerization to parasite antigens - A . lumbricoides antigens can be detected in the circulation of infected individuals [17] that may cross the placenta and engage with the foetus's developing immune system . There is evidence from animal models of helminth infection [18] and studies in humans [19] , [20] , [21] that maternal infections may induce foetal tolerance to parasites and increase susceptibility to helminth infection in offspring . We have shown previously in the same study population that maternal STH infections induce immunologic sensitization to A . lumbricoides antigens in utero [22] and that the cord blood of newborns born to infected mothers has elevated levels of the immune regulatory cytokine IL-10 compared to those born of uninfected mothers [9] . Such tolerization could have a genetic basis , and there is ample evidence that susceptibility to STH infections is associated with immune genes [23] , [24] . A recent study of children in urban Brazil provided evidence that IL-10 polymorphisms were associated with susceptibility to STH infections [25] . 2 ) The higher risk of infection in children of infected mothers may reflect a shared environment particularly during the first year of life when the child is completely dependent on the mother . This explanation is supported partly by the observation that among children whose primary carer during the first year of life was the mother , the association between maternal STH infection in pregnancy and any STH infection in offspring was stronger than for children whose primary carer was not the mother , although small numbers in the latter group yielded a very imprecise estimate . Maternal STH infections were a common risk factor - approximately 46% of mothers were infected with STH in their third trimester of pregnancy , with an estimated attributable fraction of 27 . 9% . Our observations , therefore , have identified a potentially modifiable exposure - maternal STH infections - that could be evaluated in an intervention programme using currently available and highly efficacious anthelmintic treatments . An intervention in which anthelmintic drugs are given before pregnancy for women planning to have a family , during pregnancy or soon after birth or even periodically to women of child-bearing age could substantially reduce the risk of infection and potential morbidity during early childhood . Development of immune tolerance begins after 14 weeks gestation [26] so it might be beneficial to deworm women before pregnancy or if pregnant , in the second trimester before the foetal immune system is capable of developing tolerance to parasite antigens . Clearly , deworming of mothers before pregnancy would be preferable because giving anthelmintic drugs to pregnant or lactating women carries a risk of potential adverse effects on the foetus or neonate , respectively . Albendazole has been indicated by WHO for use in pregnant women after the first trimester in areas that are highly endemic for hookworm because of the risks to mother and child of maternal hookworm anaemia [27] . We now suggest another potential benefit of deworming mothers with other STH infections – to reduce the risk of STH infection and associated morbidity in early childhood . There are still limited data on the safety of anthelmintic drugs in pregnant and lactating women although benzimidazole drugs are believed to be safe when used after the first trimester or during breastfeeding [27] . Such a strategy ( to reduce the risk of STH infection and associated morbidity in early childhood ) will only be useful in areas where A . lumbricoides and T . trichiura are endemic because hookworm is less a problem in pre-school children among whom prevalence is generally low even in highly endemic areas [28] . The use of anthelmintic treatment in mothers , therefore , to prevent child infections will require careful consideration of the balance of potential adverse effects in the mother and child versus the clinical consequences of infections in the pre-school child for which there is growing evidence of important nutritional effects [29] . Such decisions will almost certainly have to be made locally depending on the epidemiology of STH parasites in any specific area or region . An alternative strategy would be the periodic treatment of women of childbearing age . A model illustrating risk factors for childhood STH infections in our study with potential interventions is provided in Figure 4 . Risk factors , in addition to maternal STH infections , identified in our analysis were markers of low socioeconomic status such as use of untreated water for drinking , and not having a flushing toilet in the house . Having a younger mother of Afro-Ecuadorian ethnicity , being lower in the birth order , and household overcrowding were also independently associated with childhood STH infections . Previous studies have identified similar factors as being associated with risk of STH infection [4] , [17] , [30] , [31] , [32] , [33] , [34] . Among public health interventions that could be implemented to reduce STH infection risk in pre-school children are the provision of sanitation , clean water , and education of mothers in appropriate hygienic behaviours . Significant changes in hygiene practices will be required to sustain any reductions achieved by chemotherapy [35] . However , treatment of mothers ( ideally before pregnancy ) and their young children or the periodic treatment of women of child-bearing age are the only strategies that can be implemented in the short-term with immediate benefits and without major municipal investment in local infrastructure and services . The targeting of anthelmintic treatment to pregnant women attending antenatal clinics , and mothers and their young children attending public vaccination clinics that serve the most vulnerable sections of the population in STH endemic areas , provides a rapid and easily implemented strategy for the control of STH infections in pre-school children . In conclusion , our study identified risk factors for STH infection during the first 3 years of life in a birth cohort conducted in a rural District in coastal Ecuador . Over 40% of children were infected at least once with STH parasites during the first 3 years of life and risk factors for infections were those associated with poverty . We identified maternal STH infections as an important and potentially modifiable risk factor that could be evaluated in future intervention studies for the control of STH infections in pre-school children .
|
Soil-transmitted helminths ( STH ) are intestinal worms that cause significant morbidity in school age and pre-school children in developing countries . Infections are associated with poverty , particularly through lack of access to sanitation and clean drinking water . Current control strategies rely on periodic anthelmintic treatment of schoolchildren but new strategies are required for STH control in young children . There are few data on modifiable risk factors in pre-school children . We investigated environmental and socioeconomic risk factors for STH infection in the first 3 years of life in a birth cohort from an STH-endemic region of Latin America . Our data provide evidence that maternal STH infections documented during pregnancy are an important risk factor for infection in young children , raising the possibility of a novel intervention for the prevention of STH-associated morbidity in early childhood through the deworming of women of childbearing age , in particular before pregnancy .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"soil-transmitted",
"helminths",
"parasitic",
"intestinal",
"diseases",
"trichuriasis",
"infectious",
"disease",
"epidemiology",
"epidemiology",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"ascariasis",
"parasitic",
"diseases",
"helminth",
"infection"
] |
2014
|
Risk Factors for Soil-Transmitted Helminth Infections during the First 3 Years of Life in the Tropics; Findings from a Birth Cohort
|
In bacteria , one paradigm for signal transduction is the two-component regulatory system , consisting of a sensor kinase ( usually a membrane protein ) and a response regulator ( usually a DNA binding protein ) . The EnvZ/OmpR two-component system responds to osmotic stress and regulates expression of outer membrane proteins . In Salmonella , EnvZ/OmpR also controls expression of another two-component system SsrA/B , which is located on Salmonella Pathogenicity Island ( SPI ) 2 . SPI-2 encodes a type III secretion system , which functions as a nanomachine to inject bacterial effector proteins into eukaryotic cells . During the intracellular phase of infection , Salmonella switches from assembling type III secretion system structural components to secreting effectors into the macrophage cytoplasm , enabling Salmonella to replicate in the phagocytic vacuole . Major questions remain regarding how bacteria survive the acidified vacuole and how acidification affects bacterial secretion . We previously reported that EnvZ sensed cytoplasmic signals rather than extracellular ones , as intracellular osmolytes altered the dynamics of a 17-amino-acid region flanking the phosphorylated histidine . We reasoned that the Salmonella cytoplasm might acidify in the macrophage vacuole to activate OmpR-dependent transcription of SPI-2 genes . To address these questions , we employed a DNA-based FRET biosensor ( “I-switch” ) to measure bacterial cytoplasmic pH and immunofluorescence to monitor effector secretion during infection . Surprisingly , we observed a rapid drop in bacterial cytoplasmic pH upon phagocytosis that was not predicted by current models . Cytoplasmic acidification was completely dependent on the OmpR response regulator , but did not require known OmpR-regulated genes such as ompC , ompF , or ssaC ( SPI-2 ) . Microarray analysis highlighted the cadC/BA operon , and additional experiments confirmed that it was repressed by OmpR . Acidification was blocked in the ompR null background in a Cad-dependent manner . Acid-dependent activation of OmpR stimulated type III secretion; blocking acidification resulted in a neutralized cytoplasm that was defective for SPI-2 secretion . Based upon these findings , we propose that Salmonella infection involves an acid-dependent secretion process in which the translocon SseB moves away from the bacterial cell surface as it associates with the vacuolar membrane , driving the secretion of SPI-2 effectors such as SseJ . New steps in the SPI-2 secretion process are proposed .
Gram-negative pathogens use type III secretion systems ( T3SS ) to secrete effectors into the host , which promote virulence and alter host signaling functions . Salmonella enterica serovar Typhimurium encodes two T3SS on Salmonella pathogenicity islands 1 and 2 ( SPI-1 and SPI-2 ) . Their unique secreted effectors are primarily active during different phases of infection . SPI-1 effectors promote adherence and initial infection of the intestinal epithelium , while SPI-2 effectors are responsible for survival and replication in the macrophage vacuole [1–3] and bacterial spreading to distal organs [4] . The SPI-1 needle complex has been well characterized both functionally and structurally [5–7] , but the SPI-2 needle complex is fragile and not very abundant and has not been well characterized . This raises questions about the conditions that induce SPI-2 needles during Salmonella infection and about how SPI-2 needles function . In the present work , we show that the Salmonella cytoplasm is acidified both in vitro and in vivo in response to acid stress . Furthermore , acidification is necessary for OmpR activation of SPI-2–dependent secretion , but not assembly . Thus , the macrophage vacuole provides signals that activate SPI-2 expression , assembly , and secretion in vivo , and these include acidification of the bacterial cytoplasm . After entry into the macrophage , Salmonella resides in a modified intracellular compartment , the Salmonella-containing vacuole ( SCV ) . Several studies indicate that the pH of this compartment is approximately 5 [8–10] , and the SCV is comparable to a normal phagosome in terms of its biogenesis [11] . Once inside the vacuole , the EnvZ/OmpR two-component system is activated . In Escherichia coli , the EnvZ/OmpR system is a sensor of osmotic stress , regulating expression of outer membrane porins [12] . We recently discovered that the EnvZ histidine kinase responds to cytoplasmic osmotic stress rather than extracellular stress , and it is capable of sensing osmolality and activating its downstream target OmpR without being in the membrane [13] . Based on this result , we postulated that Salmonella might respond to the acidic pH of the macrophage vacuole by acidifying its cytoplasm , providing the protons that drive formation of the activated conformation of EnvZ , which promotes phosphorylation and phosphotransfer to OmpR [13] . Thus , we set out to measure the cytoplasmic pH of Salmonella while it resides in the SCV . For these experiments , we employed a novel DNA biosensor ( the “I-switch” ) that undergoes non-Watson-Crick base pairing in the presence of excess protons , resulting in fluorescence resonance energy transfer ( FRET ) [14 , 15] . We tested the I-switch in Salmonella in vitro and then used I-switch–containing bacteria to infect RAW264 . 7 macrophages . This study represents the first application of the I-switch in which it has been used to measure the unknown pH of an intracellular compartment . Our results indicate that in vitro , Salmonella acidifies its cytoplasmic compartment in response to extracellular acid stress , and during infection , the Salmonella cytoplasm rapidly acidifies in response to the low pH of the vacuole in which it resides . Thus , the bacterial cytoplasm responds positively to extracellular ( vacuolar ) pH changes . Furthermore , EnvZ and OmpR are completely required for acidification , establishing OmpR as a regulator of genes that enable Salmonella to survive the acid stress of the SCV . The known , well-characterized targets of OmpR regulation , including OmpC , OmpF , and SPI-2 [16–18] were not required for acidification , indicating that new , unidentified OmpR targets were involved . Microarray analysis in the absence and presence of ompR at pH 5 . 6 indicated that approximately 390 genes were up-regulated . Of interest was the cadaverine operon cadC/BA , which is involved in recovery from acid stress . In the macrophage vacuole , these genes were repressed by OmpR and the Salmonella cytoplasm remained acidified . In the absence of ompR , Salmonella regulated its intracellular pH and returned to near neutral in a CadC/BA-dependent manner . This activation of EnvZ/OmpR by acid stress stimulates production of the SsrA/B two-component system ( located on SPI-2 ) , and ultimately , expression of SPI-2-secreted effectors [16–18] . During Salmonella infection , bacteria switch from assembling structural components of the T3SS ( represented herein by the translocon SseB ) to secreting effectors ( SseJ in this study ) into the host cytosol to modify host functions . We used the I-switch to test an existing model that proposed a neutralization step was required for the switch from translocon secretion to effector secretion . Our results indicate that when the vacuole is not acidified , effector secretion does not occur , in agreement with some previous results [19] but differing from others [20] . Interestingly , we found that the onset of SseJ effector secretion correlated with an outward movement of the SseB translocon away from the bacterial cell surface . This outward movement coincided with an association of SseB with the vacuolar membrane . This step required vacuolar acidification , since the presence of the V-ATPase inhibitor Bafilomycin prevented vacuolar acidification and SseB remained on the Salmonella cell surface . As a result , effector secretion into the host cytosol was prevented . Our results suggest a new model where acidification of the bacterial cytoplasm drives the secretion and release of translocon protein ( s ) from the bacterial cell surface , which in turn facilitates effector secretion .
To determine the impact of acid stress on Salmonella intracellular pH ( pHi ) , we used the I-switch ( IA488/A647 ) labeled at its 5′ and 3′ termini with Alexa-488 and Alexa-647 , respectively , as a cytoplasmic pH sensor . The I–switch consists of cytosine-rich unpaired regions that form an anti-tetraplex CH+ . C by alternate Watson and Crick base pairing in the presence of protons . This leads to a “closed” conformation of DNA , enabling FRET to occur between the fluorophore pairs . This process is reversible , and at neutral pH , the I-switch dissociates into an open , extended conformation because of electrostatic repulsion between the duplex arms [14 , 15] . We first compared the fluorescence intensity of the donor alone ( 520 nm , "D" ) and the acceptor alone ( 666 nm , "A" ) with the fluorescence of the I-switch containing both donor and acceptor . This provided a determination of the cross-talk between the fluorophores , which was negligible ( S1A Fig ) . I-switch labeled with Alexa 488 and Alexa 647 was suspended in buffers ranging from pH 5 . 0 to 7 . 2 . The samples were excited at 480 nm and spectra were recorded in a Tecan spectrophotometer ( GENios ) from 510 to 700 nm . A D/A curve as a function of pH was plotted from the ratio of donor ( 520 nm ) to acceptor ( 666 nm ) intensities to generate the in vitro standard curve ( Fig 1A ) . The I-switch exhibited a sigmoidal increase over this range due to the formation of the I-motif in the closed state . We next investigated the ability of the I-switch to function inside bacterial cells by incorporating IA488/A647 into Salmonella by electroporation and verifying its incorporation by fluorescence microscopy ( Fig 1B ) . We confirmed its intracellular location by preparing spheroplasts , which eliminated the outer membrane but still retained the I-switch in the cytoplasm ( S1B Fig ) . We used 6 μM IA488/A647 for electroporation , since it showed optimum fluorescence at both donor and acceptor channels inside bacterial cells; this concentration was used in all subsequent experiments . The internal pH of Salmonella was clamped to the external pH of K+-containing media using 40 μM nigericin [21–23] . An intracellular D/A standard curve ( Fig 1A ) showed perfect correlation with the in vitro curve , indicating the functionality and integrity of the I-switch in Salmonella . The D/A profiles of cells clamped at acidic pH 5 . 1 were visibly distinct and showed a 5-fold increase from those clamped at neutral pH 7 . 1 . This study is the first application of the I-switch in bacteria . We next measured the change in cytoplasmic pH ( pHi ) in response to varying the pH of the extracellular media ( pHe ) . Salmonella containing the I-switch were subjected to acid stress and the fluorescence was measured in SPI-2–inducing ( pHe 5 . 6 ) and non-inducing ( pHe 7 . 2 ) conditions over time ( Fig 2A ) . The pHi of Salmonella decreased from 6 . 85 to 6 . 1 upon incubation at pHe 5 . 6 , but remained fairly constant when incubated at pHe 7 . 2 ( Fig 2B ) , indicating that the Salmonella cytoplasm acidifies in response to extracellular acidic pH . Addition of extracellular media at even lower pHe , e . g . , 4 . 5 , reduced the intracellular pH further , to pH 5 . 8 ( Fig 2C ) . Acidification was readily reversible upon addition of extracellular media at alkaline pH ( Fig 2C ) , which restored the near neutrality of the cytoplasm . This result further demonstrates the functionality of the I-switch; i . e . , it is not degraded , because it can reversibly switch from an open state ( low FRET at neutral pH ) to a closed state ( high FRET at acidic pH ) in response to changes in pHe . Because OmpR was implicated in the Acid Tolerance Response in Salmonella ( [24]; see also S2 Fig ) , we measured the response of an ompR null strain of Salmonella under similar acid stress conditions . In the ompR null strain , the intracellular pH did not decrease in response to extracellular pH stress ( Fig 2A and 2B ) . OmpR-dependent intracellular acidification was completely restored upon supplying ompR in trans on a plasmid ( Fig 2A and 2B ) . This result provides the first evidence for a role of OmpR as a regulator of acid stress in Salmonella by regulating genes that alter intracellular pH . In addition , the envZ null strain did not respond to acid stress , but acidification was restored upon supplying the cytoplasmic domain of envZ ( envZc ) in trans ( Fig 3A and 3B ) . The requirement of the cytoplasmic domain of the EnvZ kinase only ( EnvZc ) further indicates that the I-switch is in the cytoplasm and the cytoplasmic pH becomes acidic under these conditions . Acidification was not dependent on growth phase , but was similar at both exponential and stationary phase ( S3 Fig ) . To further ensure that the I-switch was a high-fidelity reporter of intracellular pH , we compared it to the fluorescent probe BCECF-AM , which has been widely used to measure intracellular pH in mammalian cells [25 , 26] , plant cells [27] , yeast [28 , 29] , and bacteria [23 , 30 , 31] . It allows ratiometric detection ( S4A and S4B Fig ) and is permeable to bacterial membranes . Once it gains access to the intracellular compartment , it is cleaved by intracellular esterases to create the active fluorescent form . Although the range of BCECF-AM ( pH 6–7 ) is more limited than the I-switch ( pH 5 . 1–7 . 1 ) , a similar acidification of the Salmonella cytoplasm was observed in response to extracellular acid stress ( S4C Fig ) , corroborating the use of the I-switch in reporting the pH of the Salmonella cytoplasm . Our observation that BCECF ( which is intracellular ) and the I-switch reported similar intracellular pH values corroborates the intracellular location of the I-switch , and the I-switch was also clearly visible in spheroplasts ( S1B Fig ) . Because Salmonella acidified its cytoplasm in response to external acid stress ( Figs 2 and 3 ) , we reasoned that uptake of Salmonella into the acidified SCV should also lead to a decrease in intracellular pH . This internal acidification would in turn provide a signal to EnvZ/OmpR to induce the SPI-2 type III secretion system by activating SsrA/B [16 , 18] , which is critical for the survival of Salmonella within the phagocyte . To measure the intracellular pH of Salmonella during infection in RAW264 . 7 macrophages , we used confocal microscopy to directly visualize I-switch containing Salmonella inside macrophages . At each time point , the D/A ratio of individual cells was determined as a measure of intracellular pH ( Fig 4 ) . Salmonella exhibited a remarkable , rapid intracellular pH drop ( 6 . 85 to 5 . 75 ) immediately post-infection and reached a plateau value of 5 . 65 after 3 h . Interestingly , the Salmonella ompR null mutant did not experience a decrease in intracellular pH upon entry into the SCV ( Fig 4 ) . Again , the dependence of ompR on cytoplasmic acidification could be fully complemented by supplying ompR in trans ( Fig 4 ) . This result identifies OmpR as a key regulator of Salmonella intracellular pH , not only for in vitro acid stress but also during macrophage infection . The I-switch containing bacteria in the vacuole were viable because they were capable of SPI-2 secretion ( see below ) . An alternative method employed by many Salmonella researchers for examining intracellular bacteria involves Triton solubilization at various time points post-infection to recover bacteria from macrophages . Although this method is less rigorous than direct visualization by confocal microscopy , the results obtained using this approach were in agreement with the microscopy ( S5 Fig ) . The pHi of cells recovered after solubilization was determined from the reference intracellular standard curve . The pHi of Salmonella rapidly dropped by nearly 1 unit ( from 6 . 81 to 5 . 87 ) immediately within 30 min post-infection and reached a final value of 5 . 6 at 3 h post-infection . Bacteria recovered from the vacuole were capable of growth on agar plates , further indicating that they were still viable . Taken together , both approaches firmly establish that uptake of Salmonella into the SCV leads to an OmpR-driven immediate drop in intracellular pH . This result has significant ramifications in interpreting the pH control of the gating of the SPI-2 type III secretion apparatus [20] ( see Discussion ) . Phagosomal pH is maintained between 4 to 5 by the vacuolar H+-ATPase ( V-ATPase ) and remains acidified for at least 5–6 h post-infection [9 , 32] . Treatment with the specific inhibitor Bafilomycin A1 ( BAF ) prevents SCV acidification [1] . To examine the contribution of vacuolar SCV acidification to the low intracellular pH of Salmonella during macrophage infection , we used BAF , which is known to raise the pH of endosomes to neutral in both HeLa [33] and RAW264 . 7 cells [19] . In the presence of BAF , the Salmonella cytoplasm remained approximately pH 6 . 7 throughout the infection and failed to acidify ( Fig 4 ) . The pHi in the presence of BAF was similar to the ompR null mutant . Our results with BAF clearly indicate that the acidic pH of the SCV leads to a decrease in intracellular pH of Salmonella in macrophages ( Fig 4 ) . Pre-treatment with and maintenance of 25 nM BAF throughout infection completely abrogated the intracellular pH decrease in Salmonella-infected macrophages . This effect was rapidly reversible [34] , as removal of BAF after 1 h restored the ability of Salmonella to reduce intracellular pH , suggesting a complete correlation between vacuolar acidification and an intracellular pH drop ( S6 Fig ) . Together , these data provide strong evidence that the cytoplasm of Salmonella rapidly follows the pH of the macrophage vacuole and requires OmpR ( Fig 4 ) . Because OmpR regulates the SsrA/B two-component system that , in turn , activates the type III secretion system encoded by SPI-2 [16–18 , 35 , 36] , we wanted to determine whether or not SPI-2 was the target that was essential for cytoplasmic acidification . We performed a similar analysis on a mutant Salmonella strain that does not make a type III secretory apparatus . An ssaC null strain lacks the gene encoding the outer membrane ring protein SsaC of the SPI-2 type III secretion system , and thus is not capable of type III secretion . The ssaC null strain was fully capable of cytoplasmic acidification , as shown in ( S7A and S7B Fig ) . This was not surprising , because we observed a similar OmpR-dependent response to external acid in E . coli , which lacks SPI-2 ( unpublished results ) . OmpR is best characterized for its regulation of the outer membrane proteins OmpF and OmpC [12] . An ompC mutant and the double ompC/ompF mutant lacking the genes encoding the outer membrane proteins OmpC and OmpF were similarly capable of acidification ( S7A and S7B Fig ) . A previous study identified MgtC as an inhibitor of the FoF1 ATPase , leading to acidification [37] , but an mgtC null strain was as acidified as the WT ( S7A and 7B Fig ) . Thus , OmpR regulates an unidentified target ( s ) that is responsible for cytoplasmic acidification . In order to identify genes that were regulated by OmpR , we performed a microarray analysis and compared the transcriptome of the ompR null and the WT strain under both acid and osmotic stress . We identified genes belonging to the cadC/BA operon to be highly up-regulated in the ompR mutant . We validated the microarray results by investigating the expression of the cadC/BA operon under acid stress by quantitative real-time PCR . Transcripts for genes cadC ( 2 . 3-fold ) , cadB ( 3 . 2-fold ) and cadA ( 4 . 9-fold ) were all up-regulated in the ompR mutant , suggesting that OmpR functions as a repressor at this locus ( Fig 5A ) . Known targets of OmpR activation , including ssrA and ompR ( ompB ) , were highly down-regulated in the absence of ompR ( Fig 5A ) . The ssrA transcript was down-regulated in the envZ null strain; it was restored to the WT levels upon supplying envZc in trans . This result further demonstrates that OmpR activation is via EnvZc ( Fig 5B ) . The envZ null strain showed higher expression of cadA ( 1 . 2-fold ) , cadB ( 1 . 4-fold ) , and cadC ( 5 . 9-fold ) and failed to acidify its cytoplasm upon acid stress ( Fig 3 ) . To examine whether OmpR repression was a result of a direct interaction , we performed electrophoretic mobility shift assays ( Fig 5C–5F ) . DNA sequences upstream of cadC ( -354 to -3 ) , upstream of cadB ( -362 to -1 ) and 264-bp upstream to 34-bp downstream of cadA relative to the ATG in each case were amplified using primers that were biotinylated on the 5′-ends . The resulting biotinylated DNA fragment was used in the assay with purified OmpR or OmpR~P prepared by phosphorylation from acetyl phosphate ( Fig 5C ) [38] . The formation of an OmpR~P-DNA complex is clearly visible in the presence of 0 . 3 μM OmpR~P at both cadB and cadC promoters ( Fig 5D and 5E ) . The addition of 50-fold excess of unlabeled cadB/C DNA resulted in the release of the labeled probes ( Fig 5D and 5E , OmpR~P* ) , confirming that the complexes formed were due to a specific interaction of OmpR~P with DNA . Neither OmpR nor OmpR~P bound to a 60 bp biotin end-labeled Epstein-Barr Nuclear Antigen ( EBNA ) DNA , a control for non-specific binding ( Fig 5F ) . In the absence of phosphorylation , OmpR bound to DNA ( Fig 5D and 5E ) , but with considerably lower affinity [38] . Interestingly , no binding was evident at cadA ( Fig 5C ) , suggesting cadB and cadA may be co-transcribed . Thus , OmpR represses cadC/BA by directly binding to upstream DNA at cadC and cadB . We next measured the intracellular pH of the cadBA null strain in the macrophage vacuole and compared it to the WT and ompR null strains ( Fig 5G and 5H ) . If OmpR functions as a repressor of cadC/BA ( Fig 5A ) , we would expect the cadBA null strain to be similar to the WT strain in terms of cytoplasmic acidification , i . e . , knocking out cadBA would be the same as OmpR repressing cadBA . Indeed , the pHi of the cadBA null strain was similar to WT Salmonella and unlike the bafilomycin-treated strain or the ompR null strain ( see Figs 4 , 5G and 5H ) . This result identifies the cadC/BA operon as the major system that eliminates protons upon acid stress and maintains pH homeostasis when Salmonella is in the macrophage vacuole . OmpR represses the cadC/BA system to prevent recovery from acidification . When the cadBA genes are not present , the presence of OmpR is not required for acidification ( Fig 5G and 5H ) , indicating a key role for OmpR in repressing the cad system in the vacuole . This acidification is required for activation of SPI-2 effector expression and secretion ( see below ) . Over-expression of cadBA was able to completely over-ride the OmpR-dependent acidification , and the Salmonella cytoplasm remained neutral both during macrophage infection and in response to in vitro acid stress ( Fig 5G and 5H; S8 Fig ) . The pH was similar to an ompR null strain ( Fig 2B ) or to BAF-treated cells ( Fig 4 ) , clearly confirming that cadBA enables the ompR mutant to restore pH homeostasis in the macrophage vacuole . The cadBA null strain acidified more rapidly than the WT strain , presumably because it takes some time for OmpR to repress the cad system ( S8 Fig ) . However , the intracellular pH was nearly identical to the WT strain by 120 min . The F1Fo ATP synthase employs the proton gradient to synthesize ATP [39] . Because the Salmonella cytoplasm was acidified when grown at pHe 5 . 6 , we anticipated cytosolic ATP levels would be higher than at neutral pH . Indeed , switching Salmonella from pH 7 . 2 to pH 5 . 6 increased intracellular ATP levels 1 . 8-fold compared to the ompR null strain ( S9A Fig ) . ATP was restored to WT levels when the ompR null strain was complemented with ompR supplied in trans . Control experiments with an atpB null strain or in the presence of the uncoupler carbonyl cyanide m-chlorophenyl hydrazone ( CCCP ) drastically reduced intracellular ATP . In our microarray , ATP synthase genes were down-regulated in the absence of ompR by 2- to 3-fold . We then compared the mRNA levels of atpB , a gene encoding the a Fo subunit , by qRT-PCR in the WT and ompR null strains . The transcript for atpB was up-regulated ( 2 . 1-fold ) in the WT strain , suggesting a role for OmpR in activation of atpB ( S9B Fig ) . Thus , OmpR appears to play a dual role in acid stress , by activating proton translocation via increased F1Fo ATP synthase coupled with elimination of proton consumption via repression of the cad system . Taken together , these combined effects enable WT Salmonella to reduce intracellular pH upon acid stress . We next investigated the kinetics of translocon secretion by immuno-labeling SseB in macrophages at various times post-infection . Salmonella were visualized using anti-LPS antibody up to 6 h . SseB was detectible immediately 30 min post-infection at one locus of the cell , and the number of SseB-labeled cells increased over time . After 3 . 5 h , SseB was no longer associated with a majority of Salmonella cells , but was removed from the cell surface ( Figs 6A and 7 ) . SseB was also detected in Salmonella-infected macrophages treated with BAF , but SseB remained on the bacterial cell surface ( Figs 6B and 7A ) . To measure the extent to which SseB protein was separated from the bacterial cell surface , the Distance Transformation function of Imaris Software ( version 7 . 7 . 1 ) was employed . The distance of the red SseB translocon protein to the nearest green Salmonella cell was measured by applying the distance transformation on the green channel surface to generate a new channel whose intensity indicated the shortest distance to the surface object border . Hence , for each red SseB protein spot , the minimum distance to its nearest green cell was obtained . For each time point post-infection , the distance of 50 SseB protein spots from the nearest cells were measured . A distance value of 0 indicated that SseB was attached to the cell surface , whereas a distance value larger than 0 indicated that SseB was separated from the surface . The percentage of distant SseB was plotted over time post-infection ( Fig 7A ) . Using this approach , it was evident that the vast majority of SseB ( >60% ) were no longer surface-associated by 3 . 5 h , and by 6 h post-infection this value increased to 88% ( Fig 7A ) . When acidification was prevented by treatment with BAF , the number of cells in which SseB was detached was only 5%–10% . Measurements of the distance ( in μm ) of SseB from the bacterial cell surface as a function of time post-infection are shown in ( Fig 7B ) . At 6 h post-infection , the average distance from the Salmonella cell surface was approximately 3 μm , indicating that SseB was no longer cell-surface associated . Salmonella translocated effectors ( SifA , SseF , SseG , PipB , and SpiC ) localize to the SCV and to its tubular membranous structures termed Salmonella-induced filaments ( Sifs ) [40–42] . The Sifs extend from and connect SCVs during Salmonella replication inside host cells [43] . Therefore it was possible that once SseB was separated from the cell surface , it might be in association with endosomal tubules of vacuolar membrane origin . To address this question , we used immunofluorescence to compare the localization of SseB with the endosomal membrane marker , LAMP-1 . RAW macrophages were infected with Salmonella WT harboring pFPV25 . 1 for constitutive expression of GFP and immunolabeled for LAMP-1 ( blue ) and SseB ( red ) . Confocal microscopic images indicated that when SseB was separated from the Salmonella surface , it was co-localized with LAMP-1 positive membranes , resulting in magenta staining of SseB ( Fig 8A ) . Nearly 90% of the translocon protein SseB was co-localized with LAMP-1 by 3 . 5 h and this level did not change up to 6 h post-infection ( Fig 8B ) . In the macrophage vacuole , a single focus of SseB was evident on the bacterial surface ( Fig 6 ) . This was similar to the localization of SseB that we observed in vitro ( S10A Fig , bottom panel ) . Other components of the T3SS exhibited similar patterns of localization , including the outer membrane ring protein SsaC ( top panel ) and the inner membrane ring protein SsaJ ( middle panel ) . Thus , the SPI-2 needle complex is localized primarily near to one end of the cell pole , unlike SPI-1 needles , which are distributed all over the bacterial surface [6 , 44] . An ssaC null strain was deficient for secretion and SseB was retained in the cytoplasm , substantiating that the focal assembly of the SPI-2 needle complex was competent for SPI-2 effector secretion ( S10B Fig ) . Our observation that only approximately 13% of the population expressed SPI-2 needles on their surface in vitro highlights that , in contrast to SPI-1 , SPI-2 is not abundant . To gain insight into the fate of effector secretion , we also compared the kinetics of SseJ secretion in Salmonella-infected macrophages by immunofluorescence microscopy . A strain expressing SseJ-HA was completely comparable to WT in terms of its intracellular pH response and replication in macrophages . SseJ-HA secretion was first detected 3 . 5 h post-infection in macrophages , similar to a time course of SseF-HA expression in HeLa cells reported previously [20] . The percentage of cells secreting SseJ increased between 3 . 5 and 6 h post-infection ( Fig 9 , visible as red puncta in the macrophage cytoplasm ) . In contrast to SseB secretion ( Fig 6A ) , SseJ was not translocated in the presence of BAF , indicating that cytoplasmic acidification was essential for effector secretion ( Fig 9B , lower panel; i . e . , no red puncta are visible in the BAF-treated macrophages ) . SseJ secretion was subsequent to SseB release from the bacterial surface ( Fig 6A and 6B ) . These results are in agreement with previous results [19] but significantly differ from other studies that reported that a neutralization step from the host was required for SPI-2 effector secretion [20] . Our results indicate that formation of the translocon pore is essential and precedes effector secretion . To verify that the translocon complex was required for effector secretion , we monitored the kinetics of SseJ secretion in an sseB null strain expressing SseJ-HA . No translocation of SseJ was evident , even after 6 h post-infection in the macrophages in the absence of SseB ( S11 Fig ) .
Nucleic acids have been employed as remarkable tools to create synthetic nanomachines like DNA tweezers , walkers , and DNA-hybridization motors [45 , 46] . We describe herein a successful application of an artificially designed nanomachine in Salmonella to determine pH values under relevant physiological conditions that were previously unknown . The I-switch was effectively taken up by Salmonella after electroporation with ≥80% transformation efficiency ( S12 Fig ) . However , at later time points , the limitation of the I-switch was apparent upon bacterial cell division , which resulted in dilution of I-switch positive cells ( S12 Fig ) . In the macrophage vacuole , bacterial cell division is slower and the I-switch is effective for longer periods . Nevertheless , in the present study , using the I-switch , we measured a rapid acidification of the bacterial cytoplasm within 30 min upon entry into the SCV , followed by a further decrease to pHi 5 . 65 over 3 h of infection . Our results present evidence of cytoplasmic acidification of Salmonella inside the SCV during macrophage infection . This pH drop is a result of SCV acidification upon Salmonella infection [9 , 11 , 47] . Since the pH of the SCV remained below 5 during 6 h post-infection [9] , and in the present study we show that the Salmonella cytoplasm mirrors vacuolar pH , we argue that intracellular acidification persists during the phase of macrophage infection when the translocon and effectors are being secreted . Both enumerated bacteria from macrophages and intracellular bacteria within the SCV reported similar intracellular pH values ( Fig 4; S5 Fig ) . Our results go beyond measuring the pH of the SCV and link vacuolar pH to acidification of the Salmonella cytoplasm . Our results differ substantially from studies in E . coli in suspension that reported an initial drop in intracellular pH after extracellular acid stress , followed by a rapid recovery within minutes [48] . More recent experiments in single cells on poly ( L-lysine ) -coated coverslips reported that 2%–23% of individual cells failed to recover from pHi 5 . 5 [49] , indicating significant stochastic variation . Current experiments underway in our laboratory indicate fundamental differences between Salmonella and E . coli ( Chakraborty and Kenney , manuscript in preparation ) . Previous reports have shown that bacterial recovery times varied depending on the composition of the media and required potassium , presumably to stimulate K+-dependent proton efflux [48 , 50] . This rapid response was shown to require the cadBA operon [51] , which would lead to a rapid neutralization . However , we found no evidence of K+-dependent intracellular neutralization in Salmonella ( Chakraborty and Kenney , manuscript in preparation ) . Our experiments emphasize the role of EnvZ/OmpR in controlling intracellular Salmonella pH inside the SCV during macrophage infection . The intracellular pH of both envZ and ompR null mutants were not responsive to in vitro acid stress ( Figs 2 and 3 ) . In Escherichia coli , the OmpR/EnvZ two-component regulatory system plays a pivotal role in the modulation of gene expression in response to changes in extracellular pH [52 , 53] . The OmpR/EnvZ system is essential for Salmonella replication and survival within macrophages ( [18]; see also S13 Fig ) and to render full virulence in mice [54] . The ompR mutant strain showed a replication defect in macrophages , similar to that of SPI-2 deficient ssaC null strain ( S13 Fig ) . To verify that intracellular acidification was essential for replication and survival within macrophages , we infected RAW264 . 7 macrophages with the cadBA over-expressed strain of Salmonella and monitored its ability to survive and replicate . Indeed , the cadBA over-expressed strain showed a similar replication defect as the SPI-2–deficient ssaC null strain . This result suggests that intracellular acidification increases the replication and survival fitness of Salmonella during macrophage infection . Recently , it was shown that the inner membrane protein MgtC binds to AtpB , the a subunit of the F1Fo ATP synthase , inhibiting proton translocation and ATP synthesis [37] . In our experiments , the mgtC null strain was completely comparable to the WT in terms of its intracellular pH response ( S7 Fig ) . The mgtC null strain exhibited an increase in intracellular ATP levels ( S9A Fig ) , in agreement with previous reports [37] . Under our experimental conditions , mgtC transcripts were identical in both the WT and ompR null strains ( S9B Fig ) , eliminating a role for OmpR in regulating MgtC [55] . These results suggest that different signaling pathways ( e . g . , EnvZ/OmpR versus PhoQ/PhoP ) may play significant roles under different , unique activating conditions . Our previous structural analysis of OmpR determined the physical basis for the ability of OmpR to function as a global regulator [56] . OmpR binds to AT-rich DNA , making phosphate backbone contacts , but very few base contacts [56] . Furthermore , OmpR contacts can vary at different promoters , i . e . , DNA contacts are different at the porin genes compared to the SPI-2 ssrA gene [56] . This property of OmpR enables it to become a regulator of horizontally acquired genes during the course of evolution and to play a significant role in the response to acid stress . A previous study in E . coli reported that switching from aerobic to anaerobic respiration was important for surviving acid stress and also demonstrated a key role for OmpR in this switch , although no direct targets were identified [53] . The identified genes lacked complete OmpR binding sites , although OmpR binding sites are notoriously degenerate and difficult to define [38 , 56] . This lack of specificity makes OmpR a good global regulator , because it does not require specific amino acid/base contacts . The SPI-2 T3SS is essential for survival of Salmonella within phagocytes [1 , 3] , by avoiding the terminal stages of the degradative pathway [57] . Acidic pH and other environmental factors play an important role in the induction of SPI-2 genes [18 , 35 , 47 , 58 , 59] . Co-localization of Salmonella with the lysosome-associated membrane protein 1 ( LAMP-1 ) , clearly indicated that I-switch-incorporated Salmonella were in the SCV within 30 min after infection in macrophages and remained so during 6 h of macrophage infection ( S14 Fig ) . Thus , the acidified cytoplasm of Salmonella is a result of residing in the acidified SCV and is one trigger of the onset of virulence gene expression . This view is substantiated by our results with BAF ( Fig 4; S6 Fig ) , which demonstrated that blocking SCV acidification inhibited SPI-2 secretion . SseB , SseC , and SseD form the translocon complex; SseC and SseD are membrane proteins , while SseB is soluble . Initially , SseB is on the bacterial cell surface inside the vacuole . Over time ( approximately 3 . 5 h in our experiments ) , SseB moves further away from the Salmonella surface and becomes associated with LAMP-1 positive host membranes ( Figs 6–9 ) . This event coincides with pore-opening for effector secretion and might be dependent on vacuolar membrane contact . It is logical to assume that this separation occurs as the SsaG needle elongates , emerging from the SsaC ring in the outer membrane ( Fig 10A , 10B , 10E and 10F ) . In the absence of acidification , SseB was not released from the bacterial surface and effector secretion did not occur . Neither SseC nor SseD were detected in RAW264 . 7 cells infected with an sseB null mutant Salmonella [19] , confirming an important role of SseB in forming the translocon complex . Furthermore , secretion of SseJ was not observed in RAW264 . 7 cells infected with an sseB null Salmonella strain , indicating that translocon pore formation is also essential for effector secretion . The observation that effector secretion commences approximately 3 . 5 h post-infection suggests that this is the time required for formation of the translocon pore spanning the vacuolar membrane . Our attempts to tag the needle filament protein SsaG have thus far been unsuccessful but are a focus of current research efforts . Alternately , the translocon complex may detach from the SPI-2 needle , remaining in the vacuolar membrane , and providing a pore for SPI-2 effectors ( Fig 10A–10D ) . Acidification leads to OmpR-P repression of the cadC/BA operon , resulting in cytoplasmic acidification of WT Salmonella ( Fig 10G ) . In the ompR null strain , up-regulated CadC/BA consumes intracellular protons maintaining intracellular pH homeostasis ( Fig 10H ) . Acidification was absolutely required for the secretion of effector SseJ in macrophages ( Fig 9 ) . Our results are in agreement with some previous studies [19] , but differ significantly from others [20] , in that we never observe a neutralization step in the Salmonella cytoplasm prior to effector secretion . In summary , we used a novel biosensor and determined that the Salmonella cytoplasm acidifies immediately upon entry into the macrophage vacuole . This acidification was dependent on OmpR repression of the cadC/BA operon , and drives translocon release and subsequent effector secretion .
RAW 264 . 7 cells were obtained from the American Type Culture Collection ( ATCC , Manassas , VA , United States ) and grown in a humidified 37°C , 5% CO2 tissue culture incubator . Cell culture reagents were from Life Technologies ( Carlsbad , CA , US ) unless otherwise stated . RAW 264 . 7 were maintained in Dulbecco’s Minimal Essential Media ( DMEM ) containing 2 mM L-glutamine , 1 mM sodium pyruvate and 17 . 5 mM D-glucose with 10% heat-inactivated fetal bovine serum ( FBS ) . RAW 264 . 7 cells were always cultured in antibiotic-free media . Low passage number cells ( <15 after receipt from ATCC ) were seeded in 24-well dishes ( Nunclon ) and grown O/N before infection . The cells were washed gently with PBS and removed from the flask by scraping . The cells were suspended at a density of 1 × 105 cells per well in 24-well plates . WT S . enterica serovar Typhimurium 14028s and the ompR null derivative ( ΔompR ) were used for all infection experiments unless otherwise indicated . DMSO was obtained from Sigma-Aldrich ( St Louis , MO ) . To determine the acid stress response , bacterial strains were grown as in [60] in a modified N-minimal medium ( MgM ) buffered with 20 mM Tris ( pH 7 . 2 ) or 20 mM MES ( pH 5 . 6 ) containing 7 . 5 mM ( NH4 ) 2SO4 , 5 mM KCl , 0 . 5 mM K2SO4 , 1 mM KH2PO4 , 10 mM MgCl2 , 2 mM glucose , and 0 . 1% Casamino acids . The following antibodies were used for Immuno-labeling experiments: rabbit polyclonal anti-SseB , rabbit monoclonal anti-HA ( Sigma ) , mouse monoclonal anti-Salmonella LPS ( Abcam ) , rabbit anti-LAMP-1 ( Abcam ) , mouse Alexa-488 IgG ( Invitrogen ) , rabbit Alexa-488 IgG ( Invitrogen ) , and rabbit cy5 IgG ( Invitrogen ) . I-switch sample preparation was performed as previously described [14] with the following modifications: 50 μM O1 , O2 , and O3 were mixed in equimolar ratios in 20 mM potassium phosphate buffer at the desired pH containing 100 mM KCl . Primers were annealed using touch-down PCR with a decrease in temperature of 1°C per 3 min from 95°C to 25°C and equilibrated overnight at 4°C before use . A concentration of 6 μM of I-switch DNA was used to transform electro-competent Salmonella cells . After electroporation , cells recovered for 1 h at 37°C in Super Optimal Broth ( SOB ) containing 20 mM glucose and 10 mM MgCl2 ( SOC ) unless otherwise stated . Cells were washed three times with PBS ( Life technologies ) and used for analysis . pH clamping protocols were modified based on [14] . Double-labelled I-switch ( IA488/A647 ) was diluted to 60 nM in clamping buffers ( 120 mM KCl , 5 mM NaCl , 1 mM MgCl , 1 mM CaCl , 20 mM HEPES buffered at various pH ) . The samples were excited at 488 nm and emission collected between 500–750 nm in a Tecan Spectrophotometer ( GENios ) . The fluorescence intensity at 520 nm ( D ) was divided by the fluorescence intensity at 666 nm ( A ) and the D/A ratio was plotted over the pH range . For in vivo pH clamping experiments , bacterial cells electroporated with the I-switch were incubated in a K+-rich clamping buffer at various pH at RT for 1 h in the presence of 40 μM nigericin . Cells were placed on microscope slides ( Marienfeld , Germany ) coated with 1% agarose and then imaged by wide field fluorescence microscopy . Three independent measurements were recorded for each pH value and the D/A ratios were plotted as the mean ± SEM . Salmonella were first electroporated with the I-switch as described . Cells were seeded onto overnight-grown monolayers of RAW264 . 7 cells at a multiplicity of infection of 100:1 in 24-well plates . The plates were centrifuged at 500 rpm for 10 min to synchronize the infection . The infection was conducted for 30 min at 37°C in 5% CO2 . Cells were washed three times with PBS and incubated with DMEM containing 100 μg/ml of gentamicin for 1 h at 37°C . Cells were again washed three times with PBS and incubated in DMEM containing 10 μg/ml gentamicin for the remainder of the experiment . For BAF-treated cultures , RAW264 . 7 cells were pre-treated with 25 nM BAF for 30 min prior to infection and 25 nM BAF was maintained in every step of the infection . At the stated time points , the cells were fixed using 3% PFA for 15 min at RT . The coverslips were mounted on a drop of gold anti-fade reagent ( Invitrogen ) and sealed with entellan ( Merck ) . Samples were imaged using confocal microscopy ( Nikon A1R ) and analyzed by Image J imaging software . To inhibit vacuolar acidification , 25 nM BAF or an equal volume of DMSO were added to the macrophage cells 30 min prior to infection and maintained throughout the infection process . All the wide-field images were collected using an Olympus IX71 Inverted Microscope ( Applied Precision DeltaVision Deconvolution microscope system ) equipped with 100X , 1 . 4 numerical aperture ( NA ) objective lens with a mercury arc bulb ( OLYMPUS U-RFL-T: Mercury 150W ) as the illumination source . Image capture was performed with the CoolSnap HQ , a fast , high resolution , high quantum efficiency , cooled CCD camera ( Photometrics CoolSNAP HQ2 ( CCD ) ; 1 , 392 x 1 , 040 pixels; 11 fps ) . Two sets of images were taken corresponding to ( i ) donor emission wavelength ( 525/36 nm ) upon donor excitation ( 490 ± 20 nm ) ( donor image: D ) , and ( ii ) acceptor emission wavelength ( 666 nm long pass ) ( acceptor FRET: A ) upon donor excitation ( 490/20 nm ) . Acceptor images ( I ) acceptor emission wavelength ( 666 nm long pass ) upon acceptor excitation ( 645 ± 20 nm ) were obtained for cells clamped at pH 5 . 2 and 7 . 1 . Confocal imaging was carried out on an inverted Nikon A1Rsi-Polarizing Module equipped with CFI Plan ApochromatVC 100X , 1 . 4 numerical aperture ( NA ) objective lens by using a Coherent CUBE laser at 488 excitation with set appropriate dichroics . Excitation at 470/40 nm wavelength was performed through a 510 nm dichroic mirror , and emission was recorded through a 525/50 nm band pass filter on a Andor DU897 EMCCD camera to yield the donor image ( D ) . Similarly , emission was recorded through a 700/75 nm band pass filter upon similar excitation to generate the FRET intensity ( A ) . Multiple Z sections were taken ( 0 . 5 μm apart ) and each image is represented as sum slices . Images were analyzed using ImageJ ver . 1 . 42 . All of the images were background-subtracted by taking the mean intensity over an adjacent cell-free area . Autofluorescence of unlabeled Salmonella in RAW 264 . 7 cells was measured and used to correct the determined pH values . However , there was no difference in the intracellular pH values when we corrected for auto-fluorescence compared with the uncorrected values . Because the measurement is ratiometric , auto-fluorescence did not influence the determined pH values . Images showing the D/A values were obtained by using the ratio plus plugin of ImageJ software . The high and low values were color-coded and calibrated to their respective pH values as described in the main text . For all studies in macrophages , the mean D/A intensity of at least thirty bacterial cells was measured and is represented as the mean ± SEM of three independent batches of experiments . The free acid form of BCECF was suspended in buffers ranging from pH 5 . 0 to 7 . 2 . Fluorescence intensities were measured at excitation wavelengths of 488 and 440 nm and the emission wavelength was recorded at 525 nm in a Tecan spectrophotometer ( GENios ) . This generated the in vitro standard curve . Overnight cultures of WT and the ompR null mutant of Salmonella were clamped in 100 mM potassium phosphate buffer at various pH containing 40 μM nigericin and incubated in the presence of 20 μM BCECF-AM . Paired images were obtained by using excitation wavelengths of either 440 nm ( pH-insensitive wavelength ) or 488 nm ( pH-sensitive wavelength ) . Fluorescence emission was recorded through a 525/50-nm band pass filter on a Andor DU897 EMCCD camera on an inverted Nikon A1Rsi-Polarizing Module by using Coherent CUBE lasers at 488 and 440 excitation with set appropriate dichroics . The ratios were plotted as a function of pH , and are comparable to the in vitro calibration curve . WT and the ompR null mutant of Salmonella were incubated at either acidic pHe ( 5 . 6 ) or neutral pHe ( 7 . 2 ) for 2 h . 20 μM BCECF-AM was added to the cultures for 30 min before imaging . To determine whether acidification was growth phase dependent , cells were grown until O . D . = 0 . 56 ( log phase ) or O . D . ≈ 2 . 3 ( stationary phase ) before the addition of the probe . The ratios of the fluorescence intensities of emission channel ( 525 nm ) upon 488 nm excitation and 440 nm excitation were obtained as before . Cells were placed on microscope slides ( Marienfeld Germany ) coated with 1% agarose and then imaged . Images were first corrected for background fluorescence by subtracting an image of a cell-free area and analyzed by Image J imaging software . I-switch single labeled with Alexa 647 ( IA647 ) was transformed in Salmonella as described above . Spheroplasts were prepared as described in [61] with the following modifications; Cells were resuspended after recovery in cold hypertonic solution A ( 0 . 75M sucrose 10 mM TrisHCl pH 7 . 8 ) . 5 mg/ml of lysozyme solution was gently added to the cell suspension and the mixture was kept on ice for 5 min . EDTA solution ( 1 . 5 mM EDTA pH 7 . 5 ) was then added slowly over 10 min . Cells were centrifuged at 4 , 000 rpm for 20 min and re-suspended in solution D ( 0 . 25M sucrose 3 mM Tris-HCl pH 7 . 8 , 1 mM EDTA pH 7 . 5 ) . Cells were imaged on Applied Precision DeltaVision Deconvolution microscope system . The sseJ-HA fusion or ompR DNA sequences were amplified using primers XhoI-PsseJ #F and sseJ-HA-BamHI #R or KpnI-PompR #F and envZ-HindIII #R from Salmonella 14028s genomic DNA . The PCR products and low copy promoter-less pWSK29 vector were double-digested with restriction enzymes XhoI and BamHI or KpnI and HindIII ( Thermo Fisher Scientific ) , and the digested samples were ligated with the Rapid DNA Ligation Kit ( Thermo Fisher Scientific ) . The cadA , cadB , cadC , ompC , ompF , mgtC , atpB , and sseB genes were deleted from the Salmonella chromosome using λ-Red recombination techniques as described previously [62 , 63] . A cadBA over-expressed strain was generated by cloning cadBA into plasmid pBR322 , replacing the bla gene with cadBA . SsaC-GFP and SsaJ-mCherry were constructed as follows: plasmid pWSK29 containing ssaC with the spiC promoter fused before GFP and ssaJ with the sseA promoter fused before mCherry were inserted in ssaC and ssaJ null strains , respectively . Primers are listed in S1 Table . A list of strains and plasmids are provided in S2 Table . For every time point post-infection , cells were fixed with 3% paraformaldehyde for 10 min at RT followed by three PBS washes . Cells were probed with various primary and secondary antibodies in PBS containing 2% BSA and 0 . 1% saponin . Antibody staining was done sequentially for 1 h , and after each incubation , washes were performed five times with PBS . The HA epitope tag was detected with the monoclonal HA antibody at 1:500 ( Covance ) ; LAMP-1 was detected with fluorescein isothiocyanate ( FITC ) -conjugated rabbit monoclonal antibody at 1:500 ( Research Diagnostics ) ; Salmonella was detected with mouse polyclonal anti-lipopolysaccharide ( anti-LPS ) antibodies at 1:500 ( Difco ) . Secondary antibody detection was performed with various anti-rabbit or anti-mouse antibodies conjugated to Alexa-488 and cy5 . To determine intracellular replication within macrophages , a similar protocol was followed as reported in [64] with slight modification . Salmonella was electroporated with non-labeled I-switch , and used for infection ( see text for details ) . At 2 and 16 h post-infection , cells were washed repeatedly in PBS and lysed in 0 . 1% Triton-X100 at RT for 15min . Samples were then serially diluted and plated on LB to enumerate the intracellular bacteria . Replication was described as the CFU/ml counted and represented as an average of three plates for all time points . Electrophoretic mobility shift assays ( EMSAs ) were performed using the Lightshift chemi-luminescence EMSA kit ( Pierce ) according to the manufacturer’s instructions described in [65] . The upstream regions of cadC ( 354 bp ) , cadB ( 361 bp ) , and cadA ( 264 bp upstream to 34 bp downstream ) were amplified using biotinylated oligonucleotides . Ten fmol of biotinylated DNA was used in a 15 μl reaction containing binding buffer ( 10 mM Tris , pH 7 . 5 , 50 mM KCl ) along with 2 . 5% glycerol , 5 mM MgCl2 , 0 . 05% Nonidet P-40 , and 1μg poly ( dI-dC ) ) . OmpR or OmpR~P protein was added at the concentrations indicated , and samples were separated by electrophoresis on 5% non-denaturing acrylamide gels run in 0 . 5X Tris-acetate buffer with EDTA . Following electrophoresis , DNA was electro-transferred to a nylon membrane and detected using the biotin detection system ( Pierce ) . Salmonella strains were grown in SPI-2 inducing Minimal Magnesium Medium ( pH 5 . 6 ) as described previously [60] to O . D ~0 . 6 . The total RNA was isolated using an RNeasy mini kit ( Qiagen ) . After DNase treatment of the isolated RNA , cDNA was synthesized using the iScript Reverse Transcription Supermix ( Biorad ) . Quantification of cDNA was carried out using SsoFast TM Eva Green Supermix ( Bio-Rad ) , and real-time amplification of the PCR products was performed using the iCycler iQ real-time detection system ( Bio-Rad ) . The mRNA expression level of the target gene was normalized relative to the 16S rRNA expression level . Cultures were grown in LPM media ( pH 7 . 0 ) to O . D ~1 . 5 as described in [66] with slight modifications . Unadapted cultures were immediately shifted to pH 3 . 0 acidified by HCl . Adapted cultures were shifted to pH 4 . 5 for 2 h prior to the challenge at pH 3 . 0 . Viable cells were calculated by plating aliquots of serially diluted cultures by standard plate count . Intracellular ATP levels using Tecan spectrophotometer as described with modifications [37]: Overnight cultures of Salmonella in LB were sub-cultured in MgM ( 7 . 2 ) for 24 h . The cultures were washed in MgM ( 5 . 6 ) and inoculated in 5 ml of the same media for 5 h . Cells were normalized by the OD600 and re-suspended in 500 μl of ( PBS ) . Nucleic acids were extracted by adding ice cold 1% Trichloroacetic acid ( TCA ) and 2 mM EDTA for 15 min . The extracts were neutralized with 100 μl of neutralization buffer and centrifuged for 15 min . The supernatant was diluted 2-fold with L buffer and the luciferase reaction was initiated using an ATP determination kit ( Invitrogen ) according to the manufacturer's instructions . Intracellular ATP levels ( μM ) were converted using reference to standards of known concentration . Overnight cultures of Salmonella WT and ssaC null mutant strains in LB were sub-cultured in MgM ( 7 . 2 ) for 24 h . The cultures were washed either in MgM ( 5 . 6 ) or MgM 7 . 2 and inoculated in 50 ml of the same media for 5 h . For isolation of the secreted protein , cells were removed from the culture by centrifugation ( 5 , 500 x g , 20 min , 4°C ) , and the supernatant was filtered through a 0 . 22 μm pore size binding filter ( Millipore , Billerica , MA ) . The secreted fraction was isolated by 10% trichloroacetic acid precipitation , and the protein pellet was washed three times with −20°C acetone and then air dried . Equivalent amounts of cellular protein , adjusted according to the optical densities , were used for the whole cell fraction . The protein samples were separated by 12% SDS-PAGE and transferred to PVDF membrane ( Millipore ) . The membrane was incubated with anti-SseB ( 1:10 , 000 ) or anti-GroEL ( 1:5 , 000 ) antibodies in PBS buffer ( with 0 . 05% Tween 20 and 1% Skim Milk ) followed by anti-rabbit secondary antibody ( 1:5 , 000 , Santa Cruz Biotechnology ) .
|
The human pathogenic bacteria Salmonella encounters extreme and diverse conditions during the course of host infection . Survival and adaptation inside the host requires highly regulated virulence factors . When Salmonella is engulfed by a macrophage , it forms a vacuole-type structure that is actively acidified by the macrophage in an attempt to kill or neutralize the bacteria . However , the acidic pH in this “Salmonella-containing vacuole , ” or SCV , does not kill the bacteria and is instead a major inducing signal for the expression of a specific class of virulence proteins . These proteins , encoded in the Salmonella Pathogenicity Island 2 ( SPI-2 ) , function as a nanomachine that mediates the delivery of Salmonella effector proteins that disrupt host immune defenses . In this paper , we investigate unanswered questions regarding the ability of Salmonella to survive the low pH and its consequences for bacterial growth in the SCV . Using a fluorescent biosensor , we monitored the intracellular pH of the Salmonella cytoplasm while it resides in the SCV during macrophage infection . Our results indicate that the bacteria cytoplasm acidifies in response to SVC acidity; this acidification requires the transcription factor OmpR , a known regulator of SPI-2 . OmpR represses the cadC/BA operon , which is involved in the recovery from acid stress , thus enabling Salmonella to assume the acidic pH of the macrophage vacuole . Acidification is required for the secretion of virulence factors; blocking acidification resulted in a neutralized cytoplasm that was defective for SPI-2 secretion . Our work challenges existing views that bacteria regulate their pH to maintain neutrality , and provides a new model for Salmonella virulence factor secretion and infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
A FRET-Based DNA Biosensor Tracks OmpR-Dependent Acidification of Salmonella during Macrophage Infection
|
Polycomb group ( PcG ) proteins act as evolutionary conserved epigenetic mediators of cell identity because they repress transcriptional programs that are not required at particular developmental stages . Each tissue is likely to have a specific epigenetic profile , which acts as a blueprint for its developmental fate . A hallmark for Polycomb Repressive Complex 2 ( PRC2 ) activity is trimethylated lysine 27 on histone H3 ( H3K27me3 ) . In plants , there are distinct PRC2 complexes for vegetative and reproductive development , and it was unknown so far whether these complexes have target gene specificity . The FERTILIZATION INDEPENDENT SEED ( FIS ) PRC2 complex is specifically expressed in the endosperm and is required for its development; loss of FIS function causes endosperm hyperproliferation and seed abortion . The endosperm nourishes the embryo , similar to the physiological function of the placenta in mammals . We established the endosperm H3K27me3 profile and identified specific target genes of the FIS complex with functional roles in endosperm cellularization and chromatin architecture , implicating that distinct PRC2 complexes have a subset of specific target genes . Importantly , our study revealed that selected transposable elements and protein coding genes are specifically targeted by the FIS PcG complex in the endosperm , whereas these elements and genes are densely marked by DNA methylation in vegetative tissues , suggesting that DNA methylation prevents targeting by PcG proteins in vegetative tissues .
Polycomb group ( PcG ) proteins are evolutionary conserved master regulators of cell identity and balance the decision between cell proliferation and cell differentiation [1] . PcG proteins act in multimeric complexes that repress transcription of target genes; the best characterized complexes are the evolutionary conserved Polycomb Repressive Complex 2 ( PRC2 ) that catalyzes the trimethylation of histone H3 on lysine 27 ( H3K27me3 ) , and PRC1 , which binds to this mark and catalyzes ubiquitination of histone H2A at lysine 119 [1] . Plants contain multiple genes encoding homologs of PRC2 subunits that have different roles during vegetative and reproductive plant development [2] . Whereas the EMBRYONIC FLOWER ( EMF ) and VERNALIZATION ( VRN ) complexes control vegetative plant development , reproductive development in Arabidopsis crucially depends on the presence of the FERTILIZATION INDEPENDENT SEED ( FIS ) PcG complex that is comprised of the subunits MEDEA ( MEA ) , FERTILIZATION INDEPENDENT SEED2 ( FIS2 ) , FERTILIZATION INDEPENDENT ENDOSPERM ( FIE ) and MSI1 [2] . The FIS PcG complex is required to suppress autonomous endosperm development; loss of FIS function initiates the fertilization-independent formation of seed-like structures containing diploid endosperm [3] . In most angiosperms the endosperm is a triploid zygotic tissue that develops after fusion of the homodiploid central cell with a haploid sperm cell . The endosperm regulates nutrient transfer to the developing embryo and regular endosperm development is essential for embryo development [4] . Loss of FIS function also dramatically impacts on endosperm development after fertilization , causing endosperm overproliferation and cellularization failure , eventually leading to seed abortion [5] . Thus far , only few direct target genes of the FIS PcG complex are known , among them the MADS-box transcription factor PHERES1 ( PHE1 ) [6] , FUSCA3 [7] and MEA itself [8]–[10] . All three genes are also targets of vegetatively active PcG complexes [7] , [11] , suggesting that different PcG complexes share at least a subset of target genes [7] . Similar to extraembryonic tissues in mammals [12] , the endosperm has reduced levels of DNA methylation compared to the embryo or vegetative tissues [13] , [14] . Hypomethylation is established by transcriptional repression of the maintenance DNA-methyltransferase MET1 during female gametogenesis [15] , together with active DNA demethylation by the DNA glycosylase DEMETER ( DME ) [13] , [16] . Whereas the global DNA methylation levels differ only slightly between embryo and endosperm ( ∼6% for CG methylation ) , methylation differences at transposable elements and repeat sequences are significantly more pronounced [13] , [14] . The functional significance of this genome-wide demethylation of the endosperm is not yet understood . However , it has been proposed that DNA demethylation might cause transposon activation and generation of small interfering RNAs ( siRNA ) that might move to egg cell or embryo where siRNA-mediated DNA methylation would lead to increased methylation of parasitic genomic sequences [13] . This notion is supported by the observation of accumulating 24nt siRNAs in the female gametophyte and in the endosperm [17] . However , functional loss of RNA polymerase IV , the enzyme responsible for the biogenesis of siRNAs , does not cause reactivation of most transposons [18] , suggesting the presence of redundant pathways to silence transposable elements . In this study , we profiled the H3K27me3 pattern in the endosperm and identified many target genes that were known previously to be targeted by vegetatively active PcG complexes , supporting the idea that different PcG complexes share a common set of target genes . However , we also identified endosperm-specific H3K27me3 target genes that have functional roles in endosperm cellularization and chromatin architecture , suggesting that the FIS PcG complex has endosperm-specific functions and that PcG targeting in plants has tissue specific roles . Finally and most importantly , we discovered that the FIS PcG complex in the endosperm targets transposable elements ( TEs ) that are protected by DNA methylation in vegetative tissues , implicating that DNA methylation and H3K27me3 are alternative repressive marks that may compensate for each other in the repression of a subset of TEs .
We established a transgenic line expressing PHE1 fused to the enhanced green fluorescent protein ( EGFP ) under control of the native promoter and 3′ regulatory elements . Strong EGFP fluorescence was exclusively detected in endosperm nuclei from 1 day after pollination ( DAP ) until 4 DAP , whereas only a weak signal was detectable in the chalazal endosperm at 5 DAP ( Figure 1A ) . EGFP-labeled nuclei from 1–4 DAP-old seeds were isolated with the use of a fluorescence-activated cell sorter . High-throughput techniques allowed the harvesting , nuclei isolation , and sorting of approximately 100 000 nuclei in about 4 hours . Within this time period , endosperm nuclei did apparently not undergo substantial changes in their transcriptional identity , as judged by a relatively low expression of embryo and seed coat marker genes in relation to the PHE1 gene ( Figure 1B ) . Expression of seed coat and embryo marker genes followed a similar trend in microdissected endosperm samples ( Figure 1C ) . To identify endosperm-specific PcG target genes we performed chromatin immunoprecipitation ( ChIP ) of chromatin from sorted endosperm nuclei using H3K27me3 specific antibodies followed by hybridization to high resolution whole-genome tiling microarrays ( Chip-on-chip ) . As a control , we performed ChIP with unspecific IgG antibodies . Genomic regions marked by H3K27me3 ( “H3K27me3 regions” ) were identified as continuous runs of probes with a MAT-score of at least 3 . 5 ( see Materials and Methods ) . We identified 2282 regions that were significantly enriched for H3K27me3 , covering ∼1 . 9 Mb and representing ∼1 . 6% of the sequenced genome . This corresponds to about one fourth the number of H3K27me3 regions identified in seedling tissues [11] , [19] , indicating that there are substantially fewer H3K27me3 targets in the endosperm than in vegetative tissues . Similar to the H3K27me3 distribution in Arabidopsis seedlings [11] , most H3K27me3 regions in the endosperm were located on euchromatic chromosome arms and only 17 of the 2282 regions ( 0 . 7% ) were from centromeric or pericentromeric heterochromatin ( Figure 2A ) . The distribution of H3K27me3 in endosperm over genes had a pronounced maximum in the transcribed region , similar to the distribution of H3K27me3 in vegetative tissues ( Figure 2B , [11] ) . Notably , there was a small but distinct drop of H3K27me3 at the transcriptional start and shortly after the transcriptional stop , possibly caused by localized nucleosome depletion . This interpretation would be in agreement with previous observations made in yeast and human cells , revealing nucleosome depletion at the transcriptional start and around polyadenylation sites [20]–[22] . The length of H3K27me3 regions in the endosperm was comparable to the length of H3K27me3 regions in vegetative tissues [11] , with a median region size of about 750 bps ( Figure 2C ) . MEA , PHE1 , MEIDOS ( MEO ) and FUSCA3 ( FUS3 ) as well as other genes that were previously identified as sporophytic H3K27me3 targets were among the endosperm H3K27me3 targets ( Figure 2D and Figure 3A ) , indicating that our procedure successfully identified H3K27me3 targets in the endosperm . We identified 1773 genes to be associated with H3K27me3; of those , 1533 genes ( ∼86 . 5% ) overlapped with H3K27me3 marked loci identified in seedling tissues ( “shared H3K27me3 targets” ) [11] , [19] , whereas 240 loci ( ∼13 . 5% ) were specifically enriched only in the endosperm ( “endosperm-specific H3K27me3 targets” ) ( Figure 3A and Table S1 ) . Most H3K27me3 targets in both sample sets are protein-coding genes of known or unknown functions , similar to the H3K27me3 targets in seedling tissues [11] , [19] ( Figure 3B ) . The overall distribution of H3K27me3 marked pseudogenes and TEs in the endosperm and seedling tissues was similar; TEs and transposable element genes ( TEGs; correspond to genes encoded within a transposable element ) were clearly underrepresented among H3K27me3 targets compared to the genome average ( Figure 3B ) . However , the frequency of TEs and TEGs was much higher among the endosperm-specific H3K27me3 targets than among the shared H3K27me3 targets , indicating that a subset of TEs and TEGs are specifically marked by H3K27me3 in the endosperm ( Figure 3B ) . While 16% of all TEs and 46% of all TEGs probed by the microarray are located in centromeric and pericentromeric heterochromatin , only 5% of the TEs with H3K27me3 and 16% of the TEGs with H3K27me3 were from these heterochromatic regions . Frequencies of almost all super families of TEs were similar among H3K27me3-marked endosperm-specific TEs and among all TEs detectable by the microarray ( Figure S1 ) . Among the shared H3K27me3 targets LTR/COPIA ( p<5E-4 ) , LINE/L1 ( p<0 . 05 ) , and RathE1 elements ( p<0 . 05 ) were significantly enriched , indicating non-random targeting of TEs by PcG proteins . We verified the specificity of our analysis by qPCR validation of endosperm-specific and shared H3K27me3 targets using independently prepared ChIP samples . We randomly selected 10 endosperm-specific TEGs , 9 endosperm-specific genes and 8 shared target genes and could confirm all loci in an independent ChIP experiment ( Figure S2 ) , indicating that our procedure was specific with a low false discovery rate . Shared H3K27me3 targets in the endosperm were highly enriched for genes involved in transcriptional regulation , with MADS-box transcription factors being a prominently enriched subclass of transcription factors ( p = 3 . 01E-05; Table S2 ) . However , many other GO categories were enriched among shared H3K27m3 target genes , including regulation of metabolism , flower development , cell wall organization , secondary metabolism and others ( Table S3 ) . This indicates that the FIS PcG complex acts to repress a large set of genes that are not required during early endosperm development . Among endosperm-specific H3K27me3 targets , there were many genes with potential roles in vesicle-mediated transport and cytoskeleton organization ( Table S4 ) , suggesting a specific function of the FIS PcG complex in endosperm cellularization . Furthermore , many genes with functional roles in chromatin organization , such as the PcG protein encoding genes EMF2 , VRN2 , MSI1 , the DNA glycosylase ROS1 as well as DNA helicases were among specific H3K27me3 target genes ( Table S4 ) , implicating a role of the FIS PcG complex in establishing specific chromatin architectures in the endosperm . Next , we analyzed the relation between H3K27me3 modification and gene expression . Gene expression data were derived from the peripheral endosperm of seeds containing globular stage embryos , corresponding to the main fraction of the sorted endosperm nuclei used in our ChIP-chip experiment . Consistent with the function of H3K27me3 in transcriptional silencing , the majority of shared endosperm H3K27me3 target genes were expressed at low levels ( Figure 4A ) . In contrast , a fraction of the endosperm-specific H3K27me3 targets was moderately expressed ( Figure 4A ) . Endosperm-specific H3K27me3 target genes had lower average H3K27me3 scores compared to shared targets independent of their expression level ( Figure 4B ) , suggesting that there is different efficiency of PcG protein targeting or PRC2 activity for endosperm-specific versus shared endosperm H3K27me3 targets . Using publicly available datasets we tested the tissue-specific expression of endosperm-specific H3K27me3 target genes by cluster analysis . Consistent with the idea that the FIS PcG complex is required for repression of target genes in the endosperm , genes present in clusters I , II and V ( 45% , n = 75 ) were specifically repressed in the endosperm ( Figure 4C ) . However , about half of all endosperm-specific H3K27me3 targets were expressed in the endosperm ( clusters III and IV , 55% , n = 91; Figure 4C ) , in agreement with the higher average expression levels of endosperm-specific H3K27me3 target genes compared to non-H3K27me3 target genes ( Figure 4A ) . We consider three not mutually exclusive explanations for this observation: ( i ) H3K27me3 is not necessarily connected with gene silencing in the endosperm . ( ii ) For a subset of genes only one of the alleles is marked by H3K27me3 . In this case expression of the non-marked allele would be detected , whereas the H3K27me3 allele remains silenced , as it was shown before for PHE1 and MEA [8] , [9] , [23] , [24] . However , imprinted genes predicted by Gehring and colleagues [14] were not among genes present in clusters III and IV . ( iii ) PcG target genes are differentially regulated in the different domains of the endosperm , i . e . the micropylar , peripheral and chalazal domains ) . TEs were strongly overrepresented among the endosperm-specific H3K27me3 targets compared to the shared H3K27me3 targets ( Figure 3B ) . Hence , we hypothesized that the global DNA demethylation in the endosperm [13] , [14] caused H3K27me3 to accumulate in regions that are DNA methylated in vegetative tissues and , therefore , H3K27me3-poor . This hypothesis predicts that TEs marked by H3K27me3 in the endosperm have reduced endosperm DNA methylation levels compared to all TEs . Indeed , median endosperm CG and CHG DNA methylation levels were lower at H3K27me3 marked TEs than at other TEs ( Figure 5A ) . CHH methylation levels were generally low and did not differ between H3K27me3 marked TEs and all TEs ( data not shown ) . TEs that carried H3K27me3 in endosperm and vegetative tissues were almost devoid of CG DNA methylation in endosperm and vegetative tissues . In contrast , TEs that carried H3K27me3 only in the endosperm had high DNA methylation levels in vegetative tissues while DNA methylation levels in the endosperm were markedly below the average over all TEs . Similarly , shared TEGs were almost devoid of DNA methylation in vegetative tissues and in the endosperm . Endosperm DNA methylation levels of specific H3K27me3 TEGs were comparable to the average DNA methylation levels in the endosperm of all TEGs present in the genome ( Figure 5B ) , indicating that reduced DNA methylation levels in the endosperm might allow targeting of PcG proteins to defined sequences independent of residual DNA methylation . CHG methylation followed a similar trend as CG methylation ( Figure 5B ) . In contrast , no substantial changes in CHH methylation levels were observed ( data not shown ) . Protein coding genes were generally much less DNA methylated than TEs or TEGs . Similar to shared TEs and TEGs , shared H3K27me3 target genes were almost devoid of DNA methylation in vegetative tissues and the endosperm ( Figure 5C ) . In marked contrast , endosperm-specific H3K27me3 target genes had significantly higher CG DNA methylation levels in vegetative tissues than the genome-wide average ( Figure 5C ) , supporting the idea that CG DNA methylation prevents these genes being targeted by PcG proteins in vegetative tissues . CG DNA methylation level of endosperm-specific H3K27me3 genes was reduced in the endosperm compared to vegetative tissues , again suggesting that reduced DNA methylation levels in the endosperm enable targeting of PcG proteins to selected loci . Shared and specific protein coding H3K27me3 target genes were almost devoid of CHG and CHH methylation in vegetative tissues and the endosperm ( Figure 5C and data not shown ) . Together , we conclude that DNA methylation and H3K27me3 , which both can bring about transcriptional repression of target genes , usually exclude each other at target chromatin . In the endosperm , where DNA methylation is naturally reduced , some loci that were DNA methylated in other tissues become targeted by the FIS PcG complex and marked by H3K27me3 . This hypothesis predicts that experimental reduction of DNA methylation levels in vegetative tissues will cause PcG proteins to be targeted to some loci that are usually DNA methylated . Indeed , in met1 mutants H3K27me3 was found at some TEs that did not carry H3K27me3 in wild type [25] , strongly supporting this idea . Based on their expression in the endosperm , two main clusters of protein coding genes and TEGs that were DNA methylated in vegetative tissues and carried H3K27me3 in the endosperm were apparent ( Figure 5D ) ; the first cluster contained genes and TEGs that were weakly expressed in other tissues and became specifically repressed in the endosperm , whereas the second cluster contained genes and TEGs that were mainly repressed in other tissues and became specifically expressed in the endosperm , indicating that loss of DNA methylation fostered expression of several genes and transposons in the endosperm independent of their gain of H3K27me3 . We wondered whether loss of FIS activity would cause a global deregulation of H3K27me3 target genes . Therefore , we profiled the fis2 transcriptome of seeds harvested at 3 DAP and 6 DAP and searched for deregulated genes that were marked by H3K27me3 in the endosperm . Loss of FIS function profiled at 3 DAP and 6 DAP resulted in different and largely non-overlapping gene expression profiles ( Figure 6A ) . Although the overlap of H3K27me3 target genes and genes deregulated upon loss of FIS function was significant ( p = 3 . 0E-05 and 5 . 7E-04 for 3 DAP and 6 DAP , respectively ) , expression of surprisingly few target genes ( ∼1 . 5% and ∼1 . 8% at 3 DAP and 6 DAP , respectively ) was increased upon loss of FIS function ( Figure 6A , Table S5 ) . EMF2 and VRN2 expression was not increased in fis2 seeds at 3 or 6 DAP , indicating that loss of FIS2 function is not compensated by increased expression of FIS2 homologous genes . Genes deregulated at 3 DAP and 6 DAP fell into two largely distinct clusters . Whereas most of early deregulated genes were not expressed in the wild-type endosperm until heart stage , late deregulated genes were predominantly expressed during early wild-type endosperm development and became repressed around heart stage ( Figure 6B ) , supporting the idea that the FIS PcG complex is required for the repression of a defined set of genes around endosperm cellularization [26] , [27] . Genes deregulated in fis2 at 3 DAP and 6 DAP were prominently enriched for glycosyl hydrolases ( Table S6 ) , with a strong enrichment of Family 17 of plant glycoside hydrolases at 6 DAP . Family 17 members preferentially hydrolyse the major component of endosperm cell walls , callose , [28] , suggesting that repression of cell wall degrading enzymes is a requirement for successful endosperm cellularization . Conversely , this implicates that increased expression of these genes in fis mutants might contribute to the failure of fis mutant endosperm to undergo endosperm cellularization [29] . Importantly , we did not detect increased expression of TEGs in fis2 mutants , suggesting that loss of H3K27me3 might be compensated by other repressive mechanisms . If so , we wondered whether in seeds lacking both , FIS activity and CG DNA methylation , repression of TEGs would be relieved . Therefore , we generated fis2/FIS2; met1/MET1 double mutants that contain 12 . 5% seeds homozygous for met1 and devoid of FIS activity . We randomly selected eight endosperm-specific H3K27me3 TEGs ( At4g16870 , At5g37880 , At3g32110 , At2g13890 , At5g35710 , At1g35480 , At3g28400 , At2g16010 ) that were DNA methylated in vegetative tissues and had decreased DNA methylation levels in the endosperm ( Figure S3 ) . Among those , At4g16870 , At5g37880 had increased expression levels in fis2;met1 double mutants compared to met1 and fis2 single mutants ( Figure 6C ) , whereas expression of At3g32110 equally increased in met1 and fis2; met1 double mutants . Expression of the other TEGs was not significantly changed compared to wild type ( data not shown ) . Based on these data we conclude that DNA methylation and FIS-mediated H3K27me3 can act synergistically to repress a subset of TEGs in the endosperm , but that there are additional mechanisms to silence TEGs in the absence of both mechanisms .
PcG proteins are largely viewed as general suppressors of genomic programmes that are not required in a specific tissue type or during a particular developmental stage of an organism [1] . This would predict that a large set of PcG target genes is shared in different tissues , as only a small set of genes is expressed in a tissue-specific fashion [30] . In line with this view , we found that the majority of PcG target genes identified in the endosperm are also targeted by PcG proteins in vegetative tissues [11] , [19] , suggesting that different PcG complexes share a common set of target genes during different stages of plant development . However , we identified substantially fewer PcG target genes in the endosperm than previous studies found in seedlings consisting of a mixture of many diverse cell types [11] , [19] as well as in root hair and non-hair specific cell types [31] . The low number of identified H3K27me3 target genes in endosperm correlates well with reduced expression of the critical PRC2 components MEA and FIS2 in the same tissue [8] , [27] . A reason for lower expression of PcG proteins and only few PcG protein target genes in endosperm at 1–4 DAP could be that at this time , when mitotic activity is high , the endosperm has not yet acquired its terminal differentiation status [32] . In contrast , the cells profiled in the other studies [11] , [19] , [31] were mostly fully differentiated . This is similar to the situation in mammals , where lineage-specific genes often become targeted by PcG proteins only upon cell-fate commitment [33] , leading to cell-type specific PcG target profiles and gene expression patterns [34] , [35] . Furthermore , it should be noted that the endosperm has fundamentally different developmental origin and fate than vegetative tissues; it is derived after fertilization of the diploid central cell and will not contribute any cells to embryo and the developing new plant . Therefore , it is also possible that the reduced number of H3K27me3 target genes in the endosperm might reflect a less stringent requirement of PcG-mediated gene regulation in the endosperm than in vegetative tissues . In the endosperm as well as in vegetative tissues , genes encoding for transcription factors were highly enriched among PcG target genes ( this study and [11] ) , supporting the general idea that PcG proteins regulate cell identity by controlling expression of transcription factors [36] . Importantly however , H3K27me3 target genes were also prominently enriched for pectinesterases and glycosyl hydrolases - two enzyme classes that degrade major components of plant cell walls [28] , [37] , indicating an important role of the FIS PcG complex in the regulation of endosperm cellularization . The observed deregulation of both enzyme classes in fis2 mutant seeds might be the underlying cause of endosperm cellularization failure of fis mutants [29] . Loss of FIS function caused deregulation of only few H3K27me3 genes , similar to observations made in mammalian and Drosophila cells , where only a small subset of PcG target genes were deregulated upon depletion of PcG proteins [33] , [38] , [39] . Stable repression of FIS target genes could be due to secondary epigenetic modifications that together with FIS-mediated H3K27me3 keep PcG target genes repressed and which are not alleviated in FIS-depleted cells . Alternatively , it is possible that secondary epigenetic modifications are only recruited to FIS target genes upon loss of FIS function . As a third and complementary explanation for the lack of expression of a large number of FIS target genes in FIS-depleted endosperm we propose that the promoters of many PcG target genes lack binding sites for endosperm-specific transcriptional activators required for substantially increased expression in this tissue . This last explanation would imply that those FIS target genes that are deregulated in the fis2 mutant are even in wild type expressed in the endosperm . Indeed , deregulated FIS target genes were predominantly expressed during wild-type seed development ( Figure 6B ) , supporting the hypothesis that cis-acting tissue-specific enhancers are required for full induction of FIS target genes upon loss of H3K27me3 . TEs and TEGs were most prominently enriched among endosperm-specific H3K27me3 targets . This is in contrast to the situation in vegetative tissues , where these elements are largely excluded from PcG target genes [11] . We propose that reduced levels of DNA methylation in the endosperm allow targeting of the FIS PcG complex to defined sequence elements that are protected by DNA methylation in vegetative tissues . This conclusion is supported by the following findings made in this study: ( i ) Shared H3K27me3 targets were completely devoid of DNA methylation , indicating that DNA methylation prevents targeting by PcG proteins . ( ii ) Endosperm-specific H3K27me3 protein coding genes had much higher CG DNA methylation levels in vegetative tissues compared to genome-wide average DNA methylation levels , supporting the view that DNA methylation prevents these genes being targeted by PcG proteins in vegetative tissues . ( iii ) In the endosperm , the average DNA methylation level of endosperm-specific H3K27me3 targets was reduced compared to vegetative tissues . This trend was most pronounced for TEs , where DNA methylation level of endosperm-specific TEs were much lower compared to the genome-wide average DNA methylation of TEs in the endosperm . However , also TEGs and protein-coding genes had reduced DNA methylation levels in the endosperm compared to vegetative tissues , supporting the notion that reduced DNA methylation levels in the endosperm allow targeting of the FIS PcG complex to defined sequence elements . However , DNA demethylation is a global phenomenon [13] , [14] , but only selected sequences were targeted by the FIS complex , suggesting that DNA demethylation is necessary , but not sufficient for targeting of the FIS complex . The conclusion that DNA methylation and H3K27me3 are usually exclusive epigenetic marks is strongly supported by previous studies on seedlings with experimentally altered DNA methylation . When DNA methylation was reduced , H3K27me3 localized to defined regions within heterochromatin [25] , and when DNA methylation was increased H3K27me3 levels dropped [40] . Mutual antagonistic placement of DNA methylation and H3K27me3 was also identified at the imprinted Rasgrf1 locus in mouse [41] , suggesting an evolutionary conserved basis of the underlying mechanism . Together , we conclude that DNA methylation prevents targeting of PcG proteins to sequence elements that have the potential to recruit PcG proteins .
A transgenic Arabidopsis thaliana ( Landsberg erecta ( Ler ) ) line in which endosperm nuclei were specifically marked by EGFP was established by expressing a translational fusion of PHE1 with EGFP under the transcriptional control of the PHE1 promoter ( PHE1::PHE1-EGFP ) and 3 kb regulatory 3′ sequences . A transgenic Arabidopsis ( Columbia , Col ) line constitutively expressing YFP fused to histone H3 . 2 ( 35S::H3 . 2-YFP ) served as a positive control . The fis2-1 allele ( Ler accession ) has been described previously [3] . The met1-3 ( Col accession ) allele was described in [42] . For met1; fis2 double mutant analysis the newly identified fis2-5 allele ( SALK_009910; Col accession ) was used , containing a T-DNA insertion within the first exon . The fis2-5 seed abortion ratio and mutant seed phenotypes were analyzed and found to be similar to the fis2-1 allele ( data not shown ) . Seeds were surface sterilized ( 5% sodium hypochlorite , 0 . 1% Tween-20 ) and plated on MS medium ( MS salts , 1% sucrose , pH 5 . 6 , 0 . 8% bactoagar ) . Plants were grown in a growth cabinet under long day photoperiods ( 16 h light and 8 h dark ) at 22°C . After 10 days , seedlings were transferred to soil and plants were grown in a growth chamber at 60% humidity and daily cycles of 16 h light at 22°C and 8 h darkness at 18°C . Inflorescences were harvested approximately 21 days after transfer to soil , shock-frozen in liquid nitrogen and stored at −80°C . For analysis of seedlings , seeds were stratified for 2 days at 4°C before incubation in a growth cabinet . After 10 days , whole seedling tissue was harvested , shock-frozen in liquid nitrogen and stored at −80°C before further usage . Microscopy imaging was performed using a Leica DM 2500 microscope ( Leica Microsystems , Wetzlar , Germany ) with either bright-field or epifluorescence optics . Images were captured using a Leica DFC300 FX digital camera , exported using Leica Application Suite Version 2 . 4 . 0 . R1 , and processed using Photoshop 7 . 0 ( Adobe Systems Incorporated , San Jose , USA ) . Confocal imaging was performed on a Leica SP1-2 . Nuclei were isolated from 3 . 5 g of inflorescences following the protocol described in [43] . Isolated nuclei were resuspended in 1× PBS , and proteins were crosslinked to DNA with 1% formaldehyde for 8 min . After adding glycine to 125 mM final concentration and incubation for 5 min , crosslinked nuclei were washed and resuspended in 1× PBS and stained by addition of Propidium Iodide ( PI ) or DAPI to a final concentration of 1 µg/ml or 0 . 5 µg/ml , respectively . Biparametric flow analysis of EGFP fluorescence versus nuclear DNA content was performed on a fluorescence activated cell sorter ( FACS Aria II , Becton , Dickinson , Franklin Lakes , USA ) equipped with a 70 µm flow tip and operated at a sheath pressure of 70 psi . Events were thresholded on forward scatter and samples were sorted at the event rate of 15000/sec . For EGFP and PI excitation a 488 nm laser and for DAPI excitation a 407 laser were used . The barrier filters were 610/20 nm for PI , 450/40 for DAPI and 530/30 for EGFP fluorescence . The position of the nuclei gate was defined using 6 µm beads ( Becton Dickinson ) , forwards ( FSC-A ) and sidewards scatter ( SSC-A ) and was verified by DAPI-staining ( Figure S4A ) . The position of the sort region was established by first determining the baseline of green fluorescence using inflorescence nuclei from EGFP-negative Ler control plants ( Figure S4B ) . The upper and left- and right-hand boundaries of the sort window were adjusted to include all nuclei derived from YFP-positive 35S::H3 . 2-YFP control plants ( Figure S4B ) . Sorted GFP positive nuclei from PHE1::PHE1-EGFP plants were reanalyzed to verify sorting conditions ( Figure S4C ) . For expression analysis from sorted nuclei , RNA was isolated by flow sorting nuclei directly into 450 µl of RLT lysis buffer ( Qiagen , Hilden , Germany ) and using the RNeasy Plant Mini Kit ( Qiagen ) according to the manufacturer's recommendation . For other expression analyses , siliques were harvested at the indicated time points and RNA extraction and generation of cDNAs were performed using RNeasy Plant Mini Kit ( Qiagen ) according to the supplier's instructions . For quantitative RT-PCR , RNA was treated with DNaseI and reverse transcribed using the First strand cDNA synthesis kit ( Fermentas , Ontario , Canada ) . Gene-specific primers and Fast-SYBR-mix ( Applied Biosystems , Carlsbad , USA ) were used on a 7500 Fast Real-Time PCR system ( Applied Biosystems ) . Analysis was performed using three replicates and results were analyzed as described [44] . Briefly , mean expression values and standard errors for the reference gene as well as for the target genes were determined , taking into consideration the primer efficiency that was determined for each primer pair used . Relative expression values were determined by calculating the ratio of target gene expression and reference gene expression and error bars were derived by error propagation calculation . The primers used in this study are specified in Table S7 . ChIP with 500 to 700 ng of chromatin derived from approximately 100'000 sorted nuclei was performed as described [45] using antibodies against H3K27me3 ( Millipore , cat . 07-449 ) and rabbit IgG ( Santa Cruz Biotechnology , Santa Cruz , USA , cat . Sc-2027 ) . ChIP-DNA was amplified using the WGA-4 single cell amplification kit ( Sigma-Aldrich , St . Louis , USA ) . For amplification of input DNA , 10 ng of chromatin was used . Amplified DNA was purified with the QIAquick PCR purification kit ( Qiagen ) and eluted with 50 µL of water . DNA concentration was measured using a NanoDrop 1000 ( NanoDrop Technologies , Wilmington , USA ) .
|
Cell identity is established by the evolutionary conserved Polycomb group ( PcG ) proteins that repress transcriptional programs which are not required at particular developmental stages . The plant FERTILIZATION INDEPENDENT SEED ( FIS ) PcG complex is specifically expressed in the endosperm where it is essential for normal development . The endosperm nourishes the embryo , similar to the physiological function of the placenta in mammals . In this study , we established the cell type–specific epigenome profile of PcG activity in the endosperm . The endosperm has reduced levels of DNA methylation , and based on our data we propose that PcG proteins are specifically targeted to hypomethylated sequences in the endosperm . Among these endosperm-specific PcG targets are genes with functional roles in endosperm cellularization and chromatin architecture , implicating a fundamental role of PcG proteins in regulating endosperm development . Importantly , we identified transposable elements and genes among the specific PcG targets in the endosperm that are densely marked by DNA methylation in vegetative tissues , suggesting an antagonistic placement of DNA methylation and H3K27me3 at defined sequences .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/genomics",
"molecular",
"biology/histone",
"modification",
"genetics",
"and",
"genomics/plant",
"genetics",
"and",
"gene",
"expression",
"developmental",
"biology/plant",
"growth",
"and",
"development",
"molecular",
"biology/dna",
"methylation",
"genetics",
"and",
"genomics/epigenetics",
"molecular",
"biology/chromatin",
"structure"
] |
2010
|
H3K27me3 Profiling of the Endosperm Implies Exclusion of Polycomb Group Protein Targeting by DNA Methylation
|
All skeletal muscle progenitor cells in the body derive from the dermomyotome , the dorsal epithelial domain of developing somites . These multipotent stem cells express Pax3 , and this expression is maintained in the myogenic lineage where Pax3 plays an important role . Identification of Pax3 targets is therefore important for understanding the mechanisms that underlie the onset of myogenesis . In a microarray screen of Pax3-GFP sorted cells , with analysis on Pax3 gain and loss of function genetic backgrounds , we identify Dmrt2 , expressed in the dermomyotome , as a Pax3 target . In vitro gel shift analysis and chromatin immunoprecipitation with in vivo extracts show that Pax3 binds to a conserved 286 bp sequence , situated at −18 kb from Dmrt2 . This sequence directs reporter transgene expression to the somite , and this is severely affected when the Pax3 site is mutated in the context of the locus . In Dmrt2 mutant embryos , somite maturation is perturbed and the skeletal muscle of the myotome is abnormal . We now report that the onset of myogenesis is also affected . This depends on activation , in the epaxial dermomyotome , of the myogenic determination gene , Myf5 , through its early epaxial enhancer . This sequence contains sites that bind Dmrt2 , which belongs to the DM class of DNA–binding proteins . Mutation of these sites compromises activity of the enhancer in transgenic embryos where the reporter transgene is under the control of the Myf5 epaxial enhancer . Transactivation of this site by Dmrt2 is demonstrated in vitro , and conditional overexpression of Dmrt2 in Pax3 expressing cells in the somite confirms the role of this factor in the activation of Myf5 . These results reveal a novel genetic network , comprising a Pax3/Dmrt2/Myf5 regulatory cascade that operates in stem cells of the epaxial dermomyotome to initiate skeletal muscle formation .
The Pax family of transcriptional regulators play key roles in the onset of organogenesis and cell lineage specification during development [1] . Pax3 and Pax7 regulate skeletal muscle formation . Skeletal muscle progenitors in the trunk and limbs of vertebrates derive from somites , from the dorsal compartment known as the dermomyotome . In the mouse embryo , Pax3 is expressed throughout this epithelial structure , whereas Pax7 expression is restricted to the central domain . Myogenesis is initiated by the delamination of Pax3 positive cells from the edges of the dermomyotome; at certain axial levels , cells from the hypaxial domain migrate from the somite , before activating the myogenic determination genes Myf5 and MyoD , whereas other cells , most notably in the epaxial domain , have already activated Myf5 and immediately differentiate , on delamination , to form the first skeletal muscle of the epaxial myotome . Transcription of Myf5 at this site depends on an early epaxial enhancer , which lies at −5 . 5 kb from the gene [2] , [3] . This sequence is regulated by Shh [4]–[6] and by canonical Wnt signalling [6] from the adjacent axial structures , acting through Gli and TCF binding sites . Subsequently , as the somite matures , the central dermomyotome loses its epithelial structure and cells that are positive for Pax3 and Pax7 enter the underlying muscle masses . These provide an essential population of muscle stem cells for all subsequent muscle growth . In Pax3−/−; Pax7−/− double mutants , these cells fail to activate myogenic determination genes and many of them die [7] . In the absence of Pax3 alone , cell death is also observed at the extremities of the dermomyotome , where it is most evident in the hypaxial domain [1] . Pax3 target genes that are important for the onset of myogenesis are beginning to be identified . A classic example is provided by the gene encoding c-Met , a receptor required for delamination and migration of muscle progenitor cells [8] . More recently , a Myf5 regulatory sequence required for activation of this myogenic determination gene in the limb and hypaxial somite has been identified as a direct Pax3 target [9] . The second myogenic determination gene , MyoD , has been shown to be a target of Pax3/7 in a myogenic cell line and in postnatal muscle progenitor cells [10] and indeed other Pax7 [11] , [12] targets have been identified in this cellular context . Activation of myogenic determination genes leads to entry into the myogenic programme , which is accompanied by down-regulation of Pax3/7 , to which microRNAs contribute [13] . However , maintenance of a muscle stem cell population is essential in the growing organism and this is achieved through signalling systems that maintain a balance between self-renewal and differentiation [14] . In the context of FGF signalling , Pax3 directly regulates Fgfr4 and acts genetically upstream of other components of the pathway also , such as Sprouty1 , to control self-renewal versus differentiation of Pax3 positive muscle progenitors [15] . In order to identify further Pax3 target genes acting at earlier stages within the dermomyotome , we have performed a microarray screen that has led to the identification of Dmrt2 . This gene family was first characterized as doublesex in Drosophila [16]–[18] and as mab-3 in C . elegans [19] , where these genes play important roles in sex determination [20] , [21] . Vertebrate homologues have since been identified , some of which also play a role in sexual development [22] . However Dmrt genes are also expressed outside the gonads and have been implicated in other developmental processes . Dmrt2/Terra is expressed during somitogenesis in vertebrates [23] , where transcripts are present in the presomitic mesoderm ( PSM ) and then confined to the dermomyotome of somites . In the chick embryo , Terra is expressed symmetrically in the PSM , but has a transient asymmetrical expression around the node , implicating it in left-right axis determination and normal development of bilateral somites through interaction with the segmentation clock , as indicated by morpholino experiments in zebrafish [24] . Mouse embryos lacking Dmrt2 show somite patterning defects , culminating in malformed ribs and sternum leading to postnatal death due to respiratory problems [25] . Such malformations are often associated with abnormal dermomyotome and myotome development [1] and indeed Dmrt2−/− embryos have morphological defects in these somite compartments , with perturbation of the expression of myogenic markers reported at E10 . 5 and E11 . 5 . In Dmrt2 and Pax3 compound mutants perturbations in myogenesis were also observed [26] , as evidenced by muscle differentiation markers at E10 . 5 . However cell death in the absence of Pax3 [1] complicates the interpretation . The molecular function of Dmrt2 has not been investigated . Dmrt proteins are characterised by the DM domain , an intertwined Zn finger-like motif that interacts with the minor groove of DNA [27] . The conserved DM domains of mammalian Dmrt factors bind to similar DNA consensus sequences [28] . Drosophila Dsxm acts as a transcriptional activator whereas DsxF has repressor function [29] . When human DMRT1 is fused to the VP16 activation domain , transactivation through the consensus DMRT1 binding sequence is observed , however DMRT1 alone did not show much activity [28] . Regulation of the Dmrt2 gene , present in a cluster with Dmrt1 , 3 in mice , has not been examined , although the human DMRT2 gene has been shown to encode alternatively spliced transcripts [30] . In this report , we identify Dmrt2 as a target of Pax3 , acting directly through a regulatory sequence that directs expression in the somite . We show that Dmrt2 , in turn , acts on the epaxial enhancer of Myf5 . Perturbation of this regulatory network , functioning in the dermomyotome , has consequences for the onset of myogenesis .
In a screen designed to detect Pax3 targets in the dermomyotome of interlimb somites at E9 . 5 , this region of the somite was dissected from Pax3GFP/+ heterozygote and Pax3PAX3-FKHR-IRESnlacZ/GFP gain of function embryos [31] . After dissociation , GFP positive cells were separated by flow cytometry . Microarray comparisons of the two classes of samples led to identification of a series of sequences that are up- or down-regulated in the presence of PAX3-FKHR , a constitutively active form of the Pax3 transcription factor [31] . Among these sequences , Dmrt2 was up-regulated 1 . 87 fold ( data not shown ) . Dmrt2 is transcribed in mouse somites ( Figure 1 ) , as previously reported [25] . Expression is detected as a band in presomitic mesoderm ( S0 ) and in the first somites at E8 . 5 , where transcripts accumulate in the epaxial domain , as shown for immature caudal somites at E9 . 5 ( C″ ) . Sections of a Pax3IRESnlacZ/+ embryo at this stage ( Figure 1F ) show that Dmrt2 expression is confined to the Pax3 positive dermomyotome of the somite , in the epaxial domain of the most immature caudal somites ( Figure 1G ) , and then throughout this epithelium in more mature somites ( Figure 1H ) . Already , at this stage , the most mature anterior somites ( Figure 1I ) are beginning to lose Dmrt2 expression at the epaxial and hypaxial extremities of the dermomyotome . This is detected also by whole mount in situ hybridization at E9 . 5 ( Figure 1C and 1C′ ) and at E10 . 5 ( Figure 1D ) , when labelled cells are also detectable in the mesodermal core of the 1st and 2nd branchial arches and in the forelimb bud , to which myogenic progenitor cells have begun to migrate from the somites [1] . By E11 . 5 ( Figure 1E ) , Dmrt2 transcripts are only detectable in caudal somites . The expression of Dmrt2 is similar in Pax3GFP/+ heterozygote embryos ( Figure 2A and 2D–2F ) to that in wild type embryos ( Figure 1 ) . However Dmrt2 expression levels are higher in the dermomyotome of Pax3PAX3-FKHR-IRESnlacZ/GFP gain of function embryos ( Figure 2B and 2G–2I ) , whereas they are notably lower in Pax3Pax3-En-IRESnlacZ/+ embryos , where the presence of a fusion protein in which the DNA binding domain of Pax3 is fused to the repression domain of Engrailed , results in a partial loss of function phenotype [9] , attenuating the cell death seen in the absence of Pax3 . Diminution of Dmrt2 expression is also seen in Pax3 mutant embryos at E9 . 25 ( Figure S1 ) , prior to extensive cell death . Maintenance of Dmrt2 expression in the central dermomyotome may reflect the expression of Pax7 in this domain . These results confirm that Dmrt2 lies genetically downstream of Pax3 . In order to see whether Dmrt2 is a direct Pax3 target , we examined the Dmrt2 locus on mouse chromosome 19 , which includes other Dmrt genes ( 1 , 3 ) , for non-coding sequences that are conserved between species ( Figure 3A ) . A conserved 286 bp region at about −18 kb from the Dmrt2 gene attracted our attention . This region has five putative Pax3 binding sites ( Figure 3B ) . Electrophoretic mobility gel shift assays with oligonucleotides encompassing these sites and in vitro synthesized Pax3 protein showed that site2 binds Pax3 , a result confirmed by a super-shift experiment with a Pax3 antibody ( Figure 3C ) . Chromatin immunoprecipitation of embryo extracts at E9 . 5 demonstrated that Pax3 binds specifically to the 286 bp sequence at −18 kb from the Dmrt2 gene in vivo . The function of the 286 bp sequence was tested in transgenic embryos . It directs transgene expression in the dermomyotome , where the endogenous gene is expressed , with additional ectopic expression in the ventral somite ( Figure 3E ) . Mutation of Pax3 site2 in the 286 bp sequence interferes with the expression of the transgene in the dermomyotome ( Figure 3F ) . Other conserved regions , notably at +20 kb and +37 kb did not direct reporter gene expression in transgenic embryos . A BAC transgenic analysis had shown that 150 kb of 5′ and 50 kb of 3′ genomic sequence flanking the Dmrt2 gene directed somitic expression in transgenic embryos . When the Pax3 binding site in the 286 bp sequence was mutated in the context of this BAC , somitic expression was severely affected ( Figure S2 ) . This result therefore indicates that the Pax3 site within this sequence plays an important role in the regulation of the Dmrt2 gene . Dmrt2 has been implicated in somite patterning , with morphological abnormalities reported in mature somites from E10 . 5 [25] . These include abnormal expression of myotomal markers . Since Dmrt2 is first expressed in the epaxial dermomyotome where activation of the myogenic determination gene , Myf5 , initiates the onset of myogenesis [1] , we examined Myf5 expression in Dmrt2 mutant embryos . Myf5 activation in the epaxial domain is retarded in the absence of Dmrt2 . This is detectable on whole mount in situ hybridization ( Figure 4A and 4B ) . On sections of somites at equivalent axial levels , Myf5 expression is reduced . The delay in the colonisation of the myotome reflects the reduction in Myf5 transcription in the epaxial domain . ( Figure 4D and 4F ) . This perturbation in Myf5 activation has consequences for myogenesis , as evidenced by the delay in expression of the gene for the myogenic differentiation factor , myogenin ( Figure 4G and 4H ) , in the absence of Dmrt2 . Pax3 transcription , on the other hand , is not reduced at the onset of somitogenesis ( Figure 4I and 4J ) . The initiation of Myf5 expression in the somite depends on an early epaxial enhancer [2] , [3] . This element , used here in its extended form ( Myf5-EpExt; [6] ) , is regulated by Gli and TCF binding sites , targets of Hedgehog and canonical Wnt signalling . The enhancer contains four sites that are close to the consensus for Dmrt2 binding ( [28] , Figure 5A ) . Of these , sites 1 , 3 and 4 bind Dmrt2 , as shown by electrophoretic mobility shift assays ( Figure 5B–5D ) . This is specific since it is abolished by competition with the Myf5 sequence , but not by this sequence with mutations in the Dmrt2 binding sites ( Figure 5C ) . No good antibodies are available for Dmrt2 detection , so we used a human influenza hemagglutinin ( HA ) tagged version of the protein , and anti-HA antibodies to show that the complex is disrupted ( Figure 5D ) . The Myf5-EpExt enhancer , with a TK or Myf5BA promoter region drives nlacZ transgene expression in somites . When the Dmrt2 sites are mutated in the Myf5-EpExt enhancer , this expression is very reduced in newly forming somites and is perturbed or absent in more mature somites ( Figure 5E and 5F ) . Transactivation of the Myf5-EpExt enhancer by Dmrt2 was shown by co-transfection experiments in NIH3T3 cells , with this enhancer driving a luciferase reporter and increasing amounts of a Dmrt2 expression vector; mutation of the Dmrt2 sites ( 1 , 3 , 4 ) in the Myf5-EpExt sequence showed that activity depends on Dmrt2 binding ( Figure 6A ) . Strong transactivation by Dmrt2 was also seen in HEK293 cells ( results not shown ) . In order to look at Dmrt2 activation in vivo , a transgene was constructed in which expression of a sequence encoding Dmrt2 , under the transcriptional control of a strong universal CAG promoter , depends on Cre recombinase removal of an intervening CAT sequence . When crossed with a PGK-Cre line [32] , Dmrt2 expression is seen throughout the embryo ( Figure 6C , the left hand embryo ) , as also evidenced by expression of the Tomato red reporter ( Figure 6C′ ) . When crossed with a Pax3Cre/+ line [33] , the Dmrt2 transgene is transcribed at sites of Pax3 expression , including the somites and dorsal neural tube ( Figure 6D ) , where Tomato red coloration is also detected ( Figure 6D′ ) . Expression of the endogenous Myf5 gene was monitored in these transgenic embryos at E9 . 5 ( ss 23 ) . When the Dmrt2 transgene is activated by Pax3-Cre , Myf5 expression is more extensive in the newly formed somites ( Figure 6E ) than in the control embryo ( Figure 6H ) . Myogenin transcripts are also more widely expressed ( Figure 6F and 6I ) and sections of an immature somite show premature presence of assembled laminin , which marks the myotome basement membrane ( Figure 6G , compared to Figure 6J ) . Together , these data indicate that myogenesis is accelerated in these embryos , leading to an earlier entry and organisation of cells in the myotomal space . The effect on Myf5 expression is also seen when the Dmrt2 transgene is expressed in Myf5nlacZ/+ mice ( Figure 6K , compared to Figure 6L ) and sections show that Myf5 is now ectopically expressed in the central/hypaxial dermomyotome in the presence of Pax3-Cre ( Figure 6K′ and 6K″ compared to Figure 6L′ and 6L″ ) . A similar result was seen with PGK-Cre ( Figure S3 ) . This therefore demonstrates that Dmrt2 activates the Myf5 gene in vivo in the somite , leading to ectopic expression . Acceleration of the onset of myogenesis and myotome formation are probably the consequence of over-expression of Dmrt2 and over-activation of the epaxial enhancer of Myf5 .
We establish a genetic network that is initiated in the dermomyotome by Pax3 regulation of Dmrt2 and which , through Dmrt2 regulation of Myf5 , orchestrates the onset of myogenesis in the myotome . This direct transcriptional cascade and further genetic targets discussed in this paper are presented schematically in Figure 7 . Our demonstration that Dmrt2 is a Pax3 target was initially based on a genetic analysis that showed up-regulation of this gene in the presence of PAX3-FKHR . This transcriptional activator of Pax3 target genes may introduce a bias due to the FKHR domain , however down-regulation of Dmrt2 in the presence of the Pax3-Engrailed fusion protein and in the absence of Pax3 reinforces the interpretation . This is confirmed by the identification of a Pax3 dependent regulatory sequence , located at 18 kb 5′ of the Dmrt2 gene , which directs transgene expression to the somite . Mutation of the Pax3 site shown to bind this factor in vitro and in vivo , abolishes most of the activity in the dermomyotome . Since Pax7 can replace Pax3 in the trunk [34] , it is probable that this site also binds Pax7 . Indeed in Pax3 mutants , Dmrt2 continues to be expressed in the remaining dermomyotome [26] , where Pax7 prevents cell death in the absence of Pax3 [1] . Expression of the transgene is not confined to the dermomyotome , suggesting that the 286 bp element also responds to factors in the ventral somite , where transcription of the endogenous gene is repressed by other regulatory sequences . Dmrt2 is present in a locus that includes Dmrt1 and Dmrt3 . The latter lies about 30 kb 5′ of the 286 bp sequence . However , Dmrt3 is not transcribed in the somites and there is no indication that this regulatory element directs any aspect of Dmrt3 expression , which is characteristically seen at sites in the head [35] , [36] for example , where no labelling with the Dmrt2-286-TKnlacZ transgene is observed . Other regulatory regions , namely 2 . 6 kb immediately 5′ of Dmrt1 which directs transgene expression to the testis [37] or a highly conserved sequence between Dmrt1 and Dmrt3 , associated with XY sex reversal in humans [38] , which directs Dmrt3-like transgene expression , also show regulatory properties characteristic of the adjacent downstream gene only . It is well established that the DM domains of Dmrt2 factors bind to a similar consensus sequence [28] , however direct target genes have not been identified . We now show that the Myf5 gene is targeted by Dmrt2 through the expected DM consensus sequence . The transcriptional activity of mammalian Dmrt factors is not well defined , in contrast to Drosophila Dsx which acts as a transcriptional activator or repressor , when produced from a male or female allele , respectively [29] . We demonstrate that Dmrt2 functions as a transcriptional activator on a DM consensus binding site in the Myf5 early epaxial element . This was observed in assays with the NIH3T3 fibroblast cell line , and with HEK293 cells in which transcriptional activity of Dmrt1 was not detectable under conditions in which a Dmrt1-VP16 fusion protein transactivated a luciferase reporter through DM sites [28] . Our transgenic experiments which lead to Dmrt2 over-expression in vivo also point to the action of this factor as an activator of Myf5 expression . Dmrt2 mutant phenotypes in which somite disorganisation and effects on myogenesis had been documented previously , focussed on E10 . 5 and E11 . 5 embryos [25] , [26] . We observed that Dmrt2 transcription begins to be down-regulated from E10 . 5 and transcripts are no longer detectable in most somites by E11 . 5 . The muscle recovery noted in Dmrt2 mutants at later stages [25] is probably due to the invasion of the underlying muscle by Pax3/7 positive cells , from the central dermomyotome , which takes place from E10 . 5 [7] . Indeed in Dmrt2 mutants at E10 . 5 , the absence of apically polarised N-cadherin , in the dermomyotome ( [25] , Figure 4 ) , may indicate de-epithelialization and premature disaggregation of the central domain which is the source of these muscle progenitor cells . We have concentrated our analysis on the onset of myogenesis when Dmrt2 is strongly expressed in the dermomyotome . In the absence of Dmrt2 , the onset of Myf5 expression is affected . This depends on the epaxial enhancer [2] , [3] , which we show is regulated through Dmrt2 binding sites . This enhancer is activated in the epaxial somite by signals from the adjacent axial structures [6] . The requirement for Dmrt2 will restrict this activation to the dermomyotome . Dmrt2 alone does not direct ectopic expression of Myf5 outside the somite , indicating that other factors also play a role in this restriction . Myf5 plays a key role in the initiation of myogenesis , manifested by the formation of the early epaxial myotome . The perduration of β-galactosidase , either from a Myf5nlacZ allele or from a Myf5-EpExt regulated transgene , marks these first myogenic cells . When Myf5 expression is perturbed by the lack of Dmrt2 , the formation of the myotome is affected . In contrast , over-expression of Dmrt2 promotes Myf5 expression and myotome formation . Activation of the myogenic differentiation programme is evidenced by expression of the myogenic differentiation gene , myogenin , which is delayed in somites in the Dmrt2 mutant and increased in the early myotome when Dmrt2 is over-expressed . An important aspect of myotome formation is the laying down of the basal lamina that confines this somite compartment . In the absence of Dmrt2 , the expression of myogenic markers , such as myogenin , in the myotome is more diffuse ventrally and this correlates with a lack of laminin [25] . In contrast , when Dmrt2 is over-expressed , we observed premature expression of laminin , associated with an acceleration of myotome formation . Laminin is the ligand of α6β1 integrin , required for the formation of the myotome basal lamina . Interestingly , α6β1 integrin is not expressed on myogenic cells delaminating from the dermomyotome in Myf5 mutant mice , in which cells fail to locate correctly and the early myotome does not form [39] . Since this early expression of Myf5 depends on the epaxial enhancer , Dmrt2 lies genetically upstream of both this integrin receptor and its laminin ligand ( Figure 7 ) . The Pax3-Dmrt2-Myf5 regulatory cascade , identified by genetic and molecular approaches , has consequences for the onset of skeletal muscle formation , both in terms of the activation of the myogenic regulatory programme and of the organisation of myogenic cells and their progenitors within the somite . In Pax3 mutants , epaxial myogenesis occurs , however perturbations in epaxial derivatives , such as deep back muscles , are observed [40] . Other Myf5 regulatory sequences control myotomal expression [2] . Pax3 also directly activates another Myf5 regulatory element at later developmental stages [9] . This key factor therefore , directly or indirectly influences different spatiotemporal aspects of myogenesis . Expression of Dmrt2 is not limited to myogenic progenitors in the epaxial domain , but is present throughout the developing dermomyotome , where it may be implicated in maintaining this epithelial structure [25] . We have recently identified Foxc2 as a gene that is negatively regulated by Pax3 in the dermomyotome and that promotes non-myogenic cell fates [41] . Identification of these targets begins to reveal how Pax3 controls cell behaviour in the somite before the determination of myogenic cells . Integration of Pax3 targets into a regulatory network , comprising elements such as those discussed in this paper ( Figure 7 ) , is essential for understanding how such a key regulatory factor orchestrates the progression from stem cell towards differentiated tissue . Defining such regulatory networks that control developmental processes is a major challenge , taken up by the emerging field of systems biology .
Generation and genotying of Pax3IRES-nlacZ/+ , Pax3GFP/+ [7] , Pax3PAX3FKHR-IRESnlacZ/+ [31] , Pax3Pax3-En-IRESnlacZ/+ [9] , Pax3Cre/+ [33] , PGK-Cre [32] , Myf5nlacZ/+ [42] , Dmrt2+/− , and Dmrt2−/− [25] mice was carried out as previously described . Pax3GFP/+ mice were crossed with PGK-Cre transgenic mice to obtain Pax3GFP/+; PGK-Cre females , and these females were crossed with Pax3PAX3FKHR-IRESnlacZ/+ males to obtain embryos with one Pax3GFP allele and one floxed Pax3PAX3FKHR-IRESnlacZ allele [31] . Somites were dissected from the interlimb region of E9 . 5 embryos . The GFP positive cells were collected by FACs from Pax3GFP/+ and Pax3Pax3FKHR–IRESnlacZ/GFP mutant embryos . About 3 . 0×105 cells were collected for RNA preparation , cDNA synthesis and subsequent analysis using Affymetrix microarrays . Detailed description of the microarray analysis will be published elsewhere ( Lagha and Sato , in preparation ) . The mouse Dmrt2 cDNA ( contains the complete coding region; NM145831 ) was isolated by RT-PCR from cDNA of RNA prepared from C57BL/6 mouse embryos at E9 . 5 . The Dmrt2 cDNAs with or without a 3′ sequence coding HA tag were subcloned into pBS and pcDNA3 vectors for in situ hybridization , electrophoresis mobility shift , and overexpression assays . To generate transgenes with the conserved 286 bp sequence located at −18 kb from the Dmrt2 gene ( Dmrt2-286 ) , or the extended Myf5 epaxial enhancer ( Myf5-EpExt , [6] ) , these sequences were obtained by PCR from total DNA of E9 . 5 , C57BL/6 embryos using KOD DNA polymerases ( Novagen ) and ligated into pBS plasmids to lie 5′ of TK-nlacZ or Myf5BA-nlacZ transgenic sequences [43] . The primers used were as follows , Dmrt2-286-Fwd; 5′-GCAGCGGCCGCTATCAGGAATGAATTAGCTTGTTTCCCTCC-3′ , Dmrt2-286-Rev; 5′- GCCACTAGTCTGGACCTCCTTTCCCCGATGCGTGCCCT-3′ , Myf5-EpExt-Fwd; 5′-GCAGCGGCCGCATGTAGACTCCTCACTTCTCGTTAGTAGA-3′ , Myf5-EpExt-Rev; 5′- GCCACTAGTGGATCCCGGCTGGCAAATGTTTGGTTCCCT-3′ . Transgenes with mutated Pax3 or Dmrt2 binding sites were created with the QuikChange Multi Site-Directed Mutagenesis kit ( Stratagene ) . The Myf5EpExt-TK fragment was digested with NheI and NcoI , and ligated into the pGL4 vector ( Promega ) in front of the luciferase reporter for transactivation assays . For transgenic expression of Dmrt2 , the HindIII and BamHI fragment digested from the pBS-CAG-floxedCAT vector [44] , the HindIII and XbaI fragment digested from pBS-Dmrt2 , and the XbaI and BamHI fragment digested from the pCMVTnT-IRES2-tdTomato-pA vector ( pCMVTnT; Promega , pIRES2; Clontech , tdTomato; a gift from Dr . Tsien ) were all ligated together to make the pBS-CAG-floxedCAT-Dmrt2-IRES2-tdTomato vector ( pBS-CAG-LSL-Dmrt2 ) . pBS-Dmrt2-286-TK-nlacZ , pBS-Myf5EpExt-TK-nlacZ , pBS-Myf5EpExt-BA-nlacZ and pBS-CAG-LSL-Dmrt2 vectors were linearized to produce transgenic mice according to standard techniques . The primers used for genotyping of CAG-LSL-Dmrt2 mice were as follows , F502; CTCCGGAGGCAGCAGGCCACAG , R620; ATGCTTTTGGCCAGCAAACTCG; R794; CGCGATGTCCCAAATGGACCTAA ( wildtype; 293 bp , transgene; 119 bp ) . For BAC transgenic analysis , a BAC containing Dmrt2 genomic DNA ( −150 kb/+50 kb: clone RP24-290E12 purchased from BACPAC resources center , CHORI ) was used for targeting with an nlacZ reporter into the ATG site of Dmrt2 . The Dmrt2 BAC with a mutated Pax3 binding site in the 286 bp sequence was created using the fragment from the transgene with the mutated Pax3 binding site in Dmrt2-286 . All BAC recombineering were performed with SW105 and SW106 strains [45] . Whole mount in situ hybridization analyses and X-gal staining were performed as previously described [15] . For double detection of β-galactosidase ( β-gal ) from a Pax3 nlacZ allele and Dmrt2 transcripts , the former was visualized by Red-gal ( Sigma ) , and the latter using a digoxigenin ( DIG ) -labeled probe visualized with BM purple ( Roche ) . For immunohistochemistry of laminin , the antibody against laminin ( Chemicon; AL-4 clone ) was used as described [39] . EMSA was performed using the LightShift Chemiluminescent EMSA kit ( Thermo Fisher Scientific ) . Protein extracts containing Pax3 or Dmrt2 protein were prepared from HEK293 cells transfected with pcDNA3-Pax3 [46] or pcDNA3-Dmrt2HA . Cells were then lysed on ice in extraction buffer ( 15 mM Tris pH 7 . 4; containing protease inhibitor cocktail ( Roche ) ) supplemented with 0 . 5% NP40 . Lysates were centrifuged at 2000×g for 5 min at 4°C and the supernatant was stored in 25% glycerol at −80°C until ready for use . The 5′-biotin-conjugated oligos and unlabeled probes used were as follows , Pax3BS1; 5′-CTTGTTTCCCTCCTGTAAGTGATCC-3′ , Pax3BS2; 5′-CTGTGGTGTGTGACTAATGGAGTGC-3′ , Pax3BS3; 5′-CTCTCAGGGCTGGTTTAAGACTCA-3′ , Pax3BS4; 5′-CATAGCAGGGACACAGTAAAGCCAC-3′ , mutPax3BS2; 5′- CTGTGGTGACGTCTAAATGGAGTGC-3′ , Dmrt2BS1; 5′-CTGTTTCCTAGTGTAGATCT-3′ , Dmrt2BS2; 5′- GATTTTACAACGTGTGCCGC-3′ , Dmrt2BS3; 5′-TCTGTTTTTCCTGTAACCTC-3′ , Dmrt2BS4; 5′-TCTCTGTAGCTCTCATGTAAA-3′ , mutDmrt2BS1; 5′-CTGTTTCCTTGTTTTGATCT-3′ , mutDmrt2BS3; 5′-TCTGTTTTTCCTTTTATCTC-3′ , mutDmrt2BS4; 5′-TCTCTTTTGCTCTCATTTTAA-3′ . Sense and anti-sense strands were synthesized and 40 fmol biotin-labeled double-stranded DNA was mixed with 1 µg protein from a HEK293 cell extract . Gel mobility shift assays and hybridization were performed according to the instructions of Invitrogen , using 6% DNA retardation gels ( Invitrogen ) . For ChIP experiments , somites were collected from E9 . 5 embryos with heads , neural tubes and internal organs removed . These samples were mechanically digested through a syringe and dissociated cells were fixed with 1% formaldehyde at room temperature for 10 min . The ChIP procedure was performed according to an enzymatic digestion protocol ( ChIP-IT Express; ActiveMotif ) with Pax3 antibody raised from Rabbit ( ActiveMotif and Geneka; Bajard et al . , 2006 and Lagha et al . , 2008 ) or normal Rabbit serum ( Sigma ) as a control . Input DNA and immunoprecipitated DNA were analyzed by PCR . The sequences of primers were designed as follows . The Dmrt2-18 kb conserved region ( 286 bp ) Fwd; 5′-GCTTGTTTCCCTCCTGTAAGT-3′ , Rev; 5′- GTGTAGGATCTGTGGCTTTAC-3′ and the Dmrt2+20 kb control conserved region ( +20 kb ) Fwd; 5′-GGTTCTCATAATTTACATGCT-3′ , Rev; 5′- TCCAACATCTGATTGTACTTA-3′ . For luciferase assays , pGL4-Myf5EpExt vectors were transfected into NIH3T3 cells together with a pRL-TK plasmid ( Promega; for normalizing ) and test plasmids ( Gli1 and β-catenin expression vectors were kindly provided by Dr . S . Brunelli ) . Transfected cells were cultured for 24 hours , and subjected to luciferase assays using the Dual-Luciferase Reporter Assay System ( Promega ) .
|
It is well established that skeletal muscle derives from segmented structures called somites that form on either side of the axis of the embryo . The part of the somite that contains muscle stem cells is called the dermomyotome . These cells express the transcription factor Pax3 , which regulates muscle stem cell behaviour . We now show that the Dmrt2 gene , also expressed in the dermomyotome , is directly controlled by Pax3 . Since Dmrt2 has been implicated in maintaining the integrity of the dermomyotome , this therefore indicates an upstream role for Pax3 in this structure as well as in controlling cells that form skeletal muscle . Furthermore Dmrt2 directly regulates early activation of the myogenic determination gene , Myf5 , required for the formation of the first skeletal muscle in the somite . This is a novel function for Dmrt2 and shows that this transcription factor controls both structure and cell fate . Our results reveal a Pax3/Dmrt2/Myf5 regulatory cascade through which Pax3 orchestrates the onset of myogenesis in the muscle stem cells of the dermomyotome .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/gene",
"function",
"developmental",
"biology/developmental",
"molecular",
"mechanisms",
"genetics",
"and",
"genomics/gene",
"expression"
] |
2010
|
A Pax3/Dmrt2/Myf5 Regulatory Cascade Functions at the Onset of Myogenesis
|
Non-typhoidal Salmonella ( NTS ) is a leading cause of bloodstream infections in Africa , but the various contributions of host susceptibility versus unique pathogen virulence factors are unclear . We used data from a population-based surveillance platform ( population ~25 , 000 ) between 2007–2014 and NTS genome-sequencing to compare host and pathogen-specific factors between individuals presenting with NTS bacteremia and those presenting with NTS diarrhea . Salmonella Typhimurium ST313 and Salmonella Enteritidis ST11 were the most common isolates . Multi-drug resistant strains of NTS were more commonly isolated from patients presenting with NTS bacteremia compared to NTS diarrhea . This relationship was observed in patients under age five [aOR = 15 . 16 , 95% CI ( 2 . 84–81 . 05 ) , P = 0 . 001] , in patients five years and older , [aOR = 6 . 70 95% CI ( 2 . 25–19 . 89 ) , P = 0 . 001] , in HIV-uninfected patients , [aOR = 21 . 61 , 95% CI ( 2 . 53–185 . 0 ) , P = 0 . 005] , and in patients infected with Salmonella serogroup B [aOR = 5 . 96 , 95% CI ( 2 . 28–15 . 56 ) , P < 0 . 001] and serogroup D [aOR = 14 . 15 , 95% CI ( 1 . 10–182 . 7 ) , P = 0 . 042] . Thus , multi-drug-resistant NTS was strongly associated with bacteremia compared to diarrhea among children and adults . This association was seen in HIV-uninfected individuals infected with either S . Typhimurium or S . Enteritidis . Risk of developing bacteremia from NTS infection may be driven by virulence properties of the Salmonella pathogen .
Non-typhoidal Salmonella ( NTS ) is a leading cause of bacteremia in Africa [1] . Whereas NTS commonly manifests as self-limiting diarrhea in immune-competent individuals , NTS bacteremia occurs at high rates in young children ( 175–388 cases per 100 , 000 ) [2 , 3] and HIV-infected adults ( 2000–7500 cases per 100 , 000 ) [4–8] in sub-Saharan Africa and may also be associated with malaria co-infection and malnutrition in children [9–14] . In addition to its association with certain host-factors and co-infections , recent evidence suggests that NTS bacteremia in Africa may be associated with the emergence of novel Salmonella Typhimurium genotypes that have undergone considerable genomic reduction , which may suggest human adaptation leading to invasive blood-stream infection [15–24] . The recent emergence of multi-drug resistance ( MDR ) ( defined as resistant to trimenthroprim-sulfamethoxasol [TMP-SMX] , ampicillin , and chloramphenicol ) has also been observed in dominant serovars of Salmonella in Africa , including S . Typhimurium and S . Enteritidis [16 , 17 , 19 , 25–28] . Additionally , a clone of S . Typhimurium ( ST313 ) , which has undergone considerable genomic reduction , is a common cause of bacteremia in HIV infected individuals and children across sub-Saharan Africa [16] . Since invasive NTS infections frequently occur in individuals with significant co-morbidities , including HIV infection , malaria , and malnutrition , it remains unclear whether genome reduction , and associated changes in drug resistance and virulence , has contributed to high rates of NTS bacteremia in Africa [16 , 17 , 24 , 29] . Here we test for associations between NTS bacteremia and both host and pathogen-specific factors . Using a large population-based surveillance cohort from rural Kenya , we compared individuals with NTS bacteremia to those with NTS diarrhea to elucidate host and pathogen-related risk factors for NTS invasiveness . Finally , we genotyped a subset of NTS isolates to determine the phylogenetic lineages contributing to the high burden of NTS bacteremia in western Kenya .
This study was nested within an ongoing population-based infectious disease surveillance system ( PBIDS ) conducted since 2006 by the Kenya Medical Research Institute ( KEMRI ) in collaboration with the US Centers for Disease Control and Prevention ( CDC ) , and the methods have been described previously [28 , 30 , 31] . The PBIDS protocol was reviewed and approved by the Institutional Review Boards of KEMRI and CDC . All adult subjects provided written , informed consent and a parent or guardian of any child provided informed consent on their behalf . The surveillance cohort consists of approximately 25 , 000 individuals of all ages residing in Asembo , a rural region of western Kenya located in Siaya County on Lake Victoria . The region is characterized by low population density , intense , year-round malaria transmission and an adult HIV prevalence of 15–17% [28] . Inclusion into the surveillance required the following: 1 ) currently reside within a village whose epicenter is located no more than 5 kilometers from the Lwak Mission Hospital ( LMH ) in Asembo , 2 ) have resided permanently in the area for at least four calendar months , and 3 ) have provided written informed consent/assent . PBIDS participants are visited biweekly in their households and asked about acute illnesses , healthcare seeking , and changes to the household ( i . e . births , deaths , migrations ) . Participants can access free health care for acute illnesses at LMH . At LMH , blood samples for culture are collected from patients meeting the case definition for 1 ) severe acute respiratory infection ( in children <5 years cough or difficulty breathing , plus lower chest indrawing or danger sign [unable to drink/feed , vomiting everything , convulsions , lethargy , unconscious , stridor] or oxygen saturation <90%; in those ≥5 years cough or difficulty breathing or chest pain plus temperature ≥ 38 . 0 degrees C or oxygen saturation <90% ) , 2 ) acute febrile illness ( temperature ≥ 38 . 0 degrees C ) , 3 ) jaundice , or 4 ) hospital admission ( whether or not fever was present ) . Due to the large number of individuals with febrile illnesses reporting to Lwak Mission Hospital , only the first two individuals over/under age five years presenting with acute febrile illness each day had blood collected for culture; there were no limits on blood culture collected for severe acute respiratory infection , jaundice or hospitalization . Height and weight were measured at time of clinical presentation for children under five years of age . Height-for-age ( HAZ ) and weight-for-height z-score ( WHZ ) cut-offs of -2 . 0 were used to classify stunting and wasting , respectively , in children under five years of age based on 2006 World Health Organization child growth standards [32] . Z-scores were excluded if HAZ was below -6 or above 6 , WAZ below -6 or above 5 , WHZ below -5 or above 5 [32] . HIV status was ascertained on a subset of NTS patients through two HIV home-based counselling and testing campaigns ( 2008/2009 and 2013 ) conducted on all consenting individuals ≥ age 13 years present in the disease surveillance cohort at the time of the survey . Children under 13 years of age were only tested if their parent was HIV infected . HIV data from these two community-level surveys were then linked to the clinic database through patient records . Analysis was done on the full subset of individuals linked to an HIV status , assuming that individuals who tested positive for HIV after NTS diagnosis were HIV infected at NTS diagnosis and those who tested negative for HIV prior to NTS diagnosis were HIV uninfected at NTS diagnosis . Where sample size allowed we secondarily restricted our analysis to those whose status could be confirmed at NTS diagnosis , ( i . e . , tested positive before NTS diagnosis , tested negative after NTS diagnosis , or children under 13 tested at any time prior to NTS diagnosis ) . Blood smears for malaria diagnosis were obtained from all individuals presenting with acute febrile illness and examined by trained microscopists on site . Malaria at the time of NTS diagnosis was defined as the presence of asexual stages of Plasmodium falciparum on microscopy using thick smear collected on the same day as stool/blood culture yielding NTS . Laboratory methods for culturing bacterial pathogens from blood have been described previously [28] . Briefly , blood ( 7–10 ml for adults and 1–3 ml for children ) was inoculated into BACTEC culture bottles ( Becton , Dickinson and company , Sparks , MD , USA ) and incubated in an automated BACTEC 9050 at 35°C for 1–5 days . A Gram-stained smear was prepared from any bottle with growth , and the broth sub-cultured onto standard enriched culture media for further identification . Stool specimens were processed according to standard protocols described elsewhere [33] . Colonies of Salmonella were identified and confirmed using an API 20E system ( Appareils et Procedes d’Identification , Montalieu Vercieu , France ) following manufacturers’ instructions . Commercial agglutinating antiserum ( Denka Seiken , Tokyo , Japan ) was used to serotype Salmonella isolates ( Krieg NR , Holt JG ( 1984 ) Bergey’s Manual of Systematic Bacteriology . Baltimore: Williams and Wilkins . pp 427–58 ) Antimicrobial susceptibility patterns were determined using standard Kirby-Bauer disc diffusion techniques after incubation for 16 hours [34] . Susceptibility was classified according to three MIC cutoff values ( resistant , intermediate , and susceptible ) for NTS as defined by Clinical and Laboratory Standards Institute CLSI [34] . Multi-drug resistance ( MDR ) was defined as resistance to chloramphenicol , trimethroprim-sulfamethoxasol ( TMP-SMX ) , and ampicillin [28] . For purposes of analysis , NTS isolates were classified as resistant or non-resistant whereby non-resistant included susceptible and intermediate resistance . NTS isolates obtained from a subset of individuals who provided consent after sample collection were sent to laboratories at the University of Washington for whole genome sequencing to assess the phylogenetic relatedness of the bacteremic and diarrheal NTS isolates in western Kenya . To isolate genomic DNA for sequencing , strains were grown overnight at 37°C with shaking in 3 mL of Luria Broth ( BD Biosciences , USA ) . Genomic DNA was isolated using Gentra Puregene Yeast/Bact . Kit ( Qiagen , Valencia , CA ) , according to manufacturer’s directions . For each genome standard Illumina Nextera or Nextera XT libraries were constructed according to manufacturer’s guidelines ( Illumina Inc . , San Diego , CA ) . Prior to the Nextera XT library normalization/denaturation step , double stranded libraries were normalized with the Invitrogen SequalPrep Normalization Plate Kit ( Thermo Fisher Scientific/Life Technologies , Grand Island , NY ) . Paired-end libraries for each genome were used to generate 100 bp or 150 bp reads with the Illumina HiSeq 2000 or MiSeq . Sequencing of libraries was performed according to manufacturer’s standards ( Illumina Inc . , San Diego , CA ) . Sequence types ( STs ) were generated from whole genome shotgun ( WGS ) sequence reads using the MLST scheme developed by Achtman et al . 2012 [35] , which uses fragments from seven housekeeping genes: aroC , dnaN , hemD , hisD , purE , sucA and thrA . We created a single MLST reference sequence comprised of the scheme’s seven allele template . The Burrows-Wheeler Alignment ( BWA ) algorithm [36] was used to align sequence reads from each strain to the MLST reference sequence using an edit distance of 10 , which allowed alignment of reads with up to 10 mismatches . Custom scripts were used to parse alignments to produce consensus sequences for each housekeeping gene , compare consensus sequences to the MLST allele database at EnteroBase ( http://enterobase . warwick . ac . uk ) , and generate STs based on the MLST database of 7 allele combinations . Phylogenetic analysis was restricted to S . Typhimurium isolates and was generated from whole genome sequences using the kSNP software package version 2 . 1 . 1 . , in which SNP discovery was based on k-mer analysis , i . e . single variant positions within sequences of nucleotide length k [37] . The maximum likelihood tree was constructed using 31-mers that were identified in at least 50% of the strains and was based on 3 , 058 SNPs . The 50% requirement provided phylogenetic resolution of the S . Typhimurium strains while excluding the SNPs present in only one or a small number of genomes , which are more likely to be the result of sequencing or assembly errors or from mobile elements such as phages or plasmids . Support for branch nodes was computed by FastTree version 2 . 1 . 3 [38] , which is provided in the kSNP package . Tree branches are expressed in terms of changes per total number of SNPs , not changes per site , as SNP-based trees do not include invariant sites . Local support values are based on the Shimodaira-Hasegawa test on the three alternate topologies at each split in the tree . The tree was drawn using Dendroscope version 3 . 2 . 10 [39] . NTS bacteremia cases were defined as those with at least one NTS serotype isolated from the blood . Stool samples were collected from patients presenting with diarrhea ( ≥3 looser than normal stools in a 24-hour period ) , and those with NTS isolated from the stool , ( and did not meet indication for blood culture or were blood culture negative ) , were classified in this analysis as NTS diarrheal cases . Any individuals with NTS isolated from both blood and stool culture were classified as NTS bacteremia cases . Repeat visits within one month of the original NTS diagnosis in which NTS was isolated from either blood or stool were excluded to avoid counting the same episode of NTS twice . Logistic regression with robust standard errors was used to compare the odds of host and pathogen characteristic between NTS bacteremia and NTS diarrhea cases in univariate and multivariate models . Multivariate models were stratified on age over and under five years and were adjusted for potential confounders defined a priori , including age ( continuous ) , year of diagnosis , malaria smear positivity , and multi-drug resistant phenotype . Because HIV is a major risk factor for NTS bacteremia , adjusted odds ratios were further stratified on HIV status to estimate the independent effects of pathogen-specific factors on clinical outcome in both HIV-infected and HIV-uninfected groups .
Between January 1 , 2007 and December 31 , 2014 there were 140 , 940 clinic visits to Lwak Mission Hospital among 24 , 748 unique individuals ( Fig 1 ) . Among 15 , 113 individuals who underwent blood culture , 842 had at least one bacterial pathogen isolated of which 136 were identified as contaminants , resulting in 706 ( 4 . 7% ) bacterial pathogens isolated , of which 257 ( 36 . 4% ) were NTS . Among 1 , 856 individuals whose stool specimens were tested by culture , 648 ( 34 . 9% ) grew one or more bacterial pathogens , of which 80 ( 12 . 4% ) were NTS . Among the 257 blood NTS isolates , 230 met the inclusion criteria for NTS bacteremia . Five blood NTS isolates were excluded as they were secondary blood culture NTS isolates occurring within one month of the index blood culture NTS episode . Twenty-two blood culture NTS isolates were excluded after being identified as S . Typhi through subsequent full genome sequencing of a subset of NTS isolates . Among 80 NTS isolates from stool , 73 met the criteria for diarrheal NTS . Among stool culture positive NTS isolates , 6 were excluded due to their occurrence in the same individual within one month of an index case of either blood or stool culture positive NTS . One stool NTS isolate was excluded after being identified as S . Typhi through subsequent full genome sequencing . Among the 230 NTS bacteremia cases included in the analysis , 59 ( 25 . 7% ) were linked to an HIV status ascertained from either of two population-based HIV surveys conducted by CDC-Kenya in 2008/2009 and 2013 . Among 73 NTS diarrhea cases , 28 ( 38 . 4% ) were linked to an HIV status from the same surveys . Comparison of demographic , clinical , and pathogen-specific factors between NTS bacteremia and NTS diarrhea cases . Among children under age five , those with NTS bacteremia tended to be older ( mean age 2 . 2 years ) than those with NTS diarrhea ( mean age 1 . 3 years ) , P = 0 . 003 ( Table 1 ) . Among cases age five and older , those with NTS bacteremia tended to be younger ( mean age of 26 . 9 years ) than those with NTS diarrhea ( mean age of 32 . 5 years ) , though the difference was not significant . Among cases 5 years and older with any HIV testing result , HIV prevalence was higher among bacteremia cases compared with diarrheal cases ( 71 . 7% versus 40 . 7% , P = 0 . 018 ) . This difference held when restricted to those whose HIV status was known at the time of NTS infection ( 79 . 2% versus 16 . 7% , P = 0 . 009 ) . Co-infection with malaria was more common in bacteremic versus diarrheal patients in both age-groups , though the difference was not significant . NTS bacteremia resulted predominantly from infection with Salmonella serogroups B and D ( which include the most commonly occurring NTS serovars in Africa , S . Typhimurium and S . Enteritidis , respectively [16 , 17 , 19 , 25–28] ) . In children under five , group B Salmonella was isolated from 64 . 1% of patients with bacteremia and 43 . 8% of patients with diarrhea ( P < 0 . 001 ) ; and group D Salmonella was isolated from 34 . 2% of patients with bacteremia and 18 . 8% of patients with diarrhea ( P < 0 . 001 ) . A similar pattern was observed for individuals over age five . MDR was more common among NTS isolated from bacteremic patients compared to diarrheal patients in both those under 5 years ( 83 . 3% versus 21 . 4% , P <0 . 001 ) and those ≥ 5 years of age ( 77 . 1% versus 26 . 4% , P <0 . 001 ) . Multi-locus sequence typing ( MLST ) was conducted on a subsample of 185 stored NTS isolates of which 166 were from bacteremic patients and 19 were from diarrheal patients . Among eighty-five NTS isolates from patients under age five with bacteremia 57 . 6% were S . Typhimurium ST313 , 38 . 8% were S . Enteritidis ST 11 , and 1 . 2% were S . Virchow ST16 . Among 81 NTS isolates genotyped from bacteremic patients over age five , 66 . 7% were S . Typhimurium ST313 , 29 . 6% were S . Enteritidis ST11 , 1 . 2% were S . Typhimurium ST19 , 1 . 2% were S . Enteritidis ST6 and 1 . 2% were S . Newport ST188 . Of three stool NTS samples from individuals under five years of age , one was S . Enteritidis ST11 and two were S . Virchow ST16 . Of 16 stool NTS samples from individuals over age five , 6 ( 35 . 3% ) were S . Typhimurium ST313 and 5 ( 29 . 4% ) were S . Enteritidis ST11 ( Table 1 ) . Phylogenetic reconstruction of 115 S . Typhimurium strains ( 106 from bacteremic patients , six from diarrheal patients , and three repeat tests from the same patient at separate visits within one month ) grouped almost exclusively with the ST313 phylogenetic clade associated with strain D23580 ( lineage 2 ) from Malawi [22] ( Fig 2 ) . Only nine S . Typhimurium strains grouped outside this clade: Six ST313 strains , including four bacteremic and two diarrheal strains , grouped with the chloramphenicol sensitive strain A130 from Malawi ( lineage 1 ) , which represents a ST313 lineage distinct from that of D23580 [16 , 40]; and three strains that grouped outside the branch leading to the ST313 lineages were sequence type ST19 ( Fig 2 ) . Multivariate adjusted logistic regression models , stratified by individuals under and over age five were used to identify risk factors associated with NTS bacteremia compared to NTS diarrhea . Among children under age five , a total of 104 cases of NTS bacteremia and 10 cases of NTS diarrhea with non-missing data on age , year of diagnosis , concurrent malaria parasitemia , multi-drug resistance phenotype of the NTS isolate , and serogroup of the NTS isolate were included in the multivariate logistic regression model . ( Table 2 ) . MDR phenotypes were more common among isolates cultured from NTS bacteremic patients compared to NTS diarrhea patients , [aOR = 15 . 16 , 95% CI ( 2 . 84–81 . 05 ) , P = 0 . 001] . This association was independent of age , malaria , serogroup , and year of diagnosis . We observed a positive , though non-significant association between concurrent malaria parasitemia and NTS bacteremia . We found no evidence of an age association with NTS bacteremia versus diarrhea in children under age five . Among individuals age five and over , a total of 89 cases of NTS bacteremia and 27 NTS diarrhea cases were included in the model . Similar to children under age five , we found no evidence of an association between NTS bacteremia and age , concurrent malaria parasitemia , or serogroup in multivariate models ( Table 2 ) . MDR NTS was associated with NTS bacteremia as compared to diarrhea , [aOR = 6 . 70 95% CI ( 2 . 25–19 . 89 ) , P = 0 . 001] in multivariate models . The association between MDR and NTS bacteremia , furthermore , held in secondary multi-variate analysis not adjusting for concurrent malaria , which enabled the inclusion of all individuals regardless of whether they were tested for malaria ( n = 53 NTS diarrhea and 105 NTS bacteremic patients ) , [aOR = 7 . 07 , 95% CI ( 2 . 89–17 . 33 ) , P < 0 . 001] . Among the entire sample of individuals ( including those both under and over age five ) , the association between MDR and NTS bacteremia was observed in those infected with Salmonella serogroup B , [aOR = 5 . 96 , 95% CI ( 2 . 28–15 . 56 ) , P < 0 . 001] and serogroup D , [aOR = 14 . 15 , 95% CI ( 1 . 10–182 . 7 ) , P = 0 . 042] . Serogroups C1/C2 were exclusively non-MDR and were only isolated from diarrheal patients . The association between MDR phenotype and NTS bacteremia was strongest among HIV-uninfected individuals , [aOR = 22 . 61 , 95% CI ( 2 . 53–185 . 0 ) , P = 0 . 005] , and was positive , but not significant among HIV-infected individuals [aOR = 2 . 74 , 95% CI ( 0 . 48–15 . 66 ) , P = 0 . 258] ( Table 3 ) . HIV infected individuals with either stool or blood NTS were more likely than HIV negative individuals to be infected with a MDR strain of NTS ( 72 . 0% versus 55 . 6% ) , [aOR = 2 . 97 , 95% CI ( 1 . 11–7 . 95 ) , P = 0 . 030] .
Sequence reads for all strains sequenced for this study have been submitted to the Short Read Archive ( SRA ) under BioProject PRJNA406975 .
|
Though NTS is normally associated with self-limiting gastroenteritis in humans , it is a leading cause of bloodstream infection in Africa . The biological mechanisms that contribute to invasiveness in NTS in Africa are unclear . In this paper we address which specific host and pathogen risk factors are associated with blood stream infection from non-typhoidal Salmonella in rural Kenya . We found that multi-drug resistant ( MDR ) strains of NTS were associated with NTS bacteremia , even after controlling for known host-factors including HIV , age , and NTS serogroup ( a taxonomic grouping ) . Our results suggest that multi-drug resistant NTS is associated with blood stream infection even in the immune-competent host . Salmonella Typhimurium sequence type ST313 , an emerging genotype in sub-Saharan Africa , was the most common cause of blood stream infection in children and adults , followed by Salmonella Enteritidis sequence type ST11 . The increasing prevalence of commonly circulating non-typhoidal Salmonella poses a major challenge to the control of highly pathogenic NTS serovars . The specific biological and epidemiological mechanisms driving invasiveness from infection with drug-resistant NTS warrant further investigation .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"tropical",
"diseases",
"microbiology",
"parasitic",
"diseases",
"salmonellosis",
"retroviruses",
"viruses",
"diarrhea",
"bacterial",
"diseases",
"immunodeficiency",
"viruses",
"signs",
"and",
"symptoms",
"rna",
"viruses",
"enterobacteriaceae",
"gastroenterology",
"and",
"hepatology",
"bacteria",
"bacterial",
"pathogens",
"salmonella",
"typhimurium",
"infectious",
"diseases",
"medical",
"microbiology",
"hiv",
"microbial",
"pathogens",
"salmonella",
"bacteremia",
"diagnostic",
"medicine",
"blood",
"anatomy",
"viral",
"pathogens",
"physiology",
"biology",
"and",
"life",
"sciences",
"malaria",
"lentivirus",
"organisms"
] |
2018
|
Multi-drug resistant non-typhoidal Salmonella associated with invasive disease in western Kenya
|
The aim of this study was to investigate the relationship between prior Anisakis infections and upper gastrointestinal bleeding ( UGIB ) , and its interaction with non-steroidal anti-inflammatory drug ( NSAID ) intake . We conducted a hospital-based case-control study covering 215 UGIB cases and 650 controls . Odds ratios ( ORs ) with their confidence intervals ( 95% CIs ) were calculated , as well as the ratio of the combined effects to the sum of the separate effects of Anisakis allergic sensitization and NSAIDs intake . Prior Anisakis infections were revealed by the presence of anti-Anisakis IgE antibodies specific to the recombinant Ani s 1 and Ani s 7 allergens used as the targets in indirect ELISA . Prior Anisakis infections ( OR 1 . 74 [95% CI: 1 . 10 to 2 . 75] ) and the intake of NSAIDs ( OR 6 . 63 [95% CI: 4 . 21 to 10 . 43] ) increased the risk of bleeding . Simultaneous NSAIDs intake and Anisakis allergic sensitization increased the risk of UGIB 14-fold ( OR = 14 . 46 [95% CI: 6 . 08 to 34 . 40] ) . This interaction was additive , with a synergistic index of 3 . 01 ( 95% CI: 1 . 18–7 . 71 ) . Prior Anisakis infection is an independent risk factor for UGIB , and the joint effect with NSAIDs is 3 times higher than the sum of their individual effects .
Upper gastrointestinal bleeding ( UGIB ) is a relatively frequent and potentially lethal multicausal medical emergency [1] . Gastric and duodenal ulcers are a major cause of UGIB , and bleeding from these lesions is frequently related to intake of non-steroidal anti-inflammatory drugs ( NSAIDs ) [2] . In countries where Anisakis infections are frequent , acute infections by this parasite may also provoke UGIB [3] . Anisakiasis is a worldwide re-emerging disease produced by the consumption of raw , lightly cooked , smoked or marinated fish containing the infective larvae of the Anisakis genus [4] , [5] . Most human cases of anisakiasis have been reported in Japan [6] , [7] , but there has been an increase in the frequency of reports of Anisakis infections in other parts of the world , such as Europe [8] , [9] , the USA , [10] , [11] and Canada [12] . Depending on the site of infection and the predominant clinical symptoms , acute infections by Anisakis can be classified as gastric anisakiasis , gastro-allergic anisakiasis , and intestinal anisakiasis . In gastric and intestinal anisakiasis , severe gastric or abdominal symptoms predominate , while in gastro-allergic anisakiasis , allergic symptoms ranging from mild urticaria to anaphylactic shock are more important [13] , [14] . However , recent evidence from seroepidemiologic studies undertaken in Spain indicates that the great majority of human cases of anisakiasis are asymptomatic , and that the prevalence of disease in different Spanish regions may range from a minimum of 0 . 4% [5] to more than 10% of the population [15] , [16] . In comparison with the healthy population , a high seroprevalence of anti-Anisakis antibodies has been reported in patients with GI bleeding [17] . However , the relevance of prior Anisakis infections as a risk factor for UGIB and its possible interaction with NSAID intake have never been investigated . We now report the results of a case–control study , which sought to determine the risk of UGIB associated with prior Anisakis simplex infections and any potential interaction with NSAID intake .
We based our study on data provided by a wider , multicenter , incident case-control study , which sought to analyze the influence of environmental and genetic risk factors on UGIB ( primary study ) . Three Spanish hospitals ( Complejo Hospitalario Universitario de Santiago de Compostela , Galicia; Hospital Clínico Universitario de Valladolid , Castilla-León; and Hospitales de Galdakao-Usansolo/Basurto , Basque Country ) that had stored serum samples for Anisakis determinations were included in the study . We defined cases as any patient admitted in the period 2003–2006 with primary diagnosis of UGIB and subsequent endoscopic diagnosis of duodenal or gastric ulcer , acute lesions of the gastric mucosa , erosive duodenitis or mixed lesions . To ensure that cases and controls come from the same source population , all patients were recruited from the same hospitals [18] . For each case , we selected 3 controls , matched by sex , age ( ±5 years ) , hospital and point in time . To avoid selection of controls being associated with exposure to NSAIDs , the controls were recruited from among patients in preoperative care for scheduled surgical interventions for non-painful processes such as cataracts , inguinal or umbilical hernias , and prostate adenomas . The enrolment criteria of the primary study ( cases and controls ) excluded patients who , at the starting date , had a history of cancer , coagulopathy , Mallory-Weiss syndrome or esophageal varices , and subjects who were not resident in the study area . Qualified , purpose-trained health staff interviewed cases and controls , after first obtaining written informed consent from the subjects . Pharmacologic anamnesis was comprehensive . Patients were first asked about any medications consumed during the two months prior to admission , and were then presented with a list of symptoms usually associated with NSAID consumption and asked whether they had taken any medication for any of these . Finally , patients who failed to remember the name of any medication were later telephoned at home to enable them to provide the name . We defined an NSAID consumer as any subject shown by pharmacological anamnesis to have consumed some medication belonging to this therapeutic group in the week preceding the index date . Subjects taking Acetylsalicylic acid at doses of less than 0 . 125 g/day ( considered to be antiaggregant ) were not considered as NSAIDs consumers . Also , NSAIDs and corticoids administered by ophthalmic , dermal or rectal route were not deemed to be exposures . We calculated a case index date , based on disease course and symptom-onset dates . For controls , the index date was the date of the interview . For cases , the consumption of medications between index and interview dates was not taken into account . Finally , we reviewed and assessed endoscopy reports of all cases according to whether or not they described detection of Anisakis larvae in the stomach or duodenum . We calculated odds ratios ( ORs ) and their adjusted confidence intervals ( CIs ) using hierarchical logistic model through a generalized linear mixed model [24] . For the purpose of constructing such models , patients were taken as level one , strata ( each case and their matched controls ) as level two , and hospitals as level three . In the estimation of the models we used the lmer function , implemented in the context of the lme4 R package ( version 2 . 7 . 2 ) [25] . To construct these models , a bivariate analysis of each independent variable was performed , and variables with 0 . 2 in the bivariate analysis were then included in the multivariate analysis . Independent variables with the highest level of statistical significance were successively eliminated from the original model , provided that the coefficients of the principal variables of exposure changed by no more than 10% and Schwarz's Bayesian Information Criterion ( BIC ) improved [26] . The confidence intervals of the interaction terms were calculated using the method proposed by Figueiras et al . [27] . The results of the generalized linear mixed model were validated by comparing them against results from comparable models obtained by running conditional logistic regression . We calculated the ratio of the combined effects to the sum of the separate effects of Anisakis and NSAID ( S ) ( along with its 95% confidence interval ) [28] as a measure of additive interaction , [18] since S has been shown to be the most reliable measure of additive interaction when adjusting for confounding [29] . The study protocol was approved by the following ethics committees: i ) Comité Etico de Investigación Clínica de Galicia; ii ) Comité Etico de Investigación Clínica de la Universidad de Valladolid; and iii ) Comité Etico de Investigación Clínica del Hospital de Galdakao-Usansolo . All cases and controls were required to give written informed consent , and where such approval was not forthcoming , the subject concerned was excluded from the study .
Sera from 215 cases and 650 controls were available for the study . The clinical signs most frequently presented by cases were: dizziness ( 58 . 1% ) ; black vomitus and/or vomiting of blood ( 40 . 5% ) , and melena ( 19% ) . With regard to controls , most of patients were recruited from the pre-operative unit for minor surgical procedures: 295 ( 45 . 4% ) for cataracts , and 168 ( 25 . 8% ) for inguinal hernias . The demographic and clinical characteristics of cases and controls are listed in Table 1 . The results showed that 54 ( 25 . 1% ) cases and 100 ( 15 . 4% ) controls were positive for Anisakis ( Ani s 1 or Ani s 7 ) , while 73 ( 33 . 9% ) cases and 76 ( 11 . 7% ) controls were NSAID consumers ( Table 2 ) . Considering IgE determinations in cases plus controls , 85 subjects ( 55 . 2% ) were seropositive for Ani 1 plus Ani s 7 , 40 subjects ( 25 . 9% ) positive for only Ani s 1 , and 29 subjects ( 18 . 8% ) positive for only Ani s 7 . As regards cases , 30 ( 55 . 5% ) patients were positive for both allergens , 16 ( 29 . 6% ) patients were positive for only Ani s 1 and 8 ( 14 . 8% ) patients were positive for only Ani s 7 . In the control group 55 ( 55 . 0% ) subjects were positive for both allergens , 24 ( 24 . 0% ) subjects were positive for only Ani s 1 and 21 ( 21 . 0% ) subjects were positive for only Ani s 7 . With regard to sex distribution in Anisakis seropositive patients ( either for Ani s 1 or Ani s 7 allergens ) , 35 cases ( 64 . 8% ) and 76 controls ( 76 . 0% ) were male . To investigate the effects of the interaction between Anisakis simplex IgE sensitization and NSAID intake on risk of UGIB , we calculated the ORs values obtained for both variables with and without interactions . The results in Table 2 , model 1 , show the ORs values without interactions . Anisakis seropositive subjects registered a 1 . 74 fold higher risk of suffering from UGIB than seronegative subjects ( 95% CI: 1 . 13–2 . 69 ) . However , when the effect of prior Anisakis infections was stratified by NSAID consumption ( Table 2 , model 2 ) , we observed that this had no effect among non-consumers of NSAIDs ( OR = 1 . 46 , [95%CI: ( 0 . 87–2 . 43] ) , but that there was a more than 14-fold higher risk of UGIB ( OR = 14 . 45 [95% CI: 6 . 46–32 . 33] ) among NSAID consumers than in Anisakis-negative non-consumers of NSAIDs . The interaction was additive , with a synergistic index of 3 . 01 ( 95% CI: 1 . 18–7 . 71 ) . Applying conditional logistic regression to these analyses provided very similar ORs , but with wider 95%CI range . The data in Table 3 show the location and type of gastrointestinal lesions observed by endoscopy in cases , with respect to Anisakis sensitization and NSAIDs intake . There were no differences in the number and location of bleeding lesions between Anisakis sensitized patients , NSAIDs consumers , or patients lacking these risk factors . In addition , the endoscopic reports revealed a single case of acute anisakiasis . This corresponded to a 53-year-old female with three Anisakis larvae penetrating two gastric ulcers located in the fornix region of the stomach . The patient presented with hematemesis accompanied by dyspepsia and pyrosis , and was seropositive for Ani s 1 and Ani s 7 allergens .
This is the first epidemiological study showing that: i ) prior Anisakis infections causing IgE sensitization are an independent risk factor for UGIB ( with an almost twofold increase in the risk ) and ii ) that this effect is modified by NSAID consumption , to the extent that the risk of UGIB can increase by more than 14 times through a synergic effect between Anisakis and NSAIDs , showing in turn that the joint effect of the two risk factors is 3 times higher than the sum of their individual effects . In the present study , a large number of subjects ( 15 . 4% in the control group ) tested positive for Ani s 1 or Ani s 7 allergens . However , the results appear to be accurate because the combined Ani s 1 and Ani s 7 ELISAs used in this study are highly sensitive and specific in comparison with other serological methods [22] . The present results are also consistent with previous reports [15] , [16] showing an extremely high seroprevalence for IgE antibodies to this parasite in the northern , central , and southern regions of Spain , where positive IgE values are observed in more than 10% of the population . Futhermore , other results showing that seropositive patients always had a prior history of ingestion of raw or undercooked fish [5] exclude the possibility that these high values were due to recognition of cross-reacting allergens present in other organisms such as mites [30] . Anisakis infections in Spain are mainly related to the ingestion of boquerones en vinagre ―pickled anchovies― [4] , [5] , although infections caused by eating undercooked fish ( e . g . hake ) have also been reported [4] , [15] , [31] . The presence of IgE antibodies in serum against specific secretory Anisakis allergens as Ani s 7 , and probably Ani s 1 , reveals that the patient has suffered one or more previous infections by this parasite [32] . However , for correct interpretation of the results , the effect of currently active and past Anisakis infections should be considered separately . In patients with active gastric anisakiasis , some of them may suffer erosions or hemorrhagic lesions of the mucosa , which can be detected by gastroscopy [4] . Bleeding during this phase can be explained by several causes , including: a ) the marked inflammatory allergic status of the mucosa , accompanied by massive infiltration of eosinophils , neutrophils , macrophages and lymphocytes in response to parasite excretory antigens [33]; b ) the direct erosive action of larvae moving into the gastric mucosa [34]; and c ) the activity of proteases [35] and anticoagulant [36] substances released by the parasite . In a recent study [22] we have observed that about 94% and 61% of symptomatic patients sensitized to Anisakis antigens have IgE antibodies to the Ani s 7 and Ani s 1 allergens , respectively . However , for patients that recognized both allergens the response to Ani s 1 was more prolonged in time . The data in the present study , showing that a considerable proportion of sera ( 25 . 9% ) were only positive to the Ani s 1 allergen , suggest that many positive IgE results are due to past , unnoticed , Anisakis infections . In addition , it was reported that gastric Anisakis infections are much more frequent than duodenal anisakiasis [37] . The similar number of bleeding ulcerous lesions observed in our study at gastric and duodenal level , and the fact that only one positive case of active anisakiasis was detected by endoscopy , also suggest that the increased risk of UGIB in Anisakis seropositive patients is not due to active infections . Unlike active anisakiasis , the implication of past Anisakis infections as a risk factor for UGIB is less evident . One could hypothesize , however , that effector molecules produced by defense cells previously activated in the GI tract in response to allergens and other Anisakis antigens , might provoke mucosal injury acting either alone or synergistically with other noxious factors present , such as NSAIDs . Candidate cells for mediating such action are eosinophils and , perhaps other pro-inflammatory cells that remain infiltrating the granulomatous tissue around the infecting larvae , or its debris , for long periods [37] . In particular , eosinophils are GI primary resident cells [38] which reportedly have immunoregulatory roles [39] and act as antigen-presenting cells in response to intestinal nematodes [40] , and the cytotoxic preformed cationic proteins that they produce upon activation are able to cause mucosal damage , as seen in some intestinal inflammatory diseases [41] , [42] . Eosinophils have also been reported to be present in the granulation tissue of perforated gastric ulcers in Japan , the country where Anisakis infections are most frequent , and the degree of infiltration by these cells was suggested to be a marker of perforation risk [43] . In this sense , it is thought that the matrix metalloproteinase-1 expressed in the cytoplasm of eosinophils may be able to digest collagen types I and III , which compose the stomach wall , and thus contribute to ulcer perforation [44] . Interestingly , NSAIDs also stimulate eosinophil production by downregulating PGE2 synthesis and upregulating production of cysteinyl leukotrienes [45] suggesting that the biochemical mechanisms whereby both risk factors potentiate UGIB may be interconnected . As in Anisakis infections , it can be hypothesized that other infectious agents causing chronic infections of the upper gastrointestinal tract such as Helicobacter pylori , or food hypersensitivity [46] , may also modify the risk of UGIB in NSAID consumers . In the present study , the observed synergism between NSAID consumption and prior Anisakis infections on the risk of UGIB were obtained from data adjusted by seroprevalence to H . pylori , thus discounting any possible bias caused by this confounding variable . Likewise , for food hypersensitivity to have a confounding effect there would have to be a positive correlation between food allergy and exposure to NSAIDs or infection by Anisakis larvae . However , this is not the case because there is no reason to think that subjects with food allergy may be more likely to consume NSAIDs or to be infected by the parasite . Finally , it shoud be noted that because of the low prevalence of consumption of the individual NSAIDs in the sample , the individual effect of the interaction of each particular NSAID with Anisakis could not be observed . Nonetheless , since all NSAIDs share the same mechanism of action , it is expected that all act synergistically with Anisakis to a greater or lesser extent . From the results of the present study we concluded that , in countries where there is a suspected presence of Anisakis infection , it would be wise to confirm whether or not the patient has a history of ingesting raw or undercooked fish before prescribing NSAIDs for long periods . For patients giving a positive response to this query , we recommend performing a parasite-specific IgE determination and conducting a closer follow-up during treatment with NSAIDs when the test is positive .
|
Anisakiasis is a worldwide re-emerging disease produced by the consumption of raw , lightly cooked , smoked or marinated fish containing live Anisakis larvae . In acute anisakiasis , mucosal lesions generated by the larvae may provoke upper gastrointestinal bleeding ( UGIB ) . However , the effect of past unnoticed Anisakis infections as a risk factor for UGIB , and a possible synergism with other risk factors such as NSAIDs intake , have never been investigated . In this case-control study we observed that: i ) prior Anisakis infections and NSAIDs intake are two independent risk factors for UGIB , and ii ) that both risk factors act synergistically to the extent that their joint effect is 3 times higher than the sum of their individual effects . We concluded that , in countries where Anisakis infections are frequent , it would be wise to determine parasite-specific IgE antibodies and to conduct a closer follow-up of patients who consume raw or lightly cooked fish and who are prescribed NSAIDs for long periods .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"public",
"health",
"and",
"epidemiology",
"parasitic",
"intestinal",
"diseases",
"clinical",
"epidemiology",
"epidemiology",
"gastroenterology",
"and",
"hepatology",
"parasitic",
"diseases",
"helminth",
"infection"
] |
2011
|
Synergism between Prior Anisakis simplex Infections and Intake of NSAIDs, on the Risk of Upper Digestive Bleeding: A Case-Control Study
|
Patients with chronic granulomatous disease ( CGD ) lack generation of reactive oxygen species ( ROS ) through the phagocyte NADPH oxidase NOX2 . CGD is an immune deficiency that leads to frequent infections with certain pathogens; this is well documented for S . aureus and A . fumigatus , but less clear for mycobacteria . We therefore performed an extensive literature search which yielded 297 cases of CGD patients with mycobacterial infections; M . bovis BCG was most commonly described ( 74% ) . The relationship between NOX2 deficiency and BCG infection however has never been studied in a mouse model . We therefore investigated BCG infection in three different mouse models of CGD: Ncf1 mutants in two different genetic backgrounds and Cybb knock-out mice . In addition , we investigated a macrophage-specific rescue ( transgenic expression of Ncf1 under the control of the CD68 promoter ) . Wild-type mice did not develop severe disease upon BCG injection . In contrast , all three types of CGD mice were highly susceptible to BCG , as witnessed by a severe weight loss , development of hemorrhagic pneumonia , and a high mortality ( ∼50% ) . Rescue of NOX2 activity in macrophages restored BCG resistance , similar as seen in wild-type mice . Granulomas from mycobacteria-infected wild-type mice generated ROS , while granulomas from CGD mice did not . Bacterial load in CGD mice was only moderately increased , suggesting that it was not crucial for the observed phenotype . CGD mice responded with massively enhanced cytokine release ( TNF-α , IFN-γ , IL-17 and IL-12 ) early after BCG infection , which might account for severity of the disease . Finally , in wild-type mice , macrophages formed clusters and restricted mycobacteria to granulomas , while macrophages and mycobacteria were diffusely distributed in lung tissue from CGD mice . Our results demonstrate that lack of the NADPH oxidase leads to a markedly increased severity of BCG infection through mechanisms including increased cytokine production and impaired granuloma formation .
M . bovis BCG ( Bacillus Calmette Guérin ) is an attenuated strain of M . bovis , used as a vaccine against tuberculosis . BCG vaccination has a proven efficacy only early in life ( <1 year of age ) , in particular against tuberculous meningitis and miliary tuberculosis . Thus , the WHO recommends vaccination of newborns in endemic areas [1] . However , BCG is a live vaccine , which may persist and become a pathogen . In some individuals , in particular those with immune defects , BCG vaccination may lead to severe local or to disseminated infection [2] , [3] . BCG is also used as local treatment for bladder cancer [4] , where in some cases it may lead to symptomatic infection , from cystitis to life threatening dissemination [5] . However , there is emerging evidence for increased risk of BCG infection in patients lacking the phagocyte NADPH oxidase ( chronic granulomatous disease , CGD ) [6]–[8] . Indeed studies looking at underlying risk factors in patients presenting BCG infection suggest that approximately 20% of such patients suffer from CGD [9] and in many instances , BCG infection is the first manifestation of CGD [10] . The phagocyte NADPH oxidase NOX2 is a superoxide producing enzyme , involved in the host defense against numerous bacteria and fungi . Genetic loss of function of NOX2 is a primary immunodeficiency referred as chronic granulomatous disease ( CGD ) . CGD may be caused by mutations in the gp91phox/NOX2 protein which is coded by the Cybb gene or one of its subunits , in particular p47phox , which is coded by the Ncf1 gene [11] . CGD patients suffer from severe and recurrent bacterial and fungal infections as well as from hyperinflammatory and autoimmune diseases in particular discoid lupus [12] . Until about 10 years ago , it was thought that the phagocyte NADPH oxidase was not relevant for the defense against mycobacteria [13] . Whether mice carrying CGD mutations show an increased susceptibility to infection with Mycobacterium tuberculosis remains controversial [14] , while their susceptibility to BCG infection has so far not been studied . Host defense mechanisms against mycobacteria are typically initiated by phagocytosis through macrophages , inducing inflammation and subsequently cell-mediated immunity involving Th1-type immune responses . These coordinated mechanisms result in granuloma formation . Granulomas are highly organized structures generated by interactions between myeloid and lymphoid cells that characterize the adaptive immune response to mycobacteria . In general granulomas sequester mycobacteria and thereby limit their dissemination . Granulomas are formed through cellular recruitment and are associated with production of cytokines and chemokines [15] . Among these cytokines , TNF and IFN-γ are the main players contributing to activation of macrophage host defense mechanisms [16] . Neutrophils are able to kill mycobacteria in vitro , but the in vivo relevance of neutrophils in the mycobacterial host defense remains a matter of debate [17] . Here we have first analyzed the relevance of BCG infection in CGD patients and then investigated the role of NADPH oxidase-generated ROS in experimental BCG infection . Mice lacking a functional phagocyte NADPH oxidase showed a markedly enhanced severity to BCG infection . Rescue of phagocyte NADPH oxidase function in macrophages was sufficient to reverse the phenotype to the mild disease observed in wild-type mice . We identified increased cytokine generation and poorly organized granuloma formation as mechanisms involved in the exacerbated severity of BCG infection in NADPH oxidase-deficient mice .
Mice were injected intravenously with BCG ( 107 CFU ) . Wild-type mice resisted the infection during the 4 week observation . In contrast , Ncf1 mutant mice showed early mortality: 50% of mice died after 10 days and only 33% of the mice survived after 4 weeks ( Figure 2A ) . The high mortality of Ncf1 mutant mice was associated with a rapid weight loss , which was absent in wild-type controls ( Figure 2B ) . To investigate whether the genetic background or the type of CGD mutation was responsible for the high susceptibility , we investigate other types of CGD mice . BCG-infected Ncf1 mutant mice in a C57Bl/6 background showed also a higher mortality as compared with their wild-type controls ( Figure 2C ) . Similarly to Ncf1 mutant mice in C57Bl/10 . Q background , Ncf1 mutant mice in a C57Bl/6 background showed a rapid weight loss ( Figure 2D ) . However the median survival time was around 20 days in the C57Bl/6 background , as opposed to ∼8 days in the C57Bl/10 . Q background which suggest a contribution of the mouse genetic background to BCG susceptibility . Given that the survival trends in the different genetic backgrounds appeared to be similar , suggesting a minor role for epistasis and a major role of the NOX2 subunit mutation , BCG infection was studied in Ncf1 rescue mice that we have previously characterized [31] , [32] . Ncf1 rescue as wild-type mice resisted the infection during the 4 weeks observation and no mortality or no important weight loss was observed ( Figure 2A and B ) . Finally , BCG infection in Cybb-deficient mice also led to a high mortality and weight loss as compared to wild type controls ( Figure 2E and F ) . These observations strongly suggest that in absence of ROS production by NADPH oxidase ( see below ) , mice are more susceptible to mycobacterial infection . Furthermore , the normal survival of Ncf1 rescue mice implies that ROS production in mononuclear phagocytes is crucial . To further understand the causes of the early mortality of Ncf1 mutant mice , mice were sacrificed at day 3 post-infection for lung histopathological examination . In the absence of BCG infection , no histological differences were observed between wild-type , Ncf1 mutant and rescue mice ( Figure S1 ) . However , upon BCG infection , Ncf1 mutant mice presented severe inflammatory lesions in the lungs with extended hemorrhagic lesions , intravascular thrombosis , decrease of the alveolar spaces ( Figure 3A . b ) and hypertrophy of pleural cells ( Figure 3B . b ) . Ncf1 mutant mice also showed accumulation of inflammatory cells composed essentially by neutrophils concentrated as microabscesses ( Figure 3A . e ) . In contrast , wild-type and Ncf1 rescue did not show massive hemorrhagic lesions and only moderate inflammation with mixed inflammatory cells observed in lungs ( Figures 3A . a , c , d and f ) . The surviving Ncf1 mutant mice ( 5 out of 15 ) were also analyzed at 4 weeks post-infection . Histopathological examination of Ncf1 mutant lungs revealed extensive inflammatory lesions reducing notably the alveolar space ( Figure 3C . b ) and microabscesses composed of neutrophils . Only one third of the Ncf1 mutant mice survived up to 4 weeks and the latter results might represent a survivor effect , and not necessarily be representative for all Ncf1 mutant mice . In contrast , wild-type and Ncf1 rescue did not show massive infiltrate of inflammatory cells ( Figures 3C . a–c and d–f ) . In the absence of BCG infection , the organ weight indexes for lung , liver and spleen were comparable in all mouse strains ( Figure 4A and Figure S1 ) . After BCG infection , lung weight , as a surrogate measure of lung inflammation and edema , increased only moderately in wild-type and Ncf1 rescue mice , but massively in Ncf1 mutant mice ( Figure 4A ) . At 4 weeks of BCG infection , the severity of lung pathology was also assessed by analysis of free alveolar space vs . occupied space . The occupied space was significantly increased in Ncf1 mutant as compared to wild-type and Ncf1 rescue lungs ( Figure 4B ) . This quantification corroborates with the massive obstruction of alveolar space in Ncf1 mutant mice seen in histology . Similar as seen for Ncf1 mutant mice , BCG-infected Cybb-deficient mice showed severe inflammatory lesions with extended hemorrhagic lesions and decreased alveolar space ( Figure S2 . A and B . b ) . Microabscesses of neutrophils were also present in Cybb-deficient and Ncf1 mutant in C57Bl/6 background in the lung of both sacrificed and deceased mouse while mixed inflammatory cells were observed in Ncf1 rescue and wild-type lungs ( Figure S2 . B . d and C ) . The lung weight was also increased in Cybb-deficient and Ncf1 mutant in C57Bl/6 background mice compared to wild-type , 4 weeks after BCG infection ( Figure S4 . A ) . In summary , these data support that impaired ROS production by mononuclear phagocytes is associated with increased inflammatory response involving neutrophilic microabscess formation following BCG infection . To evaluate mechanisms how ROS production protect from death by BCG infection , we assessed bacterial burden by colony forming unit quantification in different organs . Bacterial load in the lung of both the Ncf1 mutant and Ncf1 rescue mice were similar but higher than those in wild-type mice 3 days after infection ( Figure 4C ) . Four weeks post-infection , no significant differences in bacterial load were observed in lung and liver . However , bacterial counts in the spleen of Ncf1mutant and Ncf1rescue mice were significantly increased compared to wild-type ( Figure 4C ) . We further evaluated inducible nitric oxide synthase iNOS by western blot , which is crucial for clearance of BCG and mouse survival [34] . Expression of iNOS protein in the lung at 4 weeks post-infection was significantly increased in Ncf1 mutant as compared to wild-type mice ( Figure 4D ) . NO and superoxide form peroxynitrite , a highly reactive molecule , are implicated in mycobacteria killing . Upon interaction with proteins , peroxynitrite produces nitrotyrosine , which are stable biological peroxynitrite markers [35] . Interestingly , despite the increase of iNOS protein , nitrotyrosine levels measured by ELISA in lungs of Ncf1 mutant were not different from wild-type ( Figure 4D ) , presumably because Ncf1 mutant mice lack the second substrate required for peroxynitrite generation , namely superoxide . Thus , most likely CGD mice produce increased amounts of NO in response to mycobacteria , but given the lack of NOX2-generated superoxide , this is not accompanied by an increase in peroxynitrite . These results are compatible with the concept that peroxynitrite , rather than NO or ROS , is crucial for optimal mycobacterial killing . We next measured levels of selected cytokines by ELISA in lung homogenates from BCG infected mice ( Figures 5 ) . Three days post infection , the increase of TNF levels in Ncf1 mutant lung was massive ( 3 . 6-fold compared to wild-type ) and also observed at four weeks post-infection ( Figure 5A ) . TNF levels in Ncf1 rescue lung were comparable to those observed in wild-type lung . Three days , but not 4 weeks , post infection , IL-17 lung levels were increased in Ncf1 mutant mice ( Figure 5B ) . The same pattern was observed for IL-12p40 ( Figure 5C ) . The pattern was slightly different for IFN-γ: there were increased in Ncf1 mutant mice at 3 days and 4 weeks post-infection , however Ncf1 rescue mice showed even higher IFN-γ levels 4 weeks post-infection ( Figure 4D ) . We also assessed the levels of the chemokines CXCL1 ( KC , the murine IL-8 homolog ) , and CCL5 ( RANTES ) . We selected CXCL1 because it is a powerful neutrophil chemoattractant [36] and might explain the high number of neutrophils in the lung lesion in mutant mice and CCL5 because it is a leukocyte chemoattractant with a role in mycobacterial protection [37] . Ncf1 mutant mice showed a higher CXCL1 levels 3 days after BCG infection ( Figure 5E ) . CCL5 levels were increased in Ncf1 mutant mice three days and 4 weeks after infection ( Figure 5F ) . The general pattern was an increase in pulmonary cytokine and chemokine responses in Ncf1 mutant mice due to the infection which was controlled by Ncf1 rescue mice . Moreover , we also evaluated if Cybb-deficient mice would also respond with an exacerbated cytokine response using ex-vivo recall of spleen cells from BCG infected mice . Both re-infection of splenocytes or addition of BCG antigens resulted in enhanced TNF and nitrite , as an indicator of NO production , evaluated respectively by ELISA and Griess reagent , confirming that NOX2 deficiency leads to an increase response in TNF and immune mediators ( Figure S3 ) . Thus , at early time points a massive increase of pro-inflammatory cytokines was observed in BCG-infected CGD mice . At later time points , only TNF and CCL5 levels remained elevated , this had a probable importance for altered granuloma formation ( see below ) . We have previously demonstrated that ROS production in response to phorbol myristate acetate ( PMA ) or β-glucan was abolished in neutrophils , bone-marrow derived macrophages ( BMDM ) and dendritic cells ( BMDC ) in Ncf1 mutant mice [29] . We therefore investigated ROS production measured by amplex red in BMDM and BMDC exposed to BCG . Wild-type and Ncf1 rescue cells produced ROS in response to BCG , but not Ncf1 mutant cells ( Figures 6A , B ) . Diphenylene iodonium ( DPI ) , a non-specific NOX inhibitor , abolished the mycobacteria-induced ROS production in wild-type and Ncf1 rescue cells . The kinetic of ROS production was comparable in BMDM and BMDC from wild-type and Ncf1 rescue mice ( Figures 6A and B ) . We next investigated whether there are signs of ROS production in granulomas in vivo . For this purpose , liver sections from 4 weeks BCG-infected mice were stained with an antibody against 8-OHdG ( 8-hydroxydeoxyguanosine ) , a well-studied marker of DNA oxidation [38] . In wild-type and Ncf1 rescue liver , an important 8-OHdG staining was observed within granulomas ( Figure 6C ) . Note that , to the best of our knowledge , this is the first demonstration of ROS generation during granuloma formation . Importantly , no 8-OHdG staining was observed in granulomas of Ncf1 mutant mice , demonstrating that the phagocyte NADPH oxidase is the major source of ROS during BCG infection . Granuloma formation is a crucial mechanism to control mycobacterial infection . To determine the relationship between ROS production and granuloma formation , we next analyzed lung histology 3 days and 4 weeks after BCG infection using the following stainings: H/E ( general morphology ) , Ziehl-Neelsen ( mycobacteria ) , acidic phosphatase activity ( activated macrophages ) . Three days post BCG infection , lung sections from wild-type and Ncf1 rescue mice showed clusters of macrophages ( Figure 7A ) . In Ncf1 mutant mice , BCG infection induced abundant neutrophil abscesses , with a lack of macrophage clustering within restricted areas ( Figure 7B ) . Mycobacteria appeared less abundant in lung section from wild-type as compared to Ncf1 mutant mice ( Figure 7C ) . Interestingly , the Ziehl-Neelsen stain suggests a relatively high bacterial load in rescue mice , corroborating the quantitative bacterial load analysis ( Figure 4C ) . Granuloma formation is a crucial step in the mycobacterial containment and clearance . After 4 weeks of BCG infection , both wild-type and Ncf1 rescue mice had well differentiated granulomas containing multinucleated giant cells ( Figures 7D–E ) . In these mice , granulomas enclosed the mycobacteria and virtually no mycobacteria were observed outside of granulomas ( Figure 7F ) . In contrast , Ncf1 mutant mice presented large pyogranulomatous lesions with abundant neutrophil abscesses ( Figure 7D ) and diffusely distributed acid phosphatase-positive macrophages ( Figure 7E ) . Importantly , no sequestration of mycobacteria was observed in Ncf1 mutant mice ( Figure 7F ) . As seen for Ncf1 mutant mice , Cybb-deficient mice as well as Ncf1 mutant mice in a C57Bl/6 background showed larger granulomas without concise delimitations ( Figure S4B ) . Disorganized granulomas with an abnormal presence of neutrophils were mainly in lung but there are also observed in the liver and spleen of Ncf1 mutant mice ( data not shown ) . Thus , the presence of NADPH oxidase in mononuclear phagocytes is required for the formation of compact granulomas with concise delimitations and for sequestration of mycobacteria within granulomas .
Particularly interesting in this respect is the recent discovery of a family with a peculiar variant of CGD [32] . These patients lack ROS production in macrophages , but not in neutrophils and showed a high sensitivity to mycobacterial infection , in particular to BCG . The Ncf1 mutant mice used in this study , including a selective rescue in mononuclear phagocytes , provide a mirror image of the patient study [31]: selective rescue of NOX2 in macrophages protected CGD mice against BCG infection . Indeed , most of the enhanced mycobacterial pathology associated with NOX2-deficiency ( morbidity , mortality , enhanced cytokine production , abnormal granuloma formation ) could be attributed to macrophages . There is one exception to this: the moderately increased mycobacterial load , which is not reversed by Ncf1 rescue in macrophages and hence was not correlated to the outcome of infection . Thus , while our results in mice show that selective rescue of NOX2 in macrophages restores resistance BCG infection; in the above mentioned CGD patients a selective loss of NOX2 in macrophages establishes high susceptibility . Hitherto , BCG infection has never been investigated in mouse models of CGD . However , infection of CGD mice with other types of mycobacteria led to discordant results: in some cases aggravation was observed , while in other studies no effect was observed . We wanted to assure that our results are not due to a specific choice of the CGD mutation or to the genetic background . We therefore tested two different CGD mutations ( Ncf1 , Cybb ) as well as two different genetic backgrounds ( C57/B10 . Q , C57Bl/6 ) . All results concur: CGD mice are highly susceptible to BCG infection . Reconstitution of the phagocyte NADPH oxidase in mononuclear phagocytes completely reversed the neutrophil influx phenotype . Thus , it is not the lack of activity of NOX2 in neutrophils which leads to the increased number of neutrophils in inflammation . Most likely , NOX2 in mononuclear phagocytes regulates the number of invading neutrophils by controlling the release of neutrophil chemoattractants . These chemoattractants might be directly released from macrophages or possibly from other cells that depend on a macrophage signal . An alternative theory is the decreased uptake of apoptotic neutrophils by NADPH oxidase-deficient macrophages [39] . Enhanced neutrophil infiltration in CGD mice might be involved in the enhanced TNF production , thereby possibly contributing to the enhanced mortality in CGD mice . Despite enhanced levels of iNOS , our results show a small , increase of mycobacterial load in Ncf1 mutants . This suggests that the well-documented bactericidal activity of iNOS-dependent NO production is several impaired in the absence of NOX2 , compatible with the suggested role of peroxynitrite ( i . e . the reaction product of NO and superoxide ) in mycobacterial killing . Note however that NADPH-oxidase involvement in killing of mycobacteria does not necessarily signify a direct antibacterial action . For example , mycobacteria have been suggested to be sensitive to neutrophil extracellular trap ( NET ) [40] and NADPH oxidase-dependent NET formation [41] could also be a relevant mechanism limiting the multiplication of mycobacteria . It has been suggested that the increased sensitivity of CGD patients to mycobacterial infection might be linked to a ROS activation of cytokine production , in particular IL-12 ( which is secreted by macrophages to stimulate IFN-γ release by T lymphocytes [3] ) . In CGD patients , an hyperresponsiveness of neutrophils to different stimuli was usually observed [42] . In our study , we observed the opposite: CGD mice infected with BCG generated increased levels of cytokines . Interestingly , several of the cytokines increased in Ncf1 mutant mice ( in particular TNFα , IL-12 , IFN-γ , IL-17 ) are involved in the antimycobacterial defense . This might be a defense mechanism compensating for the lack of ROS generated by the NADPH oxidase . However , the high level of certain cytokines , in particular TNF , in CGD mice might also account for the high early mortality and absence of resolution of inflammation to mycobacterial infections . Our results shed new light on granuloma formation in mycobacterial infection and the role of NOX2 in this process: Taken together , our results provide strong evidence for a role of NADPH oxidase-dependent ROS generation in the fine tuning of granuloma formation . Thus , redox-sensitive signaling steps are involved in the coordinated genesis of granulomas , and the overshooting cytokine and chemokine productions observed in CGD mice probably destabilizes granulomas . To which extend do results obtained in our study apply to CGD patients ? Clearly , our analysis of the published literature demonstrates that , human CGD patients are sensitive to infection with the vaccinal BCG strain [6] , [8] . Approximately 15% of CGD patient with BCG disease will develop a disseminated form , also referred to as BCGosis . At this point , it is not clear which are the factors precipitating such disseminated disease . Genetic modifiers , type of BCG strain , inoculum size of viable mycobacteria are among the possible culprits . Similar as observed in our mouse model of disseminated BCG infection , there was a substantial mortality associated with BCGosis in CGD patients . Taken together , the results presented here not only shed new light on BCG infection in CGD , but also provide first evidence for a role of the macrophage NADPH oxidase in the coordination of granuloma formation . The vaccinal BCG strain is an important tool for the control of childhood tuberculosis in countries with a high incidence of the disease . In general , children are vaccinated at birth , because the major effect of BCG vaccination is prevention from tuberculous meningitis early in life . Thus , the vaccination occurs prior to first manifestations of immune deficiency . New algorithms need to be defined to assure vaccine protection of immunocompetent neonates , without putting immunodeficient neonates at risk .
Animal experiments complied with ethical standards of the University of Geneva and the Cantonal Veterinary Office ( Authorization No . 1005/3715/2 ) . Handling and manipulation of the animals complied with European Community guidelines . Wild-type B10 . Q , Ncf1 mutant and rescue mice , backcrossed into identical background were used ( for details of backcross see [31] , [43] ) . Ncf1 rescue mice are Ncf1 mutant animals which contain a transgenic wild-type Ncf1 gene under the control of a human CD68 promoter fragment . Ncf1 mutant with the same mutation on a C57Bl/6N background and its respective wild-type controls were used . Cybb-deficient mice and respective controls were backcrossed on C57Bl/6 background ( Jackson Laboratories ) . For all experiments , mice aged 8–12 weeks were kept in a quiet room at 25°C with a 12 h light/dark cycle and food and water were supplied ad libitum . Mice were infected intravenously with 107 living CFU of M . bovis BCG Connaught [44] , [45] . Mortality and body weights were monitored during infection . Three days and 4 weeks post-infection , mice were sacrificed and lung , liver and spleen were weighted , fixed and frozen for subsequent analyses . The number of viable bacteria recovered from frozen organs was evaluated as previously described [46] , [47] . Bone marrow primary cells were obtained from mice by flushing both the femur and the tibia as previously described [29] , [48] BMDMs and BMDCs were stimulated with BCG ( MOI 10 ) . The production of ROS by NOX2 was measured using Amplex red ( Invitrogen ) fluorescence , as described previously [49] . Lung homogenates were prepared and western blot performed as previously described [50] . Nitrotyrosine , a stable end product of peroxynitrite oxidation , was assessed in serum by enzyme-linked immunosorbent assay ( ELISA; Hycult biotechnology , Netherlands ) . Histologic analyses of lung lesions were performed at 3 days and 4 weeks after infection . Lungs embedded in paraffin for hematoxylin/eosin ( HE ) and Ziehl-Neelsen stainings . For acid phosphatase staining , cryostat tissue sections from lung frozen in liquid nitrogen were used as previously described [51] . Signs of ROS production were evaluated by 8-hydroxy-2′-deoxyguanosine ( 8-OHdG ) staining ( 1∶50 , JaICA , Shizuoka , Japan ) as previously described [52] . Evaluation of the histopathology was performed on three lung lobe sections per animal ( n = 4/group ) . Lung sections were captured on Zeiss Mirax Scan microscope system . Virtual sections were subdivided and images covering lobe sections corresponding to a surface of 21 . 50±8 . 05 mm2 per mouse , were proceeded for quantification of free space and occupied lung tissue using a specific program designed in the Metamorph software identifying cellularity , hematoxilin-eosin stain and air spaces [53] . Mice were infected with BCG , sacrificed at day 17 and spleen cells were prepared as previously described [45] . Cells were stimulated with either medium alone , living BCG ( 103 CFU/well ) , or BCG culture protein extracts ( 17 µg/ml ) . After one , three and six days of treatment , medium was harvested for nitrite and TNF determination . Nitrite accumulation , as an indicator of NO production , was evaluated by Griess reagent ( 1% sulfanilamide and 0 . 1% naphtylethylenediamide in 2 . 5% phosphoric acid ) . TNF was determined in cell supernatants as described below . Lungs were collected at different time points after BCG injection and tissue homogenate was prepared [54] . Cytokines and chemokines were measured by ELISA ( Ready&D System ) . Literature research on CGD and mycobacterial infections was done from PubMed and Google Scholar with no limitations in time . Parametric ( t -tests ) and non-parametric ( One-way analysis and Kruskal–Wallis ) tests were used . In the case of multiple comparisons , a two-way ANOVA test with Bonferroni correction was used .
|
The vaccine Mycobacterium bovis BCG is administrated to prevent early age tuberculosis in endemic areas . BCG is a live vaccine with a low incidence of complications . However , local or disseminated BCG infection may occur , in particular in immunodeficient individuals . Chronic granulomatous disease ( CGD ) , a deficiency in the superoxide-producing phagocyte NADPH oxidase , is a primary immune deficiency and one of the most frequent congenital defects of phagocyte in humans . Here we analyze the role of the phagocyte NADPH oxidase NOX2 in the defense against BCG . An extensive literature review suggested that BCG infection is by far the most common mycobacterial disease in CGD patients ( 220 published cases ) . We therefore studied BCG infection in several CGD mouse models showing that these were highly susceptible to BCG infection with a mortality rate of ∼50% . As compared to the wild type , CGD mice showed a markedly increased release of cytokines , an altered granuloma structure , and were unable to restrain mycobacteria within granulomas . Rescue of the phagocyte NADPH oxidase in macrophages was sufficient to protect mice from BCG infection and to sequester the mycobacteria within granulomas . Thus , superoxide generation by macrophages plays an important role for the defense against BCG infection and prevents overshooting release of proinflammatory cytokines .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"animal",
"models",
"of",
"infection",
"infectious",
"diseases",
"mycobacteria",
"medicine",
"and",
"health",
"sciences",
"medical",
"microbiology",
"microbial",
"pathogens",
"biology",
"and",
"life",
"sciences",
"immunology",
"microbiology",
"bacterial",
"pathogens",
"pediatrics"
] |
2014
|
Bacillus Calmette-Guerin Infection in NADPH Oxidase Deficiency: Defective Mycobacterial Sequestration and Granuloma Formation
|
The biogenesis of bacterial cell-envelope polysaccharides requires the translocation , across the plasma membrane , of sugar sub-units that are produced inside the cytoplasm . To this end , the hydrophilic sugars are anchored to a lipid phosphate carrier ( undecaprenyl phosphate ( C55-P ) ) , yielding membrane intermediates which are translocated to the outer face of the membrane . Finally , the glycan moiety is transferred to a nascent acceptor polymer , releasing the carrier in the “inactive” undecaprenyl pyrophosphate ( C55-PP ) form . Thus , C55-P is generated through the dephosphorylation of C55-PP , itself arising from either de novo synthesis or recycling . Two types of integral membrane C55-PP phosphatases were described: BacA enzymes and a sub-group of PAP2 enzymes ( type 2 phosphatidic acid phosphatases ) . The human pathogen Helicobacter pylori does not contain BacA homologue but has four membrane PAP2 proteins: LpxE , LpxF , HP0350 and HP0851 . Here , we report the physiological role of HP0851 , renamed HupA , via multiple and complementary approaches ranging from a detailed biochemical characterization to the assessment of its effect on cell envelope metabolism and microbe-host interactions . HupA displays a dual function as being the main C55-PP pyrophosphatase ( UppP ) and phosphatidylglycerol phosphate phosphatase ( PGPase ) . Although not essential in vitro , HupA was essential in vivo for stomach colonization . In vitro , the remaining UppP activity was carried out by LpxE in addition to its lipid A 1-phosphate phosphatase activity . Both HupA and LpxE have crucial roles in the biosynthesis of several cell wall polysaccharides and thus constitute potential targets for new therapeutic strategies .
The biogenesis of many bacterial cell-envelope polysaccharides ( i . e . , peptidoglycan ( PGN ) , lipopolysaccharides ( LPS ) , teichoic acids , enterobacterial common antigen ) requires the translocation , across the cytoplasmic membrane , of glycan units that are produced inside the cytoplasm [1] . Therefore , the hydrophilic sugars must be anchored to a lipid carrier ( undecaprenyl phosphate ( C55-P ) ) , yielding membrane intermediates which are translocated to the outer face of the membrane [2] . In PGN biosynthesis , these intermediates are finally cross-linked by transglycosylase and transpeptidase activities to a nascent acceptor polymer . These polymerization reactions release the lipid carrier in an “inactive” undecaprenyl pyrophosphate form ( C55-PP ) which must be recycled to participate in new rounds of cell-envelope polysaccharides biosynthesis . C55-P originates from the dephosphorylation of C55-PP , itself arising from either ( i ) cytoplasmic de novo synthesis by condensation of eight isopentenyl pyrophosphate ( C5-PP ) molecules with one farnesyl pyrophosphate ( C15-PP ) catalyzed by the essential C55-PP synthase ( UppS ) [3] or ( ii ) recycling [4] when it is released at the periplasmic side of the membrane . Two unrelated families of integral membrane proteins exhibiting C55-PP phosphatase ( UppP ) activity were identified: BacA and members of the PAP2 ( type 2 phosphatidic acid phosphatase ) super-family . Escherichia coli cells possess four UppPs: BacA enzyme which accounts for 75% of UppP activity and three PAP2 enzymes ( PgpB , YbjG and LpxT ) which ensure the remaining activity [5] . The plurality of UppPs as observed in E . coli and Bacillus subtilis [6] seems to be shared by most of the bacteria as suggested by a search for homologues , raising the question of the role of such a multiplicity . The simultaneous inactivation of bacA , ybjG and pgpB is lethal in E . coli whereas any single or double deletions had no effect on bacterial growth . Overexpression of BacA , PgpB or YbjG resulted in bacitracin resistance , and an increase of the UppP activity contained in membrane extracts [7] . Bacitracin is an antibiotic produced by Bacillus licheniformis [8] which strongly binds C55-PP , thereby inhibiting its dephosphorylation and leading to an arrest of PGN biosynthesis . When overexpressed , the UppP enzymes likely compete with the bacitracin for C55-PP binding , thus favoring its dephosphorylation . LpxT was not able to sustain growth of the triple bacA-ybjG-pgpB mutant and its overexpression did not lead to any bacitracin resistance suggesting that LpxT displays another function . Nevertheless , LpxT was shown to catalyze the transfer of C55-PP distal phosphate group onto lipid A , the lipid moiety of LPS , yielding C55-P and a pyrophosphorylated form of lipid A [9] . In addition to its UppP activity , PgpB is involved in phospholipids biosynthesis via the hydrolysis of phosphatidylglycerol phosphate ( PGP ) to form phosphatidylglycerol ( PG ) [10] . In E . coli , PgpA and PgpC are two additional integral membrane enzymes sharing the latter function . It has been shown that the co-inactivation of the three PGP phosphatases leads to a lethal phenotype [11] . Topology and structural analyses of PAP2 enzymes from E . coli and B . subtilis showed that these enzymes exhibit their active site at the interface between the plasma membrane and the periplasmic space , suggesting they are rather involved in C55-PP recycling [6 , 12 , 13] . More recently , the structure of BacA from E . coli was also resolved [14 , 15] showing here again an enzyme with its active site residues accessible from the periplasmic side . Nevertheless , the unique structure of BacA was reminiscent of that of transporters or channels raising the possibility that BacA may have alternate active sites on either side of the membrane and/or may function as a flippase allowing complete recycling of C55-P . Helicobacter pylori is a microaerophilic , spiral-shaped , flagellated , Gram-negative bacterium that colonizes the human’s gastric mucosa [16] . This pathogen is responsible of chronic gastritis and peptic ulcers [17] and is a risk factor for gastric cancer . It has been classified as a class I carcinogen by the World Health Organization in 1994 . In contrast to E . coli , H . pylori does not contain a BacA protein but has four PAP2 enzymes: HP0021 ( LpxE ) , HP0350 , HP0851 and HP1580 ( LpxF ) . LpxE and LpxF have been shown to be involved in LPS modifications . The lipid A structure of H . pylori is unique and constitutively modified . LpxE is responsible for the removal of the lipid A 1-phosphate group that is required for the further addition of a phosphoethanolamine ( PE ) group at the same position [18] . LpxF is a lipid A 4’-phosphate phosphatase . Both modifications increase the net positive charge of the lipid A , which then confers cationic antimicrobial peptide ( CAMP ) resistance and escape to the host innate immune system , thereby allowing H . pylori to survive in the gastric mucosa [19] . The aim of this study was to decipher the physiological role of the multiple PAP2 enzymes from H . pylori . We describe that HP0851 , renamed HupA , is the main UppP but is also involved in phospholipid biosynthesis by catalyzing the dephosphorylation of PGP in PG . In addition , we show that HupA has a role in CAMP resistance and is essential for colonization of the mouse stomach . In a global context of increasing resistance to antibiotics , finding new potential therapeutic targets is crucial especially against H . pylori , which colonizes half of the world’s population and is classified by the World Health Organization as a priority 2 pathogen regarding antibiotic resistance . A better understanding of essential metabolic pathways , and of the enzymes involved in such processes , could lead to the development of new antibacterials .
To characterize the PAP2 enzymes from H . pylori ( LpxE , HP0350 , HP0851 and LpxF ) , the corresponding genes were cloned on the pTrcHis30 expression vector under the control of a strong IPTG-inducible promoter ( Table 1 ) . The recombinant N-terminally His-tagged proteins were overproduced in E . coli C43 ( DE3 ) cells . The integral membrane proteins were extracted from membranes via their solubilization with n-dodecyl-β-D-maltoside ( DDM ) detergent and the PAP2 proteins were purified by affinity chromatography using Ni2+-NTA agarose beads ( Materials and Methods ) . Due to the difficulty in getting high expression levels of these membrane proteins in E . coli , they could not be purified to homogeneity . Nevertheless , LpxE , HP0350 and HP0851 were enriched to a high level as judged from SDS-PAGE analysis ( Fig 1A ) , while LpxF could not be visualized by Coomassie blue staining . However , LpxF was in the purified samples as confirmed by western blot analysis ( Fig 1B ) . We then measured the UppP activities present in these purified solutions . Considering the niche of H . pylori , which ranges from the human stomach with an acidic pH to the epithelial interphase with a neutral pH , we measured enzymatic activities on a large range of pHs ( Fig 1C ) . The uncharacterized HP0350 protein did not catalyze C55-PP dephosphorylation even under prolonged incubation time and high protein concentration . LpxE displayed an UppP activity of 900 nmol/min/mg at an optimal pH of 7 . 4 , while HP0851 exhibited a 6 . 7-fold higher UppP activity of 6039 nmol/min/mg at an optimal pH of 5 . LpxE displayed UppP activity in vitro in addition to its role as lipid A 1-phosphate phosphatase [19] . Considering the highest activity of HP0851 , this protein may account for a large part of C55-P ( re ) generation in vivo . LpxF was shown to act as a lipid A 4’-phosphate phosphatase [19] . In this study , we also detected a low UppP activity for LpxF ( Table 2 and Fig 1C ) . This activity is unlikely to arise from contaminants since HP0350 , which was purified using the same methodology , exhibited no activity as compared to LpxF . To further address the contribution of each PAP2 protein in the global UppP activity in H . pylori , we generated the four corresponding single mutants in H . pylori N6 . All four mutants were readily obtained showing that none of these proteins is essential for survival of H . pylori in vitro . We prepared DDM-solubilized membrane extracts from exponentially growing wild-type and mutant cells and we measured the residual UppP activity present in these extracts ( Fig 2 ) . The UppP activities present in membranes of hp0350 , lpxE and lpxF single mutants and wild-type strain were very similar . In contrast , the UppP activity decreased in hp0851 membranes to about 10% of residual activity as compared to the wild-type . Thus , these data confirmed the major contribution of HP0851 in C55-PP recycling in C55-P . The simultaneous inactivation of bacA , pgpB and ybjG is lethal in E . coli . The BWTsbacA strain is a thermosensitive conditional triple mutant ( ΔbacA , ΔybjG , ΔpgpB ) containing an ectopic copy of bacA on a plasmid whose replication is impaired at 42°C [7] . This mutant accumulates soluble PGN precursors and lyses after a shift from 30°C to 42°C , due to the depletion of the pool of C55-P that arrests cell-wall synthesis . The ability of the PAP2 proteins from H . pylori to restore the growth of BWTsbacA at the restrictive temperature was tested using the pTrcHis30-based plasmids previously used for protein purification ( Table 3 ) . The hp0350 gene was unable to complement BWTsbacA strain even in the presence of 1 mM IPTG . The lpxF gene was also unable to complement in the absence of inducer and was found to be toxic in the presence of IPTG as no growth was observed either at 30°C or at 42°C . In contrast , lpxE and hp0851 genes complemented BWTsbacA without the need of inducer , indicating that a basal level of the corresponding proteins allows a supply of C55-P that is appropriate for optimal growth of E . coli . The overproduction of LpxE and LpxF in E . coli , due to the addition of IPTG , was lethal possibly due: 1 ) to interference with LPS biosynthesis , since the modifications they catalyze do not normally exist in E . coli and likely generate cytotoxicity , and/or 2 ) to accumulation of large amounts of membrane proteins . These complementation assays perfectly corroborate the previous biochemical data , further demonstrating that LpxE and HP0851 act as the major and perhaps the sole C55-PP phosphatases in H . pylori . Hence , we renamed HP0851 to HupA ( Helicobacter UppP and PGPase A ) . To address whether LpxE and HupA are the only C55-PP phosphatases in H . pylori , we attempted to construct a lpxE/hupA double mutant . However , we failed to generate this strain suggesting that co-inactivation of both genes is lethal . We then performed transformation efficiency assays to test the capacity of each PAP2 from H . pylori to complement this apparent lethality . We generated four strains deleted for hupA and carrying a copy of one PAP2 encoding gene under the control of an IPTG-inducible promoter on the pILL2150 vector . We then transformed these cells with the Topo TAΔlpxE plasmid in order to replace lpxE by a gentamycin resistance cassette . The transformed H . pylori population was diluted and spread on normal ( total number of bacteria ) or selective ( ΔlpxE recombinants ) plates to measure the transformation efficiency expressed in cfu/μg of Topo TAΔlpxE plasmid ( Table 4 ) . LpxF and HP0350 were unable to complement the lpxE/hupA double mutant as no recombinant was obtained even in the presence of IPTG . In contrast , LpxE and HupA complemented the double mutant in the presence of IPTG . These data excluded LpxF and HP0350 as bona fide UppP and confirm that only LpxE and HupA are major UppPs in H . pylori . H . pylori infects the stomach and persists in the mucosa during years due to the expression of multiple virulence factors . Persistence also requires a resistance towards CAMPs , which are secreted by the host [22] . To evaluate the involvement of PAP2 proteins in CAMPs resistance , we used polymyxin B as a surrogate for CAMPs [23] . The modifications generated by LpxE and LpxF through their lipid A 1- and 4’-phosphate phosphatase activities promote resistance to different CAMPs and they further confer host innate immune system evasion [19] . We confirmed the increased sensitivity already described in [19] for the lpxE and lpxF mutants ( 32- and 128-fold , respectively , Table 5 ) . The hp0350 mutant did not show any increase of polymyxin B sensitivity , while the sensitivity of hupA mutant increased by four-fold as compared to the wild-type strain . The polymyxin B sensitivity of hupA mutant was fully restored upon complementation with an ectopic copy of hupA gene . As already mentioned , LpxE and LpxF modify lipid A structure in H . pylori . We then investigated whether the four-fold decrease in polymyxin B resistance of hupA mutant was a consequence of an altered lipid A structure . LpxE exhibits UppP activity while also acting as a lipid A 1-phosphate phosphatase . We hypothesized that , in the absence of HupA , LpxE has to cope with this very low UppP activity . This hypothesis predicted the presence of a mixture of lipid A species in hupA mutant with lipid A 1-phosphate together with the usual lipid A 1-PE . The resulting rise in net negative charges at the outermost surface of the bacterium could then account for the increased sensitivity towards polymyxin B . To test this hypothesis , the lipid A from wild-type , hupA mutant and hupA mutant transformed with pILL2150 hupA plasmid were extracted and analyzed by mass spectrometry . As shown in Fig 3A–3C , the main form of lipid A present in wild-type strain was the tetra-acylated lipid A with a PE group at position 1 and no phosphate group at position 4’ , as described [19] . Two additional minor species were observed , a penta-acylated lipid A with PE at position 1 and no phosphate at position 4’ and a hexa-acylated lipid A with PE at position 1 and with phosphate at position 4’ . In hupA mutant and its complemented derivative , the lipid A profiles were very similar to that of the wild-type strain . We then concluded that the moderate decrease in polymyxin B resistance in hupA strain was independent of the lipid A structure . In parallel , we confirmed the roles of LpxE and LpxF ( Fig 3E and 3F ) while we show that HP0350 is not involved in lipid A modifications ( Fig 3D ) . The maximal UppP activity of LpxE was reached at pH 7 . 4 ( Fig 1 ) and all complementation assays were performed at neutral pH . H . pylori resumes growth in acidic media only after buffering the environment to a pH above 6 thanks to the urease enzyme [24] , which precludes any assays at acidic pH . However , in mouse colonization , the natural environment of H . pylori is mainly acidic , ranging from 2 in the lumen to 5–6 in the mucus layer . But at low pH , we noticed that the in vitro activity of LpxE drastically decreased by 43-fold ( from 900 to 21 nmol/min/mg ) , while that of HupA increased by 1 . 7-fold ( from 3493 to 6039 nmol/min/mg ) ( Table 2 ) . Therefore , LpxE may be unable to fully complement the reduced UppP activity in the hupA mutant in an acidic environment . This issue was then addressed by mouse colonization assays . The ΔhupA mutation was first introduced into the mouse-adapted strain H pylori X47 and the kinetics of colonization of wild-type and hupA strains was followed by sacrificing mice after 1 , 4 , 7 , 15 and 32 days after bacterial inoculation . Seven mice per group were sacrificed in two independent experiments and Fig 4 illustrates the CFU/g of stomach in a time course . The hupA mutant was unable to colonize the stomach neither at early or later stages after inoculation , showing therefore that HupA is absolutely essential for viability of H . pylori in vivo , in sharp contrast to in vitro where neither of the four PAP2 mutants was affected in growth ( Fig 5 ) . HupA is an orthologue of PgpB from E . coli , which is a bifunctional enzyme acting as a minor UppP and being largely involved in phospholipid biosynthesis through the dephosphorylation of PGP in PG . In E . coli , the latter activity is also provided by two other integral membrane proteins named PgpA and PgpC , which are unrelated to the PAP2 superfamily [10 , 11 , 25] . One putative gene encoding a PgpA homolog was identified in the genome of H . pylori , hp0737 , while no PgpC encoding gene was found . The hp0737 was cloned on a pTrcHis30 vector and the N-terminal-His tagged protein was expressed in E . coli C43 ( DE3 ) cells and purified from membranes to high homogeneity in DDM micelles ( S1A Fig ) . In contrast to the PAP2 enzymes , PgpA enzymes are Mg2+-dependent , which was confirmed for HP0737 that presented an optimal Mg2+ concentration of 6 mM ( S1B Fig ) . Its PGP phosphatase activity was then fully confirmed in vitro with a specific activity of 1138 ± 174 nmol/min/mg . Noticeably , HP0737 did not display any significant UppP activity and the plasmid carrying hp0737 gene did not restore the growth of E . coli BWTsbacA strain at 42°C . Considering the activity of HP0737 both in vivo and in vitro , this protein was renamed PgpA . The pgpA null mutant was readily generated by replacement with a resistance cassette ( Table 1 ) . These disagree with a previous study in which pgpA was described as an essential gene [26] . Our data further suggested the existence of one or several other PGP phosphatases . We then assessed whether the PAP2 proteins from H . pylori are also involved in the synthesis of PG . DDM-solubilized membrane extracts obtained from wild-type and mutant cells ( lpxE , hp0350 , hupA , lpxF and pgpA ) were assayed for their PGP phosphatase activity . This activity was measured in the absence and presence of 6 mM Mg2+ ( Fig 6A ) . A drastic decrease of the PGP phosphatase activity in the hupA mutant as compared to the wild-type strain was observed , while the other mutants displayed similar activities as the control . Under these conditions , HupA accounted for 98% of the PGP phosphatase activity in the membrane of H . pylori . To further confirm the involvement of HupA in the biosynthesis of PG , we also performed complementation assays using the conditional BWPGPTs E . coli strain and the different pTrcHis30-based plasmids . Like the BWTsbacA strain , the BWPGPTs is a thermosensitive triple mutant ( ΔpgpA , ΔpgpB , ΔpgpC ) containing an ectopic copy of pgpB on a plasmid whose replication is impaired at 42°C . LpxE , HupA and PgpA fully restored the growth at 42°C of this thermosensitive strain without IPTG , while LpxF and HP0350 were unable to complement ( Table 6 ) . Thus , LpxE , HupA and PgpA carry enough of PGP phosphatase activity to support growth of the BWPGPTs strain . The apparent PGP phosphatase activity of HupA and LpxE in vivo could explain why pgpA inactivation did not lead to a lethal phenotype . To further decipher the role of PAP2 enzymes in the biosynthesis of PG in H . pylori , we tried to generate mutant strains inactivated for hupA and pgpA but we failed , while a double mutant lpxE/pgpA was readily obtained . These results suggest that HupA and PgpA are the only PGP phosphatases able to sustain growth of H . pylori in vitro . Since the HupA and LpxE accept very dissimilar lipidic substrates in vivo , we further analyzed the substrate specificity of PAP2 enzymes in vitro ( Fig 6B ) . Again , HP0350 had no visible phosphatase activity on any of the tested substrates . HupA was the most active phosphatase using all substrates tested except phosphatidic acid and the highest activity was by far obtained with the C15-PP substrate . Structural and topology analyses of integral membrane PAP2 enzymes revealed that their active site residues are oriented towards the periplasm , which is likely the same for HupA . Small molecules such as C15-PP are never exposed at the periplasmic site , therefore the C15-PP should not be a natural substrate in vivo . Nevertheless , this analysis demonstrates that HupA accepts a large range of pyro- and monophosphate substrates , in particular C55-PP and PGP . Noticeably , HupA and PgpA displayed very similar activities towards PGP . LpxE had similar substrate specificity as HupA , except for C5-PP , and much lower activities were found . In contrast , LpxF only showed very low activity towards C15-PP and PGP . Since HupA showed broad substrate specificity , we wondered whether the polymyxin B sensitivity could be related to a general membrane permeability defect of the hupA mutant . We measured the MIC of the wild-type strain and its four PAP2 mutants to four different antimicrobial , one clinically relevant CAMP , colistin , and three large antibiotics , vancomycin , teicoplanin and daptomycin that do not cross the outer membrane . Table 7 illustrates the MICs to the four antimicrobials . Absence of HupA , LpxE and LpxF affected specifically resistance to CAMPs without affecting the resistance to large antimicrobials suggesting that these three PAP2 only affect overall membrane charge and not permeability . Given the specific effect on CAMPs resistance and the broad substrate specificity of HupA , we next analyzed the phospholipid composition of the hupA mutant . As illustrated in S2 Fig , the wild-type N6 and its four PAP2 mutants showed a similar phospholipid profile . Since , HupA and LpxE are functional in E . coli , we tested if the PAP2/BacA enzymes from E . coli are also functional in H . pylori . We performed complementation assays as previously described by measuring the transformation rate of lpxE inactivation in hupA mutant containing different plasmids encoding PAP2/BacA from E . coli . We then generated eight strains deleted for hupA and carrying a plasmid ( pILL2150 or pILL2157 vector ) bearing one of the four E . coli genes ( bacA , pbpB , ybjG and lpxT ) placed under the control of an IPTG-inducible promoter . The transgene is under control of a promoter from H . pylori in pILL2157 , while it is under control of an E . coli promoter in pILL2150 . Consequently , the transgenes in pILL2157 are likely more expressed in H . pylori as compared to those in pILL2150 . Moreover , the pILL2157 promoter was previously found to be leaky in contrast to that of pILL2150 vector [27] . The transformation rates are summarized in Table 8 . As expected , LpxT was unable to complement the lack of UppP activity in H . pylori whatever the expression level . At a moderate expression level ( i . e . in pILL2150 ) , only PgpB was able to complement the double lpxE/hupA mutant ( 1 . 08 × 10−5 transformants/cfu/μg DNA topoTAΔlpxE ) in the presence of IPTG . In contrast , at a higher level of expression ( i . e . in pILL2157 ) , BacA , PgpB and YbjG were able to complement the double mutant both in the presence and in the absence of inducer . Therefore , even though BacA does not belong to the PAP2 super-family , and H . pylori does not possess UppP of the BacA type , this enzyme was functional in H . pylori .
The recycling of C55-PP is an essential step for the biosynthesis of many polysaccharides such as PGN , LPS O-antigen or teichoic acid [4] . This pathway was only studied in the Gram-positive bacterium B . subtilis [6] and more intensively in the Gram-negative E . coli . Two enzymes from H . pylori belonging to the PAP2 super-family , LpxE and LpxF , were previously demonstrated to have critical functions through the dephosphorylation of the lipid A moiety from LPS . The other two members of the PAP2 super-family , HupA and HP0350 , were not characterized . HupA and HP0350 are orthologues of PgpB from E . coli , which was described as a dual functional enzyme exhibiting both C55-PP and PGP phosphatase activities , thus involved in both cell-envelope polysaccharides and phospholipids biosynthesis [11] . Here , we demonstrated , by in vitro and in vivo analyses , that HupA constitutes the major C55-PP phosphatase in H . pylori , being responsible for 90% of the total UppP activity present in the membranes . Interestingly , HupA presented an optimal pH of 5 for its enzymatic activity , which is close to the pH that H . pylori cells face in the human stomach . To further confirm this function , we showed that HupA was also fully functional in E . coli by complementing a PAP2/BacA conditionally deficient strain . The deletion of hupA gene was not lethal suggesting that one or several other C55-PP phosphatases exist in H . pylori . Among the three other PAP2 , we showed that LpxE and LpxF exhibited UppP activity with an optimal pH of 7 . 4 and 5 , respectively . However , only LpxE was also capable of complementing the conditional E . coli strain . The inactivation of both lpxE and hupA genes was found to be lethal , which was further confirmed by complementation assays of the double mutant by ectopic copies of lpxE or hupA . All these data confirmed that HupA and LpxE are the only physiologically relevant UppP in H . pylori . In addition , we demonstrated that HupA plays critical roles in vivo . Indeed , HupA appeared important for the resistance towards CAMPs by a yet uncharacterized process . In this study , we excluded potential alteration of the structure of lipid A . More importantly , HupA was shown to be essential for stomach colonization . Indeed , the PAP2 super-family members all have their active site exposed to the periplasmic space where the pH is determined by the environmental pH . We then hypothesized , based on in vitro data , that LpxE enzyme might not be sufficiently active at acidic pH to sustain growth in the absence of HupA . Indeed , during colonization , LpxE is unlikely to function properly since it was poorly active at pH 5 in vitro . In fact , apart from the niche that is in contact with the epithelial cells where there exists a neutral pH , H . pylori cells are always exposed to an acidic environment . In these conditions , LpxE will likely be unable to provide the UppP activity required in absence of HupA enzyme . The first step for stomach colonization by H . pylori is dependent on its motility and its urease enzyme to buffer its cytoplasm [24] . H . pylori escapes the lumen towards the mucus layer to reach a more favorable environment with a higher pH . Therefore , we cannot exclude a relevant physiological role of LpxE in C55-PP dephosphorylation after the first step of colonization , once H . pylori has reached the mucus layer and the epithelial surface . Our previous work on LpxE showed that LpxE is needed for long-term colonization . However , this is likely due to its role as lipid A 1-phosphate phosphatase and escape to TLR4 signaling rather than to its UppP activity since the mutant was able to colonize TLR4 KO mice [19] . This study also highlighted the dual function of HupA , which has also a major involvement in the biosynthesis of phospholipids through the synthesis of PG from PGP . By sequence homology , only one additional putative PGP phosphatase was found in H . pylori , i . e . HP0737 , which is homologous to E . coli PgpA . A global transposon analysis performed on H . pylori reported hp0737 as an essential gene [26] . However , we were able to delete hp0737 gene by resistance cassette replacement . HP0737 could then account for the residual PGP phosphatase activity in H . pylori hupA null mutant . In E . coli , three PGP phosphatases were described , PgpA , PgpB and PgpC and only a triple mutant is lethal [11] . Even if the composition of phospholipids differs between E . coli and H . pylori , i . e . PG represents 25% and 12 . 5% of the total phospholipids , respectively [28 , 29] , we can expect that the presence of PG is also essential in H . pylori , especially because PG is the precursor of another important phospholipid , the cardiolipin . A pgpA/hupA double mutant is lethal and confirms the importance of PG in H . pylori . Members of the PgpA-family have a predicted cytoplasmic active site while PAP2-like enzymes have a periplasmic oriented active site . These results suggest that PGP is accessible to dephosphorylation on both sides of the cytoplasmic membrane similarly to E . coli . Thus , HupA and PgpA are the only two physiologically relevant PGP phosphatases in H . pylori despite the weak PGP phosphatase activity of LpxE . The involvement of HupA in phospholipid biosynthesis may explain the decrease of resistance to polymyxin B and colistin of the corresponding mutant . Indeed , these cationic peptides form pores in the cytoplasmic membrane leading to cytoplasmic leakage . Therefore , if the composition of the plasma membrane differs in the hupA mutant as compared to the wild-type strain , particularly with a possible accumulation of PGP , the membrane net negative charge might increase , favoring the binding of CAMPs , which could potentiate their capacity to insert within the membrane . However , our analysis of phospholipid composition of the wild-type N6 and its four PAP2 mutants showed similar phospholipid composition . Although TLC analysis is mainly qualitative , these results suggest that changes in membrane charge are unlikely to be due to an accumulation of PGP and suggest that accumulation of C55-PP might contribute to the mild increased sensitivity of the hupA mutant to CAMPs . LpxE and HupA have similar broad substrate specificities although LpxE exhibits a much weaker activity . In contrast , LpxF has a very narrow substrate specificity and , so far , only catalyzes the dephosphorylation of the lipid A . It would be interesting to study the molecular basis of such distinct substrate specificity . The low sequence conservation among the membrane PAP2 proteins precludes in silico modelling studies using PgpB and BsPgpB . Insights into substrate specificity will require 3D structures of HupA , LpxE and LpxF with their substrates . In this report , we described HupA as the main C55-PP and PGP phosphatase in H . pylori . This protein was also essential for colonization , likely due to its dual functionality and its central role in several lipid biosynthetic pathways , suggesting that HupA is the sole enzyme capable to support C55-PP recycling and PG synthesis in its natural niche . Hence , HupA , as well as LpxE and LpxF , constitute bona fide new targets for innovative therapeutic strategies against H . pylori , which is becoming increasingly resistant to the existing antibiotic arsenal .
Animal experiments were done according to European ( Directive 2010/63 EU ) and French regulation ( Décret 2013–118 ) under the authorized protocol CETEA 2014–072 reviewed by the Institut Pasteur Ethical Committee ( registered as number 89 with the French Ministry of Research ) . The experimental protocol was also approved by the French Ministry of Research under the number APAFIS#11694–2017100510327765 v2 . The bacterial strains and plasmids used in the study are summarized in Table 1 . Precultures of H . pylori were started from glycerol stocks routinely stored at -80°C , plated onto 10% horse blood agar medium supplemented with an antibiotic-antifungal mix [19] and incubated at 37°C for 24 h in a microaerobic atmosphere ( 6% O2 , 10% CO2 , 84% N2 ) . The glycerol storage media is composed of 25% glycerol , 38% Brain Heart Infusion liquid ( BHI ) ( Oxoid ) and 37% sterilized water . Amplification of the preculture was performed in the same conditions as precultures in new plates at 37°C for 24 h in a microaerobic atmosphere . Liquid precultures were started from amplification plates and inoculated into BHI ( Oxoid ) supplemented with 10% fetal bovine serum ( FBS ) incubated at 37°C overnight in a microaerobic atmosphere . Liquid cultures of H . pylori were started from overnight liquid culture and inoculated into BHI ( Oxoid ) supplemented with 10% fetal bovine serum . In general , liquid cultures of H . pylori were grown to an OD600nm of ~1 . 0 at 37°C under microaerobic conditions with shaking . Otherwise indicated , E . coli was routinely grown at 37°C in 2YT broth using standard conditions . A non-polar kanamycin cassette was digested out of plasmid pUC18-Km2 [30] using BamHI and KpnI and ligated into the PCR products of the 5’and 3’ flanking regions of genes lpxE , hp0350 , hupA , lpxF and pgpA using standard cloning techniques . The oligonucleotides used for the PCR amplification are described in S1 Table . The generated plasmids , TopoTA ΔlpxE:kan , TopoTA Δhp0350:kan , TopoTA ΔhupA:kan , TopoTA ΔlpxF:kan and TopoTA ΔpgpA:kan were used to create the corresponding null mutants in H . pylori N6 and/or X47 by natural transformation . Similarly , a non-polar gentamycin cassette was digested out of plasmid pUC18-Gm [31] using BamHI and KpnI to generate TopoTA ΔlpxE:Gm . For the expression of PAP2 , BacA and PgpA encoding genes from E . coli or H . pylori , the genes were amplified by PCR using appropriate primers ( S1 Table ) and cloned into the pILL2150 or pILL2157 vectors ( for expression in H . pylori ) or in pTrcHis30 vector ( for expression in E . coli ) . E . coli C43 ( DE3 ) cells carrying pTrcHis30-based plasmids were used for the overproduction of N-terminal His6-tagged proteins . Cells were grown in 1 liter 2YT-ampicillin at 37°C until the A600nm reached 0 . 9 , when the expression was induced by the addition of 1 mM IPTG and the growth was continued for 3 h at 37°C . Cells were harvested by centrifugation at 4°C for 20 min at 4000 × g , were washed and finally resuspended in buffer A ( 20 mM Tris-HCl pH 7 . 4 , 400 mM NaCl , 10% glycerol ) before disruption by sonication with a Vibracell 72412 sonicator ( Bioblock ) . The suspension was centrifuged at 4°C for 1 h at 100 , 000 × g to harvest the membranes , which were washed three times in buffer A . The membranes were finally resuspended in buffer A supplemented with 2% DDM for solubilization at 4°C for 2 h with gentle agitation . Solubilized membranes were loaded on nickel-nitrilotriacetate agarose ( Ni2+-NTA-agarose , Qiagen ) equilibrated in buffer A supplemented with 0 . 2% DDM and 10 mM imidazole . The column was washed with increasing imidazole concentration using the same buffer and the elution was performed with 2 ml of buffer A supplemented with 0 . 2% DDM and 200 mM imidazole . Desalting of samples was carried out using PD-10 desalting columns ( GE Healthcare ) and buffer A supplemented with 0 . 1% DDM . Protein concentrations were determined with the QuantiProBCA assay kit ( Sigma ) or by densitometry analysis of the gel when appropriate . H . pylori wild-type and single mutant strains were grown at 37°C in BHI medium ( 200 ml culture ) up to exponential phase . Cell free extracts , membrane free cytosol , and washed membranes were prepared as previously described [22] . Cells were harvested by centrifugation , washed and finally resuspended in 20 mM Tris-HCl , pH 7 . 4 , 200 mM NaCl ( buffer B ) . After disruption by sonication , membranes were pelleted by centrifugation at 177 420 × g and resuspended in buffer B supplemented with 2% DDM for solubilization during 1 . 5 h in the cold . The solubilized proteins were recovered by centrifugation at 177 420 × g and conserved at -20°C before activity measurement . H . pylori wild-type and four single mutant strains were grown at 37°C in BHI medium ( 50 ml culture ) up to exponential phase . Lipid extraction was performed using a new protocol ( Nozeret K et al . manuscript submitted ) . For thin-layer chromatography ( TLC ) analysis of phospholipids , the dried lipid extracts and controls were dissolved respectively in 500μl and 350μl of chloroform . 10μl of the solution was spotted onto a TLC silica gel 60 plate . The TLC plate was developed in tanks equilibrated with dichloromethane- methanol-water ( 65:28:4 [vol/vol] ) . After drying the plate , phospholipids were visualized with molybdenum blue reagent ( Sigma ) . The C55-PP and PGP phosphatase assays were carried out in a 10 μl reaction mixture containing 20 mM Tris-HCl , pH 7 . 4 , 10 mM β-mercaptoethanol , 150 mM NaCl , 0 . 2% DDM , 50 μM [14C]C55-PP or 50 μM [14C]PGP ( 900 Bq ) and enzyme . MgCl2 was added at 6 mM in the reaction mixture when PgpA activity was measured . Appropriate dilutions of purified phosphatases , or of membrane extracts , were used to achieve less than 30% substrate hydrolysis . The reaction mixture was incubated at 37°C for 10 to 30 min and the reaction was stopped by freezing in liquid nitrogen . The substrates and products were then separated and quantified by thin layer chromatography ( TLC ) analysis , as previously described for C55-PP [5] and PGP [11] hydrolysis . When the phosphatase activity was investigated at various pH values , buffering was achieved in sodium acetate ( pH 3–7 ) , Tris-HCl ( pH 7–9 ) or sodium carbonate buffer ( pH 9–11 ) . The phosphatase activity towards other non-radiolabeled substrates: C5-PP , C15-PP , diacyl ( C8 ) glycerol-PP ( DGPP ) and phosphatidic acid ( PA ) , was determined by measuring the amount of released inorganic phosphate during catalysis . The reaction mixture was as described above with 50 μM substrate in a final volume of 50 μl . After 10 min of incubation at 37°C , the reaction was stopped by the addition of 100 μl of Malachite green solution ( Biomol green , Enzo Life Sciences ) , and the released phosphate was quantified by measurement of the absorbance at 620 nm . The E . coli thermosensitive strains BWTsbacA and BWPGPTs , carrying multiple chromosomal gene deletions and harboring a bacA- or pgpB-expressing plasmid , respectively , whose replication is impaired at 42°C , have been previously described [7] . These strains were transformed by the pTrcHis30-based plasmids encoding the different H . pylori PAP2 and PgpA encoding genes . Isolated transformants were subcultured at 30°C in liquid 2YT medium supplemented with 100 μg/ml ampicillin and , when the absorbance of the culture reached 0 . 5 , the cell suspension was diluted to 10−5 in 2YT medium and 100 μl aliquots were plated onto two ampicillin-containing 2YT agar plates which were incubated at either 30°C or 42°C for 24 h . The colony forming units ( CFU ) were numerated on each plate and the functional complementation of the thermosensitive mutants was evaluated by the capacity of the transformants to equally grow at both temperatures . Precultures of H . pylori were started from glycerol stocks routinely stored at -80°C , plated onto 10% horse blood agar medium and incubated at 37°C for 24 h in a microaerobic atmosphere ( 6% O2 , 10% CO2 , 84% N2 ) . Amplification of the pre-culture was performed in the same conditions such as pre-cultures in new plates at 37°C for 24 h in a microaerobic atmosphere . 50 μl of suspension in BHI at OD600nm = 20 were made until the amplification plate and mixed with 10 ng of Topo TAΔlpxE plasmid . The whole mixture was put on a non-selective plate onto 10% horse blood agar medium and incubated at 37°C for 24 h in a microaerobic atmosphere . The transformation spot was then diluted ( 10−1 to 10−8 ) . 50 μl of the non-diluted suspension and 10 μl of the 10−1 to 10−4 dilutions were spread on 10% horse blood agar medium supplemented with chloramphenicol ( 4 μg/ml ) , kanamycin ( 20 μg/ml ) and gentamycin ( 5 μg/ml ) ± IPTG 1mM to estimate the number of transformants . The 10−5 to 10−8 dilutions were spread on non-selective plates containing 10% horse blood agar medium supplemented with chloramphenicol ( 4 μg/ml ) , kanamycin ( 20 μg/ml ) ± IPTG 1mM to estimate the total number of bacteria . The CFU of all plates incubated at 37°C for 5 days in a microaerobic atmosphere were enumerated . The transformation rate was expressed by the number of transformants/total cfu/μg of DNA TopoTAΔlpxE . Strains were grown routinely on amplification plates . Two ml of suspension in 0 . 9% NaCl at OD600nm = 0 . 75 were made from the amplification plates and spread by inundation onto Mueller Hinton medium ( Difco ) supplemented with 10% of FBS and 2 , 3 , 5-triphenyltetrazolium chloride ( TTC , Sigma ) ( 40 μg/ml ) . Polymyxin B was diluted by serial dilutions of two ( 16384 μg/ml down to 0 . 5 μg/ml ) . Once the square plates with bacterial lawns were dried , 10 μl of each dilution of polymyxin B was dropped on and allowed to dry . All plates were then incubated at 37°C for 3 days in a microaerobic atmosphere . TTC is a non-toxic dye which colors in red the alive bacteria . MICs were determined as the lowest concentration of polymyxin B leading to a clear halo of inhibition . Strains were grown routinely on amplification plates . Two ml of suspension in 0 . 9% NaCl at OD600nm = 0 . 75 were made from the amplification plates and spread by inundation onto 10% horse blood agar medium . Once the square plates with bacterial lawns were dried , Etes , Biomérieux , with different antibiotics targeting the cell wall ( Vancomycin , Teicoplanin , Daptomycin and Colistin ) were put on plate and incubated at 37°C for 48 h in a microaerobic atmosphere . MICs were determined as the lowest concentration of antibiotics leading to a clear halo of inhibition . OF1 female mice purchased from Charles River Laboratories aged 5 weeks were infected by gavage with feeding needles with X47 strain ( 2 × 108 bacteria per mouse ) . Colonization rates were determined after 1 , 4 , 7 , 15 and 32 days by enumeration of CFU per gram of stomach . Mice were euthanized with CO2 and the stomachs were ground and homogenized in peptone broth . The samples were then diluted and spread on blood agar plates supplemented with 10 μg/ml of nalidixic acid , to inhibit the growth of resident bacteria from the mouse forestomach and 20 μg/ml of kanamycin for hupA mutant strain . The CFU were enumerated after 5 days of incubation under microaerobic conditions . The results of two independent colonization experiments ( seven mice by cage ) were pooled and a one tailed Mann-Whitney test was used to determine statistical significance of observed differences ( GraphPad Prism v5 . 0 GraphPad Software , CA ) . For isolation of lipid A , H . pylori cultures were grown to an OD600 of ~0 . 6 . Lipid A chemical extraction was carried out after mild acidic hydrolysis of LPS as previously described [32 , 33] . For visualization of lipid A by mass spectrometry , lipids were analyzed using MALDI-TOF ( ABI 4700 Proteomic Analyzer ) in the negative-ion linear mode similar to previously described [34 , 35] . Briefly , lipid A samples were dissolved in a mixture of chloroform-methanol ( 4:1 , vol/vol ) , with 1 μL of sample mixed with 1 μL of matrix solution consisting of 5-chloro-2meracaptobenzothiazole ( CBMT ) ( 20 mg/mL ) resuspended in chloroform-methanol-water ( 4:4:1 , vol/vol/vol ) mixed with saturated ammonium citrate ( 20:1 , vol/vol ) , and 1 μL of sample-matrix mixture was loaded on to MALDI target plate .
|
Helicobacter pylori colonizes the human’s gastric mucosa and infects around 50% of the world’s population . This pathogen is responsible for chronic gastritis , peptic ulcers and in worst cases leads to gastric cancer . It has been classified as a class I carcinogen by the World Health Organization in 1994 . Here , we show that HP0851 , renamed HupA , is the major undecaprenyl pyrophosphate ( C55-PP ) phosphatase ( UppP ) and the major phosphatidylglycerol phosphate phosphatase ( PGPase ) . This enzyme is also involved in cationic antimicrobial peptide ( CAMP ) resistance to which H . pylori hupA mutant shows an increased sensitivity ( 4 fold ) . This mutant was unable to colonize the stomach in mouse model of infection showing that even if hupA was not essential in vitro , this gene was essential in vivo . Both HupA and LpxE have crucial roles in the biosynthesis of several cell wall polysaccharides and thus constitute potential targets for new therapeutic strategies .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2019
|
HupA, the main undecaprenyl pyrophosphate and phosphatidylglycerol phosphate phosphatase in Helicobacter pylori is essential for colonization of the stomach
|
Considerable evidence has accumulated in recent years suggesting that G protein-coupled receptors ( GPCRs ) associate in the plasma membrane to form homo- and/or heteromers . Nevertheless , the stoichiometry , fraction and lifetime of such receptor complexes in living cells remain topics of intense debate . Motivated by experimental data suggesting differing stabilities for homomers of the cognate human β1- and β2-adrenergic receptors , we have carried out approximately 160 microseconds of biased molecular dynamics simulations to calculate the dimerization free energy of crystal structure-based models of these receptors , interacting at two interfaces that have often been implicated in GPCR association under physiological conditions . Specifically , results are presented for simulations of coarse-grained ( MARTINI-based ) and atomistic representations of each receptor , in homodimeric configurations with either transmembrane helices TM1/H8 or TM4/3 at the interface , in an explicit lipid bilayer . Our results support a definite contribution to the relative stability of GPCR dimers from both interface sequence and configuration . We conclude that β1- and β2-adrenergic receptor homodimers with TM1/H8 at the interface are more stable than those involving TM4/3 , and that this might be reconciled with experimental studies by considering a model of oligomerization in which more stable TM1 homodimers diffuse through the membrane , transiently interacting with other protomers at interfaces involving other TM helices .
G Protein-Coupled Receptors ( GPCRs ) have been reported to associate in the cell membrane to form dimers/oligomers . While incontrovertible evidence exists for the constitutive dimerization of disulfide-linked family C GPCRs [1] , the interpretation of oligomerization studies of members of the largest subfamily A of GPCRs [2] has often been difficult and controversial [3] , [4] , [5] , [6] , since the majority of the techniques used to infer GPCR association in living cells are unable to conclude unambiguously in favor of direct physical interaction between receptors . Most importantly , very few GPCR oligomerization studies have been able to provide any information about the fraction of receptors that are interacting at a given time or the corresponding dynamics of the interactions , rendering it impossible to determine , with any certainty , which molecular species ( i . e . individual protomers , dimers , or higher-order oligomers ) signal through interaction with intracellular proteins . These uncertainties have fueled an ongoing debate regarding the physiological role of GPCR oligomerization , exacerbated by the evidence that individual GPCR protomers , when reconstituted into nanodiscs , can signal to G proteins [6] , [7] , [8] , [9] . Recent studies using single-molecule approaches have begun to address the details of the spatial and temporal organization of GPCR complexes in living cells . Single-molecule total internal reflection fluorescence microscopy ( TIR-FM ) was recently used to track the position of individual molecules of the M1 muscarinic acetylcholine receptor ( M1R ) labeled with fluorescent M1R antagonists in living cells [10] . Both single- and dual-color imaging experiments suggested a transient ( ∼0 . 5 seconds ) formation of M1R dimers and a dimeric fraction of only ∼30% dimers at any given time . Although similar conclusions were reached by a single-molecule study of another family A GPCR , i . e the N-formyl peptide receptor [11] , the possibility cannot be excluded that the fluorescent ligands used to image the single molecules in both studies might have altered the lifetime and preferred stoichiometry of the observed GPCR oligomers . It remains to be determined whether or not the features highlighted in these studies are the same for all GPCRs , or just specific subtypes . Recent fluorescence recovery after photobleaching ( FRAP ) studies of human β1 and β2-adrenergic receptors ( B1AR and B2AR , respectively ) [12] have raised the possibility that the strength of GPCR association may vary significantly , even among closely related receptor subtypes . Although the antibody-mediated capping approach used to immobilize receptors in these studies may have affected the interpretation of the results , B1AR was suggested to interact transiently ( on a timescale of seconds ) whereas B2AR appeared to form more stable complexes ( on a timescale of minutes ) . Fung and colleagues [13] reported data in support of spontaneous B2AR oligomerization using Förster resonance energy transfer ( FRET ) between relatively small fluorescent probes attached to purified B2AR reconstituted into phospholipid vesicles . The authors hypothesized predominant tetrameric arrangement for the B2AR , although they did note the difficulty of unambiguously determining the stoichiometry of receptor oligomers from a reconstituted system . Additional FRET saturation studies showed greatest energy transfers for H8 and smallest for TM6 , based on which , the authors proposed a preferential oligomeric arrangement of B2AR involving TM1 and H8 at the interface , similar to that previously suggested for the dopamine D2 receptor from a combination of molecular modeling and cysteine cross-linking experiments [14] . Further support for the simultaneous involvement of helices TM1 and H8 at an interface was recently provided by chemical cross-linking of endogenous cysteines in rhodopsin in disk membranes [15] . Several additional experimental studies support the direct primary involvement of TM1 , as well as TM4 in GPCR oligomerization under physiological conditions [16] . Although an alternative dimerization interface involving both helices TM5 and TM6 was recently suggested by the crystal structures of the chemokine CXCR4 [17] and μ-opioid [18] receptors , its physiological relevance has not yet been demonstrated . Here , we sought insight into the dimerization free energy of models of human B1AR and B2AR based on high-resolution inactive crystal structures interacting at two putative interfaces involving TM1/H8 or TM4/3 , using biased molecular dynamics ( MD ) simulations . Specifically , we combined umbrella sampling and metadynamics simulations to provide hypotheses of the role played by the interface sequence and configuration in imparting stability to the specific dimeric arrangements that we have simulated for both B1AR and B2AR . These studies provide a mechanistic insight into the association of GPCRs at putative dimerization interfaces at a level of molecular detail that is unattainable using current experimental techniques , yet crucial to guiding future experiments aimed at exploring the role of dimerization in receptor function .
Fig . 1A shows the reconstructed free energy surface ( FES ) as a function of the separation , r , between the COMs of the protomers for the B1AR ( red ) and B2AR ( blue ) homodimers . We note that the overall shapes of the corresponding curves are similar and their depths are equivalent , within the calculated error bars . Upon offsetting the curves to a zero value in the region beyond which the protomers were seen not to be interacting and were therefore designated as monomeric states ( r = 4 . 5–4 . 8 nm ) , we observe that the depths of each of the two minima are similar between the B1AR and B2AR homodimers . To confirm the choice of the reference state , we sampled one of the systems , specifically the B2AR interacting at the TM4/3 interface , to larger separation distance ( Fig . 1 ) . As reported in Table 1 , the primary minimum is at −4 . 8 kcal/mol and −5 . 8 kcal/mol for B1AR and B2AR , respectively . Using the same theory described in [19] , and the equations 1–3 reported in Materials and Methods , dimerization free energies ( i . e . , mole-faction standard state free energy changes ΔGX° ) of −2 . 3 kcal/mol and −3 . 7 kcal/mol were calculated for B1AR and B2AR , respectively . We proceeded to identify the relevant orientations of the protomers within each of these dimeric minima by comparing the FES calculated as a function of the angles ( see Fig . S3 ) at r = 3 . 42–3 . 48 nm for B1AR and r = 3 . 42–3 . 46 nm for B2AR . The FES as a function of the angle for the TM4/3 interface indicates minima situated approximately at Θ1 ( θa , θb ) = ( 0 . 2 , 0 . 4 ) ( or the symmetric Θ1' ( θa , θb ) = ( 0 . 4 , 0 . 2 ) value ) and Θ2 ( θa , θb ) = ( 0 . 45 , 0 . 45 ) radians , respectively . Superpositions of energetically-optimized , all-atom reconstructions of representative structures of the Θ1 and Θ2 minima , obtained using the procedure described in the Materials and Methods section , are shown in Fig . 2A for both B1AR ( red/pink respectively ) and B2AR ( blue/light blue ) . Symmetric inter-helical contacts , defined as average interaction distances between residue Cβ atoms less than or equal to the threshold distance of 11 Å during 1 ns unbiased all-atom MD simulations are listed for each of these representative structures in Table S1 . Fig . S4A shows the location of the corresponding residues involved in these contacts . Fig . 1B shows the FES resulting from the calculation of B1AR and B2AR at interfaces involving TM1/H8 as a function of their separation , r , and calculated using the CVs illustrated in the inset of Fig . 1B . As for the TM4/3 interface , the overall shape and depth of the corresponding curves are similar , albeit not identical , between the B1AR and B2AR dimers , most likely due to slight structural divergences between the corresponding reference structures . For the TM1/H8 interface , the monomeric state was defined to be r = 5 . 5–5 . 9 nm . The primary minimum for each system at this interface indicates a separation between the protomers of ∼3 . 7 nm , corresponding to −12 . 0 kcal/mol and −12 . 9 kcal/mol for B1AR and B2AR , respectively . The ΔGX° of the TM1/H8 dimers is approximately −9 . 7 kcal/mol for B1AR and −10 . 0 kcal/mol for B2AR . The reconstructed atomistic TM1/H8 dimers of B1AR and B2AR obtained in the same way as previously described , for the TM4/3 interface , are shown in Fig . 2B . Once again , the relative orientations of the protomers in these specific configurations involving the TM1/H8 interfaces were determined from the FES shown in Fig . S3 as a function of the angles , θa and θb , at r = 3 . 72–3 . 77 nm for B1AR and r = 3 . 68–3 . 72 nm for B2AR . The minima are approximately situated at Θ1 ( θa , θb ) = ( 0 . 45–0 . 50 , 0 . 45–0 . 50 ) . Contacting residues between the protomers during 1 ns of explicit simulation are listed in Table S1 and depicted in Fig . S4B .
The spatial and temporal organization of GPCRs in living cells is currently the subject of lively discussion . Although recent applications of single-molecule approaches are beginning to address the preferred stoichiometry , lifetime , and ratio of GPCR dimeric/oligomeric complexes to individual protomers in living cells , they are unable to provide the molecular details of receptor association . Biased MD simulations of the type reported here ensure thorough exploration of the interface for GPCR complex systems in an explicit lipid bilayer , that can be used to draw conclusions about relative values of dimerization free energies and dimer lifetimes . We have carried out free energy simulations of different GPCR subtypes interacting at two different interfaces that have been suggested , by experiment , to form dimers under physiological conditions . The goal of this study was not to predict the most stable interfaces of dimerization for the GPCR systems studied , which would have required comparison of all possible interfaces , but rather to investigate the effect of different sequences and/or interface configurations on the strength of GPCR dimerization at two dimeric interfaces inferred to be relevant under physiological conditions . Our study indicates that interfaces involving TM1/H8 are the most stable and the most long-lived ( minutes ) of the two simulated interfaces for B1AR and B2AR homodimers , based on estimates derived from the calculated free energies . The orientation of the protomers in the TM1/H8 interfacial arrangement is consistent with the close interaction of R333 in H8 ( at 13 Å in our dimeric model ) , which is the residue at which the greatest energy transfer was observed in the recent FRET study of B2AR [13] suggesting spontaneous B2AR oligomerization . The dimer corresponding to the TM4/3 interface appears to be significantly more transient ( hundreds of microseconds to milliseconds ) than the TM1/H8 interface for both receptors . The similar lifetimes estimated for B1AR and B2AR homodimers interacting at each of the interfaces tested suggest little difference in temporal organization between these two receptors , sharply contrasting with implications of the recent FRAP studies of human B1AR and B2AR [12] that motivated the present work . However , the possibility cannot be ruled out that the antibody-mediated capping approach used to immobilize receptors in the FRAP study might have caused the B2AR and B1AR to prefer interaction at different interfaces . While the estimated longer lifetime ( minutes ) of the B2AR dimer involving TM1/H8 at the interface may be considered in line with the views of the aforementioned FRAP study [12] , the observation contrasts with inferences from other FRAP studies on dopamine D2 receptors [21] , as well as conclusions of recent single molecule studies on muscarinic [10] or N-formyl peptide receptor [11] dimers , which suggest more transient interactions between GPCRs . Although we cannot set apart the contribution of the different membrane environments , we suggest that the results of our simulations may be reconciled with those experimental observations that imply only short-lived interactions by proposing a model of diffusion that features more stable receptor dimers , with TM1 at the interface , diffusing through the membrane and interacting transiently with one another at interfaces involving other TMs , to form short-lived tetrameric or higher-order arrangements . By superimposing the TM region of one of the two protomers of the simulated B2AR dimers on the active B2AR TM region of the recent crystal structure of the B2AR- Gs complex ( PDB ID: 3SN6 [22] ) we note interactions of the second protomer with the G-protein vary , depending on the specific dimeric arrangement of B2AR ( Fig . 3 ) . An interface involving TM4/3 would favor an exclusive interaction of the B2AR dimer with the alpha-helical domain of the nucleotide binding Gα subunit of the Gs protein ( “GαAH” in Fig . 3A ) . In contrast , in a B2AR dimeric arrangement with TM1/H8 at the interface , the second protomer would not be involved in significant interactions with any of the G-protein subunits ( Fig . 3B ) . It must be noted that these proposed conformations of the B2AR dimer in complex with the Gs were derived from simple superimposition , and thus would require additional simulations , beyond the scope of this study , to relieve the steric clashes arising from our use of the inactive conformation of the B2AR within the dimeric configurations . In summary , we have developed a protocol to assess the relative stability of GPCR dimers comprised of protomers interacting symmetrically at different TM regions that is robust within the standard caveats that apply to using a MARTINI-based CG model of proteins and membranes . We have recently published evidence for both small membrane dimeric systems [23] and GPCRs [20] that our CG simulations produce estimates of relative dimerization free energies that are in line with experimental data . Additional validations of the MARTINI model have been independently reported in the literature ( e . g . , see [24] , [25] , [26] ) . We herein demonstrate the dependence on the relative orientation of the protomers , in as much as the FES as a function of protomer separation is significantly different for B1AR and B2AR at interfaces involving either TM4 or TM1 . Although the free energy and lifetime estimates reported herein are somewhat dependent on the nature of the starting crystal structures , our calculations appear to be consistent for the different systems we have reported , and within the limits of the theories we have employed . While this manuscript was under review , a paper reporting further elaboration of simulation protocols used to study the self-assembly of rhodopsin molecules [27] , and now coupled with umbrella sampling calculations similar to those we published on delta opioid receptor [19] , [20] , has appeared in the literature [28] . Notably , the conclusions of these independent simulations about the relative stability of dimerization interfaces are in agreement with our calculations . We are confident that the protocol described herein can be generalized to begin deciphering the mechanistic details of dimerization of other GPCRs at a level of molecular detail that is unattainable using current experimental techniques . Although we do not expect to obtain an exact correspondence between the estimated lifetimes of GPCR dimers and those measured experimentally , our assessment of relative strength of association at different interfaces can be used constructively to predict specific interactions at the dimerization interface that might aid the design of experiments to assess the role of dimerization in receptor function .
Initial molecular models of human B1AR and B2AR were built using available inactive crystal structures of the turkey B1AR and human B2AR ( PDB identification codes: 2VT4 [29] , chain B , and 2RH1 [30] , respectively ) as structural templates . First , missing segments in the B1AR and B2AR crystal structures were built using Rosetta [31] . Specifically , these segments corresponded to sequence fragments 256–260 , 306–310 , and 313–317 in the B1AR and the intracellular loop 3 , sequence fragment 231–262 , which had been replaced by a T4 lysozyme in the B2AR crystal structure . To restore the broken ionic lock between TM3 and TM6 in the B2AR crystal structure , a standard MD simulation of 100 ns was carried out after embedding the receptor into an explicit hydrated POPC/10% cholesterol bilayer , following the procedure described in [32] . The homology model of the human B1AR was built using Modeller v8 [33] after alignment of the human and turkey sequences . The collective variables used in these simulations were the same as previously described in our earlier publications [19] , [20] , and are illustrated here in insets of Fig . 1 for each simulated dimeric arrangement of protomers a and b . Briefly , these CVs correspond to ( i ) the distance , r , between the centers of mass Ca and Cb of the TM regions of protomers a and b; ( ii ) the rotational angle , θa , defined as the arccosine of the inner product of the normalized vectors connecting the projections onto the plane of the membrane of the centers of mass of the specific TM ( s ) at the interface ( i . e . , TM4/3 and TM1/H8 ) and of the two TM bundles ( Ca and Cb ) , and ( iii ) the equivalent rotational angle , θb , for the second protomer . To aid simulation convergence , we restricted the exploration of θa and θb , using steep repulsive potentials , the details of which are in the SI , Table S2 . All simulations were performed using GROMACS version 4 . 0 . 5 [34] enhanced with the PLUMED plugin [35] , and the system components were represented using the MARTINI forcefield [36] , [37] , [38] ( using the parameters from version 2 . 1 for the protein beads , and version 2 . 0 for POPC and cholesterol ) , as described in our previous publications [19] , [20] . We focused on two dimeric interfaces that have received experimental validation under physiological conditions according to recent publications on GPCRs , specifically those involving TM4 or TM1 . The resulting dimeric configurations are illustrated by cartoons in the insets of Fig . 1 , and correspond to TM4/3 and TM1/H8 interfaces . Thus , initial configurations were built for homodimers of B1AR or B2AR , as described in our previous publications [19] , [20] . Subsequently , the dimers were converted to CG representation and embedded in a pre-equilibrated CG POPC/10% cholesterol membrane; the system was then solvated and counterions were added to neutralize the charge and to generate a physiological salt concentration of 0 . 1 M . An elastic network was used to restrain the protein system according to the strategy described in our previous publication [19] . Briefly , standard secondary structure constraints were introduced as per the MARTINI prescription; in addition , following a protocol put forward by Periole and colleagues [38] , we introduced elastic potentials between beads within a cutoff of 9 Å to maintain the integrity of the protein tertiary structure . In a modification of the original implementation , the force constants of the elastic network were weaker on loops ( 250 kJ/mol ) and stronger on the helical residues ( 1000 kJ/mol ) , with values chosen by matching the Cα fluctuations to those of a 50 ns , all-atom simulation of the same system [19] . The mean and standard deviation of the RMSD of the TM regions , and the whole receptor , calculated over all the simulations and reported in Table S3 for each system , demonstrate that the proteins maintained reasonably native conformations in the TM regions . Metadynamics simulations [39] were carried out to generate the starting configurations for the umbrella sampling [40] simulations . During these metadynamics simulations , Gaussian bias was only applied to the CV describing the distance between protomers ( r up to values representative of a protomeric system , see Table S2 ) , and the CVs describing the relative rotation of the protomers ( θa and θb ) were restrained to ensure the starting structures all corresponded to interactions at the interface of interest only . Thus , we limited the sampling of the two rotational angles , θa and θb , to a predefined interval using upper and lower steep repulsive restraining potentials . Specifically , the upper and lower limits of this interval were set equidistant either side of the starting values from the initial dimeric configurations ( see Table S1 for the range of θa and θb ) . Approximately 40 umbrella sampling windows , were prepared for each of the dimeric systems with a force constant of 2400 kcal/ ( mol⋅nm2 ) , ( see Table S2 for range of separation , r ) , and extra windows were included at values of r where the reweighted distribution of the distances was found to be insufficiently sampled . We combined umbrella sampling [40] with well-tempered metadynamics [39] simulations to ensure thorough exploration of both the distance and angle space available to the system , for each of the windows . Thus , in addition to the constant external harmonic bias of the umbrella sampling algorithm applied to CV1 ( i . e . , r ) , a time-dependent sum of Gaussian biases in the well-tempered metadynamics algorithm was applied to the angle CVs ( i . e . , θa and θb ) . In contrast to standard metadynamics , the bias potential in well-tempered metadynamics does not fully compensate the free energy surface , but rather depends on the underlying bias , decreasing to zero when a given energy threshold is reached [39] . Thus , not only does the computational effort remain focused on the physically relevant regions of the conformational space in these simulations , but also the convergence of the algorithm to a correct free energy profile can be proven rigorously . To ensure proper sampling , we checked that the chosen CVs could diffuse across one Gaussian size within the deposition time , so that local instantaneous equilibrium of the CVs would be satisfied . An improper choice of the bias update rate and Gaussian size would have resulted in trajectories failing to show multiple transitions across different minima and dependence on the initial starting configuration of the systems . None of the above was observed in our simulations , which showed instead increasingly fast diffusion of the CV dynamics due to the flattening of the underlying free energy surface , suggesting a proper sampling of the phase space of the systems . In all cases , the initial height of the biasing Gaussians was set to 0 . 12 kcal/mol , with a deposition stride of 10 ps , σM = 0 . 035 radians , and a bias factor of 15 . Each window simulation was run for at least 1 µs ( and up to 2 µs in regions of r thought to require additional sampling ) , resulting in a cumulative simulation time of ∼40 µs for each system , and a total of approximately 160 µs for the two receptor systems . Finally , using a model potential , we provide ( in the SI section ) validation that the accuracy of the method combining umbrella sampling and metadynamics is comparable to those of standard multidimensional umbrella sampling or metadynamics ( see both corresponding SI text and Fig . S5 ) . The well-tempered metadynamics bias acting on the angle CVs distorts the probability distribution of the distance CV , thus requiring reweighting before equilibrium Boltzmann distributions could be reconstructed with WHAM . To recover the unbiased probability distribution of the distance CV from well-tempered metadynamics , we used the reweighting algorithm originally derived in [35] and direct the reader to the SI section for a description of this algorithm , as well as for the parameters used and additional technical details . After recovering the unbiased probability distribution of the distance CV from well-tempered metadynamics , we used the well-documented WHAM technique [41] , [42] to reconstruct , for each simulated system , the free energy surfaces as a function of the separation , r , between protomers ( Fig . 1 ) ; the technical details are provided in the SI section . An error analysis of the reconstructed free energies was carried out combining recently proposed methods for the error estimation in umbrella sampling [43] and well-tempered metadynamics [44] , [45] simulations . Equations used to estimate these errors are reported in the SI section . For each of the dimers at the primary minima ( in Fig . 1 ) , we have also reconstructed the FES as a function of the angles θa and θb . To reweight the angle distribution at a fixed protomer separation r , i . e . , to remove the umbrella bias and obtain unbiased free energies as a function of the angles , we used the same algorithm as above [35] , but we accounted for the umbrella potential as an external potential . Fig . S3 shows the FES as a function of ( θa , θb ) for the homodimers of the receptors at the two different interfaces . The minima marked Θ1 , Θ1' and Θ2 are the principal minima from which representative minimum structures were extracted . Upon derivation of the FES as a function of r and the angles ( θa , θb ) , for each dimeric system , we extracted a representative frame from the trajectory that corresponded to the minima therein . We then used Pulchra [46] to convert each of the CG models to an atomistic representation . These representations were solvated in an atomistic POPC/10% cholesterol membrane , energy minimized and simulated with harmonic restraints of decreasing strength applied to Cα atoms for a few picoseconds . Where necessary , we used an adiabatic biased MD simulation ( of ∼5 ns ) to improve the integrity of the helical structure of the re-converted receptors , by steering the Cα of the transmembrane regions toward the original atomistic structure of the B1AR or B2AR ( i . e . , that prior to coarse-graining ) , before 1 ns of unrestrained , all-atom simulation , during which the analyses to obtain the lists of contacts presented in Table S1 were conducted . We can compare the strength of dimerization at each of the interfaces by the relative values of ΔGX° . In equation 1 , we remind the reader of the formulation for ΔGX° in equation 5 of [19] . Specifically , the mole fraction standard free energy change can be expressed as: ( 1 ) where R is the universal gas constant , T is the temperature in Kelvin and KX is the association equilibrium constant on the mole fraction concentration scale . Following our derivation in [19] , KX is approximately equal to: ( 2 ) where NL is the number of lipids in the membrane of area A , and KD is the dimerization constant expressed as in surface concentration units . For our membrane patch , NL/A was ≈1 . 65×106 µm−2 . Using the theory originally proposed by Roux and colleagues [47] and adapted by us to the case of GPCR dimers [19] , the dimerization constant can be expressed as a function of the free energy F ( r ) of the system constrained to a predefined angular region Ω0 , where r is the distance between the interacting protomers , Ω = ( θa , θb ) describes their relative orientation , and β = 1/kBT . This correction was necessary because the relative orientation between protomers had been constrained into a region Ω0 . In such a case , by extending the integral up to the maximum distance rD allowed for dimeric states , KD is given by equation ( 3 ) : ( 3 ) Here , ||Ω0|| , i . e . , the product of the allowed ranges for θa and θb in radians ( see SI and details in Table S2 ) , is ( maxθ-minθ-σM ) 2≈0 . 512≈0 . 26 at TM4/3 interfaces , and 0 . 462 = 0 . 22 at TM1/H8 interfaces . The above theory can be used to estimate the kinetics of dimerization by approximating the diffusion-limited association and dissociation rates , kon and koff , at long timescales , following the Smoulchowski theory in two dimensions [48] ( i . e . , the membrane plane ) . This derivation is the same as that described in [19] . kon is given by equation 4: ( 4 ) where DC is the sum of the diffusion constants of the two protomers , R is the sum of the protomer ratios , and γ is the Euler-Mascheroni constant and t refers to the experimental timescale of diffusion . Subsequently , koff = kon/KD . An initial concentration of dimers of [D]0 evolves over time to give a concentration of:M ( 5 ) and thus the half-life of dimers can be estimated using: ( 6 )
|
G Protein-Coupled Receptors ( GPCRs ) are the largest family of membrane proteins targeted by drugs in clinical practice . Despite being at the forefront of biomedical research for many years , there is still considerable uncertainty about how GPCRs function at a molecular level . Although substantial evidence exists in support of their association in cell membranes , it is unclear how general and/or long-lasting this phenomenon is and whether it plays a significant role in GPCR function . This observation highlights the importance of understanding the rules that govern receptor-receptor interactions in living cells . Here , we report the results of computer simulations from which we estimated the relative stability of dimers formed by different , yet highly homologous , prototypic GPCRs . Our results suggest overall transiency in receptor-receptor interactions at the simulated different dimerization interfaces , but a variable strength of association depending on the specific residue composition or shape of the interface . The methodology we propose is expected to provide a level of molecular detail that is unattainable using current experimental techniques . Our ultimate goal is to generate unique hypotheses of receptor-receptor inter-helical interactions that can be tested experimentally to help elucidate the role of receptor association in GPCR function .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology",
"computational",
"biology",
"biophysics",
"simulations",
"biophysics"
] |
2012
|
Assessing the Relative Stability of Dimer Interfaces in G Protein-Coupled Receptors
|
Host responses to infection encompass many processes in addition to activation of the immune system , including metabolic adaptations , stress responses , tissue repair , and other reactions . The response to bacterial infection in Drosophila melanogaster has been classically described in studies that focused on the immune response elicited by a small set of largely avirulent microbes . Thus , we have surprisingly limited knowledge of responses to infection that are outside the canonical immune response , of how the response to pathogenic infection differs from that to avirulent bacteria , or even of how generic the response to various microbes is and what regulates that core response . In this study , we addressed these questions by profiling the D . melanogaster transcriptomic response to 10 bacteria that span the spectrum of virulence . We found that each bacterium triggers a unique transcriptional response , with distinct genes making up to one third of the response elicited by highly virulent bacteria . We also identified a core set of 252 genes that are differentially expressed in response to the majority of bacteria tested . Among these , we determined that the transcription factor CrebA is a novel regulator of infection tolerance . Knock-down of CrebA significantly increased mortality from microbial infection without any concomitant change in bacterial number . Upon infection , CrebA is upregulated by both the Toll and Imd pathways in the fat body , where it is required to induce the expression of secretory pathway genes . Loss of CrebA during infection triggered endoplasmic reticulum ( ER ) stress and activated the unfolded protein response ( UPR ) , which contributed to infection-induced mortality . Altogether , our study reveals essential features of the response to bacterial infection and elucidates the function of a novel regulator of infection tolerance .
To combat infection , a host activates a combination of immune and physiological responses . While detection of microbial presence is sufficient to stimulate the innate immune response , physiological responses to infection occur as a consequence of microbial growth and virulence , and can therefore be very specific to the particular bacterium the host interacts with . Despite a growing body of literature on immunity , our knowledge of the different host processes that are activated or repressed in response to infection , and of how such responses contribute to host survival , remains limited . To identify new biological processes required to survive infection and to determine how specific or generic the immune and physiological responses to infection are , we surveyed changes in the transcriptome of Drosophila melanogaster in response to infection with 10 bacteria that span the spectrum of virulence . Drosophila is a leading model system for studying how hosts respond to infection at the organismal level . To overcome infection , the fly relies on cellular and humoral innate immune responses . The cellular response consists of phagocytosis and encapsulation [1 , 2] . The humoral response includes the pro-phenoloxidase cascade , which leads to the generation of reactive oxygen species and clotting , as well as the production of antimicrobial peptides ( AMPs ) primarily by the fat body , an organ functionally analogous to the liver and adipose tissues of mammals [3–5] . In the early 2000s , microarray studies characterizing the transcriptional response to bacterial infection were conducted in Drosophila [6–8] . These experiments were based on infection with two non-pathogenic bacteria , Micrococcus luteus and Escherichia coli . This approach successfully identified a set of genes that are differentially expressed upon infection , which became known as the Drosophila Immune-Regulated Genes ( DIRGs ) . A majority of the DIRGs were functionally assigned to specific aspects of the immune response—phagocytosis , antimicrobial peptide synthesis , and production of reactive oxygen species among others [6] . These studies also confirmed that the Toll and Imd pathways are the major regulators of the immune response in Drosophila , and that both pathways direct expression of the majority of DIRGs [7] . In this model , the host response depends on the sensing of two microbe-associated molecular patterns ( MAMPs ) : Lys-type peptidoglycan from Gram-positive bacteria , which activates the Toll pathway , and DAP-type peptidoglycan from Gram-negative bacteria , which induces the Imd pathway [9–11] . Upon activation , each pathway goes on to regulate a subset of DIRGs . More recently , new findings have expanded our insight into the Drosophila response to infection . First , the Toll and Imd pathways can also be activated by virulence factors and damage-associated molecular patterns ( DAMPs ) [12–16] . Additionally , biological processes that would not be considered as classic immunological responses , such as tissue repair and regulation of metabolism , are clearly modulated by pathogenic infection [17–20] . These observations beget the idea that microbial virulence—the relative capacity of a microbe to cause damage in a host—could be an important factor in shaping the host response , and suggest that survival from pathogenic infections may require additional biological processes beyond those that are currently known [21] . In this study , we aimed to identify a comprehensive list of genes regulated by pathogenic and avirulent infections , and to determine what responses are general or specific to each infection . To that purpose , we used RNA-seq to profile the D . melanogaster transcriptomic response to systemic infection with 10 different species of bacteria that vary in their ability to grow within and kill the host . We found that each bacterium elicits a unique host transcriptional response . However , we also identified a small set of core genes that were differentially regulated by infection with the majority of microbes . These genes are involved in a variety of immune and non-immune functions , and a fraction of them remained highly expressed even after bacteria were cleared from the host . Among the core genes was CrebA , a Creb3-like transcription factor . CrebA expression is upregulated through both Toll and Imd signaling in the fat body following infection . Knockdown of CrebA significantly increased mortality from bacterial challenge but did not alter bacterial load , indicating that CrebA contributes to host tolerance of infection . CrebA regulates multiple genes involved in the secretory pathway , and the loss of CrebA triggered ER stress upon infection . This suggests that the CrebA tolerance phenotype may arise through protection from cellular stress during the rapid and dramatic response to infection .
We began by assembling a panel of bacteria to probe the host response to infection . We selected bacteria that span the spectrum of virulence ( from 0% to 100% mortality ) , focusing on microbes that are commonly used by the D . melanogaster research community and ensuring that we included bacteria with Lys-type or DAP-type peptidoglycan ( PGN ) in each virulence level . To assess the relative virulence of each bacterium , we measured host survival and bacterial load over time following infection ( Fig 1A , S1 and S2 Figs ) . The bacteria with the lowest levels of virulence—Escherichia coli ( Ec ) , Micrococcus luteus ( Ml ) , and the Type strain of Serratia marcescens ( Sm ) —caused less than 10% mortality and did not grow past initial inoculum levels in the host . Bacteria exhibiting intermediate levels of virulence—Pectinobacterium ( previously known as Erwinia ) carotovora 15 ( Ecc15 ) , Providencia rettgeri ( Pr ) , and Enterococcus faecalis ( Ef ) —showed the ability to proliferate within the host and killed 15% to 55% of infected hosts . Highly virulent bacteria—Staphylococcus aureus ( Sa ) , Providencia sneebia ( Ps ) , Serratia marcescens strain Db11 ( Db11 ) , and Pseudomonas entomophila ( Pe ) —caused 100% mortality in less than 96 h ( Fig 1A ) . M . luteus , E . faecalis , and S . aureus are Gram-positive bacteria ( Lys-type PGN ) ; all others are Gram-negative ( DAP-type PGN ) . Bacterial load time course experiments revealed differences between bacterial species in their ability to grow and persist within the host . For example , only M . luteus and Ecc15 were eliminated from the host ( i . e . their levels fall below our detection threshold of ~30 CFU/fly ) by 132 h post-infection . In the case of Ecc15 , most but not all hosts were able to clear the infection ( S2A and S2F Fig ) . Neither E . coli nor S . marcescens Type increased in density , but the bacteria persisted inside the host at ~210 bacteria/fly even after 5 days of infection ( S2D and S2E Fig ) . P . rettgeri and E . faecalis grew during the first 24 h of infection , killing a fraction of the hosts . The flies that survived these infections remained chronically infected with ~210 to 213 bacteria per fly ( S2B and S2G Fig ) for at least 5 . 5 days . P . entomophila , S . aureus , S . marcescens Db11 , and P . sneebia all grew monotonically in the host until death occurred ( S2C , S2H , S2I and S2J Fig ) , causing complete mortality within 96 h ( S1E , S1J , S1K and S1L Fig ) . Having assembled our panel of bacteria , our next goal was to select relevant time points for transcriptomic analysis . Using our survival and bacterial load data , we identified three time points that are characteristic of different stages of infection: 12 , 36 , and 132 h . At 12 h post-infection , all flies remain alive , and they face the initial growth of microbes . Thirty-six hours represents an intermediate time point during infection , after the highly virulent bacteria have killed most or all flies and the moderately virulent bacteria have killed 15% to 55% of infected hosts . Finally , at 132 h post-infection ( 5 . 5 days ) , surviving flies are chronically infected with moderate to low levels of bacteria . To identify novel biological processes required to survive systemic infection , and to assess the level of specificity of the Drosophila response to microbes , we used RNA-seq to profile the D . melanogaster transcriptome after infection with each of our 10 experimental bacteria . We additionally included the following controls: unchallenged flies ( UC ) , flies challenged with a sterile wound ( SW ) , and flies inoculated with heat-killed E . faecalis ( Ef HK ) or heat-killed P . rettgeri ( Pr HK ) . The purpose of the controls was to distinguish the response to live bacteria from that to aseptic injury and/or inert bacterial compounds ( MAMPs ) provided by the injection of dead bacteria . The expression value dataset for the entire experiment can be downloaded or accessed online in our associated database Flysick-seq ( http://flysick . buchonlab . com ) We first determined the overall transcriptomic differences between flies infected by each of the 10 bacteria . Principal component analysis ( PCA ) showed that all three biological replicates clustered together , indicating good replicability of the response for each pathogen ( illustrated in Fig 1B and 1C for the 12 h time point and S3A Fig for the full data set ) . In total , we identified 2 , 423 genes ( 13 . 7% of the genome ) that were differentially expressed upon infection . Of these , 1 , 286 genes were upregulated and 1 , 290 genes were downregulated in response to at least one bacterial infection and time point ( Fig 2A and S3B Fig ) . Out of the total number of genes differentially regulated by all 10 live infections , more genes were upregulated than downregulated; 6 . 1% of the 1 , 286 upregulated genes were induced in all bacterial infections , while only 0 . 6% of the 1 , 290 downregulated genes were repressed by all 10 bacteria ( S3B Fig ) . We also determined that 51 . 1% of the downregulated genes were repressed in only one bacterial infection , while 38 . 6% of the upregulated genes were induced by a single bacterial condition ( S3B Fig ) . These data suggest that the host response to infection is highly specific to individual bacteria , but that there is also a core set of genes that are differentially expressed during most bacterial infections . Additionally , our data showed that downregulated genes tend to be unique to each infecting bacterium , perhaps reflecting the singular consequences of each infection to host physiology ( S3B Fig ) . In general , the largest number of differentially expressed genes was observed at 12 h post-infection . However , a substantial number of genes continued to be differentially regulated at 36 h and 132 h post-inoculation ( Fig 2A ) , presumably in part because the hosts continue to carry their bacterial infections at these later time points and/or because infection induces long-term changes in host physiology . Samples for the 36 h and 132 h time points were not available for infections with the highly virulent bacteria because they rapidly killed all their hosts . For the remaining infections , however , the number of upregulated genes at 12 h after infection was 1 . 6 times higher than the average number of genes that continued to be induced at 36 h and 132 h post-infection . Likewise , there were 2 . 8 times as many downregulated genes at 12 h post-infection than there were at later time points . These results demonstrate that the early transcriptional response to infection is larger than the sustained one , probably because the early response includes both an injury-induced transcriptional regulation and an aggressive initial immune response that is not yet tuned to bacterial titer or growth state within the host [22] . We sought to investigate the source of differences in the host response to various infections . We began by looking at the number of genes regulated by the host in response to each bacterium . The number of differentially regulated genes fluctuated considerably across bacterial infections ( Fig 2A ) . Flies inoculated with heat-killed E . faecalis and P . rettgeri , as well as flies challenged with avirulent bacteria , such as E . coli and M . luteus , induced the lowest number of genes . However , the number of genes regulated in the host did not directly correlate with the level of bacterial virulence . For example , despite the fact that both bacteria rapidly killed all flies , infection with S . aureus differentially regulated the expression of 1 , 193 genes , while P . sneebia infection altered the transcription of only 187 genes ( Fig 2A ) . In addition , there was a large variability in the number of genes regulated in response to different benign bacteria . Across all time points , M . luteus infection changed the expression of 794 genes , while E . coli infection affected only 446 genes ( Fig 2A ) . These results indicate that the breadth and the specificity of the host transcriptomic response is largely independent of virulence . Next , we aimed to identify specific genes that underlie the transcriptomic differences in response to distinct infections . We focused on the first two principal components of our PCA analysis ( Fig 1B ) , which respectively explain 34 . 0% and 27 . 2% of the variance in gene expression . We found that 73 of the top 100 genes contributing to the first principal component ( PC1 ) and 75 of the 100 genes contributing most to the second principal component ( PC2 ) are known targets of the Toll or Imd pathways ( Fig 1C and S4 Fig ) , confirming that these two pathways are key regulators of the specificity of the host response [7] . The genes that contributed most to PC1 included antimicrobial peptide genes ( Dpt , AttA , Drs , and Mtk ) as well as signaling components of the Toll ( Spz and PGRP-SA ) and Imd ( PGRP-LC , PGRP-SD , PGRP-LB , and Rel ) pathways themselves ( Fig 1C ) . Additionally , the expression of Turandot genes , stress peptides regulated by the JAK-STAT pathway , was strongly variable between infections , indicating that differential activation of the JAK/STAT pathway also contributes to PC1 . Interestingly , metabolic genes involved in lipid synthesis ( ACC ) , the Leloir pathway ( Galk ) , and trehalose and glycogen synthesis ( Tps1 , UGP , and Hex-C ) were downregulated to different levels depending on the infection , indicating that different bacteria alter host metabolism in unique ways . In general , PC1 appeared to reflect the transcriptional magnitude of the response to infection . Genes that contributed most to PC2 include target genes of the Toll pathway , including melanization and coagulation-related genes ( MP1 and fondue ) ( Fig 1C ) , as well as immune-induced proteins of the IM cluster . PC2 also included genes downregulated by infection that are involved in sugar digestion ( i . e . the Maltase cluster ) , as well as P450 enzymes known for their functions in oxidoreduction reactions ( i . e . Cyp genes ) . Flies infected with Gram-positive bacteria ( Lys-type PGN ) and Gram-negative bacteria ( DAP-type PGN ) were separated from each other on PC2 , confirming that the type of bacterial peptidoglycan is a major parameter influencing the global response to infection ( Fig 1B , S3A Fig ) [7] . A heatmap showing the expression level of genes that contribute the most to each PC can be found in S4 Fig . Subsequently , we asked whether any differentially regulated genes were unique to a specific bacterial condition . We defined unique genes as those that significantly changed their expression in one and only one infection condition , regardless of time points , thus reflecting the response to a particular bacterium rather than temporal variations in the response to this bacterium . Without exception , we found that infection with each bacterium regulates an exclusive set of genes . The number of uniquely regulated genes varied dramatically across bacterial infections ( Fig 2A ) . For instance , P . sneebia infection resulted in unique regulation of only 6 genes , whereas S . aureus infection exclusively regulated 336 genes . In order to determine what portion of the host response is specific to individual bacteria , we calculated the percentage of differentially expressed genes that were unique to each infection ( S3C Fig ) . We found that this number also differs widely between bacteria . For instance , 20 . 1% of genes upregulated in response to S . aureus were exclusive to this infection , while only 7 . 1% of genes upregulated by E . faecalis infection were unique to this condition . Evaluating Gene Ontology ( GO ) terms associated with the genes uniquely altered by individual infections revealed bacteria-specific responses in some infection conditions ( Fig 2A ) . For example , S . aureus infection induced apoptosis-related genes and downregulated genes involved in glutathione and carboxylic acid metabolism . In contrast , infection with P . entomophila upregulated genes involved in epithelial cell proliferation and strongly decreased the expression of genes associated with cellular respiration and the electron transport chain . At the same time , infection by P . rettgeri specifically downregulated genes involved in the translation machinery ( Fig 2A ) . All the GO gene categories we identified are linked to stress responses that aim to maintain cell homeostasis ( cell death and tissue repair ) or metabolic homeostasis , suggesting that the unique physiological and virulence interactions of each bacterium with the host induce a specific set of organismal responses . Altogether , our results demonstrate that the host response to infection is shaped by a combination of immune potency , metabolic impact , and physiological alteration . Next , we set out to identify the core set of genes that are regulated in response to most or all bacterial infections . We defined the core genes as those that are differentially expressed in response to 7 or more bacteria on at least one time point post-infection . We set the cutoff at 7 bacteria because we were concerned that requiring differential expression in response to all 10 infections would be overly restrictive . Specifically , we had reservations about the artificial omission of genes in cases where the bacteria are rapidly cleared from all or most hosts ( e . g . M . luteus and Ecc15 ) and in cases where the bacterium might suppress or evade the canonical response ( e . g . P . sneebia; [23] ) . Using these criteria , we identified a core response of 252 genes . This included 166 upregulated genes ( Fig 3A and S1 Table ) and 86 downregulated genes ( Fig 3B and S1 Table ) . The set of core genes is fairly robust to the criteria for inclusion , decreasing only to 135 genes induced and 54 genes repressed when inclusion required differential expression in response to 8 of the bacterial conditions . Similarly , the numbers increased only to 216 genes induced and 136 repressed when inclusion was relaxed to 6 of the bacterial infections . Within the core , 78 genes were also regulated in response to sterile wound alone or to challenge with heat-killed bacteria ( Fig 3C ) . Most of the genes regulated by injury were also regulated by challenge with live or dead bacteria ( 96/114 genes ) , which is congruent with the fact that the infection method inherently inflicts injury . However , the core response to live infection was markedly distinct from the response to heat-killed bacteria . Of our core genes , ~40% ( 105/252 genes ) were differentially expressed in response to live infections but not in response to challenge with heat-killed bacteria . Moreover , we found 493 genes that were differentially regulated by treatment with heat-killed bacteria but were not part of the core response to live infection ( Fig 3C ) . Of those 493 genes , 164 were uniquely regulated in response to heat-killed bacteria and not in response to any live infection ( S5A Fig ) . To determine whether genes exclusively regulated in response to heat-killed bacteria are simply artifacts of weak statistical detection , we relaxed the cutoff to a False Discovery Rate ( FDR ) <0 . 1 for classifying a gene as differentially expressed during infection . Even with this more lenient threshold , 61 . 6% of the 164 genes that were uniquely regulated in response to heat-killed bacteria were still not differentially regulated in response to any live infection . Our results , therefore , not only show that the response to live infections is fundamentally different from the biological challenges that simple injury and immune activation pose , but also demonstrate that challenge with dead bacteria induces a response that does not occur as a consequence of infection by live bacteria . In 2001 , a study identified a set of genes that are differentially expressed after infection with a combination of E . coli and M . luteus [6] . These genes became known as Drosophila Immune-Regulated Genes ( DIRGs ) . We compared our set of 252 core response genes to the 381 DIRGs and found that only 84 of them were previously identified as DIRGs ( S5B Fig ) . Intriguingly , the DIRGs identified in the previous study included 279 genes that were neither in our core response nor regulated by challenge with heat-killed bacteria ( S5B Fig ) , and 246 of these DIRGs were not induced in the present study even by infection with M . luteus or E . coli ( S5C Fig ) . These discrepancies may originate from differences in Drosophila genotype or rearing conditions , bacterial genotype , or experimental variation . Alternatively , they could imply that infection with a mixture of two bacteria can lead to the activation of a specific set of genes , different from each mono-microbial infection . When we compared our total number of differentially regulated genes ( 2 , 423 ) to the DIRGs , we found that our study has identified 2 , 197 novel infection response genes , including 168 new core genes . Thus , our data offer a more comprehensive list of infection-responsive genes that is expanded both because of the sensitivity of RNA-seq technology over the previous microarrays and because of the broader diversity of bacteria used in our experiment . To investigate the biological functions of our newly identified core response genes , we evaluated GO categories enriched in the core ( Fig 3A and 3B ) . Upregulated core genes were primarily annotated with immune functions , such as Toll pathway and defense response to Gram-negative bacteria . This group also included genes involved in metabolism , including glycosaminoglycan metabolic process , carbohydrate metabolism , and metal ion transport . Additionally , core upregulated genes have a role in cellular and tissue processes , with genes acting in tissue repair , response to oxidative stress , cellular homeostasis , co-translational protein targeting to membrane , and protein targeting to ER ( Fig 3A ) . The core downregulated genes were annotated with functions such as oxidation-reduction and starch and sucrose metabolism ( Fig 3B ) . Core genes can be separated into two groups: genes regulated in response to live infections only and genes regulated in response to both live infections and heat-killed bacteria ( Fig 3C ) . The 78 core genes that were also differentially expressed in the wound-only control and in the heat-killed bacteria control included genes coding for AMPs , PGRPs , Turandot ( Tot ) genes , and other classical targets of the Toll and Imd pathways [24] . Genes regulated only in response to live infection included key transcription factors of the immune system , such as Rel and dl , and were associated with biological processes such as metabolism , oxidation-reduction , regulation of iron ion transmembrane transport , and secretion . Altogether , these data indicate that heat-killed bacteria mostly trigger classically defined immune responses , while live infections regulate a set of additional biological processes that presumably reflect physiological interactions between the host and invading pathogen . These processes , including metabolic rewiring , response to stress and damage , cellular translation , and secretion , could act as physiological adaptations or buffers to the stress and damage imposed by infection . The hypothesis that D . melanogaster has a distinct response to infection by Gram-positive ( Lys-type PGN ) versus Gram-negative ( DAP-type PGN ) bacteria dominated the field for most of the 1990s and 2000s [25] . To address this hypothesis , we characterized the transcriptional response to Gram-positive versus Gram-negative bacterial infection in our study . We found that 662 genes are regulated only by infection with Gram-positive bacteria , 851 genes are regulated only by Gram-negative infection , and 1 , 063 genes are regulated by infections with bacteria of both Gram types ( Fig 3D ) . Of the 662 genes exclusively regulated by Gram-positive bacteria , only 20 ( Cyp309a1 , daw , CG31326 , etc . ) are upregulated and 8 are downregulated by all three Gram-positive bacteria . Similarly , amongst genes regulated specifically by Gram-negative bacteria , only 1 gene is upregulated ( AttD ) and no genes are downregulated in response to all 7 Gram-negative bacteria . Our data suggest that the stereotypical response to Gram-negative infection also occurs as a consequence of Gram-positive infection , such that there is no large cohort of genes responding exclusively to Gram-negative infection . To confirm this , we performed RT-qPCR on Dpt and Drs transcripts as a proxy for activity of the Imd and Toll pathways , respectively [7] . We found that infection by most of our 10 bacteria induced both pathways , although to significantly different levels ( S6 Fig ) . Our results generally confirm the notion that the Toll pathway is more responsive to infection with Gram-positive ( Lys-type PGN ) bacteria and the Imd pathway is more reactive to infection with Gram-negative ( DAP-type PGN ) bacteria , but also make clear that the differences in pathway activation are quantitative and not qualitative or binary . Since the bacteria belonging to the low and intermediate virulence categories do not kill all hosts , we followed the dynamics of gene expression in surviving hosts over several days . In particular , we aimed to contrast the sustained transcriptional response of flies that had cleared their infections to undetectable levels ( i . e . after infection with M . luteus or Ecc15 ) to that of flies carrying chronic infections ( i . e . E . coli , S . marcescens Type strain , P . rettgeri and E . faecalis ) . We hypothesized that persistent bacteria would continue to elicit a response from the host , which would be absent in flies that have cleared all bacteria . To test this idea , we determined whether genes “recover” from bacterial infection . We defined recovery in terms of gene expression: a gene that has recovered is differentially expressed at 12 and/or 36 h post-infection but returns to pre-infection levels by 132 h ( 5 . 5 days ) after inoculation . We found that , on average , a minimum of 50% of the genes that are differentially regulated by each infection returned to basal levels by our last time point ( Fig 4A ) , and this was the case even in hosts infected with persistent infections . The percentage of genes that fully recovered was substantially higher in moderately virulent infections ( P . rettgeri: 79 . 8% and E . faecalis: 71 . 1% ) than in benign infections ( E . coli: 54 . 5% and S . marcescens Type: 55 . 5% ) , perhaps in part as a consequence of the higher number of genes induced upon infection with these bacteria ( Fig 4A ) . Surprisingly , we also observed that only 56 . 5% and 58 . 7% of genes recovered in M . luteus and Ecc15 infections , respectively , even though the majority of hosts ( ≥85% ) survive these infections and the bacteria are eliminated ( i . e . their levels fall below our detection threshold ) within two days . These results demonstrate the complexities of the transcriptional response to infection . While there can be a substantial lingering transcriptional effect in flies that successfully cleared an infection , a subset of differentially regulated genes may return to basal levels even in chronically infected flies that continue to carry bacteria . Next , we evaluated how the core upregulated and downregulated genes change in expression level over time ( Fig 4B and 4C ) . We quantified the degree of recovery for each gene by comparing the fold change in expression at 132 h after infection to the fold change at either 12 h or 36 h , whichever was the highest if the gene was upregulated or the lowest if the gene was downregulated . In general , sterile wound and challenge with heat-killed bacteria resulted in the regulation of fewer core genes than live infection , and most of these genes recovered to pre-infection expression levels by 132 h post-challenge ( Fig 4B and 4C ) . Core genes induced by Ecc15 and M . luteus showed similar kinetics , and most genes had recovered or were on their way to recovery by 132 h , suggesting that the core response is not sustained in the absence of these bacteria . In contrast , core genes induced by S . marcescens Type , P . rettgeri , and E . faecalis did not recover as much , in agreement with the idea that infections with persistent bacteria continuously stimulate the core response . This paradigm was , however , not true for downregulated genes , as most downregulated genes did recover or were in the process of recovery by 132 h regardless of which bacteria was used for infection . Interestingly , we noticed that a group of genes did not recover at all in most conditions but continued to be upregulated over time ( boxed in Fig 4B ) . These included effector genes of the immune response ( AttD and Tep4 ) , regulators of iron homeostasis ( Tsf1 and MtnD ) , and negative regulators of the immune response ( PGRP-LB , nec ) . Genes like SPH93 and su ( r ) never returned to their basal expression levels in flies infected with Ecc15 or M . luteus . Additionally , while the transcript levels of most antimicrobial peptide genes decreased over time , they never returned to basal , pre-infection levels , suggesting that the effect of infection lingers for several days after bacteria are eliminated . Having identified a core transcriptional response to infection , we set out to find key regulators of that response . We used i-cisTarget to identify transcription factor binding motifs enriched in the regulatory regions of our core genes [26 , 27] . Using this approach , we found enrichment in putative binding sites for Relish ( Rel ) , Dif/Dorsal , Schnurri ( Shn ) , CrebA , Atf6 , Xbp1 , and Tbp in the regulatory regions of upregulated core genes ( Fig 5A and S2 Table ) . Dif/Dorsal and Relish are the terminal transcription factors of the Toll and Imd pathways , respectively; therefore , finding enrichment for their predicted binding sites is in agreement with the central role that these pathways play in the immune response . Our data also agree with published reports showing that the TGF-beta pathway upstream of shn and the Atf6 transcription factor are important to survive infection [28 , 29] . Transcription factor binding site enrichment analysis of the repressed genes revealed putative binding sites for the Lola and GATA transcription factors ( S3 Table ) . In addition to the i-cisTarget analysis , we searched for genes encoding transcription factors within our list of core upregulated genes . We identified 3 transcription factors in the defined core that are upregulated themselves by infection: Rel , dorsal , and CrebA ( Fig 5B ) . Although Dif is required to activate Toll pathway signaling in response to bacterial infection in Drosophila adults and dorsal is not [30] , we surprisingly found that Dif is not significantly upregulated in response to any of the 10 bacteria tested . CrebA is the single Drosophila member of the Creb3-like family of transcription factors [31] . We found the predicted DNA motif bound by CrebA ( TGCCACGT , see Fig 5A for position weight matrix [32] ) in 71 genes upregulated by infection , including 18 upregulated core genes ( S7A Fig ) . CrebA is itself significantly induced upon infection by all 10 bacteria ( Figs 3A and 5B ) . To validate our RNA-seq results on CrebA expression , we infected a new group of flies with P . rettgeri and E . faecalis and measured CrebA transcript levels at 12 h post-inoculation . In agreement with our RNA-seq data , we confirmed that CrebA expression is upregulated in response to infection with P . rettgeri ( p = 0 . 0026 ) and E . faecalis ( p = 0 . 0147 ) ( S7B Fig ) . These results demonstrate that CrebA is a transcription factor induced by infection and is potentially a key regulator of the core response . To identify the molecular mechanisms that control CrebA transcription in response to infection , we scanned 2 kb upstream and 2 kb downstream of the CrebA transcription start site for potential transcription factor binding sites using MatInspector ( Genomatix ) [33] . Within this region , we found an enrichment of putative binding sites corresponding to the transcription factors Dif/Dorsal and Relish . There were 12 predicted Relish binding sites , 16 predicted Dif binding sites , and 13 predicted Dorsal binding sites flanking the CrebA gene , suggesting that immune pathways may induce the expression of CrebA ( S7C Fig ) . To confirm regulation of CrebA by the Toll and Imd pathways , we quantified CrebA expression by RT-qPCR 12 h after infection with P . rettgeri and E . faecalis in wildtype ( WT ) flies , flies deficient for the Imd pathway ( RelE20 ) , and flies deficient for the Toll pathway ( spzrm7 ) ( Fig 5C ) . CrebA expression was significantly reduced in both RelE20 ( p = 0 . 0456 for P . rettgeri and p = 0 . 0020 for E . faecalis ) and spzrm7 ( p = 0 . 0118 for P . rettgeri and p = 0 . 0026 for E . faecalis ) mutants relative to wildtype controls , indicating that both the Imd and Toll pathways contribute to infection-induced CrebA upregulation . We then tested whether activation of the Imd or Toll pathway is sufficient to upregulate the level of CrebA expression in the absence of infection . Using the temperature-sensitive UAS/Gal4/Gal80ts gene expression system to ubiquitously drive Imd or an active form of Spz ( Spz* ) , we stimulated Imd and Toll pathway activity in adult flies [34 , 35] . Transgenic activation of either the Imd or Toll pathway in the absence of infection was sufficient to significantly increase CrebA transcript levels in D . melanogaster adults ( p = 0 . 0114 for UAS-spz* and p = 0 . 0062 for UAS-imd ) ( Fig 5C ) . Altogether , our results demonstrate that the Imd and Toll pathways are both necessary and sufficient to regulate CrebA transcription upon infection . In order to identify the tissue ( s ) and/or organ ( s ) within the fly that upregulate CrebA expression upon bacterial challenge , we infected wildtype flies with P . rettgeri and dissected out the following tissues and body parts at 12 h post-infection: head , dorsal thorax ( including wings and heart ) , ventral thorax ( including legs ) , digestive tract ( crop , midgut , and hindgut ) , Malpighian tubules , testes , and abdomen ( abdominal fat body ) . The abdomen was the only tissue that exhibited significant upregulation of CrebA as determined by RT-qPCR ( p = 0 . 0315 ) , suggesting that CrebA may be regulated in the fat body upon infection ( Fig 5D ) . We therefore knocked down CrebA expression by RNAi ( via 3 independent RNAi constructs ) using 2 separate fat body drivers , c564-Gal4 and Lpp-Gal4 , and quantified CrebA expression by RT-qPCR in whole flies 12 h after infection with P . rettgeri . The combination of 2 CrebA RNAi constructs ( B and C ) with the drivers fully prevented CrebA induction upon infection with P . rettgeri . In the case of the third RNAi construct ( A ) , CrebA was significantly upregulated by infection with P . rettgeri ( p = 0 . 0002 ) , but the induction was significantly lower ( p = 0 . 0442 ) than the expression level observed in infected wildtype samples ( S7D Fig ) . These data indicate that the cells of the fat body represent the primary site of CrebA induction . In sum , our data suggest that the Toll and Imd pathways regulate the expression of CrebA in the fat body in response to infection . We next asked whether CrebA is required for the host to survive infection . Since strong loss-of-function CrebA mutants are embryonic lethal , we tested the role of CrebA in response to infection by knocking it down in the fat body of adult flies using 3 independent RNAi constructs expressed under the control of the c564-Gal4 driver ( Gal80ts; c564-Gal4 > UAS-CrebA-IR ) and , separately , the Lpp-Gal4 driver ( Gal80ts; Lpp-Gal4 > UAS-CrebA-IR ) [36] . Because the c564-Gal4 driver expresses strongly in both the fat body and hemocytes , we additionally tested the requirement for CrebA in the response to infection in hemocytes ( Hml-Gal4 > UAS-CrebA-IR ) . All CrebA fat body knockdown flies exhibited increased susceptibility to systemic infection with P . rettgeri ( p<0 . 0001 ) ( Fig 6A and S8A Fig ) , while hemocyte-specific knockdown did not lead to any significant increase in mortality ( S8B Fig ) . When CrebA was knocked down in the fat body , nearly 100% of the flies died , and most of the death occurred during the first 24 h following infection . In contrast , almost 50% of control flies survived the infection for at least 7 days ( Fig 6A and S8A Fig ) . To confirm that the survival phenotype observed in CrebA RNAi flies upon infection was solely due to loss of CrebA expression , we co-expressed a CrebA RNAi construct and a CrebA overexpression construct in flies ( Gal80ts; c564-Gal4 > UAS-CrebA , UAS-CrebA-IR ) and infected them with P . rettgeri . We observed no significant difference between the survival of infected control flies and that of infected flies co-expressing both the RNAi and overexpression constructs , indicating that changes in CrebA expression are uniquely responsible for the lowered survival phenotype observed ( Fig 6A ) . We also infected CrebA RNAi flies with E . faecalis and found that CrebA RNAi flies were remarkably more susceptible to infection when compared to control flies ( p<0 . 0001 ) ( Fig 6B ) . In addition , CrebA RNAi flies died at a significantly faster rate than control flies when inoculated with P . sneebia ( p<0 . 0001 ) ( Fig 6C ) . Finally , infection with Ecc15 , S . marcescens Type , and E . coli also killed more flies with CrebA expression blocked in the fat body than controls ( p = 0 . 0013 for Ecc15 , p = 0 . 0004 for S . marcescens Type , and p = 0 . 0028 for E . coli ) ( Fig 6D–6F ) . None of these latter three infections were lethal to wildtype control flies , but approximately 30% of CrebA-deficient flies succumbed to infection . Collectively , our results demonstrate that CrebA is generally required to survive bacterial infection . To test whether the CrebA survival phenotype is due to a failure to control bacterial proliferation ( a resistance defect ) or a decrease in the ability to withstand infection ( a tolerance defect ) , we monitored bacterial load in individual CrebA RNAi and control flies following P . rettgeri infection [37] . We focused our sampling on 1–2 h intervals over the first 24 h of infection , as this is the time when most of the CrebA-deficient flies succumbed . We did not find a significant difference in bacterial load between wildtype and CrebA knockdown flies at any measured time point ( p = 0 . 0664 ) , indicating that CrebA RNAi flies are able to control bacterial load similarly to control flies ( Fig 6G ) . To corroborate these results , we quantified bacterial load following infection with P . rettgeri in flies where CrebA was knocked down by a different RNAi construct and in flies co-expressing a CrebA RNAi construct and a CrebA overexpression construct ( Gal80ts; c564-Gal4 > UAS-CrebA , UAS-CrebA-IR ) . Again , we did not observe any significant difference in bacterial load between wildtype and CrebA knockdown flies ( p = 0 . 3208 ) or between wildtype and CrebA rescue flies ( p = 0 . 3030 ) ( S8C Fig ) . To evaluate whether CrebA knockdown flies are less resistant to other pathogens , we measured bacterial load in individual flies following E . faecalis or Ecc15 infection . In agreement with the results of our P . rettgeri experiments , we did not find a significant difference between wildtype and CrebA-deficient flies at the time points sampled ( p = 0 . 4204 for E . faecalis and p = 0 . 7253 for Ecc15 ) ( S8D and S8E Fig ) , suggesting that CrebA knockdown flies do not have a defect in resistance to infection . We previously demonstrated that flies die at a stereotypical and narrowly distributed bacterial load , the bacterial load upon death ( BLUD ) , which represents the maximum quantity of bacteria that a fly can sustain while alive [38] . We therefore sought to determine whether CrebA RNAi flies have a lower BLUD , which would indicate a reduced tolerance of infection . We quantified the bacterial load of individual flies within 15 minutes of their death and found that CrebA RNAi flies died carrying a significantly lower bacterial load than control flies ( p<0 . 0001 ) ( Fig 6H ) . These data demonstrate that while CrebA-deficient flies control bacterial growth normally , they are more likely to die from infection , and they die at a lower bacterial load than wildtype flies . Therefore , the transcription factor CrebA acts to promote tolerance of infection . In order to identify the complete set of genes directly and indirectly regulated by CrebA upon infection , we performed RNA-seq on the fat bodies of wildtype flies and flies in which we knocked down CrebA in the fat body . We collected samples from both genotypes in unchallenged conditions and 12 h after infection with P . rettgeri . In total , we found that only 104 genes were downregulated in CrebA knockdown fat bodies compared to wildtype fat bodies following infection ( S5 Table ) . These genes were associated with GO categories such as protein targeting to the ER , signal peptide processing , protein localization to the ER , and antibacterial humoral responses . Antimicrobial peptide genes of the Cecropin gene family ( CecA1 , CecA2 , CecB , and CecC ) showed partially reduced induction when CrebA expression was disrupted . Nevertheless , they were still induced to extremely high levels ( >200-fold ) in CrebA knockdown fat bodies ( Fig 7 ) . Other antimicrobial peptide genes , such as Dpt , Drs , Def , and AttC , were expressed at similar levels in CrebA knockdown fat bodies compared to wildtype fat bodies , results corroborated by RT-qPCR analysis ( S9A–S9D Fig ) . In contrast , a number of genes including sugar transporters and multiple lipases were upregulated upon infection in fat bodies deficient for CrebA but not in wildtype fat bodies . These data suggest that CrebA regulates immune , metabolic , and cellular functions during infection . Previously , Fox and colleagues demonstrated that CrebA acts in the Drosophila embryo as a direct regulator of secretory capacity and is both necessary and sufficient to activate the expression of many secretory pathway component genes [32] . We therefore asked whether CrebA controls secretion-related genes upon infection in the adult fat body . We found that the expression level of 32 secretion-related genes significantly increased upon infection with P . rettgeri in wildtype samples . However , the induction of these secretion-related genes was significantly lower ( p<0 . 05 ) in CrebA RNAi fat body samples compared to wildtype fat body controls , a result that agrees with the findings of Fox et al . ( Fig 7 and S5 Table ) . These 32 secretion-related genes we identified included core response genes that are central components of the cell’s secretory machinery , including TRAM , ergic53 , Sec61β , and Spase25 ( Fig 7 ) . Using a separate set of samples from those of the RNA-seq , we further confirmed these findings by measuring TRAM , ergic53 , Sec61β , and Spase25 transcript levels by RT-qPCR in the fat bodies of flies infected with P . rettgeri at 12 h post-infection ( S9E–S9H Fig ) . These four genes were significantly upregulated following infection with P . rettgeri in wildtype samples . However , we were not able to detect a significant increase in the levels of TRAM , ergic53 , and Sec61β in CrebA RNAi fat bodies upon infection . The expression level of Spase25 was significantly induced by infection with P . rettgeri even when CrebA expression was inhibited by RNAi in the fat body ( p<0 . 05 ) , but the induction was significantly lower ( p<0 . 001 ) than the expression level observed in infected wildtype samples ( S9H Fig ) . In sum , our data suggest that CrebA could act to regulate an increase in secretory capacity upon infection . Since our data suggested that CrebA may promote an increase in secretory capacity in the fat body upon infection , we hypothesized that loss of CrebA expression could lead to altered protein secretion or defects in protein transport to the membrane . Accumulation of unfolded proteins or a decrease in protein secretion triggers endoplasmic reticulum ( ER ) stress , which in turn induces stereotypical pathways to limit the stress imposed on the cell . These pathways include IRE1α/XBP1- , PERK/ATF4- , and ATF6-mediated responses termed the unfolded protein response ( UPR ) [39 , 40] . Upon sensing of ER stress , Xbp1 mRNA undergoes alternative splicing via IRE1α; Xbp1 splicing is thus considered to be a marker of ER stress and of the activation of UPR [41 , 42] . To investigate whether loss of CrebA could trigger ER stress in fat body cells upon infection , we quantified the expression levels of both Xbp1t ( total ) and Xbp1s ( spliced ) in abdomens of wildtype and CrebA-knockdown flies under both unchallenged and infected conditions ( Fig 8A and 8B ) . Xbp1s levels did not change upon infection in wildtype samples or differ between wildtype and CrebA RNAi samples in the absence of infection . However , Xbp1s levels spiked dramatically in CrebA RNAi fat body samples ( p = 0 . 0289 ) ( Fig 8B ) after infection , indicating that loss of CrebA upon bacterial challenge triggers ER stress in the fat body . Our data also revealed that Xbp1t expression was significantly higher in CrebA knockdown samples compared to wildtype samples following infection ( p = 0 . 0144 ) ( Fig 8A ) . This result is in agreement with a previous study that suggested Xbp1s regulates Xbp1 transcription [43] . To determine whether ER stress is induced in fat body cells directly , we labelled fat body cells in vivo by expressing a dsRed reporter under the control of the Xbp1 regulatory sequence [44] . In agreement with our RT-qPCR experiments , we found that bacterial challenge did not induce dsRed expression in wildtype samples . However , infected CrebA RNAi fat body cells consistently expressed higher levels of dsRed compared to all other controls ( Fig 8C ) . These results demonstrate that CrebA expression prevents the occurrence of ER stress in the fat body upon infection . We next asked whether the failure of CrebA-deficient flies to prevent ER stress following infection could explain their increased susceptibility to bacterial challenge . To test this , we genetically induced ER stress in Drosophila fat bodies either by overexpression of Psn ( Gal80ts; c564-Gal4 > UAS-Psn ) , which disrupts calcium homeostasis , or by knockdown of BiP ( Gal80ts; c564-Gal4 > UAS-BiP-IR ) , a regulatory protein of the unfolded protein response [45 , 46] . Inducing ER stress in the fat body during infection made the flies more susceptible to P . rettgeri infection , phenocopying the result observed with CrebA knockdown flies ( p<0 . 0001 for both constructs ) ( Fig 8D ) . Since the increased susceptibility of CrebA RNAi flies to infection stemmed from a tolerance defect ( Fig 6G and 6H and S8C–S8E Fig ) , we sought to determine whether the increase in mortality observed in BiP RNAi and Psn overexpression flies following infection is also due to a tolerance deficiency . We monitored bacterial load in individual BiP RNAi and Psn overexpression flies following challenge with P . rettgeri . We did not observe a significant difference in bacterial load between wildtype and BiP-knockdown flies ( p = 0 . 0624 ) or between wildtype and Psn overexpression flies ( p = 0 . 6462 ) ( S10A Fig ) . Quantification of bacterial load upon death ( BLUD ) following P . rettgeri infection in BiP RNAi flies showed that BiP-deficient flies perish carrying a significantly lower bacterial load than wildtype flies ( p<0 . 0001 ) ( S10B Fig ) . Altogether , our data indicate that induction of fat body ER stress during infection decreases fly survival by lowering host tolerance of infection . Having demonstrated that CrebA-deficient flies experience fat body ER stress upon bacterial challenge and that flies with genetically induced fat body ER stress display increased mortality without a concomitant change in bacterial load following infection , thus phenocopying CrebA-deficient flies , we subsequently asked whether alleviating ER stress in CrebA-deficient flies could rescue the CrebA survival phenotype . To test this , we overexpressed BiP in fat body cells in which CrebA was knocked down by RNAi ( Gal80ts; c564-Gal4 > UAS-CrebA-IR , UAS-BiP ) . Previous work has shown that overexpression of BiP can ameliorate ER stress [47] . While overexpression of BiP alone did not alter host survival during infection , expression of BiP in CrebA RNAi flies rescued fly survival upon challenge with P . rettgeri ( Fig 8E ) . We observed no significant difference between the survival of infected control flies and that of infected flies co-expressing both the CrebA RNAi and BiP overexpression constructs ( p = 0 . 2786 ) . These data indicate that reducing ER stress is sufficient to rescue the survival phenotype of CrebA-deficient flies during bacterial challenge . Excessive and prolonged ER stress can lead to apoptosis [48] . Therefore , we investigated whether CrebA RNAi flies are more susceptible to infection due to an increase in fat body cell apoptosis . We blocked apoptosis by overexpressing the apoptosis inhibitor P35 in the fat body of CrebA-knockdown flies ( Gal80ts; c564-Gal4 > UAS-CrebA-IR , UAS-P35 ) [49] . Expression of P35 in CrebA RNAi flies did not rescue the CrebA survival phenotype upon infection ( S10C Fig ) , indicating that an increase in apoptosis is unlikely to explain the CrebA susceptibility defect . Collectively , our results show that CrebA is required in the fat body to prevent excessive and deleterious levels of ER stress upon infection .
In this study , we have characterized the transcriptomic response of Drosophila to a wide range of bacterial infections . We found that the response to infection can involve up to 2 , 423 genes , or 13 . 7% of the genome . This is a considerably greater number of genes than what has been previously reported in similar transcriptomic studies [6 , 8] . As the response to infection was highly specific to each bacterium , the larger number of genes we identified is likely a consequence of having included more bacterial species in our experiment than previous studies . Likewise , we anticipate that future studies using different species of bacteria could further increase the number of genes found to be involved in the host response to infection . Our data clearly establish that while the core response to infection is narrow and conserved , every bacterium additionally triggers a very specific transcriptional response that reflects its unique interaction with host physiology . At first , this high level of specificity may seem contrary to the traditional vision of the innate immune response . Early studies defined the innate immune system as generic , and the specificity of the Drosophila immune response was considered as a dichotomous activation of the Toll pathway by Gram-positive bacteria ( Lys-type peptidoglycan ) or the Imd pathway by Gram-negative bacteria ( DAP-type peptidoglycan ) [5 , 25] . Our data show that the host response to infection goes beyond the activation of the Toll and Imd pathways , with each bacterium also modulating host cell biology , metabolism , and stress responses in a microbe-specific manner . Although we did find that the type of bacterial peptidoglycan is a key factor shaping the response , we also found that each bacterium activates both the Toll and Imd pathways to quantitatively different levels , consistent with previous reports suggesting a much more complex coordination of the immune response [50–53] . Activation of the Toll and Imd pathways depends on recognition of microbe-associated molecular patterns ( MAMPs ) and detection of damage-associated molecular patterns ( DAMPs ) , suggesting that virulent bacteria could activate the Toll and Imd pathways to a higher degree [13 , 14] . However , we did not find a clear correlation between the virulence level of the bacterium or bacterial load sustained and the degree to which the canonical immune response is activated . In sum , our results support the notion that the response to infection comprises more than simple activation of immune functions , but instead is a function of precise physiological interactions between host and microbe . Although the response to infection appears to be largely specific , we identified a core set of genes that are regulated by infection with most bacteria . Induced genes include the classical targets of the Toll and Imd pathways , such as antimicrobial peptides and immune effectors ( TEPs and IMs ) . However , genes involved in cell and tissue biology ( translation , secretion , cell division ) were also upregulated by the majority of infection conditions , possibly indicating a response to the stress imposed by infection . On the other hand , genes involved in metabolism ( protease activity , oxidation-reduction , glucose metabolism , respiration ) , as well as digestive enzymes ( e . g . the maltase cluster ) , were downregulated , suggesting a complete reshaping of host metabolism during infection [6] . It is tempting to speculate that the majority of core genes that do not fall under the immunity category could be part of a tolerance core response . Although the subject of tolerance mechanisms has attracted a lot of interest in recent years , identifying the genes and processes that define tolerance has remained somewhat elusive [54 , 55] . Further characterization of the core genes identified here may shed light on universal tolerance mechanisms . The idea of a core response to infection has also been explored in other organisms . In Caenorhabditis elegans , for example , a study using four different pathogens to assay the transcriptional response to infection found that the core of the response included genes involved in proteolysis , cell death , and stress responses [56] . Comparative transcriptomics work in the honey bee , Apis mellifera , also revealed a core set of genes utilized in response to distinct pathogens , including genes involved in immunity , stress responses , and tissue repair [57] . In Danio rerio , immunity , metabolism , and cell killing have been implicated in host defense [58] . Collectively , these results and ours indicate that there is considerable overlap in the core response to infection across species , and that this consistency extends beyond classical immune sensing and signaling . Having a well-defined core response to infection in Drosophila will allow future studies to quantitatively assess differences in how distinct pathogens induce the core , as well as test the relative importance of various elements of the core in promoting resistance to and tolerance of infection . A surprisingly high proportion ( ~40% ) of the core response to infection was induced only by live microbes , but was not stimulated by challenge with heat-killed bacteria . One possibility is that MAMPs , such as peptidoglycan , are partially , if not fully , degraded at the sampled time points , obscuring our ability to appreciate the full extent of the response to MAMPs . An alternative explanation is that almost half of the core response to infection is a reaction to microbial activity , rather than just to the presence of MAMPs . This latter model involves the detection of the host’s own DAMPs upon infection [12] . For example , bacterial growth and secretion of toxins can inflict damage to host tissues , leading to the generation of DAMPs , such as actin , proteases , and elastases [13 , 14 , 59] . In turn , DAMPs can activate the Toll , Imd , and JAK-STAT pathways , which may trigger higher levels of signaling in these pathways beyond that which is induced by the detection of MAMPs [13 , 14 , 16 , 59] . Higher degrees of activation in these pathways could then translate into the induction of a larger set of target genes , which could partially account for the ~40% of core genes that are uniquely induced by live infections . Interestingly , our study found that gene expression levels do not always reflect the changes in bacterial load during the course of infection . In chronically infected flies , we found that most genes downregulated at 12 h post-infection had returned to baseline expression levels by 132 h after infection . Likewise , many of the induced genes also decreased in expression or returned to basal levels even while flies still harbored bacteria . It is possible that the injury inflicted to systemically infected flies generates a complex early response , which is resolved at later time points . However , we note that injury alone did not generally trigger the downregulation of genes observed in live infections . An alternative explanation is that the bacteria have entered into a less aggressive state in the late stages of infection , persisting but with a reduced impact on the physiology of their host . Yet another hypothesis is that the host’s initial response to infection is broad-spectrum and disproportionately strong , with the proactive goal of suppressing all bacteria before they can establish a highly detrimental infection . In this scenario , a subdued infection can be controlled with more nuance at later stages [22] . Finally , it is also possible that the percentage of recovered genes following infection with moderately virulent bacteria is overestimated because the RNA-seq is performed on pools of flies that may have distinct individual fates upon infection , and therefore distinct transcriptional kinetics . We have previously shown that flies infected with these same bacteria either die with a high bacterial load or survive with a low-level , persistent infection [38] . The individual flies at the 12 h RNA-seq data point comprise flies destined for both outcomes , but only persistently infected flies are sampled at the 132 h time point after mortality has occurred . If flies fated to die induce genes that are not triggered in flies destined to survive , those genes may appear to be upregulated in the pooled 12 h RNA sample that contains a mix of flies destined for both outcomes . Likewise , those same genes will appear to have returned to baseline levels at the 132 h time point when just chronically infected flies are sampled , creating the false impression that they have recovered . Future work is required to evaluate these hypotheses and to provide insight into how the complex dynamics of gene expression relate to changes in pathogen burden [60] . We also observed seemingly long-term alterations to the transcription of some core response genes , even in the case of infections with bacteria , such as M . luteus and Ecc15 , that are reduced to undetectable levels or cleared by the host . For example , the expression of several antimicrobial peptide genes ( Drs , Dro , and AttB ) as well as other effector molecules ( IM4 and IM3 ) never returned to basal levels , even multiple days after elimination of the infection . Such sustained reactions could provide long-lasting benefits in an environment with high risk of infection . Moreover , it should perhaps be considered that the baseline expression levels of these genes in laboratory-reared Drosophila are artificially low because of aseptic maintenance conditions as compared to those in natural environments . Among our core response genes , we identified CrebA as a key transcription factor that promotes host tolerance to infection . CrebA is the single Drosophila member of the Creb3-like family of transcription factors , which includes five different proteins in mammals: Creb3/Luman , Creb3L1/Oasis , Creb3L2/BBF2H7 , Creb3L3/CrebH , and Creb3L4/Creb4 [31] . A recent study demonstrated that CrebA is a master regulator of secretory capacity , capable of regulating the expression of the general machinery required in all cells for secretion [32] . Drosophila CrebA appears to have the same functional role as its mammalian counterparts . Exogenous expression of mammalian liver-specific CrebH caused upregulation of genes involved in secretory capacity and increased secretion of specific cargos [31] . Moreover , each of the five human CREB3 factors is capable of activating secretory pathway genes in Drosophila , dependent upon their shared ATB ( Adjacent To bZip ) domain [31] . In agreement with the function of CrebA and CREB3 proteins described in the literature , our study finds that CrebA regulates a rapid , infection-induced increase in the expression of secretory pathway genes in the fat body , an organ analogous to the liver and adipose tissues of mammals . Finally , it has been shown that proinflammatory cytokines act to increase the transcription of CrebH , and that CrebH becomes activated in response to ER stress [61] . Our data demonstrate that the two principal immune pathways in Drosophila , the Toll and Imd pathways , upregulate the expression of CrebA in response to bacterial challenge and that loss of CrebA in the fat body triggers ER stress upon infection . Collectively , the functions of mammalian CrebH as a regulator of secretory homeostasis under stress bear a striking resemblance to the role that we have attributed to Drosophila CrebA after bacterial challenge , suggesting that CrebH could have a similar role in mammals during infection . CREB proteins are activated by phosphorylation from diverse kinases , including PKA and Ca2+/calmodulin-dependent protein kinases on the Serine 133 residue [62] . CrebA does not contain a PKA consensus phosphorylation site , and its transcriptional activity is only slightly enhanced by cAMP [36] . Rather , we found that the Toll and Imd pathways are both necessary and sufficient to regulate CrebA expression in the fat body . Loss of CrebA leads to ER stress , further aggravating the physiological strains of infection . However , a lack of CrebA in unchallenged conditions does not lead to the induction of ER stress . We therefore propose a model in which the Toll and Imd pathways act early to upregulate CrebA in order to adapt the fat body cells for infection , thus preventing ER stress that would otherwise be triggered by the response to infection [63] ( Fig 8F ) . This interpretation would suggest that immune activation generates a massive and rapid increase in translation [64] and secretion in response to infection , and thus triggers cellular stress in the fat body . In that context , the Toll and Imd pathways would proactively induce expression of CrebA to prevent some of the stress that comes from their own activation . Lastly , CrebA knockdown flies are more likely to die from infection yet they show no increase in pathogen burden . This demonstrates that CrebA is required for tolerance of infection [65 , 66] . Considering that ER stress is induced upon infection in the absence of CrebA , our data suggest that CrebA is a tolerance gene that helps mitigate the stress imposed by the host response to infection . Fast induction of CrebA by the immune system upon infection can therefore be interpreted as an active tolerance mechanism that is generally required to survive bacterial infection .
Whole fly RNA-seq experiments were performed using wildtype strain Canton S flies . Flies were raised on standard yeast-cornmeal-sucrose medium ( 50 g baker’s yeast , 60 g cornmeal , 40 g sucrose , 7 g agar , 26 . 5 mL Moldex ( 10% ) , and 12 mL Acid Mix solution ( 4 . 2% phosphoric acid , 41 . 8% propionic acid ) per 1L of deionized H2O ) at 24°C and maintained at that temperature for the duration of the experiment . Individual males were infected with one of the ten experimental bacteria 5 to 8 days after eclosion from the pupal case . Control flies that were sterilely wounded or inoculated with heat-killed bacteria were handled equivalently . Flies were pin-pricked to generate septic injury . We standardized the initial inoculation dose across all bacteria to deliver ~3 , 000 colony-forming units ( CFU ) per fly . The following bacteria ( from overnight cultures ) were used: Micrococcus luteus ( A600 = 100 ) , Escherichia coli ( A600 = 100 ) , Serratia marcescens Type ( A600 = 1 ) , Ecc15 ( A600 = 1 ) , Providencia rettgeri ( A600 = 1 ) , Enterococcus faecalis ( A600 = 1 ) , Staphylococcus aureus ( A600 = 1 ) , Providencia sneebia ( A600 = 1 ) , Serratia marcescens Db11 ( A600 = 1 ) , and Pseudomonas entomophila ( A600 = 1 ) . Three sets of controls were included in the experiment: unchallenged and uninjured flies , sterilely wounded flies , and challenge with either heat-killed P . rettgeri or heat-killed E . faecalis . For every control and bacterial infection , with the exception of the 4 highly virulent infections , 20 flies were collected at 12 h , 36 h , and 132 h post-infection . For the 4 highly virulent bacteria , only the 12 h sample was collected because the majority of the flies had died before the later time points . Additionally , 20 unchallenged , uninjured flies were also collected at time 0 h as an extra control . Each sample of 20 flies was homogenized , and total RNA was isolated using a modified TRizol extraction protocol ( Life Technologies ) . All experiments were done in triplicate . The same methodology was employed for the RNA-seq experiment focused specifically on the fat body . Data can be downloaded from NCBI Sequence Read Archive with accession number SRP127794 . Following RNA extraction , the 3’end RNA-seq libraries were prepared using QuantSeq 3’ mRNA-Seq Library Prep kit ( Lexogen ) . The sample quality was evaluated before and after the library preparation using Fragment Analyzer ( Advanced Analytical ) . Libraries were sequenced on two lanes of the Illumina Nextseq 500 platform using standard protocols for 75bp single-end read sequencing at the Cornell Life Sciences Sequencing Core . On average , 6 million reads per sample were sequenced at their 3’ termini . This is roughly equivalent in sensitivity to 20x coverage depth under a conventional random-priming RNA-seq method . Raw reads were first evaluated by fastqc for quality control ( http://www . bioinformatics . bbsrc . ac . uk/projects/fastqc , version 0 . 11 . 3 ) and were then trimmed using Trimmomatic version 0 . 32 [67] . Trimmed reads were mapped to the D . melanogaster reference transcriptome , which was constructed with the D . melanogaster reference genome ( version 6 . 80 ) using STAR RNA-seq aligner version 2 . 4 . 1a [68] . Read depth at each transcript was then calculated using htseq ( version 0 . 6 . 1 ) [69] . Principal Component Analysis and extraction of the PC1/PC2 genes were performed by custom R scripts ( available upon request ) . The software edgeR version 3 . 10 . 5 was used to call the genes that are differentially expressed among treatments [70] . Nine samples of unchallenged flies matching the 3 different time points post-infection ( 12 h , 36 h , and 132 h ) were collapsed into a single control once it was determined that their transcriptomic profiles were very similar . Library sizes were normalized using a trimmed mean of M-values ( TMM ) approach implemented in edgeR . Genes with low counts ( count-per-million < 1 . 2 ) were filtered out prior to differential expression analysis . Genes were considered statistically differentially expressed if they were differentially expressed between unchallenged condition and an infection condition of choice at the 5% false discovery rate ( FDR ) . A fold-change cutoff was not applied to the data . Each gene was also evaluated for the number of infection conditions in which it was differentially regulated , where an “infection condition” refers to the transcriptomic profile in response to any of the live infections or controls at any point post-infection . Heatmaps were generated and clustering was performed using custom R scripts . Gene Ontology and KEGG pathway enrichment analysis was performed using the DAVID bioinformatics resource [71] and PANTHER [72] . The p-values from these analyses were corrected using the Benjamini and Hochberg procedure [73] with the FDR threshold set to 0 . 05 . The search for putative transcription factor binding sites was performed using i-cisTarget under the default parameter values [26 , 27] . For each gene under each infection condition , an expression path was assigned based on the series of inferred induction or repression of infection relative to unchallenged controls at each successive time point . Genes that were significantly induced or repressed at 12 or 36 h but then returned to basal expression levels were deemed to have “recovered” . To quantify the degree of recovery for each gene , the level of fold change at 132 h after infection was compared to the fold change in expression at either 12 h or 36 h using custom R scripts ( available upon request ) . The genes that were significantly induced ( or repressed ) at 12 h and then significantly repressed ( or induced ) at 36 h relative to the unchallenged conditions ( 1% of the genes ) were excluded , as were genes that never changed expression in any of the time points . Subsequent to the initial RNA-seq experiment , genetic manipulations of CrebA expression were performed . Flies for all of these experiments were reared at 18°C or 24°C . The RelE20 and spzrm7 stocks have been previously described [74 , 75] . For manipulation of CrebA expression level , we used the UAS/Gal4 gene expression system in combination with Gal80ts to restrict the expression of the constructs specifically to the adult stage . Male flies were collected 5 to 8 days after eclosion from the pupal case and then shifted to 29°C for an additional 8 days prior to any experiments . We used the following genotypes: 1 . c564-Gal4; tub-Gal80ts , UAS-GFP 2 . Lpp-Gal4; tub-Gal80ts , UAS-GFP 3 . c564-Gal4; tub-Gal80ts , UAS-CrebA-IR 4 . UAS-imd 5 . UAS-spz* 6 . UAS-P35 7 . UAS-Psn ( Bloomington 8305 ) 8 . UAS-BiP-IR ( Bloomington 32402 ) 9 . UAS-BiP ( Bloomington 5843 ) 10 . TRiP control line attP2 ( Bloomington 36303 ) 11 . TRiP control line attP40 ( Bloomington 36304 ) 12 . Xbp1p>dsRed 13–15 . UAS-CrebA-IR ( Bloomington 42562 ( A ) , 31900 ( B ) and 27648 ( C ) ) . Infection was done via septic pinprick to the thorax . After inoculation , death was recorded daily , and flies were transferred to fresh vials every 3 days . All experiments were performed at least 3 times . Statistical significance was determined using a Log-rank ( Mantel-Cox ) test . At specified time points following infection , flies were individually homogenized by bead beating in 500 μl of sterile PBS using a tissue homogenizer ( OPS Diagnostics ) . Dilutions of the homogenate were plated onto LB agar using a WASP II autoplate spiral plater ( Microbiology International ) , incubated overnight at 29°C , and the CFUs were counted . All experiments were performed at least 3 times . Results were analyzed using a two-way ( genotype and time ) ANOVA in Prism ( GraphPad Prism V7 . 0a , GraphPad Software , La Jolla , CA , USA ) . For all experiments utilizing RT-qPCR , total RNA was extracted from pools of 20 flies using a standard TRIzol ( Invitrogen ) extraction . RNA samples were treated with DNase ( Promega ) , and cDNA was generated using murine leukemia virus reverse transcriptase ( MLV-RT ) ( Promega ) . qPCR was performed using the SSO Advanced SYBR green kit ( Bio-Rad ) in a Bio-Rad CFX-Connect instrument . Data represent the relative ratio between the Ct value of the target gene and that of the reference gene RpL32 ( also known as Rp49 ) . Mean values of at least three biological replicates are represented ±SE . Data were normalized and then analyzed using an unpaired t-test in Prism ( GraphPad Prism V7 . 0a; GraphPad Software , La Jolla , CA , USA ) . The primer sequences used in this study are available in S6 Table . In some experiments , fat bodies were visualized microscopically . For these experiments , Drosophila abdomens were dissected and fixed in a 4% paraformaldehyde in 1X PBS solution for 45 minutes and washed 3 times with 0 . 1% Triton-X in PBS . DNA was stained in 1:50 , 000 DAPI ( Sigma-Aldrich ) in PBS and 0 . 1% Triton-X for 45 minutes . Samples were then washed three times in PBS and mounted in antifadent medium ( Citifluor AF1 ) . Imaging was performed on a Zeiss LSM 700 fluorescent/confocal inverted microscope .
|
How does an organism survive infection ? How generic or specific is the host response to diverse pathogens ? To address these questions , we infected fruit flies with 10 different bacteria that vary in their ability to kill flies and measured changes in global gene expression . In general , we found that the host response is highly specific to individual bacteria . However , we also discovered a set of genes that changed expression in response to the majority of bacteria tested . Among these genes , we determined that the transcription factor CrebA is a novel regulator of the host response to infection . We found that upon infection , the immune system induces the expression of CrebA . CrebA-deficient flies are more likely to die from infection despite carrying the same number of bacteria as wildtype flies . CrebA is expressed in the fat body , an organ analogous to the mammalian liver and adipose tissues , where it regulates the transcription of multiple secretory pathway genes . Loss of CrebA during infection triggers endoplasmic reticulum ( ER ) stress ( a type of cellular stress ) , which is sufficient to sensitize flies to infection . These results suggest that the immune system can modulate host physiology to prevent the deleterious effect of infection-associated cellular stress .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"rna",
"interference",
"gene",
"regulation",
"regulatory",
"proteins",
"immunology",
"dna-binding",
"proteins",
"animals",
"animal",
"models",
"bacterial",
"diseases",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"transcription",
"factors",
"epigenetics",
"drosophila",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"lipids",
"genetic",
"interference",
"proteins",
"fats",
"gene",
"expression",
"insects",
"immune",
"response",
"arthropoda",
"biochemistry",
"rna",
"eukaryota",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2018
|
Comparative transcriptomics reveals CrebA as a novel regulator of infection tolerance in D. melanogaster
|
The interaction with brain endothelial cells is central to the pathogenicity of Neisseria meningitidis infections . Here , we show that N . meningitidis causes transient activation of acid sphingomyelinase ( ASM ) followed by ceramide release in brain endothelial cells . In response to N . meningitidis infection , ASM and ceramide are displayed at the outer leaflet of the cell membrane and condense into large membrane platforms which also concentrate the ErbB2 receptor . The outer membrane protein Opc and phosphatidylcholine-specific phospholipase C that is activated upon binding of the pathogen to heparan sulfate proteoglycans , are required for N . meningitidis-mediated ASM activation . Pharmacologic or genetic ablation of ASM abrogated meningococcal internalization without affecting bacterial adherence . In accordance , the restricted invasiveness of a defined set of pathogenic isolates of the ST-11/ST-8 clonal complex into brain endothelial cells directly correlated with their restricted ability to induce ASM and ceramide release . In conclusion , ASM activation and ceramide release are essential for internalization of Opc-expressing meningococci into brain endothelial cells , and this segregates with invasiveness of N . meningitidis strains .
Neisseria meningitidis ( Nm , the meningococcus ) is a frequently found asymptomatic colonizer of the upper respiratory tract , which under certain circumstances may penetrate the mucosal membrane , reach the bloodstream and cause septicemia and/or meningitis . During the course of infection N . meningitidis is capable to interact with a variety of human cells including epithelial as well as peripheral and brain microvascular endothelial cells [1] , [2] . To mediate association with this wide range of host cells , meningococcci express a variety of adhesins and invasins , including type IV pili ( TfP ) [3]–[5] , the outer membrane proteins Opa and Opc and a number of newly identified minor adhesion or adhesion-like proteins [6]–[12] . As an important pathogenicity factor the integral outer membrane protein ( OMP ) Opc is particularly implicated in host cell invasion of endothelial cells [1] , [8] , [13] , [14] . Opc is a beta barrel protein with five surface loops encoded by a single gene ( opcA ) and is antigenically stable [15] , [16] . Opc expression is controlled at the transcriptional level by the length of a polycytidine stretch within its promoter region [17] . Opc is expressed by several virulent N . meningitidis lineages , but is absent from certain epidemic clones ( ET-37/ST-11 clonal complex ( cc ) ) and a few random endemic isolates [18] . Two epidemiological studies reported outbreaks where meningococcal strains of the ST-11 cc tend to cause severe sepsis with fatal outcome , but rarely meningitis [19] , [20] . For N . meningitidis uptake , Opc links the meningococcus to the extracellular matrix components and serum proteins vitronectin and fibronectin followed by binding to αvβ3 or α5β1-integrins and activation of phosphotyrosine signalling and cytoskeletal rearrangement [1] , [2] , [21]–[23] . As observed for human epithelial cells , Opc can also bind to heparin-like molecules and to cell surface heparan sulfate proteoglycans ( HSPGs ) [24] , which can mediate receptor interaction ( referred to as cis-activation ) to induce signaling and subsequent internalization [25] . Adhesion of fully encapsulated meningococci to host cells is facilitated primarily by pili . Though OMPs are partially masked by the polysaccharide capsule , they also efficiently support adhesion and invasion to eukaryotic cells especially on cells of high receptor density as would be induced in inflammatory conditions and/or lateral receptor aggregation [26] . Sphingomyelin is a major component of the outer plasma membrane layer where , under certain conditions of stress or the induction of inflammatory cytokines , it is metabolized into ceramide and phosphorylcholine by the activity of the acid sphingomyelinase ( ASM ) . Once activated , this enzyme translocates from its lysosomal intracellular storage to outer cell membrane where it is co-displayed with released ceramides [27]–[35] . Due to their biophysical properties , ceramide-enriched membrane domains fuse into extended ceramide-enriched platforms which span a few hundred nanometers to several micrometers [27] , [29] . In addition to altering membrane fluidity and rigidity , ceramide-enriched platforms serve to sort and eventually concentrate membrane receptors and membrane proximal signaling components thereby amplifying cellular responses and signal transduction [36] . Their role in enhancement of pathogen uptake has been directly revealed for N . gonorrhoeae , Pseudomonas aeruginosa , as well as rhinoviruses and measles virus [37]–[41] . In addition , Staphylococcus aureus triggers the formation of ceramide-enriched membrane platforms for induction of apoptosis [42] . It is as yet unknown whether SMase activation and ceramide release relates to N . meningitidis uptake especially in its natural target cells . In this study we now show that N . meningitidis induces ASM activation , ceramide release and formation of ceramide-enriched platforms proximal to attached bacteria within the outer layer of the membrane of brain endothelial cells . Ceramide-enriched platforms in turn serve to cluster the ErbB2 receptor underneath adherent bacteria . Opc and activation of phosphatidylcholine-specific phospholipase C ( PC-PLC ) downstream of HSPGs is critical for N . meningitidis ASM activation , which proved to be crucial for N . meningitidis uptake but not adhesion . Stressing the importance of ASM activation in N . meningitidis invasion and pathogenesis , a less invasive defined set of pathogenic isolates of the ST-11/ST-8 cc was substantially less capable of inducing ASM activation and formation of ceramide-enriched platforms .
Because uptake of some pathogenic bacteria involved formation of ceramide-enriched membrane platforms [37]–[39] , we investigated whether N . meningitidis employs a similar mechanism to infect and enter into eukaryotic cells . To analyse whether N . meningitidis stimulates surface display of ceramide on human brain microvascular endothelial cells ( HBMEC ) , cells were infected with the GFP-expressing wildtype strain MC58 ( ST-32 clonal complex ( cc ) ) , fixed and stained with an anti-ceramide antibody ( mAb 15B4 ) . N . meningitidis strain MC58 rapidly , but transiently induced formation of large extrafacial ceramide-enriched platforms , which reached a maximum within 2 hrs after infection ( Fig . 1 , Fig . S1 ) and decreased thereafter . Bacteria adhered to the cells within ceramide-enriched membrane platforms ( Fig . 1A , upper panels ) . In unexposed control cells , shallow ceramide-specific signals were visible , but these were not condensed into large platforms ( Fig . 1A , lower panels ) . Because surface accumulation of ceramides usually reflects acid sphingomyelinase ( ASM ) rather than neutral sphingomyelinase ( NSM ) activity , which acts on its substrate in the cytosolic layer , we directly assessed activation of these enzymes in response to N . meningitidis MC58 exposure directly in cell lysates or membrane preparations ( for NSM ) , respectively . Mirroring the kinetics of extrafacial ceramide release ( Fig . 1 ) , N . meningitidis MC58 caused a rapid ASM activation , which peaked within 2 hrs and subsequently declined ( Fig . 2 A ) . In contrast , NSM was not stimulated by N . meningitidis within this time frame ( Fig . S2 ) . The kinetics of ASM and ceramide surface display were monitored by flow cytometry ( Fig . 2B–D ) or by immunodetection of a ceramide-specific antibody bound to intact cells ( Fig . 2D , inset ) at various time intervals following N . meningitidis MC58 exposure . Irrespective of the method used , ASM and ceramide were readily detectable on the cell surface within 2 hrs and their expression declined thereafter . Pre-treatment of cells with the functional ASM inhibitor amitriptyline prevented surface release of ceramides after N . meningitidis MC58 exposure indicating that this process depends on ASM ( Fig . 2E ) . Finally , while ASM was not and ceramide was barely detectable on the surface of uninfected HBMEC cells , both molecules were readily co-detected 2 hrs following exposure of the cells to MC58 ( Fig . 2F ) . N . meningitidis MC58–induced ASM surface location and ceramide release were not restricted to HBMEC as revealed by experiments involving HBMEC/ciβ , which slightly differed with regard to the kinetics of peak levels ( 3 hrs rather than 2 hrs ) , but not efficiency of ceramide release ( Fig . S3 ) . Previous studies implicated a regulation of ASM by diacylglycerol ( DAG ) release in response to phosphatidylcholine-specific phospholipase C ( PC-PLC ) activity [43] , [44] . This also applied to the related species , N . gonorrhoeae , where PC consumption by PC-PLC activity released DAG resulting in ASM activation [37] . PC-PLC activity was measured in lysates of HBMEC infected with N . meningitidis MC58 over time in the presence or absence of the PC-PLC inhibitor D609 . N . meningitidis MC58 caused a significant PC-PLC activation , which peaked within 2 hrs , and this was inhibited by preincubation of HBMEC with D609 ( Fig . S4 ) . To rule out a possible role of PC-PLC also in N . meningitidis driven ASM activation , ASM activity was assessed in lysates of HBMEC that were treated with the PC-PLC inhibitor D609 prior to infection . D609 indeed blocked the activation of ASM during meningococcal infection ( Fig . 3A ) . Moreover , cell surface ceramide release was significantly reduced in the presence of D609 as demonstrated by flow cytometry 2 hrs following infection with N . meningitidis MC58 ( Fig . 3B ) . To address a potential contribution of phospholipase A2 ( PLA2 ) or phospholipase D ( PLD ) in DAG generation for ASM activation , activity of the enzyme was assessed in lysates of HBMEC treated with PLA2 inhibitor AACOCF3 or PLD inhibitor 5-fluoro-2-indolyl des-chlorohalopemide ( FIPI ) , respectively , prior to infection . Indicating that these enzymes are not involved in meningococcal ASM activation , these compounds did not affect ASM activity ( Fig . 3C ) nor cell surface ceramide display ( Fig . 3D ) . For N . gonorrhoeae , PC consumption by PC-PLC is initiated by binding of Opa-expressing strains to heparan sulphate proteoglycans ( HSPGs ) [37] . Opc is also able to mediate adhesion to and invasion of epithelial cells by binding to HSPGs [24] . To analyse whether PC-PLC and ASM are activated downstream of HSPGs , two different experimental approaches were taken: ASM activity was assessed in HBMEC infected in the presence of heparin that blocks possible Opc-HSPG interactions , or cells were treated with heparinase III , which cleaves heparan sulphate moieties , prior to infection . Both conditions significantly reduced ASM activation by N . meningitidis ( Fig . 3E ) , along with meningococcal uptake into HBMEC ( Fig . 3F and G ) , indicating an important role of this enzyme in this process . To analyze whether ASM activation was of functional importance in meningococcal uptake directly , HBMEC were treated with the ASM inhibitor amitriptyline 30 min prior to infection with N . meningitidis MC58 and an isogenic unencapsulated mutant strain MC58 siaD , and adhesion and invasion were determined over time . The unencapsulated mutant was included into these experiments due to its higher invasive capacity [1] . Interestingly , ASM activation induced by MC58 siaD after 2 h significantly exceeded that induced by wildtype MC58 ( Fig . 4A ) . As determined in gentamicin protection assays , pharmacological ASM inhibition reduced intracellular accumulation of N . meningitidis MC58 ( Fig . 4B , right panel ) and MC58 siaD ( Fig . 4C ) ( about 70% inhibition with 10 µM amitriptyline at 4 h p . i . [P<0 . 05] and about 80% inhibition at 8 h p . i . [P<0 . 05] , dose-dependent data shown for MC58 siaD only ) ( Fig . 4C ) . ASM inhibition did , however , not affect adhesion of N . meningitidis to HBMEC as determined by the estimation of cell-adherent bacteria ( Fig . 4B , left panel , for MC58 , Fig . 4C for MC58 siaD ) or bacterial growth ( data not shown ) . We took two independent genetic approaches to verify the importance of the ASM for N . meningitidis invasion . Firstly , RNAi-mediated knockdown of ASM transcripts reduced N . meningitidis uptake by HBMEC by about 40% for MC58 and >90% for MC58 siaD , while pathogen uptake was not affected by a scrambled control siRNA ( Fig . 4D and E ) . Secondly , we comparatively analyzed N . meningitidis uptake into fibroblasts derived from ASM−/− or wild-type mouse embryos ( MEFs ) . Wild-type MEFs internalised N . meningitidis with similar kinetics , but less efficiently ( one log lower ) than HBMEC , however , while ASM-deficient MEFs were significantly impaired in the ability to internalize N . meningitidis ( Fig . 4F and G; later time points could not be analyzed due to toxicity of the infection in these cultures ) . Finally , we comparatively analyzed N . meningitidis MC58 and MC58 siaD uptake into human fibroblasts generated from healthy donors or patients suffering from Niemann-Pick disease type A ( NPDA ) , a lysosomal storage disease , characterized by a lack of ASM activity . In line with the data obtained for ASM siRNA ablation and in ASM−/− MEFs ( Fig . 4D–G ) , N . meningitidis uptake was severely reduced in NPDA-fibroblasts as compared to their wild-type counterparts ( Fig . 5A and B ) . When analyzed by immunofluorescent staining , a significant proportion of cell-associated meningococci localized within fibroblasts of healthy controls , while the frequency of intracellular bacteria was substantially lower in NPDA cells 4 h p . i . ( ( Fig . 5C and Fig . S5 ) , data shown for MC58 siaD ) though equivalent amounts of cell-attached bacteria were initially seen in all cultures . Altogether , these findings clearly reveal that ASM activation by meningococci is critical for bacterial uptake , but not adhesion . To analyse whether N . meningitidis isolates belonging to different serogroups and/or sequence types might differentially activate ASM and ceramide release , we selected four pathogenic meningococcal isolates of the ST-11/ST-8 cc that belong to serogroup C ( MenC strains WUE2121 , DE7017 , FAM18 , DE6904 ) ( Table 1 ) . In gentamicin protection assays , these isolates proved to be significantly less invasive into HBMEC as compared to strain N . meningitidis MC58 , however , did not differ with regard to adhesion from MC58 ( Fig . 6A and B ) . However , these 4 ST-11 cc isolates proved to be significantly less effective at inducing ceramides in HBMEC than the invasive MC58 strain as determined by flow cytometry ( Fig . 6C ) . Indicating that invasiveness of Neisseria strains links to their ability to cause ASM activation , the activity of the ASM in HBMEC extracts prepared 2 h p . i . with four the MenC strains ( WUE2121 , DE7017 , FAM18 , DE6904 ) was significantly lower than that induced by MC58 ( Fig . 6E ) . These strains , unlike MC58 ( Fig . 1 ) , also failed to induce ceramide-enriched membrane platforms ( data shown for WUE2121 ) ( Fig . 6D ) . We extended our study and included two further isolates that belong to serogroup B ( strains DE7901 and the carrier isolate α4 ) ( Table 1 ) . Carrier isolate α4 was found to be less adherent to HBMEC than the other strains tested , however , in line with the findings observed for the serogroup C strains , both isolates were found to be significantly less invasive and effective at inducing ceramides on HBMEC ( Fig . 6B and C ) . To determine whether ASM activation and ceramide release is specific to brain endothelial cells , we extended our study to epithelial cells and included the human epithelial cell line FaDu . As shown in figure S6 , neither adherence nor invasion significantly varied among strains tested at 4 h p . i . Cermide surface levels on FaDu cells generally exceeded those detected on HBMEC , however , did not detectably increase after infection with strain MC58 or four MenC strains as determined by flow cytometry at various time points ( Fig . S6C , D ) . In order to define the meningococcal factor ( s ) involved in differential ASM activation , we focussed on genes absent from the ST-11 cc strains as compared to N . meningitidis MC58 . Based on data generated by a previous microarray comparative genome hybridization ( mCGH ) study [45] , 8 genes encoding for surface and virulence-associated proteins were identified as potentially involved in ASM activation also including the Opc outer membrane protein ( Fig . 7 ) . This study revealed that the opc gene is lacking in the three ST-11 cc strains ( WUE2121 , DE7017 , FAM18 ) , while a hybridization signal was observed for ST-8 cc strain DE6904 in the previous mCGH study , indicating the presence of opc in strain DE6904 . A detailed PCR amplification of the opc gene , however , revealed a deletion of the opc gene also in this strain resulting in lack of Opc expression ( Fig . S7 ) . To define the role of Opc for ASM activation in detail , we made use of two isogenic Opc-deficient knock out mutants MC58 opc and MC58 siaD , opc [1] . HBMEC were infected with MC58 and isogenic mutants and surface display of ceramide and ASM and total ASM activity were determined . Isogenic Opc-deficient mutants were less efficient at inducing ASM activation ( Fig . 8A ) and surface display of ASM and ceramide ( Fig . 8B and C ) compared to wildtype MC58 or isogenic unencapsulated strain MC58 siaD . These data indicate that the absence of Opc accounts for differential ASM activation . To investigate whether the bacterial factor Opc is sufficient in this process , the opc gene was cloned under the control of an IPTG inducible prokaryotic expression vector and overexpressed in E . coli BL21 . Recombinant protein was expressed at high levels in E . coli BL21 as demonstrated by immunoblot and flow cytometry analysis ( Fig . 8D ) . We next assessed the properties of the recombinant Opc expressing E . coli to adhere to and enter into HBMEC . Therefore , HBMEC were infected for 2 h with E . coli BL21 expressing Opc or the parental E . coli strain at an MOI of 30 . As expected , infection of HBMEC with Opc-expressing bacteria significantly increased internalization ( 10-fold increase of uptake ) , but not adherence ( Fig . 8D ) . HBMEC infection with Opc-expressing E . coli resulted in a 1 . 3 fold increase of ceramide release compared to the parental E . coli strain ( Fig . 8E ) . Moreover , ASM activity in response to Opc-expressing E . coli increased significantly measured at 2 h p . i ( Fig . 8F ) . Together , these data attribute an important role for the Opc protein in enhancement of bacterial uptake driven by ASM activation . There are numerous examples of cellular surface receptors including CD95 [27] , CD40 [31] or CD150 [41] , which cluster within ceramide-enriched platforms generated in response to ASM activation . We therefore hypothezised that enhancement of N . meningitidis uptake into HBMEC by Opc and HSPG driven ASM activation might relate to concentration of receptors involved in meningococcal invasion within these platforms . N . meningitidis recruit the tyrosine kinase receptor ErbB2 ( Her2/neu ) to the sites of bacterial adherence on endothelial cells [46] . Tyrosine phosphorylation of ErbB2 has been shown to be required for efficient uptake by promoting the phosphorylation of the actin-binding protein cortactin [47] . To examine whether this receptor is recruited and trapped into ceramide-enriched membrane platforms , we compared the subcellular distribution of ErbB2 in HBMEC before and after infection with N . meningitidis . In uninfected cells , ErbB2 detected in fixed , unpermeablized cells revealed an overall punctuate expression pattern ( Fig . 9 , second row ) . Infection with N . meningitidis strain MC58 for 2 h , however , caused enhanced re-distribution of ErbB2 to the site where bacteria had attached and formed microcolonies ( Fig . 9 , first row ) , where we also detected ceramide-enriched platforms . Formation of ceramide-enriched platforms upon meningococcal exposure was sensitive to ASM knockdown and remarkably , ErbB2 failed to redistribute and cluster under these conditions as well ( Fig . 9 , third row ) . To confirm involvement of HSPGs , HBMEC were pretreated with either heparinase ( 150 mU ) or heparin ( 100 µg/ml ) and infected with strain MC58 for 2 h . As shown in figure 9 ( rows 5 and 6 ) formation of ceramide-enriched platforms upon meningococcal exposure was also sensitive to interference with HSPGs and ErbB2 failed to redistribute under these conditions as well . Altogether , these data show that the reorganization of ceramide into large membrane platforms promotes lateral redistribution and clustering of ErbB2 , an important receptor involved in N . meningitidis uptake .
In this study we provide evidence for an important function of the acid sphingomyelinase ( ASM ) in Opc-mediated internalization of N . meningitidis into brain endothelial cells . Furthermore , we show that activation of ASM occurs downstream of heparan sulfate proteoglycans ( HSPGs ) and involves phosphatidylcholin phospholipase C ( PC-PLC ) activity . The importance of ASM activation in N . meningitidis uptake was strongly supported by its significant reduction upon experimental ( pharmacological and siRNA mediated interference ) or natural ( NPDA fibroblasts ) ablation of the enzyme activity . ASM activation by Opc-expressing N . meningitidis caused ceramide release resulting in formation of extended ceramide-enriched membrane platforms , where ASM and , most importantly , the tyrosine kinase receptor ErbB2 , involved in the cellular uptake of N . meningitidis , are co-displayed . It is in response to a variety of stimuli also including ligation of certain receptors that lysosomal ASM translocates to the outer leaflet of the cell membrane to catalyze breakdown of sphingomyelin into ceramide . As relevant to our studies , the presence of fatty acids , mono- , di- , and triacylglycerols as well as phosphatidylinositol has been implicated in ASM activation [48] , and this especially applies to diacylglycerol ( DAG ) released in response to by PC-PLC , activated by the related species N . gonorrhoeae [37] , [43] , [44] . Here , we show that interaction of N . meningitidis with HSPGs stimulates ASM translocation and activation , and this process also involves PC-PLC activity ( Fig . 3 ) . HSPGs function primarily as initial , low affinity co-receptors that act to concentrate pathogens on host cell surfaces , thereby increasing binding to specific secondary receptors . For N . gonorrhoeae , ASM activation has been found to be triggered upon Opa-mediated binding to HSPGs [37] , whereas ASM activation by N . gonorrhoeae in neutrophils involves Opa52-mediated binding to CEACAM receptors [38] . CEACAM-mediated activation of ASM by N . meningitidis could be excluded in our cell culture model system because this receptor is not expressed on HBMEC ( unpublished data ) . For N . meningitidis uptake , it has been established that Opc links the bacterium to vitronectin and/or fibronectin followed by binding to αvβ3 or α5β1-integrins [1] , [2] . It is quite possible that the interaction of N . meningitidis with integrins stimulates per se or contributes to ASM translocation followed by ceramide release . This can , however , only be addressed by a separate , thorough study which would exceed the scope of the present manuscript . During the interaction with endothelial cells , pilus mediated adhesion of N . meningitidis has been shown to induce the clustering and tyrosine phosphorylation of the host tyrosine kinase receptor ErbB2 [46] . Activation of ErbB2 is required to support N . meningitidis internalization [46] , by promoting the phosphorylation of the actin-binding protein cortactin and thus it is remarkable that it co-segregates with ASM within ceramide-enriched platforms generated downstream of meningococcal interaction with HSPGs ( Figs . 3 , 9 ) . This suggests that ASM and ceramide-induced clustering and lateral segregation of ErbB2 in membrane platforms function as an upstream prerequisite for ErbB2-supported meningococcal internalization . For N . gonorrhoeae , ASM activation is required for both the entry into professional and non-professional phagocytes [37] , [38] , however , the formation of large ceramide-enriched membrane platforms and clustering of receptors involved in N . gonorrhoeae uptake has not been demonstrated so far . Our findings are paralleled by the observation that measles virus ( MV ) interaction with a pattern recognition receptor ( DC-SIGN ) is followed by SMase catalyzed ceramide release and membrane display of its entry receptor CD150 within ceramide-enriched platforms on dendritic cells [41] . Thus , the generation of ceramide-enriched membrane platforms might be a general motif to mediate uptake of pathogens that applies to a very diverse range of pathogens . Whereas for MV one receptor is sufficient to promote viral uptake , it is likely that further receptors might be recruited into ceramide-enriched membrane platforms induced upon N . meningitidis infection of endothelial cells , since adhesion of N . meningitidis to epithelial cells has been shown to induce the formation of a specific molecular complex , referred to as ‘cortical plaque’ . ‘Cortical plaques’ are enriched in ezrin , moesin , tyrosine-phosphorylated proteins , ICAM-1 , CD44 and EGFR [49] , [50] . In addition , meningococcal adhesion on human endothelial cells leads to the redistribution of the Par3/Par6 polarity complex , which is recruited underneath meningococcal microcolonies , weakening the barrier properties of the blood-cerebrospinal fluid barrier [46] , [51] . It is tempting to speculate that molecules of the ‘cortical plaque’ or even polarity complex components cluster in ceramide-enriched membrane platforms underneath meningococcal microcolonies . It is important to note that ASM-inhibition prevented meningococcal uptake , but did not alter adherence regardless of whether the enzyme was inhibited by pharmacological ( amitriptyline ) or genetic approaches ( siRNA ablation , ASM−/− MEFs and NPDA fibroblasts ) ( Fig . 5 ) . Moreover , there is apparently a peak of ASM translocation at 2 hrs as measured by flow cytometry , which does , however , not resolute subcellular localization of ceramide release . This is much better reflected by detection of ceramide-enriched platforms by immunofluorescence , which , by nature , is not quantitative . Neither ASM alone nor ceramides per se are operative in supporting N . meningitidis invasion , but rather act to sort surface and membrane proximal complexes required which may not be completed , but ongoing over the time interval monitored . It is also possible that ceramide release assists in alterations of membrane curvature as required for invasion which may also be time consuming . It has , for instance been shown in an ASM dependent bead uptake assay that increase in membrane order ( reflecting ASM activity ) peaked already after 5 mins , while bead uptake in macrophages was completed only after 1 hr [52] . Previous studies have shown that P . aeruginosa and Staphylococcus aureus activate ASM and trigger the formation of ceramide-enriched membrane platforms [39] . However , the bacterial factors responsible for ASM activation have not been identified for these microorganisms so far . In our system , ASM activation and ceramide release were assigned to Opc , and the differential ability of meningococcal clonal complexes to promote ASM activation could be assigned to expression of Opc ( Fig . 6 and 8 ) . Studying the interaction of N . meningitidis with host cells is complicated by the high variability of meningococcal isolates which can be classified by multi locus sequence typing ( MLST ) based on the polymorphisms in seven housekeeping genes [53] . Moreover , meningococcal isolates can be clustered to clonal complexes comprising isolates that share identical nucleotide sequences for at least four MLST loci . Isolates from asymptomatic carriers are more diverse that those recovered from patients with invasive disease [54] , [55] . A few so called ‘hyper-invasive’ isolates are responsible for most reported diseases and may spread rapidly through human populations , resulting in countrywide epidemics of meningococcal meningitis [54]–[57] . Two epidemiological studies reported outbreaks of meningococcal strains of the ST-11 cc that tended to cause severe sepsis with fatal outcome but rarely caused meningitis [19] , [20] . Employing four MenC isolates belonging to ST-11/ST-8 cc that lack the opc gene , we demonstrated that all four MenC strains induced significant less activation and translocation of ASM in endothelial cells , followed by significant less accumulation of ceramide on the cell surface and , therefore , failed to trigger the formation of ceramide-enriched platforms ( Fig . 6 ) . In line with a proposed critical role for ASM for meningococcal uptake , all ST-11 cc strains tested in this study proved to be significantly less invasive into HBMEC compared to strain N . meningitidis MC58 . This indicated a specific stimulation of ASM by the Opc-mediated interaction of N . meningitidis with endothelial cells . The contribution of Opc was underlined by using recombinant E . coli expressing a functional form of this protein , demonstrating that Opc is sufficient to trigger ASM activation followed by accumulation of ceramide within the outer leaflet of the plasma membrane ( Fig . 8 ) . However , since the differences between Opc expressing and non-expressing strains are about 30% , further candidates might also contribute to ASM activation and thus ceramide generation , such as the PorB , NarE or the VapD-like protein , that are also absent in all ST-11/ST-8 cc strains ( see Fig . 7 ) . The critical role of the ASM/ceramide system in uptake of certain bacterial pathogens suggests this system as potential therapeutical target . Amitriptyline used to inhibit ASM in this and several other studies is in clincal use for treatment of depression , and currently tested in a randomized , double-blind , placebo-controlled phase IIb multicenter study for the therapy of patients with cystic fibrosis ( CF ) , where it was found to improve the lung function and to reduce ceramide in the lung cells of CF patients [58] . In contrast to CF , an inherited chronic disease that affects the lungs and digestive system , meningococcal meningitis develops quickly and outcome often depends on how soon antibiotics are given after the illness starts . Approximately 50–60% of patients with systemic meningococcal infection present with clinical symptoms of distinct meningitis . A potentially neuro-protective therapeutic agent blocking the ASM would be administered for a short period , should be well tolerated and , most important , must immediately inhibit the ASM in endothelial cells . Amitriptyline and structurally similar molecules functionally inhibit the ASM by inducing a degradation of the enzyme and , therefore , their action might be too slow . Thus , a drug targeting the ASM in N . meningitidis infection should inhibit directly and immediately the ASM to provide a beneficial effect in patients suffering from meningococcal meningitis .
Neisseria meningitidis strain MC58 , a serogroup B isolate ( United Kingdom , 1983 ) of the ST-32 clonal complex ( cc ) was characterized as serotype B∶15∶P1 . 7 , 16 ( kindly provided by E . R . Moxon ) . Non-encapsulated mutant MC58 siaD , Opc-deficient mutant strains MC58 opc and MC58 siaD , opc were previously described [1] ( Table 2 ) . Meningococcal isolates of the ST-11/ST-8 cc of serogroup C ( N . meningitidis strains WUE2121 , DE7017 , FAM18 , DE6904 ) and isolates of serogroup B ( α4 and DE7901 ) were taken from a collection by Schoen and colleagues [45] and are summarized in table 1 . All strains were tested for Opc , NadA and NarE expression ( Fig . S7 ) . Polyclonal antibody raised against NarE was kindly provided by Dr . F . Günther ( Medical Microbiology and Hygiene , University of Heidelberg ) . Opc expression was analyzed by PCR using primer pair RA3 5′- CATCTCAAGTCTCGTCATTCC-3′ and RA4: 5′-AGCCTGTGTAAAGATCGATAC-3′ , kindly provided by Dr . H . Claus . Lack of opc gene was verified by PCR using primer pair RA1 5′- CAAAGCGCACATCACCGTC-3′ and RA2 5′-CCATCAAATGAATATCCATACC-3′ . Confocal microscopy was performed with a derivative expressing a green fluorescent protein from the plasmid pEG2-Ery [59] , which also contains an erythromycin-resistance gene . The simian virus 40 large T antigen-transformed human brain microvascular endothelial cells ( HBMEC ) were kindly provided by K . S . Kim [60] and were cultured as previously described [1] . Cells between the 10th and 25th passages were used for infection assays at a density of 1×105 per well with bacteria at a multiplicity of infection ( MOI ) of 10–30 unless indicated otherwise as described previously [1] . Infections were carried out in the presence of 10% human serum ( HS ) supplemented RPMI medium . HS were derived from a serum pool ( voluntary staff ) and heat-inactivated for 30 min at 56°C [1] . Wildtype strain N . meningitidis MC58 and the isogenic capsule deficient mutant were tested repeatedly for pili , Opa and Opc expression before application to infection assays and after reisolation from the cell culture using Western blot analysis . HBMEC/ciβ was kindly provided by Tomomi Furihata [61] . HBMEC/ciβ were generated from a human primary BMEC-derived cell line immortalized with both the temperature-sensitive mutant of the simian virus 40 large tumor antigen ( tsSV40T ) and the human telomerase reverse transcriptase subunit ( hTERT ) [61] . Human ASM-deficient fibroblasts were obtained from patients with Niemann-Pick Disease type A ( NPDA ) and maintained in RPMI medium 1640 . Mouse embryonic fibroblasts ( MEFs ) were derived from ASM knockout mice or wildtype embryos and grown in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% fetal calf serum ( FCS ) . FaDu ( ATCC HTB 43 ) , an epithelial-like human pharynx cell line from cell carcinoma , was maintained in DMEM supplemented with 10% FCS . When indicated , cells were exposed to amitriptyline ( 5 µM , 7 . 5 µM and 10 µM ) , Heparinase III ( from Flavobacterium heparinum [H8891]; 75 mU and 150 mU ) and heparin ( heparin sodium salt; 10 µg/ml , 50 µg/ml and 100 µg/ml ) ( all: Sigma-Aldrich ( Sigma-Aldrich , Taufkirchen , Germany ) ) , phosphatidyl choline-specific phospholipase C ( PC-PLC ) inhibitor [D609] ( 100 µM ) or phospholipase A2 inhibitor AACOCF3 ( 0 . 1 µM , 5 µM and 10 µM ) or phospholipase A2 inhibitor FIPI ( 20 nM , 100 nM and 750 nM ) ( all: Tocris Bioscience , Bristol , UK ) . 6-Hexadecanoylamino-4-methylumbelliferyl-phosphorylcholine ( HMU-PC ) was purchased from Moscerdam Substrates ( Amsterdam , The Netherlands ) . D609 , AACOCF3 , FIPI and amitriptyline did not affect the viability of bacteria or HBMEC judged by survival assays and microscopic or flow cytometry examination . For flow cytometry , fluorescence microscopy and Western blotting analyses the following primary antibodies were used: mouse monoclonal antibody ( mAb ) anti-ASM IgG ( clone ab74281 , Abcam , Cambridge , UK ) , mouse mAb anti-ceramide IgM ( clone MID 15B4 , Alexis ) , polyclonal rabbit anti-ASM IgG ( H-181 , Santa Cruz ) , mouse mAb anti-ErbB2/Her2/neu IgG2a ( clone Neu ( C3 ) , Santa Cruz ) . Secondary Cy3-conjugated goat anti-mouse IgM , goat anti-rabbit IgG ( H+L ) Cy3 and Cy5 conjugates , Cy5-conjugated goat anti-mouse IgM , Fluorescein isothiocyanate ( FITC ) -conjugated goat anti-rabbit IgG ( H+L ) and Tetramethylrhodamine isothiocyanate ( TRITC ) -conjugated goat anti-rabbit IgG ( H+L ) ( all: Dianova ( Dianova , Hamburg , Germany ) ) . Alexa350-conjugated goat anti-mouse IgM ( μ-chain specific ) was from Invitrogen . Nitrocellulose membranes were incubated with horseradish peroxidase ( HRP ) -conjugated secondary goat anti-mouse IgG ( bacteria ) ( Sigma-Aldrich , St . Louis , MO ) or HRP-conjugated secondary goat-anti-mouse IgM ( spot assay ) ( Dianova ) . In order to generate a Opc-expressing Escherichia coli strain , the opc gene was amplified by PCR from chromosomal DNA of N . meningitidis strain MC58 using the primer pair OpcA-MC58-sense ( 5′-CGGATCCATGGGCAAAAAAACAGTTTTT -3′ , NcoI site ) and OpcA-MC58-antisense ( 5′- CCCGCTCGAGTCAGAATTTTATGCCGACGCG -3′ , XhoI site ) . The PCR product was digested with NcoI/XhoI and cloned in pET28a ( + ) ( Novagen ) . The pET28a ( + ) vector encoding Opc protein were transformed in E . coli BL21 ( DE3; from Novagen ) . Cloned opc gene was verified by sequence analysis . Opc expression was controlled by immunoblot analysis using monoclonal anti-Opc antibodies ( clone B306 , kindly provided by M . Achtman ) . For Western blot experiments , overnight cultures of meningococci were inoculated into Polyvitex ( 1% ( v/v ) ) supplemented Proteose Peptone medium ( PPM+ ) and grown to mid-log phase ( OD600 = 0 . 5–0 . 6 ) at 200 r . p . m and 37°C . 4×108 bacteria were harvested by centrifugation , resuspended in 50 ml of sample solution ( 25% Tris 250 mM pH 6 . 8 , 50% Glycerin , 10% SDS ( w/v ) and 25% β-Mercaptoethanol ) and heated to 100°C for 10 min . Protein concentration of each sample was measured using the BCA assay ( Thermo Scientific , USA ) and 10 µl of a 300 mg/ml final concentration of denatured protein per sample was used for electrophoresis , blotting and incubation with anti-Opc , anti-NadA or anti-NarE . Detection was performed with the Pierce chemiluminescence Western blotting kit ( Thermo Scientific , USA ) . Cell surface ceramide was detected using a spot assay previously described [41] . In brief , 5×105 cells were fixed in 3 . 7% formaldehyde for 15 min at 37°C , scraped from the culture flask , pelleted , washed twice with 1× PBS and incubated with anti-ceramide IgM ( clone MID 15B4 ) for 45 min at RT . Cell bound antibody was desorbed for 30 sec in 20 µl 100 mM glycine-HCl , pH 3 . 0 , and spotted on nitrocellulose following neutralization in 20 µl Tri-HCl , pH 8 . 0 . Antibody was detected with secondary peroxidase -conjugated goat-anti-mouse IgM ( 1∶2500 ) and after washing ( 3 times for 10 min in PBS-Tween ( PBS-T ) ) , enhanced chemiluminescence was used to visualize spots . ASM activity assays were performed using a commercial assay kit according to the manufacturer's instructions ( Sphingomyelinase fluorometric assay Kit , Cayman Chemical , Ann Arbor , MI ) . In brief , 5×105 cells/well were seeded on gelatin-coated 6-well tissue culture plates ( Sarstedt Inc . , Newton , NC ) and infected 72 hrs later at an MOI of 10 . Cells were harvested after removal of the medium and a washing step in PBS by scraping from the culture flask in 500 µl ice cold PBS and pelleted by centrifugation at 800 g for 10 min at 4°C . Pellets were resuspended in 300 µl ice cold SMase acid solution , briefly sonicated ( 20 pulses , 1 s ) , and lysates were separated from cellular debris by centrifugation at 20 , 000× g for 10 min at 4°C . Pellets were resolved in 100 µl SMase acid dilution buffer . For the reaction 10 µl sample per well was applied and incubated for 30 min with Sphingomyelin substrate at 37°C following 100 µl enzyme mixture ( as provided by the manufacturer ) for 30 min at 37°C . The subsequent analysis was performed by using an Infinite F200 Pro Reader ( Tecan Group , Maennedorf , Switzerland ) with an excitation wavelength of 540 nm and an emission wavelength of 590 nm . NSM activity assays were performed using the fluorescent substrate 6-Hexadecanoylamino-4-methylumbelliferyl-phosphorylcholine ( HMU-PC ) ( Moscerdam ) . In brief , 5×104 cells/well were seeded 72 h prior to infection on gelatin-coated 24-well tissue culture plates ( Sarstedt Inc . , Newton , NC ) , grown to a density of approximately 2×105 cells per well and infected at an MOI of 10 . Cells were trypsinised in 200 µl trypsin following addition of 800 µl RPMI medium transferred to a new 1 . 5 ml tube . The suspensions were centrifuged at 800 g for 5 min and pellets were resuspended in 100 µl lysis buffer ( 20 mM Hepes pH 7 . 4 , 2 mM EDTA pH 8 . 0 , 5 mM EGTA pH 8 . 0 , 1 mM sodium vanadate , 10 mM beta-glycerolphosphate , 5 mM DTT , protease inhibitor cocktail ) . After 5 cycles of freezing-thawing in methanol/dry ice , cell suspensions were centrifuged for 5 min at 1 , 600 rpm at 4°C . Supernatants were transferred to a 1 . 5 ml glass tube , overfilled with 500 µl PBS and pelleted for 1 . 5 h at 26 , 000 rpm at 4°C in an SW28 rotor ( Beckman ) . The pellets containing the membrane fractions were resuspended in 40 µl lysis buffer and vortexed gently . Finally , 10 µl resuspension buffer ( 20 mM Hepes pH 7 . 4 , 2 mM EDTA pH 8 . 0 , 10 mM MgCl2 , 0 . 1 mM sodium vanadate , 10 mM beta-glycerolphosphate , 5 mM DTT , 7 , 5 mM ATP and 0 . 2% Triton-X-100 ) , 10 µl HMU-PC substrate , 10 µl membrane fraction and 1 . 8 µl Tric HCl ( pH 8 . 0 ) were mixed and incubated at 37°C overnight . Reaction was stopped by addition of 200 µl stop buffer and enzyme reaction was measured on an Infinite F200 Pro Reader with an excitation wavelength of 404 nm and an emission wavelength of 460 nm . PC-PLC activity assays were performed using a commercial assay kit according to the manufacturer's instructions ( EnzChek , Invitrogen ) . Cell sonicates were prepared according to the methods described by Gomez-Cambronero et al . [62] . For the reaction 100 µl sample ( in 1X PC-PLC buffer ) per well was applied and incubated with 100 µl of 1X PC-PLC substrate ( dye-labelled glycerol-phosphoethanolamine ) ( as provided by the manufacturer ) for 30 min at RT . The cleavage releases the dye-labeled diacylglycerol , which produces a positive fluorescence signal that can be measured . The subsequent analysis was performed by using an Infinite F200 Pro Reader with an excitation wavelength of 490 nm and an emission wavelength of 520 nm . An ASM targeting siRNA were designed to generate dsRNA for post- transcriptional gene silencing of the ASM gene . The following sequences were used for siRNA knockdown experiments: ASM sense 5′-GGUUACAUCGCAUAGUGCCTT-3′ and ASM antisense 5′-GGCACUAUGCGAUGUAACCTT-3′ . SiRNA oligonucleotides were obtained from Sigma-Aldrich ( Sigma-Aldrich , St . Louis , MO ) . As negative control scrambled non-targeted control siRNA [sc-37007] ( Santa Cruz Biotechnology , Santa Cruz , CA ) were used . Control nonsilencing siRNA and ASM siRNA were transiently transfected into HBMEC growing in HBMEC medium using 3 µl of HiPerfect Transfection Reagent ( Qiagen ) according to the manufacturer's instructions . Protein knockdown efficiencies by siRNA transfection were verified by reverse transcription PCR after 72 h of transfection . RNA was extracted from HBMEC tissue using RNeasy Mini kits ( Qiagen ) . cDNA was synthesized from 1 . 0 µg of RNA using Oligo-dT primers ( Invitrogen ) using Superscript II Reverse Transcriptase ( Invitrogen ) according to the manufacturer's instructions . For PCR , the mixture was denaturated at 95°C for 3 min and the target genes were amplified by 35 cycles of reaction: ASM ( 95°C 30 s , 61°C 30 s , 72°C 1 min ) or β-actin ( 95°C 30 s , 61°C 30 s , 72°C 1 min ) . Primers used for real time PCR are as follows: forward human ASM , 5′-CCTTTTGATATGGTGTACTTGGAC-3′ , reverse human ASM , 5′-GTAATAATTCCAGCTCCAGCTCT-3′; forward human β-actin , 5′-GGACTTCGAGCAAGAGATGG-3′ , reverse human β-actin , 5′-AGCACTGTGTTGGCGTACAG-3′ . HBMEC were washed twice with PBS and resuspended in ice-cold fluorescence-activated cell sorting ( FACS ) buffer ( 5% FCS and 0 . 1% NaN3 in PBS ) . 4×105 cells were incubated for 90 min with either mouse mAb anti-ASM IgG antibody ( 1∶250 in FACS buffer ) or mouse mAb anti-Ceramide IgM antibody ( mAb 15B4 ) ( 1∶30 in FACS buffer ) , followed by washing and incubation with Cy5-conjugated goat anti-mouse IgM ( 1∶300 in FACS buffer ) or Cy5-conjugated goat anti-mouse IgG . After the incubation cells were washed and resuspended in 500 µl FACS buffer for the measurement . 10 , 000 cells were analyzed using a FACSCalibur ( BD Bioscience ) and the CellQuest Pro software ( version 5 . 2 ) . For determination of Opc expression on the surface of E . coli , E . coli BL21 ( DE3 ) pET-opc cells were grown and expression of the Opc protein was induced by adding IPTG to a final concentration of 1 mM and incubating the cells for another hour at 37°C under shaking ( 200 rpm ) . Cells were harvested by centrifugation and washed twice with PBS suspended to a final OD600 of 0 . 1/ml for further experiments . Cells were again centrifuged and resuspended in 1 ml FACS buffer . After centrifuging the cells for 10 min at 4 , 500 g ( VWR International GmbH , Darmstadt , Germany ) , the obtained cell pellet was suspended with 200 µL of mouse anti Opc antibody ( clone B306 , diluted 1∶100 in FACS buffer ) and incubated for 60 min on ice . Subsequently cells were washed twice with FACS buffer and bacterial pellets were resuspended in 20 µL of secondary Cy5-conjugated anti-mouse IgG antibody ( 1∶300 in FACS buffer ) and incubated for 30 min in the dark on ice . After washing twice , the cell pellets were finally suspended in 1 . 5 mL of FACS buffer . The samples were analyzed using a flow cytometer using a FACSCalibur and the CellQuest Pro software ( version 5 . 2 ) . Statistical differences between groups were calculated using the Student's unpaired t-test ( two-tailed ) using Excel . P-values ≤0 . 05 were considered significant , P-values ≤0 . 01 were considered highly significant .
|
Neisseria meningitidis , an obligate human pathogen , is a causative agent of septicemia and meningitis worldwide . Meningococcal infection manifests in a variety of forms , including meningitis , meningococcemia with meningitis or meningococcemia without obvious meningitis . The interaction of N . meningitidis with human cells lining the blood vessels of the blood-cerebrospinal fluid barrier is a prerequisite for the development of meningitis . As a major pathogenicity factor , the meningococcal outer membrane protein Opc enhances bacterial entry into brain endothelial cells , however , mechanisms underlying trapping of receptors and signaling molecules following this interaction remained elusive . We now show that Opc-expressing meningococci activate acid sphingomyelinase ( ASM ) in brain endothelial cells , which hydrolyses sphingomyelin to cause ceramide release and formation of extended ceramide-enriched membrane platforms wherein ErbB2 , an important receptor involved in bacterial uptake , clusters . Mechanistically , ASM activation relied on binding of N . meningitidis to its attachment receptor , HSPG , followed by activation of PC-PLC . Meningococcal isolates of the ST-11 clonal complex , which are reported to be more likely to cause severe sepsis , but rarely meningitis , barely invaded brain endothelial cells and revealed a highly restricted ability to induce ASM and ceramide release . Thus , our results unravel a differential activation of the ASM/ceramide system by the species N . meningitidis determining its invasiveness into brain endothelial cells .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"physiology",
"bacterial",
"diseases",
"infectious",
"diseases",
"infectious",
"diseases",
"of",
"the",
"nervous",
"system",
"medicine",
"and",
"health",
"sciences",
"cell",
"biology",
"pathology",
"and",
"laboratory",
"medicine",
"host-pathogen",
"interactions",
"medical",
"microbiology",
"meningitis",
"bacterial",
"meningitis",
"microbial",
"pathogens",
"biology",
"and",
"life",
"sciences",
"microbiology",
"meningococcal",
"disease",
"molecular",
"cell",
"biology",
"membrane",
"trafficking",
"pathogenesis"
] |
2014
|
Differential Activation of Acid Sphingomyelinase and Ceramide Release Determines Invasiveness of Neisseria meningitidis into Brain Endothelial Cells
|
In forebrain neurons , Ca2+ triggers exocytosis of readily releasable vesicles by binding to synaptotagmin-1 and -7 , thereby inducing fast and slow vesicle exocytosis , respectively . Loss-of-function of synaptotagmin-1 or -7 selectively impairs the fast and slow phase of release , respectively , but does not change the size of the readily-releasable pool ( RRP ) of vesicles as measured by stimulation of release with hypertonic sucrose , or alter the rate of vesicle priming into the RRP . Here we show , however , that simultaneous loss-of-function of both synaptotagmin-1 and -7 dramatically decreased the capacity of the RRP , again without altering the rate of vesicle priming into the RRP . Either synaptotagmin-1 or -7 was sufficient to rescue the RRP size in neurons lacking both synaptotagmin-1 and -7 . Although maintenance of RRP size was Ca2+-independent , mutations in Ca2+-binding sequences of synaptotagmin-1 or synaptotagmin-7—which are contained in flexible top-loop sequences of their C2 domains—blocked the ability of these synaptotagmins to maintain the RRP size . Both synaptotagmins bound to SNARE complexes; SNARE complex binding was reduced by the top-loop mutations that impaired RRP maintenance . Thus , synaptotagmin-1 and -7 perform redundant functions in maintaining the capacity of the RRP in addition to nonredundant functions in the Ca2+ triggering of different phases of release .
Synaptic vesicles are released within a few hundred microseconds of Ca2+ influx into a presynaptic terminal [1 , 2] . Exocytosis of synaptic vesicles is carried out by neuronal soluble NSF-attachment protein receptor ( SNARE ) and Sec1/Munc18-like ( SM ) proteins and triggered by Ca2+ binding to synaptotagmins [3] . To prepare for rapid exocytosis with millisecond temporal precision , synaptic vesicles undergo a series of maturation steps that result in the formation of the readily-releasable pool ( RRP ) of vesicles poised for Ca2+-triggered exocytosis . The first step that prepares synaptic vesicles for rapid exocytosis is the recruitment of vesicles to the active zone ( “tethering” ) . Morphologically , tethered vesicles abut the plasma membrane when examined by standard electron microscopy ( EM ) of chemically fixed tissues [4] . After tethering , vesicles undergo a priming process that firmly docks the vesicles at the active zone , as confirmed by EM of unfixed samples subjected to high-pressure freezing , which suggested that priming directly attaches vesicles to the presynaptic active zone downstream of tethering [4 , 5] . As a result , mutations that impair priming cause a loss of vesicle docking when viewed in rapidly frozen unfixed samples , whereas these mutations appear to have no effect on vesicle tethering when chemically fixed samples are examined [4–7] . Strikingly , the only known mutation in mammalian synapses that alters vesicle tethering as viewed in chemically fixed samples is the deletion of Rab3-interacting molecules ( RIMs ) , which are active zone proteins that mediate vesicle tethering by binding to Rab3 and Rab27 proteins on synaptic vesicles [8–11] . Priming of synaptic vesicles produces the RRP of vesicles whose capacity is generally measured by monitoring neurotransmitter release induced by hypertonic sucrose , which stimulates synaptic vesicle fusion by a Ca2+-independent , mechanical mechanism [12] . Two classes of priming factors were identified: active zone proteins such as Munc13 and RIM that are involved in organizing the machinery for synaptic vesicle exocytosis [7 , 13 , 14] , and SNARE and SM proteins that mediate membrane fusion of synaptic vesicles during exocytosis [15–21] . A widely accepted model suggests that priming results in the partial assembly of SNARE and SM protein complexes , an assembly that is catalyzed by active zone priming factors , but the mechanisms that determine the capacity of the RRP are poorly understood [3] . Extensive studies demonstrated that three vertebrate synaptotagmins ( synaptotagmin-1 [Syt1] , Syt2 , and Syt9 ) act as Ca2+ sensors for fast neurotransmitter release [22–25] . Syt1 , Syt2 , and Syt9 are generally expressed in different populations of neurons , although some overlap in expression exists , with most rostral brain neurons expressing only Syt1 and most caudal brain neurons expressing Syt2 [26] . Recent results revealed that a slower and less prominent form of Ca2+-triggered release , which becomes dominant in synapses lacking Syt1 , is mediated at least in part by synaptotagmin-7 ( Syt7 ) , which is abundantly coexpressed with Syt1 , Syt2 , and Syt9 in nearly all neurons [27–29] . A role of Syt7 as a Ca2+ sensor for slow synaptic vesicle exocytosis is supported by findings in neuroendocrine cells where Syt7 , like Syt1 , localizes to secretory granules and mediates Ca2+-induced exocytosis with slower kinetics than Syt1 [30–38] . Viewed together , these observations seem to suggest a linear progression of synaptic vesicle exocytosis from tethering to priming to Ca2+ triggering of fusion , with synaptotagmins mediating the Ca2+-triggering step . Consistent with this hypothesis , individual deletions of Syt1 , Syt2 , or Syt7 had no effect on the size of the RRP [22 , 24 , 28 , 39] . However , three puzzling observations were difficult to reconcile with this hypothesis and suggested synaptotagmin also functions upstream of Ca2+ triggering . First , loss-of-function of complexin , which is an essential cofactor of all synaptotagmins in Ca2+ triggering of fusion [40–44] , decreases the size of the RRP approximately 2–3-fold [45–47] . Second , blocking Syt1 or Syt2 function ( but not Syt7 function ) caused an increase in spontaneous “mini” release in a manner that suggested a role of Syt1 and Syt2 in clamping minis upstream of Ca2+ triggering [24 , 26 , 39 , 48–50] . Ablation of complexin also increased spontaneous mini release at least in some preparations , suggesting a common mode of action of Syt1 or Syt2 with complexin in clamping mini release [3] . Third , immunoprecipitations and pulldowns indicated that Syt1 interacts with SNARE proteins not only in a Ca2+-dependent manner but also in a Ca2+-independent manner [51–54] . Viewed together , these findings raised the possibility that synaptotagmins may perform additional functions besides Ca2+ triggering of release despite the lack of an effect of individual synaptotagmin deletions on RRP size . Based on these observations , we reasoned that synaptotagmins may have a role in exocytosis upstream of Ca2+ triggering by maintaining the RRP , and that this role may have been overlooked in analyses of single Syt1 , Syt2 , and Syt7 knockout ( KO ) neurons because Syt1 and Syt2 might be functionally redundant with Syt7 for priming of exocytosis , even though they are not functionally redundant with Syt7 for Ca2+ triggering . We thus tested in forebrain neurons ( which express little Syt2 or Syt9 ) whether simultaneous ablation of both Syt1 and Syt7 impacts the RRP . We found that loss-of-function of both Syt1 and Syt7 indeed decreased the RRP size ~2–3-fold . Moreover , we observed that the redundant function of Syt1 and Syt7 in maintaining the RRP size requires the flexible top-loop sequences of the Syt1 or Syt7 C2 domains that form their Ca2+ binding sites , even though we confirmed that maintenance of the normal RRP size itself does not require Ca2+ . We show that the Syt1 and Syt7 double loss-of-function does not detectably alter the rate of vesicle priming into the RRP , despite the fact that it dramatically decreases the RRP size . Furthermore , we demonstrate that Syt1 and Syt7 bind to SNARE complexes , that Syt1 increases SNARE complex assembly in the presence of complexin , and that Syt1- and Syt7-binding to SNARE complexes is impaired by the top-loop Ca2+-binding sequence mutations . Our data suggest that at a given synapse , different synaptotagmins mediate distinct phases of Ca2+ triggering of neurotransmitter release but redundantly maintain the capacity of the RRP upstream of Ca2+ triggering , thus enabling the organization of a fast and efficient release machinery at the synapse .
We measured the presynaptic RRP size in cultured hippocampal neurons as the synaptic charge transfer induced by a brief ( 10–30 s ) application of hypertonic sucrose [12] . Consistent with earlier studies [22 , 28 , 39] , individual Syt1 or Syt7 KO or knockdown ( KD ) manipulations had no effect on RRP size ( Fig 1A ) . When we tested hippocampal neurons that lacked both Syt1 and Syt7 , however , we observed a ~60% decrease in inhibitory synaptic transmission , quantified both as the synaptic charge transfer during the initial transient of the inhibitory postsynaptic current ( IPSC ) and as the synaptic charge transfer over the entire period of sucrose application ( Fig 1B and 1C ) . The decrease in RRP size was reversed by reintroduction of wild-type ( WT ) Syt1 or Syt7 but not of mutant Syt1 or mutant Syt7 containing amino acid substitutions in the top-loop sequences of both C2 domains , which include their Ca2+ binding sites ( Syt1C2A*B* and Syt7C2A*B*; Fig 1B and 1C ) . Overexpression of Syt1 or Syt7 rescue proteins in Syt1 KO neurons , tested as a control , had no effect on RRP size ( Fig 1B and 1C ) . The lack of rescue by the mutant Syt1 and Syt7 proteins was not due to impaired mutant protein expression , because we previously showed that these mutant proteins are well expressed in neurons [28] . To rule out the possibility of off-target effects by the Syt7 shRNA used for the Syt7 KD in these experiments , we performed the reverse experiment . We cultured hippocampal neurons from two different Syt7 KO mouse lines [55 , 56] and measured the RRP size as a function of a Syt1 KD [28 , 57] . As expected , Syt7 KO neurons containing Syt1 exhibited a normal RRP size measured as the total sucrose-evoked charge transfer . KD of Syt1 in the Syt7 KO neurons , however , severely depressed the RRP of inhibitory synapses in both Syt7 KO mouse lines ( Fig 2A and 2B ) . Similar to the Syt1 KO findings , re-expression of WT Syt7 , but not of mutant Syt7C2A*B* , restored the RRP size in Syt7 KO/Syt1 KD neurons ( Fig 2A and 2B ) . Taken together , these experiments rule out off-target effects as the origin of the observed RRP phenotype in Syt1/7 double-deficient neurons . To assess whether the redundant requirement for Syt1 and Syt7 in maintaining the RRP size is a general feature of synapses , we next measured the RRP at excitatory synapses as the excitatory postsynaptic current ( EPSC ) evoked by hypertonic sucrose . We found that the RRP size in excitatory synapses was reduced to the same extent by the Syt1 KO/Syt7 KD as in inhibitory synapses ( ~60% ) , and that this reduction could also be rescued by re-introduction of WT Syt7 but not of mutant Syt1 or Syt7 ( Fig 3A and 3B ) . Thus , Syt1 and Syt7 are redundantly required for RRP maintenance in both excitatory and inhibitory forebrain synapses . The observation that mutations in the top-loop sequences of Syt1 and Syt7 ( which also contain their Ca2+ binding sites ) blocked their ability to sustain vesicle priming ( Figs 1–3 ) was surprising because priming of vesicles into the RRP is thought to proceed by a Ca2+-independent mechanism [12] . To confirm that priming as measured by sucrose-induced synaptic transmission under our conditions is indeed Ca2+-independent , we incubated cultured neurons for 30 min in a bath solution containing 10 mM BAPTA-AM . This manipulation blocks nearly all spontaneous miniature synaptic events ( i . e . , miniature EPSCs [mEPSCs] and miniature IPSCs [mIPSCs] ) , which are mostly Ca2+-dependent [49] . When we then assessed the RRP , we found that the BAPTA-AM did not decrease the RRP ( S1 Fig ) . Thus , RRP maintenance in our conditions does not require Ca2+ , not even at resting levels . The phenotype of an apparently decreased RRP size in Syt1/7 double-deficient neurons , as revealed by the results up to this point , could be caused by an impairment in vesicle priming , a decrease in the capacity of the RRP , a decrease in the number of synapses because the Syt1/7 deficiency may produce a synapse loss , and/or a change in the ultrastructural organization of synapses . We previously showed that synapse numbers are unaltered in Syt1/7 double-deficient neurons , ruling out a loss of synapses as a cause of the phenotype [28] . To address the possibility that the simultaneous ablation of Syt1 and Syt7 alters the architecture of the nerve terminal , we performed transmission EM of cultured neurons ( Fig 4 ) . We observed no effect of the Syt1/Syt7 double loss-of-function on any basic ultrastructural parameter of synapses examined , excluding a major effect on vesicle organization in the terminal . Our results suggest that although Syt1 and Syt7 mediate separate essential functions in Ca2+ triggering and are not required for vesicle priming on their own [22 , 28 , 29] , they perform overlapping functions in maintaining the normal capacity of the RRP of synaptic vesicles . Therefore , Syt1 and Syt7 are nonredundant for Ca2+ triggering but redundant for RRP maintenance . In Syt1 , Ca2+ triggering of release is mediated primarily by Ca2+ binding to the C2B domain , although Ca2+ binding to the C2A domain contributes significantly [23 , 50 , 58–62] . In Syt7 , conversely , Ca2+ triggering of release is mediated primarily by the C2A domain [28] . Our current finding that Syt1 and Syt7 perform redundant functions in vesicle priming in addition to nonredundant functions in Ca2+ triggering prompted us to ask whether the RRP function of Syt1 and Syt7 involves the same C2 domains with a similarly asymmetric C2 domain requirement . Mutant Syt1 with top-loop substitutions in the Ca2+-binding sequence of the C2A domain ( Syt1C2A* ) rescued priming of release in Syt1/Syt7 double-deficient neurons , similar to its ability to partially rescue Ca2+-triggered fast release in Syt1 KO neurons ( Fig 5A ) [58 , 62] . Surprisingly , however , mutant Syt1 with analogous substitutions in the C2B domain ( Syt1C2B* ) also rescued priming ( Fig 5A and 5B ) . This latter result was unexpected in view of the selectively essential role of the Syt1 C2B domain Ca2+-binding sequence in Ca2+ triggering of fast release ( Fig 5B ) [23 , 58 , 60 , 62] . Thus , Syt1 exhibits distinct C2 domain requirements for RRP maintenance and for Ca2+ triggering . In contrast to Syt1 , mutant Syt7 with top-loop substitutions in the Ca2+-binding sequence of the C2A domain ( Syt7C2A* ) was unable to rescue the RRP phenotype ( Fig 5A ) . The Syt7 C2B domain mutant ( Syt7C2B* ) , however , fully rescued priming , although the presence of the C2B domain was required for Syt7 function ( Fig 5A and S2 Fig ) . Therefore in Syt7 , the top-loop sequences of the C2A domain but not the C2B domain are selectively required for both priming and triggering of release since we previously showed that the C2A domain mutant also cannot rescue slow Ca2+ triggering in Syt1/Syt7-double-deficient neurons [28] . Although the functions of Syt1 and Syt7 in priming are redundant , their mechanisms of action appear to differ in terms of their C2 domain sequences . We extended these conclusions in further experiments in which we tested point mutations in or next to the Ca2+ binding sites of Syt7 , and additionally examined swap mutations in which the entire Ca2+ binding sequences were exchanged between Syt1 and Syt7 ( S3 Fig ) . We found that the precise sequence of the Syt1 and Syt7 Ca2+ binding site was not a critical determinant of their ability to prime vesicles for release or to mediate Ca2+ triggering of release , but that a single point mutation adjacent to the Syt7 Ca2+ binding sites modestly decreased its priming function . These results corroborate the notion that the top-loop sequences support vesicle priming by a Ca2+-independent mechanism . How do Syt1 and Syt7 redundantly function in maintaining the RRP size ? It is thought that primed vesicles in the RRP are associated with partially or completely assembled SNARE complexes [3] , and Syt1 is known to bind to SNARE complexes both in a Ca2+-independent and a Ca2+-dependent manner [54] . We hypothesized that the requirement for the Syt1 or Syt7 top-loop sequences in priming may be due to an alternative , independent action of these sequences that is unrelated to Ca2+ binding , and that may be related to SNARE binding . Two recent studies using nuclear magnetic resonance ( NMR ) spectroscopy and crystallography to determine how Syt1 binds to SNARE complexes found strong Syt1 binding to SNARE complexes in the absence of Ca2+ , but detected no major interaction of SNARE complexes with the top-loop Ca2+-binding sequences [63 , 64] . The two studies mapped distinct Syt1 binding sites , however , suggesting that Syt1-binding ( and by extension , Syt7-binding ) to SNAREs may be multifaceted . To test whether Syt7 also binds to neuronal SNARE complexes and whether the top-loop mutations in Syt1 and/or Syt7 impair such binding , we measured the effect of mutations in the Ca2+ binding site sequences in Syt1 and Syt7 on the association of Syt1 and/or Syt7 with SNARE complexes in neurons . We expressed HA-tagged WT or mutant Syt1 or Syt7 in cultured neurons using lentiviruses , and immunoprecipitated the SNARE protein syntaxin-1 from these neurons . We then analyzed the immunoprecipitates by quantitative immunoblotting for the SNARE protein synaptobrevin-2 ( to assess SNARE complex assembly ) and for Syt1 or Syt7 ( to assess binding of Syt1 or Syt7 to syntaxin-1 and/or SNARE complexes; Fig 6A and 6B ) . We found that WT Syt1 and Syt7 co-immunoprecipitated with SNAREs to a similar extent , suggesting that both interact with SNARE complexes . The top-loop sequence mutations of Syt1 or Syt7 decreased their coimmunoprecipitations with syntaxin-1 by ~60% but had no effect on the coimmunoprecipitation of synaptobrevin-2 with syntaxin-1 ( Fig 6A and 6B ) . These experiments suggest that Syt1 and Syt7 both bind to SNAREs—either syntaxin-1 alone or SNARE complexes—and that this binding is impaired by the top-loop sequence mutations . How then might Syt1 and Syt7 binding promote RRP maintenance ? We first examined whether Syt1 as the paradigmatic synaptotagmin increases SNARE-complex assembly using proteins expressed in cotransfected HEK293 cells but detected no Syt1-dependent change in SNARE complex assembly ( Fig 7A ) . In contrast , complexin strongly promoted SNARE-complex assembly and stabilized individual SNARE proteins ( Fig 7B ) . Strikingly , when we expressed increasing amounts of Syt1 in the presence of constant levels of SNARE proteins and complexin , Syt1 significantly increased SNARE-complex assembly above the effect of complexin on SNARE-complex assembly ( Fig 8 ) . The effect of Syt1 was independent of Ca2+ in the buffer and thus reflects a Ca2+-independent binding activity of Syt1 ( Fig 8 and S4 Fig ) . These experiments therefore suggest that Syt1 has a direct Ca2+-independent effect on the assembly or stability of SNARE complexes which might account for the function of Syt1 in RRP maintenance , although the atomic mechanisms of this activity by Syt1 are unclear . The priming phenotype we observe could be due to a decrease in the capacity of the RRP , or to a deceleration of the rapid refilling of the RRP , a decrease that may manifest as an overall decrease in RRP size as measured by a 30 s application of hypertonic sucrose . Note that the second hypothesis differs from one suggested earlier that the Syt7 KO causes a decrease in the speed of synaptic vesicle priming [65]—we have not been able to observe any effect of the Syt7 KO alone on priming in either the current experiments ( Fig 1 ) or previous studies [28 , 29] . However , the lack of an effect of the single Syt7 KO on priming does not rule out a possible effect of the double Syt1/7 deficiency on the priming rate , which we tested by the standard approach of applying two short pulses ( 10 s ) of hypertonic sucrose separated by 40 s [66–68] . We again found that the Syt1/7 double deficiency had no effect on the rate of RRP replenishment ( Fig 9 ) . The initial RRP size was greatly decreased by the Syt1/7 double deficiency consistent with our previous results ( Fig 1 ) ; despite its smaller size , the RRP did not refill faster in Syt1/7 double-deficient synapses than in WT synapses . Note , however , that due to the stress of the double pulse of hypertonic sucrose , the access resistance is increased significantly during the second sucrose pulse ( S5 Fig ) , which suggests that these experiments , although standard in the field , should be considered with caution and can only provide clues to the relative recovery rate of the RRP after depletion .
The conclusion that Syt1 and Syt7 are required for maintaining a normal RRP size is based on the observation that simultaneous loss-of-function of both Syt1 and Syt7 decreased the RRP size , whereas loss-of-function of only Syt1 or Syt7 alone had no effect on RRP size ( Figs 1–3 ) . The fact that the RRP phenotype—different from the Ca2+-triggering phenotype [22 , 28]—was only observed upon ablating both Syt1 and Syt7 , and that it was rescued by either WT Syt1 or Syt7 but not by mutant Syt1 or Syt7 with substitutions in their top-loop sequences , demonstrates the specificity of the phenotype . We measured the RRP size via the release elicited by a brief application of hypertonic sucrose [12] . A phenotype observed with this procedure could also have been due to changes in synapse number or synapse structure , but we detected no alterations in either parameter as assessed by immunocytochemistry and EM ( Fig 5 ) [28] . We also confirmed that under our conditions , sucrose-induced synaptic responses reflect a Ca2+-independent priming process , showing that the involvement of Syt1 and Syt7 in priming cannot operate by low-level Ca2+ binding to Syt1 or Syt7 ( S1 Fig ) . Our results raise the important question whether the priming and Ca2+-triggering functions of Syt1 and Syt7 are distinct or identical , i . e . , do Syt1 and Syt7 really function as Ca2+ sensors at all , or only as priming factors ? This question is prompted by our surprising observation that mutations in the top-loop sequences of Syt1 and Syt7 , which contain their Ca2+ binding sites block their priming functions , raising the possibility that what was previously observed as a Ca2+-triggering function may in fact represent a priming function . However , three observations document that synaptotagmins perform distinct functions in RRP maintenance and Ca2+ triggering . First , the Ca2+-triggering functions of Syt1 and Syt7 are clearly apparent in their individual KOs [22 , 27–29 , 39] , whereas their priming functions are not . The Ca2+-triggering phenotype of the Syt1 KO or KD includes a block of nearly all fast synchronous release , while the Ca2+-triggering phenotype of the Syt7 KO or KD impairs delayed asynchronous release [27–29] . Second , Syt1 requires distinct sequences for its priming and Ca2+-triggering functions—the former can be mediated by either the C2A or the C2B domain top-loop sequences , whereas the latter is blocked selectively by mutations in the C2B-domain top-loop sequences ( Fig 5 ) [60–62] . Thus , Syt1 with a C2B domain mutation can still mediate priming but not Ca2+ triggering , demonstrating that the priming and Ca2+-triggering functions of Syt1 are mechanistically distinct . Third , changing the Ca2+-binding affinity of Syt1 changes the Ca2+ dependence of fast neurotransmitter release in a parallel fashion [23 , 54] . Together , these points establish that priming and Ca2+ triggering are separate processes for synaptotagmins . What exactly do synaptotagmins do with respect to the RRP ? The phenotype we observe consists of a major decrease in the RRP size without a change in the rate of RRP replenishment , suggesting that the capacity of the RRP is decreased . Vesicles in the RRP are thought to be docked by partial or complete assembly of SNARE complexes [4] . Syt1 is known to bind to assembled SNARE complexes in both a Ca2+-dependent and a Ca2+-independent manner [54] , and we show here that Syt7 also binds to SNARE complexes ( Fig 6 ) , and that Syt1 can promote SNARE complex assembly in the presence of complexin ( Figs 7 and 8 ) . It is therefore tempting to speculate that synaptotagmins bind to and even promote the formation of assembled SNARE complexes of vesicles in the RRP in a Ca2+-independent manner , and that the RRP size decreases in the absence of synaptotagmins because their binding to assembling SNARE complexes stabilizes the primed state of synaptic vesicles . This model agrees well with recent crystallography studies [64] , although the precise atomic configuration of the synaptotagmin complexes with SNARE complexes remains unclear , and it seems likely that there are multiple binding interfaces that may also serve to multimerize assemblies of synaptotagmins with SNARE–complexin complexes . Thus , we propose that other proteins—in particular Munc13 [73]—catalyze the actual priming reaction by facilitating SNARE complex assembly , but that synaptotagmins stabilize primed vesicles in the RRP , thereby enabling a normal RRP size . The Syt7 KO was reported to decrease the speed of vesicle priming [65] , but this change was not detected in other studies [29] , including the present work ( Fig 9 ) . Moreover , it is hard to imagine how Syt7 , which is highly homologous to other synaptotagmins , could have a completely different function from other synaptotagmins in presynaptic terminals , but a similar function in neuroendocrine cells [30 , 35] . Overexpression of Syt2 fragments in the calyx of Held synapse decreased neurotransmitter release , which was interpreted as a change in vesicle positioning and a disruption of the RRP [74] , but it is difficult to assess the molecular consequences of introducing large amounts of Syt2 fragments into a nerve terminal where such fragments could , for example , interfere with release by indiscriminately coating lipid membranes , or by binding to all assembling SNARE complexes . Thus , the mechanism of the inhibition of release by the overexpressed Syt2 fragments is difficult to interpret vis-à-vis priming . Finally , mutations in SNAP-25 that affect Syt1 interactions were shown in chromaffin cells to decrease priming of neuroendocrine vesicles [75] . Although this observation is consistent with a role of synaptotagmins in priming , it does not exclude the possibility that other actions by SNAP-25—which is a central component of SNARE complexes—are also affected by the SNAP-25 mutations , and it does not directly address the question whether synaptotagmins function in maintaining the RRP size . Thus , our data are broadly consistent with previous studies . Viewed together with earlier studies , our results suggest that synaptotagmins mediate three distinct sequential processes in presynaptic terminals: maintenance of the normal capacity of the RRP , clamping of spontaneous mini release , and Ca2+ triggering of evoked release ( Fig 10 ) . Syt1 likely functions in these three processes by separate molecular mechanisms , because the function of Syt1 in these three processes exhibits distinct C2 domain requirements ( Fig 5 ) [28 , 62] , and it seems likely that this also applies for Syt7 . Although Syt1 and Syt7 perform similar overall functions , they exhibit differences not only in their C2 domain requirements but also in their relative importance for Ca2+ triggering and clamping . In most synapses , the Ca2+-triggering phenotype of the Syt7 KO or KD is small , whereas that of the Syt1 KO or KD is robust . Moreover , the Syt7 KO or KD has no effect on mini release , whereas the Syt1 KO or KD unclamps most synapses . Overall , however , these differences appear minor compared to the functional overlap between Syt1 and Syt7 , as documented by their complete redundancy in RRP maintenance ( Figs 1–3 ) , and the rescue of the clamping phenotype in Syt1 KO synapses by Syt7 overexpression [28] . We propose that Syt1 and Syt7 act sequentially in multiple processes by overlapping mechanisms . Initially , synaptic vesicles are tethered to the active zone by the binding of the active zone protein RIM to the synaptic vesicle GTP-binding protein Rab3 or Rab27 [8 , 9] . At the same time , syntaxin-1 containing bound Munc18 is converted from a “closed” to an “open” conformation , most likely by RIM-bound Munc13 [73] . Together , these two processes enable SNARE complex assembly from vesicular synaptobrevin with plasma membrane syntaxin-1 and SNAP-25 , whereupon complexin and synaptotagmins bind to the partially assembled SNARE complexes in a Ca2+-independent manner . The binding of complexin and synaptotagmins to assembling SNARE complexes is proposed to both boost and stabilize SNARE complex assembly as evidenced by the effect of complexin and Syt1 on SNARE complex assembly in transfected HEK293 cells as a reduced system ( Figs 7 and 8 ) and to mediate clamping of mini release . Subsequent Ca2+ influx into the nerve terminal during an action potential is mediated by Ca2+ channels that are recruited into a location next to docked vesicles by RIMs [9] . The highly localized rapid rise in Ca2+ leads to Ca2+ binding to Syt1 and Syt7 , which ( i ) causes a rearrangement of synaptotagmin-binding to SNAREs , such that the central α-helix of complexin is displaced from the SNARE complex [41] and ( ii ) induces synaptotagmin binding to phospholipids that may cause a mechanical “pull” on these membranes [3 , 76 , 77] . Together , these two Ca2+-triggered molecular events may enable completion of SNARE complex assembly , thereby opening the fusion pore with some contribution from Munc18 . This model accounts for all of the observations described here and additionally agrees with the finding that complexin , the partner for Syt1 in synapses , is also required for vesicle priming , clamping , and Ca2+ triggering [41 , 42 , 45–47 , 78–81] . However , this model does not explain the differential C2 domain requirements and distinct kinetics of the Ca2+-triggering functions of Syt1 and Syt7 . Although this model is consistent with the notion that a macromolecular assembly composed of a synaptotagmin/complexin/SNARE complex serves as a molecular platform for all three functions of synaptotagmins , the precise atomic structure of this assembly has remained elusive . Two recent biophysical studies describe structures of the complex of the C2AB domain fragment of Syt1 with SNARE complexes but come to different conclusions about the Syt1 sequences that mediate this complex [63 , 64] . Moreover , a recent crystal structure reveals multiple interfaces with large contact areas [64] , suggesting that synaptotagmins may engage in multiple different complexes with SNARE complexes and possibly thereby multimerize SNARE complexes . Furthermore , none of the biophysical models involves the top Ca2+-binding loops of Syt1 that were functionally implicated in the present study . Finally , mutations in residues involved in the various interfaces all interfere with Ca2+-triggered release , suggesting that all of these interfaces are important . In view of these results it is , at present , not yet possible to suggest a precise atomic model for synaptotagmin function in concert with SNARE complexes and complexin , but it seems likely that multiple complexes involving all three components—synaptotagmins , complexins , and SNAREs—execute the three functions of synaptotagmins that enable synaptotagmins to act as the central regulators of exocytosis .
This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The protocol was approved by the committee on the Ethics of Animal Experiments ( IACUC ) of the Stanford University ( Protocol Numbers 29589 and 18846 ) . All mouse lines used here were reported previously [28] . Mice were bred using standard procedures and are available at Jackson Labs . All plasmids , including the Syt1 and Syt7 KD lentiviral vectors , and antibodies used were also described previously [28 , 57] . For the KD experiments , we used the following oligonucleotide sequences: Syt7 , KD606 AAAGACAAGCGGGTAGAGAAA and KD607 GATCTACCTGTCCTG GAAGAG; Syt1 , GAGCAAATCCAGAAAGTG CAA . Hippocampal neurons were cultured from WT , Syt1 KO , and Syt7 KO mice as described [39] . Briefly , hippocampi were dissected from newborn pups , dissociated by papain digestion , and plated on Matrigel-coated glass coverslips . Neurons were cultured for 14–16 days in vitro in MEM ( Gibco ) supplemented with B27 ( Gibco ) , glucose , transferrin , fetal bovine serum , and Ara-C ( Sigma ) . The production of lentiviruses and infection of neurons with lentiviruses have been described [41] . Briefly , supernatant with viruses was collected 48 hr after cotransfection of human embryonic kidney 293T cells with the lentiviral vector and three packaging plasmids . This supernatant was used to infect hippocampal neuronal cultures at DIV4 , and cultures were used for biochemical or physiological analyses at DIV14–16 . Electrophysiological recordings in cultured neurons were performed essentially as described [41 , 82] . Briefly , the resistance of pipettes filled with intracellular solution varied between 2–3 MOhm and the series resistance was 7–10 MOhm . Synaptic currents were monitored with a Multiclamp 700B amplifier ( Molecular Devices ) . The frequency , duration , and magnitude of the extracellular stimulus were controlled with a Model 2100 Isolated Pulse Stimulator ( A-M Systems , Inc . ) synchronized with the Clampex 9 or 10 data acquisition software ( Molecular Devices ) . Evoked synaptic responses were triggered by a bipolar electrode . The whole-cell pipette solution contained ( in mM ) 135 CsCl , 10 HEPES , 1 EGTA , 1 Na-GTP , 4 Mg-ATP , and 10 QX-314 ( pH 7 . 4 , adjusted with CsOH ) . The bath solution contained ( in mM ) 140 NaCl , 5 KCl , 2 MgCl2 , 10 HEPES , 10 glucose ( pH 7 . 4 , adjusted with NaOH ) and various concentrations of free extracellular Ca2+ ( 2 mM except otherwise stated ) . AMPAR-mediated EPSCs were isolated pharmacologically with picrotoxin ( 50 μM ) and AP-5 ( 50 μM ) , and recorded at a −70 mV holding potential , NMDAR-mediated EPSCs with picrotoxin ( 50 μM ) and CNQX ( 20 μM ) and recorded at a +40 mV holding potential , and IPSCs with CNQX ( 20 μM ) and AP-5 ( 50 μM ) and recorded at a −70 mV holding potential; all drugs were applied to the bath solution . Note that IPSCs were recorded with a high internal Cl- solution , resulting in large inward currents . mIPSCs and mEPSCs were monitored in the presence of tetrodotoxin ( 1 μM ) in addition to the compounds listed above . Miniature events were analyzed in Clampfit 9 . 02 ( Molecular Devices ) using the template matching search and a minimal threshold of 5 pA and each event was visually inspected for inclusion or rejection . Because the standard software was unable to measure the vastly increased mIPSC frequency in Sy1 KO neurons , we wrote a custom Matlab algorithm that analyzed the first derivative of the trace to identify potential events , followed by screening for the rise time and amplitude . For Ca2+ titrations , eIPSCs were measured for each cell at multiple Ca2+ concentrations starting at 2 mM Ca2+ , followed by measurement of the higher then lower Ca2+ concentration points . Sucrose release was triggered by a 30 s application of 0 . 5 M sucrose and was measured in the presence of 1 μM tetrodotoxin plus additional inhibitors; the synaptic charge transfer was integrated over 30 s . For all electrophysiological experiments , the experimenter was blind to the condition or genotype of the cultures analyzed . Cultured neurons were solubilized in PBS ( with 1 mM CaCl2 , 0 . 2% Triton X-100 , pH 7 . 4 ) supplemented with protease inhibitors ( Roche ) for 1 h . The lysate was cleared by centrifugation at 16 , 000 g for 10 min at 4°C . Immunoprecipitations were performed by incubating with polyclonal antibodies to syntaxin-1 ( 438B ) or preimmune sera for 1 h at 4°C , followed by incubation with 15 μl of a 50% slurry of protein-A Sepharose beads ( GE Healthcare ) for 2 h at 4°C . Beads were washed 4x with 1 ml extraction buffer , bound proteins were eluted with 2× SDS sample buffer containing 100 mm DTT and boiled for 20 min at 100°C . Coprecipitated proteins were separated by SDS-PAGE followed by detection with monoclonal antibodies against an HA epitope included in Syt1 ( HA . 11; 16B12 , Covance ) and synaptobrevin-2 ( cl . 69 . 1 , Synaptic Systems ) . To allow for quantitative detection , dye-conjugated secondary antibodies were used ( IRDye 800CW Donkey anti-Mouse IgG , Li-cor ) , membranes were scanned in an Odyssey scanner ( Li-cor ) , and quantification was performed using Image Studio software ( Li-cor ) . For complexin titration experiments , HEK293T cells were cotransfected with pCMV5 syntaxin-1a , pCMV5 SNAP-25a , pCMV5 synaptobrevin-2 ( 1:1:1 ) , and an increasing amount of pCMV5 complexin-1 ( 0 to 4-fold ) . Total DNA was kept constant by balancing the complexin-1 plasmid with pCMV5 emerald ( 4 to 0-fold ) . For Syt1 titration experiments , HEK293T cells were cotransfected as above , except that pCMV5 complexin-1 was replaced with pCMV5 Syt1 . For Syt1 titration experiments in presence of complexin-1 , HEK293T cells were cotransfected with pCMV5 syntaxin-1a , pCMV5 SNAP-25a , pCMV5 synaptobrevin-2 , pCMV5 complexin-1 ( 1:1:1:3 ) , and an increasing amount of pCMV5 Syt1 ( 0 to 4-fold ) . Total DNA was kept constant by balancing the Syt1 plasmid with pCMV5 emerald ( 4 to 0-fold ) . Two days after transfection , cells were harvested by solubilization for 1h at 4°C in 0 . 1% Triton X-100 in PBS , supplemented with protease inhibitors ( Roche ) . For Syt1 titration experiments in presence of complexin-1 , cells were solubilized in 0 . 1% Triton X-100 in TBS with or without addition of 2 mM CaCl2 , supplemented with protease inhibitors ( Roche ) . Cell lysates were analyzed either directly by quantitative immunoblotting ( for SNARE complexes; high molecular mass bands ) , or were first boiled for 20 min at 100°C ( for total SNARE proteins , complexin-1 , Syt1 , and emerald ) . Antibodies used: synaptobrevin-2 ( cl . 69 . 1 , SySy ) , SNAP-25 ( cl . 71 . 1 , SySy; SMI81 , Sternberger Monoclonals ) , syntaxin-1 ( HPC-1 , SySy ) , complexin-1 ( 122002 , SySy ) , Syt1 ( V216 ) , GFP ( T3743 ) , β-actin ( A1978 , Sigma ) . Cultured neurons were fixed with prewarmed 2% glutaraldehyde in 0 . 1 M sodium cacodylate buffer , pH 7 . 4 at room temperature for 1 h , then stored in 0 . 2% glutaraldehyde in cacodylate buffer . Samples were postfixed in 0 . 5% OsO4 , 0 . 8% potassium ferricyanide ( K3FeCN6 ) in the same buffer at room temperature for 30 min . Specimens were stained en bloc with 2% aqueous uranyl acetate for 15 min , dehydrated in a graded series of ethanol to 100% and embedded in Poly/bed 812 for 24 hr . Thin sections ( 60 nm ) were poststained with uranyl acetate and lead citrate . Samples were examined with a FEI Tecnai transmission electron microscope at 80 kV accelerating voltage; digital images were captured with an Olympus Morada CCD camera . Quantitative analyses were conducted on digital electron micrographs with magnification of 30 , 000x , and the following sample size ( images/synapses analyzed ) : Control , 59/83; Syt7 KD 77/109; Syt7 KD+WT rescue 49/71; Syt7 KD+5DA rescue 28/49 . The number of docked vesicles ( defined as vesicles touching the presynaptic plasma membrane ) , vesicles at the active zone ( defined as vesicles within 100 nm distance from the presynaptic membrane ) , number of vesicles per bouton , PSD length and bouton area were analyzed in ImageJ . Both “docked vesicles” and “vesicles at the active zone” were analyzed to sensitively detect any possible docking phenotype . Images were analyzed by an experimenter blind to the condition and were repeated two times in independent batches of cultures; the results observed were comparable . All data are shown as means ± SEM , and all statistical analyses were performed by one-way ANOVA or Student’s t test .
|
Neurons communicate with each other at specialized contact points called synapses . Presynaptic neurons store chemical neurotransmitters within presynaptic vesicles at the nerve terminal . During synaptic transmission , the presynaptic vesicles fuse with the plasma membrane , releasing their neurotransmitter content into the synaptic cleft to activate postsynaptic receptors . Neurotransmitter release is a multistage process that requires the priming of synaptic vesicles into a readily-releasable pool of vesicles . When an action potential—a transient electrical signal that travels along the neuron—invades a nerve terminal , it promotes the influx of extracellular calcium ions ( Ca2+ ) that , in turn , trigger fusion of primed vesicles , thereby causing neurotransmitter release . Previous studies established that synaptotagmins function as Ca2+ sensors for release and , additionally , inhibit spontaneous fusion of synaptic vesicles in the absence of an action potential . In most neurons of the anterior part of the brain , two synaptotagmins , synaptotagmin-1 and -7 , mediate fast and slow neurotransmitter release , respectively . We now show that in addition to their nonoverlapping roles as Ca2+ sensors and fusion clamps , synaptotagmin-1 and -7 perform an essential overlapping function in maintaining the readily-releasable pool of vesicles . This function is redundantly performed by both synaptotagmins; therefore , an impairment of the readily-releasable pool manifests only when both synaptotagmins are deleted . These results extend the functions of synaptotagmins to steps upstream of Ca2+ triggering of release and suggest that synaptotagmins , despite their simple domain structure , perform multiple sequential roles in neurotransmitter release . Thus , synaptotagmins coordinate multiple stages of Ca2+-triggered exocytosis , ensuring fast synaptic transmission for rapid information transfer between neurons at synapses .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
Synaptotagmin-1 and -7 Are Redundantly Essential for Maintaining the Capacity of the Readily-Releasable Pool of Synaptic Vesicles
|
There is limited published information on the prevalence of human cysticercosis in West Africa . The aim of this pilot study was to estimate the prevalence of Taenia solium cysticercosis antigens in residents of three villages in Burkina Faso . Three villages were selected: The village of Batondo , selected to represent villages where pigs are allowed to roam freely; the village of Pabré , selected to represent villages where pigs are usually confined; and the village of Nyonyogo , selected because of a high proportion of Muslims and limited pig farming . Clustered random sampling was used to select the participants . All participants were asked to answer an interview questionnaire on socio-demographic characteristics and to provide a blood sample . The sera were analysed using an AgELISA . The prevalence of “strong” seropositive results to the presence of antigens of the larval stages of T . solium was estimated as 10 . 3% ( 95%CI: 7 . 1%–14 . 3% ) , 1 . 4% ( 0 . 4%–3 . 5% ) and 0 . 0% ( 0 . 0%–2 . 1% ) in the 763 participants who provided a blood sample in Batondo , Pabré and Nyonyogo , respectively . The prevalence of “weak” seropositive test results to the presence of antigens of the larval stages of T . solium was 1 . 3% ( 0 . 3%–3 . 2% ) , 0 . 3% ( 0 . 0%–1 . 9% ) and 4 . 5% ( 2 . 0%–8 . 8% ) in Batondo , Pabré and Nyonyogo , respectively . The multivariate logistic regression , which included only Batondo and Pabré , showed that village , gender , and pork consumption history were associated with AgELISA seroprevalence . This study illustrates two major points: 1 ) there can be large variation in the prevalence of human seropositivity to the presence of the larval stages of T . solium cysticercosis among rural areas of the same country , and 2 ) the serological level of the antigen , not just whether it is positive or negative , must be considered when assessing prevalence of human cysticercosis antigens .
Taenia solium is a tapeworm transmitted among humans and between humans and pigs . Taeniasis is acquired by humans when eating raw or undercooked pork contaminated with cysticerci , the larval stage of T . solium . When ingested , the cysticerci migrate to the intestine of humans where they establish and become adults . These adult worms shed eggs in human feces that can infect other humans and pigs by direct contact or by indirect contamination of water or food . This can be especially problematic in developing countries where pigs are often allowed to roam freely and to eat human feces and where levels of sanitation and hygiene are poor . Ingested eggs result in larval worms which migrate to different parts of the pig or human body and form cysts . A principle site of establishment of the larvae in humans is the central nervous system . Human neurocysticercosis ( NCC ) occurs when the cysts develop in the brain or spinal cord . Seizures are believed to be the most common presentation of NCC , affecting from 66% to 90% at some stage of their disease [1] , [2] . There is limited published information on the prevalence of human cysticercosis in West Africa . In Burkina Faso , no prevalence study has ever been conducted , although NCC has been reported . In a retrospective review of the medical records of 532 persons with seizure disorder seen either as inpatients or outpatients at Yalgado Ouédraogo Teaching Hospital in Ouagadougou , 6 . 3% of the 158 cases in whom a presumed cause was identified were attributed to NCC based on clinical evidence [3] . No imaging was used to confirm the diagnosis which could have lead to an underestimation of the proportion of seizure cases attributable to NCC . In addition , no definition of “clinical evidence of cysticercosis” was provided , whichmakes this estimate very difficult to interpret . Case reports of human cysticercosis in Burkina Faso have also been published [4] . A third study reviewed 3410 histopathological samples from any location ( surgical and biopsy ) collected between 1991 and 1995 at the two reference hospitals in Bukina Faso and found 18 with evidence of current infection with T . solium larvae [5] . In neighboring pig-raising countries , community-based seroprevalence estimates of cysticercosis in humans range from 1 . 3% to 3 . 95% [6]–[9] . There have also been case reports of human cysticercosis in Ivory Coast , Ghana and Senegal [10] . The main objective of the present study is to estimate the prevalence to the antigens of T . solium cysticercosis as an indicator of current infection , in three villages in Burkina Faso . A secondary aim is to measure the association between potential risk factors and the prevalence of seropositivity to the antigens of T . solium larval stages .
Informed consents for the interviews of participants and the provision of blood samples were obtained separately . The consent process was done orally because a very large proportion of the population had never been to school ( 62 . 5% ) . Oral consent was documented on the individual consent forms by the research staff . The study protocol was reviewed and approved by the ethical committee of the Center MURAZ ( Ref . 02-2006/CE-CM ) and by the Institutional Review Board of the University of Oklahoma Health Sciences Center ( IRB# 12694 ) in regard to both human and porcine participants . Both IRBs approved the use of oral consents . The sampling of blood from pigs was approved by the OUHSC IACUC committee ( approval #06-018 ) . The pilot study was conducted in the villages of Batondo , Pabré and Nyonyogo , located close to the Capital City of Ouagadougou ( Figure 1 ) . The three villages were conveniently selected to represent three types of pig managements . The village of Batondo , located in the commune of Ténado ( province of Sanguié ) 140 km west of Ouagadougou , was selected to represent villages where pigs are owned and raised by women and are allowed to roam freely . The village of Pabré , in the commune of Pabré ( province of Kadiogo ) , located 25 km north of Ouagadougou , was selected to represent villages where pigs are raised and are usually confined for some period of time during the year . The village of Nyonyogo , located in the commune of Dapelogo ( province of Oubritenga ) was selected due to a high proportion of Muslims and hence limited pig farming . A census of all concessions ( a grouping of several households usually members of the same family ) and households in each village was first conducted . In Batondo and Nyonyogo , all concessions were included . In Pabré , 50% of the concessions were selected at random . Within each concession , all households were included and one person was randomly sampled from each household for participation in the interview and venipuncture for collection of blood samples for serological testing . The random selection was done by placing the names of each household member in a bowl and by asking a child to pick one name from the bowl . This cross-sectional study was conducted between May and October 2007 . The head of each household was first interviewed to collect information about each member of his family . In households where pigs were raised , the caretaker of the pigs was interviewed regarding pig management practices . A cooking practices interview was conducted with the wife of monogamous households and with the “senior” woman in polygamous households . The individual sampled at random in each household was asked to answer an epilepsy screening questionnaire which also included socio-demographic information and was administered by a trained member of the study staff . All questionnaires were translated from French to the local languages and back-translated to French . The questionnaires were also pilot tested among a small group of people with and without epilepsy prior to the start of the field study . Blood samples were left to decant at the end of each sampling day and the sera were put in freezers ( −20°C ) until the samples were brought to the IRSS ( Institut de Recherche en Sciences de la Santé ) in Bobo-Dioulasso where they were centrifuged and the sera kept at −20°C . The serum samples were tested for circulating antigens of the metacestode of T . solium using the enzyme-linked immunosorbent assay ( ELISA ) [11] , [12] . This test is designed to measure the presence of current infection with the larval stages of T . solium and not the history of past or present exposure . A seropositive result is indicative of current infection and may or may not be associated with symptoms . The cut-off value was calculated as described by Dorny et al . , 2004 [13] . A ratio for each test was calculated dividing the optical density of the sample by the cut-off value . The ratios were used to classify the results as negative ( ratio between 0 and 1 . 0 ) , “weak” positive ( ratio between 1 . 0 and 1 . 35 ) and “strong” positive ( ratio>1 . 35 ) . Samples with a coefficient of variation of more than 50% were considered as missing values ( n = 3 ) . The sensitivity and specificity of the AgELISA for current infection with cysticercosis has only been reported from a preliminary study conducted in Vietnam . There , the study indicated a sensitivity of 94 . 4% and a specificity of 100% for the diagnosis of current infection with cysticercosis [14] . The sensitivity and specificity of Ag-ELISA for current cysticercosis infection has been determined in pigs , with a sensitivity and specificity of 86 . 7% and 96 . 7% , respectively in Zambia [13] , and 76 . 3% and 84 . 1% , respectively in South Africa [15] . The prevalence of seropositivity to the larval stages of T . solium cysticercosis was estimated separately for “weak” and “strong” seropositivity as the number with “weak” and “strong” positive AgELISA results , respectively , divided by the number of people who provided a blood sample . We then fitted two random-effect models ( xtlogit ) with the “weak” results assumed either positive or negative to estimate the proportion of the total variance contributed by the village-level clustering ( statistic rho ) . The results from Nyonyogo were analysed separately and included only in the univariate analyses due to the small number of seropositives and the very small number of pigs raised in that village , which made this village very different from the two others . For risk factor analyses in Pabré and Batondo , only those with “strong” responses were considered as positive . Univariate associations between being positive to AgELISA and socio-demographic and pork consumption variables at the individual level , as well as pork preparation and pig management variables at the household level were first assessed . Comparisons were made by calculating a prevalence proportion ratio ( PPR ) with 95% confidence intervals ( 95%CI ) . Variables with significant or borderline significant associations with seropositivity in the univariate analyses were then included in a multivariate logistic model adjusting for the effect of village . The results are reported as prevalence odds ratios ( POR ) with 95% confidence intervals ( 95%CI ) . A random-effect logistic regression model with clustering at the concession level was also fitted to take into consideration the clustered nature of the sampling . The results of this model were identical to those from the simple model , however , and thus , only the latter are presented . All analyses were conducted in Stata 10 SE .
A total of 888 individual interviews were conducted with participants in the three villages . All sampled individuals agreed to answer the interview questionnaire . Of these , 766 ( 86 . 3% ) provided a blood sample . Table 1 shows the proportion of participants providing a blood sample by selected socio-demographic characteristics . Briefly , the proportion of people providing a blood sample varied somewhat from village to village due to a variety of factors such as the presence of visible veins , the difficulty of obtaining a blood sample , and refusal to provide a sample following the interview . Among those who provided a blood sample , 763 had a valid AgELISA test result . The prevalence of “strong” seropositive test results was estimated as 10 . 3% ( 95%CI: 7 . 1%–14 . 3% ) , 1 . 4% ( 0 . 4%–3 . 5% ) and 0 . 0% ( 0 . 0%–2 . 1% ) in Batondo , Pabré and Nyonyogo , respectively . The prevalence of “weak” seropositive test results was 1 . 3% ( 0 . 3%–3 . 2% ) , 0 . 3% ( 0 . 0%–1 . 9% ) and 4 . 5% ( 2 . 0%–8 . 8% ) in Batondo , Pabré and Nyonyogo , respectively . The random-effect models showed that the variance due to the village contributed to a large proportion of the overall variance . The rho statistic was estimated to 0 . 49 ( 95%CI: 0 . 08–0 . 92 ) when the “weak” seropositives were assumed negative and to 0 . 16 ( 95%CI: 0 . 03–0 . 55 ) when the “weak” seropositives were assumed positive . Table 2 shows the prevalence of AgELISA seropositivity within categories of several potential risk factors and stratified by village . In Nyonyogo , univariate analyses showed males to have an increased prevalence of presenting a “weak” serological results as compared to females ( PPR = 8 . 02 ( 95%CI: 1 . 01 , 63 . 86 ) ) . None of the 8 cases with “weak” results had gone to school and none reported using the toilet to defecate . However , only 18 . 2% and 11 . 3% of the population of the village had ever attended school and reported using a toilet to defecate , respectively . Nyonyongo children aged less than 16 tended to have a higher prevalence proportion , based on “weak” results , than adults with a PPR = 3 . 63 ( 95%CI: 0 . 95–13 . 95 ) , and there was a tendency for a higher prevalence in concessions of larger sizes . This latter trend by concession size was observed to a lesser extend in Batondo , but not in Pabré . Except for the association with gender , none of the variables noted in relation to seroprevalence ( “weak” only ) in Nyonyongo were observed in the other two villages ( with “strong” seropositive tests ) ( Table 2 ) . Because there were so few pigs being raised in Nyonyogo , “weak” seropositive tests were not found to be associated with pork consumption or pig raising in this village . The multivariate logistic regression , which included only Batondo and Pabré , showed that village , gender , and pork consumption habits were associated with “strong” AgELISA seropositivity ( Table 3 ) . The odds of being seropostive were considerably higher in Batondo than in Pabré ( POR = 8 . 86; 95%CI = 3 . 01 , 26 . 14 ) , in men compared to women ( POR = 2 . 34; 95%CI = 1 . 10 , 4 . 97 ) and in those eating pork either in the past ( POR = 19 . 62; 95%CI = 1 . 91 , 2010 . 95 ) or currently ( POR = 8 . 75; 95%CI = 1 . 11 , 68 . 88 ) compared to those who never ate pork . The ownership of pigs by one household member confounded the association between seropositivity and pork consumption such that when this variable was included in the model the association between eating pork and seroprevalence became significant . Although not itself statistically significant , because of this interaction the variable “pig ownership” was retained in the multivariate model .
This study is the first to estimate the seroprevalence to the presence of antigens to T . solium cysticercosis in Burkina Faso . The strengths of this study are that the results are based on a clustered-random sample of residents of three rural villages , the participation proportion for the interview was excellent ( 100% ) and very good ( 86 . 3% ) for the serology , the participants answered almost all questions in the interviews ( few missing values ) , and the majority of serological tests conducted had valid results . Our results show considerable variation in seroprevalence in the three study villages . The prevalence to the presence of antigens of T . solium cysts was nearly 8 times higher in Batondo than in Pabré . Several reasons may explain the difference between Batondo and Pabré . For example , the proportion of participants who had gone to school was much higher in Pabré ( 55 . 6% ) than in Batondo ( 30 . 0% ) , and a larger proportion of the participants used the toilet to defecate in Pabré ( 37 . 6% ) than in Batondo ( 7 . 9% ) . Pigs were also more often penned in Pabré ( 54 . 8% ) during the rainy season as compared to Batondo ( 10 . 9% ) where pigs were more often tethered ( 94 . 0% ) . Tethered pigs are more likely than penned pigs to have access to human feces as they are often moved to be able to feed , and hence , are more likely to be exposed to a contaminated site . It was also observed during the field study that pigs were living in closer contact with humans in Batondo than in Pabré . These factors , possibly in addition to other village-level variables that were not measured , may contribute to the lower prevalence of the presence of antigens of T . solium cysts in Pabré compared to Batondo . All seropositive results in Nyonyogo were “weak” . One could speculate , as an hypothesis yet to be tested , that the force of infection from the environment is lower in Nyonyogo due to the presence of very few pigs and therefore , to a lower prevalence of taeniasis among the population . It is also possible that the source of infection is different in Nyonyogo compared to that in the other villages as discussed below . Another hypothesis is that most of the infections in Nyonyogo were either very recent or old resulting in a lower density of antigens in the blood . Although based on a small number of “weak” cases , the finding that children tended to have a higher seroprevalence than adults only in Nyonyogo deserves further exploration . One hypothetical explanation for this observation is that children in Nyonyogo acquire infection through playing in the contaminated environment since few adults consume pork and thus are at lower risk for taeniasis . Sanitation was very poor in Nyonyogo with only 14 . 4% of the household having a toilet and 11 . 2% of the people using the latrine to defecate . In the other two villages , more people consume pork and the prevalence of taeniasis is probably higher , which could increase the risk of auto-infections or infection through the contamination of food and water . This possibility could also explain the difference in the strength of AgELISA optical densities between the villages . These interpretations are hypotheses which would need to be verified in a larger cohort study . To our knowledge , there has been only one other community-based sero-survey done in Sub-Saharan Africa on the presence of antigens to the larval stages of T . solium cysticercosis [9] , which reported a seroprevalence varying between 0 . 4% and 3 . 0% in the three rural communities in Menoua district , Cameroon , between 1999 and 2000 . One important limitation of this study is that sampling was done among volunteers . Nonetheless , the study's results indicate some variation in prevalence by community , but less variation than what was observed in the present study in which participants were selected according to a clustered random sampling strategy . Also , we had specifically sampled a village where the majority of the population was Muslim and where there were very few pigs being raised . In a recent hospital-based case-control study from the Kiremba area of Burundi , the prevalence of seropositivity using AgELISA in controls ( persons without epilepsy ) was estimated as 20% [16] . In this study , 80% of the controls were people being vaccinated at the Kiremba area hospital and age-matched to a group of patients with epilepsy who were being seen at that hospital . This design makes it difficult to assess the source of the controls and therefore assess to what extent they represent the general population since controls were age matched to epilepsy patients . In a study in Cameroon of people with epilepsy receiving care in rural clinics , the seroprevalence by AgELISA was estimated as 1 . 2% [17] . The estimated prevalence of seropositivity to the antigens of Taenia solium cysts among people in our sample with confirmed epilepsy was considerably higher at 15 . 2% . In the Burundi case-control study , the POR of seropositivity to the antigens of Taenia solium cysts was 2 . 5 ( 95%CI = 1 . 8 , 3 . 4 ) when comparing males to females and 1 . 7 ( 95%CI: 1 . 1–2 . 5 ) when people eating pork were compared to those not eating pork [16] , results similar to those of the present study , although the association with pork consumption in our cross-sectional study was much larger . In Batondo and Pabré , the presence of pigs within the household tended to reduce the seroprevalence and it confounded the association between pork consumption and seroprevalence . This is because there was a higher proportion of participants from households where pigs were raised who consumed pork than in household where pigs were not raised . This uneven distribution of pork consumption according to the presence of pig raising leads to an underestimation of the effect of pork consumption on the seroprevalence to the antigens of Taenia solium cysts if not adjusted . The proportion of participants aware of the link between consumption of undercooked pork and taeniasis was similar among people who raised pigs ( 15 . 1% ) and those who did not ( 14 . 2% ) . It may be that members of households where pigs are not raised are more likely to consume pork at the market than are those from household where pigs are raised , which may expose them to greater risk than those eating their meals at home . Unfortunately , our questionnaire did not include a question on the location of consumption of pork meat , which would have helped in explaining this confounding factor . This study has some limitations . First , 13 . 7% of the 888 interviewed participants did not provide a blood sample . The reasons for not providing a sample may have been linked to the difficulty in obtaining blood at the time of sampling or refusal to provide a sample following the interview . Refusals were a minority of those who did not provide a sample . Due to this , we believe that important selection bias is unlikely . Second , it would have been very interesting to re-test participants with “weak” results a few months later . This would have indicated whether those “weak” cases were new or old infections . Unfortunately , this was not feasible in the context of this project . Third , it would have also been interesting to obtain results from a valid antibody serological test . We did test the samples for the presence of antibodies according to the method of Arruda et al . 2005 [18] . However , this test is now thought to be invalid due to several cross reactions with other helminthes and protozoa such as Echinococcus , Filaria , Fasciola , Strongyloides , Schistosoma , Toxocara , Amoeba , and Plasmodium [19] . Another alternative would have been to test the samples using the well validated Western Blot test ( EITB ) for the detection of antibodies to T . solium [20] . This was not feasible due to the complexity and cost of the test . However , as a test for the presence of current infection with the larval stages of T . solium , preliminary studies indicated a sensitivity of 94 . 4% and a specificity of 100% of the Ag ELISA test in Vietnam [14] . We have shown that the prevalence to the antigens of the larval stages of T . solium can be very high in some villages of Burkina Faso and virtually nonexistent in other villages in the same region . This study illustrates two major points; first , there can be large variation in the prevalence of antigens to human cysticercosis among rural areas of the same country , and second , the serological level of the antigen , not just whether it is positive or negative , must be considered when analyzing data in order to arrive at more valid conclusions . The first point is especially relevant to studies of the burden of diseases . If all studies are concentrated in areas where pigs are roaming , the overall burden of the infection for the country would be over-estimated; the converse would be true as well .
|
Taenia solium cysticercosis is a neglected tropical zoonosis transmitted between humans and pigs . This infection is particularly prevalent in areas where sanitation , hygiene and pig management practices are poor . There is very little information about the importance of this infection in West Africa , even though pork meat is widely consumed in many areas . This pilot study , conducted in three villages of Burkina Faso , demonstrated that people living in areas where pigs are raised were more likely to be infected with cysticercosis than people living in a Muslim village in which there were very few pigs . It also demonstrated variation in the level of infection between the two villages where pigs were raised . Finally , the results suggest that the source of infection in these three villages may differ . These results are significant because they show that there is clustering of infection within villages , even if they are geographically very close to one another . This should encourage future researchers not to combine data from several villages into one summary value . In addition , more work is needed to better describe different potential sources of infection among villages .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"public",
"health",
"and",
"epidemiology/epidemiology",
"infectious",
"diseases/helminth",
"infections",
"public",
"health",
"and",
"epidemiology/infectious",
"diseases",
"public",
"health",
"and",
"epidemiology/social",
"and",
"behavioral",
"determinants",
"of",
"health"
] |
2009
|
Seroprevalence to the Antigens of Taenia solium Cysticercosis among Residents of Three Villages in Burkina Faso: A Cross-Sectional Study
|
Entomological indicators are considered key metrics to document the interruption of transmission of Onchocerca volvulus , the etiological agent of human onchocerciasis . Human landing collection is the standard employed for collection of the vectors for this parasite . Recent studies reported the development of traps that have the potential for replacing humans for surveillance of O . volvulus in the vector population . However , the key chemical components of human odor that are attractive to vector black flies have not been identified . Human sweat compounds were analyzed using GC-MS analysis and compounds common to three individuals identified . These common compounds , with others previously identified as attractive to other hematophagous arthropods were evaluated for their ability to stimulate and attract the major onchocerciasis vectors in Africa ( Simulium damnosum sensu lato ) and Latin America ( Simulium ochraceum s . l . ) using electroantennography and a Y tube binary choice assay . Medium chain length carboxylic acids and aldehydes were neurostimulatory for S . damnosum s . l . while S . ochraceum s . l . was stimulated by short chain aliphatic alcohols and aldehydes . Both species were attracted to ammonium bicarbonate and acetophenone . The compounds were shown to be attractive to the relevant vector species in field studies , when incorporated into a formulation that permitted a continuous release of the compound over time and used in concert with previously developed trap platforms . The identification of compounds attractive to the major vectors of O . volvulus will permit the development of optimized traps . Such traps may replace the use of human vector collectors for monitoring the effectiveness of onchocerciasis elimination programs and could find use as a contributing component in an integrated vector control/drug program aimed at eliminating river blindness in Africa .
Onchocerca volvulus , the etiological agent of “river blindness” , remains a major public health threat in much of Africa and in a cross-border focus between Venezuela and Brazil [1] . Recent studies suggest that mass distribution of Mectizan ( ivermectin ) may be effective in achieving local elimination of the parasite [2] , [3] and have resulted in a strategic shift from control of river blindness towards elimination [1] . The World Health Organization and other international onchocerciasis elimination programs rely on measurement of the presence and intensity of transmission of the parasite for the verification of onchocerciasis elimination , both of which are measured by surveillance of the vector population [4] . Current methods to measure parasite transmission employ human landing collections to attract and collect vector black flies ( Simulium spp . ) ; infection data from the flies collected are then used for real-time surveillance , implementation of mass drug treatment and decision-making based on the outcome of epidemiological models [5] . However , the use of such collections is problematic for several reasons . First , collections of this type pose ethical questions in endemic areas , as they might result in exposure of the collectors to O . volvulus and perhaps other uncharacterized pathogens present in these flies . Second , verifying interruption of transmission requires collecting and screening large numbers of flies . For example , the current guidelines adopted by the Onchocerciasis Elimination Program for the Americas require that sufficient flies be collected from each endemic community to ensure that the upper bound of the 95% confidence interval of the number of flies carrying the infective stage of the parasite ( L3 ) is less than 0 . 05% , or 1/2000 flies [5] . To achieve this , it is necessary to collect and screen at least 6000 flies from each community [6] , [7] . Human landing collections are labor intensive and often cannot collect such a large number of flies in a cost effective manner . For these reasons , there is a need to develop an alternative to human landing collections for onchocerciasis surveillance . Recent efforts have been made to develop traps capable of collecting vector black flies in a variety of ecological settings [8] , [9] . These studies have shown that a novel trap platform , the Esperanza Window Trap , which consists of an adhesive-coated blue or black and blue striped fabric square , when baited with CO2 and worn clothing collected numbers of vector black flies that were similar those obtained by human landing collections [8] , [9] . Both CO2 and the odor baits were found to be necessary to attract significant numbers of vector flies [8] , [9] . However , to be widely applied as a surveillance tool to replace human landing collections , it will be necessary to develop a consistent bait formulation to replace the worn clothing as bait , allowing the traps to function consistently over time and space . In the experiments presented below , we report the results of studies identifying human-derived compounds that are attractive to Simulium ochraceum sensu lato and S . damnosum s . l . The savanna dwelling sibling species of Simulium damnosum s . l . are the vectors of the blinding form of O . volvulus and represent the most important vector species on the African continent , while S . ochraceum s . l . was the primary vector in the historically largest endemic foci in Mexico and Guatemala [10] . To accomplish this goal we have utilized an approach involving a combination of electrophysiological and behavioral assays , a process that has proven to be effective for the identification of attractive volatiles for a variety of other medically important vector species [11]–[13] .
Sweat was collected from the face , armpit and groin areas of three male volunteers after each subject engaged in 30 minutes of aerobic exercise ( Fig . 1A ) . These areas were chosen as S . ochraceum s . l . usually bites on the head and upper part of the body [14] while S . damnosum s . l . targets the lower part of the body [15] , and previous studies demonstrated that sweat impregnated shirts were attractive to S . ochraceum s . l . [8] while sweat impregnated pants were attractive to S . damnosum s . l . [9] . Samples were taken by wiping the chosen areas thoroughly with sterile cotton swab . Samples were collected from each individual on three separate occasions over a three-week period . The cotton wool containing each sweat sample was immediately placed into a separate headspace-analysis vial and sealed . Starting immediately , then every 24 hours thereafter for 5 days , headspace analysis of each sweat sample was conducted by Gas Chromatography ( GC ) using carboxen-polymethylsiloxane and polyacrylate solid phase microextraction ( SPME ) fibers ( Fig . 1B ) . Samples were analyzed sequentially over a five day period to permit the development of metabolites from the skin flora , as these have been shown to be attractive in previous studies [16] , [17] . Samples were maintained at ambient temperature between analyses . Prior to introduction of an SPME fiber for each analysis , the sweat sample was incubated at 50°C for three minutes . The SPME fiber was left in contact with the headspace for five minutes before being removed for GC analysis . Control samples , comprising the full experimental sample preparation method but without application of the cotton wool to the human subjects , were analyzed separately . SPME fibers were analyzed on an Agilent 7200 GC-QToF . The sweat components were thermally desorbed in the GC inlet at 320°C , where the individual components were resolved on a HP-5ms capillary column ( 30 m , i . d . 0 . 25 mm ) using helium as the carrier gas at a flow rate of 1 . 2 mL/min . The column was heated from 60°C to a final temperature of 320°C at a rate of 20°C/min . Compounds were ionized using electron ionization ( EI ) for initial identification . Chemical ionization ( CI ) was used to confirm molecular ions using methane as the reagent gas and both positive and negative mode ( 40–110 eV , compound dependent ) to obtain optimum signal-to-noise ratio . The total ion chromatographs were deconvoluted using Agilent's Deconvolution Algorithm and the resulting EI induced fragmentation patterns were exported as CEF files and imported into Mass Profiler Professional for statistical analysis . Datasets , prepared for each time point and from both SPME fiber materials , were combined to create an entity list for each subject containing all compounds present in their sweat . Compounds in the control samples ( i . e . those not containing and sweat ) were removed from the entity lists . Spectra of entities identified as common in the sweat of all individuals were compared to the fragmentation patterns contained in the NIST/Wiley 2011–2012 mass spectra library for a tentative identification . Finally , retention times , molecular ions and fragmentation patterns of commercially available standards were compared to tentatively identified compounds for final confirmation . Host-seeking Simulium spp . females were captured by vector collectors in the field near Bodadougou , Burkina Faso ( S . damnosum sensu stricto and S . sirbanum ) and Unión Juárez , Chiapas , México ( S . ochraceum sensu lato ) , identified using morphological characters , and used for experiments in a field laboratory within 24 hours of capture ( Fig . 1D ) . Preliminary trials indicated that a preparation consisting of the thorax and head produced the best signal to noise ratio for EAG recordings . Glass electrodes ( 0 . 2 mm dia . silver chloride-coated wires ) were connected to the tip of one antenna and the exposed internal tissues of the metathorax . All test compounds were purchased from Sigma Aldrich or Fisher Scientific; only compounds with purity higher than 98% ( with the exception of lactic acid which was 90% pure ) were used . Unless specified , all chiral molecules were racemic mixtures of enantiomers . Test compounds were dissolved in hexane ( HPLC grade , ≥95% ) or distilled water to yield a solution ( 1/100 , or roughly 80 mM ) for EAG analysis . Test solutions ( 10 µL ) were applied to a strip of Whatman filter paper ( Whatman , Inc . USA ) , the solvent allowed to evaporate , and the paper strip inserted into a glass Pasteur pipette . An air stimulus controller ( CS-55 , Syntech , the Netherlands ) generated a pulse ( 0 . 2 s duration ) of filtered air that introduced headspace volatiles from the pipette into a continuous stream of humidified air ( 1000 mL/min ) that was directed at the antenna of the test insect , initiating the EAG recording . Electrodes used to monitor a stimulus were Ag/AgCl wire submerged in freshly prepared saline solution ( 750 mg NaCl , 35 mg KCl and 29 mg CaCl2•2H2O in 100 mL of distilled water ) , which was placed in each of the glass electrodes prior to analysis and changed between flies . Individual compounds were randomly assigned to one of 10 groups ( 5–7 compounds per group ) . Within each group , the order in which compounds were assayed was randomized for each trial . EAG responses of hexane ( control ) and a blank ( puff of air ) were recorded before and after each group of compounds ( Fig . 1E ) . Each compound was initially assayed in three replicates via EAG . Compounds that elicited an EAG response that did not differ from the control ( hexane ) response were not used in subsequent assays . Compounds that elicited EAG responses consistently greater than hexane in 2–3 replicates were further assayed with additional replicates . EAG responses of compounds were normalized to the preceding response of the control stimulus ( hexane ) . A t-test was used to determine if normalized EAG responses of compounds differed from that of the control ( α = 0 . 05 ) . Compounds eliciting significant EAG responses then became candidates for Y-tube evaluation . Wild-caught Simulium spp . were collected as noted above , stored in plastic vials , and transported in coolers to a local field laboratory . Experiments were conducted within 24 hours of collection . Behavioral bioassays were conducted in a Y-tube olfactometer ( Fig . 1F ) , modified from an earlier design [18] . The Y-tube was made of transparent acrylic with an internal diameter of 1 . 5 inches and consisted of four sections: release chamber , Y-split chamber , stimulus arm trap chambers , and stimulus chambers . The release chamber contained an aspirator entry point on the ventral side , a fabric mesh screen on the dorsal side , and a metal-mesh rotating door in the apical side . The mesh partitions allowed air to flow through the release chamber and the rotating door was used to release the flies after acclimation into the Y-split chamber , which connected the release chamber to the stimulus arm trap chambers . Each stimulus arm trap chamber contained a rotating metal mesh screen door , which closed after the test run to prevent any flies from entering or exiting the stimulus arm chamber . A stimulus chamber was connected to the apical end of the stimulus arm trap chamber that was separated by a fabric mesh screen . The stimulus chamber was accessible by a sliding door . During all experiments , a black cloth covered the Y-tube olfactometer completely , except the apical end of the stimulus chambers . This opening allowed light ( Utilitech fluorescent plant grow light ) to enter the apical end of the stimulus chambers and was required to initiate activation of the flies . Hydrocarbon-filtered air ( LabClear , Diamond Tool and Die , Inc . , Oakland , CA ) was supplied into the apical end of the stimulus chambers by two separate air-lines from an air pump ( Greentrees Hydroponics , Vista , CA ) . The flow of filtered air into the stimulus chambers was regulated ( Brooks Instruments , Hatfield , PA ) at a rate of 1 . 5 liters per minute . Initially , each group of flies ( n = 20 ) was acclimated to filtered air pumped through the Y-tube while held in the release chamber for 10 minutes , with no compound present in either arm . After acclimation , the flies were released and allowed 1 minute to make a choice in the Y-tube . If no preference to a particular arm was observed ( i . e . p>0 . 05 using the Chi Square likelihood ratio test based on the multinomial model described below ) , the flies were collected back into the release chamber . If a preference was observed in a particular control run , the flies were discarded and the Y-tube was cleaned and the process repeated until no preference was observed . A test run was then performed in which one arm contained the test compound and the other arm the solvent alone . Regardless of whether attraction to a given compound was observed , the Y-tube was disassembled and cleaned with a detergent water solution and ethanol after every other experiment ( i . e . after the paired control and test runs ) . A total of six replicates per compound were performed , assaying a total of 120 flies per compound . Liquid test and control stimuli were introduced into the Y-tube olfactometer by impregnating 20 µl of the test or control solution onto a 2 . 5-cm circular piece of filter paper ( Whatman grade 1 , GE Healthcare , Little Chalfont , UK ) . Sufficient time was allowed for the solvent to evaporate before the impregnated filter paper was placed in the stimulus chamber , where a clamp held the filter paper in place . Each filter paper was prewashed with hexane before the experiment . The attraction response was measured for each test run replicate and all compounds were tested using ten-fold dilutions ( 1∶10 v/v; 1∶100 v/v; 1∶1000 v/v ) to determine the overall range of attractiveness of each candidate compound . The test attraction response was calculated using the formula: Test attraction response = ( T x 100 ) / ( T+C ) where T denotes the number of flies trapped on the test stimulus side and C denotes the number of flies trapped on the control stimulus side . The same calculation was performed to find the control attraction response . The proportion of the attraction responses were transformed via arcsine transformation before the means were calculated for comparison . The proportions of flies that chose the control and the test stimuli were compared using a likelihood ratio test based on a multinomial probability model . Test attraction responses for all test stimuli were compared by means of pair-wise comparisons based on the multinomial model with alpha level adjusted to account for multiple comparisons , using a custom program written in FORTRAN 95 . The program is available upon request . Compounds found to be attractive in the Y tube assay were absorbed into plastic aroma beads ( Bitter Creek Candle Supply [www . candlesupply . com] ) for use as artificial baits . All beads were prepared at a ratio of 0 . 222 mL active compound per 1 g of aroma beads , and allowed to absorb candidate compounds over a 12-hour period . The beads were then allowed to rest for 24 hours to permit residual unincorporated solution to evaporate . To determine kinetics of compound release , two beads loaded with each compound were then placed into individual headspace vials for analysis , and incubated at room temperature . The amount of compound remaining was quantified by GC/MS at 24-hour intervals for five days . Samples were introduced into the GC inlet after being allowed to adsorb onto a carboxen-polymethylsiloxane SPME fiber from a headspace vial incubated at 35°C for 2 minutes . SPME fibers were analyzed on an Agilent 7000 GC/QqQ MS . The individual compounds were thermally desorbed in the GC inlet at 320°C , where the retention time was monitored on a HP-5ms capillary column ( 30 m , i . d . 0 . 25 mm ) using helium as the carrier gas at a flow rate of 1 . 2 mL/min . The column was heated from 30°C to a final temperature of 320°C at a rate of 40°C/min . Compounds were ionized using electron ionization ( EI ) where the total ion chromatogram was integrated and used to determine the percentage of compound remaining , using Agilent's MassHunter B . 5 . 00 software . Beads containing the attractive compounds were prepared as described above . The field-ready bait was prepared by placing aroma beads loaded with individual compounds ( 4 . 5 g per compound ) into nylon stockings , along with 4 . 5 g of powdered ammonium bicarbonate separated from the beads by a knot in the stocking . The baits were placed on each side of an EWT trap , utilizing versions of the trap that were optimized for use in Mexico [8] and Burkina Faso [9] . As previous studies had demonstrated that CO2 was necessary to induce attraction to the trap platforms baited with used clothing [9] all traps were also baited with organic CO2 , which was prepared as previously described [8] . Comparisons were made between traps with CO2 alone ( control ) and traps with CO2 + the mixture of odorant compounds ( Fig . 1G ) . For S . ochraceum s . l . , traps were set in San José , Chiapas , México , and run from 8 am to noon for 6 days . The field evaluation was conducted during the early dry season; January 16th through February 1st 2014 . The mixture of odorant compounds evaluated for the collection of S . ochraceum s . l . contained 1-octen-3-ol , 1-octanol , acetophenone , hexanal , and ammonium bicarbonate . Seven pairs of traps , located 15-20 meters apart , were used to evaluate each condition and the position of each trap was rotated daily to avoid location bias . Simulium damnosum s . l . data were collected using paired traps run at Bodajugu , Burkina Faso , over a seven-day period . Traps were run from sunrise to sunset , and their positions alternated from day to day . The trap evaluation was conducted in the dry season ( March 27th through April 1st ) . This portion of the river is located downstream of a dam , resulting in conditions that support fly breeding throughout the year . Baits evaluated for the collection of S . damnosum s . l . contained hexanoic acid , heptanoic acid , octanoic acid , nonanoic acid , 1-decanal , acetophenone , and ammonium bicarbonate . Statistical analysis of data collected in both field settings was performed using SAS Proc GENMOD to analyze the data as a Negative Binomial regression . The experiments included in this study were reviewed and approved by the Institutional Review Board for Human Subjects Research of the University of South Florida . The board declared that the project qualified for the expedited review procedure authorized by federal regulations 45CFR46 . 110 and 21 CFR 56 . 110 under category 3: “Prospective collection of biological specimens for research purposes by noninvasive means” . The review board further determined that the work qualified for a waiver for documentation of informed consent as outlined in federal regulations at 45CFR46 . 117 ( c ) , which state that an IRB may waive the requirement for the investigator to obtain a signed consent form for some or all subjects . As a result , oral consent was obtained from all participants . Participants were asked to sign a form indicating that the purpose and procedures employed in the study had been explained to them . The signed forms were retained by the principal investigator of the project .
As a first step in identifying human odorants attractive to Simulium spp . , sweat compounds were identified from samples collected from three individuals using GC-MS as described in Materials and Methods ( Fig . 2 , Panel A ) . A total of 1 , 261 compounds were identified from the three volunteers , of which the vast majority ( 90% ) were unique to one individual ( Fig . 2 , Panel B ) . Just 29 compounds were common to all three individuals . These were chosen for further evaluation in the electrophysiological and behavioral assays . In addition , 25 other compounds consistently reported in the literature to occur in sweat and/or eliciting attractive behavior in other blood-feeding arthropods [19] , [20] were included in the subsequent studies . Taken together , 54 compounds were evaluated by electroantennography ( Table 1 ) . Putative soluble attractive compounds were then assayed for neurostimulatory effects via electroantennography ( EAG ) using field-collected , host-seeking flies . Of the 54 compounds tested , nineteen were EAG-active in one or both of the vector species ( Fig . 3 , Panel A and Fig . 4 , Panel A ) . Differences were noted in the types of compounds attractive to each species . Several short chain carboxylic acids were stimulatory to S . damnosum s . l . , ( Fig . 3 , Panel A ) while short chain aliphatic alcohols proved to be stimulatory to S . ochraceum s . l . ( Fig . 4 , Panel A ) . Three compounds were found to be neurostimulatory for both species ( acetophenone , hexanal and cis-3-hexen-1-ol ) . The EAG experiments identified a suite of compounds that were neurostimulatory to both vector species , but could not evaluate what type of behavioral response , if any , these stimulatory compounds might elicit . To answer this question , the neurostimulatory compounds were evaluated in a series of choice experiments using a Y-tube olfactometer , to identify which compounds might promote host-seeking behavior in nature [21] . Hexane solutions of octanoic acid ( 1∶100; 1∶1 , 000 ) , decanal ( 1∶100 ) and acetophenone ( 1∶100 ) were found to be significantly attractive to S . damnosum s . l . ( Fig . 3 , Panel B ) . Other short chain carboxylic acids ( e . g . hexanoic acid , heptanoic acid and nonanoic acid ) were also found to elicit an attractive response , though these did not reach statistical significance ( Fig . 3 , Panel B ) . Seven compounds were attractive to S . ochraceum s . l . ( Fig . 4 , Panel B ) . These included all dilutions of 1-octen-3-ol , acetophenone ( 1∶100 ) , 3-octanol ( 1∶10 ) , nonanal ( 1∶100 ) , 1-octanol ( 1∶100 , 1∶1000 ) and hexanal ( 1∶100 , 1∶1 , 000 ) . Ammonium bicarbonate ( 1∶100 w/v ) , a solid compound not evaluated in the EAG screen for technical reasons , was also attractive to both species in the Y-tube assay ( Fig . 3 and 4 , Panels B ) . To verify that the compounds identified in the Y tube assay were also attractive in nature , slow-release baits of these compounds were prepared and field-tested as described in Materials and Methods . Beads prepared in this fashion resulted in a stable , continuous release of the odorant compounds; the release rate at 5 days post preparation remained high for almost all of the compounds , ranging from 67–95% of the initial amount released ( Fig . 5 ) . The only exception to this was hexanal , which decayed more rapidly than the other compounds , with only 28% of the original amount remaining at day 5 ( Fig . 5 ) . Pantyhose were loaded with the beads and ammonium bicarbonate and used as baits on optimized EWT platforms ( e . g . Fig . 1 , Panel G ) . Traps were baited with both CO2 and the attractant-impregnated beads , as previous studies had indicated that odor baits were not effective unless used in conjunction with CO2 [8] , [9] . The CO2 for the traps was generated from a solution of baker's yeast and sugar , as previously described [8] . Traps baited with CO2 and the aroma bead baits containing the compounds found to be attractive for S . ochraceum s . l . in the Y tube assays ( 1-octen-3-ol , 1-octanol , acetophenone , hexanal , and ammonium bicarbonate ) collected roughly twice the number of S . ochraceum s . l . in Mexico than did the traps baited with CO2 alone ( Fig . 6 , Panel A; p<0 . 001 ) . Similarly , traps baited with the compounds found to be attractive for S . damnosum s . l . in the Y tube assays ( hexanoic acid , heptanoic acid , octanoic acid , nonanoic acid , 1-decanal , acetophenone , and ammonium bicarbonate ) and CO2 collected roughly three times as many S . damnosum s . l . in Burkina Faso as did traps baited with CO2 alone ( Fig . 6 , Panel B; p<0 . 01 ) . Thus , the compounds identified as attractive in the EAG and binary choice behavioral assays were also attractive to the target species under field conditions .
Early field studies in Africa indicated that human sweat contains compounds that attract Simulium damnosum s . l . , the major African O . volvulus vector species group [22] , [23] . Unidentified compounds in soiled clothing were effective in promoting host-seeking by several sibling species in this species complex , and , of the chemical and physical variables involved in host attraction , human odor emanating from sweat was uniquely required . The studies reported here provide insight into the particular compounds responsible for this attractive response . We initially attempted to identify a suite of compounds that were common among three individuals , with the hypothesis that such common compounds would reflect a “common” human odor and thus allow us to establish a foundation for an attractive bait formulation . Surprisingly , we found a limited number of compounds that were shared among all three individuals . Just 29 compounds were found to be in common among the three individuals , of over 1200 compounds identified . This represents a smaller set of compounds than have been identified in other studies of human sweat components [24]–[26] . It is likely that many of the compounds that were not common among the individuals were derived from cosmetics and detergents , though the individuals enrolled in the study were asked to refrain from using deodorants and perfumed products during the time the collections were performed . It is also possible that some of the differences noted were due to differences in the skin microflora of the individuals , as bacterial metabolism of sweat has been previously shown to produce a number of attractive volatiles [16] , [17] . Finally , some of the differences might be due to individual differences in metabolism . For example , previous studies have suggested an association between HLA genotype and attractiveness to malaria vectors that has been linked to differences in the profile of sweat components of individuals with different HLA haplotypes [27] . Interestingly , 1-octen-3-ol , an acknowledged host-seeking stimulant for a number of hematophagous Diptera [28] , elicited a strong EAG response in both vector species , but was significantly attractive only to S . ochraceum in the Y-tube assay . 1-Octen-3-ol is a common component of human volatiles and sweat [29] , [30]; however , field studies have shown this compound to be marginally attractive for several zoophilic North American black fly species when used singly [31] , indicating this kairomone likely requires additional compounds such as CO2 to be optimally effective as a bait . The lack of attractiveness of 1-octen-3-ol to S . damnosum s . l . is therefore not unexpected , as members of this species group ( including S . sirbanum ) are mainly zoophilic and blood-feed on several host species other than humans , including cattle , donkeys , goats , sheep and dogs [32] , [33] . In contrast to S . damnosum s . l . , the cytotype of S . ochraceum s . l . found in Mexico ( cytotype A ) is highly anthropophilic and prefers humans to other animal hosts [14] . The two species groups are also quite distant both geographically ( Neotropic versus Afrotropic ) and phylogenetically ( subgenus Psilopelmia versus subgenus Edwardsellum ) and as a result have likely evolved host-seeking behaviors independently within environments that differ greatly in phenology and the range of host species available as blood sources . Simulium ochraceum ( cytotype A ) is restricted in its flight range to the montane regions of Mexico and Guatemala where onchocerciasis was formerly endemic . Simulium ochraceum s . l . also consists of far fewer sibling taxa ( n = 3 ) when compared to S . damnosum s . l . ( ≥50 ) indicating that gene flow resulting in sibling speciation has been less robust . Of the two species groups , S . ochraceum was stimulated and attracted by a larger , more chemically diverse number of human sweat components , suggesting that anthropophily as exemplified by this species may be associated with a relatively large number and variety of compounds in addition to CO2 , thereby allowing S . ochraceum to exploit humans routinely as hosts and transmit O . volvulus in the discrete Mesoamerica foci . Simulium damnosum s . l . females responded to several medium-chain carboxylic acids , particularly octanoic acid , a carboxylic acid that also evokes host-seeking behavior by other vector species in Africa . Carboxylic acids have been previously implicated as components contributing to human odor [24] , [25] and several are attractive to important mosquito vectors . Octanoic acid is EAG-stimulatory and attractive to Anopheles gambiae sensu stricto [30] , [34] and has been incorporated into a synthetic blend designed to attract various taxa in the An . gambiae species complex [11] , [12] . This compound is also EAG-stimulatory and a significant host kairomone for Culex quinquefasciatus , an important lymphatic filariasis vector in Africa and the Indian subcontinent [35] . Decanal , a kairomone for zoophilic S . damnosum s . l . , also elicits a strong EAG response in Cx . quinquefasciatus [36] and is attractive to several mosquito vectors of Rift Valley Fever virus [37] . Decanal , a component of human odor [38] , is also found on the skin of other hosts utilized by S . damnosum s . l . females , including cattle , goats and donkeys [30] , [37] . Acetophenone is less obvious as a human host kairomone , but was EAG stimulatory and elicited attraction in both species . This compound has been reported as electro-stimulatory or behaviorally-attractive to several disparate hematophagous Diptera seeking a plant sugar meal . Acetophenone is a component of some floral nectaries and is a potent attractant to Aedes aegypti and Cx . pipiens molestus [39] . Kwon et al . [40] also noted that acetophenone was stimulatory to the labellar sensillum ( S1 ) of An . gambiae . Nulliparous S . damnosum s . l . use plant sugars to fuel local host-seeking flight and sugar meals may also be essential for long-distance , migratory flight by savanna species in the S . damnosum s . l group . [41]; S . ochraceum s . l . females feed daily on plant sources and use floral carbohydrates as nutrients for flight and ovarian development [42] . However , because acetophenone also occurs as a volatile in human and bovine breath [43] , [44] as a result of food digestion , it would likely be attractive for host-seeking hematophagus insects as well . Acetophenone's status as a kairomone for both vector groups reinforces the concept of sensory parsimony , a phenomenon described only for several recently evolved blood-feeding Diptera . This process , i . e . , dissimilar behavioral patterns cued by a single chemical compound , has been reported for Stomoxys calcitrans [45] and Glossina spp . [46] but never for Simulium spp . For example , S . calcitrans may use acetophenone resulting from rumen digestion to locate hosts for blood-feeding while several species of tsetse – Glossina fuscipes , G . brevipalpis , G . pallidipes ( all obligate blood-feeders ) - responded to 4-ethyl acetophenone as a chemo-attractant to locate Lantana camara for shelter . Our report of acetophenone as stimulatory to the Simuliidae when compared to those for the higher Diptera thus dates the phenomenon of sensory parsimony occurring ≈150 million years earlier in evolutionary time among blood-feeding flies [47] . Ammonium bicarbonate was also attractive to both S . damnosum s . l . and S . ochraceum s . l . Ammonia , a major degradation product of this inorganic salt , is found in both human sweat and breath and has been shown to be attractive to a wide range of blood-sucking insects [48] . When combined with lactic acid , ammonia also serves as a synergist to attract Aedes aegypti , the Yellow Fever mosquito , to human hosts . In the current study , we concentrated upon compounds found to be in common among three different individuals , with the aim of identifying kairomones that were broadly attractive . The preliminary field trials demonstrated that the kairomones that we identified were attractive to both S . ochraceum s . l . and S . damnosum s . l . However , the number of flies captured using these compounds was relatively low , suggesting that the bait formulations will have to be optimized . Previous studies have demonstrated that the EWT , when baited with dirty clothing and CO2 collected numbers of vector flies that approached those obtained from human landing collections , suggesting that the appropriate mix of kairomones should be sufficient to allow the EWT's performance to be optimized to the point where it is a viable replacement for human landing collectors . Furthermore , it is well known that individuals differ in their attractiveness to hematophagous insects and that this difference can be traced to individual variation in odorant compounds [27] , [49] . Thus , it is possible that the baits might be made more effective through the identification of those compounds that make certain individuals particularly attractive to the black fly vectors of O . volvulus . Comparative analysis of the GC-MS patterns of host volatiles collected from more and less attractive individuals , coupled with the EAG and behavioral assays described above could be used to identify such highly attractive compounds . Once optimized bait formulations have been developed it will be necessary to conduct extensive field trials to demonstrate that the EWTs will be able to replace human landing collectors for surveillance of O . volvulus transmission and to develop models to relate the collections obtained from the traps to those obtained from human landing collectors , thereby permitting one to estimate annual biting rates from the trap collections . The EWT platforms have been designed to be inexpensive and simple to construct from materials that are commonly available in developing countries [8] , [9] . The use of plastic aroma beads as vehicles for the attractive compounds is in keeping with this design goal . Aroma beads are easily available on-line and from craft stores , and are quite inexpensive ( ca . $11 USD per kg ) . Thus , aroma bead based baits may be easily prepared locally by onchocerciasis elimination programs when needed , facilitating the adoption of the EWT trap as a surveillance toll to replace human landing collections in both Africa and Latin America . In addition to their use for surveillance and evaluation of the effectiveness of mass drug campaigns , traps baited with highly attractive lures could also be important in removing host-seeking segments of the adult black fly population in endemic onchocerciasis settings . Stochastic models indicate that the addition of vector control to community directed drug treatment with Mectizan ( ivermectin ) can have a profound effect on O . volvulus populations by changing the threshold biting rate required to maintain the parasite [50] . By using the compounds described here as baits in conjunction with CO2 , it is possible that the Esperanza Window Trap can not only increase the efficiency and safety of surveillance of O . volvulus transmission but may also assist in eliminating river blindness in Africa by reducing adult fly populations below parasite population transmission thresholds , particularly in hypo- and meso-endemic settings .
|
Human landing collections , which are the current standard for collecting the black fly vectors of Onchocerca volvulus , the causative agent of river blindness , are inefficient and pose certain ethical issues . As entomological methods are among the primary techniques recommended by the international community for verifying the elimination of onchocerciasis , there is a need to develop alternative methods to collect these vectors . Recent studies have demonstrated that traps baited with CO2 and dirty clothing have the potential to replace human landing collections for this purpose . However , for these traps to be widely applied , it will be necessary to develop a consistent bait formulation . To this end , volatile compounds from human sweat that attract the principal black fly vectors of O . volvulus in Africa and the Americas have been identified and used to optimize traps that specifically collect these insects . To achieve this milestone , we report the use of electroantennography and behavioral assays to identify human compounds that are neurostimulatory to these vectors , and demonstrate that these compounds are attractive to the vectors in field studies using previously developed trap platforms . The development of such a defined bait formulation will permit the widespread use of these traps by onchocerciasis elimination programs in Africa and the Americas .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"epidemiology",
"arthropod",
"vectors",
"parasitic",
"diseases",
"disease",
"vectors",
"nematode",
"infections"
] |
2015
|
Identification of Human Semiochemicals Attractive to the Major Vectors of Onchocerciasis
|
Metagenomic sequencing is becoming widespread in biomedical and environmental research , and the pace is increasing even more thanks to nanopore sequencing . With a rising number of samples and data per sample , the challenge of efficiently comparing results within a specimen and between specimens arises . Reagents , laboratory , and host related contaminants complicate such analysis . Contamination is particularly critical in low microbial biomass body sites and environments , where it can comprise most of a sample if not all . Recentrifuge implements a robust method for the removal of negative-control and crossover taxa from the rest of samples . With Recentrifuge , researchers can analyze results from taxonomic classifiers using interactive charts with emphasis on the confidence level of the classifications . In addition to contamination-subtracted samples , Recentrifuge provides shared and exclusive taxa per sample , thus enabling robust contamination removal and comparative analysis in environmental and clinical metagenomics . Regarding the first area , Recentrifuge’s novel approach has already demonstrated its benefits showing that microbiomes of Arctic and Antarctic solar panels display similar taxonomic profiles . In the clinical field , to confirm Recentrifuge’s ability to analyze complex metagenomes , we challenged it with data coming from a metagenomic investigation of RNA in plasma that suffered from critical contamination to the point of preventing any positive conclusion . Recentrifuge provided results that yielded new biological insight into the problem , supporting the growing evidence of a blood microbiota even in healthy individuals , mostly translocated from the gut , the oral cavity , and the genitourinary tract . We also developed a synthetic dataset carefully designed to rate the robust contamination removal algorithm , which demonstrated a significant improvement in specificity while retaining a high sensitivity even in the presence of cross-contaminants . Recentrifuge’s official website is www . recentrifuge . org . The data and source code are anonymously and freely available on GitHub and PyPI . The computing code is licensed under the AGPLv3 . The Recentrifuge Wiki is the most extensive and continually-updated source of documentation for Recentrifuge , covering installation , use cases , testing , and other useful topics .
In the case of low microbial biomass samples , there is very little native DNA from microbes; the library preparation and sequencing methods will return sequences whose principal source is contamination [5 , 6] . Sequencing of RNA requiring additional steps introduces still further biases and artifacts [7] , which in case of low microbial biomass studies translates into a severe problem of contamination and spurious taxa detection [8] . The clinical metagenomics community is stressing the importance of negative controls in metagenomics workflows and , recently , raised a fundamental concern about how to subtract the contaminants from the results [9] . From the data science perspective , this is just another instance of the importance of keeping a good signal-to-noise ratio [10] . When the signal ( inherent DNA/RNA , target of the sampling ) approaches the order of magnitude of the noise ( acquired DNA/RNA from contamination and artifacts ) , particular methods are required to tell them apart . The roots of contaminating sequences are diverse , as they can be traced back to nucleic acid extraction kits ( the kitome ) [11 , 12] , reagents and diluents [13 , 14] , the host [15] , and the post-sampling environment [16] , where contamination arises from different origins such as airborne particles , crossovers between current samples or DNA remains from past sequencing runs [17] . Variable amounts of DNA from these sources are sequenced simultaneously with native microbial DNA , which could lead to severe bias in magnitudes like abundance and coverage , particularly in low microbial biomass situations [18] . If multiplex sequencing uses simple-indexing , false assignments could be easily beyond acceptable rates [19] . Even the metagenomic reference databases have a non-negligible amount of cross-contamination [15 , 17 , 20] . Regarding the kitome , it varies even within different lots of the same products . For example , the DNeasy PowerSoil Kit ( formerly known as PowerSoil DNA Isolation Kit ) , a product that usually provides significant amounts of DNA and has been widely used , including Earth Microbiome Project and Human Microbiome Project , often yields a background contamination by no means negligible [6] . The lower the biomass in the samples , the more essential it is to collect negative control samples to help in the contamination background assessment because , without them , it would be almost impossible to distinguish inherent microbiota in a specimen —signal— from contamination —noise— . Assuming that the native and contaminating DNA are accurately separated , the problem of performing a reliable comparison between samples remains . In general , the taxonomic classification engine assigns the reads from a sequencing run to different taxonomic ranks , especially if the method uses a more conservative approach like the lowest common ancestor ( LCA ) [21] . While LCA drastically reduces the risk of false positives , it usually spreads the taxonomic level of the classifications from the more specific to the more general . Even if the taxonomic classifier does not use the LCA strategy , each read is usually assigned a particular score or confidence level , which should be taken into account by any downstream application as a reliability estimator of the classification . On top of these difficulties , it is still more challenging to compare samples with very different DNA yields , for instance , low microbial biomass samples versus high biomass ones , because of the different resolution in the taxonomic levels . This sort of problem also arises when the samples , even with DNA yields in the same order of magnitude , have an entirely different microbial structure so that the minority and majority microbes are fundamentally different between them [18] . Finally , a closely related problem emerges in metagenomic bioforensic studies and environmental surveillance , where it is essential to have a method prepared to detect the slightest presence of a particular taxon [3 , 22 , 23] and provide quantitative results with both precision and accuracy . From the beginning , the application of SMS to environmental samples supplied biologists with an insight of microbial communities not obtainable from the sequencing of Bacterial Artificial Chromosome ( BAC ) clones or 16S rRNA [24 , 25] . The scientific community soon underlined the need and challenges of comparative metagenomics [26 , 27] . MEGAN [28] , one of the first metagenomic data analysis tools , provided in its initial release a very basic comparison of samples , which has improved with an interactive approach in more recent versions [29] . In general , metagenomic classification and assembly software is more intra- than inter-sample oriented [30] . Several tools have tried to fill this gap , starting with CoMet [31] , a web-based tool for comparative functional profiling that combines different methods such as multi-dimensional scaling and hierarchical clustering analysis to predict functional differences in a collection of metagenomic samples . Soon after , a different approach appeared with the discovery of the crAssphage thanks to the crAss software [32] , which provides reference-independent comparative metagenomics using cross-assembly . The following year , Community-analyzer was released , a tool for visually comparing microbial community structure across microbiomes using correlation-based graphs to infer differences in the samples and predict microbial interactions [33] . In 2014 , yet another alternative came , COMMET [34] , a piece of software that goes a step further by enabling the combination of numerous metagenomic datasets through a scalable method based on efficient indexing . Two years later , a parallel computation method called Simka was published [35] , which performs de novo comparative metagenomics by counting k-mers concurrently in multiple datasets . In 2015 , a highly publicized report on the metagenomics of the New York subway suggested that the plague and anthrax pathogens were part of the normal subway microbiome . Soon afterward , several critics arose [36] and , later , reanalysis of the New York subway data with appropriate methods did not detect the pathogens [37] . As a consequence of this and other similar problems involving metagenomic studies , a work directed by Rob Knight [38] emphasized the importance of validation in metagenomic results and issued a tool based on BLAST ( Platypus Conquistador ) . This software confirms the presence or absence of a taxon of interest within SMS datasets by relying on two reference sequence databases: one for inclusions , with the sequences of interest , and the other for exclusions , with any known sequence background . Another BLAST-based method for validating the assignments made by less precise sequence classification programs has been recently announced [22] . The approach of Recentrifuge to increased confidence in the results of taxonomic classification engines follows a dual strategy . Firstly , it accounts for the score level of the classifications in every single step . Secondly , it uses a robust contamination removal algorithm that detects and selectively eliminates various types of contaminants , including crossovers . Recentrifuge directly supports the following high-performance taxonomic classifiers: Centrifuge [7] , LMAT [21] , CLARK [39] , CLARK-S [40] , and Kraken [41] . Other classification software is supported through a generic parser . The interactive interface of Recentrifuge enables researchers to analyze the results of those taxonomic classifiers using scored Krona-like charts [42] . In addition to the plots for the raw samples , Recentrifuge generates four different sets of scored charts for each taxonomic level of interest: control-subtracted samples , shared taxa ( with and without subtracting the controls ) , and exclusive taxa per sample . This battery of analysis and plots permits robust comparative analysis of multiple samples in metagenomic studies , especially useful in case of low microbial biomass environments or body sites . Recentrifuge enables robust contamination removal and score-oriented comparative analysis of multiple samples , especially in low microbial biomass metagenomic studies , where contamination removal is a must . Just as it is essential to accompany any physical measurement by a statement of the associated uncertainty , it is desirable to attend any read classification with a confidence estimation of the assigned taxon . Recentrifuge reads the score given by a taxonomic classification software to the reads and uses this valuable information to calculate an average confidence level for each taxon in the taxonomic tree associated with the sample analyzed . This value may also be a function of further parameters , such as read quality or length , which is especially useful in case of significant variations in the length of the reads , like in the datasets generated by nanopore sequencers . Only a few codes , such as Krona [42] and MetaTreeMap [43] , are hitherto able to handle a score assigned to the classification nodes . In Recentrifuge , the calculated score propagates to all the downstream analysis and comparisons , including the interface , an interactive framework for a straightforward assessment of the validity of the taxonomic assignments . That is an essential advantage of Recentrifuge over other metagenomic dataset analysis tools .
For each sample , according to the NCBI Taxonomy [44] , Recentrifuge populates a logical taxonomic tree , with the leaves usually belonging to the lower taxonomic levels like species , variety or form . The methods involving trees were implemented as recursive functions for compactness and robustness , making the code less error-prone . One of such methods is essential for understanding the way Recentrifuge prepares samples before any comparison or operation such as control subtraction . It recursively ‘folds the tree’ for any sample if the number of assigned reads to a taxon is under the mintaxa setting ( minimum reads assigned to a taxon to exist in its own ) , or because the taxonomic level of interest is over the assigned taxid ( taxonomic identifier ) . See Fig 1A for a working example of the method in action for two samples . The same procedure applies to the trees of every sample in the dataset . This method does not just ‘prune the tree’ , on the contrary , it accumulates the counts ni of a taxon in the parent ones np and recalculates the parent score σp as a weighted average taking into account the counts and score of both . In general , the new score of parent taxa , σ p ′ is calculated as follows: σ p ′ = 1 n p + ∑ i D n i ( σ p n p + ∑ i D σ i n i ) ∀ ( σ i , n i ) where 0 < ni < mintaxa and D is the number of descendant taxa that are to be accumulated in the parent one and σi their respective scores . This is done recursively until the desired conditions are met . This method is applied , at a given taxonomic level , to the trees of every sample before being compared in search for the shared and exclusive taxa . For a sample , the mintaxa parameter defaults to the nearest integer of the decimal logarithm of the number of reads passing the minimum score threshold ( minscore ) filter , thus growing with the order of magnitude of the effective size of the sample . However , the user can modify such automatic value for mintaxa and set it independently for control and real samples . In addition to the input samples , Recentrifuge includes some sets of derived samples in its output . After parallel calculations for each taxonomic level of interest , it adds hierarchical pie plots for CTRL ( control subtracted ) , but also for EXCLUSIVE , SHARED and SHARED_CONTROL samples , defined below . Let T mean the set of taxids in the NCBI Taxonomy and Ts the collection of taxids present in a sample s . If Rs stands for the set of reads of a sample s and Cs for the group of them classifiable , then the taxonomic classification c is a function from Cs to T , i . e . , C s → c T , where Cs ⊆ Rs and c [ C s ] = T s ⊆ T . The set L of the 32 − 1 different taxonomic levels used in the NCBI Taxonomy ( see S5 Fig ) [44] is ordered in accordance with the taxonomy , so ( L , < ) is a strictly ordered set , since form < variety < subspecies < ⋯ < domain . Then , T s = T s form ∪ ⋯ ∪ T s domain = ∪ L T s l , where T s l represents the collection of taxa belonging to a sample s for a particular taxonomic rank or level l . Related with this , we can write as T s → l the taxa of the sample s for a taxonomic level l once we have applied the ‘tree folding’ to such level l detailed in the previous subsection ( and in Fig 1A ) . For a taxonomic rank k of interest , in a series of S samples where there are N < S negative controls , Recentrifuge computes the sets of taxa in the derived samples CTRL ( CTRL T s k ) , EXCLUSIVE ( EXCL T s k ) , SHARED ( SHARED Tk ) and SHARED_CONTROL ( SHARED_CTRL Tk ) as: CTRL T s k = T s → k \ ∪ n N T n → k EXCL T s k = T s → k \ ∪ m ≠ s S T m → k SHARED T k = ∩ m S T m → k SHARED _ CTRL T k = ∩ m > N S T m → k \ ∪ n N T n → k Please see Fig 1B for examples . Finally , Recentrifuge generates in parallel a set of SUMMARY samples condensing the results for all the taxonomic levels of interest . For a taxonomic rank k , after the ‘tree folding’ procedure detailed above , the contamination removal algorithm retrieves the set of candidates T ¯ s → k to contaminant taxa from the N < S control samples . Depending on the relative frequency ( fi = ni/∑i ni ) of these taxa in the control samples and if they are also present in other specimens , the algorithm classifies them in contamination level groups: critical , severe , mild , and other . Except for the latter group , the contaminants are removed from non-control samples . Then , Recentrifuge checks any taxon in the ‘other contaminants’ group for crossover contamination so that it eliminates any taxon marked as a crossover from every sample except the one or ones selected as the source of the pollution . In detail , the algorithm removes any taxon t s k ∈ T ¯ s → k from a non-control sample unless it passes the robust crossover check: a statistical test screening for overall outliers and an order of magnitude test against the control samples . See Fig 2 for an example of this procedure . The robust crossover tests are defined as follows: Outliers statistic test ( t s k ) : f t s k > median { f t 1 k , … , f t S k } + δ Q n Order of magnitude test ( t s k ) : f t s k > 10 ξ max { f t 1 k , f t 2 k , … , f t N k } where Qn [45] is a scale estimator to be discussed below , and δ and ξ are constant parameters of the robust contamination removal algorithm . The parameter δ is an outliers cutoff factor , while ξ is setting the difference in order of magnitude between the relative frequency of the candidate to crossover contaminator in the sample s and the greatest of such values among the control samples . In Recentrifuge , δ typically ranges from 3 to 5 , and ξ from 2 to 3 . Qn is the chosen scale estimator for screening the data for outliers because of its remarkably general robustness and other advantages compared to other estimators [45 , 46] , like the MAD ( median absolute deviation ) or the k-step M-estimators . It has a 50% breakpoint point , a smooth influence function , very high asymptotic efficiency at Gaussian distributions and is suitable for asymmetric distributions , which is our case , all at a reasonable computational complexity , as low as O ( n ) for space and O ( n log n ) for time . So , here: Q n = d { | f t i k - f t j k | i < j ≤ S } ( m ) : m = ( S 2 + 1 2 ) = Γ ( S 2 + 2 ) 2 Γ ( S 2 ) = S 4 ( S 2 + 1 ) where d = 3 . 4760 is a constant selected for asymmetric non-gaussian models similar to the negative exponential distribution , m refers to the mth order statistics of the pairwise distances and Γ is the Gamma function . Recentrifuge is a metagenomics analysis software with two different main parts: the computing kernel , implemented and parallelized from scratch using Python , and the interactive interface , based on interactive hierarchical pie charts by extending the Krona [42] 2 . 0 JavaScript library developed at the Battelle National Biodefense Institute . Recentrifuge’s novel approach combines robust statistics , arithmetic of scored taxonomic trees , and concurrent computational algorithms to achieve its goals . Fig 3 is a flow diagram of Recentrifuge that clearly shows three parallel regions in the code . In each of them , the work divides into concurrent processes attending to different variables: control and regular samples in the first region , the taxonomic ranks in the second , and the specimen along with the type of analysis in the last parallel region , which summarizes the results . In any SMS study with related samples , including negative controls , Recentrifuge generates four additional sets of scored charts: the samples with the contamination subtracted , the exclusive taxa per sample , and the shared taxa with and without control taxa subtracted ( see S4 Fig ) . Fig 4 summarizes the package context and data flows . Recentrifuge straightforwardly accepts output files from various taxonomic classifiers , thus enabling a scored-oriented taxonomic visualization for metagenomics . Recentrifuge directly supports output from Centrifuge [7] , LMAT [21] , CLARK [39] , CLARK-S [40] , and Kraken [41] . Alternative taxonomic classifiers are supported through a generic interface developed to handle different file formats with comma-separated values ( CSV ) , tab-separated values ( TSV ) , or space-separated values ( SSV ) . The software also includes support for LMAT plasmid assignment system [15] . For implementation details of the Recentrifuge computing kernel please see S1 Appendix , S5 and S6 Figs . To ensure the broadest portability for the interactive visualization of the results , the central outcome of Recentrifuge is a stand-alone HTML file which can be loaded by any JavaScript-enabled browser . Fig 5 shows a labeled screenshot of the corresponding Recentrifuge web interface for an example of SMS study ( see S1 Fig ) . A vectorial screenshot in SVG format with the original font scheme is available for any sample using the “Screenshot” button of the user interface . The package also provides comprehensive statistics about the reads and their classification performance . Another Recentrifuge output is a spreadsheet collection detailing all the classification results in a compact way . This format is adequate for further data mining on the data , for example , as input for applications such as longitudinal ( time or space ) series analyzers like Dynomics ( in development ) . Besides , the user can choose between different score visualization algorithms , some of which are more interesting for datasets containing variable length reads , for example , the ones generated by Oxford Nanopore sequencers . Finally , some filters are available , like the minimum score threshold ( minscore ) , which can be set independently for the control and real samples . The minscore filter can be used to generate different output sets from a single run of the classifier with a low minimum hit length ( MHL ) setting , saving computational resources . Other filters are mintaxa , described in the scored taxonomic trees subsection , and the lists of identifiers to exclude or include a taxon and all its children in the taxonomic tree . The additional tools in the Recentrifuge package ( see Fig 4 ) can generate further products and results . Rextract is a script which helps in extracting a subset of classified reads of interest from the single or paired-ends FASTQ input files . This set of reads can be used in any downstream application , such as genome visualization and assembling . Remock is a script for easily creating mock Centrifuge samples , which is useful not only for testing and validation purposes but also for introducing a list of previously known contaminants to be taken into account by the robust contamination removal algorithm . Retest is the code used for continuous integration ( CI ) testing and algorithm verification procedures ( see Section 2 of S4 Appendix for further details and S10 Fig for its flowchart ) .
We developed a synthetic dataset carefully designed to challenge the Recentrifuge algorithms ( see S13 Fig and Section 2 . 3 of S4 Appendix for details ) , thus enabling a quantitative assessment of the capability of the method to cope with different kinds of contaminants . We also devised this mock dataset in order to evaluate the ability of the method to deal with cross-contamination between samples . This feature of Recentrifuge is one of the advantages of this novel approach . In addition , this synthetic dataset serves the purpose of the continuous integration framework of the software , as the results of processing these data are compared with a standard to check the reliability of the method after any change in the source code . Fig 6 shows a comparison of abundances of taxa included in the synthetic dataset before and after the Recentrifuge robust contamination removal algorithm . The taxa belong to species or below in the NCBI taxonomy . The left column of the figure shows the abundance histogram for seven raw samples: four real samples ( smpl1 to smpl4 ) plus three negative control samples ( ctrl1 to ctrl3 ) . Similarly , the right column shows the results after the algorithm intervention for the species taxonomic level , i . e . , the corresponding CTRL_species samples ( see ‘Derived samples’ subsection in Design and implementation ) . Native taxa are green-colored , crossover contaminants are colored in purple , and other colors indicate different classes of contaminants . The legend of S13 Fig details the complete color code . We see in Fig 6 that Recentrifuge cleared the CTRL_species samples of the different contaminants ( species and below ) found in the negative control samples while retaining the particular native taxa , which accumulated up to the species level ( see ‘Scored taxonomic trees’ subsection in Design and implementation for details ) . Examples of important contaminants removed were human reads and those belonging to Cutibacterium acnes . The algorithm also deleted more subtle contamination , such as the reads assigned to Malassezia globosa . Crossover contamination requires special mention . On the one hand , Methanosarcina mazei was ubiquitous among the samples , but it was only native to smpl1 and a contaminant in the rest . On the other hand , M . barkeri was present in the four real samples despite being only native to smpl3 , but it was scarce in the control samples , even missing from ctrl2 . Recentrifuge accurately detected which were the source sample of both Methanosarcina species , thus keeping the native reads there and clearing the cross-contamination from the rest of the samples . Furthermore , we included an additional sample ( smplH ) in the synthetic dataset containing the 241 species of a high-complexity dataset used as a gold standard for benchmarking metagenomic software [47] . As with the other samples , this specimen combined contaminants as additional taxa . In addition , we spiked the controls with low abundances of native taxa from this and the other real samples in order to simulate statistical noise in negative control samples such as low-frequency misclassifications and sequencing errors . We used the complete synthetic dataset to obtain different ROC ( receiver operating characteristic ) plots . S11 Fig shows the evolution of the sensitivity and specificity from the raw specimens to the CTRL_species samples . Basically , this ROC presented a transition from a scenario of very low specificity , on account of the contamination misidentified as native taxa , to a situation characterized by very high specificity , thanks to the correct detection of contaminants , including crossovers . For some samples , this came at the expense of a slight loss in the sensitivity . The reason for that small decline in the recall rate was the intentional introduction in the synthetic dataset of the archaea Methanobacterium formicicum with two different strains , one native to the samples ( M . formicicum DSM 3637 ) and another a contaminant ( M . formicicum JCM 10132 ) . At the species level , once the cross-contamination situation was ruled out , Recentrifuge followed a conservative strategy and deemed the archaeal species as a contaminant and , therefore , the native strain of M . formicicum became a false negative thus decreasing the sensitivity . For samples smpl1 to smpl4 and smplH , S12 Fig shows the ROC as a function of the mintaxa parameter . Results of Fig 6 , S11 and S12 Figs can be easily replicated using retest ( see Section 2 of S4 Appendix ) . To confirm Recentrifuge’s ability to analyze complex metagenomes and provide new biological insight , we considered an ambitious but severely contaminated SMS study of RNA in plasma from individuals with Myalgic Encephalomyelitis/Chronic Fatigue Syndrome ( ME/CFS ) , alternatively diagnosed chronic Lyme syndrome ( ADCLS ) , and systemic Lupus erythematosus ( SLE ) [48] . This research suffered from large batch and contamination effects and was unable to find a positive association between the plasma microbial content of sick individuals , thus highlighting the relevance of technical controls in metagenomics . More than 240 giga-base-pairs of raw genomic data distributed in 67 samples with paired-ends sequences were downloaded and analyzed using Bowtie2 [49] , Centrifuge [7] , and SAMtools [50] ( see S2 Appendix for the procedure details ) . Recentrifuge analyzed the different datasets in this study using stricter parameters than the default ones: sequencing of RNA required extra steps than sequencing of DNA , including the reverse transcription of RNA and further purifications [48] , which were additional sources of artifacts and contamination . In this case , an increase in the matching length to 60 was advisable [7] , so Recentrifuge filtered the Centrifuge output with minscore raised to an even stricter value of 75 ( unless otherwise indicated ) . To further illustrate the difficulty of the dataset of the SMS study of plasma in ME/CFS patients regarding the contamination , just a couple of results . First , affecting the sequencing batch one , Recentrifuge detected crossover contamination in the negative control samples with the source in the positive control , consisting of human metapneumovirus ( hMPV ) . Second , Recentrifuge reported quite more different taxa in the negative controls than in the normal samples: 65% and 22% more on average , respectively , for the batch two and three . The presence of generalized crossover contamination complicates the removal of the contaminants in the samples by merely excluding the taxa present in the controls . Here it is when the robust contamination removal algorithm of Recentrifuge is of great help: it detects the crossover contaminants ( hMPV and other taxa ) and removes them from all the samples except for the inferred source . Therefore , the positive control is still positive for hMPV after the contamination removal , as expected ( see S7 Fig ) . The Recentrifuge analysis of the entire collection of 67 samples revealed the presence of ubiquitous contaminants able to spread over different sequencing batches and type of samples ( see S8 Fig ) . Most of the contaminating bacteria are known contaminants belonging to the kitome [11] . Other pervasive pollutants belong to the fungi orders Eurotiales , Helotiales , Hypocreales , Pleosporales , and Saccharomycetales . The contamination by Apicomplexa , in general , and Plasmodium vivax and Besnoitia , in particular , can be linked to database contamination [15 , 20] and seems a negative hallmark of SMS RNA studies related to body fluids [8] . An interesting complementary analysis consisted of retrieving those taxa that are contaminating the negative control samples exclusively . S9 Fig shows the genera contaminating all the control samples but no other specimen along the second batch , representing contaminants which entered the workflow in some procedure or material exclusive to the control samples . In concordance with the main conclusion of the study of plasma in individuals with ME/CFS [48] , Recentrifuge did not find shared taxa after control removal ( CONTROL_SHARED empty ) when analyzing the samples rearranged in different batches and pathology/healthy groups . Nevertheless , the individual analysis of the samples after contamination removal presents interesting features in a case-per-case review . That is the case of sample 56 in Fig 7 , which belongs to an ADCLS patient . It shows a collection of taxa with a high average score ( 114 ) in the classification , implying that a majority of sequences mapped in both reads from the pair , except for the contaminant genus Besnoitia , the lowest-scored one . This set of microbes seems compatible with bacteria translocated from the buccopharyngeal cavity into blood , apparently because of an oral chronic inflammatory polymicrobial disease . However , the clinically relevant taxa in this study go far beyond those of sample 56 shown in Fig 7 . S3 Appendix portrays other representative bacteria , viruses , and fungi , present in the samples . Research in recent years is overturning the commonly accepted paradigm which stated that , in healthy individuals , the tissues and body fluids not in contact with the environment are sterile . Healthy organs once thought to be free of microbes are crawling with bacteria , archaea , viruses , and eukaryotes . The shift of paradigm has spread to more and more tissues and fluids , like the deepest layers of the skin [51] , the placenta [52] , the urine [53] , the blood [54 , 55] , the breast milk , or others [54 , 56 , 57] . The plasma is the part of the blood with the lower proportion of bacterial DNA , only 0 . 03% [55] . In the reanalyzed study of plasma in individuals with ME/CFS [48] , the intrinsic difficulties of ultra-low microbial biomass joined the handicap of an RNA sequencing technique prone to further artifacts and biases , which resulted in severe widespread contamination . With the results of the research , the classical paradigm might seem supported , that is , the idea of the absence of a plasma microbiota in healthy individuals . However , the authors of the study believed that the limitation of the current techniques prevented them from revealing the microbial component in human plasma . Indeed , with the noise in the same order of magnitude of the signal , a robust method for contamination removal was required to tackle this complex dataset . Despite all the difficulties , the analysis with Recentrifuge has unveiled a meaningful plasma microbiota in the samples ( Fig 7 and S3 Appendix ) . The results are in line with the recent research in the field , which points out the gut , the oral cavity , and the genitourinary tract as the primary sources of the blood microbiome [55 , 58 , 59] . In conclusion , thanks to the robust contamination removal and the score-oriented comparative analysis of multiple samples in metagenomics , Recentrifuge can play a key role , firstly , in the study of oligotrophic microbes in environmental samples , as it did by showing that microbiomes of Arctic and Antartic solar panels display similar taxonomic profiles [60]; secondly , in the more reliable detection of minority organisms in clinical or forensic samples . The relevant organisms found with a high score in the SMS study of plasma in ME/CFS patients [48] after the robust contamination removal are good examples . Finally , the mock dataset confirmed the worthiness of the developed methods , which demonstrated a radical improvement in specificity while retaining high sensitivity rates even in the presence of cross-contaminants .
Recentrifuge’s main website is www . recentrifuge . org . The data and source code are anonymously and freely available on GitHub at https://github . com/khyox/recentrifuge and PyPI at https://pypi . org/project/recentrifuge . The Recentrifuge computing code is licensed under the GNU Affero General Public License Version 3 ( www . gnu . org/licenses/agpl . html ) . Recentrifuge’s continuous integration ( CI ) information is public on Travis CI at https://travis-ci . org/khyox/recentrifuge . The wiki ( https://github . com/khyox/recentrifuge/wiki ) is the most extensive and updated source of documentation for Recentrifuge , including installation , testing , quick-start , and comprehensive use cases for the different taxonomic classification engines supported . In addition , Recentrifuge’s installation is explained in Section 1 of S4 Appendix , testing is detailed in Section 2 of S4 Appendix , and running Recentrifuge for Centrifuge , LMAT , CLARK flavors , Kraken , and other taxonomic classifiers are subsections of Section 3 of S4 Appendix . Similarly , Sections 4 and 5 of S4 Appendix describe running Rextract and the Recentrifuge command line , respectively . Finally , Section 6 of S4 Appendix includes troubleshooting subsections . The full Centrifuge output and the detailed Recentrifuge results for the SMS study of plasma in individuals with ME/CFS are publicly available at som1 . uv . es/plasmaCFS . Just as the biochemical profile and cell count are currently usual blood tests , metagenomic analysis of the blood will probably become a standard in a few years . The methods that will pave the way for a well-established clinical practice of metagenomics are still to come . As an open-source project , the participation of the computational biology and clinical metagenomics community will determine the future of Recentrifuge considerably . An important extension to Recentrifuge is under active development and will be released soon . It is “Regentrifuge” , the counterpart of Recentrifuge in the area of metagenomic functional analysis .
|
Whether in a clinical or environmental sample , metagenomics can reveal what microorganisms exist and what they do . It is indeed a powerful tool for the study of microbial communities which requires equally powerful methods of analysis . Current challenges in the analysis of metagenomic data include the comparative study of samples , the degree of uncertainty in the results , and the removal of contamination . The scarcer the microbes are in an environment , the more essential it is to have solutions to these issues . Examples of sites with few microbes are not only habitats with low levels of nutrients , but also many body tissues and fluids . Recentrifuge’s novel approach combines statistical , mathematical and computational methods to tackle those challenges with efficiency and robustness: it seamlessly removes diverse contamination , provides a confidence level for every result , and unveils the generalities and specificities in the metagenomic samples .
|
[
"Abstract",
"Introduction",
"Design",
"and",
"implementation",
"Results",
"and",
"discussion",
"Availability",
"and",
"future",
"directions"
] |
[] |
2019
|
Recentrifuge: Robust comparative analysis and contamination removal for metagenomics
|
A description of many biological processes requires knowledge of the 3-D structure of proteins and , in particular , the defined active site responsible for biological function . Many proteins , the genes of which have been identified as the result of human genome sequencing , and which were synthesized experimentally , await identification of their biological activity . Currently used methods do not always yield satisfactory results , and new algorithms need to be developed to recognize the localization of active sites in proteins . This paper describes a computational model that can be used to identify potential areas that are able to interact with other molecules ( ligands , substrates , inhibitors , etc . ) . The model for active site recognition is based on the analysis of hydrophobicity distribution in protein molecules . It is shown , based on the analyses of proteins with known biological activity and of proteins of unknown function , that the region of significantly irregular hydrophobicity distribution in proteins appears to be function related .
Because of the growing number of structural genomics projects oriented toward obtaining a large number of protein structures in rapid and automated processes [1–4] , there is a need to predict protein function ( or its functionally important residues ) by examining its structure . There have been a variety of efforts in this direction . Some of the techniques used to identify functionally important residues from sequence or structure are based on searching for homologue proteins of known functions [5–8] . However , homologues , particularly when the sequence identity is below 25% , need not have related activities [9–11] . Geometry-based methods have shown that the location of active site residues can be identified by searching for cavities in the protein structure [12] or by docking small molecules onto the structure [13] . The cave localization in silico has been presented on the basis of the characteristics of the normal created for each surface piece [14] . The complex analysis of protein interfaces and their characteristics versus highly divergent areas is presented by Jimenez [15] . Several experimental studies have shown that mutation of residues involved in forming interfaces with other proteins or ligands can also be replaced to produce more stable , but inactive proteins [16–19] . On this basis , several effective algorithms were developed [20 , 21] . Finally , structural analysis coupled with measures of surface hydrophobicity have been used to identify sites on the surfaces of proteins involved in protein–protein interactions [22] . The Fuzzy Oil Drop ( FOD ) model presented in this paper is based on an external hydrophobic force field [23–27] . The role of hydrophobic interactions in protein folding [28–31] as well as in protein structure stabilization [32–36] has been known since the classic oil drop model of representing the hydrophobic core in proteins was introduced by Kauzmann [37] . According to this model , the hydrophobic residues tend to be placed in the central part of the protein molecule and in hydrophilic residues on the protein's surface [38–40] . Even the recognition of native versus nonnative protein structures can be to some extent differentiated on the basis of spatial distribution of amino acid hydrophobicity [41–43] . The importance of hydrophobicity distribution has been emphasized , particularly for Rosetta development , when the description of the hydrophobic core significantly increased the performance of the Rosetta program [44] . The discrete system of ellipsoidal centroids was introduced to estimate the concentration of hydrophobic residues , in particular protein zones [44] . The nonrandom hydrophobicity distribution has been proven by Irbäck et al . [45] . However , it was suggested that the core region is not well described by a spheroid of buried residues surrounded by surface residues due to hydrophobic channels that permeate the molecule [46 , 47] . The FOD model was initially used to simulate the hydrophobic collapse of partially folded proteins . Those structural forms were assumed to represent the early stages of folding ( in silico ) ; that model is presented elsewhere [48–50] . The comparison of structures received by folding simulations with their native forms revealed , however , some unexpected results . In the case of native structures , the idealized hydrophobicity distribution satisfying the oil drop–like hydrophobicity partitioning compared with the empirically observed hydrophobicity differs in a specific manner . The high discrepancies between observed and theoretical hydrophobicities within FOD are observed in the area of the binding site [23–26] . It can even be generalized that the location of hydrophobicity differences seems to represent an aim-oriented discrepancy . This simple observation gave us the opportunity to develop a method that was able to recognize functional sites or residues in a protein structure . In this study , the FOD model is applied to 33 proteins of known function and 33 proteins of unknown function that resulted from structural genomics projects .
The 33 proteins of known biological activity ( Table 1 ) were selected to verify the reliability of the method . Most of these proteins are enzymes that have well-defined biological function and are deposited in the Catalytic Site Atlas ( http://www . ebi . ac . uk/thornton-srv/databases/CSA ) , a database of templates representing different catalytic mechanisms [51] . The residues identified in this database as active site were used as the criteria to verify the results . Two proteins of known function—rat annexin V , and ButF , the vitamin B12-binding protein , which take part in regulation [52] and transport processes [53] , respectively—are also included in the test probe . Reports from structural genomics projects [1–4] have described the solution of a number of proteins with unknown functions . The procedure for potential functional site recognition presented in this paper was performed with a set of 33 such proteins deposited in the Protein Data Bank ( PDB ) ( Table 2 ) . The multimeric proteins were represented solely by their first chain in the PDB file . All molecular visualizations were created with Pymol software [54] . The FOD hydrophobic force field is based on the assumption that the theoretical hydrophobicity distribution in proteins is represented by the 3-D Gaussian function . The value of this function in a particular j-th point within the space occupied by a protein represents the hydrophobicity density at this point: Where is the theoretical ( expected ) hydrophobicity of the j-th point , σx , σy , σz are the standard deviations , which depend on the length of polypeptide under consideration [23–26] and the point is localized in the center of coordinate system ( 0 , 0 , 0 ) of the highest theoretical hydrophobicity . This simplifies Equation 1: The molecule is oriented according to the following procedure: the longest distance between two effective atoms determines the z-axis , and the longest distance between projections on the x–y plane determines the x-axis . For this orientation of molecules in the coordinate system , the values of σx , σy , σz parameters are calculated as one-third of the highest x , y , or z coordinates of the effective atom increased by 9 Å ( cutoff distance for hydrophobic interaction ) in each direction . The values of the Gaussian function are standardized to give a value of 1 . 0 . The second component of this force field is an observed ( empirical ) hydrophobicity distribution formed by the side chains of a protein molecule , and can be expressed using the original function introduced by Levitt [55] . The j-th point collects hydrophobicity as follows: where denotes the empirical hydrophobicity value characteristic for the j-th point , N is the number of residues in a protein , represents the hydrophobicity characteristic for the i-th amino acid , rij is the distance between the j-th point and the geometrical center of the i-th residue , and c expresses the cutoff distance , which has a fixed value of 9 . 0 Å , following the original paper [55] . The observed hydrophobicity distribution is also standardized . More details concerning the FOD force field are given in recently published papers [23–27] . The similarity of the FOD-based hydrophobic scale with others commonly used for calculations ( e . g . , the Eisenberg [56] or Doolittle [40] scales ) has been shown and discussed in [57] . The differences between these scales seem to be negligible with respect to the problem under consideration . Use of these scales does not change the distribution significantly ( Equation 3 ) [57] . The introduction of the FOD-based hydrophobic scale unifies the system for proteins ( amino acids ) and molecules interacting with proteins , creating stable complexes ( ligands ) . The residues annotated in CSA as those playing roles in catalytic activity were used as the gold standard to verify the reliability of the results received according to the FOD model . To indicate the most meaningful amino acids considered by the FOD model to be located in the functional site , the calculation of percentiles was used to identify the threshold for selection of maxima , which are distinguished as belonging to the functional site . It is possible to do so , because the quantitative results expressing the level of can be taken as the criteria for discrimination . For a set of measurements arranged in order of magnitude , the p-th percentile is the value that has p percent of the measurements below it and ( 100 − p ) percent above it . In this analysis , the 95th percentile was used . In other words , among the analyzed data , 95% of values were below the 95th percentile threshold , and only the 5% above the threshold was taken into consideration . The same validation method cannot be used in the SuMo or ProFunc methods because of their different types of output data . They produce only the numbers of amino acids that potentially belong to functional sites and total scores ( based on which given set of amino acid residues is assessed and what functional site is proposed ) . This is why the percentage of commonly classified residues was calculated for each protein molecule by taking the best hit by ProFunc ( according to the score value ) and the solution most relevant to the FOD-based results by SuMo .
The proteins of known biological activity ( Table 1 ) and protein structures of unknown function that resulted from structural genomics projects ( Table 2 ) were examined for the locations of their functional sites . Table 3 summarizes the results of the method application and comparison with experimental observations ( CSA classification ) . The first column presents the protein under consideration and the list of residues recognized by CSA . For two proteins ( rat annexin V and ButF ) , residues that are in direct contact with ligand [62 , 63] and/or are part of the functional site are given [64] . In Table 3 , the columns representing FOD results show the numbers of residues recognized by this method: agreement with CSA classification ( underlined ) , and residues defined by two methods—FOD and at least one of two other methods ( SuMo , ProFunc ) as biological activity-related residues ( in bold ) . Where the position of the amino acids differed by 1 ( closest neighbors ) versus the CSA classification or versus the position found by SuMo or ProFunc , the numbers are in italics in Table 3 . The description of the SuMo and ProFunc columns in Table 3 is given below ( Comparative Analysis ) . The residues recognized as potentially responsible for binding site creation in proteins of unrecognized biological function are given in Table 4 . Profile plots of were used to identify the positions recognized by the FOD model as related to functional sites . The profile plots of were examined for proteins of known and unknown biological activity ( Figures 1 , S1 , and S2; and Figures 2 , S3 , and S4; respectively ) . The residues with the highest appeared as peaks in the profile plots and were predicted to be functionally important . The values of indicate the level of hydrophobicity irregularity . It is interpreted that the higher the value , the higher the deficiency of hydrophobicity with respect to its idealized distribution according to Gauss function . Thus , the maxima identified as being represented by a particular amino acid point out the residues in the surrounding area where the hydrophobicity deficiency is significant . In most cases , this deficiency is caused simply by the presence of a cavity or by the highly irregular distribution of side chains . The profile together with the color scale visualizes the magnitude of the irregularity . The same scale applied to the 3-D presentation of the protein molecule is able to visualize the location of high values , particularly in the protein structure . It can be seen that the residues with high values appear to be placed in close mutual vicinity , often creating a cleft , which can be responsible for ligand ( substrate ) binding . The 3-D representations for selected proteins of known function are shown in Figure 3 , and for selected proteins of unknown biological function in Figure 4 . Other proteins under consideration are presented in Figures S5–S7 and Figures S8–S10 . The color scale expressing the magnitude of is as follows: red , high ; yellow , average ; green , low and negative . The white color denotes the experimentally verified amino acids as responsible for catalytic activity ( according to the CSA database ) . In most cases , the set of amino acids selected according to the FOD model is larger than the set of residues classified by CSA . This is because the profile also selects amino acids that are close in space , which create well-defined putative cavities that accompany the residues responsible for enzymatic activity . Amino acids indicated by FOD as belonging to the binding cavity are in space filling form . The molecules presented in Figure 3A and 3B are selected to show the best results; the molecules presented in Figure 3C and 3D demonstrate the cases of low accordance . Some of the protein molecules with high values shown in Figure 3A and 3B appeared to be highly accordant to the active site location . Other proteins with high values ( Figure 3C and 3D ) are not exactly located in the positions of the amino acids that make up the catalytic site . Nevertheless , the analysis of the larger set of proteins may suggest that the specificity of the mutual location of the residues represented by high values versus the position of the enzymatic site may be classified according to enzyme specificity . One hypothesis is that the residues responsible for complex fixation ( protein and ligand or substrate ) were selected by the FOD model . Another explanation for the mismatch between experimentally identified and automatically identified residues is simply that for multimeric chains , only the first chain was present in the analysis . The results summarize the comparison of the model applied to identify the ligand-binding site and two other methods dedicated to the same purpose: ProFunc and SuMo are given in Table 3 for proteins of known biological function and in Table 4 for proteins of unknown biological function . Table 3 presents the list of proteins ( the PDB accession numbers are given ) accompanied by the amino acids identified as function-related according to CSA classification . SuMo results ( for each SuMo search in question ) show the comparison with the FOD model for only one example of a functional site found by SuMo and present the residue numbers , which appeared to be common for these two methods ( column 4 of Table 3 ) . The limitation to compare only one SuMo result for one search is caused by the specificity of output generated by the SuMo procedure , which produces an enormous number of possible solutions for one particular protein molecule ( in most cases , thousands of variants ) . Each solution is presented with regard to another protein ( PDB number given ) , the functional site of which seems to be related to that found in the molecule under analysis . This procedure proposes a list of functional sites that sometimes represent changed functionality ( e . g . , ligands of different structure/characteristics are bound ) . One functional site with a functional site of the same/closest properties is selected . The presentation of all results is impossible to present here in complete form . In column 5 of Table 3 , the ratio of commonly recognized residues to the number of all residues recognized by SuMo for that hit is shown . As we see , the total number of amino acids classified by SuMo in most cases is the same or exceeds the number identified by the FOD model . The numbers given in the last two columns ( ProFunc ) of Table 3 represent positions of amino acids recognized by ProFunc by its best hit and method score . This is why the number of commonly recognized residues ( given in bold ) is lower than in the SuMo comparison . The results describing the analysis of proteins of unknown biological function are shown in Table 4 . The presentation is similar to that for proteins of known biological function with an obvious lack of underlined positions ( no CSA classification available ) . The SuMo results are additionally characterized by the relation between the SuMo score of the solution closest to that based on the FOD model ( highest number of common positions ) and the score value of best hit , as estimated by SuMo . The comparison of the methods selected for analysis is generally very difficult . The SuMo and ProFunc methods represent the methodology of the stochastic nature . The FOD seems to be a more heuristic method . SuMo and ProFunc produce very large outputs with long lists of possible approaches . Each of them is characterized by the scoring number calculated according to the number of contacts ( pairs of amino acids ) responsible for ligand–protein interaction . However , the number of residues commonly recognized by at least two analyzed methods seems to be quite high . Taking into account a very large discrepancy in the results of one particular method , the level of mutual accordance seems to be satisfactory . Tables 5 and 6 present the results aimed toward validating the FOD model–based results . The values present error levels calculated for the methods under consideration . These calculations take into account the number of mismatched residues versus the CSA , SuMo , and ProFunc classifications . Tables 5 and 6 also include comparisons versus functional site amino acids estimated by the above the 95th percentile value . The proteins of known biological function are characterized in Table 5 , and the proteins of unknown biological function are characterized in Table 6 . The false negative ( below diagonal ) and false positive ( above diagonal ) classifications are given as average ( for all analyzed proteins ) percentages of mismatched residues . The comparison is expressed by the level of error measured in the percentage of mismatched residues . The left value in each table cell was calculated by taking into account the exact amino acid numbers . The value on the right side expresses the percentage of mismatched residues when the tolerance of ( i + 2 ) / ( i − 2 ) amino acids ( the positions of the residues ) is taken into consideration . The FOD results are based on the profile along the polypeptide chain . The search for the percentile optimally discriminating the residues belonging to those classified by CSA can be performed . The values above the 95th percentile value appeared to be the best approach of local maxima as the criteria for function-related residue classification . The results of the comparison of the 95th percentile are shown in the “FOD 95th percentile” column . The interpretation of values given in Tables 5 and 6 is as follows . For example , in FOD versus ProFunc cases , 86% of residues found by the FOD method were not selected by ProFunc ( false positives ) . Taking the amino acids with ( i + 2 ) / ( i − 2 ) tolerance , the level decreases to 73% . In false negative cases , 81% of residues selected by ProFunc were not selected by FOD ( 65% when closest neighbors were taken into account ) . This study is not designed to give a thorough comparison of functional site tools , nor is it meant to review the current advances in this field . Therefore , the mutual comparisons between SuMo and ProFunc , SuMo and CSA , and ProFunc and CSA are not presented here . Additional analysis summarizing the applicability of the presented method is also shown in Table 7 . It is shown that the correctness of the FOD model depends on the enzyme class . Values in Table 7 express the percentages of the residues identified by the FOD method versus those identified by CSA . The highest agreement was found for the EC . 3 category ( hydrolases ) , where almost 70% of residues classified by CSA were found by the FOD model . The functional sites in enzymes belonging to the EC . 4 ( lyases ) and EC . 6 ( ligases ) classes were recognized quite well ( more than 60% ) . The lowest agreement was found for the EC . 2 class ( transferases ) , where the percentage of correctly predicted amino acids ( versus CSA classification ) was about 20% ( this seems nonrepresentative due to the low number of proteins under consideration in this class ) . The specificity of the active sites in particular enzymatic classes will be analyzed in future publications with respect to the FOD methodology . The larger number of proteins belonging to particular enzyme classes will be taken into consideration in the prospective analysis with respect to the applicability of the FOD model as the tool for functional site recognition . The increased number of proteins representing a particular enzyme class may clarify also the applicability of the method in relation to the detailed type of reaction catalyzed .
The recognition of functional sites in protein molecules is important for the identification of biological activity . The fully automatic method is highly expected . In analogy to the methods applied for protein structure prediction , the ligand-binding site can be recognized on the basis of comparative methods ( according to CASP [critical assessment of structure prediction] classification ) . The alternate possibility is to search for a ligand-binding site using new fold ( according to CASP classification ) techniques that use only the structure of individual proteins . The FOD method presented here identifies the potentially function-related amino acids . In contrast to SuMo and ProFunc , which are based on comparative analysis , the FOD method is of heuristic form , taking as its criterion the individual local hydrophobicity deficiency in a particular protein body . The ligands' ( as cofactors or cosubstrates ) presence makes the biological activity possible for some proteins . The enzymatic activity also requires substrate binding . The presence in the cavity of high specificity versus ligand/substrate is needed for this kind of interaction . The location of the cavity ( dependent on the protein character ) in protein molecules seems to be well recognized by the FOD model . The part of the protein molecule with high hydrophobic deficiency is recognized as a possible ligand-binding site ( or active site ) . Some results received according to the FOD model seem to be quite satisfactory ( Figure 1A and 1B and Figure 3A and 3B ) . The catalytic mechanisms of enzymes are quite differentiated and require appropriate molecular structures . The analysis of their specificity may clarify the origin of failure ( Figure 1C and 1D and Figure 3C and 3D ) . The possible protein–protein complex creation ( not taken into consideration in this analysis ) may significantly influence the results ( e . g . , Figures S1 and S6 ) . Two proteins ( in Figures 1C and 3C , and in Figures 1C , 3C , S2 , and S7 ) of common enzymatic specificity ( disulphide isomerase ) have been recognized on the basis of the FOD method as highly similar with respect to the mutual orientation of residues involved in cavity creation . The specificity of enzymes with respect to their active site construction is the aim of prospective analysis , which will be published soon , as well as analysis of proteins responsible for biological functions other than enzymatic ( e . g . , proteins responsible for transport as given in Table 3 ) . The calcium-binding sites in annexin V are not recognized by FOD , although the ion channel–creating residues are pointed out by this method according to expectations for the method of biological function recognition . The FOD model may also represent the specific hydrophobic environment for protein folding and was initially aimed at the simulation of the hydrophobic collapse of partially folded proteins . The heuristic model of protein folding , according to which the folding polypeptide is directed to follow the hydrophobicity distribution , is represented by the 3-D Gaussian function . The external force field may direct the folding process toward the hydrophobic core creation . The resulting structure appeared to be dissatisfactory , particularly because of the absence of a ligand-binding site in the final structural form . The presence of a ligand in the folding environment may ensure the specific binding cavity creation . Thus , it seems to be important or even necessary . The comparative analysis of the results of the FOD-based method with the results of SuMo and ProFunc ( Tables 3–6 ) reveals the very high similarity of obtained results . The methods use different criteria for classification . The exhaustive comparative analysis of the results obtained by the application of different methods seems to be necessary and has been taken into consideration; this will be published soon together with explanation of the source of these differences . The proteins shown in this paper represent mostly enzymes of varying biological activity , the relation of which to the character of the results will be the object of independent research . It is generally accepted that globular proteins consist of a hydrophobic core and a hydrophilic surface [36 , 40] . However , the core region is not well described by a spheroid of hydrophobic residues surrounded by hydrophilic residues due to channels that permeate the molecule [46 , 47] . The FOD model , when applied to protein structure , characterizes the hydrophobicity density in a continuous form by pointing out the irregularities in a hydrophobic core construction disturbing the regularity of hydrophobicity distribution [23–26] . Those irregularities seem to be good markers for ligand-binding sites or functionally important residues . Methods dedicated to active site recognition have been widely developed: SARIG [65] , Q-SITE FINDER [66] , HIPPO ( SPROUT ) [67 , 68] , FEATURE [69–71] , THEMATICS [72–74] , APROPOS [75] , DRUGSITE [76] , and LIGSITE [77] , to mention just a few . Limitation to two methods ( SuMo and ProFunc ) for comparative analysis in this paper is due to the very large variability of the models when applied . The method described in this paper is assumed to be applied for active site identification for a large set of proteins , the structure of which is planned to be generated using different methods ( FOD and ROSETTA [78] ) . The project geared toward biological activity identification in never born proteins ( NBPs ) is assumed to deliver the molecules of pharmacological application [79 , 80] . This is the main scientific goal for pharmacology application in the EuChinaGrid project . The FOD method is available at http://bioinformatics . cm-uj . krakow . pl/activesite .
The Protein Data Bank ( http://www . rcsb . org/pdb ) accession numbers for the proteins discussed in this paper are: rat annexin V ( 1A8A ) , ButF ( 1N2Z ) , phosphomannose isomerase ( 1PMI ) , triosophosphate isomerase ( 1TPH ) , protein disulfide isomerase ( 1MEK ) , 7 , 8-dihydroneopterin aldolase ( 2DHN ) , protein identified in the Pseudomonas aeruginosa genome ( 2AZP ) , protein identified in the Thermotoga maritima genome ( 2EWR ) , protein originated in the Thermus thermophilus genome ( 2D4R ) , protein originated in the Staphylococcus aureus genome ( 2FFM ) , myeloperoxidase ( 1MHL ) , and riboflavin synthase ( 1KZL ) .
|
We present here a method of defining functional site recognition in proteins . The active site ( enzymatic cavity or ligand-binding site ) is localized on the basis of hydrophobicity deficiency , which is understood as the difference between empirical ( dependent on amino acid positions ) and idealized ( 3-D Gauss function , or Fuzzy Oil Drop model ) distribution of hydrophobicity . It is assumed that the localization of amino acids representing a high difference of hydrophobic density reveals the functional site . The analysis of the structure of 33 proteins of known biological activity and of 33 proteins of unknown function ( with comparable polypeptide chain lengths ) seems to verify the applicability of the method to binding cavity localization . The comparative analysis with other methods oriented on biological function is also presented . The validation of predictability accuracy is shown with respect to the enzyme classes .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion",
"Supporting",
"Information"
] |
[
"chicken",
"eubacteria",
"computational",
"biology"
] |
2007
|
Prediction of Functional Sites Based on the Fuzzy Oil Drop Model
|
During the past twenty years , a number of molecular analyses have been performed to determine the evolutionary relationships of Onchocercidae , a family of filarial nematodes encompassing several species of medical or veterinary importance . However , opportunities for broad taxonomic sampling have been scarce , and analyses were based mainly on 12S rDNA and coxI gene sequences . While being suitable for species differentiation , these mitochondrial genes cannot be used to infer phylogenetic hypotheses at higher taxonomic levels . In the present study , 48 species , representing seven of eight subfamilies within the Onchocercidae , were sampled and sequences of seven gene loci ( nuclear and mitochondrial ) analysed , resulting in the hitherto largest molecular phylogenetic investigation into this family . Although our data support the current hypothesis that the Oswaldofilariinae , Waltonellinae and Icosiellinae subfamilies separated early from the remaining onchocercids , Setariinae was recovered as a well separated clade . Dirofilaria , Loxodontofilaria and Onchocerca constituted a strongly supported clade despite belonging to different subfamilies ( Onchocercinae and Dirofilariinae ) . Finally , the separation between Splendidofilariinae , Dirofilariinae and Onchocercinae will have to be reconsidered .
The Onchocercidae ( Spirurida ) , commonly referred to as filariae , are a family of parasitic nematodes characterised by a wide host-range in squamates , crocodilians , amphibians , mammals and birds [1] . As several filarial species are agents of human and veterinary diseases [2–4] , the Onchocercidae have been the subject of numerous studies . Presently , the family is divided into eight subfamilies , including 88 genera [5–8] . They are nematodes with an evolved life cycle , involving blood- or skin-inhabiting first-stage larvae transmitted by haematophagous arthropods vectors [1 , 9] . A pronounced regression of morphological features and a number of convergences resulting from a parasitic life style , as well as the absence of fossilised material , make it difficult to produce phylogenetic hypotheses for this family [10] . Thus , major questions regarding the classification of the Onchocercidae , their origin and evolution remain as yet unresolved . In numerous morphological and biological studies an attempt has been made to elucidate the evolution of the Onchocercidae [1 , 10] . First drawn up by Wehr in 1935 [11] , the classification of the family has been reconstructed several times [12 , 13] , with the most comprehensive analysis being that of Anderson and Bain in 1976 [6] . In 1994 [14] , the development of molecular techniques made the construction of the first molecular phylogeny of the Onchocercidae , based on the 5S rDNA sequences , possible . This pioneering study focused on ten species from six genera , representing two subfamilies ( Dirofilariinae and Onchocercinae ) . With the rise of the DNA barcoding approach to distinguish between species , studies on filarial nematode phylogeny have been predominantly based on 12S rDNA and coxI gene analyses [15–18] , and phylogenies for selected onchocercid species have been proposed . However , biological material has been scarce , impeding broad taxonomic sampling , and the markers used were not suitable to resolve the internal nodes which describe the evolution of the Onchocercidae [19] . Finally , a few studies have been conducted using alternative genes , but were either focused on nematodes in general [20–22] , a specific genus [23 , 24] or based on the mitochondrial genome [25] , all including a limited number of species . The aim of the present study was to propose a robust phylogenetic hypothesis of the relationships within the Onchocercidae based on the concatenation of seven loci ( two mitochondrial and five nuclear genes ) of 48 belonging to seven subfamilies . This hitherto most comprehensive sampling of the Onchocercidae casts doubt on the biological validity of some of the classically defined subfamilies .
Sixty-two specimens comprising 48 species belonging to 25 genera and representing seven of eight onchocercid subfamilies were analysed ( Table 1 and S1 Table ) [8 , 9 , 26–59] . Due to a lack of material , the Lemdaninae could not be included in this study . All procedures were conducted in compliance with the rules and regulations of the respective competent national ethical bodies ( S2 Table ) . Nematode specimens and DNA samples were deposited in the National Nematode collection of the Muséum National d’Histoire Naturelle ( MNHN ) , Paris , France; accession numbers are recorded in Tables 1 and S1 . Below we briefly list the specimens used in this study , followed by the source/collector in parentheses . For the author ( s ) and year of parasite and host ( authors only ) species , the reader is referred to Table 1 and S1 Appendix . DNA , L3 , microfilariae . From the following species , DNA had been extracted from adult worms for previous [18 , 19 , 41] and deposited in the MNHN collection ( Table 1 ) : Madathamugadia hiepei , Rumenfilaria andersoni , Cercopithifilaria rugosicauda , Aproctella alessandroi , Loxodontofilaria caprini , Mansonella ( Cutifilaria ) perforata , Onchocerca dewittei japonica , Onchocerca eberhardi , Onchocerca skrjabini and Ochoterenella sp . 1 . DNA of adult Setaria labiatopapillosa ( Dr Ben Makepeace , UK ) was given to us . DNA of Cercopithifilaria bainae had been extracted from infective larvae from Ixodes ricinus ( Prof Domenico Otranto , Italy ) . From the following species , DNA had been extracted from microfilaria: Brugia timori ( Dr Tri Baskoro , Indonesia ) ; . Mansonella ( Mansonella ) ozzardi ( Prof . Christian Raccurt , Haiti ) ; Dirofilaria ( Dirofilaria ) immitis ( Bayer Animal Health GmbH , Germany ) . From the following species , DNA had been extracted from adult females: Acanthocheilonema viteae , Brugia malayi and Brugia pahangi ( NIAID/NIH Filariasis Research Reagent Resource Center ( MTA University of Wisconsin Oshkosh—SJ 770–12 ) ; Loa loa ( Prof Jean Dupouy Camet , Hopital Cochin , France ) ; Litomosoides sigmodontis ( experimentally infected jirds , Meriones unguiculatus; MNHN ) . Additional specimens of adult filariae were recovered during dissections of their vertebrate hosts captured in the wild from different geographic areas ( Table 1 ) . They were mainly recovered from body cavities and lymphatic vessels or extracted from the dermis , subcutaneous or connective tissue , or tendons of limbs . Identification were made by several experts ( OB , KJ , RG , YM , YK , SL , and AR ) , based on morphological studies . A few species have not yet been named , as their description is still in progress . Samples were fixed and stored in 70% ethanol . The anterior and posterior parts of worms were used for morphological studies , whereas the median part was processed for molecular analysis . Jirds are maintained in the animal facilities of UMR7245 , MNHN , as hosts of L . sigmodontis; they are inoculated intraperitoneally with 70 infective larvae and sacrificed 60 days post infection , upon which adult worms are recovered from the peritoneal cavity . All experimental procedures were carried out in strict accordance with the EU Directive 2010/63/UE and the relevant national legislation , namely the French “Décret no 2013–118 , 1er février 2013 , Ministère de l’Agriculture , de l’Agroalimentaire et de la Forêt” , National licence number 75–1415 approved animal experiments: protocols were approved by the ethical committee of the Museum National d’Histoire Naturelle ( Comité Cuvier , Licence: 68–002 ) and by the “Direction départementale de la cohésion sociale et de la protection des populations” ( DDCSPP ) ( No . C75-05-15 ) . Some non-human vertebrates were captured for experimental procedures , subject to the ethics approval of the relevant national bodies ( S2 Table ) , while others were obtained at abattoirs or donated to the MNHN by hunters or veterinarians ( S2 Table ) . The MNHN does neither solicit nor compensate for these donations . Human samples were not collected specifically for this study . They were provided by third parties . Human blood samples were collected with the ethics approval of the relevant national bodies ( S2 Table ) . The adult female of Loa loa had been surgically removed from a male patient in Hospital Cochin , France . The adult subject had given oral consent to Prof J . Dupouy Camet , head of parasitology department at Hospital Cochin , to donate the worm to the MNHN ( S2 Table ) . The MNHN did not and will not compensate for the donation . DNA was extracted using a commercial kit , following manufacturer’s instructions ( QIAamp micro kit , Qiagen , Germany ) . A preliminary step of disruption for two cycles of 30 seconds at 30 Hz using a TissueLyser II ( Qiagen , Germany ) was added . PCR screening of filarial nematodes was based on seven partial sequences of seven different genes: two mitochondrial genes , 12S rDNA ( approximately 450 base-pair ( bp ) sequence ) and cytochrome oxidase subunit I ( coxI; approximately 600 bp ) ; five nuclear genes , 18S rDNA ( approximately 740 bp ) , 28S rDNA ( approximately 900 bp ) , the myosin heavy chain ( MyoHC; approximately 785 bp ) , RNA polymerase II large subunit ( rbp1; approximately 640 bp ) , 70 kilodalton heat shock proteins ( hsp70; approximately 610 bp ) . Amplification of the 12S rDNA and coxI sequences was conducted according to Casiraghi et al . [15] . For the remaining five genes , primer pairs were designed ( S3 Table ) based on regions that had been found conserved among nine species for which drafts of or complete genome were accessible: from NCBI database Brugia malayi ( PRJNA27801 ) , Loa loa ( PRJNA37757 ) , Onchocerca flexuosa ( Weld , 1856 ) ( PRJEB512 ) , Dirofilaria immitis ( PRJEB1797 ) Onchocerca ochengi ( PRJEB1809 ) ; from Nematode Genomes from the Blaxter lab , University of Edinburg ( www . nematode . org ) Acanthocheilonema viteae ( nAv . 1 . 0 ) , Litomosoides sigmodontis ( nLs . 2 . 1 ) ; and from Filarial worms Sequencing Project , Broad Institute of Harvard and MIT ( http://www . broadinstitute . org/ ) Onchocerca volvulus ( Leuckart , 1893 ) and Wuchereria bancrofti ( Cobbold , 1877 ) . For each gene , primer pairs adapted to nested PCR were designed ( S3 Table ) . PCRs were processed in a final volume of 20 μl under the conditions summarised in S3 Table . Obtained PCR products were purified using the SV Wizard PCR Purification Kit ( Promega , USA ) and sequenced directly . A total of 345 sequences were deposited in the GenBank Data Library: KP760116 to KP760460 ( S4 Table ) Sequences generated during the current study and previously published sequences from draft/complete genomes were aligned using MAFFT [60] . To check for the absence of stop codons , the alignment of coding genes was translated using EMBOSS Transeq [61] , and a comparison with available transcript sequences made . Using the corrected version of the Akaike Information Criterion ( AICc ) , JModelTest analysis [62] was performed to establish the evolution model best adapted to the sequences alignment for each individual gene and for the concatenation of all genes . The General time-reversible plus Invariant sites plus Gamma distributed model ( GTR+I+Γ ) offered the best fit for coxI , hsp70 , 12S rRNA and 28S rRNA sequences as well as the concatenated alignment; the General Time-Reversible plus Gamma distributed model ( GTR+Γ ) for MyoHC and rbp1 sequences alignment; and the Kimura 2-parameter Invariant sites plus Gamma distributed model ( K80+I+Γ ) for 18S rRNA sequences alignment . To root the trees , two species were included as outgroups: Filaria latala ( Spirurida: Filariidae ) and Protospirura muricola ( Spirurida: Spiruridae ) . Phylogenetic relationships between onchocercid taxa , based on the concatenated dataset , were performed by Bayesian inference using MrBayes [63] . A partitioned model was implemented to estimate evolution parameters separately for each gene . Two runs were performed using five millions steps with four chains , with tree sampling every 1 , 000 generations; the first 1 , 250 points were discarded as burn-in and Posterior Probabilities were calculated from these post-burning trees . In addition , Maximum Likelihood ( ML ) was used to infer a phylogenetic tree based on the partitioned concatenated dataset , and was executed with 1000 slow bootstrap replicates using RaxML [64]; presented in Supplementary data ( S1 Fig ) . All new sequences generated in this study were deposited in GenBank ( http://www . ncbi . nlm . nih . gov/genbank ) under the accession numbers: KP760168 to KP760211 for coxI , KP760314 to KP760357 for 12S rDNA , KP760116 to KP760167 for 18S rRNA , KP760212 to KP760262 for MyoHC , KP760410 to KP760460 for hsp70 , KP760263 to KP760313 for rbp1 , KP760358 to KP760409 for 28S rRNA .
An important consideration when proposing a comprehensive phylogeny for any taxonomic group is the suitable rooting of the tree and , thus , the choice of appropriate outgroups . Previous studies of the Onchocercidae , based on coxI and 12S rRNA gene analyses , have used spirurid representatives of the Thelaziidae [15 , 16] and Filariidae [16] as outgroups . Furthermore , some phylogenetic analyses of nematodes based on SSU rDNA [21] or complete mitochondrial genome sequences [22] suggested other spirurids as the Physalopteroidea or Diplotriaenoidea to be appropriate outgroups for the Onchocercidae . In the present study , P . muricola ( Spiruridae ) and F . latala ( Filariidae ) were investigated as potential outgroups . Our multi-gene analyses indicate that this choice is reasonable ( Fig 1 and S1 Fig ) . To support this result , the average genetic divergence between the outgroup and ingroup taxa was about 20% ( 21% for P . muricola; 25% for F . latala ) , and only about 14% between onchocercid specimens themselves . It was , therefore , concluded that both P . muricola and F . latala are appropriate outgroups for the present multi-gene dataset . In congruence with previous systematic analyses [5 , 6] , the chosen rooting clearly supported the monophyly of the Onchocercidae . All onchocercid taxa examined in the present study formed a monophyletic group and five strongly supported clades within the family were identified ( Fig 1 ) . Based on their monophyly and to facilitate the description of tree topology , these clades will be referred to as ONC1 to ONC5 ( Fig 2 ) . In the present phylogenetic analysis , the Oswaldofilariinae , Icosiellinae and Waltonellinae occupied a commun position as a highly supported clade ( ONC1 ) and sister group of all other taxa studied within the Onchocercidae ( Fig 1 . ) . The three subfamilies appeared to be closely related , with ONC1 having evolved independently from the remaining onchocercid representatives ( Fig 2 . , S1 Fig ) . Previous morphological studies revealed a number of ancestral characters in the Icosiellinae and Oswaldofilariinae , such as infective larvae possessing cephalic spines ( Icosiellinae ) , and a vulva that is positioned posterior to the oesophagus ( Oswaldofilariinae ) [10 , 65] . Therefore , these subfamilies have traditionally been considered “ancient” [1 , 10 , 65] . In addition , it has been suggested that their Gondwanian distribution is an indication of an ancient evolution which occurred prior to the supercontinent’s break-up [1 , 10] . Interestingly , the Oswaldofilariinae appear to have evolved independently from the remaining two subfamilies of clade ONC1 and were placed as sister group to a clade that consists of Ochoterenella spp . ( Waltonellinae ) and Icosiella neglecta ( Icosiellinae ) ( S1 Fig ) . Indeed , previous studies emphasised the presence of numerous plesiomorphic characters ( long oesophagus , large buccal capsule and the presence of deirids ) within Oswaldofilaria spp . [10] . Moreover , the host spectrum of these subfamilies differs: oswaldofilariines are parasites of squamates , whereas members of Icosiellinae and Waltonellinae are exclusively parasites of anurans [5] . The systematic position of the Setariinae among the Filarioidea superfamily has long been questioned . Initially the Setariinae had been placed into Filariidae because they share certain morphological similarities with representatives of this family ( spines on anterior end of first stage larva and adult cuticular elevation ) [11 , 13] . Nevertheless , based on other biological and morphological traits ( e . g . viviparous , long microfilaria with pointed spineless tail ) , it would have been equally possible to consign them to the Onchocercidae . At a later stage , the Setariidae was proposed as a family with the purpose of removing the Setariinae from the Filariidae[12] , and current classifications list the Setariinae as a subfamily within the Onchocercidae [6] . Such an affiliation between the Setariinae and Onchocercidae has since been supported by many molecular phylogenies [16 , 19 , 25 , 65] . The same was true for the present study , where the two Setaria species grouped in one clade ( ONC2 ) with good support ( Fig 2 ) . Excepting ONC1 , the clade consisting of Setaria spp . was sister to all remaining onchocercid taxa sampled in this study . In the past , the Setariinae , Oswaldofilariinae and Waltonellinae were grouped together on the previous phylogenetic trees despite an unresolved topology [16 , 65] . For some time , morphological studies of infective stages from Setaria spp . ( long larvae , presence of deirids ) have supported the hypothesis that Setariinae may have emerged early in the evolution of Onchocercidae [10 , 59] . In contrast , the study of larval development of S . labiotopapillosa suggested that Setariinae , like the remaining Onchocercidae , could have been derived from habronematid ancestors , but that speciation leading to the Setariinae occurred later and independently from the remaining Onchocercidae [1 , 66] . For example , the development of the glandular cells of the oesophagus in the Setariinae was shown to be closed to the development observed in some potential ancestors of the Onchocercidae , namely the Habronematidae [66 , 67] . The current phylogeny strongly supported the position of Setariinae as a group derived from an independent speciation to the examined Onchocercidae . The remaining taxa of the Onchocercidae analysed in this study grouped in three distinct and well supported clades ( ONC3 , ONC4 and ONC5 ) ( Fig 2 ) , including representatives of the Dirofilariinae , Onchocercinae and Splendidofilariinae . It is noteworthy that the tree topology demonstrated here is not congruent with classic systematic delineations: One of the main difficulties in comparing the results of the present study with those of earlier molecular analyses was either the lack of broad taxonomic sampling or incompletely resolved deeper phylogenetic relationships in preceding studies [15 , 16 , 18 , 19] . However , an equivalent to clade ONC5 clade had been identified before , in the first phylogeny including L . loa , B . malayi and W . bancrofti and Mansonella ( M . ) perstans ( Manson , 1891 ) [14] . A similar clade , including B . malayi , W . bancrofti , L . loa and Chanderella quiscali Linstow , 1904 , a Splendidofilariinae , was confirmed by a recent analysis of mitochondrial genomes [25] . To summarise , the present phylogeny did neither support the monophyly of the Dirofilariinae , Onchocercinae nor Splendidofilariinae ( see ONC3 and ONC5 , Fig 2 ) . Interestingly , however , the relationships exhibited between the various taxa in clades ONC3 and ONC5 , although traditionally included in different subfamilies , are supported by morphological as well as biological features ( see discussion above ) . Based on the combined analysis of molecular , structural and developmental characters , we , thus , conclude that a reassessment of the boundaries between these three subfamilies is called for . The parasitic life style of the Onchocercidae has led to the reduction of certain morphological traits and promoted convergent evolution in others [10] , reducing the usefulness of structural characters in developing a phylogenetic framework for this family . One might thus look at the role of vectors and hosts for further clues as to the speciation of the Onchocercidae . Three ( ONC1 , ONC2 and ONC3 ) of the five clades exhibited on the current phylogenetic tree use exclusively dipterans as vectors . In addition , dipterans are involved in the transmission of members of clades ONC4 and ONC5 , supporting a hypothesis that old evolutionary bonds exist between this order and the Onchocercidae ( Fig 3 , S1 Table ) . The diversity of vectors used in clades ONC4 and ONC5 ( Fig 3 ) , on the other hand , suggests that the evolution of filarial nematodes is not strictly constrained by their vector but does permit changes to other , more appropriate groups of vectors , to ensure parasite transmission , colonisation of new ecological niches and diversification . Clade ONC4 shows an adaptation to a new group of vectors–ticks and mites . It should , however , be emphasised that some Acanthocheilonema spp . ( ONC4 ) are known to use fleas , hippoboscid flies or sucking lice as vectors [9] , and Dipetalonema spp . are transmitted by ceratopogonid biting midges ( Fig 3 ) . Considering relationships of the Onchocercidae with their definitive hosts , the present analysis revealed a group of parasites adapted exclusively to cold blooded tetrapods , namely squamates , crocodilians and amphibians ( ONC1 ) , while their sister group diversified in various groups of mammals ( ONC2-5 ) , as well as birds and squamates ( ONC5 ) ( Fig 4 ) . It has been suggested that the parasite-host relationships of M . hiepei ( parasite of the gecko Chondrodactylus turneri ) and F . candezei ( parasite of lizards ) may be the result of secondary capture with the narrow geographical localisation of Foleyella and Madathamugadia , i . e . Africa and Madagascar , supporting the hypothesis of capture phenomena [74] . The present phylogenetic analysis does not support a close relationship between parasites of birds , but representation of avian parasites was limited . All three parasites of birds formed part of clade ONC5 , which presented the most heterogeneous host range within the Onchocercidae , comprising parasites of three classes of vertebrates , reptiles , birds and mammals . The diverse mammalian hosts represented in clades ONC3-5 indicate multiple events of host-switching and radiation within the evolutionary history of the group . In summary , the overall picture revealed by the current analysis does not suggest a broad pattern of onchocercid parasite-host/vector coevolution and is consistent with earlier studies suggesting that parasitic nematode speciation could be related to a variety of events of host switching and host acquisition and/or geographical and ecological drift [75–77] . The present study casts doubt on the validity of of the currently eight traditional onchocercid subfamilies . Traditional classifications of the Onchocercidae have been established mainly on the basis of morphological characters of adult worms . However , as suggested by previous authors [10 , 68] , biological features , such as larval biology or the morphology of the third-stage larva , may offer a better phylogenetic resolution . It is hoped that an approach combining broad molecular taxonomic sampling and traditional morphological as well as life history studies , will eventually lead to the development of more comprehensive phylogenetic hypotheses for this fascinating group of nematodes .
|
Filariae are predominantly tissue-dwelling nematodes of the Onchocercidae ( Spirurida ) . They are parasites of terrestrial vertebrates and some of them are particularly well known as agents of human diseases in tropical environments ( e . g . onchocercosis , lymphatic filariosis , loaosis ) . Because of their predilection for host tissue sites , filariae are difficult to collect and their diversity is far from being well-investigated . Given the lack of fossilized material , it is not possible to formulate a comprehensive evolutionary hypothesis for this group based exclusively on morphological studies . Molecular analyses on the other hand have so far been based on a limited number of filarial species only or have used genetic markers convenient for species differentiation but not for retracing filarial history . Thus a consistent evolutionary framework for this family is still not in place . The originality and strength of the present study rest within its ability to access an unusually large range of biological samples , comprising 48 species representing seven of the eight onchocercid subfamilies . Using a new multi-gene dataset analysis , five major clades within the family were defined , including i ) a group of ancestrally-derived subfamilies , ii ) Setariinae as a sister group to all remaining subfamilies and iii ) one large clade comprising genera of the Dirofilariinae , Onchocercinae and Splendidofilariinae .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results/Discussion"
] |
[] |
2015
|
Shaking the Tree: Multi-locus Sequence Typing Usurps Current Onchocercid (Filarial Nematode) Phylogeny
|
The mechanisms of stress tolerance in sessile animals , such as molluscs , can offer fundamental insights into the adaptation of organisms for a wide range of environmental challenges . One of the best studied processes at the molecular level relevant to stress tolerance is the heat shock response in the genus Mytilus . We focus on the upstream region of Mytilus galloprovincialis Hsp90 genes and their structural and functional associations , using comparative genomics and network inference . Sequence comparison of this region provides novel evidence that the transcription of Hsp90 is regulated via a dense region of transcription factor binding sites , also containing a region with similarity to the Gamera family of LINE-like repetitive sequences and a genus-specific element of unknown function . Furthermore , we infer a set of gene networks from tissue-specific expression data , and specifically extract an Hsp class-associated network , with 174 genes and 2 , 226 associations , exhibiting a complex pattern of expression across multiple tissue types . Our results ( i ) suggest that the heat shock response in the genus Mytilus is regulated by an unexpectedly complex upstream region , and ( ii ) provide new directions for the use of the heat shock process as a biosensor system for environmental monitoring .
The majority of molluscan species go through two principal developmental phases , a larval embryo ( motile phase ) followed by a clumping structure ( sessile phase ) , when they are permanently attached to an underwater substrate . This lifecycle , common amongst marine invertebrates , poses challenges for adaptation and tolerance for a wide range of conditions at the littoral zone , including steep salinity or temperature gradients . Key model organisms for molluscan biology include species from the genus Mytilus , in particular M . edulis , M . galloprovincialis and M . californianus . Crucially , the latter species is a target organism for a genome sequencing project , whose results are eagerly expected by the community ( http://www . jgi . doe . gov/sequencing/why/3090 . html ) . The Mytilus species group provides an ideal model both for fundamental questions of animal adaptation to stress response , as well as biotechnological applications , primarily as a pollution biosensor [1] . Its use extends into biomimetics [2] , in particular protein-based medical adhesives [3] , with potential applications in fields such as dentistry [4] . Moreover , its relatively complex developmental structure and higher taxonomic status as an invertebrate , combined with the fact that it can suffer from mussel haemic neoplasia , renders this organism a potential model for human leukemia and an ideal biomarker for pollution-induced disease [5] . In this context , it is important to understand the mechanisms by which mussels tolerate and cope with environmental stress , given that their behavioral options are highly restricted , due to the sessile phase of their lifecycle . In the past , comparisons between motility and sessility for higher organisms have been primarily confined to animals versus plants [6] , with follow-up studies focusing on comparisons between large animals , e . g . humans , versus large plants , e . g . trees , and the trade-offs for the tree body plan [7] . Less attention has been paid to adaptations by sessile animals , in particular intertidal invertebrates ( or , “plant” equivalents ) [8]–[9] , and the molecular mechanisms through which they achieve tolerance to stress . One exception is represented by heat shock response , a key factor for temperature adaptation that has been studied in this context to a certain extent [8] , and specifically in Mytilus with regard to the Hsp70 [10] and Hsp90 [11] genes . Transcriptional regulation can be achieved either by an extensive repertoire of paralogs and transcription factors ( ‘gene content strategy’ ) or a complex structure of promoters ( ‘gene structure strategy’ ) . Analysis of comprehensive datasets has clearly demonstrated that transcription factors ( TFs ) and transcription-associated proteins ( TAPs ) are not universally distributed but highly taxon-specific and that relative TF gene content increases with the taxonomic scale [12]–[13] . Such comparisons have been later extended by follow-up studies that analyzed TAP complements and their expansion rates in plants [14]–[15] . Thus , it is now known that one way by which plants , sessile organisms par excellence , achieve a finer degree of regulation is by the expansion of TF/TAP complements and a ‘gene content strategy’ . Yet , it is unclear whether similar trends are followed in sessile animals , since entire genome sequences for those are lacking so far , limiting the range of comparative genome-wide studies that can be performed . As far as paralogs are concerned , recent studies that have focused on the heat shock response in plants , and in particular Arabidopsis thaliana , have revealed that the process involves up to 21 known TFs and four heat shock protein ( Hsp ) families ( Hsp20/70/90/100 ) [16]–[17] . Despite a cursory resemblance to mammals , in Drosophila thermal sensing is achieved by a unique repertoire of genes [18] , including thermostat systems not exclusively involving heat shock proteins [19] . In other words , and probably for different reasons , a gene content strategy might prevail in both model organisms for plants ( A . thaliana ) and motile invertebrates ( Drosophila ) . Thus , it is worth examining what are the mechanisms through which stress response is regulated in sessile marine invertebrates in general , and the Mytilus genus in particular , and which strategy dominates gene expression . We focus on the Hsp90 family as a case study for stress response in sessile animals and examine the structure and function of the Hsp90 upstream region in M . galloprovincialis . Previously , two distinct Hsp90 genes with the same genomic organization have been isolated from M . galloprovincialis [11] , herein called Mghsp90 genes . Detailed sequence analysis revealed that the two genes contain nine exons and exhibit great similarities in both the 5′ non-coding and the coding regions but differ in their 3′ non-coding regions , as well as in three introns , due to the presence of repeated sequences [11] . The 5′ non-coding region of both genes contains a non-translated exon and multiple binding sites for various transcription factors , highly suggestive of potential interactions of these factors with the Hsp90 promoter and subtle patterns of gene regulation [11] . A comparative analysis of Hsp90 gene content across all taxa with available sequence data has clearly shown that invertebrate genomes contain a relatively small number of Hsp90 genes ( 3–4 genes ) , compared to those of vertebrates ( >5 genes ) [20] . Thus , it appears that the Mytilus genome might contain a relatively small number of TFs ( e . g . heat shock factors or HSFs – no such factors can be detected in the Mytilus californianus EST collection , not shown ) and/or Hsp90 genes , raising the question how the expression of Hsp90 and other heat shock genes is regulated in sessile invertebrates . In the present work , we perform a detailed analysis of the Mghsp90 upstream region in terms of structure and expression , and reveal the presence of previously undetected sequence elements of unknown function . Based on tissue-specific expression data , we also delineate the potential associations of Mghsp90 with another 174 genes that are involved in a complex pattern of expression across tissues . These two discoveries are discussed within the context of existing knowledge and are expected to contribute towards a deeper understanding of the heat shock response in sessile organisms .
The comparison of the 5′ upstream region of Mghsp90 genes to their homologs in two model organisms for which there is extensive genomic evidence and humans reveals an increase of complexity in TF binding sites including heat shock elements ( HSEs , binding sites for HSFs – see Methods ) . The M . galloprovincialis Hsp90 region exhibits a peculiar degree of unexpected complexity with regard to its phylogenetic context , not only in terms of quantity of predicted elements but also in fine structure of the promoter ( Figure 1 ) . The Mytilus region contains more regulatory sites than the D . melanogaster region ( namely , 14 sites vs . 8 ) , a total count similar to that of the human Hsp90 beta gene ( 17 sites – Figure 1 ) . Moreover , it is host to two newly identified elements ( Gamera and a genus-specific sequence ) , both of unknown function ( represented by blue bars , Figure 1 ) , followed by a HSE-rich region with a CAAT binding site and a putative p53 binding site ( see also below , and Figure 1 in Protocol S1 ) . Curiously , upstream of the first exon ( 1158 nucleotides ) of Mghsp90 , there exists a 201-base pair ( bp ) sequence element with a putative GAGA factor binding site ( Figure 1 ) , 69% identical over 181 nucleotides to the medaka fish Oryzias curvinotus LINE-like repetitive sequence Gamera [21] . The similarity extends over positions 1907–2085 of the O . curvinotus 4493-bp sequence entry ( Genbank accession number AB081572 , GI:19570857 ) and more specifically over the ‘open reading frame’ b ( defined at positions 1353–3052 ) [21] . Thus , this region of approximately 200 nucleotides is only a fraction of the putative ORF b and , to our knowledge , it is the first time this segment is reported outside the Oryzias genus and its closest relatives [21] ( Figure 2 ) . Moreover , multiple copies of this region can also be identified in the genome of the blood fluke Schistosoma mansoni [22] ( Figure 2 , Figure 2 . 1 in Protocol S1 ) . Fragments of this sequence are also present in ( i ) the Expressed Sequence Tag ( EST ) database , more specifically in the neural transcriptome and thus genome of the gastropod Aplysia californica [23] , the termite Hodotermopsis sjoestedti [24] , the African cichlid fish Oreochromis niloticus ( Lee et al . , unpublished , GI: 253867024 ) , the mollusc Lymnaea stagnalis [25] and the sea anemone Nematostella vectensis [26] ( in that order of sequence similarity – Figure 2 . 2 in Protocol S1 ) ; ( ii ) the unfinished high-throughput genomic sequence database ( Figure 2 . 3 in Protocol S1 ) , in the genome of sea urchin Strongylocentrotus purpuratus [27] and ( iii ) the Whole-Genome-Shotgun Sequence database ( Figure 2 . 4 in Protocol S1 ) in the genome of the hemichordate Saccoglossus kowalevskii ( unpublished ) . The functional significance of this element is not clear , yet given that the region can be identified in at least ten – highly unrelated and primarily aquatic – species , the presence of a transposable element of a highly mobile nature ( or its evolutionary relic ) is indicated ( Figure 2 ) . In M . galloprovincialis , it has also been shown that mobile elements reside within introns of the Hsp70 genes [10] , however there is no detectable sequence similarity between those elements and the Mghsp90 Gamera-like sequence presented here . Another feature of the Mghsp90 upstream region is a genus-specific sequence , approximately 100-bp long , located 787 positions before the first exon of Mghsp90 genes ( Figure 1 ) . This region is much more phylogenetically restricted than the Gamera element , found only in the genus Mytilus , namely the M . galloprovincialis mytilin B precursor gene [28]–[29] – ( accession number: AF177540 . 1 , positions 777–815 antisense strand , non-coding region ) , a lysozyme gene ( AF334662 . 1 , positions 1016–1050 sense strand , second intron ) [30] and a cDNA ( AM878017 . 1 ) both from M . edulis , and a cDNA sequence from M . californianus ( GE753693 . 1 ) ( Figure 3 ) . This genus-specific sequence does not contain any transcription factor binding sites ( Figure 1 ) , thus its functional significance is not known at present . It is worth noting that similarly to the mytilin B gene , another antimicrobial peptide gene , the M . galloprovincialis defensin 2 ( MGD2 ) gene , contains a 160-bp long element with similarities to the M . edulis lysozyme gene ( fourth intron ) , two glycosidase gene introns ( endo-1 , 4-beta-D-glucanase – AJ308548 . 1 , 2nd intron; endo-1 , 4-mannanase – AJ271365 . 2 , 5th intron ) , the 3′-UTR of the M . galloprovincialis Hsp70-1 gene , all being similar to an ISSR sequence ( AJ938114 ) , indicating the presence of a transposable element [31] . The above mentioned genes all have catabolic roles and might indeed be connected to defense mechanisms , broadly associated with stress . Further study is required in order to understand the role of these genus-specific sequences in the molecular physiology of the above mentioned loci . A putative binding site for p53 is located between two HSEs in the 5′ regulatory region of the Mghsp90 genes [11] ( Figure 1 ) , being identical to the consensus binding site of human p53 to retinoblastoma susceptibility gene [32] . This binding site is evidently absent from other species , including C . elegans and D . melanogaster , but present in the human Hsp90 beta gene [33] ( Figure 1 ) . The p53 proteins from two Mytilus species exhibit very high similarity to their human homologs , and especially in the DNA binding domain , the transcriptional activation domain ( TAD ) and the nuclear localization signal . In addition , residues mutated in various human cancers are also conserved in the Mytilus p53 proteins [34] . It should be noted that p53 is phylogenetically restricted to animals while the molluscan versions ( Decapodiformes , Bivalvia and Haliotis sp . ) exhibit a very high similarity to the vertebrate sequences ( not shown ) . The prediction of the p53 binding site in Mytilus is based on the known association of p53 with the upstream region of the human Hsp90 beta gene [33] , the conservation of the Mytilus p53 genes [34] and the observation that an identical site is present in human Hsp90 ( an Mghsp90 homolog ) [11] . In order to further establish the validity of the predicted p53 binding site in a phylogenetic context , we have searched the non-redundant nucleotide database with the Mghsp90 genes as queries ( see Methods ) . We subsequently identified 215 homologous target regions , with the closest sequence-similar entries carefully selected to exclude cDNA clones or partial coding sequences , across a wide taxonomic spectrum ( Figure 1 in Protocol S1 ) . These sequences were scanned for putative p53 binding sites ( 732 matches in total , see Methods ) , conditioned on the p53 phylogenetic distribution mentioned above; in other words , sites found in organisms known to encode for p53 were considered as positive cases ( 727 in total ) , while the remainder were treated as negative cases ( 5 in total ) . Despite well-understood limitations , e . g . the under-representation of certain species in terms of comparable Hsp90 sequence data and the over-representation of others in terms of redundant sequences , it is evident that p53-containing species exhibit a high number of predicted p53 binding sites ( primarily chordates ) , while other organisms ( such as fungi or plants ) , present a sporadic pattern of false positive hits , as expected . The exception in this otherwise consistent picture is the molluscs ( Bivalvia and Haliotis sp . ) , having a small number of predicted p53 binding sites ( Figure 1 in Protocol S1 ) . The shortage of sequence information for molluscs , coupled with a possibly non-canonical sequence motif , leaves the question open for the unambiguous detection and experimental confirmation of the elusive molluscan p53 binding site . The presence of a putative p53 binding site in the promoter region of the Mytilus Hsp90 genes raises questions about the possible implication of Hsp90 proteins in molluscan leukemia . Very recent studies on the association of p53 with heat shock response [35] , the differential expression of p53 in mussel haemic neoplasia [5] , and the impact of pollutants on p53 expression [36] underline the potential involvement of p53 in both heat shock response and neoplasia and its irregular similarity to vertebrate homologs [37] , as well as its potential use as a marker for environmental monitoring [34] . In other species , namely soft-shell clams , certain results also indicate that environmentally induced alterations in p53 might contribute to leukemia [38]–[39] . Indeed , expression studies have established that Hsp genes and a p53-like gene are abundant in M . galloprovincialis [40] , especially in pollutant exposed mussels [41] , now searchable through the Mytibase resource [42] . Moreover , there is evidence from proteomics studies that Hsp proteins are expressed in stress conditions and can potentially be used as pollution biomarkers [43]–[44] or temperature biosensor [45] . In order to investigate co-expression patterns for Mghsp90 genes , we have extracted tissue-specific gene expression data available in Mytibase , encompassing 3840 cDNA sequences [42] . Following normalization ( see Methods ) , we detected 547 genes ( 14% of total , in the ‘original’ network ) that are differentially expressed across all four tissue types under investigation ( namely gills , gonads , foot and digestive gland – Figure 4 ) . A two-way clustering across genes and tissues confirms that the four tissue types can be accurately detected ( Figure 4A ) . This step also suggests that the 547 differentially expressed genes can be clustered into four distinct classes corresponding to the four tissues , with relatively low overlap ( Figure 4A ) . A Principal Component Analysis of the original network further confirms the inter-replicate reproducibility and tissue specificity , indicating the high quality and consistency of the initial gene expression data ( Figure 4B ) . To infer gene associations via co-expression profiles , PCCs ( see Methods ) were computed for all possible pair-wise gene permutations of the original network . High PCC values correspond to a large similarity in expression profiles across four tissue types . Only those gene pairs with PCC>0 . 90 were further considered . This step yielded a global co-expression network , defined as the ‘inferred’ network , containing 3692 nodes and 57697 edges ( Figure 5 ) . The inferred network represents 96% of all cDNA clones in the original network . Such high coverage may be explained by the limited number of experimental replicates provided in the dataset . To ensure that only significant associations are considered , MCL clustering ( see Methods ) was performed to produce a ‘clustered’ network with 1719 nodes and 43286 associations ( Table 1 ) . The clustered network represents a subset of the inferred network enriched with the most highly connected genes with the strongest co-expression values ( Figure 5 ) . Interestingly , of the 547 differentially expressed genes obtained initially , 271 ( ∼50% ) are present in the clustered network , signifying a sufficient coverage of differential expression . This enriched network thus maintains 75% ( 43286/57697 ) of network edges , from which more reliable associations can then be extracted . To delineate the involvement of the wider Hsp class of genes in normal M . galloprovincialis tissue , 8 cDNA sequences corresponding to 4 distinct Mytilus Hsp genes , labeled as Hspa5 ( Grp78 homolog ) , Hsp70 , Hspa90 ( Mghsp90 ) , and Ankrd45 ( similar to heat shock 70 KD protein C precursor ) were identified in the clustered network ( Figure 5 ) . The “Ankrd45”-like sequence ( e . g . XP_290882 . 1 ) warrants description: its N-terminal part contains ankyrin repeats most similar to the ankyrin repeat domain of the human p53 binding protein [46] , while its C-terminal part is similar to Grp78 , a homolog of Hsp70 ( Figure 3 in Protocol S1 ) . Structural evidence indicates that the ankyrin repeats of p53 binding proteins ( 53BP2 ) bind to the L2 loop of p53 [47] , implicating a configuration of ankyrin repeats such as the one found in Ankrd45 , in a potentially mediated p53-Hsp70 domain interaction . In fact , since the initial discovery that the Hsp70 promoter is regulated by p53 [48] , there is mounting evidence that these two proteins are involved in various processes , including oral dysplasia [49] , endometrial carcinomas [50] , gastric cancers [51] , ischemia [52] and wound healing [53] . These interactions have been reviewed elsewhere [54]–[55] . Similarly , it has been demonstrated that p53 requires the activity of Hsp90s [56] and the structural [57] and biochemical [58] basis of this interaction has been deciphered . In fact , it appears that p53 , Hsp70 and Hsp90 are involved in a complex interplay during carcinogenesis [59] . To examine Hsp-related associations in greater detail , the nearest-neighbor members of 8 Hsp cDNA clones were selected , defined as the Hsp network ( Figure 5 ) . This network contained 174 genes and 2226 associations , accounting for 4 . 5% of genes in the original network ( Tables 1 and 2 in Protocol S1 – node labels refer to MyArray1 . 0 identifiers , see Methods ) . The Hsp network contains clones with similarity to perlucin ( a biomineralization-associated protein ) [60] and the M . edulis polyphenolic adhesive protein [61] , among others ( Table 1 in Protocol S1 ) ; it is curious that in this set , there is also a clone highly similar to the M . edulis gene for endo-1 , 4-mannanase , discussed above . Remarkably , 30/547 ( 5 . 5% ) of differentially expressed genes are found to be co-expressed with the Ankrd45 clone . This suggests that members of the Hsp class are involved in complex transcription patterns across multiple tissue types rather than a single one . Indeed , the closest co-expression neighbors of Mghsp90 are two cDNAs for calreticulin – a calcium-binding chaperone ( AJ624756/AJ625361 ) known to be associated with Hsp proteins [62] ( Figure 6 ) . Given the high-quality , yet limited data , the gene expression analysis outlined here strongly indicates that the known Hsp-associated genes in Mytilus are involved in intricate ways with each other , are possibly controlled by a small number of TFs over a number of tissues and conditions . It is thus possible that a mechanism for heat response might involve a ‘gene structure’ strategy , with few genes involved in a multitude of gene expression pathways . In this study , we have dissected computationally the upstream region of the Mghsp90 genes to investigate its structure and function . The structural complexity of this region strongly suggests that the transcription of Hsp90 stress response is tightly regulated via a dense region of heat shock elements and other regions of varying phylogenetic dispersion ( Figure 1 ) . Compared to other model organisms , such as C . elegans and D . melanogaster , this regulation appears to be achieved through a ‘gene structure’ strategy , i . e . a complex gene structure . In addition , expression analysis of the heat shock response indicates that a handful of key molecules belonging to the heat-shock class , exhibit a differential tissue-specific expression profile , possibly in gills and the digestive gland , while at the same time maintaining a multitude of associations through a complex co-expression network ( Figure 5 ) . Our results are consistent with current knowledge about chaperone function both within molecular [63]–[64] and ecological contexts [65]–[66] , and demonstrate the efficacy of both comparative genomics and systems biology for the elucidation of complex relationships between genotype , environment and phenotype . The nature of sessile animals , with the Mytilus genus as a model organism , can shed light into their metabolic capabilities [67] , stress responses [68] and resilience of evolutionary extinction [69] . The stress response for sessile animals is of particular interest , especially in cases where different ecological niches can be compared for close relatives , e . g . different growth potential in varying hydrostatic pressure or temperature [70] . Heat shock proteins in particular are used as indicators of thermal stress [68]; for instance , in the case of marine snails ( Tegula genus ) , the time course and magnitude of the heat shock response was measured in a field study by monitoring the synthesis of heat shock proteins [71] . In another field study on the Oregon coast , M . californianus and its predator Pisaster ochraceus were examined for production of the Hsp70 heat shock proteins; it was found that while mussels ( a sessile species ) have an increased production of Hsp70 , its starfish predators ( a mobile species ) do not , potentially exhibiting decreased heat shock adaptation compared to their prey [72] . Sessile marine invertebrates have been studied in the context of rising sea temperatures , including M . edulis [73] and Rhopaloeides odorabile , a common Great Barrier Reef sponge [74] . In the future , the thermal ecology of stress response can potentially inform policy decisions for environmental management in the context of climate change [75] – including the analysis of biogeographical range shifts [76] , particularly important for sessile animals , the understanding of complex prey-predator interactions e . g . the above mentioned pair of P . ochraceus and M . californianus [77] , and instigate a more integrated approach that will eventually include both weather records and niche-level measurements [78] . Currently , more established approaches for the use of Mytilus relate to its use as a biosensor system for the environmental monitoring of coastal water pollution [79] , heavy metals or organic pollutants [80] – including manufactured substances such as fiberglass [81] . In conclusion , this work forms a basis upon which the stress response in Mytilus will be better understood at the molecular level .
The M . galloprovincialis Hsp90 sequence was analyzed as previously [11] . The Hsp90 upstream regions from three other representative animal species , namely Caenorhabditis elegans , Drosophila melanogaster and Homo sapiens , were analyzed in a similar fashion . Previously published data concerning Hsp90 genes from these species were also taken into consideration for annotation purposes [82]–[88] . Sequence database searches were performed by BLAST ( for nucleotide sequences , blastn , version 2 . 2 . 22 ) [89] . Sequence alignments were computed using ClustalW [90] and visualized by JalView [91] . Regulatory elements , in the 5′ non-coding regions of the Mghsp90 genes were identified with Alibaba2 [92] , P-MATCH [93] and the Transcription Element Search System ( TESS ) [94] . An extensive comparative analysis for p53 binding sites was performed using the Matrix Search analysis tool of the TRED database [95] , scanning query sequences against the p53-specific sequence matrix ( cut-off score 2 ) from the JASPAR collection [96] . Data were obtained from the Gene Expression Omnibus ( GEO ) database , representing one spotted cDNA array ( Accession number: GSE2176 ) for gene expression in normal mussels ( M . galloprovincialis ) , using the MyArray 1 . 0 platform targeting 1712 clones with a total of 3840 cDNA sequences [42] . This dataset encompasses the total RNA isolated from gills ( n = 2 ) , gonads ( n = 2 ) , foot ( n = 2 ) , and digestive gland ( n = 2 ) [42] . Data normalization was performed by taking the binary logarithm ( log2 ) of normalized intensities ( defined as test signal/reference signal ) . Normalized data ( ‘original’ network ) was subsequently subjected to statistical validation . Gene associations were identified by computing the Pearson Correlation Coefficient ( PCC ) for all gene pairs in the raw network . Gene pairs with positive correlations indicated by a PCC>0 . 90 were considered to be co-expressed . Co-expression patterns were represented as networks where each node corresponds to a unique gene and each edge represents a co-expression association . The final network ( ‘inferred’ network ) was clustered using the Markov Clustering Algorithm ( MCL ) in order to both filter noisy associations and identify biologically meaningful clusters ( ‘clustered’ network ) , as previously described [97] . The inflation parameter for MCL was set to 3 . 0 . Only clusters with >10 genes were further analyzed for biologically meaningful associations . Differential expression analysis was performed by applying Analysis of Variance ( ANOVA ) to all genes across four distinct tissue types . Only those genes with the overall p-value bellow 0 . 05 were considered as differentially expressed . Two-way unsupervised hierarchical clustering of differentially expressed gene signals was performed using Euclidean distance as a similarity measure . Principal component analysis ( PCA ) was also performed to confirm the validity of the analysis for the four tissue-specific datasets . All statistical analyses were performed with MATLAB ( The MathWorks , Natick , MA – www . mathworks . com ) .
|
Adaptation of sessile animals , such as molluscs , to stress is achieved by a number of molecular mechanisms , few of which are clearly understood . Insights from this research can provide clues about stress tolerance both for sessile and mobile organisms . The Mediterranean mussel , of the genus Mytilus , is a model organism for the study of stress at the molecular level , with sufficient gene structure and function data available . We have thus investigated a key stress response gene , Hsp90 , and in particular its upstream region , using a combination of sequence and expression analysis approaches . We demonstrate that this region , responsible for the regulation of heat shock-associated gene expression , exhibits an unparalleled structural and functional complexity compared to other model organisms , as well as subtle gene expression patterns across multiple tissues . These results form the basis upon which the heat shock response can be used as a molecular biosensor for environmental monitoring in the future .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] |
[
"genetics",
"and",
"genomics/comparative",
"genomics",
"computational",
"biology/sequence",
"motif",
"analysis",
"computational",
"biology/transcriptional",
"regulation",
"genetics",
"and",
"genomics/gene",
"expression",
"physiology/genomics",
"computational",
"biology/comparative",
"sequence",
"analysis",
"genetics",
"and",
"genomics/bioinformatics",
"genetics",
"and",
"genomics/disease",
"models",
"marine",
"and",
"aquatic",
"sciences/theoretical",
"biology",
"computational",
"biology/signaling",
"networks",
"oncology/hematological",
"malignancies",
"computational",
"biology/systems",
"biology"
] |
2010
|
Promoter Complexity and Tissue-Specific Expression of Stress Response Components in Mytilus galloprovincialis, a Sessile Marine Invertebrate Species
|
Retrotransposition of endogenous retroviruses ( ERVs ) poses a substantial threat to genome stability . Transcriptional silencing of a subset of these parasitic elements in early mouse embryonic and germ cell development is dependent upon the lysine methyltransferase SETDB1 , which deposits H3K9 trimethylation ( H3K9me3 ) and the co-repressor KAP1 , which binds SETDB1 when SUMOylated . Here we identified the transcription co-factor hnRNP K as a novel binding partner of the SETDB1/KAP1 complex in mouse embryonic stem cells ( mESCs ) and show that hnRNP K is required for ERV silencing . RNAi-mediated knockdown of hnRNP K led to depletion of H3K9me3 at ERVs , concomitant with de-repression of proviral reporter constructs and specific ERV subfamilies , as well as a cohort of germline-specific genes directly targeted by SETDB1 . While hnRNP K recruitment to ERVs is dependent upon KAP1 , SETDB1 binding at these elements requires hnRNP K . Furthermore , an intact SUMO conjugation pathway is necessary for SETDB1 recruitment to proviral chromatin and depletion of hnRNP K resulted in reduced SUMOylation at ERVs . Taken together , these findings reveal a novel regulatory hierarchy governing SETDB1 recruitment and in turn , transcriptional silencing in mESCs .
Long terminal repeat ( LTR ) retrotransposons , also called endogenous retroviruses ( ERVs ) , are the relics of ancient and more recent germline retroviral integrations , comprising ~8–10% of the mouse and human genomes , respectively [1] . De novo retrotransposition of these parasitic elements is responsible for ~10% of spontaneous mutations in mice [2] . Among the remaining transcriptionally competent ERVs in the mouse genome , many class I Moloney murine leukemia virus ( MLV ) and class II intracisternal A-type particle ( IAP ) and MusD elements are transiently expressed and subsequently silenced in the early embryo [3] . Distinct epigenetic mechanisms cooperate to maintain ERV silencing including DNA methylation , covalent histone modifications , chromatin remodelling and non-coding RNAs [4] . Although DNA methylation suppresses ERV transcription in differentiated somatic cells [5] , pluripotent stem cell lines derived from the inner cell mass of the blastocyst , such as murine embryonic stem cells ( mESCs ) utilize a DNA methylation-independent pathway to maintain ERV silencing [6] . Key effectors in this silencing pathway are the conserved Krüppel-associated box zinc finger proteins ( KRAB-ZFPs ) , the largest family of C2H2 zinc finger transcription factors in vertebrate genomes [7] . Earlier experiments utilizing the MLV-based retroviral vectors harbouring a proline tRNA primer binding site ( PBSPro ) revealed that KRAB-ZFPs bind to specific proviral sequences such as the PBS , to direct the recruitment of a large silencing complex that includes the obligate co-repressor KAP1 ( also called TRIM28/TIF1β ) [8 , 9] and the lysine methyltransferase SETDB1 ( also called ESET/KMT1E ) , which deposits H3K9me3 to maintain a repressive chromatin state [10 , 11] . Interestingly , the KRAB-ZFP/KAP1 pathway also functions to protect the human genome against retroviral activity [12] , indicating that this silencing pathway is conserved in primates . Although prototypical KRAB-ZFP candidates for this pathway have been identified , such as ZFP809 and ZFP819 [9 , 13] , it remains unclear whether PBS binding is a general property of most KRAB-ZFPs or only a select few . Consistent with observations that PBS sequences alone are insufficient to confer SETDB1/KAP1-mediated silencing [14] , the transcription factor YY1 was shown to be required for silencing of the newly integrated MLV-based retroviruses in F9 embryonal carcinoma cells and mESCs [13] , revealing that additional sequence-specific factors may collaborate with KRAB-ZFPs . In addition , KAP1 is apparently recruited to IAP elements via sequences in the 5’UTR downstream of the PBS [14] . In mESCs but not embryonic fibroblasts , both class I and II ERVs and newly integrated MLV-based retroviral vectors are marked with H3K9me3 by a SETDB1/KAP1-containing complex [11] . During DNA methylation reprogramming in E13 . 5 primordial germ cells ( PGCs ) ERVs are also marked by H3K9me3 and are silenced in a SETDB1-dependent manner [15] . Conditional knockout of Setdb1 in undifferentiated mESCs or E13 . 5 PGCs abolishes H3K9me3 at ERVs and leads to reduced levels of DNA methylation and increased 5-hydroxymethylation [16] , concomitant with pervasive de-repression of distinct class I and II ERV families including MLV , IAP , MMERVK10C and MusD elements [11 , 15 , 17] . A similar phenotype is apparent upon deletion of Kap1 in mESCs [14] . Indeed , KAP1 is required for SETDB1 recruitment , since depletion of KAP1 leads to a loss of SETDB1 binding and H3K9me3 at ERVs and newly integrated MLV-based vectors [11 , 14] . The Small ubiquitin-like modifier ( SUMO ) paralogue SUMO1 is conjugated to KAP1 via the autocatalytic SUMO E3 ligase activity of the plant homeodomain ( PHD ) zinc finger towards the bromodomain at the major lysine acceptor sites K554 , K779 and K804 to direct SETDB1 recruitment and H3K9 methylation [18 , 19] . However , the role of SUMOylation in SETDB1-mediated repression of ERVs and the involvement of additional factors in SUMO-dependent SETDB1 targeting have not been addressed . Here , we identified the RNA-binding protein and transcription co-factor heterogeneous nuclear ribonucleoprotein K ( hnRNP K ) as a novel binding partner of the SETDB1/KAP1 complex in mESCs . Depletion of hnRNP K in these cells leads to a reduction of H3K9me3 and de-repression of class I and II ERVs , proviral reporter constructs and a cohort of germline-specific genes targeted by SETDB1 . Strikingly , hnRNP K is required for SETDB1 but not KAP1 recruitment through its influence on SUMOylation levels at ERV chromatin . Taken together , our data reveal a novel RNA-independent role for hnRNP K in regulating recruitment of SETDB1 to KAP1-bound targets and in turn H3K9me3-dependent transcriptional repression in mESCs .
To identify novel factors involved in SETDB1-dependent transcriptional repression , we characterized endogenous SETDB1-containing complexes from mESCs by immunoprecipitation ( IP ) and mass spectrometry ( MS ) , utilizing conditions that minimize de-SUMOylation of proteins given the SUMO-dependent interactions between SETDB1 and KAP1 [18] . Indeed , the Sentrin/SUMO-specific proteases SENP1 and SENP7 can de-SUMOylate KAP1 [20 , 21] and are expressed in mESCs [22] . To enrich for candidate SUMO-dependent binding partners of SETDB1 , we performed an anionic exchange step which efficiently depleted SENP1 followed by IP of endogenous SETDB1 with a specific N-terminal antibody [23] ( Fig . 1A and 1B ) . MS analysis revealed the specific enrichment of KAP1 along with the previously described SETDB1 co-factor MCAF1 ( also called mAM/ATF7IP ) ( Table 1 ) , which directly interacts with SETDB1 independent of SUMOylation [18 , 24] . Detection of MCAF1 and KAP1 supports the validity of this approach to identify candidate SUMO-independent and SUMO-dependent binding partners . MS analysis of a SETDB1 IP without prior SENP depletion identified a different set of polypeptides associated with SETDB1 ( S1A Fig . ) . While MCAF1 was identified in this direct IP approach , KAP1 was not ( S1A Fig . ) , indicating that the presence of SENPs in mESC nuclear extracts precludes the association of SETDB1 with its SUMO-dependent binding partners , including KAP1 . Among the novel SETDB1-associated proteins detected in the SENP-depleted but not the direct IP , we chose to focus on hnRNP K ( Table 1 ) , a ubiquitously expressed protein that functions as a DNA/RNA-binding transcriptional co-activator or co-repressor [25] . Notably , Hnrnpk is highly expressed in the inner cell mass and in mESCs relative to earlier stages of development in the preimplantation embryo [22] and was previously reported to directly interact with the KRAB-ZFPs Zik1 and Kid1 [25 , 26] . We further validated the interaction between hnRNP K and SETDB1 in mESCs using a combination of co-IP , immunostaining and co-sedimentation assays . Both KAP1 and hnRNP K were detected in FLAG-tagged SETDB1 complexes immunopurified from mESCs in the presence of the cysteine protease inhibitor N-ethylmaleimide ( NEM ) , which blocks SENP activity [27] ( Fig . 1C ) . Moreover , using a specific antibody raised against an internal epitope of SETDB1 [28] , both KAP1 and hnRNP K co-precipitated with SETDB1 from mESC nuclear extract only in the presence of NEM ( Fig . 1D ) . The association of hnRNP K and SETDB1 was also apparent by immunostaining , which revealed that hnRNP K and SETDB1 colocalize in the nucleus and to a lesser extent the cytoplasm of mESCs upon a short incubation with NEM ( S1B Fig . ) . Reciprocally , SETDB1 co-precipitated with both KAP1 and hnRNP K in the presence of NEM and hnRNP K and KAP1 also co-precipitated with each other ( Fig . 1E ) , indicating that these proteins are present in a single complex . Notably , the IP of KAP1 was clearly more efficient in the presence of NEM ( Fig . 1E ) , revealing that SENP inhibition may stabilize KAP1 oligomeric state , as KAP1 is known to form oligomers [29] . In addition , although hnRNP K binds to both DNA and RNA sequences [30 , 31] , the interaction between SETDB1 and hnRNP K was not perturbed in the presence of RNAse A and DNase I ( S1C Fig . ) indicating that it is not dependent upon nucleic acid . Consistent with the finding that the KAP1 IP was more efficient in the presence of NEM ( Fig . 1E ) , sucrose gradient ultracentrifugation of mESC nuclear extracts revealed that SENP inhibition promotes the stability of SETDB1/KAP1/hnRNP K complexes , which migrated at higher density compared with the profile of purified GST-KAP1 and GST-hnRNP K ( S1D–S1E Fig . ) . Although most of the hnRNP K remained uncomplexed with SETDB1/KAP1 , a fraction of total nuclear hnRNP K clearly co-sedimented with SETDB1 and KAP1 at a higher density in fractions 9–11 in the presence of NEM ( S1E Fig . ) , compared with GST-hnRNP K in fraction 5 ( S1D Fig . ) . Together these results confirm that hnRNP K is associated with the SETDB1/KAP1 complex in mESCs . To determine whether hnRNP K directly binds to SETDB1 , we performed GST pulldown assays with recombinant SETDB1 or Ubc9 as a positive control protein for hnRNP K [32 , 33] . Although KAP1 is SUMOylated in the SETDB1 complex under standard tissue culture conditions [18] , hnRNP K was also reported to be SUMOylated but only following DNA damage [32 , 33] . Indeed , whereas SETDB1 complexes from mESCs contained both SUMOylated and unmodified KAP1 , we found no evidence of SUMOylated hnRNP K ( S2A Fig . ) and thus used unmodified hnRNP K in subsequent pulldown assays . In contrast with Ubc9 , which bound to SUMO2 and hnRNP K , SETDB1 bound to SUMO2 but not hnRNP K ( Fig . 2A ) . In addition , no interaction was detected between FLAG-tagged SETDB1 and T7-tagged hnRNP K upon co-expression and FLAG IP from 293T cells ( S2B Fig . ) . Together these data indicated that hnRNP K does not directly interact with SETDB1 . To determine whether hnRNP K directly binds SUMOylated and/or unmodified KAP1 , we prepared in vitro SUMO1-conjugated GST-tagged KAP1 , GST-p53 as a positive control binding partner of hnRNP K [33] , or a GST-tagged fragment of RanGAP1 as a model SUMO1 substrate for GST pulldown assays with recombinant SETDB1 or hnRNP K baits . Using purified SUMOylation cascade components , we achieved efficient mono-SUMOylation of RanGAP1 at K526 [34] and mono- , di- , tri- and tetra-SUMOylation of KAP1 ( Fig . 2B ) at its major SUMO acceptor lysines including K554 , K676 , K779 and K804 [18 , 19] . p53 was mono-SUMOylated at K386 [35] although this was less efficient in the absence of a SUMO E3 ligase ( Fig . 2B ) . While SETDB1 directly bound to p53 independently of SUMOylation , it bound to KAP1 in a SUMO1-dependent manner ( Fig . 2C ) , consistent with previous observations [18] . HnRNP K binding to p53 was enhanced by SUMOylation but surprisingly , its binding to KAP1 was decreased upon SUMOylation ( Fig . 2D ) . KAP1 harbours several functional domains that participate in protein-protein interactions , including an N-terminal RING-B-box-coiled-coil ( RBCC ) domain , which mediates binding to KRAB-ZFPs and other proteins [36–38] , a proline-x-valine-x-leucine ( PxVxL ) motif , which binds to HP1 proteins [39] and a C-terminal PHD finger-bromodomain that binds to Ubc9 and chromatin-modifying factors , including SETDB1 and CHD3 upon SUMOylation [10 , 18 , 40] . While hnRNP K bound to wt full-length KAP1 , it failed to bind to the deletion fragments containing only the PxVxL motif or only the PHD finger-bromodomain ( Fig . 2E ) , revealing that hnRNP K binding requires the N-terminal RBCC domain . Taken together , these observations indicate that hnRNP K and SETDB1 indirectly interact with each other via their binding to unmodified or SUMOylated KAP1 subunits . Consistent with this model , SETDB1 complexes in mESCs contain both unmodified and SUMOylated KAP1 ( S2A Fig . ) , despite SETDB1 exhibiting binding affinity for only SUMOylated KAP1 ( Fig . 2C ) . Furthermore , the interactions between hnRNP K and KAP1 in mESCs were unperturbed upon depletion of SETDB1 ( S2C Fig . ) , confirming that they interact in a SETDB1-independent manner . In addition , endogenous KAP1 also co-precipitated with hnRNP K from 293T cell extracts ( S2D Fig . ) , indicating that this interaction is not limited to mESCs . Finally , the observation that hnRNP K colocalized with KAP1 throughout the nucleus in mESCs in the absence of SENP inhibitor ( S2E Fig . ) is consistent with the model that they form complexes in the absence of KAP1 SUMOylation in cells . We next investigated whether loss of hnRNP K compromises SETDB1-dependent transcriptional silencing of ERVs in mESCs . Using siRNA-mediated knockdown ( KD ) , hnRNP K was efficiently depleted at the protein level by 24 h post-transfection ( S3A Fig . ) . Notably , KD of hnRNP K in mESCs significantly reduced their proliferation by 72 h post-transfection ( S3B Fig . ) . However , there was no gross effect on cell cycle distribution at this time-point and only minimal effects on expression of the pluripotency marker SSEA1 ( S3C–S3D Fig . ) . Furthermore , reduced proliferation was not associated with induction of apoptosis , as determined by Annexin V staining ( S3D Fig . ) . Thus hnRNP K KD does not result in overt differentiation or apoptosis at this time-point . To determine the influence of hnRNP K depletion on proviral silencing , we used previously established proviral GFP reporter mESC lines , including the murine stem cell virus bearing a glutamine tRNA PBS ( MSCV-PBSGln ) GFP line [11] and the HA36 mESC line , which harbours a silent IAP LTR-PBS-5’UTR region driving GFP transgene integrated into a defined genomic locus [41] ( Fig . 3A ) . In both lines , proviral silencing is dependent upon H3K9me3 deposited by the SETDB1/KAP1 complex [11 , 41] . Transfection of siRNAs specific for Setdb1 or Hnrnpk effectively reduced expression of the relevant mRNAs to ~10–25% of the control siRNA-transfected cells ( Fig . 3B ) . While only ~2–3% of SSEA1+ cells were also GFP+ in the untransfected ( MSCV and IAP ) and siRNA transfected controls , KD of Setdb1 de-repressed both reporters , resulting in ~37% and ~20% SSEA1+; GFP+ cells , respectively ( Fig . 3C ) . Strikingly , KD of Hnrnpk also consistently de-repressed both the MSCV and IAP reporters , resulting in an average of ~29% and ~20% SSEA1+; GFP+ cells , respectively ( Fig . 3C ) . We also interrogated the role of the SETDB1 co-factor MCAF1 , which facilitates conversion of H3K9me2 to H3K9me3 by SETDB1 [24] . Interestingly , Setdb1 KD cells showed a ~4-fold upregulation of Mcaf1 expression ( S4A Fig . ) , indicating that the level of Mcaf1 expression is sensitive to the level of SETDB1 . The MSCV proviral reporter was also de-repressed in Mcaf1 KD cells , ( S4A–S4B Fig . ) , revealing that this catalytic co-factor of SETDB1 also plays a role in proviral silencing . We next determined whether KD of hnRNP K disrupts silencing of ERVs . In contrast to the proviral reporter lines , KD of SETDB1 or hnRNP K in TT2 wt mESCs resulted in only modest de-repression ( ~2-fold ) of class I and II ERVs by 72 h post-transfection ( S4C–S4D Fig . ) , with the exception of robust induction of MMERVK10C elements in SETDB1 KD cells , despite efficient depletion of the protein ( S4C Fig . ) . Surprisingly , class III MERVL elements , which are repressed by KAP1 in a SETDB1-independent manner [42] , were strongly induced in hnRNP K KD cells ( S4D Fig . ) . We have shown previously that DNA methylation also plays a role in transcriptional repression of ERVs , particularly of IAP elements , in mESCs cultured in serum [17] . To preclude the influence of DNA methylation , we knocked down Setdb1 or Hnrnpk in Dnmt3a; Dnmt3b; Dnmt1 triple KO ( Dnmt TKO ) mESCs [43] ( Fig . 3D ) , which are devoid of DNA methylation but maintain SETDB1 binding and H3K9me3 at ERVs [11] and thus solely rely on the SETDB1/H3K9me3 pathway for silencing of these elements . The absence of DNA methylation alone did not perturb silencing of MLV , MMERVK10C and MusD elements , but yielded a ~6-fold upregulation of IAP elements ( Fig . 3E ) , consistent with the finding that IAP elements are modestly upregulated in Dnmt TKO cells [11] . In contrast , KD of Setdb1 expression resulted in a substantial induction of ERVs in these cells ( Fig . 3E ) . These ERVs were also de-repressed upon Hnrnpk KD , with IAP elements showing an increase in expression of ~18-fold , ~3-fold greater than the control KD in the Dnmt TKO line ( Fig . 3E ) . Depletion of Mcaf1 in the Dnmt TKO cells ( Fig . 3D ) also resulted in upregulation of class I and II ERVs beyond what was observed in the Dnmt TKO line alone ( S4E Fig . ) . Importantly , although IAP elements are strongly induced in DNA methylation-deficient , differentiated Dnmt1-/-; Oct4-negative mESCs [6] , levels of Oct4 mRNA was not appreciably reduced in any of these KD cultures compared to the control siRNA Dnmt TKO cells ( S4F Fig . ) , indicating that ERVs were not induced as a secondary consequence of an increase in the number of differentiated cells in culture . Taken together , these results reveal that depletion of hnRNP K disrupts SETDB1/H3K9me3-mediated silencing of ERVs in mESCs . To investigate whether depletion of hnRNP K disrupts SETDB1-dependent repression of genes , we performed mRNA-seq from two biological replicates of TT2 mESCs transfected with control or Hnrnpk siRNA ( S4C Fig . ) . A total of 290 genes were consistently misregulated upon hnRNP K KD , 264 genes were upregulated ≥ 2-fold in both KD lines while only 26 were downregulated by ≥50% ( Fig . 4A and S1 Table ) . Gene ontology ( GO ) analysis revealed that the upregulated genes were enriched for “apoptosis” ( S5A Fig . ) indicating that although these KD cells do not show high levels of Annexin V staining at this time-point , their progressive proliferation block ( S3B Fig . ) may coincide with induction of the apoptotic pathway . Although hnRNP K regulates the expression of pro-apoptotic genes Bcl-Xs and Bik under certain conditions [44] , these genes were not upregulated in hnRNP K KD mESCs . Nevertheless , Btg2 , Anxa8 , Perp , Trp73 , Cdkn1a and Casp14 were among the 16 apoptosis-associated genes identified by GO analysis . In addition there was an enrichment of genes involved in “lung and respiratory system development” ( S5A Fig . ) including the primitive endoderm and mesoderm lineage transcription factors Gata6 , Gata3 , Tbx3 , Tbx20 , Foxa1 , Nkx2–9 and Nkx2–2 ( S1 Table ) . Previous ChIP-seq data indicates that these transcription factor genes harbour the bivalent chromatin state of H3K4me3 and H3K27me3 [45–47] and are subject to polycomb repressive complex 2 ( PRC2 ) -mediated silencing [48] . The de-repression of Gata6 and Gata3 , which were upregulated ~15-fold and ~10-fold in hnRNP K KD cells , respectively ( Fig . 4B ) , indicates that hnRNP K KD could eventually lead to a loss of pluripotency , since the overexpression of these transcription factors is sufficient to drive endoderm lineage differentiation [49 , 50] . RNA-seq analysis of ERVs in the TT2 hnRNP K KD cells ( S1 Table ) generally confirmed our qRT-PCR analysis from the same cells ( S4D Fig . ) in that class I and II elements were only modestly de-repressed ( ≤2-fold ) while MERVL elements were strongly de-repressed ( ≥14-fold ) . We next compared the list of genes upregulated in hnRNP K KD mESCs to our list of upregulated genes in Setdb1 KO mESCs [17] , which revealed 54 genes in common ( S2 Table ) . We previously identified a cohort of 33 germline lineage genes that are directly repressed by SETDB1-dependent H3K9me3 and DNA methylation [17] . Notably , many of these direct SETDB1 target genes were consistently upregulated >2-fold in hnRNP K KD cells ( 15 of these genes are shown in S5B Fig . ) For example , the promoter of the male germline gene Dazl harbours a peak of SETDB1 binding and SETDB1-dependent H3K9me3 and is upregulated in both Setdb1 KO and hnRNP K KD cells ( Fig . 4C ) . In addition , of the 134 SETDB1-bound genes that are upregulated in Setdb1 KO mESCs [17] , 30 were consistently de-repressed in hnRNP K KD cells ( Fig . 4D and S2 Table ) . Quantitative RT-PCR analysis of a subset of these genes , including the male germline-specific genes Dazl , Fkbp6 , Mael and Taf7l confirmed that they are indeed upregulated in hnRNP K KD cells ( Fig . 4B ) . Furthermore , levels of H3K9me3 at the promoters of these genes were reduced in hnRNP K KD cells to a similar extent as in SETDB1 KD cells ( Fig . 4E ) . A comparison of the genes upregulated in hnRNP K KD and Kap1 KO cells [14] also revealed a significant overlap ( S5C Fig . ) and included lineage-restricted genes such as Gata6 , Arg2 and Dkk1 ( S2 Table ) . We identified 33 genes that are commonly de-repressed in Setdb1 KO , Kap1 KO and hnRNP K KD mESCs , several of which were direct SETDB1 targets and were expressed in a lineage-dependent fashion , including the imprinted gene Igf2 and liver-specific gene Cml2 ( S2 Table ) . The promoter of Cml2 lies immediately downstream of an intact ETn family retroelement that is bound by SETDB1 and marked by SETDB1-dependent H3K9me3 , which spreads into the Cml2 promoter ( S5D Fig . ) indicating that this gene is silenced by the spreading of H3K9me3 from the intact ERV . In conclusion , these results support a role for hnRNP K in transcriptional repression of genes regulated by SETDB1 and KAP1 as well as PRC2 , the latter via an undefined pathway . We next determined whether hnRNP K is bound at de-repressed ERVs . Since the presence of NEM increased the sensitivity of KAP1 and SUMO1 chromatin immunoprecipitation ( ChIP ) , improving the enrichment of both at MLV and IAP 5’ LTRs where KAP1 binding is high , but not at MERVL 5’ LTRs where KAP1 binding is low [42] ( S6A Fig . ) , we performed subsequent ChIP assays in the presence of NEM to preclude a refractory effect of SENP activity on the binding of these factors at ERV 5’LTRs and other loci ( Fig . 5A ) . Under these conditions , hnRNP K was enriched at the promoters of the SETDB1-bound , H3K9me3-marked germline genes Fkbp6 , Dazl , Mael and Taf7l , with the highest level of enrichment detected at Mael ( Fig . 5B ) , indicating that these loci are direct targets of hnRNP K in mESCs . Relative to the germline gene promoters , the 5’LTRs of class I and II ERVs showed lower enrichment of hnRNP K , with ETn/MusD and MLV elements showing the highest and lowest levels , respectively ( Fig . 5B ) . Importantly , the signal at ERVs and the germline gene promoters was specific , since it was reduced upon hnRNP K KD . In contrast , there was no enrichment of hnRNP K at the Egr1 promoter ( Fig . 5B ) , which is active in mESCs and was shown to be bound by hnRNP K only upon serum stimulation in the HCT116 colon cancer cell line [51 , 52] . Furthermore , RNAse did not perturb hnRNP K enrichment at ERVs ( S6B Fig . ) , indicating that hnRNP K is recruited to class I and II ERVs in an RNA-independent manner . We next determined whether hnRNP K is required for SETDB1-dependent H3K9me3 deposition at ERVs . As shown previously [11] , SETDB1 KD resulted in depletion of H3K9me3 at MLV , IAP , MMERVK10C and ETn/MusD 5’LTRs ( Fig . 5C ) . Strikingly , this effect was phenocopied upon KD of hnRNP K ( Fig . 5C ) . Importantly , the reduction of H3K9me3 at ERVs was apparent in hnRNP K KD cells as early as 48 h post-transfection , similar to the kinetics of H3K9me3 perturbation in SETDB1 KD cells ( S6C Fig . ) , indicating that this phenotype is not a secondary consequence of the loss of proliferation that commences at ~72 h post-transfection in hnRNP K KD cells ( S3B Fig . ) . Furthermore , as early as 24 h after hnRNP K KD , there was a clear reduction of H3K9me3 at the MSCV 5’LTR-PBS and Gfp regions ( Fig . 5D and 5E ) . The levels of H4K20me3 , a mark deposited by SUV420H1/2 enzymes in a SETDB1/H3K9me3-dependent manner [11] , were also reduced at the MSCV provirus in hnRNP K-depleted cells ( Fig . 5F ) . H3K9me3 was dramatically reduced at both ERVs and the MSCV proviral reporter in MCAF1 KD cells ( S6C–S6D Fig . ) , consistent with its role as a catalytic co-factor of SETDB1 [24] . Importantly , siRNA-mediated depletion of hnRNP K or SETDB1 did not affect global H3K9me2 or H3K9me3 levels ( S6E Fig . ) , indicating that the effect of hnRNP K depletion on H3K9me3 at ERVs is not the result of a general reduction of H3K9me2/3 . Thus we concluded that hnRNP K is required for SETDB1-dependent H3K9me3 deposition at proviral chromatin . To determine whether hnRNP K is required for SETDB1 recruitment , we next conducted ChIP analysis of SETDB1 in cells depleted of hnRNP K ( Fig . 6A and 6B . A reduction of SETDB1 enrichment was apparent at all class I and II ERV LTRs in SETDB1 KD cells , confirming the specificity of our antibody ( Fig . 6B ) . Strikingly , KD of hnRNP K also reduced the level of SETDB1 enrichment at ERVs ( Fig . 6B ) , likely explaining the reduction of H3K9me3 observed at these loci following hnRNP K KD ( Fig . 5C ) . SETDB1 enrichment was also reduced at the MSCV 5’LTR-PBS and Gfp internal region upon depletion of hnRNP K ( S7A Fig . ) , revealing a link between loss of H3K9me3 , de-repression of the MSCV proviral reporter and reduced SETDB1 recruitment . In contrast , KD of MCAF1 did not perturb SETDB1 enrichment at ERVs or the MSCV provirus ( S7B Fig . ) . Importantly , neither SETDB1 nor KAP1 protein levels were reduced in hnRNP K KD cells ( Fig . 6A ) and SETDB1 was still localized to the nucleus in hnRNP K-depleted cells ( S7C Fig . ) . KAP1 is the only factor known to be required for SETDB1 recruitment to proviral chromatin [11 , 14] . While KAP1-depleted cells showed reduced levels of KAP1 enrichment at ERVs confirming antibody specificity , KD of hnRNP K did not affect KAP1 enrichment levels ( Fig . 6C and 6D ) . In contrast , KD of KAP1 substantially reduced hnRNP K enrichment at ERVs ( Fig . 6E ) . Taken together , these data reveal that hnRNP K is recruited in a KAP1-dependent manner and facilitates subsequent SETDB1 binding at proviral chromatin . Previous studies have shown that KAP1 SUMOylation is necessary for recruitment of SETDB1 and H3K9 methylation to promote silencing of heterologous promoters in transformed cell lines [18 , 19 , 53] . To determine whether a functional SUMOylation pathway is also necessary for SETDB1 recruitment to ERVs in pluripotent stem cells , we used either anacardic acid to inhibit SUMO E1 activating enzyme [54] or siRNAs to KD Ubc9 ( also called Ube2i ) in the MSCV-GFP cell line and assayed for de-repression of the proviral LTR by flow cytometry ( Fig . 7A ) . In accord with the inhibitory effect of anacardic acid on the activity of SUMO E1 activating enzyme Aos1/Uba2 and histone H3 acetyltransferases such as p300 [55] , this compound blocked both KAP1 SUMOylation and bulk histone H3 acetylation ( Fig . 7B ) . While , anacardic acid treatment did not affect bulk H3K9me3 ( Fig . 7B ) , it consistently de-repressed the proviral reporter in a dose-dependent manner , resulting in ~15% GFP+ cells at 100 μM ( Fig . 7C ) . Using siRNAs , we depleted Ubc9 mRNA to ~35% of the control ( Fig . 7D , inset graph ) . As Ubc9 is essential for early embryogenesis [56] , we monitored changes in MSCV expression at 48 h post siRNA transfection . KD of Ubc9 expression consistently de-repressed the proviral reporter resulting in an average of 23% GFP+ cells ( Fig . 7D ) . Notably , ChIP analysis revealed that SUMO1 levels at the MSCV 5’ LTR were dramatically reduced in Ubc9-depleted cells ( S8A Fig . ) indicating that the loss of SUMOylation on proviral chromatin correlates with de-repression . Furthermore , there was a reduction of SETDB1 enrichment at the MSCV provirus in Ubc9 KD cells ( Fig . 7E ) , confirming that SUMOylation of chromatin proteins associated with ERVs enhances SETDB1 recruitment . Strikingly , SUMO1 levels were greatly reduced at the 5’ LTRs of MLV , IAP , MMERVK10C and ETn/MusD elements by 24 h post-transfection of hnRNP K siRNAs ( Fig . 7F ) . Moreover , this effect persisted in hnRNP K KD cells at 72 h post-transfection both at ERVs and the MSCV provirus ( S8B–S8C Fig . ) , coinciding with the timeframe in which SETDB1 recruitment to proviral chromatin was compromised ( Fig . 6B ) . Although the loss of SUMOylation at ERV chromatin upon hnRNP K KD could be a consequence rather than a cause of reduced SETDB1 recruitment , KD of SETDB1 , which was sufficient to de-repress the MSCV LTR ( S8D Fig . ) , did not concomitantly attenuate SUMOylation on proviral chromatin ( S8E Fig . ) . Taken together these results are consistent with the model that hnRNP K is necessary for SUMOylation of proteins such as KAP1 on ERV chromatin , which is required for SETDB1 recruitment and in turn proviral silencing .
KRAB-ZFP/KAP1 complexes [9 , 57] are thought to play a central role in repression of ERV transcription in pluripotent stem cells via SETDB1 recruitment [11 , 14] . In this work , we have identified hnRNP K as a novel co-factor , which is required for recruitment of SETDB1 to proviral chromatin and in turn for efficient proviral silencing . HnRNP K is a highly conserved , multi-functional protein involved in transcription regulation , mRNA splicing and translation [25] . Studies in flies , yeast and in mammalian cell lines reveal that hnRNP K plays important roles in development and gene regulation [58 , 59] . HnRNP K was reported to directly interact with chromatin regulatory proteins , such as the PRC2 subunit EED [60] and KRAB-ZFPs Zik1 and Kid1 [25 , 26] , indicating that it may regulate Polycomb and/or KRAB-ZFP/KAP1 complexes . Our results reveal a role for hnRNP K in the KRAB-ZFP/KAP1-based silencing pathway acting on ERVs and retroviral vectors in pluripotent stem cells . Based on these findings , we propose a novel model for the SETDB1/KAP1 proviral silencing pathway incorporating hnRNP K ( Fig . 8 ) . In wt mESCs , KRAB-ZFPs recruit KAP1 in an oligomeric state , possibly as a homotrimer [29 , 36] , to proviral chromatin and unmodified KAP1 may recruit hnRNP K . HnRNP K may promote KAP1 SUMOylation on chromatin , which then serves as a ligand for the SETDB1/MCAF1 complex [18 , 61] , eliciting SETDB1-dependent H3K9me3 deposition at SUMOylated KAP1-bound regions ( Fig . 8A ) . In hnRNP K-deficient cells , SUMOylation of KAP1 on chromatin may be compromised , leading to reduced SETDB1 recruitment at ERVs , diminution of H3K9me3 and eventual transcriptional de-repression ( Fig . 8B ) . This model is consistent with recent ChIP-seq analyses of SUMO1 , SUMO2 and Ubc9 in human fibroblasts [62] , which show co-occupancy with sites of KAP1 , SETDB1 and H3K9me3 at the 3’ ends of KRAB-ZFP genes [63] indicating that these SETDB1/KAP1-bound , SUMOylated loci are sites of active KAP1 SUMOylation on chromatin [62] . Although a possible contraindication to this model is our observation of the differing binding affinities of SETDB1 and hnRNP K for SUMOylated KAP1 in vitro ( Fig . 2C and 2D ) , this could be rationalized by: 1 ) the existence of multiple KAP1 subunits in each complex such that some are SUMOylated while others are unmodified , providing binding sites for both SETDB1 and hnRNP K simultaneously , and/or 2 ) the observation that hnRNP K directly binds to certain KRAB-ZFPs [25 , 26] and therefore may still indirectly interact with SUMOylated KAP1 . Indeed , consistent with the former possibility , rather than solely containing SUMOylated KAP1 , we found that SETDB1 complexes contained predominantly unmodified KAP1 with only a minority SUMOylated KAP1 under conditions where we could preserve mono- and di-SUMOylated KAP1 in mESC nuclear extracts ( S2A Fig . ) . This observation indicates that SETDB1 binding to KRAB-ZFP/KAP1 complexes in vivo may only require a small proportion of the total KAP1 in the complex to be SUMOylated . KAP1 SUMOylation is highly dynamic and previous investigations have relied on overexpression of SUMO paralogues to detect it [19 , 20 , 38 , 64] . Therefore , it is also possible that hnRNP K facilitates transient KAP1 SUMOylation events in a cell cycle-dependent manner , such as during S-phase when chromatin modifications must be re-established . How hnRNP K promotes the SUMOylation of KAP1 remains to be determined , although given that hnRNP K is a SUMO target itself and can directly interact with Ubc9 [32 , 33] , it may facilitate recruitment of this SUMO E2 enzyme to KAP1-bound loci . Alternatively , hnRNP K might also counteract SENP activity toward KAP1 providing an additional layer of regulation over KAP1de-SUMOylation . Since KAP1 is constitutively phosphorylated at Ser824 in pluripotent stem cells [65] , another intriguing possibility is that hnRNP K counteracts the activity of the SUMO-targeted ubiquitin ligase RNF4 , which conjugates ubiquitin to Lys676 SUMOylated , Ser824 phosphorylated KAP1 promoting its degradation [64] . Similar to SETDB1 KD mESCs [17] , KD of hnRNP K only resulted in modest upregulation of class I and II ERVs in wt mESCs cultured in serum . A likely explanation for this observation is the relatively high level of DNA methylation in mESCs cultured in serum relative to two-inhibitor ( 2i ) media . Under the latter conditions , mESCs adopt a “naïve” hypomethylated state , more reflective of the inner cell mass of the E3 . 5 blastocyst [46] . Consistent with this model , depletion of hnRNP K in DNA methylation-deficient cells led to a more robust upregulation of class I and II ERVs as compared with wt cells and previous work has shown that IAP elements are synergistically upregulated upon KD of both SETDB1 and DNMT1 in serum-cultured mESCs [17] . Thus siRNA KDs in serum-cultured mESCs are likely not robust enough to elicit loss of DNA methylation at ERVs controlled by SETDB1 , despite losses of H3K9me3 . In contrast with ERVs , for reasons that are not entirely clear , knocking down SETDB1 in serum-cultured mESCs harbouring a newly integrated silent MSCV provirus results in losses of both H3K9me3 and DNA methylation at the 5’LTR and subsequent de-repression [11 , 41] . In addition to hnRNP K , we also identified a crucial role for MCAF1 in SETDB1-mediated proviral silencing , consistent with its role in enhancing SETDB1 catalytic activity towards H3K9me2 to generate H3K9me3 [24] . In contrast to hnRNP K- and KAP1-depleted cells , SETDB1 recruitment is maintained but H3K9me3 is no longer efficiently deposited at proviral chromatin in MCAF1-deficient mESCs ( Fig . 8C ) . This phenotype is consistent with the observation that the catalytic activity of SETDB1 is crucial for full ERV repression [11 , 41] and a previous report showing that the MCAF1 orthologue Windei is necessary for dSETDB1/Eggless function in the Drosophila germline [66] . Intriguingly , class III MERVL elements , which are silenced by H3K9me2 deposited by the lysine methyltransferases G9a/GLP [42] , were strongly induced in hnRNP K KD cells . Since these elements are also de-repressed in Kap1 KO but not Setdb1 KO mESCs [14 , 42] , hnRNP K may play a role in SETDB1-independent chromatin regulatory pathways with KAP1 and G9a/GLP . Further experiments are required to address whether hnRNP K has direct role in MERVL silencing in mESCs . In addition to ERVs , a cohort of SETDB1/H3K9me3-repressed male germline-specific genes [17] are bound at their promoters by hnRNP K , show reduced H3K9me3 and increased expression upon hnRNP K KD , indicating a role for hnRNP K in SETDB1/H3K9me3-mediated gene repression . How SETDB1 may be targeted to these promoters by hnRNP K remains unclear , since these genes are not upregulated in Kap1 KO cells [14] . One possibility is that hnRNP K promotes SUMOylation of proteins other than KAP1 on chromatin , leading to SETDB1 binding and transcriptional silencing . In addition to SETDB1/H3K9me3 , hnRNP K may also promote PRC2/H3K27me3-mediated gene repression in mESCs , since a cohort of PRC2 target genes , including Gata6 and Nkx2–9were strongly upregulated in hnRNP K KD cells . This is consistent with a previous report showing that hnRNP K can promote PRC2-dependent repression via recruitment of the subunit EED to a heterologous promoter [60] . Further studies will be necessary to clarify the contribution of hnRNP K to SETDB1- and PRC2-mediated transcriptional silencing at specific genes in mESCs . In conclusion , our results reveal novel mechanistic insights into the transcriptional silencing of class I and II LTR retrotransposons and genes by SETDB1/H3K9me3 in pluripotent stem cells . Notably , both hnRNP K and SETDB1 have been identified as bona fide oncogenes and are aberrantly overexpressed in a variety of human cancers including melanoma [67 , 68] , prostate carcinoma [69 , 70] and lung carcinoma [71 , 72] . A greater understanding of how hnRNP K regulates the recruitment of SETDB1 to promoters may ultimately provide new targets for anti-oncogenic therapeutics .
Mouse ES cell lines used in this study included: TT2 wt , TT2 33#6 Setdb1lox/-harbouring the randomly integrated silent MSCV ( PBSGln ) -GFP [11] , TT2 33#6 Setdb1lox/-expressing 3XFLAG-Setdb1 [11] , HA36 harboring the silent IAP LTR-GFP construct [41] , J1 wt and Dnmt3a-/-; Dnmt3b-/-; Dnmt1-/-[43] . All ES cell lines were cultured under standard feeder-free conditions on gelatinized tissue culture dishes in standard mESC media: DMEM high glucose containing 15% fetal bovine serum , 20 mM HEPES , 1 mM L-glutamine , 100 U/ml penicillin-streptomycin , 1 mM nonessential amino acids , 1 mM sodium pyruvate , recombinant LIF and 0 . 1 mM β-mercaptoethanol . HEK293T cells were cultured in DMEM high glucose containing 10% fetal bovine serum and 100 U/ml penicillin-streptomycin . All cell lines were cultured at 37°C with 5% CO2 . RNAi was performed essentially as described [41] using predesigned siRNA SMARTpools from Dharmacon ( ThermoFisher ) . Briefly , cells were seeded in ES media lacking antibiotics to achieve 70–80% confluence and 24 h later were transfected with 100 nM of SMARTpools for Setdb1 , Hnrnpk , Mcaf1 ( also called Atf7ip ) , Kap1 , Ubc9 or non-targeting siRNA #2 ( control siRNA ) using DharmaFECT Reagent #1 . A second round of transfection was performed 48 h later using 100 or 50 nM siRNAs . Plasmids were transfected into HEK293T cells in antibiotic-free media using lipofectamine 2000 ( Life Technologies ) and harvested 48 h post-transfection . For blocking SUMO E1 activity , anacardic acid ( Sigma-Aldrich ) diluted to 5–100 μM in DMSO was incubated with cells in complete ES media for 18 h prior to harvest . Indirect immunofluorescence staining was performed using standard methods . Cells were grown on coverslips or harvested by trypsinization were crosslinked with 4% formaldehyde , permeabilized with 0 . 25% triton-X-100 and blocked with 1% bovine serum albumin ( Sigma-Aldrich ) . Cells were then incubated with anti-SETDB1 H300 ( Santa Cruz Biotechnology sc-66884 ) , anti-hnRNP K 3C2 ( Abcam 39975 ) , anti-hnRNP K ( Abcam 70492 ) or anti-KAP1 20C1 ( Abcam 22553 ) at 37°C for 1 h or overnight at 4°C and subsequently incubated with Alexa Fluor 488 and 594-labeled secondary antibodies ( Life Technologies ) . DNA was counterstained with Hoescht 33342 ( Sigma-Aldrich ) . Flow cytometry analysis of GFP-fluorescing cells was performed as previously described [41] . Briefly , cells were resuspended in 0 . 5 μg/ml propidium iodide ( Sigma-Aldrich ) in FACS buffer ( phosphate buffered saline containing 3% fetal bovine serum ) and analyzed on a BD LSRII flow cytometer using BD FACS Diva software . Cells were successively gated on forward and side scatter , then PI- ( live cells ) and lastly GFP+ cells , using the untransfected mESC line ( either MSCV-GFP or IAP-GFP ) as a GFP- population to set the gates . SSEA1 and Annexin V staining were detected on mESCs using 1:400 anti-SSEA1 PE-conjugate ( BD Pharmigen ) or 1:1000 anti-Annexin V Alexa Fluor 488-conjugate ( Life Technologies ) . Where indicated , cells were gated for the SSEA1+ population prior to GFP gating to identify the SSEA1+; GFP+ ( double-positive ) population . Cell cycle analysis was performed according to standard methods where cells were harvested and fixed for >2 h in ice-cold 70% ethanol , permeabilized with 0 . 25% triton-X-100 and stained with 10 μg/ml propidium iodide ( Sigma-Aldrich ) . Cell cycle profiles were analyzed by the Dean-Jett-Fox Model using FlowJo software ( Tree Star ) . Nuclear extracts were prepared from mESCs as previously described [42] with or without 10 or 20 mM NEM and clarified by centrifugation . For immunoprecipitation of SETDB1 complexes after anionic column fractionation , approximately 12–15 mg of TT2 mESC nuclear extract ( 4 ml ) was prepared without NEM , diluted with 2 volumes with 56 mM HEPES pH 7 . 9 , 5% glycerol and passed over a 2 ml column of Macro HiQ anionic exchange media ( BioRad ) in an equilibration buffer ( 50 mM HEPES pH 7 . 9 , 100 mM KCl , 10% glycerol ) . Bound proteins were washed with 5 column volumes of equilibration buffer and then eluted stepwise in 2 volumes of buffer containing 250 mM KCl , then 2 volumes of buffer containing 500 mM KCl . The 500 mM KCl fraction containing SETDB1 and depleted of SENP1 ( 4 ml ) was then diluted with 2 volumes IP dilution buffer ( 20 mM HEPES pH 7 . 9 , 0 . 5% NP-40 , 10% glycerol containing 2 mM PMSF ) and divided into two equal aliquots and immunoprecipitated overnight at 4°C with protein G sepharose beads crosslinked with ~100 μg of rabbit IgG ( Sigma Aldrich ) or rabbit anti-SETDB1 H300 ( Santa Cruz Biotechnology ) using dimethylpimelimidate . Beads were washed extensively with a wash buffer ( 20 mM HEPES pH 7 . 9 , 200 mM KCl , 1% NP-40 , 0 . 1% sodium deoxycholate , 10% glycerol ) and eluted by boiling in SDS-PAGE loading buffer . For direct SETDB1 IP from mESC nuclear extract , ~7–8 mg of nuclear extract ( 1 . 5 ml ) was diluted with 2 volumes of IP dilution buffer as above and incubated with 30 μg rabbit IgG or anti-SETDB1 H300 overnight at 4°C . Immunocomplexes were captured on protein G dynabeads , washed extensively with wash buffer as described above except omitting deoxycholate , eluted with 0 . 1 M glycine pH 2 . 5 and neutralized with 1 . 5 M Tris pH 8 . 8 . Immunoprecipitated samples were analyzed by SDS-PAGE , western blot and silver staining . For mass spectrometry , IgG and SETDB1 IP samples were resolved by SDS-PAGE and stained with colloidal coomassie . The IgG heavy and light chain bands were removed first and discarded then the rest of each gel lane was excised and subjected to in-gel digestion [73] . Extracted peptides were then analyzed by nano-flow liquid chromatography-tandem mass spectrometry ( LC-MS/MS ) on a LTQ-Orbitrap Velos Pro mass spectrometer ( ThermoFisher ) [74] . Tandem mass spectra were searched against the UniProt mouse database using Mascot ( v2 . 4 , Matrix Science ) . Each IP sample was analyzed independently twice . The final refined hit list of proteins was filtered for nuclear proteins with enrichment ratios of SETDB1 IP/IgG IP ( medium/light ) of >2 , >2 unique peptides and >2 independent spectra . Native whole-cell extracts for immunoprecipitation were prepared from mESCs and 293T cells by lysing cells in 20 mM HEPES , 200 mM KCl , 1% NP-40 , 1 mM EDTA , 10% glycerol , containing 1 mM DTT , 10 or 20 mM NEM , complete EDTA-free protease inhibitor cocktail ( Roche ) and PhosStop phosphatase inhibitor cocktail ( Roche ) . For preparing cell extracts for western blotting , cells were lysed in RIPA buffer ( 50 mM Tris pH 8 . 0 , 150 mM NaCl , 1% NP-40 , 0 . 25% deoxycholate , 0 . 1% SDS ) . Histones were isolated from mESCs for westerns by acid-extraction of nuclei with 0 . 2 M HCl or by boiling cells in SDS-PAGE loading buffer . For IP , nuclear extract ( 100 μl ) was diluted with 2 volumes of IP dilution buffer ( 20 mM HEPES pH 7 . 9 , 1 . 5 mM MgCl2 , 0 . 5% NP-40 , 10% glycerol ) . Protein samples were immunoprecipitated overnight at 4°C with anti-SETDB1 ( kind gift from H . H . Ng reported previously [28] ) , anti-hnRNP K 3C2 ( Abcam ) , anti-KAP1 20C1 ( Abcam ) , anti-DYKDDDDK ( FLAG , GenScript A00187–200 ) , or rabbit or mouse IgG ( Sigma-Aldrich I8140 and I8765 ) . Whole-cell extracts were immunoprecipitated by adding antibodies and incubating overnight . Immunocomplexes were captured on protein A or protein G dynabeads ( Life Technologies ) , washed three times in IP wash/whole-cell extraction buffer and eluted by boiling in SDS-PAGE loading buffer . For IP of FLAG-SETDB1 complexes from mESCs , Setdb1 deletion was induced with tamoxifen in the 33#6 cell line expressing 3XFLAG-Setdb1 , as previously [11] . As a negative control , the 33#6 line lacking the Setdb1 transgene was used without inducing Setdb1 deletion . Nuclear extracts were prepared as above with 10 mM NEM and immunoprecipitated overnight with anti-DYKDDDDK ( FLAG ) antibodies ( GenScript ) . Immunocomplexes were captured on protein G dynabeads , washed with IP wash buffer containing 0 . 1% NP-40 and eluted with phosphate-buffered saline ( Dulbecco ) containing 0 . 1% Tween-20 and 500 μg/ml 3XFLAG peptide ( Sigma Aldrich ) . Western blotting was performed as previously described [42] using anti-SETDB1 H300 ( Santa Cruz Biotechnology ) , anti-hnRNP K 3C2 ( Abcam ) , anti-KAP1 20C1 ( Abcam ) , anti-GAPDH ( Millipore AB2302 ) , anti-Ubc9 ( Santa Cruz sc-5231 ) , anti-SUMO1 ( Santa Cruz sc-9060 ) , anti-SENP1 ( Novus Biologicals NB100–92101 ) , anti-H3K9me2 ( Abcam ab1221 ) , anti-H3K9me3 ( Active Motif 39161 ) , anti-pan H3ac ( Millipore 06–599 ) , anti-H4 ( Millipore 04–858 ) , anti-GST ( GenScript A00097–100 ) , anti-DYKDDDDK ( FLAG , GenScript ) and anti-T7 ( Millipore 59622 ) . Primary antibodies were detected using IRDYE-conjugated secondary antibodies and scanning on the Odyssey imager ( LiCOR Biosciences ) . The pSG5 plasmid harbouring FLAG-tagged mouse Setdb1 cDNA [75] was a kind gift from L . Yang . The pcDNA3 . 3-T7-HNRNPK plasmid [32] was kindly provided by A . Srebrow . The pET16b-HNRNPK plasmid expressing 6X-His-tagged human hnRNP K was a kind gift from A . Ostareck-Lederer and was expressed and purified from the BL21 ( DE3 ) E . coli strain as previously [76] . GST-tagged hnRNP K and GST-tagged KAP1 and KAP1PxVxL ( residues 379–524 ) were purchased from Novus Biologicals . GST-KAP1PB ( residues 624–811 ) was from Cayman Chemical . Purified GST was from Sigma-Aldrich , the C-terminal GST-RanGAP1 fragment ( residues 419–587 ) was from Enzo Life Sciences and GST-p53 was from Millipore . Purified FLAG-tagged SETDB1 protein was from Active Motif . In vitro SUMOylation assays were performed according to previous methods [27] with minor modifications . Approximately 500 ng of GST-fused proteins were mixed with 125 ng Aos1/Uba2 heterodimer ( Enzo Life Sciences BML-UW9330–0025 ) , 500 ng Ubc9 ( Enzo Life Sciences BML-UW9320–0100 ) and 2 μg 6X-His-tagged SUMO1 ( Enzo Life Sciences ALX-201–045-C500 ) and incubated in 20 μl of 1X SUMOylation buffer ( 50 mM Tris pH 8 . 0 , 50 mM KCl , 5 mM MgCl2 , 1 mM DTT , 1 mM ATP ) for 90 minutes at 30°C . Negative control reactions were performed by omitting SUMO1 . Following this , reactions were stopped either by addition of SDS-PAGE loading buffer for western blotting or prepared for pulldown assays . For pulldown assays , GST-tagged proteins were immobilized on glutathione magnetic beads ( GenScript ) , washed twice with pulldown buffer ( 50 mM Tris pH 8 . 0 , 100 mM NaCl , 0 . 1 mM EDTA , 1 mM DTT , 0 . 01% Tween-20 , 10% glycerol ) incubated with 0 . 5–5 μg recombinant prey proteins ( SETDB1 , Ubc9 or hnRNP K ) in 150 μl pulldown buffer for 1 . 5 h at 4°C . For pull-downs with GST-tagged KAP1 mutant baits and 6X-His-tagged hnRNP K prey , BSA was included in the binding reaction at 1 mg/ml . Beads were washed again three times pulldown buffer for SETDB1 and Ubc9 or in pulldown buffer containing 300 mM NaCl for hnRNP K and subsequently eluted with SDS-PAGE loading buffer for western blotting . Alternatively , glutathione elution buffer was used to elute bound proteins ( 50 mM Tris pH 8 . 0 , 20 mM reduced L-glutathione , 1 mM DTT ) . Ultracentrifugation of proteins over sucrose gradients was performed according to previous methods [77 , 78] . Approximately ~2 mg of mESC nuclear extract or ~4 μg of recombinant proteins in 500 μl was layered onto a 5 ml linear 5–50% gradient and centrifuged in parallel with identical gradients containing purified molecular weight standards ( blue dextran 52 . 6S/~2 MDa , thyroglobulin 19 . 4S/670 kDa , catalase 11 . 4S/250 kDa , BSA 4 . 3S/67 kDa all from Sigma-Aldrich ) at 27 , 500 rpm ( ~91 , 900 g ) in a SW-55Ti rotor ( Beckman Coulter ) at 4°C for 18 . 5 h . Fractions of 200 μl were collected from top to bottom including the pellet fraction and 20 μl samples were assayed by western blot . Peaks for migration of the molecular weight standards were determined by absorbance at 280 nm . For native ChIP , mESCs were harvested by trypsinization and lysed in NChIP lysis buffer ( 20 mM HEPES pH 7 . 9 , 50 mM KCl , 1 mM MgCl2 , 3 mM CaCl2 , 1 mM DTT , 0 . 5% NP-40 , 10% glycerol ) containing protease inhibitors on ice . Chromatin was digested with MNase ( Worthington Biochemicals ) to produce predominantly mono and di-nucleosomes and stopped by addition of EDTA and EGTA to 5 mM each , respectively . Salt concentration was adjusted to 150 mM KCl and native chromatin was immunoprecipitated overnight with anti-H3K9me3 ( Active Motif 39161 ) or anti-H4K20me3 ( Active motif 39180 ) . Immunocomplexes were captured on protein A and G dynabeads ( Life Technologies ) washed extensively in RIPA buffer and eluted with 100 mM sodium bicarbonate buffer containing 1% SDS , 20 mM DTT . DNA was RNAse A-treated and purified over spin columns and analyzed by qPCR using primers indicated in S3 Table . Crosslinked ChIP of SETDB1 , KAP1 and SUMO1 was performed according to a previously described method [42] with or without 10 mM NEM . Chromatin was immunoprecipitated overnight at 4°C using anti-SETDB1 H300 , anti-KAP1 20C1 or anti-SUMO1 ( Santa Cruz sc-9060 ) . ChIP for hnRNP K was performed according to a previous method with minor changes [51] . Briefly , cells were crosslinked in 1 . 45% formaldehyde for 15 minutes at room temperature , quenched with glycine and collected by centrifugation . Cells were lysed in modified RIPA ( 50 mM Tris pH 8 . 0 , 150 mM NaCl , 1% NP-40 , 0 . 5% triton-X-100 , 5 mM EDTA , 1 mM DTT , 10 mM NEM , 10% glycerol containing protease inhibitors ) and sonicated to yield predominantly 150–600 bp fragments . Chromatin lysate was precipitated overnight with 10 μg/ml anti-hnRNP K ( Abcam ab70492 ) . RNAse treatment of sonicated chromatin prior to ChIP was performed according to a previous method [79] . Samples were washed and eluted as described above and purified DNA was analyzed by quantitative PCR using ChIP primers indicated in S3 Table . Total RNA was extracted from mESCs with the GenElute RNA kit ( Sigma-Aldrich ) and reverse transcribed with SuperScript III ( Life Technologies ) . Quantitative RT-PCR was performed as previously [42] . Expression levels were normalized to endogenous control genes Gapdh or β-actin ( Actb ) . Primers used for qRT-PCR are listed in S3 Table . Strand-specific , paired-end mRNA-seq on poly ( A ) RNA was performed as previously described [42] . Libraries were sequenced on the Illumina HiSeq 2000 . Reads per kilobase per million mapped reads ( RPKM ) was calculated and genes up- or downregulated relative to control siRNA KD cells were determined by applying fold-change threshold of 2 and minimum read count of 25 for genes up ( down ) regulated in hnRNP K KD ( control ) cell lines . Gene ontology analysis was performed with DAVID bioinformatic resource version 6 . 7 at http://david . abcc . ncifcrf . gov/home . jsp .
|
Retroelements , including endogenous retroviruses ( ERVs ) , pose a significant threat to genome stability . In mouse embryonic stem ( ES ) cells , the enzyme SETDB1 safeguards the genome against transcription of specific ERVs by depositing a repressive mark H3K9 trimethylation ( H3K9me3 ) . Although SETDB1 is recruited to ERVs by its binding partner KAP1 , the molecular basis of this silencing pathway is not clear . Using biochemical and genetic approaches , we identified hnRNP K as a novel component of this silencing pathway that facilitates the recruitment of SETDB1 to ERVs to promote their repression . HnRNP K binds to ERV sequences via KAP1 and subsequently promotes SETDB1 binding . Together , our results reveal a novel function for hnRNP K in transcriptional silencing of ERVs and demonstrate a new regulatory mechanism governing the deposition of H3K9me3 by SETDB1 in ES cells .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
hnRNP K Coordinates Transcriptional Silencing by SETDB1 in Embryonic Stem Cells
|
The opa genes of the Gram negative bacterium Neisseria meningitidis encode Opacity-associated outer membrane proteins whose role is to promote adhesion to the human host tissue during colonisation and invasion . Each meningococcus contains 3–4 opa loci , each of which may be occupied by one of a large number of alleles . We analysed the Opa repertoire structure in a large , well-characterised collection of asymptomatically carried meningococci . Our data show an association between Opa repertoire and meningococcal lineages similar to that observed previously for meningococci isolated from cases of invasive disease . Furthermore , these Opa repertoires exhibit discrete , non-overlapping structure at a population level , and yet low within-repertoire diversity . These data are consistent with the predictions of a mathematical model of strong immune selection upon a system where identical alleles may occupy different loci .
The Opacity ( Opa ) proteins of the bacterial pathogen Neisseria meningitidis mediate adhesion to and invasion of the human nasopharyngeal epithelium [1] via interaction with cell surface saccharides [2] and members of the carcinoembryonic antigen cell adhesion molecule ( CEACAM ) family of proteins [3] , [4] . The opa gene repertoire comprises 3–4 loci per meningococcus ( opaA , opaB , opaD and opaJ ) [5]–[8] . These are constitutively transcribed and their expression is controlled by stochastic changes in a phase variable , pentameric repeat tract within the reading frame of the genes [9] . Varying numbers of opa loci may be expressed at different times and in different combinations , providing both functional flexibility and a possible mechanism for immune evasion . Opa proteins are highly diverse [8] , [10] with the majority of sequence changes localised in three regions which correspond to surface exposed loops in the proposed protein structure . It is thought that different sequences in the semivariable ( SV ) and two immunodominant hypervariable ( HV ) regions [10] , [11] confer different receptor specificities to the protein [12] , [13] . Diversity is generated by gene conversion , mosaicism and also modular exchange of variable regions , with the consequence that different opa loci in the same meningococcus may encode identical , similar or diverse HV regions [14] , [15] . It has been shown that the Opa repertoire is highly structured among the hyperinvasive lineages of meningococci that are responsible for the majority of global disease [16] . Isolates from the same hyperinvasive clonal complexes ( as defined by MLST ) have been shown to possess similar and often identical Opa repertoires , despite being sampled from disparate geographical locations and temporal periods [8] . Little information is available , however , on the extent of the diversity of the Opa repertoire in carried populations of meningococci which contain the majority of meningococcal biodiversity . In this investigation , we analysed the Opa repertoires of a geographically and temporally related collection of asymptomatically carried meningococci to determine whether the association between clonal complexes and particular combinations of these adhesins , as observed in hyperinvasive lineages , was present in non-disease causing meningococci . We analysed the data using a theoretical model of immune selection which incorporated the particular features of this antigenic system including its phase variable nature and the modular exchange of variable regions within genotypes . We found the patterns of diversity evident at both the population level and within individual repertoires to be indicative of strong immunological selection acting in addition to the forces of functional adaptation in influencing the structure of the Opa repertoire .
The four known opa loci were analysed in the 216 meningococcal isolates from a carried population sample from the Czech Republic: a total of 864 loci . In 784 loci ( 90 . 74% ) an intact opa sequence was detected; these contained a total of 222 alleles ( nucleotide p distance: 13 . 59% ) . These encoded 76 HV1 variants ( amino acid p distance: 47 . 8% ) which fell into 19 families and 93 HV2 variants ( amino acid p distance: 37 . 6% ) which fell into 21 families . A total of 212 opa loci were also analysed in a contemporaneous collection of 53 isolates from invasive disease . In 185 loci ( 87 . 26% ) an opa sequence was detected , these contained a total of 75 alleles ( nucleotide p distance: 14 . 26% ) . These encoded 41 HV1 variants ( amino acid p distance: 48 . 4% ) which fell into 15 families and 44 HV2 variants ( amino acid p distance: 40 . 1% ) which fell into 17 families . In both the carriage and disease collections , we found that genetically related isolates , whether belonging to hyperinvasive clonal complexes or not , often had identical Opa repertoires ( see Text S1 and Tables S1 and S2 ) . For example , for the ST-11 complex , the opa gene alleleic repertoire opaA 83 , opaB 11 , opaD 132 and an insertionally inactivated opaJ locus was present in 27 of 32 carried isolates and 16 of 20 disease isolates . The remaining isolates from this complex in each collection had highly related repertoires , differing at only one or two loci . These repertoires were highly similar to those of isolates belonging to the same clonal complexes observed in a global collection of hyperinvasive meningococci ( see Text S1 and Tables S1 and S2 ) [8] . The modular exchange of the immunodominant HV regions among different opa loci [14] , [15] makes the system unusual in the context of immune selection , since the same hypervariable region variants may be present at multiple opa loci within the same isolate , as well as in different isolates . We extended a multi-locus mathematical model ( see Materials and Methods for details ) developed by Gupta et al . [17] to incorporate this feature by allowing the two HV regions ( HV1 and HV2 ) at each locus to contain two possible amino acid sequence epitopes ( ‘a’ and ‘b’ for HV1 and ‘x’ and ‘y’ for HV2 ) as shown in Figure 1 . Thus , possible combinations of HV regions in Opa proteins expressed by different meningococci could be: ‘ax’ , ‘ay’ , ‘bx’ , ‘by’ , ‘axy’ , ‘bxy’ , ‘abx’ , ‘aby’ , and ‘axby’ . The behaviour of this model under different levels of immune selection is shown in Figure 2 . These simulations indicate that the system shows a tendency to self organise at a population level into discrete antigenic types as the strength of immune selection increases , as previously observed [17] for multi-locus systems without modular exchange of variable regions . When immunological selection ( as measured by cross-protection between pathogen types sharing variable regions ) is weak , all antigenic types coexist at the similar abundances as shown in Figure 2a . By contrast , when immunological selection is high , a subset of two strains expressing two non-overlapping HV1/HV2 region combinations ( for example , ‘ax’ and ‘by’ ) dominates , excluding all other strains , as exemplified by Figure 2c . Between these two extremes , we observe cyclical dynamics with strains expressing subsets of non-overlapping HV variants successively dominating the population ( Figure 2b ) . Figure 3 shows the combinations of HV1 and HV2 present at all loci for all isolates . Each opa locus is treated independently , so each isolate can contribute more than one combination . Variants for which only a single isolate was found were excluded from this analysis ( see Table S3 for full details ) . To determine which of these population structures best described these data , a simple metric ( f* ) was developed to assess the extent of overlap between two epitopes among different isolates ( see Materials and Methods for the derivation , and Text S1 for model validation ) : f* scores close to 1 indicate a highly non-overlapping structure , expected when cross-immunity is high , whereas scores close to 0 occur when strains have completely overlapping antigenic repertoires . Scores obtained from the opa loci in the dataset were compared to scores from housekeeping genes belonging to the same isolates . The f* metric for the data shown in Figure 3 is 0 . 9737 , whereas pairwise comparisons of the housekeeping gene loci yielded a mean f* score of 0 . 3453 and a maximum of 0 . 4578 . These scores indicate the non-overlapping nature of the Opa HV1/HV2 combinations as compared to the housekeeping loci , and reflect the diagonal pattern observed in the figure . A total of 124 HV1/HV2 combinations were observed out of a possible total of 7068 ( 76 HV1 variants multiplied by 93 HV2 variants ) . Discrete , non-overlapping combinations of HV1 and HV2 are clearly dominant , despite the presence of rare combinations generated by frequent recombinational exchange . These observations support the model structure described above in which strong immune selection is responsible for structuring Opa repertoires . Another important feature of the simulations presented in Figure 2 which is unique to a system with modular exchange between loci is that immune selection paradoxically leads to a reduction in diversity within individual opa repertoires . In other words , if more than one opa locus is expressed in vivo , selection will favour those strains expressing multiple loci encoding the same combination of HV regions , rather than different , more diverse variants . It is evident from Figure 2a that even under low levels of immune selection , the prevalence of strains expressing three or four antigenic determinants is suppressed . The magnitude of suppression increases with the degree of cross-protection ( represented in the model by the parameter γ ) , such that these more diverse types are entirely absent in Figure 2c . This suppression occurs because strains expressing more than two HV variants are less likely to encounter hosts who have not previously been exposed to one or more of their epitopes , and are therefore at a disadvantage within the population . Thus , at very high levels of immune selection , we observe only meningococci that expressed a single opa locus , or multiple loci encoding the same combination of HV regions ( i . e . ‘ax’ at locus 1 and ‘ax’ at locus 2 ) . This is because the pathogen population , and therefore the background of host immunity , is dominated by two non-overlapping strains , say ‘ax’ and ‘by’ , so that more diverse strains ( such as ‘axy’ ) are more likely to be recognized by hosts who have encountered either one of the dominant strains . To investigate the effect of host immunity on the structuring of the HV region repertoire diversity in individual isolates , we analysed the HV combinations of different Opa proteins within the same isolate for those that had full opa sequences at more than one locus ( not including those that had frame-shift mutations or insertional inactivations ) . Figure 4 shows the proportion of isolates with identical HV1/HV2 combinations at different opa loci within the same isolate , compared to a hypothetical pathogen population in which the same combinations found in the data were distributed randomly within and among isolates . Only unique Opa repertoires were included in the analysis , to control for bias due to particularly prevalent sequence types and clonal complexes . Our results showed that significantly more isolates contained two or more of the same HV1/HV2 combination than would be expected by chance given the same overall prevalence of variants ( p<0 . 0001 ) . Furthermore , they were not always the same HV1/HV2 combinations that were identical , with 28 different combinations occurring more than once within isolates . Finally , the probability that two or more were identical increased with the number of opa loci at which a full length opa allele was detected for each isolate ( see Figure 4 ) .
The Opa repertoire structure observed in the carried meningococci from the Czech Republic , and its relationship to clonal complex , was similar to that previously described in an isolate collection representing the diversity of meningococci causing disease globally in the latter half of the 20th century [8] . This does not imply that the Opa repertoire has no role in meningococcal pathogenesis since other factors , such as differences in expression patterns among meningococci and host susceptibility , are likely to influence the outcome of infection . Despite evidence for extensive recombination of opa loci among meningococci , only a fraction of all possible combinations of HV1 and HV2 were observed . . These combinations exhibited a non-random and non-overlapping structure , which was consistent with a model of immunological selection in which competition between pathogen types leads to a pathogen population dominated by non-overlapping combinations of antigenic variants [17]–[19] . The low frequency off-diagonal elements shown in Figure 3 may be attributed either to the point prevalent nature of the data set ( ie . that these combination are shortlived ) or reflect the fact that certain variants possess immunological similarities , and are therefore equally likely to occur in combination with certain others . It is also possible that there are functional constraints in operation here since particular HV1/HV2 combinations influence receptor tropism and potentially also avidity [12] , [13] , [20] . It has been suggested that expression of CEACAM on host cell surfaces may allow evasion of antibody responses by Opa-mediated entry into epithelial cells [4] and modulation of the host immune responses by interaction with CD4+ T cells [21] . The specificity of these interactions is likely to constrain allowable HV1/HV2 combinations and may explain why particular combinations are entirely absent in our data . Non-overlapping patterns of epitope combinations have also been observed among meningococcal PorA variable regions [17] , [19] . Unlike PorA however , the Opa repertoire is a four-locus system [5] , [8] , and has been suggested to play a role in immune evasion [22] , [23] . For the Opa proteins , individual repertoires exhibited more identical HV variants than would be expected under the assumption that antigenic diversity of a surface component prolongs infection ( Figure 4 ) . This result is , however , consistent with the predictions of a mathematical model of strong immune selection upon a system where identical alleles may occupy different loci . Within this framework , isolates expressing diverse repertoires are at a disadvantage because they are more likely to encounter hosts with previous exposure to one or more of their epitopes . This results in selection for identical variants at multiple loci: in other words , a reduction in the diversity of the Opa repertoires within individual meningococci . The prevalence of identical Opa variants within repertoires implies that multiple opa loci are expressed in vivo; if expression were restricted to a single opa locus , there would be no selective disadvantage of carrying a diverse repertoire . An alternative explanation for the low within-repertoire diversity is that identical HV combinations reflect genetic duplication events that are followed by specialisation of duplicates for new functions [24] , [25] . An exception to the pattern of diversity within the Opa repertoires in most clonal complexes was that of the ST-11 complex which did not have any identical HV variants among its loci . This may be due to the recent entry and rapid spread of this clonal complex into the population of the Czech Republic , where it was responsible for a rapid increase in the incidence , mortality and morbidity of invasive meningococcal disease in 1993 [26] . Retrospective monitoring of isolates since 1970 suggested that this strain was not present in the country before 1993 and consequently the Czech population may have been immunologically naïve , allowing these meningococci to spread through the population . Thus , high Opa repertoire diversity may be selectively advantageous for the invasion of new communities of hosts with variable immunological backgrounds . During prolonged carriage in the same host population however , increased diversity may become costly as the proportion of immunologically naïve hosts decreases . This would inevitably cause a reduction in the range of receptor tropism , but this would be offset by the gain in probability of transmission . To date , the majority of Opa proteins tested bind at least CEACAM1 [27] , suggesting that the repertoire retains binding of this major receptor . Intriguingly , the number of opa loci differ among the Neisseria species , with 3–4 in Neisseria meningitidis [6] , [7] , 11–12 loci in Neisseria gonorrhoeae [28] and two in Neisseria lactamica [29] . The reasons for these differences are unclear , but our analyses in this study suggest a theory based on population prevalence and immunological cross-protection . For example , whereas N . meningitidis is transmitted by aerosol inhalation , N . gonorrhoeae is transmitted sexually and consequently has a much lower population prevalence . The likelihood of N . gonorrhoeae encountering an immunologically naïve host may be much higher , therefore , and the diversity-reducing effect from the host population's immunological responses less pronounced than for the meningococcus . A more diverse Opa repertoire with more loci may be more advantageous in these circumstances . Further information on the antigenic diversity of the gonococcal Opa repertoire and immunological responses against both pathogenic Neisseria species would be required to test this hypothesis . In conclusion , this analysis demonstrates that particular Opa repertoires are associated with meningococcal clonal complexes irrespective of geographic or temporal sampling , whether isolated from asymptomatic carriers or invasive disease cases . The repertoires exhibit discrete , non-overlapping structure on a population level and low within-repertoire diversity , indicating that immune selection plays an important role in shaping Opa repertoires .
A total of 216 meningococcal isolates were obtained from an asymptomatically carried population of meningococci collected in the Czech Republic between March and June of 1993 [30] . A full description of these meningococci , including year and location of isolation , MLST and antigen gene sequencing data appears online at http://pubmlst . org . Genomic DNA was prepared by culturing isolates as previously described [30] before extracting with a DNA mini kit ( Qiagen , Crawley , UK ) according to the manufacturer's instructions . The opa loci were isolated in separate , locus-specific PCR amplifications , their nucleotide sequences determined at least once on each strand and their variable regions identified as previously described [8] . Nucleotide and amino acid sequence data are available in an online database located at http://neisseria . org/nm/typing/opa/ . For analyses of diversity , by uncorrected nucleotide or amino acid percentage ( p ) distance , sequences were aligned and diversity calculated using the program DAMBE: Data Analysis and Molecular Biology and Evolution [31] . A non-overlapping strain structure results in a matrix of allelic associations between two antigenic loci in which each allele at locus 1 should be predominantly associated with only one allele at locus 2 , and vice versa . This means that the most prevalent strains should dominate both the ‘row’ and ‘column’ of their allelic association matrix . The level of overlap within such a matrix can therefore be measured by assessing the dominance of the most prevalent allele combinations;The dominance of the most prevalent allele combination in each column ( fa ) is calculated , where locus 1 expresses allele i , and locus 2 expresses allele j . fi is the frequency of the most prevalent strain expressing allele i at locus 1 , with respect to all strains expressing that allele ( ie . the ‘column’ dominance ) , fj is the frequency of that strain with respect to all strains expressing allele j at locus 2 ( ie . the ‘row’ dominance ) , and fij is the frequency of allele combination ij overall . These are calculated as follows:The sum over all fa gives the overall overlap between two loci:such that f* varies between 0 and 1 . For a completely non-overlapping matrix , with no combinations found except for those that do not overlap , f* will be exactly one . As this structuring breaks down , the f* score will decrease rapidly . Three differential equations , based on a model by Gupta et al . [17] , [18] , describe the system:The model states that once infected with a particular strain , the host gains partial immunity to any other strains expressing shared antigenic determinants ( the subset of strains i' above ) as specified by the parameter γ . For each strain i , the host population consists of three overlapping compartments; the proportion infectious to other hosts , xi; the proportion exposed ( and therefore immune ) to strain i , zi , and the proportion exposed to any strain sharing antigenic determinants with i , wi . It was assumed that the duration of infectiousness ( 1/σ ) was short compared to the average host life-span ( 1/μ ) , and that immunity was life-long . All strains were assumed to have the same transmission coefficient , β . The effect of recombination was not explicitly included in the model , however all possible strains were present from the start in order to investigate the competitive interactions between them . Note that in this model there was no dose-dependence; two loci expressing Opa proteins with identical HV regions was taken as being the same as if only one locus expressed the protein .
|
Neisseria meningitidis is a globally important pathogen that causes 2 , 000–3 , 000 cases of invasive meningococcal disease annually in the United Kingdom . The meningococcal Opa proteins are important in mediating adhesion to and invasion of human tissues , and are important for evasion of the host immune response . They are encoded by a repertoire of 3–4 genomic loci in each meningococcus and exhibit high levels of sequence diversity . Here we analyzed the Opa repertoires of a large , well-characterised , asymptomatically carried meningococcal isolate collection . We found that the Opa repertoires were specific to individual meningococcal genotypes , similar to that observed in isolates from cases of invasive disease . These repertoires exhibited discrete , non-overlapping structure at a population level , and yet low within-repertoire diversity . These data were consistent with the predictions of a mathematical model of strong immune selection , suggesting that the collective immune response of the host population shapes the antigenic diversity of the meningococcal Opa repertoire . This study provides new insights into Opa-mediated meningococcal pathogenesis and the effect of host population immunity on the biodiversity and population structure of bacterial pathogens . These data may also have implications for the design of new meningococcal vaccines based on surface proteins .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"computational",
"biology/population",
"genetics",
"genetics",
"and",
"genomics/microbial",
"evolution",
"and",
"genomics",
"evolutionary",
"biology/microbial",
"evolution",
"and",
"genomics",
"microbiology/immunity",
"to",
"infections",
"computational",
"biology/molecular",
"genetics",
"microbiology/microbial",
"evolution",
"and",
"genomics",
"computational",
"biology/evolutionary",
"modeling",
"evolutionary",
"biology/bioinformatics",
"infectious",
"diseases/bacterial",
"infections",
"immunology/immunity",
"to",
"infections",
"genetics",
"and",
"genomics/bioinformatics",
"infectious",
"diseases/epidemiology",
"and",
"control",
"of",
"infectious",
"diseases",
"genetics",
"and",
"genomics/population",
"genetics"
] |
2008
|
The Effect of Immune Selection on the Structure of the Meningococcal Opa Protein Repertoire
|
The macromolecular assembly required to initiate transcription of protein-coding genes , known as the Pre-Initiation Complex ( PIC ) , consists of multiple protein complexes and is approximately 3 . 5 MDa in size . At the heart of this assembly is the Mediator complex , which helps regulate PIC activity and interacts with the RNA polymerase II ( pol II ) enzyme . The structure of the human Mediator–pol II interface is not well-characterized , whereas attempts to structurally define the Mediator–pol II interaction in yeast have relied on incomplete assemblies of Mediator and/or pol II and have yielded inconsistent interpretations . We have assembled the complete , 1 . 9 MDa human Mediator–pol II–TFIIF complex from purified components and have characterized its structural organization using cryo-electron microscopy and single-particle reconstruction techniques . The orientation of pol II within this assembly was determined by crystal structure docking and further validated with projection matching experiments , allowing the structural organization of the entire human PIC to be envisioned . Significantly , pol II orientation within the Mediator–pol II–TFIIF assembly can be reconciled with past studies that determined the location of other PIC components relative to pol II itself . Pol II surfaces required for interacting with TFIIB , TFIIE , and promoter DNA ( i . e . , the pol II cleft ) are exposed within the Mediator–pol II–TFIIF structure; RNA exit is unhindered along the RPB4/7 subunits; upstream and downstream DNA is accessible for binding additional factors; and no major structural re-organization is necessary to accommodate the large , multi-subunit TFIIH or TFIID complexes . The data also reveal how pol II binding excludes Mediator–CDK8 subcomplex interactions and provide a structural basis for Mediator-dependent control of PIC assembly and function . Finally , parallel structural analysis of Mediator–pol II complexes lacking TFIIF reveal that TFIIF plays a key role in stabilizing pol II orientation within the assembly .
In humans , the transcription initiation machinery consists of Mediator , pol II , TFIIA , TFIIB , TFIID , TFIIE , TFIIF , and TFIIH and approximates 3 . 5 MDa in size . This large assembly can exist in various structural and functional states [1] . When not in an activated state that supports transcription initiation , this assembly is best described as a Pre-Initiation Complex ( PIC ) [2] . At 1 . 2 MDa , Mediator represents a major component within the human PIC and based upon biochemical assays , Mediator helps assemble and stabilize the PIC [3] , [4] . Human Mediator is known to functionally interact with most PIC components , including TFIIB , TFIID , TFIIE , TFIIH , and pol II itself [3] , [5]–[9] . Although Mediator appears critical for controlling the assembly and activity of the PIC , a structural basis for these observations has not been established . Structural analysis of the human transcription initiation machinery has been hindered by several factors , including the large size and complexity of the machinery itself . Although it is well-established from biochemical assays that pol II physically interacts with human Mediator [5] , [7] , little is known about the pol II–Mediator interface . Attempts to structurally define the pol II–Mediator interface have been made in yeast , but these have been limited to incomplete assemblies of Mediator and/or pol II [10] , [11] . Perhaps as a consequence , these studies have provided inconsistent predictions of pol II orientation relative to Mediator itself [12] , [13] . Because structural data are not available for the complete Mediator–pol II assembly , even the most basic information about human PIC structure remains unknown , such as how PIC components might assemble together with Mediator at a promoter . For instance , it is not known how pol II orients upon interaction with Mediator . Because the location of other PIC factors ( TFIIA , TFIIB , TBP , TFIIE , TFIIF , TFIIH ) has been determined relative to pol II itself [14]–[18] , identifying the pol II orientation when bound to Mediator would help define the structural organization of the entire 3 . 5 MDa human PIC . Thus , structural analysis of the Mediator–pol II assembly represents an essential , yet missing , link to defining the molecular architecture of the human PIC . It is also unclear how the interaction of Mediator and pol II permits simultaneous assembly of the large , 1 . 1 MDa TFIID complex as well as other PIC components—such as TFIIB , TFIIE , and TFIIF—that interact directly with pol II during transcription initiation . Finally , Mediator is required for TFIIH-dependent pol II CTD phosphorylation within the human PIC [9] , yet it is not established how the pol II CTD might track within the PIC , nor is it known what structural features within the Mediator–pol II assembly could allow for regulation of TFIIH-dependent pol II CTD phosphorylation . The large size , low-abundance , and dynamic features of the human Mediator complex prevent an analysis using high-resolution techniques such as X-ray crystallography or NMR spectroscopy . However , structural analysis of Mediator is well-suited for cryo-EM studies , which require sub-microgram quantities of purified protein and can potentially resolve alternate conformational states of macromolecular complexes . We purified two different sub-assemblies within the 3 . 5 MDa PIC: the 1 . 8 MDa Mediator–pol II binary complex and the 1 . 9 MDa Mediator–pol II–TFIIF assembly . In each case , Mediator was bound to the activation domain of VP16 . Cryo-EM analysis of each assembly revealed the overall structural organization of the entire human PIC and identified a role for TFIIF in stabilizing Mediator–pol II interactions . Our results establish Mediator as the scaffold around which the entire human PIC assembles and reveal a pol II-induced structural shift within Mediator that likely precludes Mediator binding to the CDK8 submodule . Collectively , these observations provide a structural basis for initiation and post-initiation regulatory events and further define how Mediator coordinates PIC assembly and function .
In order to assemble the Mediator–pol II–TFIIF complex or the Mediator–pol II binary complex , we first purified Mediator , pol II , and TFIIF independently . Human TFIIF was purified following recombinant expression in E . coli , whereas pol II and Mediator were each isolated as endogenous complexes from HeLa cells . Mediator purification involved an affinity resin using the activation domain of VP16 ( residues 411–490 ) , yielding VP16-bound Mediator complexes [19] . Each complex ( TFIIF , VP16-Mediator , and pol II ) was purified to near-homogeneity , as shown in Figure 1A–C . We completed mass spectrometry analysis of Mediator and pol II , primarily to confirm that these purified complexes contained each of their consensus subunits: 26 subunits within Mediator and 12 subunits for pol II ( Table 1 ) . With the purified , 26-subunit Mediator complex and the 12-subunit pol II complex in hand , we next tested whether Mediator and pol II and/or Mediator , pol II , and TFIIF would associate to form a stable assembly that could be isolated and imaged using electron microscopy . To isolate the Mediator–pol II–TFIIF assembly , pol II and TFIIF ( added in excess to pol II ) were incubated together for 1 h at 4°C . Mediator was then added and all three factors were incubated together for an additional hour at 4°C . After incubation , the sample containing Mediator , pol II , and TFIIF was loaded onto a glycerol gradient ( Figure 2A ) . The gradient was designed such that the complete , 1 . 9 MDa Mediator–pol II–TFIIF assembly would migrate and concentrate within the final 1–2 fractions , whereas free Mediator or pol II would mostly sediment within earlier gradient fractions ( Figure 2B and unpublished data ) . As a 100 kDa dimer , free TFIIF sedimented much earlier in the gradient ( Figure 2B ) . The presence of Mediator , pol II , and TFIIF within the final gradient fraction—denoted fraction A—was confirmed by immunoblotting experiments ( Figure 2E ) . Because TFIIF alone sediments much earlier than fraction A , this immunoblotting result provided biochemical evidence that Mediator , pol II , and TFIIF were forming a stable , ternary complex . To investigate the potential impact of TFIIF on the Mediator–pol II structure , we also isolated a Mediator–pol II binary complex using a method similar to that described for isolation of Mediator–pol II–TFIIF ( Figure 2C ) . The glycerol gradient corresponding to the Mediator–pol II experiment showed a silver-stain pattern consistent with the presence of both Mediator and pol II in the last fraction ( fraction B , Figure 2D ) , as expected . Western blot experiments confirmed the presence of Mediator and pol II in this fraction , whereas TFIIF—which was not added in this protocol—was not detected ( Figure 2E ) . To further confirm that TFIIF was present together with Mediator and pol II in gradient fraction A ( Figure 2B ) , but absent in gradient fraction B ( Figure 2D ) , we used an in vitro transcription assay consisting of purified and recombinant human factors [20] . Because this assay requires reconstitution of the transcription machinery from purified components , transcription initiation will not occur if a PIC component ( e . g . TFIIF ) is not added to the reaction . An outline of the transcription assay is shown in Figure 2F . Following addition of activator ( GAL4-p53 ) , the general transcription factors TFIIA , IIB , IID , IIE , and IIH were added to chromatin templates , together with fraction A or fraction B . As shown in Figure 2G , transcription reactions that were not supplemented with a glycerol gradient fraction ( and therefore lacked Mediator , pol II , and TFIIF ) were inactive , as expected ( lane 1 ) . Upon supplementation with gradient fraction A , transcription was activated in a dose-dependent manner ( lanes 2–4 ) . By contrast , supplementation with gradient fraction B , which contains Mediator and pol II but lacks TFIIF , was unable to support transcription ( lanes 5–7 ) . Taken together , the data in Figure 2G further demonstrated TFIIF was present with Mediator and pol II within glycerol gradient fraction A , whereas TFIIF was absent from glycerol gradient fraction B . Given the functional ( Figure 2G ) and biochemical ( Figure 2E ) evidence that Mediator , pol II , and TFIIF formed a stable assembly , we next examined this sample ( fraction A ) using EM . We first imaged negatively stained samples and collected both untilted ( 0° ) and tilted ( 25°–45° ) images to produce an ab initio random conical tilt reconstruction [21] . Examination of untilted micrographs revealed a relatively homogenous field of particles with a size and shape consistent with an intact , 1 . 9 MDa Mediator–pol II–TFIIF assembly ( Figure S1A ) . Reference-free 2D classification followed by back-projection and cross-correlation of the corresponding 3D model structures established a homogenous data set that represented 36% of all single-particle images . This data set was then used to generate an initial reference volume , which was subjected to iterative projection matching [22] to produce a final , refined structure at 42 Å resolution ( Figure S1B–D ) . A number of reprojections of the refined , Mediator–pol II–TFIIF structure are compared with those of Mediator alone in Figure S1E to highlight the significant change in protein density within the Mediator head region . Based upon this comparison , it was evident that pol II interacts with the Mediator head domain , as observed previously with yeast Mediator–pol II complexes [10] , [23] . This Mediator–pol II–TFIIF structure , obtained from negatively stained samples using the random conical tilt methodology , served as a starting point for cryo-EM refinement of Mediator–pol II–TFIIF assemblies . Our studies using negatively stained samples indicated that VP16-Mediator , pol II , and TFIIF could form a stable assembly that was amenable to 3D reconstruction using single-particle methods . We next obtained EM data from the same sample ( i . e . fraction A , Figure 2B ) using cryo-EM techniques , which allows for samples to be imaged in a fully hydrated state , offering the potential for higher resolution structural information and implementation of a powerful 3D variance technique [24] to assess potential structural variability within the sample . Cryo-EM micrographs of Mediator–pol II–TFIIF were collected and screened for astigmatism and sample stage drift ( see Materials and Methods ) . The best 106 micrographs were selected , from which 10 , 856 single-particle images were obtained for image processing . A representative micrograph and its corresponding power spectrum is shown in Figure S2A and S2B . The initial Mediator–pol II–TFIIF structure , generated from negatively stained samples , was low-pass filtered to 57 Å resolution and used as an initial reference volume for iterative projection matching refinement . Initial refinement with the cryo-EM data improved the structure , but the resolution leveled off at around 48 Å . This suggested some conformational or compositional heterogeneity within the cryo-EM data set . Such heterogeneity was in fact expected , as VP16-Mediator was added in excess of pol II during the isolation of the Mediator–pol II–TFIIF assembly and a fraction of complexes within the cryo-EM data set should correspond to free VP16-Mediator . To partition the cryo-EM data into distinct complexes we carried out multi-reference refinement using as references the negative stain reconstructions of the free VP16-Mediator structure [19] and the Mediator–pol II–TFIIF assembly structure , each filtered to 57 Å ( see Materials and Methods ) . Using this protocol , individual cryo-EM images were aligned to re-projections of each distinct reference ( VP16-Mediator or Mediator–pol II–TFIIF ) and partitioned to the structure that yielded the highest cross-correlation . In this way , the cryo-EM data set was separated into two , more homogenous groups . The free VP16-Mediator cryo-EM structure that resulted from this refinement is shown in Figure S3 , whereas different views of the Mediator–pol II–TFIIF assembly are shown along the left panel of Figure 3A ( see also Movie S1 ) . Importantly , separation of free VP16-Mediator images from Mediator–pol II–TFIIF significantly improved the resolution of the Mediator–pol II–TFIIF reconstruction from 48 Å to 36 Å ( Figure S2C ) , based upon the FSC criterion ( or 26 Å using the 3σ criterion ) . The distribution of particle orientations within the Mediator–pol II–TFIIF data set was fairly isotropic , although proportionally fewer end-on views were observed ( Figure S2D ) . To further assess the quality of the Mediator–pol II–TFIIF cryo-EM 3D reconstruction , we generated reference-free 2D class averages from the cryo-EM data set , using a k-means clustering algorithm [25] . As expected , 2D classes resembling Mediator–pol II–TFIIF or the free VP16-Mediator structure were observed . These 2D class averages—generated without any reference bias—were then compared with 2D class averages derived from re-projections of the Mediator–pol II–TFIIF assembly shown in Figure 3A . As shown in Figure S4A , the reference-free 2D class averages closely matched reference-based 2D class averages derived from re-projections of the refined Mediator–pol II–TFIIF structure , supporting the validity of the cryo-EM reconstruction . Similarly , the reference-free 2D class averages closely matched 2D class averages derived from re-projections of the free VP16-Mediator structure ( Figure S4B ) . We also completed protocols to ensure that refinement of Mediator–pol II–TFIIF was not negatively impacted by model bias [26]; these experiments are described in Materials and Methods . Additional strategies were implemented to further refine the Mediator–pol II–TFIIF cryo-EM structure; however , improvement of the resolution beyond 36 Å was not achieved , likely because of flexibility inherent within the 1 . 9 MDa assembly ( see below ) . Importantly , the 36 Å resolution assembly structure is sufficient for accurate pol II docking studies ( Figure S5 ) . We utilized two complementary approaches to determine the orientation of pol II within the Mediator–pol II–TFIIF assembly . We began by docking a crystal structure of yeast pol II into the cryo-EM density map using the program Situs [27] . Docking pol II with Situs represents a rigorous , unbiased way in which to probe pol II orientation , and regardless of where the pol II crystal structure was initially positioned ( e . g . outside the cryo-EM map or within the leg domain or within the head domain ) , Situs calculated the same docking result . Rigid body docking of the pol II atomic model indicated a best fit within the Mediator head region ( Figure 3A , center panels; see also Movie S1 ) , as expected based on comparison of the structure with that of Mediator alone . Note the pol II atomic model shown throughout this article is PDB 1Y1V [28] . This model was chosen because it most closely matches the human pol II structure determined by cryo-EM [29] . However , docking calculations completed with over a dozen distinct pol II crystal structures yielded the same results ( Table S1 ) . The orientation of the pol II cleft in this docking model is perpendicular to the long axis of Mediator such that downstream DNA would extend from the “top” of the assembly . This pol II orientation can readily accommodate binding of other PIC factors ( see Discussion ) . As an alternate means of determining the orientation of pol II within the Mediator–pol II–TFIIF assembly , we used a projection matching strategy ( see Materials and Methods ) in which 2D projections of the human pol II cryo-EM structure [29] were aligned and cross-correlated with 2D projections of the human Mediator-pol II-TFIIF assembly . This independent analysis resulted in the same pol II orientation within the Mediator–pol II–TFIIF assembly as that calculated by docking of the yeast pol II crystal structure ( Figure S6 ) . Given the consistent pol II docking calculations , corroborated by the projection matching data , it was evident that pol II adopted a stable orientation within the Mediator–pol II–TFIIF assembly . The pol II orientation calculated from these alternate methods indicated the pol II cleft was exposed at one end of the assembly ( Figure 3A ) . Because 2D projections of a 3D volume allow an assessment of protein density throughout the volume , 2D projection views from the “top” and “bottom” of the Mediator–pol II–TFIIF structure should provide an additional means to probe the location and orientation of the pol II cleft . As shown in Figure 3B and Figure 3C , the 2D projection views reveal an area deficient in protein density that overlaps precisely with the pol II cleft , offering an additional verification of the pol II docking and projection matching results . The cryo-EM structure of the entire Mediator–pol II–TFIIF assembly shown in Figure 3A was reconstructed using 46% of the data . Single-particle images were included in the reconstruction based upon a cross-correlation threshold . Two additional 3D reconstructions were completed de novo in which a greater percentage of the cryo-EM data was included , based upon adjusting the cross-correlation threshold ( see Materials and Methods ) . As before , the multi-reference refinement protocol was implemented . Single-particle images were free to align to reference projections derived from either of the major entities present in the sample: the free VP16-Mediator structure or the Mediator–pol II–TFIIF complex . In each case ( 62% or 55% of the data was included in the analysis , instead of 46% ) , the Mediator–pol II–TFIIF assembly refined to essentially the same structure as shown in Figure 3A , including an identical docking solution for pol II . However , the resolution did not improve , and small areas of structural discontinuity became evident with the larger data sets , likely due to inclusion of alternate conformational states . These results suggested that inherent flexibility within the Mediator–pol II–TFIIF assembly was limiting the ultimate resolution of the reconstruction . In support of this , we further probed for alternate Mediator–pol II–TFIIF structural states using 3D variance analysis [24] , which yielded no structure distinct from that shown in Figure 3A ( see Materials and Methods ) . A comparison of the free VP16-Mediator structure ( Figure S3 and [19] ) with that of the Mediator–pol II–TFIIF assembly indicates that Mediator itself undergoes significant structural shifts upon binding pol II–TFIIF . Structural shifts occur not only at the Mediator–pol II interface , but also throughout the complex , including the Mediator leg domain ( Figure S7 ) . As a consequence , difference map calculations ( e . g . VP16-Mediator with or without bound pol II–TFIIF ) are not informative . Despite the limited sequence conservation between yeast and human Mediator ( Table S2 ) , it is notable that global structural shifts are also observed upon pol II binding to yeast Mediator [10] . A scheme outlining human VP16-Mediator structural shifts that occur upon pol II–TFIIF binding is shown in Figure S7 . Upon completion of the cryo-EM reconstruction of Mediator–pol II–TFIIF , we next analyzed the Mediator–pol II sample ( Figure 2D , fraction B ) using EM . As with the Mediator–pol II–TFIIF assembly , EM analysis of fraction B began with analysis of negatively stained samples to generate an initial Mediator–pol II structure . As expected , extra density in this structure was apparent within the head domain of Mediator , indicating that pol II associates with the Mediator head domain even in the absence of TFIIF . A description of the EM image processing of the Mediator–pol II data with random conical tilt , negatively stained samples is provided in Materials and Methods . The 3D reconstruction of the Mediator–pol II binary complex obtained from negatively stained samples was used as a starting model for cryo-EM analysis of the complex 141 cryo-micrographs , screened for astigmatism and sample drift , were used ( Figure S8A ) . As with the Mediator–pol II–TFIIF cryo-EM reconstruction , image processing was initiated using a multi-reference approach . The low-pass filtered ( to 57 Å ) free VP16-Mediator structure [19] and the Mediator–pol II structure ( generated from negatively stained samples ) were used as initial references . Throughout the refinement it was evident the Mediator–pol II data set was more structurally heterogeneous than the Mediator–pol II–TFIIF sample ( see Materials and Methods ) . As a result , initial refinements of Mediator–pol II contained regions of discontinuity , suggesting the presence of multiple conformational states within the Mediator–pol II data set . By contrast , data that partitioned to the free VP16-Mediator structure—the second volume included in this initial multi-reference protocol—refined normally: its resolution improved throughout the refinement and its structure matched the previously published structure of VP16-Mediator [19] . To probe for additional conformational states within the Mediator–pol II cryo-EM data , we implemented a 3D variance and focused classification procedure [24] . This identified a region of peak structural variance near the pol II binding site ( region 1 , Figure S9 ) . Cryo-EM images within the Mediator–pol II data set were then sorted into two groups based on focused classification within this region ( see Materials and Methods ) . Two new Mediator–pol II reference structures that resulted from this classification were then used for angular refinement against the cryo-EM data set . Thus , a new multi-reference angular refinement was completed that partitioned the data into one of three reference volumes: free VP16-Mediator , or two distinct Mediator–pol II substructures . Structure refinement improved substantially with this revised multi-reference procedure . In particular , each of the two Mediator–pol II substructures refined to an improved resolution ( 34 Å for substructure 1; 36 Å for substructure 2 ) , and structure discontinuity was eliminated . The 3D reconstruction of each Mediator–pol II structure is shown in Figure 4; see also Figure S8B–E . To define the orientation of pol II within each Mediator–pol II structural state , docking experiments were completed using Situs [27] . In contrast to the Mediator–pol II–TFIIF structure , however , a high-confidence docking result could not be attained for either Mediator–pol II complex ( substructure 1 or substructure 2 ) . Although pol II localized consistently to the Mediator head domain region in each substructure , its orientation was variable and undefined ( Figure 4 ) . Projection matching experiments provided similar results in that a defined , stable pol II orientation was not evident for Mediator–pol II substructure 1 or substructure 2 ( unpublished data ) . Because the same Mediator sample and the same pol II sample were used to assemble both Mediator–pol II–TFIIF and Mediator–pol II ( see Materials and Methods ) , it appears the absence of TFIIF is solely responsible for the dramatically different structure of the Mediator–pol II complex relative to Mediator–pol II–TFIIF . From these data , we conclude that pol II can associate with Mediator in the absence of TFIIF , but pol II does not adopt a stable orientation . By contrast , pol II does adopt a stable orientation within the Mediator–pol II–TFIIF assembly , suggesting TFIIF helps orient and stabilize pol II when bound to Mediator .
Several labs have used X-ray crystallography or crosslinking experiments to determine where the general transcription factors TFIIA , TFIIB , TBP , TFIIE , TFIIF , and TFIIH reside in the PIC relative to pol II itself [14]–[18] , [33] , [34] . Because pol II adopts a defined orientation within the Mediator–pol II–TFIIF assembly , the structural data from previous studies can now augment the Mediator–pol II–TFIIF cryo-EM map . By merging these data , the structural organization of the entire human PIC can be proposed ( Figure 5A ) . Importantly , the location and orientation of pol II within the cryo-EM assembly is completely consistent with previous structural studies that examined pol II and other general transcription factors . For example , surfaces along pol II shown to be required for interaction with TFIIB and TFIIE are exposed and accessible in the Mediator–pol II–TFIIF assembly [15]–[17] . Furthermore , no significant structural rearrangements appear necessary to accommodate the large , multi-subunit TFIID and TFIIH complexes ( Figure S10 ) [35] , [36] . The pol II cleft , which interacts with promoter DNA , is also accessible in the Mediator–pol II–TFIIF assembly and an unobstructed path for the upstream and downstream DNA can be envisioned ( Figure 5A ) . Along the surface of the pol II subunit RPB2 , the entry site for NTPs ( the pore and funnel ) is accessible as is the docking site for potential interactions with TFIIS ( Figure 5B ) [28] , [37] . The fact that previous structural models of pol II together with TFIIB , TFIID , TFIIE , or TFIIH can be incorporated within the human Mediator–pol II–TFIIF cryo-EM map offers further validation of the cryo-EM structure and the pol II docking results . By contrast , these same structural models are not compatible with the various structures proposed for yeast Mediator-yeast pol II ( Figure S10 ) [10] , [11] , [13] . At least four factors might contribute to the differences observed between the human and yeast PIC models . First and foremost , the cryo-EM study outlined here involves the entire , 26-subunit human Mediator complex and the complete , 12-subunit human pol II enzyme . No structural study in yeast has examined the entire 12-subunit pol II enzyme together with Mediator; moreover , pol II docking calculations—which represent the most rigorous , unbiased means to determine pol II orientation within an EM map—have not been completed with yeast factors . One yeast PIC model is based upon a recent EM study that examined a 7-protein yeast Mediator head module and two subunits of yeast pol II , RPB4 and RPB7 [11] . Yet biochemical experiments indicate the 7-subunit yeast head module does not interact with the pol II CTD , nor can it stably interact with the entire pol II enzyme [38]; additional Mediator subunits are required for pol II binding . The pol II–Mediator interface is extensive and clearly involves more than the 7-subunit Mediator head module and does not require the RPB4/7 subunits [10] , [23] . Consequently , an alternate pol II orientation proposed from EM analysis of the 7-subunit yeast Mediator and RPB4/7 dimer likely derives from the fact that interactions required for stable pol II–Mediator binding could not occur [11] . Further supporting this idea , pol II orientation within the yeast pol II-yeast Mediator assembly appears to shift substantially when the entire yeast Mediator complex is examined together with a 10-subunit pol II enzyme that lacks RPB4/7 [10] , [23] . Modeling the RPB4/7 stalk into the EM structure of yeast Mediator bound to the 10-subunit pol II enzyme orients RPB4/7 toward the middle/tail domain of yeast Mediator [13] , whereas the study with the 7-subunit yeast head module proposes that RPB4/7 physically interacts with the Mediator head domain [11] . A second potential reason for the structural differences proposed for yeast Mediator-pol II complexes is that structural studies in yeast have not examined Mediator and pol II together with TFIIF; the structural data outlined here indicate TFIIF plays a major role in orienting human pol II relative to Mediator itself . Third , our structural studies examined human Mediator bound to the activation domain of VP16 , whereas structural studies with yeast Mediator and yeast pol II did not examine Mediator bound to any activation domain . Lastly , whereas pol II and several other general transcription factors ( e . g . TFIIB , TFIIE , TFIIF ) are relatively well-conserved , Mediator is poorly conserved between yeast and humans ( Table S2 ) . Thus , it is plausible that the human transcription initiation machinery may adopt a distinct architecture relative to that coordinated by yeast Mediator . A summary comparing structural studies with yeast and human Mediator–pol II complexes is shown in Table 2 . The human pol II enzyme contains a C-terminal domain ( CTD ) within its RPB1 subunit that is approximately 500 residues in length and is largely unstructured . The pol II CTD interacts with the human Mediator complex; in fact , the pol II CTD can be used to affinity purify Mediator from partially purified extracts [5] . Immediately adjacent to the RPB4/7 stalk region is the point from which the pol II CTD extends from the pol II enzyme ( asterisk , Figure 5B ) . Because the pol II CTD is unstructured , it cannot be reliably localized within the Mediator–pol II–TFIIF cryo-EM map , nor has it been resolved from pol II crystal structure data . Based upon existing biochemical and biophysical studies , however , we propose the region highlighted green in Figure 5B represents the probable location of the pol II CTD within the PIC . This region corresponds to the site of pol II CTD–Mediator interaction identified previously using EM coupled with antibody labeling experiments with human Mediator bound to the pol II CTD [5] . Furthermore , this proposed pol II CTD location ( Figure 5B ) is proximal to the putative Cyclin H/CDK7 kinase module site within the human PIC ( Figure 5A ) . Phosphorylation of the pol II CTD correlates with transcription initiation and elongation , and CDK7—a TFIIH subunit—is the major pol II CTD kinase within the PIC . The XPB/ERCC3 subunit within human TFIIH has been shown to interact with DNA elements just downstream of the transcription start site [14] , which places TFIIH in an orientation such that its CDK7/Cyclin H lobe would be positioned near the proposed location of the pol II CTD in the Mediator–pol II–TFIIF structure . Although additional structural studies will be required to confirm the precise orientation of TFIIH within the human PIC , such TFIIH–pol II CTD co-localization also supports biochemical data that indicate a Mediator requirement for TFIIH-dependent pol II CTD phosphorylation within promoter-bound , human transcription complexes [9] . During transcription initiation , the newly transcribed RNA exits the pol II enzyme along the pol II RPB4/7 stalk [39] , [40] . Within the Mediator–pol II–TFIIF assembly , the RPB4/7 stalk is oriented such that the nascent RNA could extend unobstructed from the PIC ( Figure 5B ) . Thus , the architecture of the assembly ensures the transcript is readily accessible for capping enzymes and other RNA processing factors . Similarly , the pol II CTD emerges from the pol II enzyme at an exposed site adjacent to the RPB4/7 stalk ( Figure 5B ) . In addition to binding Mediator within the PIC , the pol II CTD serves as an assembly platform for many RNA processing factors ( e . g . capping , splicing , cleavage , and poly-adenylation factors ) and is critical for generating stable , mature transcripts [41] . Based upon the Mediator–pol II–TFIIF assembly structure , RNA processing factors would have unhindered access to the pol II CTD . Upon the transition from initiation to elongation , pol II must break contacts with the PIC . From a structural standpoint , it is currently unclear how pol II makes this transition; however , such a transition likely involves additional structural alterations within Mediator . Based upon the Mediator–pol II–TFIIF assembly structure , a portion of the Mediator head domain ( 1 , Figure 5B ) could pivot with a simple hinge-like motion back toward domain 2 ( Figure 5B ) to facilitate pol II promoter escape . In this way , Mediator could help regulate the pol II transition from initiation to elongation . Interestingly , activator-induced structural shifts within Mediator have been linked to activation of promoter-bound pol II complexes to a productively elongating state , indicating that activators likely contribute to this regulation [9] . Incorporation of additional PIC factors ( e . g . TFIIE , TFIIH ) might also trigger structural shifts in Mediator to facilitate pol II promoter escape . A conformational shift also occurs within pol II itself upon its transition to an elongating state and also when single-stranded DNA enters the active-site cleft [42] . These structural shifts may also disrupt pol II–Mediator contacts to favor promoter clearance and elongation . Further structural and functional studies will be required to better define how pol II–Mediator contacts are affected during the early stages of initiation . Perhaps most striking about the Mediator–pol II–TFIIF structure is that the majority of Mediator would remain exposed even upon assembly of the entire PIC ( Figure 5A ) . Most of the surface area within the “body” of Mediator and all within the “leg” domain would remain accessible for potential protein-protein interactions . One well-established interaction involving the Mediator leg domain is with the 600 kDa CDK8 subcomplex , and biochemical and functional assays reveal the CDK8 subcomplex and pol II interact with Mediator in a mutually exclusive fashion [20] , [32] . For example , the CDK8 subcomplex will not bind Mediator–pol II . The interface between the CDK8 subcomplex and the leg domain of Mediator is extensive and requires the Med13 subunit within the CDK8 subcomplex [20] . In fact , the CDK8 subcomplex contains a hook-like structural domain ( Figure 6A ) that interfaces with a complementary-shaped surface within the leg domain of VP16-Mediator ( Figure 6B ) . This structural complementarity is abolished upon interaction with pol II , despite the fact that pol II binds the Mediator head domain over 100 Å from the leg domain–Med13 interface ( Figure 6C ) . The pol II-induced structural shift within the leg domain is also observed without TFIIF—that is , even when pol II is not stably oriented within Mediator ( Figure 6D and 6E ) . This structural shift in the leg domain likely occludes the Med13 interaction site , thereby regulating Mediator–CDK8 subcomplex interactions . The mutually exclusive CDK8 subcomplex/pol II interactions with Mediator suggest a dynamic exchange at actively transcribing genes . Detachment of pol II from Mediator likely accompanies promoter clearance and transcription elongation and would allow subsequent CDK8 subcomplex–Mediator association [20] . This association would act to prevent a second pol II enzyme from immediately re-engaging the promoter . CDK8 subcomplex–Mediator association following pol II promoter clearance might also enable CDK8-Mediator to regulate transcription elongation . The human CDK8 subcomplex was recently identified as a regulator of transcription elongation for genes within the serum response network [43] , and CDK8-Mediator appears to interact with elongation factors , including P-TEFb [32] . Potentially , pol II might remain in proximity to Mediator during elongation [44] , which would allow a means by which CDK8-Mediator could simultaneously prevent re-initiation of transcription while affecting ongoing elongation events . Although further studies are required to explore this possibility , it is notable that ChIP data indicate Mediator occupancy within coding regions of active genes [45] , [46] , suggesting a juxtaposition of Mediator and pol II elongation complexes .
VP16-Mediator and pol II were individually purified as endogenous complexes from HeLa nuclear extracts , whereas the two subunits of TFIIF , Rap74 and Rap30 , were expressed recombinantly in E . coli and purified as described [20] . For each purification protocol ( Mediator–pol II–TFIIF , Figure 2A , or Mediator–pol II , Figure 2C ) , the same Mediator and pol II samples were used . That is , a single , purified VP16-Mediator sample and a single , purified pol II sample were each split in half , with half of each sample used for the Mediator–pol II–TFIIF experiment and half for the Mediator–pol II experiment . The C-terminal domain ( CTD ) within the largest subunit of pol II , RPB1 , can be extensively phosphorylated and this phosphorylation can negatively impact pol II association with Mediator [47] . Although the majority of the purified pol II sample appeared to be hypo-phosphorylated , we incubated pol II over a phosphatase resin ( Sigma P0762 ) for 4 h at 4°C to ensure complete de-phosphorylation of the pol II CTD . This thoroughly de-phosphorylated pol II sample was used for the purifications outlined in Figure 2A and Figure 2C . The parallel Mediator–pol II and Mediator–pol II–TFIIF preparations were applied to separate glycerol step gradients . The gradient was designed to concentrate the full assemblies in the bottom fraction , while dispersing smaller complexes in earlier fractions . The gradient contained the following amounts of glycerol ( in 20 mM HEPES , 100 µM EDTA , 150 mM KCl , 0 . 02% NP-40 ) from bottom to top: 100 µL 35% , 300 µL 30% , 800 µL 25% , and 800 µL 15% . The gradients were centrifuged at 55 , 000 rpm for 6 h at 4°C . The bottom ( 35% ) glycerol gradient fraction was used for EM studies . Mediator was detected in Western blotting experiments using an antibody to Mediator subunit MED23 ( Bethyl Cat . #A300-425A ) . Pol II was detected using an antibody to RPB1 ( Santa Cruz sc-899 ) , which detects both the hyper- and hypo-phosphorylated forms . TFIIF antibody ( Rap74 ) was purchased from Austral Biologicals ( Cat . #TM-101D-55 ) . The purified VP16-Mediator complex ( ∼1 µg ) and pol II complex ( ∼2 µg ) fractions were precipitated at 4°C using 20% ( v/v ) TCA , 0 . 067 mg/mL insulin , and 0 . 067% ( w/v ) deoxycholate . Precipitated protein pellets were washed twice with −20°C acetone and air dried . Proteins were trypsin digested using a modified Filter-Aided Sample Prep ( FASP ) protocol [48] . Briefly , protein pellets were suspended with 4% ( v/v ) SDS , 0 . 1 M Tris pH 8 . 5 , 10 mM TCEP , and incubated 30 min ambient to reduce disulfides . Reduced proteins were diluted with 8 M Urea , 0 . 1 M Tris pH 8 . 5 . and iodoacetamide was added to 10 mM and incubated 30 min in total darkness . Reduced and alkylated proteins were then transferred to a Microcon YM-30 spin concentrator and washed twice with 8 M Urea , 0 . 1 M Tris pH 8 . 5 to remove SDS . Three washes with 2 M Urea , 0 . 1 M Tris pH 8 . 5 were performed , then trypsin and 2 mM CaCl2 was added and incubated approximately 2 h in a 37°C water bath . Digested peptides were eluted and acidified with 5% ( v/v ) formic acid . Peptides were desalted online and fractionated with a Phenomenex Jupiter C18 ( 5 µm 300 Å; 0 . 25×150 mm ) column using a two-dimensional LC/MS/MS method ( Agilent 1100 ) . Seven steps of increasing acetonitrile ( 3 , 6 , 9 , 12 , 16 , 20 , and 100% B; A: 20 mM ammonium formate pH 10 , 4% acetonitrile; and B: 10 mM ammonium formate pH 10 , 65% acetonitrile ) at 5 µL/minute eluted peptides for a second dimension analysis on a Dionex Acclaim PepMap C18 ( 3 µm 100 Å; 0 . 075×150 mm ) running a gradient at 0 . 2 µL/minute from 5% to 25% B in 100 min for steps one through six and 10% to 30% B in 100 min for step seven ( A: 4% acetonitrile and B: 80% acetonitrile , both with 0 . 1% formic acid pH∼2 . 5 ) . PepMap eluted peptides were detected with an Agilent MSD Trap XCT ( 3D ion trap ) mass spectrometer . All spectra were searched with Mascot v2 . 2 ( Matrix Sciences ) against the International Protein Index ( IPI ) database version 3 . 65 with two missed cleavages and mass tolerances of m/z ±2 . 0 Da for parent masses and ±0 . 8 Da for MS/MS fragment masses . Peptides were accepted above a Mascot ion score corresponding to a 1% false discovery rate ( 1% FDR ) determined by a separate search of a reversed IPI v3 . 65 database . Peptides were then filtered and protein identifications were assembled using in-house software as described [49] . A listing of all polypeptides identified by MS is shown in Table S3 and Table S4 . Reconstituted transcription reactions were completed on a DNA template with tandem GAL4 binding sites assembled into chromatin , as described [20] . Sample ( either Mediator–pol II–TFIIF or Mediator–pol II ) was applied to a glow-discharged carbon-coated copper EM grid ( EMS cat . #CF400-Cu ) and washed twice with 5% trehalose buffer ( 20 mM HEPES , pH 7 . 9 , 0 . 1 mM EDTA , and 100 mM KCl ) . Grids were floated on a droplet of water and then stained with 2% uranyl acetate in water . Images were recorded on Kodak SO-163 film using a Tecnai F20 microscope operated at 200 kV . Untilted ( 0° ) and tilted ( 25°–45° ) specimen images were collected at 29 , 000× magnification with a defocus range of −1 . 0 to −3 . 5 µm . The film was digitized with a sample-scale pixel size of 4 . 29 angstroms . Individual particle images were windowed into 161×161 pixel boxes using the Web interface of the SPIDER image processing software [50] . The untilted images were subjected to unsupervised ( reference-free ) 2D classification based upon a k-means clustering algorithm [25] . For the Mediator-pol II-TFIIF data set , a total of 8 , 923 tilted and untilted pairs were selected . Each 2D class—derived from the untilted data set—contained dozens to hundreds of individual single-particle images . Particles grouped within the same class represented complexes with a similar orientation on the EM grid; particles in different 2D classes represented an alternate orientation of the assembly ( e . g . “side” or “top” views ) or potentially might reflect an alternate conformational state . The corresponding tilted images within each 2D class were back-projected to generate 3D model structures , which were then cross-correlated and subjected to hierarchical clustering using the statistical program package R [51] . One branch of the cluster dendrogram contained well-correlating volumes with images that could be combined into a single converging reference volume , and these data served as the negative stain data set . The tilted and untilted data were used for iterative 2D projection matching to refine the structure . The initial Mediator–pol II–TFIIF reference volume was generated from 3 , 215 tilted images . This reference was then refined with 6 , 430 total images ( tilted + untilted ) using projection matching with angular steps of 15° , 10° , and 5° . The final refinement step included 5 , 047 ( 78% ) of these images . All half-volumes were individually masked to 2 MDa ( to generate non-identical masks ) prior to resolution assessment using the 0 . 5 Fourier shell criterion [52] . The Mediator–pol II reference volume was generated from 3 , 134 tilted images . This reference was then refined with 6 , 268 total images ( tilted + untilted ) . The resolution of this structure improved more slowly , so angular steps of 15° , 10° , 5° , 4° , 3° , and 3° were used . The final refinement step for this volume included 4 , 864 ( 78% ) of these images and the 0 . 5 Fourier shell criterion was applied to determine the resolution [52] . Because an excess of VP16-Mediator was included in the sample preparation , we anticipated some free VP16-Mediator would be present within each data set . To deal with this heterogeneity , efforts were taken to generate negative stain structures of Mediator–pol II–TFIIF and Mediator–pol II free from potential contamination with unbound VP16-Mediator . Images of free VP16-Mediator were removed using a single round of projection matching ( 15° angular step ) with two input reference volumes—the negative stain volume described above and a previously determined structure of free VP16-Mediator [19] . Particle images with a higher correlation to free VP16-Mediator were removed from the data set , and both the Mediator–pol II–TFIIF and Mediator–pol II structures were refined again . This refinement was identical to that described above , except that only untilted images were used for projection matching . The Mediator–pol II–TFIIF structure was refined using 15° , 10° , 5° , and 5° steps with 2 , 390 ( 74% ) images included in the final volume . The Mediator–pol II structure was refined using 15° , 10° , 5° , 5° , and 4° steps with 2 , 223 ( 71% ) images included in the final volume . Cryo-negative stain samples for electron microscopy were prepared largely as described [53] . Purified complexes were added to a glow-discharged thin carbon-coated holey carbon copper mesh grid ( EMS cat . #CF424-50 ) . A buffer containing 5% trehalose , 20 mM HEPES , 100 mM KCl , and 100 µM EDTA was used to wash the grid 3 times to remove excess glycerol . The sample grid was then floated on a drop of stain ( 1 . 2 M ammonium molybdate pH 7 . 5 ) . Excess stain was blotted away and the grid was plunge frozen in liquid ethane to vitrify the sample . Imaging was carried out at liquid nitrogen temperatures under low-dose conditions on a Tecnai F20 FEG microscope ( 200 kV ) . Images were recorded at 29 , 000× magnification on Kodak SO-163 film with a defocus range of −1 . 0 to −4 . 5 µm . Negatives were then digitized to 4 . 29 angstroms per pixel using a Microtek Scanmaker i900 for Mediator-RNA pol II-TFIIF and 4 . 22 angstroms per pixel using a Nikon Super Coolscan 9000 ED for Mediator–pol II . Micrographs were screened for astigmatism and drift using ctfit ( within the EMAN [54] software package ) by removing those with distorted or poorly defined Thon rings in the power spectrum . Individual particle images were manually selected from high-quality micrographs before being windowed into 161×161 pixel boxes for further processing . The contrast transfer function parameters were estimated for each micrograph using the program CTFFIND3 [55] . Using these parameters , individual images were CTF-corrected for phase on a per-micrograph basis . The cryo-EM data set for the Mediator–pol II–TFIIF assembly included 10 , 856 images . The cryo-EM data set for the Mediator–pol II assembly included 7 , 962 images . The appropriate negative stain structures or cryo-EM structures ( generated by three-dimensional variance and sub-classification of images—see below ) , Butterworth filtered to 57 Å , served as initial references for multi-reference projection matching refinement [56] , [57] . Multi-reference refinements were completed as described [57] . The resolution of each reconstruction was calculated using the 0 . 5 Fourier shell criterion [52] or the 3σ-threshold criterion [58] . For Mediator–pol II–TFIIF , using the 0 . 5 Fourier shell criterion , the resolution was found to be 36 Å , whereas using the 3σ-threshold criterion indicated 26 Å resolution . For Mediator–pol II substructure 1 , the 0 . 5 Fourier shell and 3σ criteria specified a resolution of 34 Å and 29 Å , respectively . For Mediator–pol II substructure 2 , the 0 . 5 Fourier shell and 3σ criteria indicated 36 Å and 31 Å resolution , respectively . The reconstructions were displayed after filtering to the values obtained using the 0 . 5 Fourier shell criterion . In order to estimate the degree of structural homogeneity in the cryo-EM Mediator–pol II–TFIIF data set , several multi-reference refinements were completed using different percentages of the data . Volumes were created using 46% , 55% , 62% , or 78% of the 10 , 856 cryo-EM images , based upon their cross-correlation coefficient . The structure generated using 78% of the data converged poorly and resulted in a highly discontinuous structure . The assembly structures generated using 46% , 55% , or 62% of the data converged to very similar solutions that cross-correlated in the 0 . 94–0 . 98 range . Despite containing additional data , the 55% and 62% reconstructions did not result in an improved resolution , and trace discontinuities in the volumes appeared at the 1 . 8 MDa threshold . Consequently , the assembly reconstruction produced using 46% of the data set was used for crystal structure docking and projection matching experiments . Similar refinement trials were used to determine that 59% of the data would be included in the final 3D reconstruction of the Mediator–pol II binary complex . To further probe for alternate structural states within the Mediator–pol II–TFIIF assembly and to potentially separate out structurally distinct conformers , we carried out a 3D variance analysis , developed by Penczek and co-workers [24] . This methodology identifies regions of structural variability by comparing the variance within the structure pixel-by-pixel , relative to background . Focused classification within a region of high variance has the potential to segregate structurally distinct assemblies , especially in cases in which the variance is well-localized and there are a few clearly distinguishable conformational states . Importantly , this technique has proven effective in identifying structural flexibility within a number of multi-subunit complexes , including human transcription complexes [22] , [29] , [57] . Implementation of this approach to the Mediator–pol II–TFIIF data ( i . e . data that partitioned to the Mediator–pol II–TFIIF structure during the multi-reference refinement ) , however , did not improve the resolution of the structure overall . In fact , this approach—which will partition the Mediator–pol II–TFIIF data to generate two substructures—resulted in two virtually identical structures with consistent pol II docking orientations . This is likely indicative of flexibility within the assembly and suggests a relatively stable Mediator–pol II–TFIIF conformational state about which the structure oscillates . Taken together with the de novo cryo-EM reconstruction results described above , these results indicate that although the structure shown in Figure 3A is stable and represents a major entity within the Mediator–pol II–TFIIF cryo-EM data set , the flexible nature of the assembly precludes further improvement of spatial resolution . The 3D variance within the refined Mediator–pol II–TFIIF or Mediator–pol II cryo-EM structures was estimated by creating and comparing 500 re-sampled volumes using a bootstrap technique as described [24] . Focused classification within the area of the 3D reconstruction displaying the highest variance was then used to separate the data into more homogenous groups [22] . Specifically , the refined structure was projected in 28 directions ( angular interval of 25 degrees ) and each image used in the final round of angular refinement was matched to the highest correlating projection . This classification step produced 28 groups of images , each representing a similar orientation of the assembly , or “projection groups . ” The highest peak in the 3D variance density map was then used to generate a mask for classification by projection in each of the directions corresponding to the 28 groups of images . For each projection group , k-means classification within the mask was used to separate the images into two subclasses , which in turn were sorted by visual inspection into one of two groups . Group 1 contained subclasses with less apparent density , and group 2 contained subclasses with more apparent density . The original structure was then subjected to one round of refinement using each group of images to generate two new reference structures . Finally , multi-reference projection matching was used to refine these substructures . Whereas the generation of reference-free classes that closely resembled reprojections of either Mediator–pol II–TFIIF or free VP16-Mediator indicated that both entities were present within the cryo-EM data set ( Figure S4 ) , we wanted to further confirm that each structure ( Figure 3 and Figure S3 ) did not result from model bias [26] during angular refinement . To do this , reciprocal refinements were completed for Mediator–pol II–TFIIF and free VP16-Mediator . In one case , the free VP16-Mediator reference was refined using the data that had previously been partitioned into the Mediator–pol II–TFIIF data set . As expected , pol II density was built into the structure and the general shape of the Mediator–pol II–TFIIF assembly began to emerge during angular refinement . For example , the cleft between lobes 1 and 2 and the cleft between lobe 3 and the body/leg of Mediator became re-defined in the structure ( Figure S7 and unpublished data ) . This suggested that structural features observed for Mediator–pol II–TFIIF do not result from initial model bias . In the second case , the data representing free VP16-Mediator was used to refine an initial Mediator–pol II–TFIIF reference structure . Again , as expected , density representing pol II disappeared from the Mediator head region and the pocket domain located between lobes 1 , 2 , and 3 of Mediator emerged during the refinement ( Figure S7 and unpublished data ) . Because key structural features of the Mediator–pol II–TFIIF and free VP16-Mediator models were preserved even when a reference volume lacking these features was used as a starting point for refinement , model bias did not negatively impact the cryo-EM reconstructions of either the Mediator–pol II–TFIIF assembly or the free VP16-Mediator structure . A yeast crystal structure of pol II ( PDB 1Y1V ) [28] , with the TFIIS fragment removed , was roughly fit into the desired EM structure using Chimera [59] . This visual docking was refined using the FFT-Accelerated 6D Exhaustive Search program of Situs [27] . Using the search tool Colores [60] , an exhaustive search of translational and rotational space was performed and as expected we did not see any dependence of the best docking fit upon the initial position . Because the 12-subunit yeast pol II structure 1Y1V was shown previously to fit well into the human pol II EM structure [29] , the 1Y1V structure was chosen to be displayed . Docking was also completed using each of the complete 12-subunit pol II structures found in the RCSB Protein Data Bank ( 1NT9 , 1PQV , 1WCM , 1Y1W , 1Y1Y , 1Y77 , 2B8K , 2B63 , 2JA5 , 2JA6 , 2JA7 , 2JA8 , 2VUM , 3FKI , 3HOU , 3HOV , 3HOW , 3HOX , 3HOY , 3HOZ , 3K1F ) , with identical results . The orientation of pol II was also determined using 2D projection matching [10] of projection averages of each cryo-EM structure ( Mediator–pol II–TFIIF or Mediator–pol II ) and re-projections generated from a previously published cryo-negative stain human pol II 3D model [29] . The refined 3D reconstruction ( Mediator–pol II–TFIIF or Mediator–pol II ) was projected into 84 directions ( 15°intervals ) and each image included in the reconstruction was matched to the best-correlating projection . Images from each projection group were aligned in 2D and averaged to yield a “view average . ” Because many projections of the Mediator–pol II–TFIIF or Mediator–pol II assembly structures contain overlapping Mediator and pol II densities , density contributions from pol II alone often could not be distinguished . To ensure the most accurate comparison of pol II features , the view average corresponding to the projection of the assembly that had the least amount of Mediator density overlapping with pol II density was used for this analysis . Re-projections of the human pol II structure were generated using a 5° angular step for a total of 1 , 596 projections . Each of these projections was matched to the area of the assembly 2D projection average containing pol II ( as determined by Situs docking ) . A plot was generated to visualize areas of best correlation for a single projection average . The axes of the plot are the phi and theta angles . Larger points indicate higher correlation between a particular 2D re-projection of pol II and the projection average of the assembly structure . For ease of viewing the peak correlations , the radius of each point is proportional to 10 raised to the studentized residual ( 10 ( value - mean ) /standard deviation ) . Each cluster of large points indicates a well-correlating orientation of pol II . The highest correlation was observed with an orientation of pol II that closely matched the 3D docking result ( Figure S6 ) .
|
Transcription initiation in humans is regulated by a macromolecular complex formed by the RNA polymerase II enzyme ( pol II ) , Mediator , and the general transcription factors TFIIA , TFIIB , TFIID , TFIIE , TFIIF , and TFIIH . Collectively , these factors are known as the Pre-Initiation Complex ( PIC ) . Although the 1 . 2 MDa Mediator seems to have a major role in regulating assembly and function of the PIC , a structural understanding of the complex has yet to be established . This study outlines a cryo-EM analysis of the Mediator–pol II assembly in the presence or absence of the dimeric TFIIF complex . We observe that TFIIF is required to stably orient the pol II enzyme within the Mediator–pol II assembly , indicating a novel structural role for TFIIF in transcription initiation . Additionally , we accurately dock the pol II crystal structure within the human Mediator–pol II–TFIIF cryo-EM map . The locations of TFIIH , TBP ( a subunit within TFIID ) , TFIIA , TFIIB , TFIIE , and TFIIF relative to the pol II enzyme itself have been determined by previous studies . These data in combination with the Mediator–pol II–TFIIF structure described here allow us to propose the structural organization of the entire 3 . 5 MDa human PIC .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biophysics/macromolecular",
"assemblies",
"and",
"machines",
"molecular",
"biology/transcription",
"initiation",
"and",
"activation",
"biochemistry/macromolecular",
"assemblies",
"and",
"machines",
"biochemistry/transcription",
"and",
"translation",
"biophysics/transcription",
"and",
"translation"
] |
2011
|
Molecular Architecture of the Human Mediator–RNA Polymerase II–TFIIF Assembly
|
Invasive fungal infections by Candida albicans ( Ca ) are a frequent cause of lethal sepsis in intensive care unit patients . While a contribution of type I interferons ( IFNs-I ) in fungal sepsis remains unknown , these immunostimulatory cytokines mediate the lethal effects of endotoxemia and bacterial sepsis . Using a mouse model lacking a functional IFN-I receptor ( Ifnar1−/− ) , we demonstrate a remarkable protection against invasive Ca infections . We discover a mechanism whereby IFN-I signaling controls the recruitment of inflammatory myeloid cells , including Ly6Chi monocytes and neutrophils , to infected kidneys by driving expression of the chemokines CCL2 and KC . Within kidneys , monocytes differentiate into inflammatory DCs but fail to functionally mature in Ifnar1−/− mice , as demonstrated by the impaired upregulation of the key activation markers PDCA1 and iNOS . The increased activity of inflammatory monocytes and neutrophils results in hyper-inflammation and lethal kidney pathology . Pharmacological diminution of monocytes and neutrophils by treating mice with pioglitazone , a synthetic agonist of the nuclear receptor peroxisome proliferator-activated receptor-γ ( PPAR-γ ) , strongly reduces renal immunopathology during Ca infection and improves mouse survival . Taken together , our data connect for the first time the sepsis-promoting functions of IFNs-I to the CCL2-mediated recruitment and the activation of inflammatory monocytes/DCs with high host-destructing potency . Moreover , our data demonstrate a therapeutic relevance of PPAR-γ agonists for microbial infectious diseases where inflammatory myeloid cells may contribute to fatal tissue damage .
Fungal sepsis is a frequent cause of death in the intensive care unit , with Candida spp . being among the most common microbial pathogens isolated from septic patients . Worldwide , C . albicans ( Ca ) represents the most prevalent isolate recovered from human candidemic patients . The inflammatory response and disease progression of invasive candidiasis presents a similar clinical picture as seen for bacterial sepsis , with some patients developing septic shock and organ dysfunction , the most common of which is acute renal failure [1] . Type I interferons ( IFNs-I ) constitute a family of pleiotropic cytokines that regulate resistance to viruses , enhance innate and adaptive immunity , and modulate cell survival and apoptosis [2] . The most relevant members include IFN-β , with only one member in humans and mice , and the IFN-α family encompassing more than 10 members . While IFNs-I strongly protect against viral infections , they can have either protective or detrimental host effects in bacterial infections depending on the pathogen in question [3] . In comparison , little is known about their contributions in fungal infections . We and others have recently reported that mouse bone marrow-derived dendritic cells ( BM-DCs ) produce IFN-β in response to Candida spp . in vitro [4] , [5] . Notably , the IFN-I receptor subunit IFNAR1 is among the highest upregulated genes in blood leukocytes in a mouse model of invasive candidiasis [6] . Furthermore , similar to bacterial infections , recent reports suggest that IFNs-I are implicated in the in vivo response to fungal pathogens , albeit with opposing effects for the host [7] , [8] . The divergent functions of IFNs-I may relate to their versatile effects on antimicrobial immunity and to their ability to trigger either inflammatory or anti-inflammatory responses depending on the particular pathological situation [9] . Besides modulating innate and adaptive immune responses , IFNs-I play a critical role in promoting lethal endotoxemia and sepsis . For instance , mice lacking the IFN-I receptor ( Ifnar1−/− ) are highly insensitive to LPS- or TNF-induced lethal shock [10] , [11] . Notably , they also show improved survival in a model of septic peritonitis [12] . However , the concept of IFNs-I as adverse mediators of sepsis has been challenged by a recent study using a low lethality model of cecal ligation and puncture-induced sepsis [13] . In this model , Ifnar1−/− mice show an increased late mortality , underlining the complex effects of IFNs-I . The apparent discrepancies might relate to differences in the temporal regulation of the IFN-I response , the amount or subtype of cytokines produced or the microbial species used within a certain model [14] . This work addresses the pathophysiological role of IFN-I signaling during Ca infections using an intravenous ( iv ) mouse challenge model . As in humans , the mouse kidney is the prime target organ , as progressive sepsis concomitant with renal failure account for mortality in that model [15] . Since the severity of kidney tissue damage is quantitatively related to the level of host innate response , it has been suggested that an uncontrolled inflammatory response rather than Ca itself may worsen disease outcome . Indeed , massive infiltrations of neutrophils are commonly observed and believed to contribute significantly to host tissue destruction [16] . Another highly inflammatory cell type frequently associated with host immunopathologies are Ly6Chi inflammatory monocytes [17]–[19] . The chemokine CCL2 recruits inflammatory monocytes to infected body sites , where they exert direct anti-microbial activities or further differentiate into inflammatory DCs [20] . This DC subtype participates in the protective innate response and is characterized by high production of TNF-α and iNOS [21] . Interestingly , IFNs-I have been shown to regulate Ly6Chi monocyte recruitment during viral infections [22] , [23] , as well as during chronic inflammation in mice [19] by inducing Ccl2 . However , the role of inflammatory monocytes in the host response to Ca infections remains unknown . Here , we demonstrate a pivotal role of IFNs-I in triggering the inflammatory host response and immunopathology in experimental candidiasis . We establish a mechanism through which IFNs-I mediate the lethal effects of Ca-induced sepsis . IFN-I signaling stimulates the recruitment of both Ly6Chi inflammatory monocytes and neutrophils from the bone marrow to infected kidneys . Furthermore , IFN-I signaling is required for the subsequent maturation of monocytes into functional inflammatory DCs . Thus , mice competent for IFN-I signaling ( WT ) suffer from unrestrained hyper-inflammation , resulting in lethal kidney pathology . In sharp contrast , Ifnar1−/− mice are remarkably resistant to otherwise lethal Ca infections , and show reduced activity of inflammatory myeloid cells . Strikingly , the pharmacological suppression of inflammatory monocytes and neutrophils by the anti-diabetes drug pioglitazone , a PPAR-γ agonist , strongly reduces renal immunopathology and improves survival of mice , suggesting a novel therapeutic option to combat fungal sepsis . Our data provide a molecular mechanism explaining the manifestation and progression of systemic fungal infections . We suggest that interfering with the activity of inflammatory monocytes and neutrophils provides beneficial effects for disease outcome by suppressing fatal kidney pathology during Ca infections . Importantly , our results are the first report of a central role of the IFN-I-driven CCL2/CCR2 pathway in controlling inflammatory monocyte trafficking during fungal infections .
We have recently shown that Candida spp . induce an IFN-β response in mouse BM-DCs [4] . To test for in vivo production of IFNs-I upon systemic Candida infections , we infected WT and IFNAR1-deficient mice with Ca by lateral tail vein injection ( iv ) . Both infected mouse genotypes produced similar levels of IFN-α with increasing cytokine levels upon disease progression ( Figure 1A ) . We were unable to detect serum IFN-βlevels under the same experimental conditions , since IFN-β is notoriously known for its low expression levels in vivo and thus remained below the detection limit . Nevertheless , BM-DCs from both WT and IFNAR1-deficient mice produced equal levels of IFN-β upon Ca stimulation ( Figure S1A ) . Unlike WT cells , Ifnar1−/− BM-DCs failed to respond to IFN-β as evident from the absence of STAT1 phosphorylation . ( Figure S1B ) . To examine a contribution of IFN-I signaling on the infection outcome and fungal dissemination , we compared the survival of WT and Ifnar1−/− animals infected with various Ca loads . Interestingly , at a dose of 105 Ca colony-forming units ( cfus ) the lack of IFNAR1 caused a remarkable protection to otherwise lethal infections , which became apparent after one week of injection ( Figure 1B ) . The same phenotype was observed after infection with lower fungal doses of 0 . 5×105 Ca cfus ( Figure S1C ) , whereas increasing the fungal loads to 5×105 Ca cfus obliterated the protective effect of IFNAR1-deficiency ( Figure S1D ) . Thus , the beneficial effect of lacking an IFN-I response is only evident under infection conditions that do not overburden host protective capacity . To test whether the improved survival of Ifnar1−/− mice was a result of increased fungal clearance , we determined fungal burdens in the major organs of infected animals . To avoid premature death of WT mice before sample collection , we infected mice with 0 . 5×105 Ca cfus . At indicated time points , spleen , liver , brain and kidneys were collected and analysed for the presence of fungal cells by cfu counting . During the first week of infection , Ca cfus remained high only in the kidneys ( Figure 1C ) , which is in agreement with previous findings [24] , [25] . In all other organs , fungal burden was controlled by the immune system and either stayed low ( brain ) or progressively decreased over time ( spleen and liver ) suggesting successful clearance ( Figure S1E ) . For any organ investigated , we did not observe a significant difference in Ca cfus between WT and Ifnar1−/− mice . We also investigated the in vitro Ca killing capacity of host phagocytes with suspected or known functions in the clearance of fungal infections . Again , we did not observe significant differences in the in vitro Ca killing capacities of phagocytes from WT or Ifnar1−/− mice ( Figure 1D ) . In conclusion , we demonstrate that Ca-induced IFN-I signaling mediates detrimental host effects during disseminated and invasive infections , though not by altering fungal clearance . These data suggest that other mechanisms confer resistance to systemic Ca infections in a host with defective IFN-I signaling . We hypothesized that Ifnar1−/− mice may be more resistant to Ca infections because IFN-I signaling is able to enhance deleterious inflammatory responses . Therefore , we quantified sepsis-relevant inflammatory mediators in serum and kidney , the critical target organ , during the first week of Ca infection . In support of our hypothesis , Ifnar1−/− mice showed significantly reduced serum levels of inflammatory TNF-α and IL-6 when compared to WT mice ( Figure 2A ) . Furthermore , blood counts revealed considerably lower leukocyte numbers in Ifnar1−/− mice , indicating an attenuated infection-induced haematopoiesis . In particular , granulocyte counts were significantly lower ( Figure 2B ) , suggesting an impaired or delayed granulopoesis in knock-out animals . By contrast , total lymphocyte numbers did not differ significantly at any time point ( Figure 2B ) . Taken together , these results demonstrate that IFN-I signaling triggers an early systemic release of inflammatory cytokines and contributes to the Ca-induced robust granulopoesis . As observed for the systemic inflammatory response , early IL-6 levels were significantly reduced at day 1 post infection in kidneys of Ifnar1−/− mice when compared to WT animals ( Figure 2C ) . We also quantified expression levels of additional inflammation mediators with a known involvement in acute kidney injury [26] , including the major adhesion molecules ICAM-1 and P-Selectin . In agreement with attenuated inflammation , we detected reduced renal expression of these adhesins in Ifnar1−/− mice ( Figure 2C ) . The decrease of these critical cytokines and adhesion molecules in kidneys of knock-out mice strongly suggested a reduced immune cell infiltration in the infected organ . To test this notion , we isolated leukocytes from Ca-infected kidneys and evaluated the total number of immune cells ( CD45+ ) and myeloid cells ( CD11b+ ) . In line with the reduced granulopoesis , Ifnar1−/− kidneys displayed significantly lower numbers of infiltrating leukocytes , which could be attributed to the selective reduction of myeloid cells within the tissue ( Figure 2D ) . In summary , our results indicate that IFN-I signaling strongly promotes both systemic and acute local inflammatory responses in Ca-infected mice by enhancing the expression of pro-inflammatory mediators , as well as the recruitment of innate immune cells . To investigate the pathological consequences of the IFN-I-mediated inflammatory response , we examined the immunohistopathology of infected kidneys from WT and Ifnar1−/− mice . Histopathological inspection at day 1 after infection revealed typical Ca-containing abscesses in the renal cortex including massive phagocyte infiltrates . There were no obvious differences in the quality and quantity of fungal lesions in WT vs . Ifnar1−/− mice at this early stage of infection ( Figure 3A ) . However , at day 3 , Ca-containing abscesses were still abundantly present in WT mice but were mainly cleared in IFNAR1-deficient animals ( Figure 3A ) . At day 7 , the inflammatory process in WT animals spread to the renal tubules and pelvis with extensive tubular cast formation and distortion of the renal architecture . By contrast , there were almost no fresh cellular casts detectable in Ifnar1−/− mice ( Figure 3A ) . Notably , at that time point , no more Ca cells were detectable within the renal cortex of WT or Ifnar1−/− mice . Thus , the continuous immune cell recruitment to the renal cortex can be considered an over-reactive host response that may promote kidney pathology . To test whether the observed damage of renal architecture also impaired kidney function , we measured the expression level of kidney injury molecule-1 ( Kim-1 ) in kidneys , as well as urea levels in serum of infected animals . Kim-1 expression strongly increases in de-differentiated renal proximal tubular epithelial cells upon tissue damage , and is thus a suitable biomarker for early kidney injury [27] . Strikingly , Kim-1 expression levels peaked at day 3 of infection and were much higher in WT mice when compared to mice lacking IFNAR1 ( Figure 3B ) . The peak of Kim-1 expression coincided with the presence of Ca lesions in the WT kidney , preceding the subsequent kidney damage , which was evident from the increased urea levels at day 7 ( Figure 3C ) . Taken together , the data let us conclude that IFN-I signaling drives unrestrained inflammatory responses in Ca-infected kidneys , as indicated by the prolonged presence of immune cell infiltrates/fungal abscesses and the abundant formation of cellular casts at later stages of infection . This hyper-inflammatory response increases organ damage as demonstrated by higher expression of Kim-1 and serum urea levels . To identify the cell types contributing to the kidney damage , we thought to determine kidney recruitment kinetics for major immune cell types frequently implicated in inflammatory conditions , including neutrophils [28] , inflammatory monocytes [29] , [30] and T cells [31] , [32] . Therefore , we enriched infiltrating leukocytes from kidneys of Ca-infected WT or Ifnar1−/− mice to characterize and quantify the different leukocyte populations by flow cytometry , using common immune cell markers ( Table S1 in Text S1 ) . As previously observed by others , kidney infections by Ca followed a two-phase innate response [25] . At day 1 , neutrophils and inflammatory monocytes accumulated to equal amounts in kidneys of WT mice ( Figure 4A ) . The influx of inflammatory monocytes was transient , peaking at day 1 and declining thereafter . In contrast , neutrophils showed a second wave of massive infiltration starting between day 3 and 5 post infection ( Figure 4A ) . T cells numbers also increased throughout disease , but remained only a minor fraction of the total cell population within the infected kidney ( Figure S2A and B ) . Whereas the recruitment pattern of CD8+ and CD4+ T cells was comparable between WT and Ifnar1−/− mice ( Figure S2A and B ) , inflammatory phagocyte infiltrates were significantly less in kidneys of knock-out mice ( Figure 4A ) . Interestingly , Ifnar1−/− mice displayed significantly lower numbers of inflammatory monocytes and neutrophils during early infection stages and lacked the late massive neutrophil influx ( Figure 4A ) . We further confirmed the lack of neutrophil influx by measuring MPO levels in kidney homogenates . MPO is an oxidative granular enzyme found primarily in neutrophils and is therefore used to quantify tissue neutrophil content [33] . In agreement with reduced neutrophil numbers , MPO levels remained low in Ifnar1−/− kidneys at day 7 post infection ( Figure S2C ) . We further confirmed the apparent recruitment defect of monocytes and neutrophils using an intraperitoneal ( ip ) Ca infection model , which is commonly used to assess leukocyte infiltrations to sites of infection [34] , [35] . As expected , we observed significantly reduced peritoneal cell infiltrations in IFNAR1-deficient mice ( Figure 4B ) . Since the cell frequencies of peritoneal Ly6Chi monocytes and neutrophils were similar between WT and Ifnar1−/− mice ( Figure 4B ) , we concluded that the recruitment of both cell types is equally affected by the absence of IFN-I signaling . Altered expression levels of the respective cell-specific chemokines also reflected the impaired innate cell recruitment to infected kidneys . Coinciding with the peak of inflammatory monocytes , Ccl2 expression levels were strongly reduced in kidneys of Ifnar1−/− mice ( Figure 4C ) . Likewise , the early expression of the neutrophil chemokine KC was diminished in knock-out animals ( Figure 4D ) . In contrast , the late neutrophil influx seemed to be independent of KC signaling , since both WT and Ifnar1−/− mice displayed comparable levels of the chemokine at day 7 . Locally produced IL-1β represents another chemo-attractant signal for neutrophils during inflammatory conditions [36] . As expected from the increased neutrophil accumulations , we also detected elevated expression levels of IL-1β in kidneys of WT mice ( Figure 4E ) . Increased IL-1β levels coincided with the peak of KIM-1 expression and occurred just prior to the second wave of neutrophil influx . Together , these data indicate that IFN-I signaling promotes the early recruitment of neutrophils and inflammatory monocytes into Ca-infected kidneys . Elevated numbers of inflammatory cells during the first days of infection seem to drive a late massive influx of neutrophils , which is absent in Ifnar1−/− animals . The regulation of inflammatory monocyte trafficking and anti-microbial defense during fungal infections remains unexplored . Emigration of Ly6Chi monocytes from the bone marrow ( BM ) is known to require signaling via the CCR2 receptor . Notably , the induction of the cognate chemokine ligands such as CCL2 , CCL7 , and CCL8 are regulated by IFNs-I in certain infectious diseases [22] , [23] . Since we had observed a strongly reduced expression of CCL2 in Ca-infected kidneys of Ifnar1−/− mice , we thought to investigate the IFNAR1-dependent expression of the chemokine in more detail . Monocytes are known as high producers of chemokines , including CCL2 [22] . Therefore , we stimulated GM-CSF-differentiated BM-DC cultures , which contain about one third Ly6C+ monocytes ( Figure 5A ) , with heat-inactivated Ca and determined the release of the two major monocyte-attracting chemokines CCL2 and CCL7 [37] . Whereas BM-derived WT Ly6C+ monocytes released CCL2 and CCL7 upon Ca challenge , IFNAR1-deficient cells were strongly impaired in their chemokine response ( Figure 5A ) . The release of IFN-β was similar under all conditions , ensuring equal responsiveness of cells ( data not shown ) . We confirmed the IFNAR1-dependent induction of CCL2/CCL7 in WT cells pre-treated with an anti-IFNAR1 blocking antibody prior to Ca stimulation ( Figure S3A ) . As a control , pre-treatment of WT cells with an unspecific isotype IgG did not alter CCL2 production ( data not shown ) . In line with the kidney chemokine data , we also observed a reduced expression of two neutrophil-attracting chemokines , KC and MIP-2 , in Ca-stimulated Ifnar1−/− BM-DC cultures ( Figure 5A ) , although neither chemokine is a classical IFN-stimulated gene . Interestingly , simply blocking IFNs-I signaling with the anti-IFNAR1 antibody was not sufficient to reduce the release of KC and MIP-2 ( Figure S3A ) , suggesting that the physical presence of a functional IFN-I receptor , rather than its signaling function , cooperates with other pathways for full activation of neutrophil-specific chemokines . During infections with microbial pathogens , the initial CCL2 production is triggered in the BM , where it mediates inflammatory monocyte egression into the blood stream [4] . Therefore , we examined whether Ca could induce the local production of CCL2 in the BM . To determine whether Ca resides within the BM during infection , we measured cfu counts of infected WT and Ifnar1−/− mice . In accordance with previous reports [38] , we confirmed that fungal cells are present in the BM . As for other organs , the BM fungal burden did not differ between WT and knock-out animals ( Figure 5B ) . To test whether Ca-infected BM releases CCL2 , we isolated BM cells from infected WT and Ifnar1−/− mice at day 1 and cultured them for a total of 3 days . At the time of BM isolation cellular composition and viability of cells were the same between both mouse genotypes ( data not shown ) . Thereafter , we quantified the daily release of CCL2 into the media . Again , we observed a requirement of IFNs-I for full CCL2 production in Ca-infected BM , which became even more evident after longer periods of culture ( Figure 5C ) . Similar results were obtained for the gene expression levels of Ccl2 and Ccl7 in infected BM . At day 3 post infection , the induction of monocyte-attracting chemokines was strongly reduced in absence of IFNAR1 signaling ( Figure S3B ) . We hypothesized that the diminished CCL2/CCL7 production in the BM of Ifnar1−/− mice may result in a mobilization defect of inflammatory monocytes . Therefore , we determined the number of circulating inflammatory monocytes in the blood of infected WT and Ifnar1−/− mice . Indeed , we found significantly reduced inflammatory monocyte counts in the blood of IFNAR1-deficient mice ( Figure 5D and 5E ) . In summary , these data establish CCL2/CCL7 as IFN-I-dependent chemokine signals driving the host response during systemic candidiasis . Through regulating the expression of these chemokines in the BM and the kidneys , IFNs-I contribute to the mobilization of inflammatory monocytes into the blood stream and their subsequent migration to the target organ . In the context of inflammation , Ly6Chi monocytes differentiate into inflammatory DCs within the tissue [39] . Thus , we were interested to test if monocyte-derived inflammatory DCs are also generated during invasive Ca infections . Therefore , we further characterized Ly6Chi kidney monocytes for common inflammatory DC surface and activation markers using flow cytometry . Cells from both WT and Ifnar1−/− mice expressed high levels of CCR2 , the hallmark receptor of inflammatory monocytes , as well as the two major DC markers CD11c and MHCII ( Figure 5F ) . In infected WT kidneys , inflammatory DCs showed upregulated expression of the activation marker PDCA1 ( Figure 5F ) , as well as extensive intracellular iNOS staining ( Figure 5G ) . Strikingly , in the absence of IFN-I signaling , PDCA1 expression and iNOS+ cell numbers were significantly reduced . The defect in inducing iNOS gene expression was also confirmed by qPCR in infected kidneys of Ifnar1−/− mice ( Figure 5H ) . We confirmed these results in vitro by stimulating GM-CSF-differentiated BM-DCs with heat-inactivated Ca . As for inflammatory DCs in the kidney , we observed a lack of PDCA1 and iNOS expression in the IFNAR1-deficient Ly6C+ monocyte population ( Figure S3C ) . Similar results were obtained in WT cultures that had been pre-treated with an anti-IFNAR1 blocking antibody prior to Ca-challenge ( Figure S3D ) . Again , the levels of CD11c or MHC class II were not affected by the absence of IFN-I signaling ( Figure S3C and D ) . All together , we demonstrate that Ly6Chi monocyte differentiate into inflammatory DCs during invasive Ca infections . Whereas the initial differentiation into DCs seems to be independent of IFN-I signaling , our results suggest a novel role for IFNs-I in the functional maturation of inflammatory DCs based on the appearance of specific activation markers . IFNs-I promote the inflammation-associated lethal kidney pathology during systemic Ca infections by stimulating the recruitment and activation of Ly6Chi monocytes and neutrophils . However , both cell types are essential for fungal clearance and total elimination renders mice hyper-susceptible to infection [40] . Based on our findings , we hypothesized , that interfering with the activity of these cell types rather than the total elimination would ameliorate host tissue damage and thereby improve the overall outcome of infection . Published evidence suggests that the drug pioglitazone , a synthetic agonist of the nuclear receptor PPAR-γ , impairs inflammatory DC trafficking and associated lung pathology in an influenza mouse model [17] . To test our hypothesis , we treated mice daily with 5 mg/kg pioglitazone or vehicle starting on the day of Ca infection . Strikingly , pioglitazone-treated mice were strongly protected against lethal Ca challenge and experienced a reduced weight loss when compared to non-treated animals ( Figure 6A ) . No significant differences in kidney fungal loads where found between the two mouse groups ( data not shown ) . To determine whether the protective effect of pioglitazone correlated with an inhibition of inflammatory myeloid cell recruitment , we determined Ly6Chi monocyte and neutrophil numbers in blood and kidneys . In the bloodstream , pioglitazone treatment specifically reduced inflammatory monocyte numbers , while total granulocyte numbers were not affected ( Figure S4A ) . However , in the kidneys , early accumulation of both inflammatory monocytes and neutrophils was equally reduced in drug-treated animals ( Figure 6B ) . In addition to cell recruitment , we investigated the effect of pioglitazone treatment on the activation of inflammatory DCs in infected kidneys . In line with previous reports identifying iNOS as one of the main target genes of PPAR-γ mediated repression [41] , we detected significantly reduced iNOS+ inflammatory DCs in kidneys of drug-treated mice ( Figure 6C ) , suggesting that pioglitazone also interferes with the functional maturation of inflammatory DCs . The inhibition of iNOS expression could be reproduced in vitro , since stimulating pioglitazone pre-treated BM-DCs with heat-inactivated Ca ( Figure S4B ) led to similar reductions in iNOS expression . Strikingly , the reduced early cell infiltration in kidneys resulted in the absence of the late detrimental neutrophil influx ( Figure 6D ) . Hence , pioglitazone treatment improves kidney damage by diminishing local inflammation , which was evident by reduced serum urea levels at day 5 ( Figure 6E ) . Although there was a general decrease in the magnitude of inflammation in drug-treated animals , the measurements for IL-6 and Ccl2 in blood and kidney ( Figure S4C ) did not reach statistical significance due to high variability between mice . Nevertheless , pioglitazone pre-treatment significantly reduced CCL2 and CCL7 release from Ca-stimulated BM-DCs in a dose-dependent manner ( Figure 6F ) , whereas production of KC and MIP-2 was not influenced by the treatment ( Figure S4D ) . A cytotoxic effect of pioglitazone was excluded by live-dead staining of cells after drug treatment ( Figure S4E ) . Interestingly , in addition to monocyte-attracting chemokines , pioglitazone also decreased the initial release of IFN-β by BM-DCs ( Figure 6F ) . This observation prompted us to investigate whether the reduced release of CCL2/CCL7 results from the impaired IFN-β production . Therefore , we pre-treated Ifnar1−/− BM-DCs with pioglitazone and examined if the remaining release of chemokines from these cells was further decreased by the drug . Pioglitazone treatment was able to even further suppress the production of CCL2 in IFNAR1-deficient cells ( Figure S4F ) . Therefore , the suppressive function on chemokine expression must be a direct effect of the PPAR-γ-mediated transcription inhibition of those genes [42] , without involving secondary IFN-β signaling . Taken together , our results suggest that the usage of PPAR-γ agonists might represent a novel therapeutic option to improve survival of a host in the setting of an emerging fungal sepsis . On the basis of our collective data , we therefore propose the following model for the IFN-I-regulated detrimental activity of inflammatory myeloid cells during experimental candidiasis ( Figure 7 ) . Recognition of Ca by innate immune cells in the bone marrow initiates the IFN-I–CCL2 cytokine-chemokine cascade and stimulates mobilization of inflammatory monocytes from the bone marrow to the infected organs . Simultaneously , infection-triggered granulopoiesis induces the proliferation and mobilization of neutrophils . Monocytes and neutrophils migrate towards the infected kidneys where CCL2 and KC are produced in an IFNAR1-dependent manner . Inside the renal tissue , inflammatory monocytes acquire a DC-like phenotype through a process not requiring IFN-I signaling . However , the subsequent activation of inflammatory DCs to become high producers of iNOS strictly depends on a functional IFN-I response . High abundance and activity of inflammatory cells causes early tissue damage , leading to the subsequent massive accumulation of neutrophils , whose destructive power ultimately leads to kidney failure . The genetic deficiency of IFNAR1 or the pioglitazone-mediated pharmacological suppression of inflammatory cell recruitment and activation strongly improves Ca-mediated immunopathology and survival of the host .
The present study demonstrates a detrimental effect of IFNs-I during invasive experimental candidiasis . We show that IFN-I signaling stimulates the recruitment of inflammatory innate immune cells , including Ly6Chi monocytes and neutrophils , into the infected kidney , which promote lethal hyper-inflammatory and tissue-damaging immune responses in mice . Importantly , we establish a pharmacological approach by interfering with the activity of those inflammatory cells , which ameliorates disease progression and improves survival of otherwise lethal fungal infections . In many microbial infections , a disregulated or overshooting immune response rather than the pathogen itself can cause fatal host damage . Likewise , studies comparing host responses to attenuated and virulent Ca strains suggest that fungus-induced inflammation contributes considerably to tissue damage and mortality [16] . The lack of IFN-I signaling in Ifnar1−/− mice constrains hyper-inflammatory immune reactions and improves survival of mice infected with Ca . Hence , our data support the notion that uncontrolled host responses rather than the fungal growth are a primary cause of death in the murine Candida infection model . Since Ifnar1−/− mice exhibit similar fungal burdens in critical organs when compared to WT animals , the sepsis resistance phenotype may result from increased host tolerance to the pathogen burden . Therefore , the Ifnar1−/− mouse model would represent an excellent tool to study virulence factors of Ca , allowing for the mechanistic distinction between tissue damage caused by fungal dissemination versus the hyper-inflammatory host response . Similar to bacterial infections , IFNs-I can exert both beneficial and detrimental effects during fungal diseases [7] , [8] . We have previously reported a role for IFNs-I in supporting persistence of Candida glabrata ( Cg ) within the host . Contrasting with their effects in our Ca infection model , fungal persistence of Candida glabrata may arise from an IFN-I-mediated attenuated host response , which facilitates immune evasion by the fungus [4] . This of course implies that specific “patterns of pathogenesis” are exploited by different pathogens , including the use of distinct virulence strategies to generate pathogen-specific contextual cues during infection , which decide about the balance of pro- versus anti-inflammatory functions of IFNs-I . Although a recent study identified Ifnar1−/− mice as hyper-susceptible to disseminated Ca infections , different fungal strains and infection doses , as well as mouse genetic backgrounds , may explain the discrepancy [5] . Indeed , different microbial loads and injection routes may influence the outcome of infection regardless of the genetic phenotype in the host , which normally determines the capacity of immune surveillance [43] , [44] . IFNs-I may dictate a pro- or anti-inflammatory response in host cells depending on the initial IFN-I concentrations targeting responder cells [45] , [46] . We report severely reduced levels of inflammatory mediators and innate immune cells in blood and kidneys of Ifnar1−/− mice upon Ca infection , suggesting a pro-inflammatory role of IFNs-I in this infection model . Notably , IFNs-I have been reported to exert an inhibitory effect on inflammasome activation and the subsequent release of pro-inflammatory IL-1βin mice . The resulting impairment of inflammation-mediated protective immunity renders mice hyper-susceptible to invasive Ca infections [47] . In this mouse model , however , massive IFN-I production was pre-induced with repeated injections of poly-IC prior to fungal infection . We believe that the mechanistic difference to our model is explained by the level of IFNs-I induced between the different challenges , thereby changing the ratio of inflammatory versus anti-inflammatory effector functions . Indeed , in our infection model , where IFNs-I are not produced prior to fungal challenge , we are unable to detect an IFN-I-mediated inhibition of IL-1βproduction . Consistent with our model , we observe unchanged or even lower concentrations of IL-1β in kidney homogenates of Ifnar1−/− mice . Invasive Ca infections progress mainly in the kidneys and the consequent organ damage likely contributes to the overall pathology of disease [48] . Thus , we focused on the immune response and pathology of the kidney to further investigate the organ-specific inflammatory response to Ca . Interestingly , we show a significantly reduced early accumulation of myeloid cells , including inflammatory monocytes and neutrophils , and a lack of the second massive neutrophil infiltration in kidneys of IFNAR1-deficient animals . While IFN-I signaling seems to play a general role in promoting monocyte recruitment , as evident from other studies [19] , [22] , [49] , their effects on neutrophil trafficking and function are quite divergent between different infection models [12] , [13] , [50] . Similar to our Ca infection model , IFNs-I mediate the recruitment of both inflammatory monocytes and neutrophils into kidneys during experimental glomerulonephritis [51] . Also in this model , the consequent higher abundance of myeloid cells worsens disease pathology , indicating the detrimental role of monocytes and neutrophils in kidney function . In future studies it will be interesting to delineate the individual contributions of monocytes and neutrophils to Ca-induced kidney immunopathology by adopting antibody-mediated depletion strategies . One may speculate that early tissue damage caused by the activity of inflammatory innate cells triggers the second wave of neutrophil influx in our infection model . There is evidence that danger signals generated through cellular damage at the site of inflammation attract neutrophils . Many of these danger signaling pathways converge on IL-1β as a key orchestrator of inflammation and cell recruitment [36] . Indeed , we detect higher IL-1β expression levels in kidneys of WT animals when compared to Ifnar1−/− mice . Likewise neutrophils , inflammatory monocytes/DCs have been implicated in the immunopathology of various diseases , including infectious or autoimmune diseases [52] . In other fungal infection models , including respiratory aspergillosis and cryptococcosis , monocytes contribute to the instruction of protective T cell immunity and Th cell differentiation [53] , [54] . Inflammatory monocytes emigrate from the BM through CCR2 receptor-mediated signaling and accumulate at the sites of infection [37] , [55] . We show here that IFNs-I control the recruitment of Ly6Chi monocytes by inducing CCL2/CCL7 expression both in the BM and the target organ , strongly increasing inflammatory monocyte counts in blood and kidneys of Ca-infected WT mice . Bone marrow mesenchymal stem and progenitor cells are the initial producers of CCL2 and inducers of monocyte emigration during infections [56] . Notably , during infections with Listeria monocytogenes , IFNAR1 does not contribute to the early monocyte emigration from the BM within the first 24 hours [57] . Our data confirm the early IFNAR1-independent induction of monocyte-attracting chemokines in the BM during the early response . However , at later stages of infection , IFNs-I become central mediators of CCL2/CCL7 production and monocyte egression . Hence , the IFN-I-driven CCL2/CCR2 pathway is of general importance for controlling inflammatory monocyte recruitment during microbial infectious diseases . Due to a defect of monocyte egression from the BM , mice lacking CCL2 or the cognate receptor CCR2 retain most monocytes in the BM [37] , [55] . In contrast to Ccr2−/− or Ccl2−/− mice , Ifnar1−/− mice do not show this retention phenotype , most likely because IFN-I signaling might also affect the primary differentiation or proliferation of monocytes within the BM [58] . Upon arrival in the infected tissue , Ly6Chi monocytes further differentiate into inflammatory DCs to execute their anti-microbial defense [21] , [55] . In addition to the impaired recruitment of inflammatory monocytes , Ifnar1−/− mice show defects in the activation of inflammatory DCs , as evident by their reduced PDCA1 and iNOS expression . Whereas iNOS is a well-established functional marker of inflammatory DCs , a PDCA1+ inflammatory-like DC subset has been only recently associated with higher expression of pro-inflammatory cytokines and increased T cell-stimulatory potential [59] . IFNs-I have been known for their capacity to differentiate human blood monocytes in vitro into a special DC subset , termed IFN-DCs [60] . Interestingly , the ability of IFNs-I to tip the balance in monocyte terminal differentiation has been also noted during chronic inflammation [19] , where IFNs-I sustain the inflammatory conditions by inhibiting the terminal differentiation of monocytes into anti-inflammatory tissue macrophages , perhaps by promoting their differentiation into inflammatory DCs . Since inflammatory monocytes seem to carry out several central functions in anti-fungal immunity , including propagation of inflammation and T cell instruction , it will be of particular interest to establish a possible role of these cells during fungal diseases in the patient setting , including the search for genetic polymorphisms in CCR2 or CCL2 that might modulate the outcome of fungal infections . Notably , SNPs in STAT1 , a key mediator of both IFN-I and IFN-II ( IFN-γ ) signaling , are implicated in the outcome of invasive fungal infections in humans [61] . Here , we show that pharmacological suppression of monocytes and neutrophils as achieved by treating mice with pioglitazone , a synthetic agonist of the nuclear receptor PPAR-γ , can rescue animals from Ca-mediated immunopathology . Stimulating PPAR-γ antagonizes inflammatory responses by repressing the transcriptional activation of NFκB target genes , including CCL2 , TNF-α and iNOS [62] , [63] . Interestingly , a prophylactic pioglitazone treatment has been recently adopted to reduce detrimental inflammatory DC infiltration into the lungs during influenza virus infections [17] . Strikingly , we demonstrate here that pioglitazone treatment of Ca-infected mice reduces the early accumulation of inflammatory monocytes , as well as neutrophils in kidneys . Furthermore , treatment also strongly impairs the neutrophil influx at later stages of infection . The drug also suppresses the activation of inflammatory DCs , thereby diminishing organ inflammation . Thus , pioglitazone treatment precisely phenocopies the observations in Ifnar1−/− mice , including the reduced inflammatory cell recruitment and functional maturation of inflammatory DCs . Activation of individual nuclear receptors is believed to repress specific subsets of inflammatory target genes with different functions , resulting in distinct biological consequences for the host response [64] . Notably , our data indicate that pioglitazone represses a specific gene subset in vivo , which seems to overlap with the IFN-I target genes driving the lethal inflammation during invasive candidiasis . Previous studies have revealed that pioglitazone treatment rescues mice from lethal influenza virus infections and reduces disease activity of septic peritonitis or murine lupus [17] , [62] , [65] . In support of the beneficial effects of pioglitazone , we show that treatment of Ca-infected mice ameliorates renal immunopathology and improves animal survival . Our data suggest that the modulation of Ly6Chi monocyte and neutrophil numbers , as observed for Ifnar1−/− and pioglitazone-treated mice , has beneficial effects for the host by dampening the hyper-inflammation . The detrimental immunopathology most likely results from the biphasic recruitment and activation of inflammatory monocytes and neutrophils , both of which contribute to the fatal kidney pathology . In support of our hypothesis , depletion of Gr1+ cells at day 7 of Ca infection , which removes both monocytes and neutrophils , increases survival of mice [40] . Taken together , our work identifies IFN-I signaling as a central mediator of inflammatory innate immune cell migration into infected organs , demonstrating a pivotal yet detrimental role for IFNs-I in fungal pathogenesis in vivo . The inflammatory cascade driven by the sustained expression of IFN-I-regulated inflammatory genes substantially contributes to tissue damage observed in infected mice . The pharmacological suppression of inflammatory monocytes and neutrophils shows that interfering with those cell types during invasive Ca infections improve immunopathology and disease outcome . Thus , this work expands the spectrum of detrimental inflammatory immune cells during Ca infections beyond neutrophils . Our study provides a novel mechanism for the role of IFNs-I in sepsis progression , coupling IFN-I target genes such as CCL2 and iNOS to the recruitment and activation of inflammatory monocytes/DCs with considerable host-destructive potential . The direct connection between IFNs-I and inflammatory monocytes might be of general importance in microbial diseases where an IFN-I response is detrimental for the host . We propose that the beneficial effects of PPAR-γ agonists may also apply to other infectious diseases where inflammatory myeloid cells promote tissue damage , including parasitic infections or tuberculosis [18] , [66] .
All animal experiments were discussed and approved through the University of Veterinary Medicine Vienna institutional ethics committee and carried out in accordance with animal experimentation protocols approved by the Austrian law ( GZ 680 205/67-BrGt/2003 , GZ-BMWF-68 . 205/0233-II/10b/2009 and GZ-BMWF-68 . 205/0231-II/3b/2011 ) . The Candida albicans ( Ca ) strain used in this study was the standard clinical isolate SC5314 [67] . Fungal cells were grown to the logarithmic growth phase in single-use , pyrogen- and endotoxin-free sterile flasks . For detailed information see supplemental text and materials . Ifnar1−/− mice on C57BL/6 background [68] were bred at Biomodels Austria , University of Veterinary Medicine . C57BL/6 wild type controls were purchased from Charles River . Mice were housed under specific pathogen-free conditions according to FELASA guidelines . Male mice were challenged on day 0 via the lateral tail vein with 1×105 Ca colony-forming units ( cfus ) per 21 g body weight , if not otherwise stated . Fungal load was adjusted to individual mouse weights . For survival experiments , mice were monitored for 14–35 days . Groups of mice were sacrificed on different days post infection ( p . i . ) for analysis of macroscopic and histological changes , fungal organ burden , and changes in the levels of cytokines and other inflammatory mediators in blood as well as organ homogenates . Mice were treated daily with 5 mg/kg pioglitazone in 0 . 5% methylcellulose/PBS via ip injections starting on day 0 with the Ca infections . The control mice received vehicle only . Weight loss was monitored every other day as a measure of morbidity . Mice were sacrificed and spleen , liver , kidneys , and brain were removed aseptically at necropsy , rinsed with sterile PBS , weighted , and placed in 1 . 5 ml sterile tissue lysis buffer ( 200 mM NaCl , 5 mM EDTA , 10 mM Tris , 10% glycerol , 1× protease inhibitor cocktail ( Roche ) ) on ice . The organs were aseptically homogenized using an Ika T10 basic Ultra-Turrax homogenizer ( Ika , Staufen ) . Serial dilutions of homogenates were plated in triplicate on YPD ( 1% yeast extract , 1% peptone , 2% dextrose ) plates containing ampicillin , tetracycline , and chloramphenicol . Colonies were counted after 48 h of incubation at 30°C . The fungal burden was calculated as cfus per gram of tissue . To assess kidney tissue damage , levels of blood urea ( blood urea nitrogen - BUN ) were determined in serum samples by a routine veterinarian diagnostics laboratory ( InVitro GmbH ) . For haematology , blood samples were collected in K-EDTA-coated tubes ( Sarstedt ) and analysed with an automated blood counter ( V-Sight , A . Menarini ) . For histology , parts of organs were fixed with buffered 4% paraformaldehyde , and paraffin-embedded sections were stained with hematoxylin-eosin ( HE ) or periodic acid-Schiff ( PAS ) stain according to standard protocols . The amount of IFN-β released in cell culture supernatants was assayed using the Verikine mouse IFN-β ELISA kit ( R&D systems ) . Serum IFN-α was measured using the Luminex system with Procarta Cytokine Profiling Kits; TNF-α and IL-6 in serum were determined using the Mouse CBA flex sets ( BD Biosciences ) ; CCL7 , KC , and MIP-2 using Procarta Immunoassays ( Panomics-Affymetrix ) ; CCL2 using a commercial ELISA Set ( Biolegend ) ; all according to the manufacturer's instructions . For cytokine quantification in tissues , organ homogenates of Ca-infected mice prepared as described for fungal burden determination were centrifuged twice ( 1 , 500× g , 15 min , 4°C ) and diluted prior to measurement . Myeloperoxidase ( MPO ) was determined by the Mouse MPO ELISA kit ( Hycult Biotechnology ) ; IL-6 and KC chemokine were measured using the Mouse CBA flex sets ( BD Biosciences ) ; all conditions were according to the manufacturer's instructions . Blood preparation and leukocyte enrichment from kidneys were performed as described in the supplemental material . Blood and kidney leukocytes were stained with the appropriate combination of FITC-labeled anti-Ly6G ( 1A8 , Biolegend ) or anti-CD3 ( 145-2C11 , BD Biosciences ) , PE-labeled anti-PDCA1 ( eBio129c , eBioscience ) , anti-CD4 ( RM4-5 , BD Biosciences ) , APC-Cy7-labeled anti-CD45 ( 30-F11 , BD Biosciences ) , PerCP-Cy5 . 5-labeled anti-CD11c ( N418 , Biolegend ) , Pacific Blue-labeled anti-Ly6C ( HK1 . 4 , Biolegend ) , BD Horizon-labeled CD11b ( M1/70; BD Biosciences ) or anti-CD8 ( 53-6 . 7 , BD Biosciences ) , after blocking of Fc receptors with anti-CD32/CD16 ( 93 , eBioscience ) . Intracellular staining of iNOS was performed with anti-NOS2 ( M-19 , Santa Cruz ) and DyLight 649-conjugated anti-rabbit IgG ( Jackson Immuno Research ) according to the application guidelines of BD Bioscience . Anti-CCR2 ( MC-21 ) was kindly provided by Matthias Mack . Data were acquired using a FACSAria ( BD Biosciences ) and analysed using the FlowJo software ( Tree Star ) . Leukocyte characterization was performed on gated CD45+ cells . Leukocytes were further identified according to cell-specific markers as listed in Table S1 in Text S1 . For the preparation of exudate neutrophils and peritoneal macrophages ( Mphs ) , C57BL/6 mice were ip-injected with 0 , 5 ml 10% proteose peptone ( Sigma , St . Louis , MO , USA ) /PBS . After 4 h ( for neutrophils ) or 3 days ( for peritoneal macrophages ) , the peritoneum was flushed with 7 ml PBS containing 50 U/ml heparin to collect cells . For isolation of resting Mphs , peritoneum of untreated mice was flushed . For isolation of bone marrow resident neutrophils , bone marrow was collected from mice . After lysis of red blood cells , samples were separated on a discontinuous Percoll gradient: from bottom to top 78% , 69% , and 52% Percoll ( GE Healthcare ) . Gradient was centrifuged at 1500×g and 4°C for 30 min . Cells accumulating in the interphase between the 78% and 69% Percoll layer were collected as neutrophils . All different types of immune cells were plated in RPMI supplemented with 10% heat-inactivated FCS and used for co-culture with fungi . For the in vitro Ca killing assay , innate immune cells were plated in replicates at a density of 1×105 cells/well of 96-well plates . Cells were incubated with Ca at indicated MOIs and for indicated time . After incubation , mammalian cells were lysed by addition of Triton X 100 to a final concentration of 1% . After lysis , wells were extensively scrapped , 2× washed with PBS and surviving Ca was determined by plating serial dilutions of the collected media and washes in duplicates on YPD plates containing ampicillin ( Sigma ) . The percentage of killing was calculated according to the following formulas ( df = dilution factor ) : Male mice were injected intraperitoneally with 1×107 Ca colony-forming units ( cfus ) . After 4–6 hours , peritoneal cells were collected with sterile PBS , and the total cell number was assessed in a CASY counter . Cells were stained for flow-cytometry analysis with FITC-labeled anti-Ly6G ( 1A8 , Biolegend ) , Pacific Blue-labeled anti-Ly6C ( HK1 . 4 , Biolegend ) , and BD Horizon-labeled CD11b ( M1/70; BD Biosciences ) . RNA sample preparation , reverse transcription and real-time PCR were performed as described in the supplemental material . Relative quantification was performed with the ΔΔCt-method . Expression level of the genes of interest were normalised to the expression level of the housekeeping gene HPRT . Real-time PCR data are expressed as fold increase of mRNA expression over baseline levels ( uninfected mice ) . All primers used in this study are listed in Table S2 in Text S1 . Statistical analysis of data was performed using the Prism graphing and analysis software ( Graphpad ) . Survival data were compared using the logrank test . Candida cfu data were analysed using the non-parametric Mann-Whitney-test . Time-kinetic comparisons of WT and Ifnar1−/− mice data at every time point were performed using Two-way ANOVA followed by a Bonferroni post-test . Two-group comparisons were done with the Student's t test . In all cases , P<0 . 05 was considered significant . *p<0 . 05; **p<0 . 01; ***p<0 . 001; ns , not significantly different . Bst2 ( PDCA1 ) : ENSMUSG00000046718 , Ccl2: ENSMUSG00000035385 , Ccl7: ENSMUSG00000035373 , Ccr2: ENSMUSG00000049103 , Cd4: ENSMUSG00000023274 , Cd8a: ENSMUSG00000053977 , Cxcl1 ( KC ) : ENSMUSG00000029380 , H2-D1 ( MHC II ) : ENSMUSG00000073411 , Havcr1 ( KIM-1 ) : ENSMUSG00000040405 , Icam1: ENSMUSG00000037405 , Ifna2: ENSMUSG00000078354 , Ifnar1: ENSMUSG00000022967 , Ifnb1: ENSMUSG00000048806 , Il1b: ENSMUSG00000027398 , Il6: ENSMUSG00000025746 , Itgam ( CD11b ) : ENSMUSG00000030786 , Itgax ( CD11c ) : ENSMUSG00000030789 , Ly6c1: ENSMUSG00000079018 , Mapk14 ( p38 ) : ENSMUSG00000053436 , Mpo: ENSMUSG00000009350 , Nos2 ( iNOS ) : ENSMUSG00000020826 , Ptprc ( CD45 ) : ENSMUSG00000026395 , Selp ( P-Selectin ) : ENSMUSG00000026580 , Stat1: ENSMUSG00000026104 , Tnf: ENSMUSG00000024401
|
Inflammation constitutes a major host response in many microbial infections . Innate immune cells orchestrate the inflammatory response to kill pathogens and clear infections . However , invasive infections by pathogenic microbes including the fungus Candida albicans , can result in an uncontrolled hyper-inflammatory response , leading to severe host damage and sepsis . Type I interferons constitute a hallmark of protective innate immunity in viral and bacterial infections , but at the same time have been notoriously known for their sepsis-promoting effects in numerous experimental inflammation models . Here , we show that type I interferon-signaling mediates the lethal hyper-inflammatory response during systemic mouse infections with C . albicans . Following fungal infections , type I interferons promote the recruitment and activation of inflammatory monocytes and neutrophils to infected organs . The high abundance and activity of inflammatory phagocytes lead to fatal tissue damage . Remarkably , we show that the pharmacological suppression of these inflammatory cells with the drug pioglitazone reduces immunopathology and sepsis-related lethality , suggesting a novel therapeutic option to combat fungal sepsis . In conclusion , our data couple the sepsis-promoting role of type I interferons to the host-destructive activity of inflammatory monocytes and neutrophils . We propose that therapeutic approaches dampening hyper-inflammation might be of general importance in microbial diseases where deleterious immunopathology occurs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"infectious",
"diseases",
"inflammation",
"clinical",
"immunology",
"immunity",
"immunology",
"biology",
"fungal",
"diseases"
] |
2012
|
Type I Interferons Promote Fatal Immunopathology by Regulating Inflammatory Monocytes and Neutrophils during Candida Infections
|
DNA double-strand breaks are repaired by multiple mechanisms that are roughly grouped into the categories of homology-directed repair and non-homologous end joining . End-joining repair can be further classified as either classical non-homologous end joining , which requires DNA ligase 4 , or “alternative” end joining , which does not . Alternative end joining has been associated with genomic deletions and translocations , but its molecular mechanism ( s ) are largely uncharacterized . Here , we report that Drosophila melanogaster DNA polymerase theta ( pol theta ) , encoded by the mus308 gene and previously implicated in DNA interstrand crosslink repair , plays a crucial role in DNA ligase 4-independent alternative end joining . In the absence of pol theta , end joining is impaired and residual repair often creates large deletions flanking the break site . Analysis of break repair junctions from flies with mus308 separation-of-function alleles suggests that pol theta promotes the use of long microhomologies during alternative end joining and increases the likelihood of complex insertion events . Our results establish pol theta as a key protein in alternative end joining in Drosophila and suggest a potential mechanistic link between alternative end joining and interstrand crosslink repair .
DNA double-strand breaks ( DSBs ) and interstrand crosslinks pose serious threats to cell survival and genome stability . Because these lesions compromise both strands of the double helix , they impede DNA replication and transcription and therefore must be removed in a timely and coordinated manner . Interstrand crosslink repair has been shown to involve a DSB intermediate in some cases ( reviewed in [1] ) . Therefore , there may be substantial mechanistic overlap in the processes used during repair of these two lesions . Error-free repair of DSBs can be accomplished through homologous recombination ( HR ) with an undamaged homologous template ( reviewed in [2] ) . However , in contexts where suitable templates for HR do not exist , error-prone repair mechanisms are also used . For example , non-homologous end joining ( NHEJ ) frequently creates small insertions and deletions during DSB repair , particularly in cases where the broken ends cannot be readily ligated ( reviewed in [3] ) . Analogously , the use of translesion DNA polymerases during interstrand crosslink repair can result in mutations , due to the reduced fidelity of these polymerases [4] , [5] . Accumulating evidence suggests that NHEJ is composed of at least two genetically distinct mechanisms . Classical NHEJ ( C-NHEJ ) involves the sequential recruitment of two highly conserved core complexes ( reviewed in [6] ) . First , the Ku70/80 heterodimer recognizes and binds to DNA ends in a sequence-independent manner , thereby protecting them from degradation . In many eukaryotes , Ku70/80 also recruits DNA-PKcs , forming a synaptic complex that can recruit additional processing enzymes such as the Artemis nuclease and the X family DNA polymerases mu and lambda . These proteins expand the spectrum of broken ends that can be rejoined . The second core complex , composed of DNA ligase 4 , XRCC4 , and XLF/Cernunnos , catalyzes ligation of the processed ends . Depending on the substrate , C-NHEJ can result in perfect repair of broken DNA , or it can result in small deletions of 1–10 nucleotides and/or insertions of 1–3 nucleotides [7] . Although C-NHEJ can repair blunt-ended substrates , a subset of C-NHEJ products appear to involve annealing at 1–4 nucleotide microhomologous sequences on either side of the break . Alternative end-joining ( alt-EJ ) is defined as end-joining repair that is observed in cells or organisms lacking one or more C-NHEJ components ( reviewed in [8] ) . Alt-EJ in yeast is associated with deletions larger than those typically created by C-NHEJ , together with an increased tendency to repair by annealing at microhomologous sequences . Ku and ligase 4-independent end joining observed in mammalian cells also displays an increased tendency towards use of short microhomologies compared to C-NHEJ [9] , [10] . Therefore , alt-EJ is sometimes called microhomology-mediated end joining ( MMEJ ) [11] . However , the relationship between MMEJ and alt-EJ is still unclear , and alt-EJ may comprise one or more C-NHEJ-independent repair mechanisms [8] . The importance of alt-EJ repair is highlighted by multiple studies that suggest it may promote chromosome instability and carcinogenesis . Alt-EJ produces chromosome translocations in mouse embryonic stem cells lacking Ku70 [12] and the use of alt-EJ during V ( D ) J recombination in C-NHEJ-deficient murine lymphocytes causes complex chromosome translocations and progenitor B cell lymphomas [13] . Furthermore , alt-EJ has been implicated in various translocations associated with chronic myeloid leukemia and human bladder cancer [14] , [15] . Importantly , alt-EJ also operates during V ( D ) J rejoining in C-NHEJ-proficient B lymphocytes [16] , suggesting that its role in DSB repair is not limited to situations where C-NHEJ is defective . However , alt-EJ is frequently masked by more dominant repair processes that are essential for vertebrate development , making it difficult to study . Therefore , its molecular mechanisms and the proteins involved remain largely unknown . Several lines of evidence demonstrate that Drosophila is an excellent model system in which to study alt-EJ in a metazoan . The Drosophila genome lacks several mammalian C-NHEJ components , including DNA-PKcs and Artemis . This may predispose flies towards non-C-NHEJ repair . Consistent with this , we have previously shown that a DSB caused by excision of a P element transposon in flies is readily repaired by a DNA ligase 4-independent end-joining process [17] . Interestingly , although Drosophila orthologs for the Pol X family DNA polymerases mu and lambda have not been identified [18] , we and others have found evidence for polymerase activity in Drosophila end-joining repair [17] , [19] , [20] . Specifically , end joining in flies is often associated with the insertion of nucleotides at repair junctions , frequently involving imperfect repeats of 5–8 nucleotides . Full or partial templates for the insertions , occasionally possessing mismatches , can often be identified in adjacent sequences , suggesting the action of an error-prone polymerase . Similar templated nucleotides ( T-nucleotides ) have previously been identified at translocation breakpoints in human lymphomas [21]–[23] . Therefore , T-nucleotides could represent a signature of alt-EJ and may be informative regarding its molecular mechanisms . Additional insight into alt-EJ is provided by recent reports suggesting a mechanistic link between alt-EJ and interstrand crosslink repair . For example , a study of two Chinese hamster ovary cell lines sensitive to the crosslinking agent mitomycin C found that they were also deficient in alt-EJ [24] . Furthermore , certain interstrand crosslink-sensitive cell lines from Fanconi Anemia patients are also impaired in DNA-PKcs-independent rejoining of linearized plasmids [25] . Based on these reports , we hypothesized that additional mechanistic insight into both interstrand crosslink repair and alt-EJ could be gained by searching for mutants defective in both processes . To test this , we have screened Drosophila mutants that are sensitive to DNA crosslinking agents for additional defects in alt-EJ repair . In this work , we describe our studies with one such mutant , mus308 . The mus308 ( mutagen sensitive 308 ) mutant was originally identified by its extreme sensitivity to interstrand crosslinking agents but normal resistance to alkylating agents [26] . Subsequently , mus308 was found to code for DNA polymerase theta , which is most similar to A family DNA polymerases such as Escherichia coli Pol I [27] . Orthologs of polymerase theta ( hereafter referred to as pol θ ) are found in many metazoans , including Caenorhabditis elegans , Arabidopsis thaliana , and Homo sapiens , but not in unicellular eukaryotes , including the yeasts [28]–[30] . Several lines of evidence suggest that pol θ may play an important role in maintaining genome stability . Similar to flies , C . elegans with mutations in POLQ-1 are defective in repair of interstrand crosslinks [28] . Mice lacking pol θ ( chaos1 mutants ) have a high frequency of spontaneous and mitomycin C-induced micronuclei in erythrocytes , consistent with genomic instability [31] . In addition , vertebrate pol θ orthologs have been implicated in a wide range of repair processes , including base excision repair , bypass of abasic sites , and somatic hypermutation of immunoglobulin genes [32]–[36] . Finally , upregulation of pol θ is observed in a variety of human tumors and is associated with a poor clinical outcome , suggesting that its overexpression may contribute to cancer progression [37] . Pol θ is unusual in possessing an N-terminal helicase-like domain and a C-terminal polymerase domain . Although pol θ purified from human cell lines and Drosophila has error-prone polymerase activity and single-stranded DNA-dependent ATPase activity , helicase activity has not been demonstrated in vitro [30] , [38] , [39] . Therefore , it remains unclear exactly how the structure of pol θ relates to its multiple functions in DNA repair in different organisms . We report here that in addition to its role in DNA interstrand crosslink repair , Drosophila pol θ is involved in end-joining repair of DSBs . This alt-EJ mechanism operates independently of both Rad51-mediated HR and ligase 4-dependent C-NHEJ . Genetic analysis using separation-of-function alleles provides support for distinct roles of both the N- and C-terminal domains of pol θ in alt-EJ . Collectively , our data support a model in which helicase and polymerase activities of Drosophila pol θ cooperate to generate single-stranded microhomologous sequences that are utilized during end alignment in alt-EJ .
Drosophila mus308 mutants were initially identified based on their sensitivity to low doses of chemicals that induce DNA interstrand crosslinks [26] . To confirm this phenotype , we assembled a collection of previously identified mus308 mutant alleles [40] , [41] and measured the ability of hemizygous mutant larvae to survive exposure to the crosslinking agent mechlorethamine ( nitrogen mustard ) . Of the mutants that we tested , four were unable to survive exposure to 0 . 005% mechlorethamine: D2 , D5 , 2003 , and 3294 ( data not shown ) , consistent with their inability to repair interstrand crosslinks . To determine the molecular lesions responsible for mechlorethamine sensitivity , we sequenced the entire mus308 coding region of flies hemizygous for each mutant allele . Pol θ possesses both a conserved N-terminal helicase-like domain and a C-terminal pol I-like polymerase domain ( Figure 1 , Figure S1 ) [30] . Three of the four alleles contain unique sequence changes that are predicted to affect pol θ primary structure ( Figure 1 , Figure S2 , and Figure S3 ) . The 2003 allele is a nonsense mutation upstream of the polymerase domain , while the D5 allele is a missense mutation that alters a highly conserved proline in the conserved N-terminus . The 3294 allele changes an invariant glycine in the helicase domain to serine . Interestingly , this residue is conserved in the related mus301 helicase , but not in other DNA helicases ( data not shown ) . No mutations were found in the coding sequence of the D2 allele . Because homozygous D2 flies have undetectable levels of pol θ protein [38] , the D2 mutation may affect a regulatory region of mus308 . One explanation for the extreme sensitivity of mus308 mutants to mechlorethamine could be a defect in the repair of certain types of DSB intermediates that are created during crosslink repair . To test this , we exposed flies hemizygous for each mutant allele to increasing doses of ionizing radiation ( IR ) . Although IR creates many different types of lesions , unrepaired DSBs are thought to be the main cause of cell death following irradiation . All four mus308 mutants survived IR exposures as high as 4000 rads ( Figure 2A ) , although bristle and wing defects characteristic of apoptotic cell death were frequently observed at high doses . Drosophila lig4 mutants , which are completely defective in C-NHEJ , also survive IR doses in excess of 4000 rads [17] . However , spn-A mutants , which lack the Rad51 recombinase required for strand invasion during homologous recombination initiation [42] , are highly sensitive to IR [17] . Thus , in Drosophila , HR is the dominant mechanism used to repair IR-induced DSBs . To test whether pol θ acts to repair IR damage in the absence of HR , we created mus308 spn-A double mutants and exposed them to doses of 125–1000 rads . Strikingly , doses as low as 125 rads resulted in almost complete killing of mus308 spn-A mutants ( Figure 2B ) . In contrast , lig4 spn-A double mutants are only slightly more sensitive than spn-A single mutants to IR [17] . Thus , in the absence of HR , pol θ participates in a process crucial for repair of damage caused by ionizing radiation . Because interstrand crosslink repair and alternative end joining have been shown to have partially overlapping genetic requirements in mammals [24] , [25] , we hypothesized that the extreme sensitivity of mus308 spn-A mutants to IR might relate to a role of pol θ in an alternative end-joining mechanism . To explore this hypothesis , we tested each mus308 mutant allele using a site-specific double-strand break repair assay that can distinguish between synthesis-dependent strand annealing ( SDSA , a specific type of HR ) and end joining ( EJ ) ( Figure 3A ) [43] . We have previously shown that the majority of end joining observed in this assay occurs independently of DNA ligase 4 , and is therefore a form of alt-EJ [17] . In this system , excision of a P element ( P{wa} ) located on the X chromosome is catalyzed by P transposase , resulting in a 14-kilobase gap relative to an undamaged sister chromatid . The DNA ends remaining after excision each have 17-nucleotide non-complementary 3′ single-stranded overhangs [44] . These ends are highly recombinogenic and repair by SDSA is initially favored . However , because repair synthesis in this system is not highly processive , most repair products that are recovered from wild-type flies result from incomplete repair synthesis from one or both sides of the break , followed by end joining of the nascent DNA ( SDSA+EJ events ) [45] . To quantitate the percentage of repair events that derive from each mechanism , repair events are recovered from male pre-meiotic germline cells by mating individual males to females homozygous for the P{wa} element . Each of the resulting female progeny represents a single repair event that can be classified by eye color . Red eyed-females inherit a repair event involving homology-dependent synthesis that generated complementary single-stranded regions that subsequently anneal ( repair by SDSA ) . Yellow-eyed females inherit a chromosome that was repaired by EJ or SDSA+EJ mechanisms ( these repair events are hereafter referred to as ( SDSA ) +EJ; for further details , see Materials and Methods ) . Overall , the results from the P{wa} assay indicated that mus308 mutants are defective in end-joining repair of DSBs . We observed no decrease in the percentage of red-eyed progeny recovered from mus308 mutant males ( Figure 3B ) , suggesting that SDSA repair is not impaired when pol θ is missing or defective . In contrast , all four mus308 mutant alleles resulted in a significantly decreased percentage of yellow-eyed progeny relative to wild type ( p<0 . 001 , Kruskal-Wallis test ) . Because yellow-eyed progeny can only result from a repair mechanism involving end joining , these data suggest that pol θ is involved in an end-joining process . To further demonstrate that pol θ is not involved in DNA synthesis during SDSA , we recovered independent ( SDSA ) +EJ events in males , isolated genomic DNA , and used PCR to estimate the approximate amount of DNA repair synthesis that occurred prior to end joining . The amount of repair synthesis in ( SDSA ) +EJ repair products did not differ significantly between wild-type and mus308 mutant flies ( Figure 3C ) . We conclude that pol θ is not required for DNA synthesis during SDSA , but plays an important role in end-joining repair following aborted SDSA . Mutations that abolish end joining in flies cause an increased frequency of genomic deletions during repair of site-specific DSBs [46] , [47] . To determine whether mutation of mus308 also results in repair-associated deletions , we took advantage of the fact that deletions can be easily scored in the P element excision assay . Because P{wa} is inserted in the essential scalloped ( sd ) gene , repair events that delete into sd exons cause a scalloped-wing phenotype when recovered in heterozygous females and lethality in hemizygous males [48] , [49] . We observed a substantial increase in the percentage of deletion-associated repair events isolated from mus308 mutant males relative to wild type ( Figure 3D ) . Overall , the total percentage of end-joining repair events involving deletions recovered from mus308 mutants was elevated from 3- to 26-fold over wild type , depending on the mus308 allele tested . Previously , we observed a similar deletion-prone phenotype in flies lacking the DmBlm protein , which is involved in repair of DSBs by SDSA [48] . Because our data did not support a role for pol θ in homologous recombination , we expected the deletion-prone phenotype of mus308 mutants to persist even in SDSA-deficient flies . To confirm this , we assayed repair following P{wa} excision in mus308 mutants lacking the Rad51 protein , which renders them unable to carry out HR repair [42] , [45] . As expected , PCR analysis of repair products showed that SDSA was abolished in both spn-A and spn-A mus308 mutants ( data not shown ) . Approximately 17% of P{wa} chromosomes recovered from spn-A mutant males showed evidence for end joining at the 17-nucleotide overhangs that are created by P transposase ( Figure 4A and Table 1 ) ; the other 83% of P{wa} chromosomes recovered were presumably uncut . We observed a 30–50% decrease in end-joining repair products in spn-A mus308 double mutants compared to spn-A mutants ( p<0 . 001 , Kruskal-Wallis test ) , confirming a unique role for polθ in end joining when HR is absent . Importantly , mutation of mus308 still caused an increased percentage of deletions in the absence of Rad51 ( Figure 4B ) . From these data , we conclude that the deletions formed during break repair in mus308 mutants are not the result of aborted SDSA . Rather , they are a consequence of a deletion-prone repair mechanism that operates in the absence of both SDSA and pol θ-dependent end joining . During the course of these experiments , we made a number of observations suggesting that Rad51 and pol θ act in parallel and distinct DSB repair mechanisms . First , we recovered fewer spn-A mus308 double mutant males than would be predicted from Mendelian ratios . For example , in crosses between mus308D2 and mus308D5 heterozygotes , 38% of the progeny were mus308D2/mus308D5 compound heterozygotes . In contrast , only 16% of progeny recovered from parallel matings between spn-A057mus308D2 and spn-A093mus308D5 mutants were spn-A mus308 compound heterozygotes ( P<0 . 05 , Fisher's exact test; Figure 5A ) . This difference in viability between mus308 and spn-A mus308 mutants was even more extreme in flies in which excision of P{wa} was occurring ( P<0 . 01; Figure 5A ) . In addition , we observed heightened male sterility in various combinations of spn-A mus308 mutants undergoing P{wa} transposition , with 51% of the double mutant males unable to produce viable progeny in the most severe allele combination ( Figure 5B ) . Finally , we observed morphological abnormalities , specifically abdominal closure defects and aberrant cuticle banding patterns , in 100% of spn-A093mus308D5/spn-A057mus308D2 double mutants ( Figure 5C ) . These defects were more severe in the double mutants experiencing P{wa} transposition , but were not apparent in either mus308 or spn-A single mutants . From these data , we conclude that Rad51 and pol θ participate in independent pathways required for repair of DSBs that arise during both endogenous developmental processes and during P element transposition . P element ends are good substrates for DNA ligase 4-independent end joining [17] . Based on the results presented above , it seemed likely that pol θ is involved in an end-joining process different from C-NHEJ . To formally test this , we repeated the P{wa} assay in lig4 mus308 double mutants that lack DNA ligase 4 and are unable to repair DSBs by C-NHEJ . Unlike spn-A mus308 mutants , we observed no viability , fertility , or morphological defects in lig4 mus308 mutants . We also observed no defect in HR repair in the double mutants ( Figure 4C ) , consistent with results obtained using mus308 single mutants . In contrast , we observed a further decrease in the percentage of end joining repair products recovered from lig4 mus308 double mutants relative to mus308 mutants , from 3 . 0% to 1 . 3% ( P<0 . 01 , Kruskal-Wallis test ) . Previously , we have shown that end-joining repair of DSBs induced by P{wa} excision is unaffected in lig4 mutants [17] . Therefore , the removal of pol θ-mediated end joining reveals a previously hidden role for DNA ligase 4 in the repair of DSBs created by P transposase . Strikingly , although only 50% of end-joining products isolated from mus308 mutants involved large , male-lethal deletions , 100% of end-joining products recovered from lig4 mus308 mutant males were associated with large deletions ( Figure 4D ) . From these results , we conclude that at least three distinct mechanisms for end-joining repair exist in Drosophila . One , which corresponds to C-NHEJ , requires DNA ligase 4 and other canonical NHEJ proteins , including XRCC4 , Ku70 , and Ku80 [46] , [47] , [50] . Another mechanism , which is at least partially independent of DNA ligase 4 , is defined by a requirement for pol θ and corresponds to alt-EJ . Interestingly , alt-EJ appears to be used more frequently than C-NHEJ , at least for the repair of P element-induced breaks . In the absence of these two repair processes , a Rad51-independent backup mechanism characterized by extensive genomic deletions operates at low efficiency . Alt-EJ repair in Drosophila is frequently associated with annealing at microhomologous sequences of more than four nucleotides and with long DNA insertions at repair junctions [8] . To determine whether pol θ-dependent end joining involves either of these types of repair , we sequenced repair junctions obtained from spn-A and spn-A mus308 double mutants following P{wa} excision . Because we sequenced only one junction per male germline , each junction analyzed represents an independent repair event . Five distinct junction types were identified . Three of these types are characteristic of junctions arising from C-NHEJ in mammalian systems [7]: junctions involving small , 1–3 base pair insertions , junctions involving annealing at 1–3 nucleotide microhomologies , and junctions for which no microhomologies can be identified ( apparent blunt end junctions ) . The other two types of junctions , characteristic of alt-EJ [8] , involve annealing at 5–10 nucleotide microhomologous sequences or insertions of more than three base pairs . Approximately 58% of junctions from spn-A mutants showed structures considered typical of C-NHEJ repair , while 29% involved annealing at 5–10 nucleotide microhomologies and 13% had insertions of greater than three base pairs ( Figure 6A and Table 1 ) . Potential templates for the larger insertions could almost always be identified in flanking sequences . These insertions may be analogous to T-nucleotides that have been observed at translocation breakpoint junctions isolated from certain human cancers [21]–[23] . When we sequenced repair junctions from spn-A mus308 mutants , we observed two distinct patterns , depending on the mus308 alleles used . For both the D2/2003 and D5/2003 allele combinations , the percentage of junctions involving annealing at long microhomologies was significantly decreased ( P<0 . 01 , Fisher's exact test; Figure 6A , Table 2 , and Table 3 ) . Only 12% of D2/2003 junctions possessed an insertion greater than three base pairs , compared to 44% of junctions recovered from males with the D5 and 2003 alleles . In addition , most insertions isolated from D5/2003 males were highly complex and had multiple copies of imperfect repeats of T-nucleotides . Similar results were obtained with the D5/3294 allele combination ( data not shown ) . An overall comparison of insert length showed that flies with wild-type mus308 alleles had an average insert length of 5 . 5 nucleotides , compared to 3 . 8 nucleotides for D2/2003 mutants and 13 . 3 nucleotides for D5/2003 mutants . In summary , both mus308 mutant combinations significantly abrogated annealing at long microhomologies during alt-EJ repair . However , we observed a distinct difference in repair junctions recovered from males harboring the D2 allele , which greatly reduces overall pol θ protein levels [38] , compared to flies with the D5 allele , which alters a conserved residue near the helicase-like domain . These results suggest that pol θ has two distinct functions in alt-EJ: one that promotes the annealing of long microhomologous sequences during end alignment , and another that is responsible for complex T-nucleotide insertions . Flies with the D2 allele are impaired in their ability to carry out both the annealing and insertion functions , whereas flies possessing the D5 separation-of-function allele cannot perform the microhomology annealing function but can still produce complex insertions . P element-induced breaks are unique in that they possess 17-nucleotide non-complementary ends that are poor substrates for C-NHEJ . To test whether the results obtained with P elements can be generalized to other types of breaks , we used the I-SceI endonuclease and the previously characterized [Iw]7 reporter construct [50] to create site-specific DSBs in wild-type flies and flies lacking either DNA ligase 4 or pol θ . I-SceI produces a DSB with 4-nucleotide complementary overhangs that can be directly ligated through a C-NHEJ mechanism [50] , [51] . Accurate repair regenerates the original I-SceI recognition sequence , which can then be cut again , while inaccurate end-joining repair abolishes further cutting . We utilized an hsp70 or ubiquitin-driven I-SceI construct integrated on chromosome 2 to drive high levels of I-SceI expression [50] , [52] . Nearly 100% of repair events that we recovered involved gene conversion ( HR repair from the homologous chromosome ) or inaccurate end-joining ( data not shown ) . In the [Iw]7 system , both gene conversion events and large deletions that remove the white marker gene are phenotypically indistinguishable . PCR analysis confirmed that many repair events recovered from mus308 mutants involved large deletions ( >700 base pairs , data not shown ) . Our subsequent analysis focused on the characterization of repair events involving smaller deletions . Twenty-three percent of I-SceI repair junctions isolated from wild-type flies possessed insertions of more than 3 base pairs ( Figure 6B ) . This percentage was significantly increased to 46% in lig4 mutants ( P<0 . 01 , Fisher's exact test ) , consistent with increased use of alt-EJ in the absence of C-NHEJ . If pol θ plays a general role in insertional mutagenesis during alt-EJ repair , one would predict that the frequency and length of insertions following I-SceI cutting should decrease in mus308 mutants . Indeed , the percentage of large insertions decreased to 9% in mus308 mutants ( P = 0 . 03 , Fisher's exact test ) . Wild-type flies had an average insertion length of 7 . 6 base pairs , compared to 4 . 2 base pairs for mus308 mutants . Strikingly , no mus308 insertion was longer than twelve base pairs , while insertions of more than twenty base pairs occurred in both wild type and lig4 mutants . Because microhomologies of greater than four base pairs are not present near the I-SceI cut site in this construct , repair involving annealing at long microhomologies was not observed . Surprisingly , the total percentage of repair junctions with short , 1–3 base pair insertions was not decreased in lig4 mutants relative to wild type ( 17% vs . 13% , respectively ) . Furthermore , the percentage of junctions involving annealing at 1–3 nucleotide microhomologies was also similar between the two genotypes ( 25% for lig4 mutants vs . 34% for wild type ) . These two types of junctions have historically been associated with ligase 4-dependent C-NHEJ repair . Our results suggest that this may not be the case . Indeed , a fine-level sequence analysis of I-SceI repair junctions that we have recently conducted suggests that alt-EJ may produce C-NHEJ-like junctions in certain sequence contexts [53] . Nevertheless , our data obtained using two independent site-specific DSB repair assays strongly suggest that C-NHEJ and alt-EJ represent at least partially independent mechanisms for the repair of DSBs and that pol θ plays an important role in the generation of T-nucleotide insertions during alt-EJ repair of both P element and I-SceI-induced breaks .
Pol θ orthologs characterized from a variety of metazoans possess both helicase-like and DNA polymerase domains [27]–[31] , [35] . Pol θ purified from both Drosophila and human cells has a Pol I-like polymerase activity and single-stranded DNA-dependent ATPase activity [30] , [38] . However , DNA helicase activity of the purified protein remains to be demonstrated . Although our experiments did not formally test for helicase activity of pol θ , our results are consistent with pol θ having a DNA unwinding or strand displacement function . Flies with the D5 and 3294 mutations ( located in or near the conserved helicase domain ) produce repair products with complex T-nucleotide insertions but not products involving annealing at long microhomologies . The D5 and 3294 alleles may therefore encode proteins that retain polymerase activity but lack unwinding activity , resulting in an inability to expose internal microhomologous sequences . Because the microhomologies used in repair following P element excision are often located in the 17-nucleotide 3′ single-stranded tails , pol θ may also be important for the unwinding of secondary structures that form in single-stranded DNA . Alternatively , the DNA-dependent ATPase activity demonstrated by pol θ might represent an annealing function of the protein that is required during alt-EJ . Such an annealing activity was recently described for the human HARP protein , which is able to displace stably bound replication protein A and rewind single-stranded DNA bubbles [61] . One notable aspect of alt-EJ in Drosophila is the large percentage of repair junctions with templated insertions . These insertions may be “synthesis footprints” that are formed during the cell's attempt to create microhomologous sequences that can be used during the annealing stage of alt-EJ when suitable endogenous microhomologies are not present or are not long enough to allow for stable end alignment . Indeed , analysis of the insertions from I-SceI repair junctions suggests a model involving local unwinding of double-stranded DNA and iterative synthesis of 3–8 nucleotide runs [53] . The P{wa} repair junctions isolated from mus308D2/mus3082003 mutants are consistent with an important ( but not exclusive ) role for the polymerase domain of pol θ in the synthesis of T-nucleotides . We speculate that pol θ may be involved in both DNA unwinding and repair synthesis during alt-EJ ( Figure 7 ) . Linking these two activities in one protein would provide a convenient mechanism for creating longer microhomologies that could increase the thermodynamic stability of aligned ends prior to the action of a DNA ligase . Studies based on the crystal structure of a dual function NHEJ polymerase-ligase protein found in Mycobacterium tuberculosis suggest that a synaptic function for an NHEJ polymerase is plausible [62] . Because ligase 4 is not involved in alt-EJ in Drosophila , another ligase must be involved in the ligation step . Studies from mammalian systems have identified DNA ligase 3 as a likely candidate [63] , [64] . Pol θ was originally identified in Drosophila based on the inability of mus308 mutants to survive exposure to chemicals that induce DNA interstrand crosslinks . A crucial question posed by our findings is whether pol θ performs a common function during the repair of both DSBs and interstrand crosslinks . The C . elegans pol θ ortholog , POLQ-1 , is also required for resistance to interstrand crosslinks and acts in a pathway that is distinct from HR but depends on CeBRCA1 [28] . In S . cerevisiae , several repair mechanisms are utilized during interstrand crosslink repair , including nucleotide excision repair ( NER ) , HR , and translesion synthesis [65] , [66] . Given our results and the findings from C . elegans , it seems unlikely that the role of pol θ in interstrand crosslink repair involves a function in HR . In human cells , exposure to agents that induce interstrand crosslinks causes a shift in repair mechanisms that leads to increased use of non-conservative pathways associated with complex insertions and deletions [67] . Furthermore , interstrand crosslinks can cause frequent recombination between direct repeats [68] , [69] , suggesting that single-strand annealing may provide a viable mechanism for interstrand crosslink repair . The single-strand annealing model of interstrand crosslink repair posits that NER-independent recognition and processing of the crosslinked DNA is followed by generation of single-stranded regions flanking the crosslink and annealing at repeated sequences . Because alt-EJ frequently proceeds through annealing at short direct repeats , it is tempting to speculate that the role of pol θ in interstrand crosslink repair might be to expose and/or promote the annealing of microhomologous single-stranded regions that flank the crosslinked DNA . Consistent with this model , the initial incision step made after recognition of the interstrand crosslink remains normal in mus308 mutants [26] . Alternatively , pol θ might utilize its polymerase activity and nearby flanking sequences as a template to synthesize short stretches of DNA that could be used to span a single-stranded gap opposite of a partially excised crosslink . Such a model has been proposed to explain the formation of microindels in human cancers [70] . We are currently testing these two models using helicase- and polymerase-specific mus308 mutant alleles . Although it seems counterintuitive , alt-EJ likely functions in some situations to promote genome stability . As evidence of this , we found that DSB repair following P element excision in mus308 mutant flies frequently results in genomic deletions of multiple kilobases . A similar deletion-prone phenotype was previously observed in mus309 mutants , which lack the Drosophila BLM ortholog [43] , [71] . Epistasis analysis demonstrated that the mus309 deletion phenotype depends on Rad51 , implying that DmBlm acts after strand invasion during HR and that the deletions observed in mus309 mutants are likely a result of failed SDSA [48] . In contrast , the deletions observed in mus308 mutants do not depend on Rad51 , demonstrating that the function of pol θ in DSB repair is independent of HR . The deletion phenotype is exacerbated in lig4 mus308 double mutants , suggesting that C-NHEJ and alt-EJ represent two parallel mechanisms to prevent deletions . In the absence of these two end-joining options , resection at the broken ends may continue unchecked , resulting in extensive genomic deletions that are generated by an unknown Rad51-independent repair mechanism . Therefore , both C-NHEJ and alt-EJ function to prevent overprocessing of broken DNA ends and extreme degradation of the genome . Microhomology-mediated end joining , which shares many features with alt-EJ , has been proposed to perform a similar function in urothelial cells [72] . Nonetheless , alternative end-joining repair can also be genome destabilizing , as demonstrated by an increasing number of reports linking it to cancer . We have shown that complex insertions observed in alternative end-joining products are more frequent in flies possessing pol θ . These insertions , which are often combinations of nucleotides derived from several templates inserted in both direct and reverse-complement orientations , are remarkably similar to T-nucleotide insertions found in translocation breakpoints reconstructed from follicular and mantle cell lymphomas ( reviewed in [73] ) . Therefore , if pol θ also functions in alternative end joining and T-nucleotide generation in mammals , it might be an important factor involved in translocation formation . A recent study suggests that pol θ levels are tightly regulated in humans and that loss of this regulation may promote cancer progression [37] . The protein is primarily found in lymphoid tissues but is upregulated in lung , stomach , and colon cancers . Furthermore , high levels of pol θ expression correlate with poorer clinical outcomes . Intriguingly , pol θ is regulated by endogenous siRNAs in Drosophila [74] , [75] , although the significance of this regulation is currently unclear . We suggest that polθ-mediated alt-EJ serves as a medium-fidelity repair option used by cells when precise repair cannot be carried out for any number of reasons . As such , it prevents extreme loss of genetic information . However , its error-prone nature requires tight regulation , which , when lost , may lead to excessive inaccurate repair and ultimately , carcinogenesis . The results described here establish that Drosophila pol θ plays two distinct roles in an alternative end-joining mechanism operating in parallel to canonical DNA ligase 4-mediated C-NHEJ . This novel finding lays the groundwork for future studies focusing on the specific roles of the pol θ helicase-like and polymerase domains in alt-EJ and DNA interstrand crosslink repair . Whether pol θ plays a similar role in alt-EJ in other organisms , including mammals , remains to be determined . Regardless , these studies reveal an unexpected role for DNA polymerase θ that is required for genomic integrity in Drosophila and possibly other metazoans .
All flies were maintained on standard cornmeal-based agar food and reared at 25°C . The mus308 D2 and D5 stocks were obtained from the Bloomington Stock Center and the 2003 and 3294 stocks were from the Zuker collection [76] . To identify mutations in these stocks , genomic DNA was isolated from flies harboring each allele in trans to Df ( 3R ) Exel6166 and PCR was performed with primers specific to overlapping regions of the entire coding sequence . PCR products were sequenced and the sequence was compared to the Drosophila reference sequence release 5 . 10 . Sequence changes unique to each allele were verified by sequencing in both orientations . The lig4169a [17] , spn-A093 and spn-A057 [42] stocks harbor null alleles of DNA ligase 4 and Rad51 , respectively . For mechlorethamine sensitivity assays , balanced , heterozygous parents were crossed to Df ( 3R ) Exel6166 and allowed to lay eggs in vials containing 10mL of food for three days , after which they were moved to new vials for two additional days . The first vials were treated with 250µL of 0 . 005% mechlorethamine dissolved in ddH2O , while the second vials were treated only with ddH2O . Survival was calculated as the number of homozygous mutant adults divided by the total number of adults that eclosed within 10 days of treatment . Ratios were normalized to untreated controls for each set of vials ( five to eight sets of vials were counted for each experiment ) . For ionizing radiation sensitivity assays , heterozygous parents laid eggs on grape-juice agar plates for 12 hr . Embryos developed at 25°C until larvae reached third-instar stage , at which point they were irradiated in a Gammator 1000 irradiator at a dose rate of 800 rads/min and larvae were transferred to food-containing bottles . Relative survival rates were calculated as above . Repair of DNA double-strand breaks was monitored after excision of the P{wa} transposon as described previously [43] , [77] . P{wa} was excised in males using a second chromosome transposase source ( CyO , H{w+ , Δ2–3} ) and individual repair events were recovered in female progeny over an intact copy of P{wa} . Females with two copies of P{wa} have apricot eyes [78] . Progeny with red eyes possess a repair event involving HR with annealing of the copia LTRs . A fraction of apricot-eyed females also possess HR repair events , but these cannot be distinguished from chromosomes in which no excision event occurred ( using the CyO , H{w+ , Δ2–3} transposase source , ∼80% of apricot-eyed female progeny inherit a non-excised P{wa} element ) . Yellow-eyed females harbor a repair event in which repair is completed by end joining . For each genotype , at least 50 individual male crosses were scored for eye color of female progeny lacking transposase . The percentage of progeny from each repair class was calculated on a per vial basis , with each vial representing a separate experiment . Statistical comparisons were done with a Kruskal-Wallis non-parametric ANOVA followed by Dunn's multiple comparisons test using InStat3 ( GraphPad ) . For analysis of HR synthesis tract lengths , genomic DNA was purified and PCR reactions were performed as in [43] , using primer pairs with the internal primer located 250 , 2420 , and 4674 base pairs from the cut site at the 5′ end of P{wa} . For deletion analysis , the percentage of females with scalloped wings was calculated relative to all yellow-eyed females counted . The percentage of male lethal and small ( 0 . 1–3 . 6 kb ) deletions was calculated based on a subset of yellow-eyed females ( one from each original male parent ) that were individually crossed to males bearing the FM7w balancer . Vials for which no white-eyed male progeny were recovered were scored as male lethal . Some of the male lethal events also caused a scalloped-wing phenotype in heterozygous females . For those that did not , testing to ensure that the male lethality was due to deletion of scalloped coding sequence was performed by recovering the repaired chromosomes in trans to the hypomorphic sd1 mutation [79] and scoring for a scalloped-wing phenotype . Repair events which could be recovered in males were subjected to PCR analysis , using primers internal to P{wa} [43] , to detect small deletions into one or both introns of sd . Repair of I-SceI mediated DNA double strand breaks was studied in the context of the chromosomally integrated [Iw]7 construct [52] , which contains a single target site for the I-SceI endonuclease . DSBs were induced in the male pre-meiotic germline by crossing females harboring [Iw]7 to males expressing the I-SceI endonuclease from a second chromosomal location under the control of either the hsp70 promoter ( 70[I-SceI]1A ) [52] or the ubiquitin promoter ( UIE[I-SceI]2R ) [50] . Independent inaccurate end-joining repair events from the male pre-meiotic germline were recovered in male progeny and DNA was isolated for analysis [80] . PCR was performed using primers PE5′ ( GATAGCCGAAGCTTACCGAAGT ) and jn3′b ( GGACATTGACGCTATCGACCTA ) to amplify a 1 . 3 kb fragment of the [Iw]7 construct including the I-SceI target site . Products were gel purified ( GenScript ) and sequencing of PCR products was performed using the PE5′ primer . Sequences were aligned using ClustalW or by manual inspection against sequence obtained from an uncut [Iw]7 construct . Statistical comparisons were done using Excel and SPSS .
|
DNA double-strand breaks , in which both strands of the DNA double helix are cut , must be recognized and accurately repaired in order to promote cell survival and prevent the accumulation of mutations . However , error-prone repair occasionally occurs , even when accurate repair is possible . We have investigated the genetic requirements of an error-prone break-repair mechanism called alternative end joining . We have previously shown that alternative end joining is frequently used in the fruit fly , Drosophila melanogaster . Here , we demonstrate that a fruit fly protein named DNA polymerase theta is a key player in this inaccurate repair mechanism . Genetic analysis suggests that polymerase theta may be important for two processes associated with alternative end joining: ( 1 ) annealing at short , complementary DNA sequences , and ( 2 ) DNA synthesis that creates small insertions at break-repair sites . In the absence of polymerase theta , a backup repair mechanism that frequently results in large chromosome deletions is revealed . Because DNA polymerase theta is highly expressed in many types of human cancers , our findings lay the groundwork for further investigations into how polymerase theta is involved in repair processes that may promote the development of cancer .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/chromosome",
"biology",
"molecular",
"biology/dna",
"repair"
] |
2010
|
Dual Roles for DNA Polymerase Theta in Alternative End-Joining Repair of Double-Strand Breaks in Drosophila
|
Adenosine-to-inosine ( A-to-I ) editing is hypothesized to facilitate adaptive evolution by expanding proteomic diversity through an epigenetic approach . However , it is challenging to provide evidences to support this hypothesis at the whole editome level . In this study , we systematically characterized 2 , 114 A-to-I RNA editing sites in female and male brains of D . melanogaster , and nearly half of these sites had events evolutionarily conserved across Drosophila species . We detected strong signatures of positive selection on the nonsynonymous editing sites in Drosophila brains , and the beneficial editing sites were significantly enriched in genes related to chemical and electrical neurotransmission . The signal of adaptation was even more pronounced for the editing sites located in X chromosome or for those commonly observed across Drosophila species . We identified a set of gene candidates ( termed “PSEB” genes ) that had nonsynonymous editing events favored by natural selection . We presented evidence that editing preferentially increased mutation sequence space of evolutionarily conserved genes , which supported the adaptive evolution hypothesis of editing . We found prevalent nonsynonymous editing sites that were favored by natural selection in female and male adults from five strains of D . melanogaster . We showed that temperature played a more important role than gender effect in shaping the editing levels , although the effect of temperature is relatively weaker compared to that of species effect . We also explored the relevant factors that shape the selective patterns of the global editomes . Altogether we demonstrated that abundant nonsynonymous editing sites in Drosophila brains were adaptive and maintained by natural selection during evolution . Our results shed new light on the evolutionary principles and functional consequences of RNA editing .
Genomic mutations are the major sources for phenotypic changes and adaptation [1–4] . In diploid multicellular organisms , a nonsynonymous DNA mutation ( a mutation that alters the amino acid sequence of a protein ) will permanently affect the protein products in all the cells ( soma or germline ) that express the mutant allele . The “all-or-none” property of DNA mutations might incur pleiotropic effects that are antagonistic among cell types , tissues , developmental stages , sexes , or other aspects of life history [5–7] , which would constrain the available genetic diversity for a species . Given the prevalence of pleiotropic effects in the genomes [8–10] , the sequence space might be inaccessible to many mutations , which potentially slows down the rate of phenotypic evolution and adaptation [11] . However , the transcriptomic or proteomic diversity limited by mutation sequence space could be expanded by the alteration of RNA sequences in an epigenetic approach , such as RNA editing , which was hypothesized to facilitate adaptation [12–14] . In addition , RNA editing has the advantage to quickly respond to environmental stress and adjust the activity of final protein products accordingly [15 , 16] . RNA editing is an evolutionarily conserved mechanism that alters RNA sequences at the co-transcriptional or post-transcriptional level [13 , 17–23] . Among various RNA editing systems in animals , the base substitution from adenosine ( A ) to inosine ( I ) , termed A-to-I editing , is the most common form [13 , 20] . Due to the high level of structural similarity between inosine ( I ) and guanosine ( G ) , the cellular machineries , such as ribosomes , spliceosomes or the microRNA ribonucleoprotein complex ( miRNP ) , would recognize I as G during translation [13 , 20 , 24 , 25] , splicing [26–29] , microRNA target recognition [30–32] , or other RNA biological processes [14] . Therefore , A-to-I RNA editing usually produces a change similar to an A-to-G DNA change in particular tissues or developmental stages , which potentially increases phenotypic plasticity without the alteration of genomic sequences [13 , 20 , 24 , 25] . The adenosine deaminase acting on RNA ( ADAR ) family are the enzymes that convert adenosine ( A ) to inosine ( I ) in pre-mRNAs [33–36] . Although multiple Adar genes are encoded in the genomes of mammals and worms , there is only one Adar locus in Drosophila [37 , 38] , which is predominately expressed in the nervous system [39] . The substrates of ADAR are usually double-stranded RNAs [34 , 36 , 40–42] . A-to-I editing plays essential roles in many biological processes [18 , 19 , 43–45] , and the abolition of Adar in D . melanogaster severely affects its viability and behavior [33 , 34 , 46] . Previous studies have identified thousands of A-to-I editing sites in different developmental stages , adult heads or whole animals of Drosophila [47–52] . In addition , A-to-I editing has been extensively characterized in other organisms , such as humans [53–58] , macaques [59 , 60] , mice [61] , worms [62] , and squids [63] . Despite these intriguing advances , only a few examples of the advantageous effects conferred by RNA editing have been demonstrated [13 , 14 , 20 , 63] . For example , the A-to-I editing events in Kv1 mRNA provide numerous adaptive amino acid changes that allow the octopus to adapt to extremely cold temperatures [64] . The functional consequences of the majority of A-to-I editing events , however , remain to be explored . In fact , comparative genomics has demonstrated that only a small fraction of the human A-to-I editing events were evolutionarily conserved [65–68] . Furthermore , it was nicely demonstrated that the editing events in primate coding regions were generally non-adaptive [60 , 67 , 68] . Nevertheless , the targets of RNA editing might have evolved rapidly across species because A-to-I editing in mammals predominantly occurs in repetitive sequences [53–55 , 61 , 65 , 66] , while the editing events in Drosophila are mainly located in coding regions of genes encoding neurotransmitters or ion channels [47–50 , 69 , 70] . Therefore , the evolutionary forces acting on A-to-I RNA editing might be different between Drosophila [50] and primates [60 , 67 , 68] . If A-to-I editing indeed facilitates adaptation by expanding proteomic diversity , we expect to observe predominant signals of adaptation in the editing sites . A recent study [50] reported that although signals of positive selection could be found in genes of the nervous system , the A-to-I RNA editing events were overall subject to purifying selection in Drosophila . Additionally , the overall effect of natural selection on the editome is different across Drosophila developmental stages [50] . Despite these intriguing discoveries , it still remains a mystery whether or not we can find evidence to show that the whole editome is overall adaptive . Specifically , we are interested in testing whether the nonsynonymous A-to-I editing events in Drosophila brains , the core component of the nervous system , are predominantly adaptive . Furthermore , several other fundamental questions on editing deserve to be further investigated: 1 ) Do editing sites preferentially increase sequence space of evolutionarily conserved genes ? 2 ) Why does the global editome of different tissues or developmental stages show differential selective patterns ? 3 ) How does temperature shape the global editomes ? Answers to these questions will help understand the evolutionary principles and functional consequences of RNA editing . In this study we addressed these questions by systematically sequencing the transcriptomes and deciphering A-to-I editing in the female and male brains of three Drosophila species at different temperatures . With evolutionary analysis from different perspectives , we provided lines of evidence to demonstrate that the nonsynonymous editing events in coding regions are generally adaptive in brains of Drosophila . Then we identified a set of gene candidates that had nonsynonymous editing events favored by natural selection . Overall our results demonstrated that abundant nonsynonymous editing events in Drosophila brains were adaptive and maintained by natural selection during evolution .
To comprehensively characterize the A-to-I editing landscapes in brains of Drosophila , we set out to dissect the brains of 1- to 5-day-old or 1- to 14-day-old female and male adults of the inbred ISO-1 strain of Drosophila melanogaster that were constantly raised at 25°C , or raised at 25°C and treated at 30°C for 14 hours or 48 hours ( Table 1 ) . Next we selected the poly ( A ) -tailed mRNAs , fragmented them , ligated the mRNA fragments with adaptors , and deep sequenced the transcriptome of each brain sample ( Materials and Methods ) . We obtained 13 . 9–21 . 6M reads mapped on the reference genome ( see Table 1 and S1 Table for detailed statistics ) , and the median coverage on an exonic site in a library ranges from 5 to 9 reads ( S1 Fig ) . As justified previously [71] , the mRNA fragmentation library preparation procedure we employ minimizes the bias of non-uniform sequencing read coverage along mRNAs , which would reduce the bias in detecting editing in 3' ends of mRNAs . It is a challenging task to reliably distinguish the A-to-I editing events from SNPs [72–76] , therefore , the ISO-1 strain used in this study , which was inbred and sequenced to assemble the reference genome of D . melanogaster [77] , enables us to detect DNA-RNA differences with high accuracy and sensitivity . We employed a two-step strategy to identify editing sites in the brains of D . melanogaster ( Fig 1A ) . First , we used the GATK RNA-Seq variant calling pipeline [78] to identify the candidate A-to-I editing sites in each brain library ( i . e . , the A-to-G differences in the final sequencing results ) . We identified 1 , 531 unique sites with A-to-G DNA-RNA differences in these brain libraries , and such differences accounted for 81 . 5% ( with a standard error of 0 . 97% ) of all the differences detected by the GATK pipeline in each library ( S2A Fig ) . In contrast , the proportion of A-to-G DNA differences ( reference vs . alternative allele ) was only 9 . 9% out of all the mutations ( S2B Fig ) in the global populations of D . melanogaster [79] . This comparison justified the reliability and accuracy of our procedures in defining the candidate A-to-I editing sites . Second , we retrieved a total of 5 , 389 editing sites characterized in D . melanogaster in previous studies ( 972 in Graveley et al . [47] , 1 , 350 in Rodriguez et al . [49] , 3 , 581 in St . Laurant et al . [48] , and 1 , 298 in Yu et al . [50] ) . Altogether , we obtained 5 , 925 unique candidate sites ( 986 sites overlapped between GATK and the four previous studies , S2 Table ) . For each candidate site in a brain library k , we calculate the probability that the A-to-G difference ( if detected ) is caused by editing with Pk ( E1 ) = 1 − Pk ( E0 ) , where Pk ( E0 ) is the probability that the difference is solely caused by sequencing error ( ε ) , by incorporating the sequencing coverage ( Ck ) and the number of G allele ( Lk ) at that site . Next , we utilize the multiple-sample information and calculate the joint probability that this site is edited in at least one library , P ( E1 ) = 1 − P ( E0 ) , where P ( E0 ) is the probability that the A-to-G differences observed in that site across all the applicable libraries are entirely caused by sequencing errors ( Materials and Methods ) . Among the 5 , 925 candidate sites , we did not detect expression of 664 sites in any brain library . For the remaining 5 , 261 expressed sites , we divided them into five exclusive classes with decreasing confidence based on P ( E1 ) , sequencing coverage , and the number of libraries in which the editing events were detected . Briefly , Class I ( 1 , 702 sites ) were defined with the following criteria: 1 ) at FDR of 0 . 001 , 2 ) the maximum sequencing coverage across all the libraries ( Cmax ) ≥ 10 , 3 ) the total coverage across all the libraries ( Ctotal ) ≥ 40 , and 4 ) editing was detected in at least two libraries . Among the remaining sites that had editing events detected in at least two libraries , we defined Class II ( 447 editing sites ) with these criteria: 1 ) at FDR of 0 . 01 , 2 ) Cmax ≥ 5 and 3 ) Ctotal ≥ 16 ( we also employed other cutoffs to define editing sites in Class I and II , and obtained results not very different from the results reported here; see S3 Table for details ) . 824 sites do not meet the aforementioned two criteria but have P ( E1 ) > 0 . 99 , which suggests they might also be edited , although with lower confidence in brains of D . melanogaster ( Class III ) . Moreover , 131 sites have editing detected in at least one library but have P ( E1 ) ≤ 0 . 99 ( Class IV ) . Notably , we detected mRNA-Seq reads covering the remaining 2 , 157 candidate sites but none of them has editing events detected ( Class V ) . The detailed information about these sites is presented in S2 Table . It is not surprising that the sequencing coverage decreases in the order of Class I , II , and III in each library [the median coverage in each library is 17 . 4±1 . 4 ( mean ± s . e . throughout this study ) , 3 . 87±0 . 35 and 1±0 raw reads , respectively; S3A Fig] . Interestingly , although the sites in Class II have significantly lower coverage compared to sites in Class I ( P < 0 . 05 in each library , Kolmogorov-Smirnov tests ) , the editing levels are even significantly higher in Class II than in Class I ( the median editing level in each library is 0 . 24±0 . 02 vs . 0 . 16±0 . 01 in Class II vs . I , S3B Fig ) . From another perspective , 45 . 5% of the Class I sites were edited in all the eight brain libraries , meanwhile , only 8 . 5% of the Class II sites were edited in all the eight brain libraries ( P < 0 . 01 , Fisher’s exact test; S3C Fig ) . Compared to Class I and II , sites in Class III have both lower coverage and editing levels ( S3A and S3B Fig ) . Sites in Class IV are extremely lowly edited and sites in Class V do not have any editing event detected in our samples; however , these two classes do not have the lowest sequencing coverage compared to the other three classes ( the median coverage in a library is 18 . 6±1 . 1 and 6 . 1±0 . 35 for Class IV and V respectively , S3A Fig ) , suggesting they might have negligible editing in brains of D . melanogaster . To estimate the false positive rates of the editing sites in each class , we analyzed the RNA-Seq datasets from paired samples of wild-type strain w1118 vs . Adar5G1 mutant of D . melanogaster as conducted previously [50 , 51] . We found 1 , 145 , 161 , and 103 editing sites in Class I , II , and III respectively that have editing events detected in w1118 heads , and correspondingly , 33 , 2 , and 7 of these sites were detected in the heads of Adar5G1 mutant , yielding a false positive rate of 2 . 88% , 1 . 24% and 6 . 80% for Class I , II , and III , respectively . Therefore , the sites in Class I and II captured the editing events in brains of D . melanogaster with high accuracy , and Class III sites were not considered in the down-stream analysis due to the high positive rate . For the sites in Class I and II , we identified 1 , 630 ( 1 , 243 in Class I and 387 in Class II ) sites overlapped with previous studies [47–50] , and 519 sites ( 459 in Class I and 60 in Class II ) are novel in this study ( S2 Table ) . It is not uncommon that many editing sites are not overlapping between studies in Drosophila [47–50]: on average 30 . 4±3 . 6% of the editing sites are shared in pairwise comparisons ( ranging from 12 . 8% to 54 . 7% , S4 Table ) ; and we observed comparable proportions of shared sites between our study and the previous ones: 21 . 8% , 36 . 7% , 61 . 1% and 28 . 2% of the Class I+II sites in our study are overlapping with Graveley et al . [47] , Rodriguez et al . [49] , St . Laurant et al . [48] , and Yu et al . [50] , respectively ( S4 Table ) . Importantly , when we pooled Class I and II together , we found the novel sites have comparable false positive rates as the common ones in the w1118 vs . Adar5G1 mutant analysis ( 8/242 = 3 . 31% vs . 27/1064 = 2 . 54% for the novel vs . common sites ) . Furthermore , 111 of the novel sites are annotated in Ramaswami et al . [52] , which is independent from this study . Taken together , we identified 2 , 114 “high-confidence” editing sites after combining sites in Class I and II ( 35 sites that have A-to-G difference in Adar5G1 mutants were removed ) , including 1 , 603 ( 75 . 8% ) sites overlapped with sites identified by previous studies [47–50] and 511 ( 24 . 2% ) novel sites ( Fig 1B ) . The novel sites have slightly higher sequencing coverage than the common sites in the brain libraries ( the median coverage is 25 . 1±1 . 5 and 20 . 5±1 . 1 for the former and latter , respectively , P < 0 . 05 in each library , KS tests; S3D Fig ) , but generally lower editing levels ( the median is 0 . 18±0 . 006 vs . 0 . 36±0 . 006 , P < 10−16 in each library , KS tests; Fig 1C ) . Moreover , compared to the common editing sites , the novel sites are generally edited in fewer brain samples: 42 . 9% of the common ones were detected in all the eight brain libraries , while only 20 . 0% of the novel sites were detected in all the libraries ( P < 10−16 , Fisher’s exact test; S3E and S3F Fig ) . Altogether these results suggest that these novel editing sites are genuine but lowly edited in the brains , and were probably diluted in the samples of previous studies that were carried out in heads or whole flies [47–50] . Among the 2 , 114 high-confidence sites ( Fig 1D ) , 235 ( 11 . 1% ) are in intergenic regions , 42 ( 2 . 0% ) are in ncRNAs , and 1 , 837 ( 86 . 9% ) are in 517 protein-coding genes , including 20 ( 0 . 95% ) in 5' UTRs , 550 ( 26 . 0% ) in introns , 414 ( 19 . 6% ) in 3' UTRs , 678 ( 32 . 1% ) nonsynonymous ( in CDS regions and causes amino acid changes when edited , abbreviated as N throughout this study ) , and 144 ( 6 . 8% ) synonymous ( in CDS regions but do not cause amino acid changes when edited , abbreviated as S ) , and one editing site ( chr3R:18806029 ) that putatively disrupts the stop codon of CG18208 ( UAG>UGG ) . The detailed annotation in each library was presented in S5 Table . The gene ontology analysis revealed that the high-confidence exonic editing sites were significantly enriched in genes that encode transporters , synaptic vesicles or neurotransmitters ( S6 Table and S7 Table for male and female brains , respectively; and the top 50 genes that had the largest number of editing sites were presented in S8 Table ) . For the exonic editing sites , the editing levels ( averaged across libraries ) decrease in the order of N ( 0 . 319±0 . 010 ) , S ( 0 . 214±0 . 017 ) , 3' UTRs ( 0 . 168±0 . 008 ) , and 5' UTRs ( 0 . 133±0 . 020 ) , with levels in N sites significantly higher than those in the other three categories in the brains of D . melanogaster ( P < 0 . 001 , KS test; Fig 1D ) , suggesting that high levels of nonsynonymous editing events are overall favorable . Among the 550 intronic editing sites , 167 of them might also be exonic due to alternative splicing ( we only used annotations of the canonical transcript and some intronic sites in the canonical transcripts might be coding in the non-canonical transcripts ) , and the coverage and editing levels ( 0 . 336±0 . 013 ) are comparable to the N sites ( Fig 1D ) . Interestingly , the remaining 383 authentic intronic sites generally have significantly lower coverage than the coding regions ( Fig 1D ) , however , high editing levels in these sites ( 0 . 418±0 . 006 ) were observed , supporting previous results that editing is exerted co-transcriptionally [49] . We uncovered a tendency that A-to-I editing events were more readily detected in the genes with higher expression levels ( or adenosine sites with higher mRNA-Seq coverage ) . In each brain sample , when we grouped the expressed genes into 20 bins with increasing expression levels ( only genes with RPKM ≥ 1 were considered ) , we found a significant positive correlation between the editing density ( hereafter defined as the number of edited out of the total adenosine sites ) and the gene expression level ( P < 0 . 001 in each library; S4A Fig ) . Similar patterns were observed if we grouped all the adenosine sites with increasing mRNA-Seq coverage in each sample ( only sites ≥ 5X coverage were considered; S4B Fig ) . Analogous results were obtained if we weighted each editing site with its editing level ( “level-weighted density of editing sites” , see S4A and S4B Fig ) . Consistent with previous observations [47–49] , we found the editing density was significantly increased from the 5' to 3' of pre-mRNAs . After dividing the adenosine sites ( ≥ 5X coverage ) into 20 equal bins along their positions in pre-mRNAs , our meta-gene analysis indicated that the editing density in each bin was significantly positively correlated with the relative distance of that bin to the transcriptional start sites ( P < 0 . 005 in each library; S4C Fig ) . Despite our experimental optimization , the poly ( A ) selection procedure still caused slightly increased coverage bias towards 3' ends of mRNAs ( S4D Fig ) . However , we found the coverage difference between 5' and 3' of mRNAs was not the main cause of elevated editing density in the 3' ends of mRNAs with two analyses . First , in each library , we split each gene ( RPKM ≥ 1 ) into two equal parts , calculated the RPKM values for each half-gene separately , ranked all the half-genes with increasing RPKM values , and grouped them into 20 bins . Next , in each bin , we combined the 5' and 3' half-genes independently and calculated the editing density in the 5' and 3' half . We found within each bin the editing density in the 3' half-genes are significantly higher than the 5' half genes ( P < 0 . 05 in each library; paired t tests , S4E Fig ) . To further reduce the coverage variation within the half-genes , we ranked all the adenosine sites ( ≥ 5X ) with increasing coverage and binned them into 20 groups , and in each group we calculated the editing density in the 5' ( front ) half and 3' ( rear ) half of pre-mRNAs independently . We also found the editing density were significantly higher for sites in the rear half compared to sites in the front half of pre-mRNAs ( P < 0 . 001 in each library; paired t tests , S4F Fig ) . Taken together , the increasing editing density along mRNAs is not likely caused by detection bias , but more likely shaped by the recruitment of ADAR to the transcription elongation complex , as previous functional studies demonstrated [49 , 80] . We predicted that 591 ( 50 . 1% ) of the 1 , 179 exonic editing sites were located in stable local hairpin structures of mRNAs ( Materials and Methods ) , such as Adar ( S5A Fig ) , rtp , DIP1 , rdgA , CG43897 , and CG42540 ( editing events in these genes were verified with Sanger sequencing; S5B and S5C Fig ) . In contrast , we obtained only 363 exonic sites ( 332–393 sites within 95% CI ) located in stable hairpin structures after comprehensively folding all the transcripts expressed in brains and randomly sampling the equal amounts of editing sites ( Materials and Methods; Fig 1E ) . Similar results were obtained when we focused on the N or S editing sites individually ( P < 0 . 002 in simulations for both cases; Fig 1E ) . In addition , we found 181 intronic editing sites located in stable hairpin structures when we folded the pre-mRNA sequences . Long-range pseudoknots are another class of RNA substrates recognized by ADAR [81] . By extensively folding the flanking sequences of the editing sites ( see Methods for details ) , we inferred 260 ( 22 . 1% ) exonic editing sites that were located outside stable hairpin structures but were located in stems of long-range pseudoknots in pre-mRNAs of genes , such as the 3' UTR of Adar ( S5D Fig ) , nrm , B52 , nAchRbeta1 , CG8034 and roX1 ( S6 Fig; the editing events in nrm were verified by Sanger sequencing of the cDNA and genomic DNA , S6B and S6C Fig ) . Taken together , our results systematically demonstrated that at least 874 ( 74 . 1% ) of the exonic A-to-I editing sites in the brains of D . melanogaster were located in pre-mRNA regions that formed stable secondary structures . These results also well explain why the A-to-I editing sites are located in clusters , as commonly observed in previous studies [41 , 47–49 , 62] . By clustering the editing sites with distances smaller than 100 nucleotides , we identified a total of 1 , 320 editing sites that form 413 clusters in brains of D . melanogaster ( S7 Fig ) , and unusually large editing clusters were frequently observed , such as in NaCP60E and CaMKII ( the Sanger verification was presented in S8 Fig ) . To characterize the A-to-I editing events that were evolutionarily conserved ( i . e . , commonly observed ) across species , we deep sequenced the poly ( A ) -tailed transcriptomes of female and male brains of 1- to 5-day-old D . simulans and D . pseudoobscura that were accommodated at the same temperature conditions as D . melanogaster ( six libraries for each species ) . The mapped reads range from 8 . 7–16 . 4M in each library of D . simulans , and 10 . 9–17 . 8M in D . pseudoobscura ( Table 2 , Table 3 and S1 Table for detailed information ) , and the median sequencing coverage on an exonic site in a library ranges from 5 to 9 reads in D . simulans , and ranges from 4 to 5 in D . pseudoobscura ( S1 Fig ) . D . simulans diverged from D . melanogaster ~5 . 4 million years ago ( Fig 2 ) while D . pseudoobscura diverged from D . melanogaster approximately 55 million years ago [82] . Comparing A-to-I editing across these three species will help us understand the role of natural selection in shaping the brain editomes during evolution . To exclude SNPs in the RNA editing characterization , we also deep sequenced the genomic DNA of the same strain of D . simulans ( the median coverage per site is 46 , totally 313 , 133 SNPs , S9A Fig ) and D . pseudoobscura ( the median coverage per site is 47 , totally 489 , 828 SNPs , S9B Fig ) and masked all the SNPs ( Materials and Methods ) . For each high-confidence editing site in brains of D . melanogaster , we employed two complementary approaches to search for the orthologous sites in D . simulans and D . pseudoobscura . First , we used liftOver [83] to convert the genomic coordinates of the orthologous sites between D . melanogaster and D . simulans , or between D . melanogaster and D . pseudoobscura , based on the pairwise genome alignments as previously conducted [51] ( termed “g_align” approach , Materials and Methods ) . Second , we parsed out the genomic coordinates with the pairwise CDS alignments that were made based on the protein alignments between D . melanogaster and the other species ( “c_align” approach ) . We pooled orthologous sites by the two approaches together . For each site in each species , we calculated the joint probability that this site is edited in at least one library P ( E1 ) . At FDR of 0 . 05 , we identified 996 sites edited in D . simulans ( S9 Table ) , and 451 sites edited in D . pseudoobscura ( S10 Table ) , and 367 sites edited in both D . simulans and D . pseudoobscura ( Fig 2 ) . We present the editing sites evolutionarily conserved in the same gender under the same temperature conditions in D . simulans ( Table 2 and S11 Table ) and D . pseudoobscura ( Table 3 and S12 Table ) . For the editing sites we characterized in the brains of D . melanogaster , 34 . 3–44 . 0% of them have editing events detected in the matched samples of D . simulans ( Table 2 ) , and 22 . 3–24 . 1% of them have editing in the matched samples of D . pseudoobscura ( Table 3 ) . Note the proportion of editing sites in D . melanogaster that have editing events detected in brains of other species varies across libraries since we required the sites are edited in both paired samples . In general , with divergence increases , the level of conserved editing sites decreased , suggesting the editing events are evolutionary dynamic . Notably , for the 996 editing sites with conserved events in both D . melanogaster and D . simulans , and the 451 editing sites with conserved events in both D . melanogaster and D . pseudoobscura , we found 416 ( 41 . 8% ) and 78 ( 17 . 3% ) of them are located outside the coding regions , respectively ( Tables 2 and 3 ) , which is consistent with a recent study [84] and suggests a possible functional role for these sites , such as influencing alternative splicing [26–29] , microRNA targeting [30–32] , or other cellular processes related to RNAs [14] . Comparing the editing sites with conserved and non-conserved events revealed two interesting features . First , in each brain library , the editing levels are significantly higher in the sites with evolutionarily conserved editing events than in the remaining sites ( the mean level in a library is 0 . 340±0 . 008 vs . 0 . 252 ± 0 . 009 in the D . melanogaster/D . simulans comparison , and 0 . 323±0 . 012 vs . 0 . 187±0 . 009 in the D . melanogaster/D . pseudoobscura comparison; P < 0 . 01 in each comparison , KS tests , S10 Fig ) . Second , the N sites are significantly enriched in the editing sites that are evolutionarily conserved: 72 . 9% ( 494 out of 678 ) N sites compared to 35 . 0% ( 502 out of 1436 ) of the remaining sites that have evolutionarily conserved events between D . melanogaster and D . simulans ( P < 10−10 , Fisher’s exact test ) , and 47 . 9% ( 325 out of 678 ) N sites compared to 8 . 77% ( 126 out of 1436 ) of the remaining sites that have evolutionarily conserved events between D . melanogaster and D . pseudoobscura ( P < 10−10 , Fisher’s exact test ) , suggesting the nonsynonymous editing events are generally maintained and regulated by different evolutionary forces compared to the other sites . There are 84 editing sites that have editing events detected in both D . melanogaster and D . pseudoobscura but without editing events confidently identified in D . simulans . Nevertheless , this does not necessarily mean these sites are not edited in D . simulans ( for 60 of these sites we did not find the orthologous sites in D . simulans , and for the 24 remaining sites , 10 of them have low level of editing but are undistinguishable from sequencing errors; S13 Table ) . Sampling bias frequently causes the sites with low expression or low editing levels to yield no editing signals in the sequencing libraries . Therefore , next we only focused on the sites with high sequencing coverage to explore the possible gain and loss patterns of editing events . We obtained 87 editing sites that have minimal editing level of 0 . 05 in D . melanogaster and have at least 200 raw reads ( across all the libraries ) in both D . simulans and D . pseudoobscura . We found 52 sites with editing events reliably detected in all the three species . For each of the remaining 35 sites , in case no editing event was discovered at a site in a sample m in D . simulans ( or D . pseudoobscura ) , we calculate Pm ( D0 ) , the probability that this observation happens by sampling bias or because the editing signal was abolished by sequencing error ( ε ) , given a depth of Cm and an assumed editing level lm at that site ( Materials and Methods ) . We assumed the orthologous sites in the other species have the same editing level as in D . melanogaster and calculated the joint probability P ( D0 ) that a site was edited despite zero edited allele was detected in all the libraries . After correcting for multiple testing , at FDR of 0 . 05 , we found 20 sites with editing present in both D . melanogaster and D . simulans but absent in D . pseudoobscura , and 3 sites with editing specifically present in D . melanogaster . The most parsimonious interpretation is that the brain editome in Drosophila is expanding during evolution ( Fig 2 ) . We did not find any convincing case that editing was detected in both D . melanogaster and D . pseudoobscura but was absent in D . simulans , suggesting that the established editing events , at least for the set we examined here , are well maintained by natural selection during evolution . In contrast to previous observations that nonsynonymous editing events were generally non-adaptive in mammals [67 , 68] and in Drosophila [50] , our analysis revealed the nonsynonymous editing events in Drosophila brains were predominantly adaptive . The ratio of nonsynonymous ( N ) to synonymous ( S ) editing sites ( N/S ) in different brain samples of D . melanogaster ranges from 4 . 79 with 95% CI ( 3 . 85 , 6 . 20 ) to 6 . 25 ( 4 . 70 , 8 . 67 ) , all of which is significantly higher than the ratio expected under neutrality ( 3 . 80 ) that was calculated similarly as previously described [67] ( Materials and Methods; P < 0 . 03 in each library , Fisher’s exact tests; Table 1 ) . In other words , in the brains of D . melanogaster , the rate of nonsynonymous A-to-I editing is significantly higher than the rate of synonymous editing . Given the observed and expected N/S ratios under neutrality ( randomness ) , a conservative estimation is that 20 . 7% [with 95% CI ( 1 . 3% , 38 . 7% ) ] of the N sites in the brains of D . melanogaster might be adaptive ( Table 1 ) . Moreover , we obtained higher N/S ratios in each brain library when we increased the cutoff of editing level ( S11 Fig ) . Our analysis is essentially the same as the classical dN/dS ( the number of nonsynonymous changes per nonsynonymous site over the number of synonymous changes per synonymous site ) test of DNA sequences in molecular evolution [85] , and provides compelling evidence that the nonsynonymous editing events in Drosophila brains are overall beneficial and favored by natural selection . We observed significantly negative correlations between the sequencing coverage ( C ) and editing level ( l ) in each brain library or when we pooled the library together ( P < 10−10 in each case , S12 Fig ) . These patterns do not necessarily mean that lowly expressed sites have higher editing levels , but rather suggest that the sites of lower sequencing coverage have stronger sampling bias: editing events in such sites are either not detected or detected with over-estimated editing levels . Furthermore , although the S and N editing sites have comparable coverage ( Fig 1D ) , the S sites are generally edited at lower levels compared to the N sites ( Fig 1D ) . Therefore sampling bias would affect S sites more severe than the N sites , which potentially causes over-estimation of the N/S ratios in the above analysis . However , to what extent the N/S ratio is over-estimated due to sampling bias remains unclear . Our “joint probability method” in detecting editing across multiple libraries allows the editing sites with low coverage or with low editing levels in a single library to be efficiently identified with the aid of information from other libraries ( such sites are usually filtered out if only based on information of a single library ) , and the full list of editing sites across libraries enable us to test whether and how our observed N/S ratios are affected by sampling bias . We conducted simulations with two different methods , both of which considered the observed distribution of sequencing coverage and editing levels among sites . In the first method , we focused on all the high-confidence sites present in a library that had sequencing coverage C ≥ Cmin and editing level l ≥ lmin . In each round of simulation , for a site j that had an observed depth Cj , we randomly sampled Cj reads ( with replacement ) from all the sequenced reads covering that site and calculated the simulated editing level lsj with the obtained reads of the edited allele , we counted the site only if lsj ≥ lmin , and then we pooled all the counted sites together and calculated the N/S ratio for this round of simulation . For each brain library , we tried different combinations of Cmin ( ranging from 5 to 20 , with Cmin = 20 roughly accounting for 50% of all the editing sites in a library ) and lmin ( 0 . 01 , 0 . 02 and 0 . 05 ) values , and performed the simulations for 1 , 000 replicates . At different lmin cutoffs , both the observed and simulated N/S ratios were generally higher at lower Cmin values; and the median simulated N/S ratio is usually higher than the observed one , but the extent of elevation is very modest , in most cases much smaller than 10% ( see Fig 3A and 3B for all the eight brain libraries at lmin of 0 . 02 , and S13 Fig for other lmin values ) . Similar results were obtained when we pooled all the libraries together and performed the simulations ( S14 Fig ) . These results suggest that detection bias of editing sites would slightly increase the observed N/S ratio , however , the over-estimation caused by such a bias is very modest compared to the large difference between the observed and the expected N/S ratio under neutral evolution ( Fig 3B ) . Importantly , the observed N/S ratio is significantly higher than the neutral expectation even after we adjusted the bias ( S14 Table ) . In the second method , we pooled all the eight libraries together and randomly sampled ( with replacement ) a fraction ( f ) out of the total reads , and after that , we calculated the N/S ratio for the sites that have simulated editing levels lsj ≥ lmin . We tried different combinations of f ( from 0 . 05 to 1 with a step size of 0 . 025 ) and lmin ( 0 . 01 , 0 . 02 and 0 . 05 ) values , and performed the simulations for 1000 replicates . In agreement with the first method , the simulated N/S ratios were higher at lower depth ( smaller f values ) . For example , when f is set at 0 . 05 , which is less than half the size of a library we sequenced , the N/S ratio would be elevated by roughly 10% due to sampling bias ( Fig 3C ) . However , the simulated N/S ratios approached to the observed N/S ratio ( calculated based on the pooled libraries ) rapidly with increasing f ( Fig 3C ) . In summary , the simulations revealed that N/S ratios tend to be overestimated at lower sequencing coverage , however , the degree of over-estimation was usually small given our sequencing depth . Our conclusion that the N/S ratios in the brain libraries are significantly higher than the neutral expectation is not affected by the possible detection bias . It is notable that several X-linked genes such as cac , CG42492 and Sh harbor multiple N editing sites ( 12 , 10 and 3 respectively ) while very few S editing sites ( 1 , 1 , and 0 respectively ) . In fact , the N/S ratio is substantially higher for the X-linked than the autosomal genes in all the eight brain libraries of D . melanogaster , indicating the signal of adaptation is generally stronger for the X-linked genes in Drosophila brains ( S15A Fig ) . This observation is essentially congruent with the fast-X evolution observed for nonsynonymous DNA mutations under positive selection , by which the X-linked advantageous effect is more readily manifested compared to the autosomal counterparts [86 , 87] . Furthermore , we identified a set of brain-expressed gene candidates that had N editing sites favored by natural selection , which were termed “PSEB” ( Positively Selected Editing in Brains ) gene set ( we required each gene to have a N/S editing ratio > 5; totally 223 genes met this criteria and 49 of them were X-linked; S15 Table ) . There are 683 high-confidence editing sites in the PSEB genes in brains of D . melanogaster , including 447 N and only 8 S sites , yielding a N/S ratio of 55 . 9 . 80 ( 35 . 9% ) of the PSEB genes are overlapping with the type III genes that have editing events positively selected by Yu et . Al . [50] . The N editing sites in the PSEB genes , which are significantly enriched in chemical and electrical neurotransmission pathways ( Fig 4 ) , are very likely the targets of positive selection . Notably , the expression levels of the PSEB genes are significantly higher than the non-PSEB genes ( S16 Fig ) , which suggests that the higher N/S ratios in the PSEB genes are not likely caused by sampling bias of the S sites . Moreover , we ranked all the genes with editing events in each brain library with increasing expression level ( RPKM ) and equally divided those genes into “Highly” and “Lowly” expressed groups . In both groups , the N/S ratios are significantly higher for the editing sites in the PSEB compared to non-PSEB genes ( S16 Table ) . Furthermore , we conducted random sampling simulations and confirmed that neither the higher N/S ratios in the PSEB genes nor lower N/S ratios in the non-PSEB genes was caused by detection bias ( Fig 3D ) . Notably , the editing levels of the PSEB sites are higher than the remaining sites in all the brain libraries , and four of them are statistically significant ( S17 Fig ) , which further suggests that editing in these sites are favored by natural selection . We observed higher N/S ratios in the sites that have editing events commonly observed across Drosophila species , partially caused by the bias that synonymous adenosine sites are less constrained during evolution . However , after we contrasted the observed N/S ratios to the expected values ( 4 . 18 ) calculated from the evolutionarily conserved adenosine sites ( Materials and Methods ) , we still detected signals of adaptation in the sites with conserved editing events between D . melanogaster and D . simulans: the N/S ratio ranges from 7 . 11 ( 4 . 97 , 10 . 95 ) to 8 . 73 ( 5 . 90 , 14 . 29 ) across libraries , and all ratios are significantly higher than the neutrally expected ratio ( P < 0 . 005 in each comparison , Fisher’s exact tests; Table 2 ) . We also observed higher N/S ratios for the conserved editing sites in X chromosome compared to those in the autosomes ( S15B Fig ) . Importantly , the adaptation signals are primarily detected in the PSEB genes , and considerably lower N/S ratios ( ranging from 2 . 73 to 3 . 60 ) were observed in the non-PSEB genes ( S11 Table ) , suggesting the conserved N sites in the non-PSEB genes are unlikely favored by natural selection . Analogous results were obtained for the conserved editing sites between D . melanogaster and D . pseudoobscura ( S12 Table ) . Altogether , our results suggest that the N editing sites in the PSEB genes are favored and maintained by natural selection while the S editing sites , which are putatively neutral , might degenerate during long-term evolution , which generates even higher N/S ratios in the sites with editing events conserved between species [50 , 68] . Next we evaluated the effect of local nucleotide contexts on the comparison of observed vs . expected N/S ratios . For each of the 2 , 114 high-confidence editing sites , we extracted the upstream and downstream 3 nucleotides ( we also used other number of nucleotides and obtained similar results ) , counted the number of nucleotide at each position ( S17 Table ) , and developed a position probability matrix ( S18 Table ) . Consistent with observations in primates that local nucleotide contexts affect the editing efficiencies [60 , 88] , we also found that G immediately upstream a focal editing site was generally not favored , the nucleotide immediately downstream the editing site was slightly biased toward G , and other flanking nucleotides were generally not important , although the overall patterns of preferences were weak in D . melanogaster ( Fig 5A; we presented the frequencies of the tri-nucleotides centered with the editing sites and background adenosines in S19 Table ) . We scanned mRNAs of the edited genes with the position probability matrix and scored each 7-mer sequence that was centered with adenosine ( Materials and Methods ) . We chose the score cutoff that specified the top 90% quantile of the high-confidence editing sites ( -0 . 622 ) , and found 75 . 4% of the background 7-mer sequences were above this score cutoff ( Fig 5B ) . For the background adenosine sites with scores above the cutoff , the expected N/S ratio was 3 . 11 ( even lower than 3 . 80 , which was the expected N/S ratio based on all the adenosine sites ) , significantly lower than the observed N/S ratio ( 4 . 62 ) for all the high-confidence editing sites that have scores above the threshold ( P < 0 . 0001 , Fisher’s exact test ) . Therefore , our conclusions based on the comparisons between the observed and expected N/S ratios are not affected after considering effect of the local nucleotides contexts on editing . We calculated the divergence of all the protein-coding genes between D . melanogaster and D . simulans ( Materials and Methods ) , and we found PSEB genes have significantly lower dN and dS values compared to the non-PSEB genes expressed in brains of D . melanogaster ( dN is 0 . 0107 ± 0 . 0012 vs . 0 . 0193 ± 0 . 0003 for PSEB vs . non-PSEB , and dS is 0 . 1025 ± 0 . 0025 vs . 0 . 1262 ± 0 . 0005 for PSEB vs . non-PSEB; P < 10−8 in both comparison , KS test ) , possibly due to the anti-correlation between expression levels and evolutionary rate [88–91] since the PSEB genes are usually expressed at higher levels ( S16 Fig ) . These observations are consistent with the hypothesis that A-to-I editing increases the mutation sequence space of protein-coding genes . Next we ask whether we can detect higher editing densities in the evolutionarily conserved genes . We ranked all the protein-coding genes expressed in Drosophila brains with increasing dN values between D . melanogaster and D . simulans and grouped the genes into 20 bins . The editing density of the N sites ( ≥5X coverage ) was significantly inversely correlated with the dN value in each bin ( rho ranges from -0 . 824 to -0 . 734 , P < 0 . 001 in each library , S20 Table; see Fig 6A for 1- to 5-day female and male brains ) . We observed similar patterns when we ranked all the possible nonsynonymous adenosine sites ( ≥ 5X coverage was required ) with increasing phyloP scores ( higher scores mean higher conservation levels ) and grouped them into 20 bins: in each library , the density of N sites is significantly positively correlated with the median phyloP score of that bin ( rho ranges from 0 . 711 to 0 . 832; P < 0 . 001 in each library , S21 Table; Fig 6B for 1- to 5-day female and male brains ) . Analogous but weaker correlations were observed for the S sites when we grouped the genes with dN value ( rho ranges from -0 . 560 to -0 . 402 for each individual library , P < 0 . 1 in each library , S20 Table ) or grouped the synonymous editing sites with phyloP scores ( rho ranges from 0 . 519 to 0 . 738; P < 0 . 05 in each library , S21 Table ) . However , one potential pitfall of our analysis is that the conserved genes ( with low dN ) or sites ( with high phyloP scores ) usually have higher expression levels [88–91] , which would lead to detection bias as higher editing densities were found in genes ( or sites ) with higher sequencing coverage ( S4A and S4B Fig ) . To exclude such a possibility , in each library , we ranked all the nonsynonymous adenosine sites with increasing sequencing coverage ( ≥ 5X was required ) and grouped them into 20 bins . Within each bin , we further divided the sites into two equal-sized subgroups based on the phyloP scores . In each library , the editing density in the nonsynonymous adenosine sites is significantly higher in the conserved subgroup compared to the non-conserved subgroup ( P < 0 . 001 in each comparison , paired t test , Fig 6C and S18 Fig ) . Analogous patterns were observed for the synonymous editing sites as well ( S19 Fig ) . Therefore , the elevated occurrences of A-to-I editing events in the evolutionarily conserved genes ( or sites ) in Drosophila are not likely caused by detection bias due to gene expression levels ( or bias of sequencing coverage ) . Altogether , our results further support the adaptation hypothesis of RNA editing [12–14] since the sequence space of the evolutionarily conserved genes is generally inaccessible through DNA mutations , while RNA editing provides an epigenetic approach to expand proteomic diversity temporally and spatially . So far our analysis revealed prevalent beneficial editing sites in brains of Drosophila and the majority of them were enriched in the PSEB genes . Next we asked whether we can observe similar patterns in whole bodies of D . melanogaster adults . We deep sequenced the poly ( A ) -tailed transcriptomes of female and male adults from five strains of D . melanogaster that were collected from five continents [79] ( Materials and Methods ) . We sequenced 15 . 4–28 . 6M reads that were mapped on the reference genome in each library ( S1 Table ) , and the median sequencing coverage on an exonic site in a library ranges from 17 to 31 reads in female , and ranges from 11 to 21 in male adults ( S20 Fig ) . We masked all the SNPs in these five and other related strains which were sequenced previously [79] , so that we only focused on the DNA sites that were exclusively adenosines across the five strains of D . melanogaster . For each site with A-to-G difference in a strain k , we calculated the probability that this site was edited Pk ( E1 ) , and then we calculated the joint probability that this site was edited in at least one strain P ( E1 ) . We analyzed female and male adults independently and required each site to have at least 10 raw reads in each library . At FDR of 0 . 05 , we identified 910 candidate editing sites in female and 1 , 458 candidates in male adults in exons . We obtained a false positive rate of 4 . 17% ( 26 out of 624 ) in female and 2 . 99% ( 29 out of 969 ) in male adults with the w1118 vs . Adar5G1 mutant analysis . After removing the false positive sites in Adar5G1 , we obtained 875 exonic editing sites in female ( S22 Table ) and 1 , 422 exonic sites in male adults ( S23 Table ) ( 719 overlapped between female and male adults ) . We obtained mixed results in detecting signals of adaptation when we compared the observed N/S ratios to the expected under neutrality ( 3 . 80 ) in both female and male adults ( compared to the neutral expectation , the observed N/S ratio was significantly higher in strain B12 and N10 , significantly lower in T07 and ZW155 , and no significant difference was observed in I17 , Table 4 ) . As observed in brains , the N/S ratio was significantly higher in PSEB genes compared to neutral expectation in all the five strains ( P < 0 . 001 in each comparison; Table 4 ) . In contrast , in all the five strains the observed N/S ratio was significantly lower in the non-PSEB genes compared to neutral expectation ( P < 0 . 001 in each comparison; Table 4 ) . The patterns held when we separated the genes with editing events into “Highly” and “Lowly” expressed groups based on their expression levels ( S24 Table ) . We also conducted simulations by randomly sampling the reads covering each site with increasing cutoffs of coverage ( Cmin ) or editing level ( lmin ) , and our simulation results constantly confirmed these observed patterns ( S21 Fig ) . Our results indicate the N/S ratio is much higher in editing sites in the PSEB genes while strong purifying selection is acting on editing sites in the non-PSEB genes . We uncovered significant but not high correlations in editing levels between strains in female ( pairwise Pearson’s r was 0 . 718 ±0 . 018 , S25 Table ) and male adults ( r was 0 . 804 ±0 . 013 , S26 Table ) , suggesting that editing levels were variable across strains of D . melanogaster as previously observed [41] . Notably , we found a considerable number of sites at which editing events was readily detected in certain strains while absent in the other strains . We hypothesize that editing in such sites might be polymorphic in the populations , although reliably identifying such sites is challenging . We took a multiple-step procedure to identify the polymorphic editing sites from all the sites we detected ( 875 in female and 1 , 422 in male adults ) . First , we filtered the sites that have editing events detected in D . simulans or D . pseudoobscura . Second , we required a site to have editing reliably detected in at least one strain k with Pk ( E1 ) > 0 . 999 and to have no editing detected in at least one strain m . Third , we calculated Pm ( D0 ) , the probability that the editing was not detected at depth Cm due to sampling bias or sequencing error in the strain m . Finally we calculated the joint probability P ( D0 ) if no editing was observed in multiple strains at that site ( Materials and Methods ) . For each site , a required parameter in calculating Pm ( D0 ) in the strain m which has no editing detected is the authentic editing level in that strain . We used two levels of stringency to calculate Pm ( D0 ) . Level I: at each site , we assumed editing level is identical across all the strains , and the averaged level from the strains with reliably editing signals was used . Level II: we assumed the editing level is 0 . 05 in the strain with no editing detected if the averaged level from the detected strains is greater than 0 . 05 , and the averaged level was used if it was smaller than 0 . 05 . We conducted the analysis in females and males independently . Finally under Level I we obtained 165 and 179 candidate polymorphic editing sites in females and males respectively ( 57 sites overlapped ) . The averaged editing level for a site ( based on the strains with reliable editing events ) is 0 . 125±0 . 010 in females and 0 . 095±0 . 008 in males ( we obtained 117 and 125 candidate sites under Level II in females and males respectively , and 44 sites were overlapping; S27 and S28 Table ) . Interestingly , among these putatively polymorphic editing sites , we did not find signal of adaptation in editing of the PSEB genes ( in females the N/S ratio was 4 . 67 and 3 . 5 under Level I and II respectively; and in males the N/S ratio was 3 and 2 . 25 under Level I and II respectively; Table 5 ) . Moreover , we observed even lower N/S ratios in editing of the non-PSEB genes ( 1 . 47 and 1 . 48 under Level I and II respectively in females; 1 . 96 and 1 . 74 under Level I and II respectively in males; Table 5 ) . We observed similar patterns when we individually examined the N and S sites with respect to the number of strains in which such editing events were detected ( S29 Table ) . In contrast , we observed significantly higher N/S ratios in the editing sites of PSEB genes that were fixed in D . melanogaster . To reliably detect the sites with editing events fixed in the populations of D . melanogaster , we first identified the sites at which the probability of editing in each strain Pk ( E1 ) > 0 . 95 ( k was B12 , I17 , N10 , T07 and ZW155; female and male adults were studied separately ) . Due to the small number of strains we used , we further sequenced the transcriptomes of female and male adults of D . simulans and required the orthologous sites to be edited in the same gender of D . simulans . We obtained 181 sites ( 101 in PSEB and 80 in non-PSEB genes , S30 Table ) fixed in female and 373 editing sites ( 194 in PSEB and 179 in non-PSEB genes , S31 Table ) fixed in male adults ( 171 overlapped between female and males ) . The editing levels are higher in the fixed sites compared to the polymorphic ones: ( the averaged editing level per site is 0 . 149±0 . 014 in female and 0 . 197±0 . 011 in male adults ) . In Table 5 we show the fixed editing sites have significantly higher N/S ratios in PSEB genes ( 44 in females and 34 . 2 in males ) compared to the neutral backgrounds ( 4 . 18 , calculated with the adenosine sites that are in the genes expressed in adults and conserved between D . melanogaster and D . simulans ) . Furthermore , in the PSEB genes the N/S ratio is significantly higher in the fixed editing sites compared to the sites showing polymorphic editing patterns in females and males after adjusting the background difference ( Table 5 ) . Nevertheless , strong signals of purifying selection were observed in the fixed editing sites in the non-PSEB genes ( Table 5 ) . We also did not find significant difference in the N/S ratio between the fixed and polymorphic editing sites in the non-PSEB genes in both female and male adults ( Table 5 ) . Previous studies have demonstrated that mutations influence mRNA secondary structures and hence the efficiency of RNA editing in natural populations of D . melanogaster [41 , 92] . Here we ask whether we can find SNPs associated with the variations in editing levels across the five strains of D . melanogaster . To increase the statistical power , we only focused on the 58 editing sites that are polymorphic ( under Level I ) in both female and males ( Materials and Methods ) . Finally , we found 39 out of the 58 sites are associated with SNPs that are within 10kb flanking the editing sites , and on average each editing site is associated with 22 . 4 ± 3 . 7 SNPs , and the nearest distance between a SNP and the editing site is 770 ± 140 nts ( S32 Table ) . Meanwhile , we conducted the same analysis on the 171 sites that have editing events fixed in both females and males , however , we did not observe any of those sites have editing levels associated with SNPs under the same criteria . These comparisons suggest cis-regulatory elements affect the levels or status of editing across strains , as previously observed [41 , 84 , 92] . Compared to hard-wiring a particular codon change in the genome , A-to-I editing may give the flexibility to quickly respond to environmental stress and adjust the activity of final protein product accordingly [15] . For example , the A-to-I editing level can be regulated by external stimuli [93 , 94] , energy and nutrient [95] , and hypoxic conditions [96 , 97] . Notably , temperature increases would reduce the thermo-stabilities of mRNA secondary structures and down-regulate the expression level of Adar in Drosophila , both of which would reduce the global editing efficiency [16] . The editing levels of Adar and a handful of other genes ( totally 54 sites ) have been examined under elevated temperatures [16] . Nevertheless , how temperature affects the editing sites at the genome-wide level remains unclear . Herein , we compared the editing levels in the brains of 1- to 5-day-old fly adults constantly raised at 25°C with those in the brains of flies raised at 25°C but treated at 30°C for 48 hours ( i . e . , B2 vs . B4 for female , and B6 vs . B8 for male brains of D . melanogaster in Table 1 ) . The changes of editing levels were significantly positively correlated between D . melanogaster and D . simulans ( S22 Fig; see Table 2 for information of D . simulans ) , consistent with previous observation that editing level changes under higher temperature were evolutionarily conserved[16] . The editing levels in general decreased under elevated temperatures for all functional categories of editing events , however , the decrease in editing levels was considerably smaller for the nonsynonymous compared to the silent sites ( Fig 7A ) . Interestingly , when we folded the flanking sequence of each editing event ( 100 nts at each side ) , we uncovered the nonsynonymous editing sites were located in more stable secondary structures than all the silent events ( P < 10−16 , Kolmogorov-Smirnov test , Fig 7B ) . This comparison suggests that Drosophila has evolved mechanisms to maintain or even enhance the levels of some nonsynonymous editing sites under elevated temperature , which might be important for temperature adaptation since they expanded proteomic diversities . We observed high correlations in editing levels between pairwise brain libraries ( Pearson’s r ranges from 0 . 849 to 0 . 933 , S33 Table ) , and interestingly , the highest correlation coefficients were usually observed between brains of flies that were maintained at the same accommodation conditions but not of the same gender . The pattern was more pronounced when we clustered the samples based on the editing levels of the high-confidence sites ( totally 391 sites that had at least 20 raw reads in each brain library ) : Females and males of 1–14 day old ( B1 and B5 , see Table 1 for annotations ) clustered together; females and males of 1–5 day old that were constantly raised at room temperature ( B2 and B6 ) , treated at 30°C for 14 hours ( B3 and B7 ) , and treated at 30°C for 48 hours ( B4 and B8 ) always clustered together ( Fig 7C; similar patterns were observed when the coverage cutoff was set 15 or 25 raw reads , S23A Fig ) . Moreover , we did not find any site with editing level significantly different between female and male brains of D . melanogaster under the same accommodation conditions ( i . e . , B1 vs . B5 , B2 vs . B6 , B3 vs . B7 or B4 vs . B8 ) after multiple testing corrections . Therefore , our results indicate that temperature plays a more important role than gender effect in shaping the global brain editomes . However , we found the species effect is generally stronger than the temperature effect on the editing levels when we clustered the samples of D . melanogaster and D . simulans with 289 sites that have at least 20 raw reads in each library ( Fig 7D ) or when we clustered the samples of D . melanogaster and D . pseudoobscura with 152 sites that have at least 20 raw reads in each library ( S23B Fig ) : the samples of the same species always clustered first , then the temperature conditions , and the gender effect was still very weak . Gene expression plasticity is a strategy organisms evolved for adapting to new environments [97] . Yet it remains elusive whether ( and how ) RNA editing and gene expression plasticity coordinately participate in temperature stress responses . We detected hundreds of genes that were significantly differentially expressed in the brains of D . melanogaster that were constantly raised at 25°C compared to those raised at 25°C and treated at 30°C for 48 hours ( B2 vs . B4 for female , and B6 vs . B8 for male brains ) . The down-regulated genes under elevated temperatures were enriched in the “oxidative phosphorylation” pathway in both female and male brains , while the up-regulated genes were enriched in the “ATP binding , ” “translation” and “response to temperature” functional categories in female brains and in the “ATP binding” and “regulation of transcription” pathways in male brains ( S34 and S35 Tables ) , which suggested gene expression plasticity was involved in the temperature stress responses but in a sexually dimorphic manner [98 , 99] . Interestingly , in both female and male brains of D . melanogaster , the changes in editing levels of the nonsynonymous sites were weakly but significantly positively correlated with the changes in expression levels of the host genes under various cutoffs of expression levels ( S24 Fig ) , suggesting these correlations were not artifacts caused by gene expression cutoffs . For example , a nonsynonymous A-to-I editing site ( chrX:1781840 ) in Adar mRNA causes a Ser ( AGU ) to Gly ( IGU ) change ( abbreviated as S>G change ) , and a previous study [16] and our data both indicated that the editing level of this S>G change ( S5A Fig ) and expression level ( S25 Fig ) of Adar mRNA were reduced in both female and male brains of D . melanogaster under elevated temperatures . In contrast , for synonymous sites we did not observe significant correlations between changes in editing levels and gene expression levels ( S24 Fig ) . Altogether , these results suggest that the nonsynonymous editing events might interplay with gene expression changes in temperature adaptation , although detailed mechanisms remain to be further explored .
By extensively characterizing RNA editing sites in brains of D . melanogaster and two related Drosophila species , we identified a considerable number of N sites in Drosophila brains that were adaptive and maintained by natural selection during evolution . Our analysis revealed the N/S ratios in the editomes of Drosophila brains were significantly higher than the neutral expectation . In contrast , we did not observe such a pattern in the editing sites of different developmental stages or whole flies of D . melanogaster that were identified by the modENCODE Project [47] ( Fig 8A ) or re-analyzed by Ramaswami et al . [52] ( S26 Fig ) . We also obtained mixed results when we compared the overall N/S ratios to the expected ratio under neutral evolution in female and male adults from five strains of D . melanogaster ( Table 4 ) . We found the significantly higher N/S ratios in the whole editomes of brains are mainly contributed by the N sites in the PSEB genes , which are favored by natural selection . The N/S ratio for editing sites in the PSEB genes is significantly higher than the neutral expectation in most developmental stages except in early embryos ( 0–16 hours ) or larvae ( Fig 8A ) , while N/S ratio for editing sites in the non-PSEB genes is lower than neutrality in all those samples ( Fig 8A ) . Importantly , in brains , ~60% of the N sites were contributed by PSEB genes , while only ~5% of the S sites were from PSEB genes , which considerably elevated the overall N/S ratio in the brain editomes ( Fig 8A ) . In contrast , in the late embryo , pupae and adults , although the N/S ratios for the PSEB sites were significantly higher than neutral expectation , less than 40% of the N sites were contributed by PSEB genes in each stage/tissue ( Fig 8B ) , and hence , the signatures of adaptation in PSEB sites are masked by the non-PSEB sites when pooling all the editing sites together ( Fig 8A ) . These patterns were constantly observed when we independently considered the “Highly” and “Lowly” expressed genes that have editing events in these samples ( S27 Fig ) . We also observed a trend that the editing levels in the PSEB sites were increased during Drosophila development ( S28 Fig ) . Furthermore , we observed a significant positive correlation between the expression level of Adar and the number of editing sites during Drosophila development in the modENCODE Project [47] ( Spearman’s rho = 0 . 413 , P = 0 . 023; S29A Fig ) . And the expression level of Adar was also positively correlated with the cumulative editing levels ( i . e . , summing up the editing levels of all the sites , as following Ref . [63] ) in the modENCODE samples ( rho = 0 . 487 , P = 0 . 006; S29B Fig ) . Importantly , the expression level of Adar is higher in brains compared to all the modENCODE developmental stages ( Fig 8C ) , and accordingly , we observed the highest number of editing sites in brains compared to other developmental stages ( S30 Fig ) . Notably , in early embryos ( 0–16 hours ) and larvae , either Adar was lowly expressed , or only a few PSEB genes were expressed in these stages ( Fig 8C ) , which putatively explains why we did not detect significantly higher N/S ratio for editing sites in PSEB genes in these stages ( Fig 8A ) . Taken together , our results demonstrate that the expression level of Adar , together with the expression profiles of the PSEB genes that have editing sites favored by natural selection , are important in shaping the overall N/S ratios in the global editomes at different developmental stages ( or tissues ) of D . melanogaster . We were able to predict secondary structures for ~74% of the exonic editing sites , including ~22% as long-range pseudoknots , which were often very challenging to find . Our results help understanding the molecular basis by which the editomes are regulated and maintained . First , our analysis reveals a considerable number of editing sites could be clustered due to promiscuous editing of multiple adenosines by ADAR simultaneously . Second , we found a significantly higher proportion of the N sites were located in stable hairpin structures of mRNAs than the silent sites ( 55 . 0% vs . 41 . 7% , P < 0 . 001 , Fisher’s exact test ) , and similarly , the flanking sequences of the N sites have significantly lower MFE compared to those of the silent sites ( Fig 7B ) . Since ADAR recognize double-stranded RNAs to exert editing , these findings provide the structural basis for the observed higher editing levels and excessive occurrences of the N sites . The density of editing events was significantly higher in the evolutionarily conserved genes than in the non-conserved ones ( Fig 6C , S18 Fig ) , which supports the hypothesis that editing increases the mutation sequence space . However , another competing explanation for this observation is that the secondary structures ( stable hairpins or pseudo-knots ) of mRNAs , which are ADAR substrates for RNA editing , constrain the evolutionary rates of these genes . Indeed , when we separately calculated the evolutionary rates between D . melanogaster and D . simulans for the CDS regions that formed secondary structures and harbored editing events in Drosophila brains ( termed “structured” ) and the remaining CDS regions after masking the secondary structures ( “structure masked” ) , we uncovered both dN and dS were significantly lower in the “structured” than “non-structured” CDS regions ( P < 10−10 for both dN and dS comparisons , S31 Fig; totally 172 genes were included in the comparisons ) . Nevertheless , the length of the “structured” CDS regions in general only account for 4 . 36 ± 0 . 24% of the full lengths of CDS regions , and we still observed significantly lower dN and dS values in the “structure masked” CDS regions compared to the genes without editing events ( “Unedited” ) in Drosophila brains ( P < 10−10 for both dN and dS comparisons , S31 Fig ) . Hence , although the editing-associated secondary structures considerably constrain mRNA sequences , they generally have negligible impact on the evolutionary rates of the total CDS regions . In summary , our results support the hypothesis that A-to-I editing increases the proteomic diversity for the genes that are highly conserved due to functional constraints . It is worth noting that compared to humans [67 , 68] and macaques [59 , 60] , we observed strong signals of adaptation in the editing sites in Drosophila , especially in the brain editomes . Furthermore , the A-to-I editing events we identified in Drosophila were significantly enriched in evolutionarily conserved genes while the editing sites in human coding regions showed an opposite pattern [67] . The different observations between Drosophila and primates might be shaped by the difference in the underlying molecular mechanisms and selective forces . First , there are two catalytically active ADAR enzymes ( ADAR1 and ADAR2 ) in primates and both enzymes are expressed in many tissues , which potentially cause promiscuous editing events that might be neutral or deleterious [59 , 60] . Second , the targets of editing are mainly repetitive non-coding sequences in primates [13] . Although the coding editing events conserved between human and mouse are adaptive [68] , the majority of the editing events in coding regions might be solely by-products of the over-activity of ADARs and hence selected against [67] . In contrast , there is only one Adar locus in Drosophila [37 , 38] , which is predominately expressed in the nervous system and preferentially edits pre-mRNAs of neural genes [39] . Third , the effective population size is much larger for D . melanogaster than primates , which makes natural selection more efficient in the former than in the latter [100] . Therefore , the adaptive editing events , once originated , will be more effectively spread and fixed in Drosophila than in primates . On the other hand , the detrimental effects incurred by RNA editing , will be more efficiently selected against in D . melanogaster than in primates . Besides providing proteomic diversity , our results also suggest that mRNA editing interplayed with gene expression plasticity to fine-tune gene expression activity under temperature stress responses , which supports previous hypothesis that RNA editing might be a driving force for environmental adaptation [15 , 16] . Interestingly , we also found several editing events in D . melanogaster that compensated for the G-to-A DNA mutation in the D . melanogaster lineage after splitting with its sibling species ( S32 Fig ) , which suggests that RNA editing events are advantageous because they reverse the deleterious effects caused by G-to-A DNA mutations as previously proposed [67 , 101] . Taken together , our evolutionary analyses , combined with our functional genomic studies , shed new light on the molecular mechanisms and functional consequences of RNA editing in Drosophila .
Flies were grown in 12 hour light: 12 hour dark cycles at 25°C . The ISO-1 strain of D . melanogaster , the sim4 strain of D . simulans , and one lab strain of D . pseudoobscura were gifts from Dr . Andrew G . Clark’s lab at Cornell University . In the temperature stress experiments , 1- to 5-day-old flies were transferred from 25°C incubators to 30°C incubators and treated for 14 hours or 48 hours , and the humidity and light conditions were maintained at the same levels . The 1- to 5-day-old and 1- to 14-day-old flies were separately sexed , and the brains were dissected in RNAlater solution ( Ambion ) . We also separated the heads and bodies of the female or male adults of the sim4 strain of D . simulans with a fine sieve and extracted the total RNAs from heads and bodies of each gender . Total RNA was extracted using TRIzol reagent ( Invitrogen ) according to the manufacturer's protocol . Poly ( A ) + mRNA was isolated from 15 μg total RNA with oligodT25 DynaBeads ( Thermo Fisher ) . Next , the mRNA was fragmented and size selected from 40 nts to 80 nts by 15% TBU gels . Following 3' dephosphorylation , 3' ligation with a 3' adaptor , 5' phosphorylation and 5' ligation with a 5' adaptor , and size-selected mRNA fragments were reverse transcribed with SuperScript III ( Invitrogen ) . The sequence of the 5' adaptor was 5'GUUCAGAGUUCUACAGUCCGACGAUC3' and the 3' adaptor was 5'TGGAATTCTCGGGTGCCAAGG3' . All cDNA was amplified by 14 PCR cycles with Phusion High-Fidelity DNA polymerase ( NEB ) with the TruSeq index adapters , and the products within the correct size ranges were collected from 20% TBE gels for the quality tests ( Fragment Analyzer , Agilent Technologies ) and sequencing ( Platform: Illumina HiSeq 2500; read length: 50 bp , single-end ) . The 3' adaptor sequences were clipped by the Cutadapt program [102] . The remaining reads were aligned to the reference genome of D . melanogaster ( r6 . 04 ) , D . simulans ( r1 . 4 ) or D . pseudoobscura ( r3 . 2 ) using STAR [103] ( mapping statistics were summarized in S1 Table ) . The genome sequences and annotations of the three Drosophila species were downloaded from FlyBase ( www . flybase . org ) . First , for each brain library , we employed the GATK RNA-Seq variant calling pipeline [78] to detect the A-to-I editing events , where the transcriptomic base is Adenosine and the sequencing read variant base is Guanine ( Inosine ) . We were able to identify 1 , 531 editing sites at this preliminary stage . To retrieve the editing sites that were potentially excluded by the pretty strict GATK pipeline , we pooled all the editing sites in four previous studies [47–50] and the GATK candidates altogether , constructing a list of 5 , 925 candidate sites . Second , for each candidate site in a brain library k , we discarded the reads with mapping quality lower than 10 and the reads with mismatches other than A-to-G , and extracted the sequencing coverage ( Ck ) and the number of edited allele that shows A-to-G difference ( Lk ) with SAMtools [104] and calculate the probability that a site is edited in a library k given the observed sequencing data Pk ( E1 ) = 1 − Pk ( E0 ) , where Pk ( E0 ) is the probability the A-to-G difference is solely caused by sequencing error with a rate of ε . We define Pk ( E0 ) =∑i=LkCk ( Cki ) ∙εi∙ ( 1−ε ) Ck−i . The Illumina HiSeq platform generally has an error rate ε0 ranging from 0 . 2% to 0 . 6% [105–107] and we used ε0 = 0 . 5% in the analysis . As we focus on the sites with solely A-to-G but not A-to-C or A-to-T mismatches , the error rate of A-to-G would be scaled to ε=ε0/31− ( 2/3 ) ε0≈0 . 00167 . The probability that a site is edited in none of the n libraries is thus P ( E0 ) =∏k=1nPk ( E0 ) , where n is the number of brain libraries , and accordingly , the probability that this site is edited in at least one of these libraries is P ( E1 ) = 1 − P ( E0 ) . For each of the 5 , 261 sites that have sequencing coverage in our brain libraries of D . melanogaster , we calculated P ( E1 ) and P ( E0 ) , and corrected for multiple testing with Benjamini & Hochberg method [108] . The procedures are summarized in Fig 1A . The functional annotations of the editing sites were conducted with the software SnpEff [109] . The transcriptome data for heads of the Adar5G1 mutant and paired wild type w1118 of D . melanogaster were downloaded from NCBI SRA under accession numbers SRR629970 and SRR629969 [51] . We mapped the transcriptome data to the reference genome of D . melanogaster using STAR [103] and extracted the A and G alleles with SAMtools [104] . To estimate the false positive rates of the editing sites we detected , we first examined the number of the editing sites we identified in D . melanogaster that also had editing events detected in the heads of w1118 strain ( N1 ) , and then we counted the number of these shared sites that also showed A-to-G difference in the Adar5G1 mutant ( N2 ) . The false positive rate is estimated by N2/N1 as conducted previously [50 , 51] . We folded all the transcripts in genes expressed in brains of D . melanogaster with RNALfold [110] and identified the exonic editing sites that were located in the stable local hairpin structures ( z score < -1 . 5 , ΔG < -15 kcal/mol , and the stem length > 50 nts ) . To obtain the expected numbers of editing sites in the hairpin structures by randomness , we randomly sampled the same number of exonic editing sites under study with replacement for 1000 replicates and counted the median and 2 . 5% and 97 . 5% quantiles in the simulations . For each introinc editing site , we folded the flanking sequences for each site ( 100 nts at each side ) with RNALfold and examined whether this intronic site was located in the stable hairpin structures with the same criteria . For each of the editing events detected in D . melanogaster , the flanking sequences ( 100 nts at each side ) were folded In the long-range pseudoknot searches , the full-length pre-mRNA or flanking sequences ( 2 , 000 nts at each flank , totally 4 , 000 nts for long pre-mRNAs ) were folded with RNAfold [110] and putative long-range pseudoknots were parsed if the pairing region in each stem of the pseudoknot was > 40 nts and the distance between the two pairing stems was > 100 nts . We also fold the flanking sequences for each editing site ( 100 nts at each side ) and calculated the MFE ( kcal/mol ) with RNAfold [110] . The protein and CDS sequences of D . melanogaster ( r6 . 04 ) , D . simulans ( r1 . 4 ) and D . pseudoobscura ( r3 . 2 ) were downloaded from FlyBase . The reciprocal best orthologous genes were obtained based on pairwise BLASTP [111] between D . melanogaster and D . simulans ( r1 . 4 ) , and between D . melanogaster and D . pseudoobscura . The protein sequences of the orthologous genes were aligned with the clustalw [112] program , and the CDS alignments were produced with the tranalign [113] program based on the corresponding protein alignments . The yn00 program in the PAML [114] package was employed to calculate the dN and dS values for each gene between D . melanogaster and D . simulans . The phyloP score for each site of D . melanogaster were downloaded from UCSC Genome Browser ( genome . ucsc . edu ) . For each editing site in D . melanogaster , we employed two complementary approaches to search for the orthologous sites in D . simulans and D . pseudoobscura . First , we used liftOver [83] to convert the genomic coordinates of the orthologous sites in coding and non-coding regions between D . melanogaster and D . simulans , or between D . melanogaster and D . pseudoobscura as previously conducted [51] ( termed “g_align” approach ) . The pairwise genome alignments between D . melanogaster and D . simulans , and between D . melanogaster and D . pseudoobscura were downloaded from UCSC Genome Browser ( genome . ucsc . edu ) and used to identify the evolutionarily conserved adenosine sites . Second , we parsed the genomic co-ordinates with the pairwise CDS alignments between D . melanogaster and D . simulans , and between D . melanogaster and D . pseudoobscura ( termed “c_align” approach ) . The g_align approach efficiently identified both the coding and non-coding orthologous sites , and the c_align approach is powerful in identifying orthologous sites in coding regions between distantly-related species . Among the 2 , 114 high-confidence editing sites in brains of D . melanogaster , we identified 1 , 499 orthologous sites in D . simulans that were also adenosines and had sequencing coverage in at least one brain library of D . simulans ( 1 , 443 by g_align , 577 by c_align , and 521 by both ) , and 892 sites in D . pseudoobscura that were also adenosines and had sequencing coverage in at least one brain library of D . pseudoobscura ( 707 by g_align , 527 by c_align , and 342 by both ) . To exclude SNPs in the RNA editing characterization , we also deep sequenced the genomic DNA of the same strains of D . simulans ( the median coverage per site is 46 ) and D . pseudoobscura ( the median coverage per site is 47 ) that were used for RNA-editing detection . We mapped the DNA reads on the reference genomes with BWA[115] , excluded reads with mapping quality lower than 10 , and called the SNPs with SAMtools ( 313 , 133 SNPs in D . simulans and 489 , 828 SNPs in D . pseudoobscura ) . After masking the SNPs , for each site in each species , we calculated P ( E1 ) , the joint probability that this site is edited in brains . At FDR of 0 . 05 , we identified 996 sites edited in D . simulans ( 947 by g_align , 495 by c_align and 446 by both ) , and 451 sites edited in D . pseudoobscura ( 340 by g_align , 326 by c_align , and 215 by both ) . Our experimental designs ( brain samples of the same gender and the same age under the same accommodation conditions for different species ) and the combinations of g_align and c_align approaches enabled us to identify more evolutionarily conserved editing events compared to Yu et al . [50] which mainly focused on the conserved editing events in the coding regions . For an editing site with coverage Cm and editing level lm in a library m , no edited ( G ) allele would be detected if the edited RNA molecules were not sampled or all the edited signals were abolished by sequencing errors . Therefore , we estimated the probability of observing zero edited reads at this site in a sample m as Pm ( D0 ) =∑i=0Cm ( Cmi ) ∙lmi∙ ( 1−lm ) Cm−i∙εi∙ ( 1−ε ) Cm−i , where ε is the scaled sequencing error rate ( 0 . 00167 ) . Then the joint probability that this site is edited but not detected in n libraries would be P ( D0 ) =∏m=1nPm ( D0 ) . The expected N/S ratio for the editing events under neutral evolution was calculated with a similar procedure described previously [67] . Briefly , for each adenosine site in the CDS regions of the genes that have at least one editing event in brains of D . melanogaster , we tested whether it cause an amino acid change ( nonsynonymous , N ) or not ( synonymous , S ) when edited . To obtain the expected N/S ratios for the editing events that are evolutionarily conserved , we conducted similar analysis only on the conserved adenosine sites between D . melanogaster and D . simulans , or between D . melanogaster and D . pseudoobscura , as previously conducted between human and mouse [68] . The N/S ratio expected under neutral evolution is 3 . 80 for the editing sites in D . melanogaster , 4 . 18 for the editing sites with events conserved between D . melanogaster and D . simulans , and 7 . 11 for the editing sites with events conserved between D . melanogaster and D . pseudoobscura . We obtained the 95% confidence intervals by random sampling the same number of observed editing sites with replacement and calculated N/S ratios from the simulated data for 1000 replicates . The detailed procedures of the two methods in evaluating the effect of detection bias on N/S ratio are fully described in TEXT . The processes of these simulations ( and other simulations and statistical tests in this study ) were performed using R ( www . r-project . org ) . For each of the 2 , 114 high-confidence editing sites , we extracted the upstream and downstream 3 nucleotides flanking this site , counted the number of nucleotide at each position of the 7-mer , and developed a position probability matrix . Sequence logo for this motif was generated with WebLogo ( http://weblogo . berkeley . edu ) . In order to get a well-controlled set of genomic background sites , we scanned the mRNA regions for other adenosine sites in genes where the 2 , 114 editing events were detected . Then we scored both the 2 , 114 high-confidence editing sites and the background adenosine sites with the same position probability matrix . For a given sequence N-3N-2N-1AN1N2N3 ( Ni ∈ ( A , T , C , G ) ) , the score for this sequence was calculated as ∑ilog2Pi ( Ni ) 0 . 25 , where Pi ( Ni ) is the probability of observing base Ni at position i ( -3 , -2 , -1 , 1 , 2 , 3 ) based on the position probability matrix . We chose the score cutoff that specified the bottom 10% quantile of the high-confidence editing sites ( -0 . 622 ) . With that score cutoff , we chose the sites with scores above this cutoff ( -0 . 622 ) in the high-confidence editing sites and background adenosine sites , and calculated the N/S ratio in each dataset . The 1–14 day old female and male adults from five strains ( B12 , I17 , N10 , T07 and ZW155 ) of D . melanogaster were sexed , the poly-A tailed mRNAs were isolated from females and males independently with the procedures described above . The libraries were prepared with NEBNext® Ultra™ Directional RNA Library Prep Kit for Illumina and the sequencing was carried out at Illumina HiSeq 2500 ( read length: 100 bp , paired-end ) . These strains , kindly provided by Dr . Andrew G . Clark at Cornell University , were originally collected from five continents [79]: Beijing , China ( abbreviated B ) ; Ithaca , NY USA ( abbreviated I ) ; Netherlands , Europe ( abbreviated N ) ; Tasmania , Australia ( abbreviated T ) ; and Zimbabwe , Africa ( abbreviated Z ) . These flies were grown in 12 hour light:12 hour dark cycles at 25°C . The A-to-G editing sites in the exonic regions were called with the “joint probability” method as above described , and the SNPs in these five strains and other 79 related strains based on the whole-genome re-sequencing information [79] were masked in the down-stream analysis . Finally we identified 875 editing sites in female and 1 , 422 exonic sites in male adults . To identify the editing sites with putative polymorphic events across these five strains , we first filtered the sites that have editing events detected in brains of D . simulans or D . pseudoobscura , or whole files of D . simulans . And then we required a site to have editing reliably detected in at least one strain k with Pk ( E1 ) > 0 . 999 and to have no editing detected in at least one strain . Next we calculated Pm ( D0 ) , the probability that the editing was not detected at depth Cm due to sampling bias or sequencing error in a strain m as above described . Finally , we calculated the joint probability P ( D0 ) if no editing was observed in multiple strains at that site . Under Level I in calculating Pm ( D0 ) , we assumed the strain m that has no editing event detected have the same editing level at that site as in the other strains which had reliably editing event detected ( the mean value was used ) . Under Level II , we used 0 . 05 as the editing level in calculating Pm ( D0 ) if the mean editing level from the other strain is >0 . 05 . To detect the editing events that were fixed across these five strains , we employed two complementary approaches: First we identified the sites at which the probability of editing in each strain Pk ( E1 ) > 0 . 95 ( k was B12 , I17 , N10 , T07 and ZW155; we studied female and male adults separately ) ; and then we sequenced the transcriptomes of female and male adults of D . simulans ( the sim4 strain ) and required the orthologous sites to be edited in the same gender of D . simulans . To identify the SNPs that are associated with the variation of the editing levels in female or male adults across these strains , we only focused on the editing sites that are polymorphic in both female and males ( totally 58 sites ) or fixed in both females and males ( 171 sites ) . For each site in females , we retrieved the SNPs within 10kb regions that flanked the focal editing sites in the five strains and conducted the association test between the editing level in each strain and the genotypes of a SNP ( we assumed the reference allele as 0 and the alternative allele as 1 ) . We also performed the same analysis in male adults . We required the reference and alternative allele of a SNP to be associated with editing levels of the same site in the same direction in females and males and P < 0 . 05 in both tests . Total RNA from the female or male brains was prepared independently using Sanger sequencing procedures . The total RNA was treated with RNase-free DNase I ( Invitrogen ) to remove genomic DNA . Reverse transcription was performed using random primers , and cDNA was amplified using target-specific primers ( the primers sequences are presented in S36 Table ) . The final PCR products were sequenced with the Sanger method at the Ruibiotech Sequencing Company . qRT-PCR was performed with SYBR Green Master Mix ( Thermo Fisher ) in a 20 μL reaction volume and monitored on a StepOnePlus Real-Time PCR System ( Thermo Fisher ) . rp49 was used as the internal control . The primers for the real-time PCR assay are listed in S37 Table . The raw NGS reads for each gene were counted with the htseq-count program [116] , gene expression levels were normalized , and differentially expressed genes were detected with the edgeR package [117] . The RPKM for each gene or half-gene was calculated with CuffLinks [118] . All of the gene ontology analyses were performed by DAVID [119] and all the brain-expressed genes were used as background list . The sequence data in this study have been submitted to the NCBI Sequence Read Archive ( SRA , http://www . ncbi . nlm . nih . gov/sra ) under accession number SRP074828 and SRP068882 . All other relevant data are within the paper and SI files .
|
Adenosine-to-inosine ( A-to-I ) RNA editing is an evolutionarily conserved mechanism that alters RNA sequences at the co-transcriptional or post-transcriptional level . RNA editing is hypothesized to facilitate adaptation in that it expands the transcriptomic and proteomic diversity . However , evidence for adaptation of RNA editing at the whole editome level is still lacking . In this study we systematically identified A-to-I RNA editing sites in female and male brains of three Drosophila species at different temperatures . With evolutionary analysis from different perspectives , we provide lines of evidence to demonstrate that the nonsynonymous editing sites in Drosophila brains are generally adaptive . The signals of adaptation for the editing sites are significantly enriched in genes related to chemical and electrical neurotransmission . We show that the RNA editing events might interplay with gene expression plasticity in temperature stress responses . Furthermore , we demonstrated that the expression level of Adar , together with the expression profiles of a set of genes that have editing sites favored by natural selection , were important in shaping the overall selective patterns of the global editomes at different developmental stages ( or tissues ) of D . melanogaster . Altogether our results support the hypothesis that A-to-I editing provides a driving force for adaptive evolution in Drosophila from different aspects .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"invertebrates",
"glycosylamines",
"computational",
"biology",
"animals",
"invertebrate",
"genomics",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"genome",
"analysis",
"drosophila",
"adenosine",
"research",
"and",
"analysis",
"methods",
"genomic",
"libraries",
"gene",
"expression",
"evolutionary",
"genetics",
"insects",
"animal",
"genomics",
"arthropoda",
"biochemistry",
"rna",
"rna",
"editing",
"nucleic",
"acids",
"nucleosides",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"evolutionary",
"biology",
"glycobiology",
"organisms"
] |
2017
|
Adaptation of A-to-I RNA editing in Drosophila
|
Rift Valley fever ( RVF ) is a re-emerging zoonotic disease responsible for major losses in livestock production , with negative impact on the livelihoods of both commercial and resource-poor farmers in sub-Sahara African countries . The disease remains a threat in countries where its mosquito vector thrives . Outbreaks of RVF usually follow weather conditions which favour increase in mosquito populations . Such outbreaks are usually cyclical , occurring every 10–15 years . Recent outbreaks of the disease in South Africa have occurred unpredictably and with increased frequency . In 2008 , outbreaks were reported in Mpumalanga , Limpopo and Gauteng provinces , followed by 2009 outbreaks in KwaZulu-Natal , Mpumalanga and Northern Cape provinces and in 2010 in the Eastern Cape , Northern Cape , Western Cape , North West , Free State and Mpumalanga provinces . By August 2010 , 232 confirmed infections had been reported in humans , with 26 confirmed deaths . To investigate the evolutionary dynamics of RVF viruses ( RVFVs ) circulating in South Africa , we undertook complete genome sequence analysis of isolates from animals at discrete foci of the 2008–2010 outbreaks . The genome sequences of these viruses were compared with those of the viruses from earlier outbreaks in South Africa and in other countries . The data indicate that one 2009 and all the 2008 isolates from South Africa and Madagascar ( M49/08 ) cluster in Lineage C or Kenya-1 . The remaining of the 2009 and 2010 isolates cluster within Lineage H , except isolate M259_RSA_09 , which is a probable segment M reassortant . This information will be useful to agencies involved in the control and management of Rift Valley fever in South Africa and the neighbouring countries .
Rift Valley fever ( RVF ) , a mosquito-borne viral disease , affects humans and some species of ruminants including sheep , cattle , goats and buffalos . The causative agent , Rift Valley Fever virus ( RVFV ) , belongs to the genus Phlebovirus in the family Phenuiviridae [1] . The disease in livestock is characterized by abortion storms and high mortality of young animals [2] . In humans , it manifests as febrile illness , resulting in retinal degeneration , severe encephalitis , haemorrhage; fatal hepatitis occurs in less than 1% of patients [3] . Protection of animals from the disease can be conferred by vaccination; however , there is currently no approved vaccine for use in humans . Transmission of RVF can be prevented by eliminating the mosquito vectors and avoiding human contact with tissues of infected animals . The RVF virus has been isolated from more than 30 species of mosquitoes , belonging to at least six genera ( Aedes , Culex , Anopheles , Eretmapodites , Mansonia and Coquillettidia ) . In a number of mosquito species , the virus has been isolated from both the insects and their eggs , suggesting different modes of transmission [4 , 5] . Furthermore , the virus has been isolated from unfed mosquitoes reared from eggs obtained during inter-epidemic periods in Kenya and South Africa [4 , 5 , 6] . Eggs from floodwater mosquitoes can remain viable for considerable time between outbreaks , hatching during conducive climatic conditions associated with periods of high rainfall [4] . Changing climate and farming systems can create conditions favorable for mosquito breeding , resulting in unexpected outbreaks of the disease [7] . The disease is endemic in eastern and southern Africa , but outbreaks have been reported in Egypt , Madagascar , Mauritania , Saudi Arabia , Sudan and Yemen [8 , 9 , 10] . No significant antigenic differences have thus far been demonstrated among isolates of this virus from different geographic locations , confirming the existence of a single RVFV serotype . Despite the single serotype , differences in virulence and pathogenicity of the virus have been observed , necessitating the need for detailed genetic characterization of the various isolates [11 , 12] . The genome of RVFV , like that of the other bunyaviruses , consists of three-segmented , single-stranded negative- and ambi-sense RNAs with a total size of 12kb . The L ( Large ) segment codes for the viral RNA polymerase . The M ( Medium ) segment encodes a single precursor protein which is cleaved to produce the envelope glycoproteins G1 and G2 , and two non-structural proteins of 78kDa and 14 kDa . In contrast , the ambisense S ( Small ) segment codes for the nonstructural protein NSs in the genomic sense and the nucleocapsid protein N in the antigenomic sense [13 , 14 , 15] . Because of its segmented structure , the genome of RVFV is thought to undergo recombination through reassortment , thereby contributing to its evolutionary dynamics [12 , 16] . In general , the RVFV genome is characterized by low genetic diversity ( ~5% ) ; consequently , it is difficult to statistically detect intragenic recombination events [11 , 15] . Similar to other arboviruses , all the genes of RVFV are under purifying selection and have evolved at distinct rates by accumulating mutations at 1 . 9 x 10−4 to 2 . 5 x 10−4 substitutions per site per year [12 , 16] . The previously estimated time to most recent common ancestor ( TMRCA ) is at around 124 to 133 years . This coincides with the importation of highly susceptible European breeds of cattle and sheep into East Africa where the disease was first reported [16 , 17 , 18 , 19] . Despite the high percentage sequence identity , nucleotide sequences of complete or partial segments from viruses isolated from various countries over the last 60 years , have been grouped into 15 lineages [20] . Based on the topologies of phylogenetic trees constructed from nucleotide sequences representing each of the three genome segments , it is suggested that reassortment , specifically of a 2010 isolate from a patient in South Africa , contributed to antigenic shift during outbreaks . The individual was accidentally co-infected with live RVF animal vaccine and a RVF virus in lineage H [20] . In 2009 , a RVF outbreak with unusual clinical presentation in animals was observed in South Africa . This outbreak had two distinguishing features: first , it occurred atypically in the absence of abnormally high rainfall; secondly , in addition to causing abortion storms , it had a high mortality among pregnant adult cattle [21] . The first case of RVF in South Africa occurred in the summer of 1950–1951 in animals and subsequently it was diagnosed in humans in 1951 [22 , 23] . Three major outbreaks of the disease occurred in South Africa in 1950–1951 , 1974–1976 and , most recently , in 2008–2011 . There were minor incidents in the inter-episodic periods interspersing these outbreaks [24] . Confirmation of suspected cases of RVF in animals in South Africa is normally done at Agricultural Research Council–Onderstepoort Veterinary Research ( ARC-OVR ) . Over time , the institute has accumulated a large collection of RVFV isolates from a majority of reported cases of the disease in South Africa . In order to obtain comprehensive information on the genetic composition of the RVF viruses ( RVFVs ) circulating in South Africa , we performed full genome sequence analysis of some of the viruses isolated from animals at discrete foci of the outbreaks which occurred during the 2008–2010 period . The genome sequences of these viruses were compared with those of other RVFVs from earlier outbreaks in South Africa and other countries where the disease has occurred . Furthermore , the genome sequence data generated add to the repertoire of the RVFV sequences available in the public domain databases . Such data are necessary for studies required to find new or improved technologies for management of this zoonotic disease . Overall , data presented here add to the understanding of epidemiology and ecology of RVF . The information will be useful to agencies involved in the control and management of Rift Valley fever in South Africa and the neighbouring countries .
Organ biopsies or blood specimens from buffalo , cattle or sheep in 7 South African provinces were collected from sick or dead animals during the 2008 to 2010 RVF outbreaks . The specimens were brought to the Agricultural Research Council–Onderstepoort Veterinary Research ( ARC-OVR ) , to be subjected to confirmatory laboratory tests for RVF . The specimens were stored at 4°C for less than 24 hours before being processed for the tests , which include virus isolation . An outbreak operationally refers to a case of RVF confirmed in a laboratory by isolation of RVF virus , detection of RVF viral RNA or IgM antibody to RVF viral protein , in a tissue specimen from an animal found at a specific location during a season [24] . Multiple outbreaks in a season constitute an epidemic . Vero cells in T25 tissue culture flasks ( approximately 2 . 0×106 cells/ml ) were inoculated with 1/10 clarified suspensions of blood or homogenized tissue in DMEM supplemented with 100 IU/ml Penicillin , 100μg/ml Streptomycin and 0 . 25μg/ml Amphotericin B ( Lonza , BioWhittaker ) , and 2% foetal bovine serum ( Gibco , LifeTechnologies ) . The cells were incubated at 37°C with 5% CO2 in a humid chamber for 1 hour . Thereafter the inoculum was discarded , the cell monolayers washed twice with media , 5ml of the same medium replenished and the flasks returned in the incubator , where they were monitored daily for development of cytopathic effect ( CPE ) . Incubation lasted 4–6 days , and up to three passages were made per sample [25] . Presence of RVFV nucleic acids in the isolates was confirmed by real-time PCR using a slight modification of an established method [26] . The isolates were aliquoted in 500μl– 1ml quantities and stored at -80°C until further use . Optimum conditions for efficient infection of Baby Hamster Kidney ( BHK 21 ) cells ( obtained from AATC ) with RVF virus were established empirically using isolate M35/74 , the challenge strain of RVFV [27] . The BHK 21 cells were grown in DMEM-F12 supplemented with 5%FBS ( LONZA ) and 1% pen/strep Amphotericin B ( LONZA ) . These conditions were applied to infect BHK cells at a MOI resulting in the highest viral load . The infected cells were pelleted by centrifugation at 2 500 rpm for 5 minutes and the supernatant recovered . The supernatant was committed to sequence independent single primer amplification ( SISPA ) [28] . Briefly , the viral particles in the supernatant were treated with 100U DNase I and 4g RNase at 37°C for 2h to remove possible host nucleic acids contamination . Viral RNA was extracted using TRIZOL LS kit ( Invitrogen ) according to the procedure provided by the supplier ( Invitrogen ) . The RNA was recovered and used as the template in the first strand cDNA synthesis primed with FR26RV-N ( 5’GCC GGA GCT CTG CAG ATA TCN NNN NN3’ [28] . The single-stranded cDNA was the template for double-stranded cDNA synthesis using random 20mer primers and Klenow fragment of E . coli DNA polymerase . These products were subjected to PCR amplification using the 20-mer region of the above primer ( FR26RV: 5’GCC GGA GCT CTG CAG ATA TC3’ ) in a reaction incubated in a thermocycler programmed to denature at 94°C , 2 min then 35 cycles of 94°C , 30 sec; 55°C , 30 sec; 68°C , 30s; with a final extension at 68°C , for 10 min . The SISPA products were resolved by electrophoresis in 1% agarose gels . The SISPA products ranging in size from 0 . 2kb to 1 . 5kb were recovered from agarose gels and used in the preparation of library for sequencing reactions on the Next Generation Sequencing ( NGS ) platforms exactly as described by the manufacturer ( Roche Applied Science ) . The sequencing was done on the Genome Sequencer 454 platform ( GSFLX; 454 Life Sciences , Roche Applied Science; http://www . 454 . com ) . The sequence data obtained was processed and assembled into contigs using the appropriate software set to default values ( Roche/454 Newbler for 454 Life Sciences Corporation , Software Release: 2 . 8–20120726_1306 or CLC Genomics Workbench , QIAGEN Bioinformatics ) . The data was subjected to further analyses using a combination of bioinformatics software . The nucleotide sequences were aligned using Clustal W [29] within the Molecular Evolutionary Genetics Analysis ( MEGA version 6 ) [30] set to optimum parameters for each sequence type . The best fitting nucleotide substitution model was determined for each genome segment using MEGA 6 and then applied in all the subsequent analyses . The aligned nucleotide sequences were used in calculating the mean pairwise distances and to derive phylogenetic trees using Maximum likelihood under 1000 bootstrap iterations [31] . Evidence for possible intragenic recombination events among the isolates was sought using different methods available from RDP3 [32] . Rates of molecular evolution for individual genome segments were estimated using Bayesian Markov Chain Monte Carlo ( MCMC ) implemented in the BEAUTI v1 . 8 . 1 , BEAST v1 . 8 . 1 , Tracer and FigTree packages [33] . The substitution rates were estimated using both strict and relaxed uncorrelated lognormal molecular clock under General Time Reversible ( GTR ) model with gamma distribution ( T4 ) . The general Bayesian skyline coalescent prior was used and the MCMC allowed to run for sufficient number of generations ( > 10 million ) with sampling every 1000 states , to ensure convergence of all parameters [33] . The nucleotide sequences of all the segments of the RVF isolates analyzed in the current study have been deposited in GenBank with accession numbers indicated in Table 1 .
Sequence alignments were generated for each of the three segments using all the available RVFV sequence data in GenBank . The alignments , which included full genome sequences of 120–140 virus isolates depending on the segment , were used in evaluating the evolutionary dynamics acting on each of the three segments . Generally , sequence diversity among the segments were <5% among S or L segments , and <6% among M segments . Bayesian coalescent estimations of RVF genomes indicated that the segments evolve at a mean rate between 3 . 9 x10-4 and 4 . 17 x10-4 substitutions per site per year , regardless of the molecular model used . This is in agreement with previous Bayesian estimations [12 , 20] . Similarly , the estimated Time to Most Recent Common Ancestor ( TMRCA ) supports previous estimations of between 1880 and 1890 [12] . In order to determine the influence of substitution rate on biological function , we estimated the effect of differential selection pressures by calculating the rate of non-synonymous ( dN ) to synonymous ( dS ) substitutions . All the coding regions were found to be under purifying selection pressure ( dN/dS <1 ) . Using Maximum likelihood trees , the phylogenetic relationship of the 23 RVFV isolates was assessed in relation to 50 other isolates the genome sequences of which were already in GenBank ( S1 Table ) . Incongruences among the tree topologies of the individual genome segments were observed ( Fig 2A–2C ) , prompting an investigation of possible influence of recombination and reassortment . We found no evidence for intragenic recombination among any of the three segments . This is probably due to the low genetic diversity among the sequences [12 , 30] . Reassortment has been described for RVFV [12 , 16] and was therefore investigated utilizing the current data . Previous work using nucleotide sequences of the three segments assigned 15 lineages to the available RVFVs [20] . Our data indicate that one of the 2009 and all the 2008 isolates from South Africa and Madagascar ( M49/08 ) clustered in Lineage C or Kenya-1 ( Fig 2A , 2B and 2C ) . The remaining 2009 and 2010 isolates clustered within Lineage H [20] , except isolate M259_RSA_09 . The latter originated from serum of a bovine in Upington , Northern Cape Province , South Africa . Both its segments L and S cluster in Lineage K together with that of isolate JQ068143 from Kakamas ( also in the Northern Cape Province ) ; however , its segment M clustered in Lineage H along with the rest of the 2009–2010 isolates ( Fig 2B ) . This indicates that isolate M259_RSA_09 is probably a segment reassortant from a coinfection with RVFVs in Lineages H and K . Whether this event occurred in an insect vector or an animal host is not clear . Segments L and S of isolate M1955_RSA_55 cluster in Lineage I together with the 1951 South African Van-Wyck isolate ( DQ380158 ) ( Fig 2A and 2C ) . However , segment M of this isolate ( M1955_RSA_55 ) places it in Lineage L with the isolates of the 1974 and 1975 outbreaks in South Africa ( Fig 2B ) . Thus , isolate M1955_RSA_55 was the second RVFV in this study , which had sequence features suggesting that it may be a segment reassortant . Since both these putative reassortment events relate only to Segment M , which encodes two glycoproteins ( Gc and Gn ) , the segment was subjected to additional analysis . The amino acid sequences of the glycoproteins encoded by the M segment of different RVF virus isolates from the 2008–2010 outbreak were compared to those previously published ( S1 Table ) . The predicted amino acid residues are conserved with <3% differences in the percentage sequence identity . Of the amino acid changes , 55% are conservative , 9 . 8% result in loss of a charge , 17% in gain of a charge and 2 . 7% in change of a charge . The positions of amino acid substitutions relative to the proportion of sequences with that change and those resulting in a change of charge are shown in Fig 3 . Even though the majority of the substitutions are at the C-terminal region of the glycoprotein Gn , they are only observed in a small number of the sequences and the majority of them were conservative . One exception was a change from D ( Aspartic acid ) to N ( Asparagine ) at the amino acid position 95 , which is prominent in the 2008–2009 isolates in Lineage C , Fig 2B . To investigate the possible influence of the amino acid changes on the antigenic properties of the viruses , we performed antigenicity predictions using Welling [36] , with a window size of 11 . Antigenicity plots for isolate M33_RSA_10 , M37_RSA_08 and T1 ZH-501 isolated in Egypt in 1977 are shown in Fig 3 . These isolates represent Lineages H , C and A , respectively . Virus ZH501 from the 1977 outbreak in Egypt has been shown to be associated with increased virulence in rats [37] . Although this observation in rodents may not necessarily hold for other susceptible animal species , this isolate was included in the analysis here for comparison with isolates from Lineages H and C [20] . The major differences in antigenicity predictions between ZH501-77 and the South African isolates are at positions 60 and 631 ( Fig 3 ) . Isolate ZH501-77 has a valine at both of these positions in contrast to the South African isolates which have isoleucine . Association of virulence with amino acid substitutions at positions 595 , 631 , 659 and 1059 has been shown in previous studies [37] .
South Africa has experienced three major periods of RVF outbreaks , the first in 1950–1951 , the largest in 1974–1976 and lastly in 2008–2011 . In 2008 a total of 15 outbreaks were reported , localized to the central provinces of Limpopo , Mpumalanga , North West and Gauteng [34] . Complete genome sequence analysis of viruses isolated from the 2008 outbreak clusters them in Lineage C together with isolates from a 2007 outbreak in Kenya , known as Lineage Kenya-1 , Fig 2A , Fig 2B and Fig 2C [12 , 20] . In contrast to the localized outbreak of 2008 , 19 outbreaks which were reported in 2009 were widespread , with single cases in Mpumalanga and Gauteng , and the rest in KwaZulu-Natal , the Eastern Cape or along the Orange River in the Northern Cape [34 , 38] . Similar to the isolates from the 2008 outbreaks , the 2009 isolate from Gauteng clustered in Lineage C , within Lineage Kenya-1 ( Fig 2A and 2B ) . This 2009 Gauteng outbreak appears to have been caused by the 2008 RVF viruses present in that region . Isolates from both the geographically distinct KwaZulu-Natal and Northern Cape outbreaks of 2009 , clustered in Lineage H ( Fig 2A ) . This lineage includes isolate M259_RSA_09 , a segment M reassortant , whose segments L and S cluster with the 2009 Kakamas isolate ( JQ068142 ) in Lineage K ( Fig 2B ) . It is therefore possible that the RVF viruses associated with the majority of the outbreaks in 2009 originated from a single source . In 2010 a total of 484 outbreaks were reported . These were found in every province of South Africa , except KwaZulu-Natal [24] . Initial reports of outbreaks during this time were from Bultfontein and Brandfort , both in the Free State; subsequently , cases were reported from across the country [34] . Similar to the virus isolates from the 2009 outbreaks in KwaZulu-Natal and the Northern Cape , all the isolates from the 2010 outbreaks clustered in Lineage H ( Fig 2A and 2B ) . The eight isolates from the 2010 outbreak analyzed in this study ( Table 1 and Fig 2 ) are not necessarily statistically representatives of the 14342 cases reported in that year , but analyses of their nucleotide sequence data support speculation that the 2010 outbreak was a continuation of the 2009 in KZN and Northern Cape outbreaks . The clustering of isolates in lineage H ( Fig 2A ) gives an indication that new strains do evolve due to nucleotide substitution ( Fig 3 ) , albeit at slow/low rate . It is possible that these viruses were not introduced from elsewhere outside South Africa , but rather that they mutated over time and caused outbreaks when suitable conditions prevailed . This study has contributed full genome sequence of RVFVs M57_RSA_74 isolated during the 1974 outbreaks and M1955_RSA_55 isolated from one of the 28 outbreaks in 1955 [34] . The largest RVF epidemic reported in South Africa was between 1973 and 1976 , with mortality rates of 95% and cases reported from every province [34 , 39] . Previous studies have clustered the 1973–1975 isolates into Lineage L along with a 1970 isolate from Zimbabwe and a 1956 isolate from Kenya [12 , 20]; as expected isolate M57_RSA_74 clustered with these ( Fig 2A , 2B and 2C ) . In contrast , Segment S and Segment L of isolate M1955_RSA_55 clustered with a 1951 South African isolate known as van-Wyck in Lineage I ( Fig 2A and 2C ) [16] , but Segment M clustered with isolates from the 1973–1976 outbreaks in Lineage L ( Fig 2B ) , making this 1955 isolate a segment M reassortant . The occurrence of segment M reassortment in M1955_RSA_55 indicates that multiple RVF virus lineages can co-circulate , leading to reassortant viruses re-emerging decades later causing disease outbreaks . The evolutionary dynamics of RVFVs are characterized by low substitutions rates ( 3 . 9 x10-4 and 4 . 17 x10-4 substitutions per site per year ) under strong purifying or negative selection with the major genomic diversity resulting from reassortment [12 , 15 , 35] . Similar evolutionary dynamics have been described in other arboviruses such as bluetongue virus and Epizootic haemorrhagic disease virus , due to the obligatory replication of the virus in both its insect vector and mammalian host [40] . The majority of reasortment events described in RVFV involve the exchange of segment M , resulting in antigenic shift due to the two glycoproteins Gn and Gc encoded by this segment [12 , 16 , 20] . Although RVF virus is antigenically homogenous , some isolates of the virus differ in the severity of disease they cause in the mammalian host [37 , 40] . Whereas some of these differences may be attributable to the individual host , others are inherent to the virus . Differences in virulence and lethality of RVF virus isolates have been observed during the experimental infection of BHK cells [41] , mice [37] , sheep [42] and cattle [41] . Significant differences associated with the severity of RVF in humans have also been observed [43 , 44] . An increase in the severity of RVF since the 1977 outbreak in Egypt to the devastating outbreak during the 2006–2008 in East Africa have been reported [45] . Although a definitive association between genotype and lethal phenotype has not been established for RVFV , various amino acid substitutions have been implicated in this phenotype [41] . The most prominent substitution in the glycoproteins are 595 I>V , 605 R>K , 631 I>V , 659 V>A located in Gn and 1059 S>T within Gc [12] . Another variation was identified in ZH501 , isolate from a human in Egypt during an outbreak in 1977 , which resulted in the change of Glycine to Glutamic acid at position 277 . The virus with the Glutamic acid displayed an increased virulence in mice , compared to the virus with Glycine in the same position [37] . The majority of RVF viruses analysed in this study had Glutamine at position 277 , except wild type isolate 763/70 from a foetus aborted during an outbreak of the disease in Zimbabwe in 1970 [12] . This study identified additional substitutions between the lethal isolate ZH501-77 from Egypt and isolates belonging to Lineage H from 2010 in South Africa ( Fig 3B ) . The substitutions included 602 V>I , 987 D>E and 1131 T>I . The possible influence of each of these substitutions on the pathogenicity of RVFVs remain to be investigated . One caveat with the dataset analyzed here is that the isolates might not be representative of the RVF viruses circulating during the 2008–2010 outbreak . This is inherent in the way the study was done: samples brought for testing at the ARC-OVR are opportunistic and are not necessarily representative of cases of RVF in animals in South Africa . During this period , the RVFVs whose genomes could be analyzed were the viruses that could infect BHK 21 cells growing in culture media; and finally , good quality sequence data could not be obtained from all RVFVs , which were isolated in cell culture . A different picture of RVF viruses and their potential quasispecies might emerge when genome analyses are performed on viruses obtained directly from representative proven clinical cases . This is currently not possible in our system , but determining the entire RVF viral genome sequence directly from clinical samples is being investigated .
|
A single RVF virus serotype exists , yet differences in virulence and pathogenicity of the virus have been observed . This necessitates the need for detailed genetic characterization of various isolates of the virus . Some of the RVF virus isolates that caused the 2008–2010 disease outbreaks in South Africa were most probably reassortants resulting from exchange of portions of the genome , particularly those of segment M . Although clear association between RVFV genotype and phenotype has not been established , various amino acid substitutions have been implicated in the phenotype . Viruses with amino acid substitutions from glycine to glutamic acid at position 277 of segment M have been shown to be more virulent in mice in comparison to viruses with glycine at the same position . Phylogenetic analysis carried out in this study indicated that the viruses responsible for the 2008–2010 RVF outbreaks in South Africa were not introduced from outside the country , but mutated over time and caused the outbreaks when environmental conditions became favourable .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"rift",
"valley",
"fever",
"virus",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"tropical",
"diseases",
"microbiology",
"geographical",
"locations",
"viruses",
"rift",
"valley",
"fever",
"south",
"africa",
"rna",
"viruses",
"neglected",
"tropical",
"diseases",
"bunyaviruses",
"veterinary",
"science",
"africa",
"research",
"and",
"analysis",
"methods",
"microbial",
"genomics",
"sequence",
"analysis",
"infectious",
"diseases",
"viral",
"genomics",
"veterinary",
"diseases",
"zoonoses",
"bioinformatics",
"medical",
"microbiology",
"epidemiology",
"microbial",
"pathogens",
"people",
"and",
"places",
"dna",
"sequence",
"analysis",
"virology",
"viral",
"pathogens",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"genomics",
"organisms"
] |
2019
|
A comparative genome analysis of Rift Valley Fever virus isolates from foci of the disease outbreak in South Africa in 2008-2010
|
Microbial tropism , the infection of specific cells and tissues by a microorganism , is a fundamental aspect of host-microbe interactions . The intracellular bacteria Wolbachia have a peculiar tropism for the stem cell niches in the Drosophila ovary , the microenvironments that support the cells producing the eggs . The molecular underpinnings of Wolbachia stem cell niche tropism are unknown . We have previously shown that the patterns of tropism in the ovary show a high degree of conservation across the Wolbachia lineage , with closely related Wolbachia strains usually displaying the same pattern of stem cell niche tropism . It has also been shown that tropism to these structures in the ovary facilitates both vertical and horizontal transmission , providing a strong selective pressure towards evolutionary conservation of tropism . Here we show great disparity in the evolutionary conservation and underlying mechanisms of stem cell niche tropism between male and female gonads . In contrast to females , niche tropism in the male testis is not pervasive , present in only 45% of niches analyzed . The patterns of niche tropism in the testis are not evolutionarily maintained across the Wolbachia lineage , unlike what was shown in the females . Furthermore , hub tropism does not correlate with cytoplasmic incompatibility , a Wolbachia-driven phenotype imprinted during spermatogenesis . Towards identifying the molecular mechanism of hub tropism , we performed hybrid analyses of Wolbachia strains in non-native hosts . These results indicate that both Wolbachia and host derived factors play a role in the targeting of the stem cell niche in the testis . Surprisingly , even closely related Wolbachia strains in Drosophila melanogaster , derived from a single ancestor only 8 , 000 years ago , have significantly different tropisms to the hub , highlighting that stem cell niche tropism is rapidly diverging in males . These findings provide a powerful system to investigate the mechanisms and evolution of microbial tissue tropism .
The evolutionary interests of males and females are frequently divergent . Sexual conflict arises when phenotypes that enhance the reproductive success of one sex reduces the fitness of the other sex [1] . A well-characterized example in Drosophila is sperm competition between males . Sperm competition results in rapid evolution of sperm proteins which up-regulate females' egg-laying rate and reduces her desire to re-mate with another male . However , these proteins also shorten the female's lifespan reducing her fitness [reviewed by 2] . Vertically transmitted reproductive parasites , such as Wolbachia , can also cause sexually divergent phenotypes in males and females . Wolbachia are obligate intracellular bacteria present in a large fraction of insects , as well as spiders , mites , crustaceans , and filarial worms . They are primarily vertically transmitted from mother to offspring in a manner analogous to mitochondrial inheritance , although there is extensive evidence of horizontal transmission in nature [3] , [4] . For intracellular bacteria , vertical transmission often favors infected female hosts , which is also the case for Wolbachia [5] . There are several Wolbachia-induced phenotypes favoring the infected female , including parthenogenesis , feminization , male killing , and cytoplasmic incompatibility [6] . Each of these phenotypes ultimately results in the spread of more infected female hosts . In such cases , maternally transmitted bacteria can act as selfish genetic elements driving sexual conflict [5] . For successful vertical transmission , Wolbachia need to be present in the eggs laid by infected females . It has been shown in Drosophila that Wolbachia display a strong tropism for the germline , in particular , the oocyte , to ensure a high percentage of vertical transmission [7]–[10] . Although vertical transmission is prevalent , Wolbachia also can spread horizontally across individuals and species [3] , [11] , [12] . Colonization of the germline is a prerequisite for the infection to become successfully established into a population . We have previously shown that upon recent infection , Wolbachia colonize the stem cell niches in the Drosophila ovary , favoring vertical transmission after horizontal transfer [13] . Furthermore , stem cell niche tropism in the ovary is a highly evolutionarily conserved phenotype across the Drosophila genus , present in 100% of ovaries analyzed [14] . Wolbachia also infect the putative stem cell niches in the ovaries of other species , such as the bedbug and leafhopper [15] , [16] indicating that the selective pressure for Wolbachia targeting of ovarian stem cell niches to favor transmission extends beyond the Drosophila genus . Wolbachia have also been shown to display tropism to the stem cell niche present in the testis in Drosophila mauritiana [17] . However , the conservation of this phenotype across the Drosophila genus is unknown . Here we show that the evolutionary conservation of stem cell niche tropism present in females is not maintained in the male lineage . In fact , Wolbachia niche tropism in the testis , compared to the female results , represents a pronounced sexual dimorphism in the evolutionary history of Wolbachia stem cell niche tropism . Furthermore , we′ve identified that both Wolbachia and host factors modulate hub tropism in this system . Finally , we show that very closely related Wolbachia strains infecting the same host differ significantly in the densities at which they colonize the hub , indicating that hub tropism is a rapidly diverging phenotype in males .
In the testis , the germline stem cells ( GSCs ) and cyst stem cells ( CySCs ) reside at the “hub” , a structure at the apical tip of the testis ( Fig . 1A ) . The hub is a group of 10 to 16 somatically derived cells forming the microenvironment supporting the stem cells , referred to as the niche [18] . It has been shown that the GSCs receive maintenance signals from both the hub and the CySCs , hence both are considered to be part of the stem cell niche for the GSCs . However for the context of this study , niche tropism in the testis refers to Wolbachia infection of the hub only . To investigate whether Wolbachia niche tropism is as pervasive in the hub , as previously shown in the ovary [14] , we surveyed various Drosophila species infected with different strains of Wolbachia ( Fig . 1; S1 Dataset; see S1 Table for the sources for the stocks used in this analysis ) . Using confocal imaging and immunohistochemistry , we analyzed the density of Wolbachia infection in the hub cells as compared to the density of Wolbachia in the surrounding tissue ( see Materials and Methods ) . We found that Wolbachia target the hub at varying frequencies and densities across the Drosophila genus ( Fig . 1 , S2 Table , S1 Dataset ) . 3 out of 9 species showed very little to no Wolbachia infection in the hub ( Fig . 1 H–J , quantification in K ) , indicating that hub tropism is not pervasive across the Drosophila genus . 6 out of 9 species analyzed , however , did have Wolbachia tropism to the hub , ranging from 17% of niches infected to 95% of niches infected ( Fig . 1 B–G , K , see also Materials and Methods ) . The 6 Drosophila species- Wolbachia strain pairs with hub tropism fall into two groups with significantly different frequencies and densities of tropism: 3 had very high frequencies and densities of hub infection: D . ananassae wAna , D . melanogaster wMel , and D . mauritiana wMau; and 3 had moderate frequencies of Wolbachia tropism to the hub: D . yakuba , wYak , D . tropicalis wWil , and D . simulans wRi . In the ovary , tropism to the somatic stem cell niche is found at high frequencies in every individual of all Drosophila species analyzed [14] . In contrast , tropism for the hub is found in only a fraction of the species analyzed . Similar to the results for hub tropism , the frequency of tropism to the germline stem cell niche ( GSCN ) in the ovary was shown to be variable across the Drosophila genus ( Fig . 2A and [14] ) . We reasoned that Wolbachia tropism to the hub in the testis could simply be a byproduct of GSCN targeting in the ovaries . Interestingly , however , the presence of hub tropism does not correlate with the presence GSCN tropism ( S3 Table , Correlation Test , p = 0 . 773 ) . Although tropism in males and females is correlated in some strains ( 5 out of 9 , e . g . wMau displays high frequencies of both hub tropism and GSCN tropism and wSh does not have tropism to either the hub or the GSCN ) , there are also others that do not ( 4 out of 9 ) . The Wolbachia strain displaying one of the highest frequencies of GSCN tropism in the ovary ( wNo , 99% [14] ) , displays no tropism to the hub ( 0% , Fig . 1 I and K ) . Conversely , a Wolbachia strain displaying a high frequency of tropism to the hub ( wMel , 71% , Fig . 1 C and K ) does not target the GSCN in the ovary ( 1% , [14] ) . These data reveal that Wolbachia stem cell niche tropism does not correlate with GSCN tropism in the female . Previously , we have shown that the pattern of GSCN tropism is evolutionarily conserved across the Wolbachia lineage ( [14] and Fig . 2 ) . To assess whether hub tropism was also conserved across the Wolbachia lineage , we aligned the frequencies of hub tropism on the Wolbachia phylogenetic tree ( Fig . 2 ) . We quantified the correlation of hub tropism pattern with the Wolbachia phylogeny using a computer simulated model of randomized character distributions to compare with the distribution of niche tropism pattern on each of the phylogenies , as previously described [14] . We found that it is highly probable that the distribution of hub tropism is completely independent of the Wolbachia phylogeny ( S2 Fig . ) . Similarly , when we compared hub tropism to the Drosophila phylogeny , we found no clear correlation between the two ( S3 Fig . ) . Quantification of the relationship revealed that frequency of hub tropism bears no correlation with the Drosophila phylogeny ( S4 Fig . ) . An important Wolbachia related phenotype that also bears no correlation with host or microbial phylogenies is cytoplasmic incompatibility ( CI ) . CI is a reproductive phenotype resulting in reduced embryo hatching when a Wolbachia infected male mates with an uninfected female . We examined the possibility of a correlation between tropism to the hub and CI by comparing our tropism data to previously published reports on the levels of CI across the Drosophila genus ( S4 Table ) [19]–[23] . This analysis shows that some species with high levels of CI have different levels of tropism ( i . e . wSh and wRi have 0% and 17% hub tropism , respectively ) . Conversely , some species with low levels of CI also have a wide range of hub tropism phenotypes ( i . e . wTei and wMau have 2 . 3% and 71% hub tropism frequencies , respectively ) . Although hub tropism is highly divergent even amongst closely related strains of Wolbachia , similar to CI , there does not seem to be a correlation between these two phenotypes ( S4 Table , Correlation test , p = 0 . 267 ) . We next aimed to elucidate if host or bacterial factors influence the highly dynamic nature of the hub tropism phenotype . To investigate this question , Wolbachia strains backcrossed into a different host were used to assess Wolbachia strain versus host background influence on hub tropism , as previously described [14] . D . mauritiana wMau , which displays hub tropism ( Fig . 1D and Fig . 3 ) and D . sechellia wSh , which does not display hub tropism ( Fig . 1J and Fig . 3 ) and their hybrid offspring were utilized in this study ( See Material and Methods ) . Wolbachia strain wSh , infecting its native host , D . sechellia , and its non-native host , D . mauritiana , displays no hub localization , regardless of host genetic background ( Fig . 3 , S5 Table ) . This result suggests that Wolbachia wSh is incapable of hub tropism in either species . However it does not rule out the possibility that the hosts share a mechanism for excluding wSh from the hub . Therefore , a lack of tropism in both hosts cannot provide insight into whether the host or microbe is providing factors contributing to hub tropism . The analysis of wMau hub tropism allows further probing into this question . Wolbachia strain wMau infecting its native host , D . mauritiana , and its non-native host , D . sechellia , displays tropism for the hub , suggesting that the Wolbachia strain is driving this phenotype . However , the frequency of targeting in the hybrid host is 3-fold lower than in the native host ( Fig . 3C , green bars ) . Statistical analysis of frequency data indicates that both host genetic background and Wolbachia strain can significantly affect the frequency of hub tropism ( Fisher's exact test , p = 8 . 309×10−5 and p = 2 . 267×10−10 , respectively ) . These results are in contrast to previous data in the ovaries where only the Wolbachia strain drives tropism . wMau can efficiently target the GSCN in the ovary of both its native and hybrid host , greater than 80% of niches infected , regardless of the host genetic background [14] . The wMau frequency data in the male support the hypothesis that the Wolbachia strain is directing hub tropism . However , because the frequency of targeting is not as robust in the hybrid host compared to its native host , a role for the host is also implicated . In relation to Wolbachia density in the hub , the data indicate that the Wolbachia encoded factors play a major role in both native and hybrid hosts . The overall density at which wMau infect the hub is conserved ( Fig . 3 B and C , native host solid green bar , hybrid host hatched green bar , S4 Table ) . Similarly , wSh hub titers , compared to the surrounding tissue , are less than 1 in both native and hybrid hosts ( Fig . 3 B and C , native host solid red bar and hybrid host hatched red bar , S4 Table ) . Linear regression analysis of density data indicates that the Wolbachia strain , rather than the host genetic background , modulates Wolbachia density in the hub ( P = 0 . 045 and P = 0 . 56 , respectively ) . With respect to both frequency and density , the overall data reveal that factors encoded by both the host species and the Wolbachia strain influence hub tropism in the Drosophila testis . To further investigate the role of Wolbachia on hub tropism , we then analyzed different Wolbachia strains in the same host species . We took advantage of D . simulans , which is a host to many different Wolbachia strains . We investigated two strains of D . simulans flies differentially infected with wRi and wNo and their backcrossed offspring . Flies were backcrossed to account for any genomic divergence between host strains , as previously described [14] . D . simulans flies infected with Wolbachia wRi display hub tropism in about 33% and 43% of hubs analyzed for the parental and backcrossed hosts , respectively ( Fig . 4 , S6 Table ) . D . simulans wNo displays hub tropism infrequently ( 2% and 15% of hubs highly infected for the parental and backcrossed hosts , respectively , Fig . 4 , S6 Table ) . Although the frequencies of hub tropism for each Wolbachia strain increase in the backcrossed hosts , the general trend remains , where wRi targets the hub at a higher frequency than wNo . To quantify the relative contributions of host and bacterial factors towards hub tropism , logistical regression was performed . Wolbachia factors have a significant effect on hub tropism as compared to no significance of the host genetic background in the D . simulans hybrid flies ( p = 0 . 0000552 and p = 0 . 927 respectively ) . These results indicate that when host factors are kept constant , Wolbachia strain factors are sufficient to significantly modulate the frequency of hub tropism . In the previous analyses of hybrid crosses , hub tropism of distantly related Wolbachia strains were compared , first with different host species ( Fig . 3 ) , then within the same host species ( Fig . 4 ) . These results indicate that although the fly host can play a role in hub tropism , Wolbachia can significantly affect tropism on its own . In both cases , we were comparing Wolbachia strains from the A and B supergroups . We next investigated if the observed diversity of niche tropism is still present between more closely related Wolbachia strains . To address this question , we analyzed hub tropism of several Wolbachia strain variants infecting Drosophila melanogaster that diverged from a single ancestor within the last 8 , 000 years [24] , [25] . Hub tropism of wMel-like ( wMel , wMel2 , and wMel3 ) and wMelCS-like ( wMelCS , wMelCS2 , and wMelPop ) Wolbachia strains were analyzed . These Wolbachia strains were introgressed into the same D . melanogaster genetic background with the same microbiota [25] . The data reveal that the three wMel-like Wolbachia strains have significantly different tropism phenotypes from the wMelCS-like strains ( Fig . 5 , S7 Table ) . The wMel-like strains target the hub at similar frequencies , between 25% and 50% , and at similar densities , about 1 . 5-fold higher than the surrounding tissue . The wMelCS-like strains target the hub at significantly higher frequencies ( P<0 . 05 ) and densities ( P<0 . 001 ) than the wMel-like strains . Within the wMelCS-like group , wMelPop targets the hub at a significantly higher frequency ( 100% ) than wMelCS2 ( 77%; P = 0 . 005 ) , but not wMelCS ( 90% ) . However , wMelPop targets at a significantly higher density than both wMelCS and wMelCS2 ( P<0 . 0001; S1 Movie ) . Interestingly , wMelPop densities increase to the point where the hub cells burst open in approximately 20% of hubs ( S5 Fig . and S2 Movie ) . The finding that the wMel-like and wMelCS-like Wolbachia variants , all derived from a single ancestor only 8 , 000 years ago , have significantly different frequencies and densities of targeting indicates that hub tropism is a rapidly diverging phenotype .
A fundamental aspect of Wolbachia-host interactions is the type of tissue preferentially infected by the bacteria . We have previously shown that Wolbachia tropism to the stem cell niches in the female Drosophila ovaries is important for vertical transmission , and that this tropism is ubiquitous across the Drosophila genus . Furthermore , closely related Wolbachia strains tend to display the same patterns of tropism in the ovary , indicating the importance of maintaining this phenotype for vertical transmission [14] . If the major role of niche tropism is related to Wolbachia transmission , evolutionary theory predicts that there should be reduced selective pressure to maintain niche tropism in males , since Wolbachia is not transmitted through the sperm . Patterns of Wolbachia niche tropism in the filarial nematode ( B . malayi , D . immitis , L . sigmondontis , M . unguiculatus , and O . dewittei japonica ) support this concept , where Wolbachia colonization of the distal tip cell ( the nematode equivalent of the stem cell niche ) and subsequent germline invasion occurs only in females [26] . In agreement , the results shown here indicate a reduced level of conservation of hub tropism phenotype , contrasting with previous observation in females [14] . The stem cell niches in the ovary and testis are well characterized and have several signaling pathways in common [27] . The robust sexual dimorphism in the evolutionary conservation of niche tropism , indicates that Wolbachia could be recognizing novel sex specific differences in these cells [28] . Wolbachia-induced host phenotypes related to stem cell biology and testis physiology have been previously described [17] , [23] . We investigated whether hub tropism correlates with those known Wolbachia-related reproductive phenotypes . Because GSCN tropism in the ovary was shown to not be ubiquitous across the Drosophila genus , we reasoned that hub tropism could simply be a byproduct of GSCN tropism in the female . However , the frequencies of GSCN and hub tropism only correlate in 5 out of the 10 species and are not statistically significant . On the cellular level , another phenotype we have previously shown was a Wolbachia-dependent increase in the rate of germline stem cell division ( GSCD ) in the ovaries of D . mauritiana . Although a similar trend exists in the D . mauritiana testis , the up-regulation of GSCD was not shown to be significant , showing a lack of conservation of a phenotype derived in the females to boost their spread [17] . A third important Wolbachia mediated phenotype , cytoplasmic incompatibility ( CI ) , is a consequence of Wolbachia modification of sperm during spermatogenesis , causing embryonic lethality of uninfected eggs fertilized by sperm from infected males [reviewed by 29] . Although the precise mechanism is not well understood , the sperm from infected males is modified ( mod+ ) and an infected egg with the appropriate rescue factor ( resc+ ) is required for embryo viability [30] , [31] . Several lines of evidence suggest that the modification of the sperm occurs at the chromatin level [32]–[34] . Extensive analyses of Wolbachia population dynamics and localization during spermatogenesis have demonstrated that CI is a non-cell autonomous effect caused by a diffusible Wolbachia factor during spermatogenesis [35] . Interestingly , local factors secreted by the hub can act on the germline stem cell . Since niche factors are extrinsic to the stem cell , they can affect the testis germline stem cell and consequently their sperm-forming progeny in a non-cell autonomous fashion . Niche factors have also been shown to cooperate with chromatin remodeling complexes towards control of germline stem cell maintenance and differentiation [36] . Therefore , we attempted to correlate our tropism data with published data regarding CI levels of several Wolbachia strains across the Drosophila genus . However , we found no correlation between Wolbachia hub tropism and CI , suggesting that Wolbachia's presence in the hub is not required for the CI effect . This suggests that either Wolbachia factors modify the sperm later in spermatogenesis or if Wolbachia-derived factors are affecting early spermatogenesis events towards CI , it is independent of Wolbachia infection of the niche . Literature shows that both the host species and Wolbachia strains have rapidly evolving aspects that could contribute to the dynamic evolutionary changes in Wolbachia hub targeting shown here . Regarding the host , several testis specific genes , male seminal fluid proteins , and spermatogenesis genes have been shown to be rapidly evolving [37] . Furthermore , proteins related to GSC biology are also undergoing recurrent positive selection [38] . From the perspective of the bacteria , Wolbachia genomic analyses suggest that these bacteria have one of the most highly recombining intracellular bacterial genomes , with many genomic differences between closely related strains [39]–[42] . We investigated the relative contribution of both host and bacterial factors towards hub tropism phenotype . Unlike in the ovary where host derived factors did not play a role [14] , in the testis , host factors could not be ruled out . When comparing distantly related Wolbachia strains and host species ( D . mauritiana and D . sechellia hybrid lines ) , the data indicate that both host and Wolbachia derived factors contribute to the differences in hub tropism . One possibility is that there is selective pressure on the host driving rapid evolution of the hub intracellular environment to counteract negative effects of Wolbachia colonization of the testis niche . Although there is no evidence in the literature for positive selection of hub proteins , genes in the neighboring germline stem cell have been shown to be undergoing positive selection [38] , [43] . Independent of differential host factors , we were able to confirm Wolbachia's role in hub tropism . By comparing distantly related Wolbachia strains in the same host species ( D . simulans lines ) , we were able to confirm that Wolbachia derived factors significantly modulate hub tropism . To assess how quickly this modulation of hub tropism can evolve , we investigated if very closely related Wolbachia strains that have recently diverged could display diverse hub tropism phenotypes . Several variants of the wMel strain of Wolbachia naturally infecting D . melanogaster exist [44] , [45] . Due to strict maternal transmission , congruent Wolbachia and mitochondrial lineages made it possible to trace these lineages back to a single common D . melanogaster ancestor existing around 8 , 000 years ago [24] , [25] . We investigated hub tropism of wMel-like ( wMel , wMel2 , and wMel3 ) and wMelCS-like ( wMelCS , wMelCS2 and wMelPop ) Wolbachia strains which have been shown to induce differential protection against viruses [25] . The wMel-like and wMelCS-like subgroups can be separated into three statistically distinct groups based on their density of hub infection ( 1: wMel , wMel2 , and wMel3; 2: wMelCS and wMelCS2; 3: wMelPop ) , indicating that they have evolved distinct cellular tropisms . These data demonstrate that hub tropism is a rapidly diverging phenotype . The fast paced changes in the hub tropism phenotype during the evolution of these different Wolbachia strains raises the questions of what mechanisms are driving these rapid changes and is adaptive evolution occurring . If Wolbachia tropism for the hub is causing an unfavorable phenotype in the host , a molecular arms race will result where both the host and microbe will rapidly evolve [46] , [47] . We did not find any correlation of hub tropism with CI , germline stem cell division , or with other obvious testis related phenotypes . It is possible that hub tropism may have a phenotypic effect on the host , but at the moment these are unknown and we have no evidence supporting adaptive evolution in response to a host-microbe arms race driving rapid changes in hub tropism in wMel strains . Another possibility is that genetic drift is driving the extreme divergence in hub tropism that we report here . At every generation , from embryonic development through the mature egg , Wolbachia undergoes several bottlenecks: only the Wolbachia present in the germplasm of the embryo will colonize the primordial germ cells [8] , [10] . Within the germline , only the Wolbachia present in the oocyte is transmitted to the progeny [7] , [9] , [10] . This effectively reduces the genetic effective population sizes and increases the rate of fixation of mutations by drift . There are several studies highlighting the role of genetic drift driving high rates of genome sequence evolution in vertically transmitted endosymbionts [reviewed by 48] . The data presented here suggest that mutations that are neutral regarding niche targeting in the female may affect niche tropism in the male . If these mutations do not affect Wolbachia overall fitness in the females and do not interfere with transmission , they can be fixed by drift and result in significant niche tropism evolution in males . At the moment it is difficult to identify the specific molecular underpinnings resulting in the differences in niche tropism phenotypes between these strains . A possible molecular player involved in hub tropism could be encoded by the gene region known as ‘octomom’ . This region was found to be amplified several times in wMelPop , and contains genes predicted to be involved in DNA replication . It has been proposed to be responsible for the wMelPop over-replication phenotype [25] , although there are conflicting reports [49] . This could explain the highest titers present in wMelPop-infected hubs . However , there are other unknown factors contributing to the range of hub tropism phenotypes observed in the other wMelCS-like and wMel-like strains , since they have only once copy of the octomom region . The wMel variants are defined by several polymorphic genetic markers [25] , [44] , [45] , [49] . There are 108 single nucleotide polymorphisms ( SNPs ) , a tandem duplication , and seven insertion-deletion polymorphisms between the wMel and wMelCS-like ( wMelPop ) strains [25] . Further characterization of niche tropism of different strains in the same host genetic background , together with additional sequencing of diverse strains , will allow the correlation of Wolbachia genomic features with patterns of niche tropism . Future identification of Wolbachia proteins modulating the different levels of hub tropism will provide insights into the evolutionary mechanism driving this rapid divergence in males and the robust sexual dimorphism of stem cell niche targeting . Here we presented tropism differences in Wolbachia strains well characterized at the genomic level in a Drosophila species with a large repertoire of transgenic and genetic tools . These findings provide the foundation to dissect the molecular mechanisms involved in Wolbachia hub tropism . Furthermore , the differences in stem cell niche tropism between males and females may reveal sex specific differences in the biology of stem cell niche being recognized by Wolbachia . Identification of the Wolbachia factors involved in tissue tropism is fundamental in understanding how bacteria spread and infect their hosts in nature and will provide additional tools towards vector and disease control .
Fly stocks used in this analysis and their sources are listed in S1 Table . Drosophila species naturally infected with Wolbachia comprising the melanogaster subgroup were selected , along with two additional species outside the melanogaster subgroup: D . tropicalis and D . ananassae , belonging to the willistoni and ananassae subgroups , respectively . Introgression crosses for hybrid analysis experiments were performed as previously described [14] . D . melanogaster flies infected with the several wMel Wolbachia variants were introduced into the same genetic background as described elsewhere [25] . Flies were raised at room temperature and fed a typical molasses , yeast , cornmeal , agar food , with the exception of D . sechellia flies which were supplemented with reconstituted Noni Fruit ( Hawaiian Health Ohana , LLC ) [50] . For consistency and proper comparison to previous analysis of niche tropism in the female , males in this study were aged to seven days at room temperature ( with the exception of the D . simulans hybrids for Fig . 4 , which were dissected upon eclosion , see Toomey et al , 2013 for details ) . At least 20 flies were dissected for each sample , and total N's of hubs analyzed are listed in the Supplemental tables for each experiment . Testis were fixed using a 4% paraformaldehyde solution and subjected to immunostaining as previously described [13] . The mouse anti-hsp60 ( Sigma , 1∶100 ) antibody was used to visualize Wolbachia . Hub markers were either rat anti-α-catenin ( DSHB , DCAT1 , 1∶40 ) or rat anti-DE-Cadherin ( DSHB , DCAD2 , 1∶20 ) . Nuclei were counterstained with Hoechst ( 1 µg/ml , Molecular Probes ) . Images of the hub were acquired using a FV1000 confocal microscope . Wolbachia signal intensity in the hub and surrounding area were measured in Z-stacks of images using MatLab software for image quantification . Manual masks were drawn around the hub structure as well as the surrounding soma and germline using only the hub marker and DNA . Wolbachia density was measured within each mask and Wolbachia infection of the hub was considered tropism if the density relative to the surrounding soma and germline was at least 1 . 5-fold increased . A 1 . 5-fold threshold for tropism was previously determined to best represent what visually appears to be a higher density of Wolbachia in the niche versus the surrounding tissue [14] . Raw data showing density ratios is provided in S1 Dataset . We utilized a computer simulation model of randomized character distributions to compare with the distribution of niche tropism pattern on each of the phylogenies to quantify the correlation of niche tropism pattern to the Wolbachia and Drosophila phylogenies ( S1 and S3 Figs . ) [51] . We used tree length as a measurement for goodness of fit for the distribution of a character , such as the tropism pattern , as aligned with the phylogeny . Tree length is defined as the total number of steps required to map a data set onto a phylogenetic tree . To determine the three significant groups for tropism in Fig . 1 , a two-sample test for proportions was used on frequency data ( Fig . 1K ) and T-tests were used for density data ( Fig . 1L ) . A Bonferroni correction was applied to account for multiple comparisons . To determine the significance of host genetic background versus Wolbachia strain ( Fig . 4 ) on the frequency of hub tropism a logistical regression was performed on frequency data as previously described ( Fig . 4B ) [14] . When “zero” frequencies are present , logistic regression analysis was replaced by a Fisher Exact Test ( Fig . 3B ) . For density data , a linear regression was performed ( Fig . 3C ) . To determine if the frequencies of targeting between Wolbachia strains were significantly different ( Fig . 5B ) , a two-sample test for proportions was used . If there were more than two strains being compared a Chi-square test was performed . To determine if the differences in densities were significant , pair-wise t-tests were performed ( Fig . 5C ) .
|
Microbes evolve to infect structures favoring their transmission in host populations . A large fraction of insects are infected with Wolbachia bacteria . Usually Wolbachia are transmitted the same way we inherit our mitochondria , via the eggs from the mother . In fruit flies , to favor maternal transmission , Wolbachia infect the microenvironment containing the egg producing stem cells , called the “stem cell niche” . Targeting of the stem cell niche is evolutionary conserved in female fruit flies , observed in all Wolbachia strains analyzed to date . Remarkably , in males , we find many Wolbachia strains not infecting the stem cell niche present in the testis . We report a surprising diversity in stem cell niche infection in males , contrasting with extreme conservation in females . We further show that even closely related Wolbachia strains in D . melanogaster display rapidly evolving patterns of stem cell niche targeting in males . Understanding the molecular mechanisms driving these differences will identify sex specific features of stem cell niche biology . Because Wolbachia promote insect resistance against human diseases transmitted by mosquitos , Wolbachia are becoming a valuable tool in the control of several diseases , including Dengue and malaria . Knowledge emerging from this research will also provide novel tools towards Wolbachia based strategies of disease control .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"invertebrates",
"symbiosis",
"animals",
"wolbachia",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"stem",
"cells",
"bacteria",
"drosophila",
"research",
"and",
"analysis",
"methods",
"animal",
"cells",
"stem",
"cell",
"niche",
"insects",
"arthropoda",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"species",
"interactions",
"organisms"
] |
2014
|
Extreme Divergence of Wolbachia Tropism for the Stem-Cell-Niche in the Drosophila Testis
|
According to the Red Queen hypothesis or arms race dynamics , coevolution drives continuous adaptation and counter-adaptation . Experimental models under simplified environments consisting of bacteria and bacteriophages have been used to analyze the ongoing process of coevolution , but the analysis of both parasites and their hosts in ongoing adaptation and counter-adaptation remained to be performed at the levels of population dynamics and molecular evolution to understand how the phenotypes and genotypes of coevolving parasite–host pairs change through the arms race . Copropagation experiments with Escherichia coli and the lytic RNA bacteriophage Qβ in a spatially unstructured environment revealed coexistence for 54 days ( equivalent to 163–165 replication generations of Qβ ) and fitness analysis indicated that they were in an arms race . E . coli first adapted by developing partial resistance to infection and later increasing specific growth rate . The phage counter-adapted by improving release efficiency with a change in host specificity and decrease in virulence . Whole-genome analysis indicated that the phage accumulated 7 . 5 mutations , mainly in the A2 gene , 3 . 4-fold faster than in Qβ propagated alone . E . coli showed fixation of two mutations ( in traQ and csdA ) faster than in sole E . coli experimental evolution . These observations suggest that the virus and its host can coexist in an evolutionary arms race , despite a difference in genome mutability ( i . e . , mutations per genome per replication ) of approximately one to three orders of magnitude .
Host–parasite coevolution has been a topic of intense research interest in various fields from basic science of molecular evolution to agricultural and medical applications [1]–[5] . According to the Red Queen hypothesis or arms race dynamics , coevolution leads to complex but continuous change , adaptation , and counter-adaptation of the phenotypes of interacting organisms [2] , [6] , [7] . Futuyma and Slatkin suggested that investigation of coevolution could raise and help provide answers to many questions regarding the history of evolution , e . g . , whether parasites tend toward specialization or toward benign or even mutualistic relationships with their hosts [8] . There have been many previous observational and theoretical studies on natural host–parasite dynamics . With regard to the relationships between bacteria and phages , Rodríguez-Valera et al . proposed the constant-diversity dynamics model in which the diversity of prokaryotic populations is maintained by phage predation [9] . Moreover , an observational study supported the model by analyzing the dynamics of bacteria and phages in four aquatic environments using a metagenomics method and showed that microbial strains and viral genotypes changed rapidly [10] . In addition , experimental models in simplified environments have been employed to analyze the ongoing process of coevolution . Various pairwise combinations of bacteria and phages and one with Caenorhabditis elegans and bacteria have been subjected to long-term laboratory cultivation [11]–[15] . These studies indicated that coevolution proceeded on a laboratory time scale [11]–[14] , accelerated molecular evolution of parasites [16] , [17] , and broadened the host range of parasites [14] . However , the changes in genetic information and phenotype of parasites and their hosts through coevolution remain to be elucidated , and the changes in host specificity and virulence of the parasites through the arms race have not been determined in sufficient detail because ongoing adaptation and counter-adaptation in simplified experimental model systems have not been analyzed at the levels of population dynamics and molecular evolution . To examine the ongoing changes driven by host–parasite interactions , we have constructed a coevolution model consisting of Escherichia coli and the lytic RNA bacteriophage Qβ ( Qβ ) in a spatially unstructured environment . Qβ is a simple RNA bacteriophage that infects and lyses E . coli cells , taking about 1 h for its burst without escaping into a lysogenic state . It has a single-stranded RNA genome of 4 , 217 bases encoding four genes for A2 , A1 ( read-through ) , coat protein , and RNA replicase β subunit [18] . Due to a high misinsertion rate and lack of a proofreading mechanism , ribovirus RNA replicase ( including that of Qβ ) has a high mutation rate [18]–[22] , which allows us to monitor the evolutionary changes on a laboratory time scale . Here , we report that in coevolution through 54 daily copropagations of the parasite and its host , E . coli first evolved partial resistance to infection and later showed acceleration of its specific growth rate , while the phage counter-adapted by improving release efficiency with a change in host specificity and a decrease in virulence . Fitness analysis indicated that these phenotypic changes occurred within an arms race , i . e . , accompanied with a monotonic fitness increase of either the parasite or its host . Whole-genome analysis indicated that the phage accumulated 7 . 5 mutations mainly in the A2 gene 3 . 4-fold faster than in Qβ propagation evolution where the phage was transferred daily to freshly prepared E . coli cultures , while E . coli showed fixation of two mutations ( in traQ and csdA ) faster than in sole E . coli experimental evolution . The results indicated ongoing adaptation and counter-adaptation through a host–parasite arms race .
Evolution experiments were carried out with copropagation of E . coli and Qβ and with propagation of Qβ only ( Figure 1A ) . In the copropagation experiment , the ancestral E . coli strain HL2 ( Anc ( C ) ) and Qβ derived from cloned Qβ cDNA [23] ( Anc ( P ) ) were mixed , cultivated , and diluted so that the next daily culture was initiated at approximately 1×107 E . coli cells/ml . We calculated the replication generations of Qβ genome as the cumulative generations of each passage , ( Nfinal/Ninitial ) = 2g , where Nfinal and Ninitial represent final and initial free phage density of each passage in plaque forming units ( PFU/ml ) , respectively , and g represents replication generation . We also calculated E . coli cell generations as the cumulative generations of each passage , ( Nf/Ni ) = 2n , where Nf and Ni represent the final and initial colony forming units ( CFU/ml ) of each passage , respectively , and n represents cell generation . In the very early phase of the copropagation experiment , the cell generation was underestimated due to cell lysis by infection . The copropagation experimental population was divided into two on day 18 , equivalent to 59 replication generations and 62 cell generations . Culture was continued to a total of 54 days ( lines 1 and 2 ) , equivalent to 163 replication generations and 163 cell generations for line 1 , and 165 replication generations and 164 cell generations for line 2 ( Figure 1B ) . Two Qβ propagation experiments , lines 3 and 4 , were conducted in parallel for 18 days , equivalent to 169 and 168 replication generations where the phage population was separated daily by centrifugation from the host and transferred into fresh logarithmic cultures of the host Anc ( C ) ( Figure 1B ) . The population dynamics of the copropagation experiment demonstrated the coexistence of E . coli and Qβ ( Figure 2 ) , although Qβ is lytic and has no lysogenic state . The daily Qβ population density fluctuated over the course of the copropagation experiment , while the E . coli population density was stable probably due to the constant initial density of the host at each daily coculture . The degree of phage amplification in the copropagation experiment ( 2–20-fold per single coculture ) was substantially lower than that in the Qβ propagation experiment ( approximately 1 , 000-fold ) , even though the initial multiplicity of infection ( MOI ) in each passage was approximately 0 . 5 ( approximately 107 phages/ml over 2×107 E . coli cells/ml ) for the Qβ propagation experiment and was not higher than that for the copropagation experiment ( Figure 2B and 2C ) . These observations suggested that the biotic environment for phage amplification , i . e . , the cellular state of the host E . coli , changed during the copropagation experiment . Cross-cocultures were conducted to determine the changes in fitness of E . coli and Qβ in the copropagation experiment . Four hosts ( Anc ( C ) , M54 ( C ) , M163 ( C ) , and M165_2 ( C ) ) , and the four corresponding phages ( Anc ( P ) , M54 ( P ) , M163 ( P ) , and M165_2 ( P ) ) at the 1st , 54th , 163rd , and 165th replication generations in the copropagation experiments of lines 1 and 2 were cocultured to measure fitness in each pairwise combination . Here , the fitness of E . coli is defined as the ratio of the initial to the stationary optical density at 600 nm ( OD600 ) , while the fitness of the phage is the ratio of the initial to the stationary free phage density ( PFU/ml ) ( Table 1 , Table 2 and Figure S1 ) . The host E . coli evolved partial resistance along with its increase in fitness ( Table 1 ) . The evolved hosts M54 ( C ) , M163 ( C ) , and M165_2 ( C ) showed phage amplification ratios two to three orders of magnitude lower than Anc ( C ) , regardless of whether the ancestral or evolved phage was used ( host: Anc ( C ) , M54 ( C ) , M163 ( C ) , and M165_2 ( C ) , parasite: Anc ( P ) , M54 ( P ) , M163 ( P ) , and M165_2 ( P ) : one-way ANOVA F3 , 24 = 213 , P<0 . 01; post hoc Tukey–Kramer test , P<0 . 01; Table 2 ) . The resistance was only partial , allowing phage amplification of only approximately one order of magnitude . On the other hand , the host E . coli infected with Anc ( P ) gradually showed an increase in amplification ratio along with host evolution ( one-way ANOVA F2 , 3 = 469 , P<0 . 01; post hoc Tukey test detected significant differences between all combinations: Anc ( C ) vs . M54 ( C ) and Anc ( C ) vs . M163 ( C ) , P<0 . 01; M54 ( C ) vs . M163 ( C ) , P<0 . 05; Table 1 ) . The growth curve of M163 ( C ) inoculated with the phages became similar to that of the uninfected host ( Figure S1A , third from left ) , suggesting that the host population evolved , increasing its fitness , to become almost oblivious to the phages . Despite the development of partial resistance by the host , the phage also increased its fitness through changes in host specificity ( Table 2 ) . The evolved phage M54 ( P ) and M163 ( P ) showed higher fitness on the evolved host M54 ( C ) with partial resistance than Anc ( P ) on the same host ( one-way ANOVA F2 , 3 = 19 . 7 , P<0 . 05; post hoc Tukey test , P<0 . 05; Table 2 ) . In line 1 and line 2 , the most evolved Qβ , M163 ( P ) or M165_2 ( P ) showed the highest fitness on corresponding E . coli , M163 ( C ) or M165_2 ( C ) , respectively . Briefly , there was a significant difference in fitness among the host–parasite combinations ( host: M163 ( C ) or M165_2 ( C ) , parasite: Anc ( P ) , M54 ( P ) , M163 ( P ) , or M165_2 ( P ) , one-way ANOVA F2 , 3 = 60 . 8 , P<0 . 01; post hoc Tukey test , P<0 . 05 for M163 ( C ) ; F2 , 3 = 37 , P<0 . 01; post hoc Tukey test , P<0 . 05 for M165_2 ( C ) ; Table 2 ) . The phage evolved through natural selection to show greater amplification on the corresponding host , although the amplification ratio itself decreased from approximately 104 to 101 . The phage , while responding adaptively to the evolutionary changes of its host , showed a decrease in amplification ratio on the ancestral host strain , leading to a decrease in virulence . The amplification ratios of the phage on the host Anc ( C ) gradually decreased over the course of the copropagation experiment ( host: Anc ( C ) , parasite: Anc ( P ) , M54 ( P ) , M163 ( P ) , and M165_2 ( P ) , one-way ANOVA F3 , 4 = 17 . 4 , P<0 . 01; post hoc Tukey test , P<0 . 05; Table 2 ) . A decline in phage amplification was also observed as a reduction in plaque size ( Figure 3 ) . Consequently , the evolved phage showed less cell killing effect against the ancestral strain Anc ( C ) , resulting in better growth of the ancestral bacterial strain ( host: Anc ( C ) , parasite: Anc ( P ) , M54 ( P ) , M163 ( P ) , and M165_2 ( P ) , one-way ANOVA , F3 , 4 = 340 , P<0 . 01; post hoc Tukey test , P<0 . 01; Table 1 and Figure S1A , left ) , i . e . , the phage showed a decrease in virulence . In addition , the phage evolved in the Qβ propagation experiment ( S94_3 ( P ) ) showed the greatest amplification ratio ( host: Anc ( C ) , parasite: Anc ( P ) , M54 ( P ) , M163 ( P ) , M165_2 ( P ) , and S94_3 ( P ) , one-way ANOVA F4 , 5 = 37 . 8 , P<0 . 01; post hoc Tukey test , P<0 . 05; Table 2 and Figure S1B , left ) and similar virulence against Anc ( C ) with Anc ( P ) ( host: Anc ( C ) , parasite: Anc ( P ) and S94_3 ( P ) , Welch's t test , t = 12 . 7 , P = 0 . 45; Table 1 and Figure S1A , left ) . These results suggest that the decrease in virulence was not due to simple degeneration through the long-term passage experiment , but was probably at the expense of increasing the fitness of the phage in the arms race with its host . To examine how E . coli and Qβ improved their fitness during coevolution , free phages , infected E . coli cells , and total E . coli cells from the copropagation of line 1 were monitored hourly by determining the numbers of PFUs in the supernatant and pellet after centrifugation and CFU , respectively ( see Materials and Methods ) . E . coli was found to first evolve partial resistance to Qβ , which was followed by a later increase in the specific growth rate . After 3 hours of incubation with Anc ( P ) , almost all of the ancestral host Anc ( C ) cells were infected , while infection ratios of the evolved hosts M54 ( C ) and M163 ( C ) were only 0 . 03% and 0 . 08% , respectively ( Figure 4A , 4B , and 4D , left ) . The observed partial resistance was likely due to a very low adsorption rate of E . coli cells to the phage ( Figure 5 ) . As most of the evolved host cells remained uninfected , they were able to proliferate , while the ancestral cells could not . In addition , the uninfected cells of the most evolved hosts , M163 ( C ) and M165_2 ( C ) for lines 1 and 2 , respectively , showed higher specific growth rates than those of M54 ( C ) ( see legend of Figure 6 for specific growth rates , ANCOVA , F2 , 24 = 18 . 0 , P<0 . 001; post hoc Tukey test , P<0 . 001 ) . Although the OD600/CFU seemed to have changed over the copropagation experiments ( Figure 6 ) , the same conclusion was obtained using specific growth rates based on CFU values ( data not shown ) . Briefly , M54 ( C ) eliminated Anc ( C ) from the population by developing partial resistance to phage infection , and M163 ( C ) and M165_2 ( C ) finally took over the population due to acceleration of specific growth rate . The phages evolved to show increased release efficiency , i . e . , the number of phages released from a single infected cell per unit time . As the phage evolved , the speed of free phage amplification for either M54 ( C ) or M163 ( C ) increased ( the amplification rates of free phage density of Anc ( P ) and M54 ( P ) on M54 ( C ) were 0 . 11 h−1 ( r2 = 0 . 43 ) and 0 . 31 h−1 ( r2 = 0 . 93 ) , respectively , two-tailed t test t = 3 . 24 , P<0 . 01; Figure 4B and 4C , right , and those of Anc ( P ) , M54 ( P ) , and M163 ( P ) on M163 ( C ) were 0 . 25 h−1 ( r2 = 0 . 79 ) , 0 . 24 h−1 ( r2 = 0 . 89 ) , and 1 . 38 h−1 ( r2 = 0 . 93 ) , respectively , ANCOVA; F2 , 18 = 35 . 3 P<0 . 01; post hoc Tukey test , P<0 . 01; Figure 4D , 4E , and 4F , right ) , while the infection efficiency , i . e . , the rate of increase in infected cells , did not change significantly for the same hosts ( the rates of increase in infected cells of M54 ( C ) infected with Anc ( P ) or M54 ( P ) were 1 . 21 h−1 ( r2 = 0 . 89 ) and 1 . 26 h−1 ( r2 = 0 . 98 ) , respectively , two-tailed t test , t = 0 . 39 , P = 0 . 70; Figure 4B and 4C , left , and those of M163 ( C ) infected with Anc ( P ) , M54 ( P ) , or M163 ( P ) were 2 . 40 h−1 ( r2 = 0 . 98 ) , 2 . 80 h−1 ( r2 = 0 . 92 ) , and 2 . 72 h−1 ( r2 = 0 . 99 ) , respectively , ANCOVA , F2 , 12 = 0 . 95 , P = 0 . 41; Figure 4D , 4E , and 4F , left ) . The acceleration of free phage amplification rate could be attributed to either an increase in burst frequency per unit time or burst size . There was no significant difference in burst size between M163 ( P ) and Anc ( P ) on M163 ( C ) determined by the method of analysis of burst sizes in single cell [24] ( data not shown ) . Therefore , the phage seems to have evolved to burst more frequently from infected cells per unit time . It is noteworthy that the most evolved phage , M163 ( P ) , inoculated onto M163 ( C ) showed a marked increase in number of free phage at 3 h , probably leading to further infection of surrounding uninfected hosts and an increase in proportion of infected cells beyond the inoculated free phage concentration ( 4 . 8×105 PFU/ml ) ( Figure 4F ) . Other phages stopped increasing the number of infected cells at around the inoculated free phage concentration ( Figure 4D , 4E , and 4G ) . We performed whole genome sequence analyzes of all of the Qβ populations indicated in Figure 1B to determine how molecular evolution of the phage proceeded in response to the adaptation of the hosts . First , mutations were shown to be accumulated in a biased manner in the A2 gene , which encodes a multifunctional protein related to infection and cell lysis ( Figure 1B and Table 3 ) . The A2 gene , accounting for 30% of the whole genome , accumulated 65 . 5% of all mutations , and this bias was shown to be statistically significant ( P<0 . 05 , two-tailed binomial test ) . A similar substantial accumulation of mutations in genes related to host infection was demonstrated previously in an evolution experiment using the DNA bacteriophage Φ2 [16] . The mutation fixation rate in phage was higher in the copropagation experiment ( 1 . 0×10−5±6 . 0×10−7 per base per generation ) than that in the Qβ propagation experiment ( 3 . 2×10−6±5 . 1×10−7 per base per generation ) ( two-tailed Welch's t test t = 4 . 3 , P<0 . 01 ) , suggesting that the phage showed accelerated molecular evolution through coevolution with its host ( Figure 7 ) . Whole-genome analysis of E . coli revealed the process of molecular evolution in the host cells . We analyzed the whole genome sequence of M163 ( C ) using an Illumina Genome Analyzer IIx ( GAIIx; Illumina , San Diego , CA ) and confirmed the mutations in M163 ( C ) together with Anc ( C ) and M54 ( C ) by the dideoxynucleotide chain termination sequencing method [25] . A single mutation in traQ ( S21P ) encoded on the F plasmid was detected in M54 ( C ) and an additional mutation was detected in csdA ( D340N ) in M163 ( C ) ( Table S1 , Table S2 ) . As discussed below , the protein products from these genes may contribute to resistance to phage infection and the increase in fitness of E . coli .
In the copropagation experiment , E . coli adopted a simple strategy with only two mutations , while the phage accumulated more mutations within its small genome as counter-adaptation against the evolutionary changes in the host . The host first developed resistance to phage infection via a non-synonymous mutation in traQ . This gene encodes TraQ , the conjugal transfer pilin , which is a component of the F pilus , and is a chaperone for inserting propilin into the inner membrane . Propilin was reported to be unstable in traQ− cells [26] , and amino acid 21 of TraQ where the mutation was detected in this study interacts with propilin [27] . F pilus assembly from membrane F-pilin requires many Tra proteins [28] . As no mutations were detected in other Tra protein genes in the copropagation experiment , the mutation on TraQ may result in a decrease in the amount of inserted propilin , leading to the partial resistance observed in this study . E . coli then showed further mutation in csdA , which encodes CsdA , an enzyme related to Fe/S biogenesis and a new sulfur transfer pathway that is related to the fitness of these cells , especially in stationary phase [29] . Therefore , this mutation could be beneficial as the host was passaged daily at the stationary phase in the evolutionary experiment . On the other hand , the phage evolved to increase release efficiency by accumulating mutations mostly in the gene encoding the A2 protein . A2 is a multifunctional protein with roles in host cell lysis , adsorption to the F pilus of E . coli , RNA binding during capsid assembly , protection of the 3′ terminus , penetration into the cytoplasm of the host , and blockage of cell wall biosynthesis by inhibiting the catalytic step from UDP-GlcNAc to UDP-GlcNAc-EP catalyzed by MurA [18] , [30]–[32] . Due to the cell lytic activity of A2 , it is unsurprising that these mutations might have resulted in an increased burst frequency and release efficiency . In fact , the experimental lag period between infection and detection of the increase in free phage became shorter by approximately 1 hour in the cross-culture experiment ( e . g . , 3 h for Anc ( P ) and 2 h for M163 ( P ) on M163 ( C ) , Figure 4D and 4F , right , respectively ) . It should be noted that the uninfected M163 ( C ) reached the stationary phase , which was not susceptible to phage infection , approximately 1 hour earlier than the other hosts ( see legend of Figure 6 for the time to reach the stationary phase , one-way ANOVA , F2 , 3 = 693 . 8 , P<0 . 001; post hoc Tukey test , P<0 . 001 ) . Thus , it is possible that the increased burst frequency of M163 ( P ) for the earlier phage release evolved as a counter-adaptation on the evolved host M163 ( C ) due to the shorter period available for infection . Previous experiments using DNA bacteriophages indicated that shorter latent periods were favored in the presence of a high density of highly susceptible host cells [33] , [34] . It is of interest that Qβ evolved to show reduced virulence toward the ancestral host . Many studies have indicated that phages with low or moderate virulence were favored in vertical transmission or in structured environments [35]–[37] , while Qβ has no lysogenic state and evolved reduced virulence in this experiment . The decrease in virulence observed in this study may have been a side effect of the increase in burst frequency . If fact , the evolved phage M163 ( P ) with increased burst frequency on the evolved host M163 ( C ) showed lower virulence and lower fitness on Anc ( C ) than Anc ( P ) on Anc ( C ) ( Figure 4 and Figure S1 left ) , suggesting that Qβ may have co-evolved to increase the burst frequency in reducing some benefits that can be gained if the host reverts to the Anc ( C ) -like phenotype . The single non-synonymous mutation at position 221 of the A2 gene found in M54 ( P ) seems to have resulted in reduced virulence and a change in host specificity . As the non-synonymous mutation was only observed in the copropagation experiment and two others were also observed in Qβ propagation and the deposited sequence ( NCBI accession no . AY099114 ) , the mutation at 221 and/or the combinations with the mutation and two other mutations may have resulted in the decrease in virulence and the change in host specificity observed in M54 ( P ) . In coevolution between Qβ and its host , E . coli , the phage showed accelerated molecular evolution ( Figure 7 ) . In the Qβ propagation experiment , the molecular evolution of the phage proceeded but seemed to slow down after the 94th generation . On the other hand , the phage coevolving with E . coli retained a 3 . 4-fold faster molecular evolution rate throughout the copropagation experiment . The higher evolution rate may be attributable to the changes occurring in the host E . coli . If the host had stopped evolving , e . g . , at the 54th generation , the M163 ( P ) or M165_2 ( P ) phage would not have been fixed into the population as it had fitness similar to or less than that of M54 ( P ) , leading to deceleration of evolutionary rate . It should be noted that neutral mutations cannot be fixed in the copropagation experiment because 163 replication generations is too short for them to become fixed . The fixation of neutral mutations is known to require generations approximately as long as the effective population size ( Ne ) [38] . The effective population size in the copropagation experiment was roughly estimated as the bottleneck size of the population ( approximately 103 phages ) assuming that 1% of the minimum initial 106 phages infect and burst to release approximately 107 phages . Thus , even synonymous mutations observed here were positively selected [38] , consistent with the influence of the RNA secondary structure on Qβ genome replication reported previously [39]–[42] . Some synonymous mutations may have physiological impacts on phage growth because of genomic secondary structure; it has been reported that some synonymous mutations or mutations in intergenic regions show lethal effects in Qβ [42] . It is noteworthy that the fixation rate of the E . coli genome in the copropagation regime ( 2 . 6×10−9 per bp per generation ) calculated as 2 mutations in 4 . 73 Mbp per 163 generations was one order of magnitude higher than that under conditions of E . coli sole passage , maintaining log phase at 37°C ( 1 . 7×10−10 per bp per generation ) [43] or 20 , 000 generations ( 1 . 6×10−10 per bp per generation ) ( Poisson distribution , P<0 . 01 ) [44] . In summary , these observations indicated that molecular evolution rates of both the parasite and its host were accelerated through adaptation and counter-adaptation . Based on the observed fitness changes in the host E . coli and in the Qβ phage , we propose a plausible coevolution path to depict the arms race between Qβ and the host E . coli . As the order of phage fitness on the ancestral E . coli Anc ( C ) was Anc ( P ) >M54 ( P ) >M163 ( P ) , the population in the coculture seemed to first take a route not in the direction of phage evolution ( upward ) but in the direction of host evolution ( right ) , increasing host fitness by increasing its resistance to Qβ ( Figure 8 ) . The arrows in Figure 8 reflect the experimentally determined finesses changes ( Table 1 and Table 2 ) . Arriving around the pair position of M54 ( C ) vs . Anc ( P ) , the population could take either the upward or rightward direction , but happened to take the direction of phage evolution due to the occasional appearance of a single non-synonymous mutation at position 221 in the phage genome that was detected only under copropagation conditions ( Table 3 ) . The population of M54 ( P ) and M54 ( C ) could not fix a phage mutant like M163 ( P ) with the same fitness as M54 ( P ) , but fixed the E . coli mutant M163 ( C ) with fitness higher than that of M54 ( C ) . Due to the host change from M54 ( C ) to M163 ( C ) accompanied with an additional single non-synonymous mutation in csdA , the phage mutant M163 ( P ) was fitter than M54 ( P ) and was therefore fixed in the final population . Taken together , these findings indicated that the evolutionary path seemed to be an arms race involving adaptation of E . coli and counter-adaptation of the phage . We showed that parasites , such as RNA viruses , and hosts , such as E . coli , have the potential to coexist even in an arms race . When a parasite encounters its host , the host may become extinct through the evolution of high parasite virulence , or the parasite may become extinct through the evolution of host resistance . However , both may also change their phenotypes by genomic mutation in a synchronized manner and thus coexist . The results of the present study indicated that a host with a larger genome size ( 4 . 6 Mbp ) with a low spontaneous mutation rate ( 5 . 4×10−10 per bp per replication ) [45] and a parasite with a smaller genome size ( 4 , 217 bases ) and a higher spontaneous mutation rate ( 1 . 5×10−3–10−5 per base per replication ) [18]–[22] , despite the large difference in mutability of their genomes ( approximately one to three orders of magnitude difference ) , were capable of changing their phenotypes to coexist in an arms race . Further studies linking the phenotype mutability and genome complexity will help to elucidate the dynamic host–parasite relationship .
The E . coli HL2 strain was used as the coculture host strain and A/λ [46] was used as an indicator strain for the titer assay . The E . coli HL2 strain was constructed by conjugation with DH1ΔleuB:: ( gfpuv5-Kmr ) [43] and HB2151 [47] . We mixed log-phase DH1ΔleuB:: ( gfpuv5-Kmr ) and HB2151 for 2 . 5 hours and screened for kanamycin-resistant clones on LB agar medium supplemented with 25 µg/ml kanamycin . F′ retention of HL2 was checked by PCR with the primers TraU_f ( 5′-ATGAAGCGAAGGCTGTGGCT-3′ ) and TraU_r ( 5′-GCAGCTTGAACGCCATGCGT-3′ ) and the ability of HL2 to amplify Qβ was confirmed . Before the evolution experiments , HL2 was grown in mM63gl ( 62 mM K2HPO4 , 39 mM KH2PO4 , 15 mM ammonium sulfate , 1 . 8 µM FeSO4·7H2O , 15 µM thiamine hydrochloride , 2 . 5 mM MgSO4·7H2O , 0 . 04% glucose , and 1 mM l-Leu ) for several passages until the specific growth rate had become stable , and the strain with stable growth rate was used as the ancestor strain ( Anc ( C ) ) . The OD600 of stationary-phase Anc ( C ) cultured in mM63gl medium was approximately 0 . 4 ( approximately 3×108 CFU/ml ) because of glucose limitation . Qβ was kindly provided by Dr . Koji Tsukada ( Osaka University , Japan ) , which was generated from Qβ genomic cDNA [23] . Qβ particles were diluted with LB medium and plaque assay was performed according to the standard method [48] . Polypropylene centrifuge tubes ( 15 ml , No . 430791; Corning Incorporated , Corning , NY ) treated with 0 . 1% BSA for at least 15 minutes to prevent attachment of phages to the tube walls were used for all the experiments as culture tubes . Copropagation experiment: 4 . 8×107 cells Anc ( C ) and 5 . 1×107 PFU Anc ( P ) were mixed and copropagation was started in a culture volume of 3 ml at 37°C with shaking at 160 rpm . Mixed cultures were divided into 2 lines on the 18th day , equivalent to 59 replication generations , and propagated independently for a further 36 days ( line 1 and line 2 in Figure 1B ) . Serial transfer was conducted by daily transfer of the cultures with cells and phages . The portion of cultures calculated based on the final OD600 were transferred into fresh medium with dilution to an initial OD600 of 0 . 05 . Daily culture samples were divided into thirds: one for preparing −80°C frozen stocks with 15% glycerol , one for CFU determination by dilution and spreading on low divalent cation mM63gl agar medium with 0 . 2 mM MgSO4·7H2O , and the other for PFU analysis using the supernatant after centrifugation . Qβ propagation experiment: Two lines ( line 3 and line 4 in Figure 1B ) were independently propagated from Anc ( P ) for 18 days , equivalent to 168–169 replication generations , at 37°C with shaking at 160 rpm . Serial passages consisted of infection of a host culture , followed by about 6 h of phage growth , and extraction of the phage from the culture . Each serial passage was performed as follows: uninfected Anc ( C ) cultures were grown at 37°C overnight and transferred into new medium with dilution to OD600 of 0 . 03 . When OD600 became 0 . 06–0 . 07 ( approximately 1×107 CFU/ml ) after 2–2 . 5 h , cells were infected with phage to approximately 1 . 0–2 . 0×107 PFU/ml from the previous passage . The cultures were grown for about 6 h . E . coli cells were removed by centrifugation , and the supernatant was subjected to filtration with 0 . 2 µm syringe filters ( Minisart RC15 filters; Sartorius Stedim Biotech , Goettingen , Germany ) , and phage solution was stored at 4°C for infection on the next serial passage . The replication generation number of the phage population ( n ) was calculated as n = ln2 ( Nf/Ni ) , where Ni and Nf are the phage density ( PFU/ml ) at the initial and final time points of each passage , respectively . The initial value ( Ni ) was calculated by dividing the Nf of the previous passage by the dilution rate . The evolved E . coli populations ( M54 ( C ) , M163M ( C ) , and M165_2 ( C ) ) and Qβ phage populations ( M54 ( P ) , 163 ( P ) , and M165_2 ( P ) ) were purified from mixed cultures to analyze the phage genome sequence and to determine their fitness . To purify the evolved E . coli population , cultures stocked at −80°C including evolved E . coli and phage were streaked on mM63gl agar medium and then passaged several times in low divalent cation medium , 0 . 2 mM MgSO4·7H2O mM63gl , to prevent further phage adsorption to E . coli . We checked the purity of evolved E . coli by confirming that no plaques were observed in the passaged and chloroform-treated cultures . To purify the evolved phage population , cultures stocked at −80°C including evolved E . coli and phage were cultured in mM63gl at 37°C with shaking at 160 rpm for 1 day and filtrated with 0 . 2 µm syringe filters ( Minisart RC15 filters; Sartorius Stedim Biotech ) . These particles were used for RNA genome sequencing analysis as described below . For analysis of phage fitness , these filtrated particles and Anc ( P ) were dialyzed to remove carry-over glycerol from the −80°C stock using Microcon centrifugal filter devices with 10 , 000 nominal molecular weight limit membranes ( Millipore , Billerica , MA ) . We confirmed that the dialysis step did not affect plaque forming ability . The RNA genomes of the Qβ population noted in Figure 1B , i . e . , 8 kinds of genome derived from approximately 108 PFU particles of the Anc ( P ) , M54 ( P ) , M109 ( P ) , M163 ( P ) , M165_2 ( P ) , S94_3 ( P ) , S169_3 ( P ) , and S168_4 ( P ) phage populations , were extracted using a QIAamp Viral RNA mini kit ( Qiagen , Hilden , Germany ) according to the manufacturer's instructions . To analyze the full-length RNA genome sequence , samples were prepared as follows . Poly ( A ) was added at the 3′ end using poly ( A ) polymerase ( Applied Biosystems/Ambion , CA ) . cDNA was synthesized using the primer Qt ( 5′-CCAGTGAGCAGAGTGACGAGGACTCGAGCTCAAGCTTTTTTTTTTTTTTTTT-3′ ) with SuperScript™ III Reverse Transcriptase ( Invitrogen , Carlsbad , CA ) , and then RNA was degraded with RNaseH . The first-strand cDNA was purified and poly ( A ) was added at the 3′ termini of the cDNA with terminal deoxynucleotidyl transferase ( Roche Diagnostics , Basel , Switzerland ) and dATP . The cDNA with poly ( T ) at the 5′ terminus and poly ( A ) at the 3′ terminus was purified . To obtain the 5′ end of the Qβ phage genome sequence , second-strand DNA was prepared using Qt primer and PfuUltra II Fusion HS DNA Polymerase ( Stratagene , La Jolla , CA ) . PCR was performed separately with high fidelity DNA polymerases for the whole Qβ genome divided into 5 regions as shown in Table S3 . In total , PCR products were obtained from cDNA template derived from approximately 106 PFU particles . The templates , primers , and polymerase used for PCR and the primers used for sequencing are listed in Table S3 . Sequencing was performed by the dideoxynucleotide chain termination method [25] on both strands , but we conducted direct sequencing in one direction using 2 sets of primers only for the 5′ end of the genome ( Table S3 ) . When a double peak appeared in the sequencing chart , positions where the height of the smaller peak was over half that of the larger peak were defined as polymorphic sites . Genomic DNA was extracted from over 109 cells of Anc ( C ) , M54 ( C ) , and M163 ( C ) using a DNeasy Blood & Tissue Kit ( Qiagen ) according to the manufacturer's instructions . The genomic DNA of M163 ( C ) was sequenced with an Illumina GAIIx ( Illumina ) using 51-bp of single-read format by Hokkaido System Science Co . , Ltd . ( Sapporo , Hokkaido , Japan ) . The GAIIx produced 18 , 718 , 549 reads and 954 , 646 kb . All reads were aligned to the reference sequence of E . coli DH1 genome sequence ( GenBank accession number , CP001637 . 1; genome size , 4 , 630 , 707 bp ) and F plasmid sequence ( GenBank accession number , NC_002483 . 1; size , 99 , 159 bp ) using MAQ [49] guaranteed to find alignments with up to 2 mismatches in the first 24 bp of the reads and 5 mismatches in 51 bp . Mean depth was 196 dividing 927 , 788 kb mapped bases by 4 , 730 kb of total reference sequences . After mapping and consensus base calling , SNPs were filtered with the same threshold values as reported previously [49] . The alignment view was also confirmed with Mapview [50] , and SNPs were also scored with the following parameters: Phred quality score ≥20 , variant frequency ≥0 . 40 , coverage sum ≥5 . In addition , Tablet [51] also has the ability to align short reads to the reference sequence , thus allowing us to score sites with a deletion , insertion , and/or regions with large insertions or deletions . SNPs , insertions , and deletions were scored for the F plasmid and for the M163 ( C ) genome , and they were confirmed by sequencing using the dideoxynucleotide chain termination method for the genomes derived from over 107 cells of Anc ( C ) and M163 ( C ) using the primers listed in Table S1 . Two different positions observed in Anc ( C ) and M163 ( C ) were sequenced for M54 ( C ) by the dideoxynucleotide chain termination method . The cross-coculture experiment , time course analysis after infection , and adsorption rate constant analysis were conducted according to the methods described in Text S1 . Fitness and the time to reach the stationary phase were compared by one-way ANOVA with the post hoc Tukey test [52] . The virulence of Anc ( P ) and S94_3 ( P ) to Anc ( C ) was compared with Welch's t test . The differences in growth rates calculated from semi-logarithmic plots of E . coli or phage densities were tested by two-tailed t test and ANCOVA with the post hoc Tukey test [53] .
|
To examine the ongoing changes driven by host–parasite interactions , we have constructed a coevolution model consisting of Escherichia coli and the lytic RNA bacteriophage Qβ ( Qβ ) in a spatially unstructured environment . In coevolution through 54 daily copropagations of the parasite and its host , E . coli first evolved partial resistance to infection and later accelerated its specific growth rate , while the phage counter-adapted by improving release efficiency with a change in host specificity and a decrease in virulence . Whole-genome analysis of E . coli and Qβ revealed accelerated molecular evolution in comparison with Qβ propagation in this study and E . coli sole passage reported previously . The results of the present study indicated that , despite the large difference in mutability of their genomes ( approximately one to three orders of magnitude difference ) , a host with larger genome size ( 4 . 6 Mbp ) and a lower spontaneous mutation rate ( 5 . 4×10−10 per bp per replication ) and a parasite with a smaller genome size ( 4 , 217 bases ) and a higher mutation rate ( 1 . 5×10−3 to 1 . 5×10−5 per base per replication ) were capable of changing their phenotypes to coexist in an arms race .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"organismal",
"evolution",
"evolutionary",
"ecology",
"population",
"genetics",
"microbiology",
"escherichia",
"coli",
"mutation",
"prokaryotic",
"models",
"model",
"organisms",
"coevolution",
"microbial",
"evolution",
"forms",
"of",
"evolution",
"biology",
"evolutionary",
"genetics",
"adaptation",
"natural",
"selection",
"genetics",
"evolutionary",
"biology",
"genomic",
"evolution",
"evolutionary",
"processes",
"genetics",
"and",
"genomics"
] |
2011
|
Ongoing Phenotypic and Genomic Changes in Experimental Coevolution of RNA Bacteriophage Qβ and Escherichia coli
|
Abnormalities in glycan biosynthesis have been conclusively linked to many diseases but the complexity of glycosylation has hindered the analysis of glycan data in order to identify glycoforms contributing to disease . To overcome this limitation , we developed a quantitative N-glycosylation model that interprets and integrates mass spectral and transcriptomic data by incorporating key glycosylation enzyme activities . Using the cancer progression model of androgen-dependent to androgen-independent Lymph Node Carcinoma of the Prostate ( LNCaP ) cells , the N-glycosylation model identified and quantified glycan structural details not typically derived from single-stage mass spectral or gene expression data . Differences between the cell types uncovered include increases in H ( II ) and Ley epitopes , corresponding to greater activity of α2-Fuc-transferase ( FUT1 ) in the androgen-independent cells . The model further elucidated limitations in the two analytical platforms including a defect in the microarray for detecting the GnTV ( MGAT5 ) enzyme . Our results demonstrate the potential of systems glycobiology tools for elucidating key glycan biomarkers and potential therapeutic targets . The integration of multiple data sets represents an important application of systems biology for understanding complex cellular processes .
Glycosylation , a broad term covering the addition of oligosaccharides ( glycans ) to proteins and lipids followed by their subsequent modification during transit through the secretory apparatus , is an intricate intracellular process whose complexity hinders ready interpretation from mass spectral and other data sets . Nonetheless , three decades of research has made it clear that the glycosylation of healthy and diseased cells often diverges resulting in glycan changes that contribute to pathological progression [1] , [2] , [3] , [4] , [5] . A prime example of the contribution of glycan analysis to the understanding of a pathological process and the development of clinically relevant biomarkers is provided by prostate specific antigen ( PSA ) [6] , [7] , [8] , [9] , [10] . Changes in the glycosylation status of this widely used biomarker for prostate cancer screening have been useful in improving its specificity and ability to distinguish benign forms of this disease from highly malignant cancer [11] , [12] . While considerable progress has been made from decades of painstaking research focused on PSA , efforts to identify additional glycan markers of disease suffer from the difficulties in identifying specific glycosylation changes . However , with the current proliferation of high throughput ‘omics’ approaches , opportunities are at hand to develop and implement methodologies that analyze the resulting large data sets in order to provide critical glycan signatures of disease; for example to expand analyses from PSA to additional prostate cancer biomarkers and , more broadly , from prostate cancer to the numerous cancers and diseases known to have abnormalities in glycosylation . Unfortunately , the disparate sets of data needed to fully characterize glycosylation –including expression profiles of the enzymes involved in glycosylation , the activities of the resulting enzymes , and finally the large number of glycans actually produced by these enzymes – cannot be directly compared and there is yet no facile way to integrate the data to generate meaningful biological insights . Transcriptional profiling of mRNA allows quantitative global assessment of the many genes involved in glycan biosynthesis i . e . glycosyltransferases , the enzymes responsible for generating glycans . A wealth of data is also becoming available from the detailed assessment of the glycans using mass spectrometric techniques [13] . Despite progress on both ‘omics’ fronts , useful bioinformatics tools to identify glycan structural data and also to link these findings with transcriptional profiles of the enzymes that produce these sugars have lagged . For example , a common approach for mass spectrometry-based glycoprofiling involves a one-to-one data base matching of particular mass spectrometry measurements to specific glycans from a known glycan library in order to annotate the mass spectra [14] , [15] . Statistical database-driven approaches have attempted to relate gene expression levels to the abundance of specific glycan linkages [16] , [17] , [18]; however these approaches do not provide quantitative predictions of detailed glycan distributions . As a consequence , there is no clear understanding of how mRNA levels relate to the actual amount and distribution of glycans found within a healthy or diseased cell . In addition , current bioinformatic techniques only consider each mass spectral peak in isolation and do not consider other relevant peaks when making an identification or quantification . In this work , we address this void with a novel systems biology model that interconnects glycan structural data obtained from mass spectrometry with changes in gene expression obtained from mRNA profiling of the relevant glycan processing enzymes . The glycobioinformatics approach interprets mass spectral data in terms of the activity of glycan-processing enzymes and then compares these values to those indicated from gene expression profile . The model identifies a number of unrecognized glycan structures and their abundances from mass spectral peaks by analyzing the entire mass spectrum in concert instead of considering each mass peak in isolation . The model can also process the relative enzyme transcript gene expression levels , and translates them into a synthetic mass spectrum and a quantitative glycan profile . This model approach has been applied to uncover subtle differences in the glycan signatures between two sets of cancer cells , specifically low and high passage , androgen dependent and independent ( respectively ) LNCaP prostate cancer cells . This effort has yielded insights into glycan-specific changes associated with malignant progression in this disease . In addition , this model approach enables a comparison of the result from the two ‘omics’ platforms and enables identification of consistent and inconsistent patterns across the two media . Moreover , this systems biology methodology allows users to gain insights into the complex multi-step cellular glycosylation process from disparate data sets and will serve as a critical step along the path towards the identification of key glycan biomarkers and therapeutic disease targets .
In previous publications we applied a comprehensive mathematical model that incorporates a kinetic network for enzyme processing of N-glycans to interpret mass spectral and other glycan analytical data ( HPLC ) in terms of detailed glycan structures as well as specific enzyme activities [19] , [20] . This analysis was useful for screening differences in glycan profiles and enzyme activities between different cell types . In this study we present an integrative glycan systems modeling approach that considers mRNA gene expression profiles for the glycosyltransferases and other enzymes involved in glycan synthesis together with matching MALDI TOF ( Matrix assisted laser desorption ionization time of flight ) mass spectral data . This data integrative modeling approach provides a thorough characterization of the changes in the glycan structural profile and abundances through the mass spectra . Model sizes used in this study are typically limited to about 10 , 000 to 25 , 000 glycan structures based on the implementation of a molecular mass cutoff and a network pruning method . This allows prediction of the complete glycan profile and its abundances for any set of assumed enzyme concentrations and reaction rate parameters . A schematic representation and explanation of how the model integration of mass spectrometric and gene expression data works is shown in Figure 1 ( for more details see Materials and Methods ) . High and low passage LNCaP cells provide a model for cancer progression from the androgen-dependent to the androgen-independent state [21] . The MALDI TOF mass spectrometry data for the low and high passage human prostate LNCaP cells are available at the Consortium of Functional Glycomics ( CFG ) database [22] and under supplementary material in Dataset S1 and Dataset S2 . The C-33 cells , or low passage cells , include cells between passages 25 and 35 and serve as a model for androgen-dependent cells . The C-81 cells or high passage type were derived from the low passage cell line and have diverged into an androgen non-responsive state; they include cells between passages 81 and 125 [21] . The comparison of model generated synthetic mass spectra to experimental MALDI TOF mass spectra ( Figure 2 and Figure 3 ) requires processing of the raw mass spectra , including baseline correction , mass calibration adjustment , peak integration and filtering of isolated spikes ( individual peaks without isotopic satellites ) , software development for that end is described in Materials and Methods . Fitting of both MALDI TOF experimental data sets to synthetic mass spectra obtained through solving our N-glycosylation computational model as described in Figure 1 generates a set of glycan structures and abundances . The parity plot in Figure 2 gives the agreement of the calculated and measured experimental mass spectrometric data in the range of 1400–4000 for both high and low passage LNCaP cells . Peaks in agreement are located on the parity line; in general , a good fit is obtained for many of the glycans . The experimental mass spectrum extends to 5000; however , for this set of data few molecular masses are significant in the 4000 to 5000 range and thus the model was limited to the 4000 range . This model approach can be readily implemented to assign a group of glycans with specific details on their associated structures and abundances to each peak in the MALDI TOF mass spectra of diverse mammalian cell types . For example , the good agreement obtained between the measured and synthetic mass spectra obtained from the model as indicated in Figure 2 for both LNCaP cell lines , is translated in Figure 3 as a close alignment of the experimental mass spectral levels ( blue line ) with the model predictions ( red line ) for most of the peaks . Overall , in Figure 3 , we present a selected portion of the mass spectra for the low and high passage cell lines in the range of 2150 to 2750 with the dominant glycan structures producing each peak indicated by schematic structural diagrams . The comparative glycoprofiling of both cell lines is discussed in the next section and the complete set of glycans annotated across the 1500–4000 MS range is provided in Figure S1 . High and low passage LNCaP cells that provide a model of cancer progression from the androgen-dependent to androgen-independent state exhibited considerable similarity in N-glycan patterns ( Figure 3 ) ; this result was expected because both cell lines come from the same progenitor and only differ in the number of passages . However , the comparative glycoprofiling shown in Figure 3 establishes that there are also several differences between the two cell lines . For example , at the peak envelope starting at a monoisotopic mass of 2418 . 21 ( the lowest mass of the isotope group ) , the high passage cells have about twice the signal of the low passage cells , partially due to the appearance of a structure containing the blood group H ( II ) epitope ( Fucα1 , 2Galβ1 , 4GlcNAcβ ) in the high passage glycans ( See Figure 4 for N-glycan processing diagram ) . Another peak envelope starting at a mass of 2592 . 30 is about four times higher for the high-passage cells , due to the increased abundance of terminal fucose groups ( Fucα1 , 2Galβ ) . Predictions of the abundances of glycans based on the model analysis of the complete mass spectral data is shown in terms of types of glycans in Figure 5 and glycan moieties associated into blood group categories in Figure 6 . The mass spectra for both high and low passage prostate cancer LNCaP cells are most abundant in high mannose glycoforms ( Figure 5 ) . Indeed , prostate-specific membrane antigen ( PMSA ) protein derived from LNCaP cells was found to contain high mannose structures [23] which indicates the generation of these structures from this cell line . In addition , a series of complex glycans with tetraantennary structures being the most abundant followed by biantennary and triantennary glycans are observed in both low and high passage cell lines ( Figure 5 ) and were reported in [24] . Here we see that the more metastatic high passage cells have higher levels of hybrid glycans and lower levels of complex glycans , especially tetraantennary . Nearly 70%–80% of the hybrid and complex glycan structures ( % based on hybrid and complex not total glycans ) in both low and high passage LNCaP cells are core fucosylated ( Figure 5 ) . Indeed , previous studies have reported core fucosylated glycan structures as characteristic of PSA from LNCaP cells [8] , [10] , [25] . Complex glycans are mostly identified as non sialic acid capped complexes of the bi , tri and tetraantennary type . However , limited levels of monosialylated , bisected glycans structures and lactosamine repeats ( Galβ1–4GlcNAc ) are predicted . Glycan moieties of type II are predominant in both cell lines ( Figure 5 and Figure 4 ) . Note that a single glycan can contain more than one of a particular glycan moiety , so the abundances of some moieties can exceed 100% of the total number of glycans , used as the basis for this percentage . In terms of blood group structures which include antigens A , B , Lea , Leb , Ley , Lex and H in Figure 4 , the H ( II ) and the Ley epitopes , containing α1 , 2-fucose linkages , are predicted in these types of prostate cancer cells , with greater abundance in high passage LNCaP cells ( Figure 5 ) . Indeed , these epitopes have been reported as characteristic markers for prostate cancer [7] , [10] , [26] , [27] , [28] , [29] , [30] . The concentrations of the glycan-processing enzymes were adjusted in the computational model until satisfactory agreement was obtained between the in silico mass spectral profile with the experimental mass spectral profile for both data sets . Shown in Table 1 ( columns 4 and 5 ) are the adjusted enzyme activities that provided the best fit with both experimental mass spectral data sets and produced the glycan abundances in Figure 5 and 6 . A comparison of the observed changes in the enzyme levels between high and low passage LNCaP cells provides a succinct way of interpreting the differences in glycan structural profiles between these LNCaP cells lines . In general we observe that model predicted enzyme activities are in qualitative agreement with available published enzyme values in terms of increasing and decreasing levels between cases . A description of some of the trends that were observed in the enzymatic activities corresponding to the resulting glycan profiles for the two LNCaP cell lines are described in following sections . Mapping of model enzymes Table 1 ( column 1 ) to gene probes ( columns 6 and 7 ) on the Consortium of Functional Glycomics ( CFG ) Glycogene microarray ( Glycochip version 3 , CFG ) was performed for both low and high passage human prostate LNCaP cell line data available at the Consortium of Functional Glycomics website ( MAEXP_291_040606 ) and also included in Dataset S3 [22] . Listed in columns 8 and 9 of Table 1 are the observed changes in expression levels of these genes as determined from mRNA analysis of microarrays for the low passage and high passage LNCaP cells . These expression signals were obtained by averaging three replicate samples for each glycogene in the microarray and represent the average relative abundance of a transcript . The glycogenes were assigned calls by CFG of present ( P ) , marginal ( M ) or absent ( A ) ( more information on the classification criteria is found in Materials and Methods ) . In Table 1 the marginal or absent calls are indicated with numbers in bold . Note that more than one gene can encode the same type of enzyme activity . In addition to the genes for N-Glycan processing as discussed for Table 1 , the glycochip version 3 includes genes for many other glycosylation-related genes . However , in some cases not all the genes that encode for a given enzyme are included on the CFG version 3 microarray . Interestingly , the largest shifts in gene expression observed between the two types of prostate cancer cell types from the glycochip are those that encode for the enzyme glucuronosyltransferase ( EC 2 . 4 . 1 . 17 ) , which is involved in androgen/estrogen metabolism but has no effect on N-glycan structure [36] , [37] , [38] . In general transcript expression levels from the microarray are consistent with the enzyme activities resulting from the model , at least in terms of the presence or absence of enzymes in the model and on the gene expression array . Enzymes whose genes are classified as marginal or absent in the microarray data include GnTIII , a6SiaT , GalTB and GalNacT-A ( Table 1 columns 8 and 9 in bold ) . Interestingly , enzymatic activity levels for this same collection of enzymes were also predicted by the model to be low or zero independently based on fits of the mass spectral data ( Table 1 columns 4 and 5 ) . The most significant percentage shift in enzyme activity that differentiates high and low passage LNCaP cells was increased FucTH activity in high passage or androgen independent LNCaP cells . Although the model based solely on mass spectrometry measurements deduced this finding of increased FucTH activity , the uptick in FucTH expression was also correlated with mRNA microarray expression data for the FUT1 gene ( Table 1 columns 8 and 9 ) . The FUT1 gene has been experimentally identified in LNCaP cells and prostate cancer tissues as responsible for production of the H ( II ) epitope [10] , [28] . In addition model-fitted mass spectra data indicate a lower capacity for generating type I glycans relative to type II glycans as b3GalT activity is lower than b4GalT activity in Table 1 . This finding also correlated with the relative gene expression data in that the expression of b3GalT genes encoded by B3GALT1 , B3GALT2 , and B3GALT5 are marginal in both data sets from both cell lines ( Table 1 columns 8 and 9 ) . In contrast , gene expression levels for the B4GALT1 , B4GALT2 , B4GALT3 and B4GALT5 ( Table 1 columns 8 and 9 ) encoding b4GalT activity for type II glycans is robust in both cell lines as it is the generation of type II glycans as indicated by the model fitting of mass spectral data and predicted b4GalT enzymatic activity . While there were many consistencies between model interpretation of mass spectrometry and gene expression profiles , there were also some disagreements between the model-calculated enzyme levels and the expression levels obtained from mRNA , such as for GnTV , a3SialT and FucTLe . For some enzymes , these differences can be attributed to shortcomings in the mRNA profiling chips . A case in point is that the inactivity of a probe set for MGAT5 ( one of two genes encoding GnTV activity ) in the glycochip version 3 from CFG led to a reported lack of expression of this gene despite the presence of GnTV activity in the results modeled from the mass spectral data ( Table 1 column 4 and 5 ) . Interestingly , by reviewing all experimental cases run with the glycochip version 3 for all cell lines posted on the CFG we found that the probe set for MGAT5 gene was inactive . Furthermore , the presence of GnTV indicated by the model interpretations of the mass spectral data agrees with a previous experimental study that reports positive GnTV enzymatic activity using a zymography assay and active expression of MGAT5 through RT-PCR in LNCaP cell lines [32] . Results from model testing of this proposed deficiency in the glycochip version 3 are discussed in Text S2 , Figure S2 , and Figure S3 . These findings support the interpretation that the inactivity of the GnTV in the microarray data is due to a defect in this specific probe on the microarray . For other enzymes , the lack of model agreement with gene expression data is likely due to the scope of the current model . For example , the microarray data indicates significant levels of α3SiaT mRNA for both LNCaP cell passages , although the enzymatic model interpretation from MALDI TOF indicated that LNCaP cells have low a3SialT activity . Interestingly , analysis of expression of genes associated with sialic acid biosynthesis , a feature not included in the current model , indicates that the transcript levels for the GNE gene ( that encodes the bifunctional UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase ) for both cell lines are interpreted as absent . Bifunctional GNE catalyzes two critical steps involving sequential reactions in the biosynthesis of the sugar nucleotide CMP-Neu5Ac ( CMP-sialic acid ) ( Figure S4 , Text S3 ) , which is the co-substrate for a3SiaT . The α2 , 3- or α2 , 6-linked sialylated N-glycans are generated by the transfer of the sialic acid ( Neu5Ac ) group from the nucleotide donor sugar CMP-Neu5Ac onto the oligosaccharide acceptor ending in a galactose ( Gal ) residue . Thus , the formation of a limited number of sialylated glycoforms may be due to low levels of the nucleotide sugar substrate , CMP-Neu5Ac , rather than due to a limitation in the a3SiaT activity as currently manifested in the model . While the current model has focused on glycosyltransferase activities , it can be readily expanded to include other metabolic reactions such as the generation of sugar nucleotides including CMP-Neu5Ac and others . For the FucTLe enzyme , the model predictions show lack of activity , while the microarray data shows gene expression signal for FucTLe . Both FucTLe and a3FucT enzymes catalyze the addition of α-3 fucose to GlcNAc residues on type II chains as shown in Figure 4 . They only differ on Type I chains , in the generation of Lea and Leb structures . The lack of these Le structures makes it difficult for the model to separate the two enzyme activities and at least part of the activity the model shows for a3FucT may be due to FucTLe . As an alternative predictive approach , the potential also exists for using gene expression profiles to estimate changes in enzyme activities . Appreciating that mRNA levels do not always reflect enzyme levels directly , we assumed shifts in gene expression data were related to shifts in enzyme levels . These shifts in gene expression were then used to predict the corresponding shifts in glycan profile and abundances for high and low passage LNCaP cells . To get the transcript expression level of the enzymes in the model , the average of all mRNA-microarray gene signals corresponding to each enzyme was used . Next the ratios of average signals ( high/low passages ) for each enzyme were used as inputs to the N-glycosylation model based on the assumption that they are estimates of the relative enzyme levels for high and low passage LNCaP cells . The implementation of this methodology into the model also required adjusting the enzyme levels in the model to match the experimental mass spectra data for one case ( low passage for this study ) while maintaining the enzyme activity ratios to be equal to the experimental values obtained from the microarray data ( Table S1 ) . The ratios of expression levels of high to low passage cells were then used to predict the enzyme levels for the high passage case and the resulting glycan profile predicted . No constraint was placed on the concentrations of GnTV or ManI in the model as these enzymes have either a probe defect ( GnTV ) or are missing in the glycochip version 3 ( Man I ) . The results of keeping the enzyme ratios constant and equal to the microarray ratios are shown in Figure 7 , which show the abundances of different categories of glycans . The percentage of different structures predicted by the model using MALDI-TOF mass spectral data alone ( green bars ) are compared to the abundances obtained after fixing the gene expression mRNA ratios ( orange bars ) . Predictions for the low passage , androgen dependent LNCaP cells are indicated by light green bars for mass spectral data and light orange bars for mRNA data . Similarly , predictions for high passage , androgen independent LNCaP cells are indicated by dark green bars for mass spectra data and dark orange bars for mRNA data . This approach allows the model to predict changes in glycan structure profile and glycan abundances based on comparative N- glycosylation enzyme gene expression data . Although constraining model fitting to fixed gene expression ratios resulted in a higher RMS error ( lower model agreement ) with respect to the model fitting to the mass spectra data alone ( as would be expected for additional model constraints ) , the trends in glycan structures and abundances are comparable in both cases with a few minor exceptions . In general , glycan structure prediction from both data sets ( gene expression and mass spectra vs . mass spectra alone ) show consistency in terms of presence , absence and increased or decreased abundances of glycans in high passage with respect to low passage LNCaP cells . For example , both model predictions with mass spectra and gene expression data suggest an increase in abundance of hybrid and biantennary structures in the high passage cells and a corresponding decrease in the tetraantennary structures . Both mRNA data and models predict high levels of type II chains in both low and high passage LNCaP cells , reaffirming that these cell lines contain predominantly type II glycans . The model based on gene expression data suggested lack of type I glycans , principally due to imposing a strict restriction on b3GalT which in reality may not be as strict as the mRNA data for the genes encoding this enzyme are classified as either marginal or absent .
In this study , a systems biology computational model that connects diverse experimental data sets was used to evaluate N-glycan data from MALDI-TOF mass spectra and mRNA expression arrays for androgen independent , high passage LNCaP cells , and androgen dependent , low passage , LNCaP cells . Most significantly , insights into the N-glycosylation processing for LNCaP high and low passage cells were found based on model predictions of enzyme activities , glycan structures and gene expression profiles . The model was also useful for identifying consistencies as well as incongruities between glycan structural information and gene expression data . The N-glycosylation model identified and quantified glycan structural details not typically derived from single-stage mass spectral or gene expression data , such as the type of fucosylation ( Fuc-α1 , 2-Gal vs . Fuc-α1 , 3-GlcNAc and Fuc-α1 , 6-GlcNAc ) , the predominance of Type II chains ( Gal-β1 , 4-GlcNAc ) versus Type I chains ( Gal-β1 , 3-GlcNAc ) or the number of antennae . This is possible by analyzing the total mass spectrum in terms of the underlying processing events and enzyme activities that generate both the individual structures and the assemblage of structures resulting in the complete mass spectrum . For example , Fuc-α1 , 2-Gal can be differentiated from Fuc-α1 , 3-GlcNAc because they include the different molecular weight linked sugars of Gal and GlcNAc . More relevant to the current modeling approach , Fuc-α1 , 3-GlcNAc and Fuc-α1 , 6-GlcNAc can be differentiated because , even though both include the same molecular weight linkages , Fuc-α1 , 6-GlcNAc is added to the glycan core early in processing , while the Fuc-α1 , 3 is only added later to GlcNAc on one of the N-glycan branch extensions . The presence of a collection of mass spec peaks provides a fingerprint that the model can interpret to indicate whether this fucose is added earlier ( Fuc-α1 , 6-GlcNAc ) or later ( Fuc-α1 , 3-GlcNAc ) in the N-glycan processing pathway . Thus the model creates a picture of the complete N-glycan processing , including enzyme activity levels acting sequentially through the secretory apparatus that is consistent with the entire collection of glycan peaks across the molecular weight spectrum . Comparison of the underlying enzyme activities derived by the model from the mass spectral data with changes in gene expression levels measured by the CFG glycogene microarray show them to be consistent for most enzymes , which tends to verify that the model-derived shifts are meaningful . The agreement also suggests that the model could be used to predict shifts in glycan structure from a well-defined base case based on changes in microarray expression data when MS data for other cases is not available . It is important to note that aberrant N- and O-glycosylation are important features of cancer cells . Indeed a number of the genes included in the model act on both N-glycans and O-glycans , while the model infers changes in the total enzyme activities based only on the observed shifts in N-glycan structures . Thus if the levels of O-glycan structures that compete with N-glycan structures for a number of enzymes changes significantly , the fraction of those enzymes activities that are available for N-glycan processing could also change , distorting the comparison between model predicted enzyme activities and enzyme gene expression levels . While our N-glycosylation model gave very good agreement between the model-predicted and measured mass spectra , it is expected that the incorporation of O-glycosylation together with N-glycosylation will improve the model predictability . In addition , these competing reactions will be better modulated in the model . Implementation of O-glycosylation in the current model framework is possible since kinetic parameters for the corresponding O-glycan enzymes as well as experimental data are available to tune the model . Of course , processing larger data sets including O-glycans may very well elucidate limitations in the model that will need to be addressed through appropriate modification of model parameters . Interestingly , the most significant difference found between high and low passage prostate cancer cell lines was the increase in expression of α1 , 2-Fuc-transferase ( FucTH ) enzyme in the high passage LNCaP cells , as predicted by the model based on mass spectrometry and verified by the gene expression data . The microarray data indicates that the FUT1 gene is predominant in high passage LNCaP cells with respect to low passage LNCaP cells . The FUT1 gene has been experimentally identified in LNCaP cells and prostate cancer tissues and associated to the H ( II ) epitope ( Fucα1 , 2Galβ1–4GlcNAc ) [28] . Moreover , the presence of Fucα1-2Gal residues that results from the enzymatic action of FuTH has been reported in PSA from LNCaP cells [10] , [28] , [31] , [35] . The high passage LNCaP cells in this work were obtained from low passage LNCaP cells ( androgen dependent ) after successive passages and they have diverged into an androgen independent state . Our finding that high passage LNCaP prostate cancer cells ( androgen independent ) have increased levels of the enzyme FucTH ( FUT1 gene ) responsible for α1 , 2-fucosylation and the H ( II ) and Ley epitopes with respect to low passage LNCaP cells may represent a potential marker of higher malignancy or androgen refractory prostate cancer cells and may be relevant in diagnosing prostate cancer stage . For example it may be possible to compare glycan mass spectra of PSA concentrated from blood serum , presumably originating in cancer cells , to PSA from semen samples , derived mostly from normal prostate cells , to evaluate the stage of the cancer . Our results also indicate that b4GalT is present in both high and low passage LNCaP cells . The most expressed member of the family is the B4GALT1 gene followed by the B4GALT3 gene . Evidence of b4GalT in prostatic cancer samples is found by detection of Galβ1 , 4GlcNAc ( Type II ) structure with the Erithrina cristagalli lectin ( ECL ) in PSA from prostate cancer serum and PSA from LNCaP medium as compared to seminal plasma ( normal control ) [25] . This type II structure was also detected with a set of lectin-immobilized columns together with enzyme-linked immunosorbent assays ( ELISA ) on prostate cancer serum PSA and LNCaP cells PSA as compared to benign prostate hyperplasia ( BHP ) serum PSA [31] . Interestingly , screened experimental data on prostate cancer predominantly reports the presence of Galβ1 , 4GlcNAc ( Type II ) glycans and some of its derivatives and almost no information is found for Galβ1 , 3GlcNAc ( type I ) glycans . In agreement with that , the model predicted the presence of H type II ( Fuc1 , 2Galβ1 , 4GlcNAc ) and Ley- ( Fuc1 , 2Galβ1 , 4GlcNAc Fuc1-3 ) glycans and also increased levels of these epitopes in the more metastatic high passage LNCaP cells . Importantly , several previous studies have reported type II based epitopes H [10] , [26] , [27] , [28] and Ley as blood group antigens as characteristic of prostate cancer [7] , [29] , [30] , [39] . For example , lectin histochemistry comparisons between normal human prostate and prostatic carcinoma tissues show increased expression of galactose ( using DSA lectin suggesting presence of Galβ1 , 4GlcNAc ) , and fucose [6] ( using UEA-I , a marker for the H antigen ) . Also , investigations of PSA serum from 40 patients revealed an increase in the glycans containing Fucα1 , 2Galβ1 , 4GlcNAc and GalNAcβ1 , 4GlcNAc for patients with prostate cancer as opposed to those with benign prostatic hyperplasia ( BPH ) [31] . Moreover the production of the H ( II ) epitope ( Fuc1 , 2Galβ1–4GlcNAc ) has been associated with the potential for carcinogenic cell proliferation ( Marker et al , 2001 ) . Most importantly , these findings reflect trends predicted by both gene expression data and mass spectral data . A corollary glycan signature predicted by the model from mass spectra is lower relative levels of the b3GalT enzyme , which was even more pronounced in the gene expression data as indicated by the low levels of transcript signals for the genes encoding for b3GalT . In addition to the lower abundance of type I glycans in both cell lines , derivatives including Lea and Leb epitopes were also absent . These observations are in agreement with studies reporting low levels or the complete absence of type I based antigens Lea and Leb [34] in prostatic carcinoma . Additionally , the A and B blood group antigens from type I and Type II glycans , were predicted absent or minimal in agreement with [26] , [27] . Moreover , the comparisons of model predictions from glycan structural data with gene expression findings pointed to deficiencies in the mRNA microarray , such as a lack of a sensitive probe for the MGAT5 gene for GnTV . This was further confirmed by computationally suppressing GnTV enzyme activity and demonstrating that the modified model could not regenerate the experimental mass spectra ( Text S2 , Figures S2 and S3 ) . Indeed , the presence of GnTV indicated by the computational model agrees with a previous experimental study that reports GnTV enzymatic activity using zymography assay together with detection of expressed MGAT5 using RT-PCR in LNCaP cell lines [32] . In summary , this study demonstrates the potential of systems glycobiology approaches as means to connect and interpret disparate data sets obtained with widely different experimental methods , in this case mass spectral data and gene expression profiles . The resulting model approach allows users to better understand N-glycosylation processing events in a prostate carcinoma cell line and also helps to define consistent patterns and incongruities between data obtained from mass spectrometry and microarrays . Consistent patterns observed in data sets from multiple methods may represent potential glycan biomarkers for cancer and other diseases . Furthermore , by using these computational tools , we have been able to show how changes in the mRNA data can be used to describe glycan patterns in low and high passage LNCaP cells . In this way , the model can increase the value of current mRNA profiling data as a useful tool for indicating changes in glycan processing . This effort will contribute significantly to the current need for bioinformatics and systems biology tools in glycobiology . The method allows users to interpret , integrate and compare multiple complex data sets in order to identify and validate critical biomarkers involving N-glycosylation processing in normal and diseased cells and tissues .
The model has been extensively validated with several public available mass spectra data sets ( http://www . functionalglycomics . org ) of mammalian human and CHO cells . Methods demonstrated in previous publications of the model ( Krambeck 2005 [20] , and Krambeck 2009 [19] ) with other experimental data sets enable using the model comparatively between a control case and other case/s . These methods have the advantage of allowing a common adjustment to the model for two or more cases while using only the enzyme concentration levels to differentiate the samples . This principle limits case-to-case variations as much as possible to just the enzyme concentrations while holding almost all other model parameters to uniform values for all cases . In this respect the accuracy of this approach depends on how sensitive the predictions are to the assumed values of unknown parameters in the model . For example sensitivity analysis on the effect of total glycan concentration on the enzyme levels shows that the predicted effects of changes in enzyme concentration drive similar shifts in glycan profiles for different total glycan concentrations . Similar studies have been applied to analyze sensitivity effects to assumed values of other parameters . As the range of experimental data the model can accommodate expands , the more robust and reasonable the model results will become . Thus establishing a wide database of analyzed glycans from numerous cells and tissues is essential to improving model robustness . In the current model framework , glycan structures are expressed using a condensed version of IUPAC linear formulas [40] with some minor modifications . The first modification is to order the branches at a branch point based just on the branch locants ( the carbon atom numbers of the attachment points of each branch ) without regard to the lengths of the branches as is used by the IUPAC scheme . In addition the sugar abbreviations have been replaced with the shorter abbreviations of the Linear Code [41] , but we have not used the complicated branch ordering scheme of the Linear Code . Figure 8 shows the sugar symbols used in glycan structures for our model as well as an example of the condensed linear formulas used to represent glycan structures for a 9 -mannose glycan . This scheme provides linear formulas that are general , are easily readable by humans , are unique for each glycan structure , and allows the model to apply to N-glycans , O-glycans and glycolipids . Table 2 gives a list of the enzymes included in the current model and the set of reaction rules for each enzyme . These are sufficient to produce most of the N-glycans present in human cells . The basic idea is that the “Substrate” column is a substring of the linear formula that must be present for the enzyme to act . The “Product” column specifies what the substrate string is replaced with through action of the enzyme . The “Constraint” column specifies a set of additional conditions that must be satisfied for the enzyme to act . These conditions are usually the presence or absence of another substring somewhere in the substrate formula . These are combined using the “not” operator ( ∼ ) , the “and” operator ( & ) and the “or” operator ( or ) , with parentheses as appropriate . To simplify these expressions a number of additional conventions have been added . All substrate formulas are assumed to be enclosed in parentheses before searching for the substrate substring . Thus an initial “ ( ” always indicates the terminal end of a branch . Other codes and abbreviations used in formulating the reaction rules are summarized in Table 3 . ( See example in Table S2 and further detail in Text S1 ) . The model kinetic reaction network ( Table S3 ) is generated by a series of string searches and substitutions which begin with a list of starting structures ( 9 and 8 -mannose glycans e . g . Figure 8 ) . Each substrate rule and corresponding constraint rule is then applied to each structure in the list to determine which structures are substrates for each rule . For structures that satisfy the rules , the product structure is determined , essentially by substituting the product substring of Table 2 for the substrate substring , taking the various abbreviations into account . If the structure is not already in the list of structures it is added to the list . At the same time , the new reaction is added to a reaction list . The reaction list includes the enzyme , substrate , cosubstrate , product and coproduct strings for each reaction . This process is repeated until no new reactions can be generated . A molecular mass cutoff to limit the size of the glycans generated to those observable in the mass spectral data is included . To reduce the size of the model further a network pruning method was used based on roughly estimating the abundances of the structures and dropping those structures of negligible abundance . Starting with the 9-mannose glycan shown in Figure 8 , and including an inert structure with an additional glucose residue ( Ga3 ) , the rules of Table 2 generate a reaction network containing 10 , 809 structures and 28 , 797 reactions . The maximum mass cutoff used to generate this network was 4000 on a permethylated basis and network pruning was enabled . Consider the glycosylation of a glycan Pi with a monosaccharide S catalyzed by an enzyme E . Assuming that the donor cosubstrate is UDP-S , ( µM ) the overall reaction is shown in equation ( 1 ) : ( 1 ) Assuming that the product Pi+1 competes for the same enzyme site as the substrate Pi ( µM ) , that the donor cosubstrate UDP-S occupies a second site on the enzyme , and that the reaction is reversible , the Michaelis-Menten kinetic equation takes the form shown in equation ( 2 ) . ( 2 ) Here [Et] is the Enzyme concentration , µM , kf ( min−1 ) and kr ( min−1 µM−1 ) are the forward and reverse rate coefficients , Kmi and Kmd are the dissociation constants for the substrate and donor cosubstrate in µM , and is the apparent equilibrium constant for the overall reaction . The symbols in equation ( 2 ) denote equilibrium concentrations . The subscript “j” in the summation in the denominator is taken over all the substrates that compete for the same enzyme . A derivation for equation ( 2 ) is given in the KB2005 model [20] . The values of the kinetic parameters kf , Km and Kmd for a given enzyme can vary significantly for different substrates . This was accommodated by selecting base values for these parameters for each reaction rule and adding a set of structure-dependent adjustment rules . Development of these parameter values and adjustments for CHO and human cell enzymes are detailed in [19] , [20] . The base parameter values currently used for each of the reaction rules in Table 2 are shown in Table 4 . Adjustment rules for the parameters are given in Table 5 . Each adjustment rule includes a condition on the substrate structure and multipliers to apply to each of the three parameters whenever the condition is satisfied . The Golgi compartments were modeled as well-mixed reactors with bulk flow of the contents from each compartment to the next compartment in line while the enzymes remain fixed in the compartments . At steady state the concentrations of the various structures satisfy the balance equations . ( 3 ) where cij is the concentration of structure i in compartment j , τj is the residence time of compartment j , and rij is the net rate of production of structure i per unit volume of the compartment due to all the biochemical reactions that occur . The total concentration of N-glycans , ctot , is the same in each compartment and is given by pτ1/v1 , where p is the total production rate of glycans and v1 is the volume of compartment 1 . In the balance equations for the first compartment the concentration ci0 is given by the fraction of total glycans that initially have structure i multiplied by ctot . Using the above Michaelis-Menten kinetics for the glycosylation reactions , equations were derived to solve for the concentrations of each of the individual glycan structures in each of the Golgi compartments . The model equations are nonlinear algebraic equations which are solved for the concentrations of each of the structures in each of the four Golgi compartments . These are solved using a constrained Newton-Raphson method with the Harwell MA28 sparse linear solver ( HSL 2002 ) . The efficiency of a sparse linear solver for large numbers of variables depends on the problem Jacobian being sparse . The Michaelis-Menten denominator terms in Equation ( 2 ) involve a large number of species that compete for each of the enzymes . This could make the Jacobian matrix rather dense . To avoid this , the denominator terms for each enzyme are formulated as separate variables with equations added to specify how the denominators are calculated . This confines the equations with large numbers of variables to only one for each enzyme . Analytical derivatives were used . While each compartment could be solved separately in sequence to give four subproblems , each one fourth the size , this was not found to be necessary . In addition to solving the model for a given set of model parameters , provision was also made to adjust parameters to match a given set of data . This was done using the Marquardt-Levenberg method with analytical derivatives [42] . The same method was used for optimizing model parameters to achieve a desired distribution of glycan structures . The Marquardt-Levenberg method is typical of nonlinear optimization algorithms in that it makes use of a sequence of local linear approximations to the nonlinear model to converge to a solution that is a local optimum . Except in special cases there is no way to determine whether the nonlinear problem possesses an even better solution far removed from this point . Experience in using this method shows , however , that if a reasonably good fit is obtained with the local optimum it is also a global optimum for the parameter estimation problem . Robust solution methods were devised to allow simultaneous solution of the approximately 45 , 000 nonlinear equations for the concentration of each of the glycan structures in each of the four compartments of the model . Other parameters needed for the calculations , include compartment residence times , enzyme distributions between compartments , compartment volumes , total glycan concentration , and donor cosubstrate concentrations . These were estimated based on a variety of literature sources as detailed in our previous publications [19] , [20] . It should be emphasized that these numbers are intended to be reasonable initial estimates subject to further refinement . Several steps are involved in generating the synthetic spectrum: The most significant isotope peaks for each glycan ( those amounting to at least 10−6 of the total for the glycan ) are calculated and stored in a database . The experimental MALDI mass spectra require processing before comparison with the synthetic mass spectra through baseline correction , mass calibration adjustment and peak integration . The baseline correction method was adapted from Williams et al . ( 2005 ) [43] . The mass calibration was done by finding the linear mass adjustment that maximizes the sum of the experimental peaks interpolated to the theoretical masses of the model-predicted glycan peaks . An approximate area for each peak in the baseline-corrected and mass-calibrated spectrum was calculated as follows: First the nearest local maximum to the theoretical mass for each peak was determined to give a “peak height” . Then a “peak width” was determined for the 50 largest peaks by finding the point on either side of the maximum with an intensity of exp ( -π/4 ) ( or 45 . 6% ) of the peak height . Note that multiplying this peak width by the peak height would give the exact area of a Guassian peak and also approximates the area of a skewed peak , such as a relatively narrow gamma distribution . The peak widths for the largest peaks so determined are then correlated as a linear function of peak molecular mass to accommodate the broadening of mass spectrometer peaks with increasing mass . The linear correlation of peak width vs . peak molecular mass is then used to calculate a peak width for every peak in the spectrum . The calculated peak width is multiplied by the peak height to estimate peak area . These peak areas are then normalized to add up to 100% . Examples of processed experimental spectra and calculated synthetic spectra are shown in Figure 3 and in Figure S1 . The points on this plot are the area of each peak plotted against the mass at the peak maximum . Thus the curves on the plots are isotope envelopes . After processing the experimental mass spectra still contain a significant number of minor peaks ( actually isotopic satellite groups of peaks ) , which do not correspond to any glycans in the model . In most cases they do not correspond to any known N-glycans . Presumably these are artifacts of the sample processing , perhaps fragments produced in the mass spectrometer . In any event to avoid confounding of the model parameter adjustment step the preprocessed experimental spectra were further adjusted by projecting them onto only the masses contained in the model by means of a nonnegative linear regression method . This allows us to match the model parameters to only that part of the experimental mass spectrum explained by the model . However in visually comparing the model spectrum to the experimental spectrum , for example in Figures 2 and 3 , the original unprojected experimental spectra have been used . Figure 9 shows the comparison of the model-generated mass spectra with the projected experimental spectra . Both experimental glycan structure measurements via MALDI TOF mass spectrometry and gene expression measurements using mRNA microarray are available at the Consortium of Functional Glycomics ( CFG ) website [22] and also in supplementary material Datasets S1 , S2 , and S3 , which contains such results for a series of low , medium and high passage human prostate LNCaP cells . These cell types , with N-glycan mass spectra and the microarray data ( MAEXP_291_040606 ) provided by Pi-Wan Cheng , represent a progression from androgen dependent cells to an androgen independent state . The human prostate cancer LNCaP C-33 and C-81 cell model used was established and characterized by [21] . C-33 cells include cells between passages 25 and 35 , and C-81 cells include cells between passages 81 and 125 . Three independent experimental mRNA data sets ( MAEXP_291_040606- Dataset S3 ) for each of the high passage and low passage cell lines were used [22] . Expression levels were detected with the CFG Glycochip version 3 , a custom designed gene chip that uses the Affymetrix technology and contains probe sets for over 1 , 000 human glycogenes . Each targeted mRNA transcript sequence was interrogated by 11 probe pairs of 25 base oligonucleotides each . Each probe pair consists of one perfect match ( PM ) - oligonucleotide complementary to a given portion of the targeted gene- and one mismatched ( MM ) -oligonucleotide identical in sequence to the PM probe , except for a single mismatched base- . The difference between the PM and MM probe signals among all probe pairs for a given gene was used to calculate the hybridization signal . This signal is a weighted average calculated for each probe set that represents the relative abundance of a transcript . In addition , a one-sided Wilcoxon Signed Rank Test is applied to this probe-pair intensity distribution to generate the p value . Thus , ( p<0 . 04 ) is called present ( P ) , p above 0 . 06 is called absent ( A ) , and ( 0 . 04<p<0 . 06 ) is called marginal ( M ) [22] .
|
Glycans are the sugar attachments that are found on proteins and lipids . These highly variable and structurally diverse sugar chains confer distinctive characteristics to the cell surface . Recent research has revealed that these glycan profiles can represent important signatures of disease states and thus understanding glycan processing and structures in cells is an important systems biology goal . Glycan structures are often characterized through mass spectral analysis while their glycosylation processing enzymes are characterized using gene expression profiling . Unfortunately , due to the complexity of glycosylational processing , it has been difficult to relate these disparate data sets until now . In this paper we demonstrate for the first time the ability of a systems glycobiology model to link glycan structural data obtained from mass spectral analysis with mRNA expression data in terms of enzyme activities catalyzing the glycosylation reactions in the cells . We show that such a systems biology model enables identification of distinctive and subtle glycan fingerprints differences between prostate cancer cell stages ( androgen-dependent and more metastatic androgen independent ) . This systems approach will enable us to use high throughput glycomics and gene expression data sets in order to specify glycan-based signatures as important diagnostic markers of disease and potential therapeutic targets .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"systems",
"biology",
"biology",
"genetics",
"and",
"genomics"
] |
2013
|
Integration of the Transcriptome and Glycome for Identification of Glycan Cell Signatures
|
Visceral leishmaniasis ( VL ) has become an important opportunistic infection in persons with HIV-infection in VL-endemic areas . The co-infection leads to profound immunosuppression and high rate of annual VL recurrence . This study assessed the effectiveness , safety and feasibility of monthly pentamidine infusions to prevent recurrence of VL in HIV co-infected patients . A single-arm , open-label trial was conducted at two leishmaniasis treatment centers in northwest Ethiopia . HIV-infected patients with a VL episode were included after parasitological cure . Monthly infusions of 4mg/kg pentamidine-isethionate diluted in normal-saline were started for 12months . All received antiretroviral therapy ( ART ) . Time-to-relapse or death was the primary end point . Seventy-four patients were included . The probability of relapse-free survival at 6months and at 12 months was 79% and 71% respectively . Renal failure , a possible drug-related serious adverse event , occurred in two patients with severe pneumonia . Forty-one patients completed the regimen taking at least 11 of the 12 doses . Main reasons to discontinue were: 15 relapsed , five died and seven became lost to follow-up . More patients failed among those with a CD4+cell count ≤ 50cells/μl , 5/7 ( 71 . 4% ) than those with counts above 200 cells/μl , 2/12 ( 16 . 7% ) , ( p = 0 . 005 ) . Pentamidine secondary prophylaxis led to a 29% failure rate within one year , much lower than reported in historical controls ( 50%-100% ) . Patients with low CD4+cell counts are at increased risk of relapse despite effective initial VL treatment , ART and secondary prophylaxis . VL should be detected and treated early enough in patients with HIV infection before profound immune deficiency installs .
Visceral leishmaniasis ( VL ) is a fatal-but treatable- disease caused by a protozoan belonging to the Leishmania donovani complex . While the Indian-subcontinent , East-Africa and Brazil share the major disease burden , it was long known as a rare pediatric disease in the Mediterranean basin . However , in the HIV-era , VL resurged in Southern Europe in adults with HIV co-infection [1] and has been a clinical challenge until highly-active antiretroviral therapy ( ART ) was introduced [2 , 3] . Today the co-infection is reported from 35 countries [4] and VL is an important opportunistic infection of HIV [5 , 6] . The profound immune deficiency in HIV/VL co-infection leads to poor treatment outcome and frequent recurrence of VL . A few case series studies showed 50% to 100% relapse in a year period without antileishmanial secondary prophylaxis [7–11] compared with less than 10% relapse in those without HIV [10] . Individuals with multiple episodes of VL described as active chronic VL were also reported [12] . Secondary prophylaxis for the prevention of VL relapse is recommended in international guidelines [13 , 14] based on a few case series and small sample size studies from Europe where VL is due to L infantum and transmission is zoonotic [8 , 9 , 11 , 15 , 16] . In northwest Ethiopia , where VL is caused by L donovani and transmission is anthroponotic , the HIV co-infection rate reaches 20 to 30% with up to 56% relapse in a year in patients on ART but without secondary prophylaxis [17] . Patients with low CD4+cell count and/or multiple relapse had an increased risk of ( subsequent ) relapse [17] . Using first line antileishmanial drugs ( sodium stibogluconate , liposomal amphotericin B , paromomycin , miltefosine ) as secondary prophylaxis risks for resistance development that can easily be transmitted in anthroponotic transmission regions [4] . Thus , we chose pentamidine , an aromatic diamidine that is not used in first intention because of toxicity but that was found to be safe when used as prophylaxis at a lower dose ( 3–4mg/kg every 3–4 weeks ) than the daily ( 4mg/kg ) therapeutic dosage [16 , 18 , 19] . The objective of the study was to assess the effectiveness , safety and feasibility of this intervention .
The protocol of the study was approved by the Ethiopian Regulatory Authority ( Food , Medicine , Health Care Administration and Control Authority , FMHACA ) , the National Research Ethics Review Committee ( NRERC ) and the Institutional Review Board of University of Gondar in Ethiopia . Additionally , it was also approved by the Ethics Review Board of Médecins Sans Frontières , and the Ethics Committee of Antwerp University Hospital , Belgium . All subjects were included in to the study after written informed consent was signed . Free treatment was provided . Patients were compensated for transport and food during their visits to the study sites . All study documents were kept confidential and were accessible for the study team , monitors and inspectors . Trained clinical trial monitors carried out two pre-study visits , one initiation visit and 6 monitoring visits according to the WHO and ICH Good Clinical Practices standards . Regulatory inspection was carried out by FMHACA at both sites during the study period . The independent Data and Safety Monitoring Board met five times during the study and assessed the progress of the study when every quarter of total sample recruitment was achieved . The protocol was registered in Clinicaltrials . gov ( code NCT01360762 ) . This was an open label , single arm trial designed to investigate the effectiveness , safety and feasibility of monthly pentamidine prophylaxis to prevent VL relapse in patients with HIV . The study has three phases , an initial 12 months of monthly pentamidine ( main study period ) , six months extended treatment period ( with monthly pentamidine ) for those who remain with CD4 count less than 200cells/μl at the end of 12months follow-up , and a subsequent 12months follow-up after the prophylaxis to assess long term outcomes . The findings of the latter two phases will be published in the future . The study was conducted in Northwest Ethiopia–at the Leishmaniasis Research and Treatment Center ( LRTC ) at University of Gondar Hospital ( UoGH ) and at the Abdurafi Health Centre . They are the largest leishmaniasis treatment centers in the region and are supported by the non-governmental organizations Drugs for Neglected Diseases initiative ( DNDi ) and Médecins sans Frontières respectively . Recruitment of the patients for the study proceeded in two steps . During pre-screening , age 18 or more years , parasitological diagnosis of VL , documented HIV test result and acceptable distance of residence from the trial centres for monthly follow-up were checked . Eligible patients were then approached for consent . There were three types of study participants considered at increased risk of relapse . Patients presenting with active VL disease during the recruitment period were classified into two groups . Current primary cases were those presenting with VL disease for the first time and current relapse cases were patients presenting with two or more episode of VL . Those with active VL disease were admitted to the treatment centres for VL treatment and combination ART ( initiated or continued ) . The drugs used to treat VL were sodium stibogluconate alone or in combination with paromomycin and Liposomal amphotericin B alone or in combination with miltefosine . Treatment of VL was prolonged or changed from one regimen to another when there was no cure with the initial regimen used . The current primary VL cases were included in the study after VL cure if they had a CD4+cell count less than 200 cells/μl or a WHO stage 4 HIV/AIDS defining condition ( other than VL ) while the current relapse cases were included in the study regardless of the CD4+cell and WHO stage of the co-morbidities . Patients who were treated for VL before the start of the study recruitment but in follow-up ( taking ART ) were defined as past VL cases and were included if their CD4+cell count at the time of screening for the study was less than 200 cells/μl or if they had a WHO stage IV-defining illness on presentation . All cases were included after ascertaining parasitological cure ( no parasite on tissue aspirate microscopy ) . Renal dysfunction ( creatinine above twice the normal reference ) , diabetes , pregnancy and lactation , and chronic medical conditions ( e . g . cardiac illnesses ) were exclusion criteria . Pentamidine isethionate ( provided by Sanofi-Aventis ) was started one month after VL cure for the current VL patients; and soon after the inclusion criteria were met for past VL cases . The monthly infusion was continued until the primary end points are met . Pentamidine isethionate 300mg powder was reconstituted with 5ml distilled water and the 4mg/kg ( maximum 300mg ) was drawn and further diluted in 200ml normal saline for infusion over one hour . Frequent blood pressure monitoring was done during infusion . Any adverse events observed were documented . At every visit weight , height , temperature , blood pressure , pulse , spleen and liver size and nutritional status were checked . Monthly full blood count , creatinine , liver function tests and blood glucose; every fourth month electrocardiography tracing and amylase level; and every sixth month CD4+cell count were monitored . HIV viral load determination was done only when clinically indicated and logistically possible . All adverse events were documented , and all the serious adverse events were reported to the sponsor and concerned Ethics Committee via a fast track procedure . Adherence to ART and co-trimesaxole was monitored by patient interview and pill count . The nationally recommended definition for ART adherence as “Good” missing less than three doses; “Fair” missing less than nine doses and “Poor” missing more than nine doses per month was used . Patients found with additional opportunistic infections and/or ART failure were managed according to the national Ethiopian guidelines [20] . During the monthly scheduled visits or the unscheduled visits of the patient the possibility of VL relapse was assessed ( fever , weight loss , organomegaly , reduction in hematological profiles ) . Tissue aspiration and microscopy was done when VL relapse was suspected . HIV was diagnosed followed the national diagnostic algorithm of using two sequential positive rapid test results; KHB ( Shanghai Kehua Bio-engineering Co-Ltd , Shanghai , China ) followed by STAT-PAK ( Chembio HIV1/2 , Medford , New York , USA ) . In case of discrepancy between the two tests , the Uni-Gold ( Trinity Biotech PLC , Bray , Ireland ) was used as a tie breaker in Gondar . In Abdurafi , the confirmation was done by ImmunoComb ( Orgenics ImmunoComb II , HIV 1&2 Combfirm ) after two positive rapid tests . VL was diagnosed by tissue aspirates ( spleen or bone marrow ) and microscopy of the giemsa stains for Leishmania amastigotes . Tissue aspiration was repeated at the end of treatment to assess parasitological cure defined as no parasite on microscopy from the tissue aspirate . Splenic aspiration was avoided whenever the patients had bleeding tendency or the platelet count was less than 50 , 000/μl . CD4+ T lymphocyte count was done at recruitment and every sixth months during follow-up using FACS counter ( BD FACSCalibur flow cytometer , 2009 , USA ) . The haematological analysis was done by a haematology analyser–Beckman Coulter AcT diff , Beckman Coulter Inc . , 2003 , USA . We report here the outcomes as assessed by the end of 12 months . The primary effectiveness outcome was time to relapse or death ( all causes ) ; while the primary safety outcome was the proportion of patients with pentamidine related serious adverse events ( SAEs ) or pentamidine related adverse events that led to the discontinuation of the drug . An adverse event was considered drug-related when the relationship was judged as possibly , probably or definitely related according to the treating physician . The primary outcome for feasibility was the proportion of patients that completed at least 90% of the scheduled visits ( i . e . , 11 out of 12 pentamidine administrations without experiencing relapse or drug-related SAEs ) . Secondary variables of interest related to safety were ‘any drug related adverse event and ‘any SAE’; while feasibility-related variables included ‘the number of treatment discontinuations , interruptions , and additional clinical/therapeutic interventions needed’ . Causes of death were analyzed as tertiary end points . Sample size was calculated with a required precision of 10% for primary effectiveness , 7 . 5% for main safety analysis and 12 . 5% for tolerability . The anticipated main analysis outcomes were a failure rate of 20% and a frequency of drug related SAEs of 10% . With these assumptions , the required sample size was 65 . Allowing for 10% of patients lost to follow-up , the final sample size was calculated to be 72 . A CONSORT diagram ( Fig 1 ) and checklist ( S1 Checklist ) were used to present the patient accounting–total screened , screening failures , enrolled , discontinued and the outcomes . All patients who received a single dose of pentamidine were included in the analyses and the results presented for the three groups: primary VL , relapse and past VL . Baseline characteristics were presented in terms of medians and interquartile ranges for continuous characteristics and using counts and percentages for categorical characteristics . Effectiveness: Effectiveness was analyzed using Kaplan-Meier survival analysis with time to relapse or death as outcome measure . "Failed" means the patient died or parasites were detected in tissues aspirates . Aspirates were taken in case of clinical suspicion of relapse . All other patients were considered free of relapse and were censored at 12 months ( for patients who completed follow-up ) , or at their last visit for patients lost to follow-up . In principle , death was defined as all-cause mortality . The results were given as probability of relapse free survival with 95% confidence interval at 6 and 12 months . Patients who were lost to follow-up or discontinued the treatment for reasons not related to VL relapse before the end of the follow-up period were censored in the main study analysis , but were included as treatment failures in a “worst-case” scenario . Safety: Adverse events were coded using the Medical Dictionary for Regulatory Activities ( MEDDRA ) and were analyzed based on counts of patients with a specific category and not on counts of individual AEs . Primary safety outcome was described in terms of counts of patients with drug-related SAEs or adverse events that led to study drug discontinuation . Counts ( % ) of patients with any SAE and any drug-related adverse events were presented as secondary safety outcomes . Feasibility: The primary outcome for feasibility was the proportion of patients that completed at least 11 of the 12 monthly doses of the prophylaxis according to the protocol without experiencing relapse or drug-related SAEs expressed in percentage . The secondary feasibility endpoint was computed as the number ( % ) of patients who interrupted ( temporarily or permanently ) , and the number of clinical interventions and/or therapeutic procedures needed .
From a total of 176 HIV/VL patients , 74 were recruited into the study ( 38 at Abdurafi and 36 at Gondar ) ( Fig 1 ) in the period from November 2011 to September 2013 . Most were male migrant workers , with a median age of 32 years . Sixty patients were current VL cases ( 25 primary and 35 relapsed VL cases ) , while the rest ( 14 ) were past VL cases . Demographics and baseline characteristics were similar among the three groups ( Table 1 ) . Most were malnourished , 56/74 ( 76% ) ( body-mass-index less than 18 . 5kg/m2 ) and with a history of VL relapse , 43 ( 58% ) . The median duration on ART was 7 months . Tenofovir , lamivudine and efavirenz combination was the common ( 74% ) ART regimen used . While the median CD4+cell count at ART initiation was 70cells/μl , it was 127 cells/μl at inclusion into the study , with 60 ( 85 . 7% ) having a CD4+cell count below 200cells/μl ( Table 2 ) . There were 5/7 ( 71 . 4% ) failures among patients with a CD4+cell count ≤ 50cells/μl , whereas 2/12 ( 16 . 7% ) failed in those with a CD4+cell count greater than 200cells/μl ( p = 0 . 005 ) ( Table 5 ) . Age , body mass index , presence of previous relapse or the number of VL episodes , the antileishmanial drug used to treat the most recent VL episode , duration of ART ( less than or greater than 6 months ) , adherence to ART ( S3 Table ) and diagnosis of tuberculosis did not show statistical significance with chemoprophylaxis failure . HIV viral load was done only for eight of the patients who failed ( 1 death and 7 relapse cases ) and it was undetectable in five of them .
The probability of failure ( relapse or death ) from secondary VL chemoprophylaxis within 1 year was 29% which is lower than the 50% to 100% reported in case series without prophylaxis in Europe [7–11] . The annual probability of VL relapse was 56% in a cohort of patients with HIV on ART , but without secondary prophylaxis in northwest Ethiopia [17] . In a meta-analysis of studies conducted in the L infantum region , the relapse rate was reduced from 67% to 31% with chemoprophylaxis [21] . Our study endpoint was relapse and death , while only the relapse rate was reported from the other studies . Our data corroborate the risk of relapse associated with low CD4+ cell counts , but not with previous multiple relapses as seen before in a study in Ethiopia [17] . However , this is a subgroup analysis and the sample size is not adequately powered to make conclusions . The majority ( 75% ) of the failures occurred in the first six months of the follow-up . In east Africa , VL relapse usually occurs in the 3 to 9 months following the initial treatment [17] . Early relapse may actually be a treatment failure that was missed due to the inherently less sensitive microscopy leading to a false verdict parasitological cure . Or else it is related to a deficient cellular immunity to control remaining parasites after treatment , resulting in a regeneration of the parasite and another episode of disease . Additionally , the protective serum level of pentamidine might not be reached in the first few months . The optimal dose of pentamidine for prophylaxis is also not clearly known . While 4mg/kg dose was meant to be for base-moiety , guidelines did not specify the need for dose modification of the different salt preparations [22] . Similar to other studies , pentamidine prophylaxis was found to be safe [8 , 16 , 23–25] . Only one patient developed transient hyperglycemia . Although renal failure occurred in two patients ultimately leading to death , the patients were having severe infections and it was difficult to attribute the cause of renal failure solely to pentamidine . Other adverse events were mild . Seven ( 9 . 5% ) of the study participants were lost to follow-up and four ( 5% ) of the patients interrupted more than one dose before the primary end point was met . Despite the fact that our study patients belonged to a highly mobile and difficult to trace population group ( migrant workers ) , the proportion of lost to follow-up did not exceed the 10% accounted in the initial sample size calculation . Our study has several limitations . It is not a randomized controlled trial because there was no other antileishmanial drug available or recommended by the Ethiopian national guidelines to compare with . Secondly , because international guidelines recommend secondary prophylaxis to prevent VL recurrence in HIV equipoise was hard to claim [13 , 14] . We did not systematically monitor HIV viral load and we did not determine pentamidine serum levels . Future research need to include pharmacokinetics and resistance testing for anti-leishmania drugs . The efficacy and safety of higher doses of pentamidine and/or more frequent dosing should also be explored . In conclusion , longer VL relapse free survival was achieved using pentamidine as secondary prophylaxis in people with HIV infection . However , patients with profound immune deficiency were still at risk of relapse . Thus , there is a need to investigate additional treatment options for this group of patients . Early VL case detection ( before profound immune deficiency ) is crucial for effective management and prevention of relapses [26] .
|
Relapse of visceral leishmaniasis ( VL ) among HIV co-infected patients occurs universally . Evidence on the use of secondary prophylaxis especially in anthroponotic transmission regions was lacking . It was found out now that secondary prophylaxis in addition to antiretroviral therapy for VL in people with HIV infection is useful to decrease the relapse rate . However , this intervention is more effective when started before profound immune deficiency . Patients with low CD4 cell counts continued to relapse significantly despite the use of secondary prophylaxis as compared to those with high CD4 cell counts . Earlier VL case detection and management is crucial . This is the first adequately powered trial that has addressed the use of secondary prophylaxis for prevention of VL relapse in HIV co-infected patients .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Use of Pentamidine As Secondary Prophylaxis to Prevent Visceral Leishmaniasis Relapse in HIV Infected Patients, the First Twelve Months of a Prospective Cohort Study
|
Trachoma , the worldwide leading infectious cause of blindness , is due to repeated conjunctival infection with Chlamydia trachomatis . The effects of control interventions on population levels of infection and active disease can be promptly measured , but the effects on severe ocular sequelae require long-term monitoring . We present an age-structured mathematical model of trachoma transmission and disease to predict the impact of interventions on the prevalence of blinding trachoma . The model is based on the concept of multiple reinfections leading to progressive conjunctival scarring , trichiasis , corneal opacity and blindness . It also includes aspects of trachoma natural history , such as an increasing rate of recovery from infection and a decreasing chlamydial load with subsequent infections that depend upon a ( presumed ) acquired immunity that clears infection with age more rapidly . Parameters were estimated using maximum likelihood by fitting the model to pre-control infection prevalence data from hypo- , meso- and hyperendemic communities from The Gambia and Tanzania . The model reproduces key features of trachoma epidemiology: 1 ) the age-profile of infection prevalence , which increases to a peak at very young ages and declines at older ages; 2 ) a shift in this prevalence peak , toward younger ages in higher force of infection environments; 3 ) a raised overall profile of infection prevalence with higher force of infection; and 4 ) a rising profile , with age , of the prevalence of the ensuing severe sequelae ( trachomatous scarring , trichiasis ) , as well as estimates of the number of infections that need to occur before these sequelae appear . We present a framework that is sufficiently comprehensive to examine the outcomes of the A ( antibiotic ) component of the SAFE strategy on disease . The suitability of the model for representing population-level patterns of infection and disease sequelae is discussed in view of the individual processes leading to these patterns .
Trachoma is the leading infectious cause of blindness in the world; 8 million people are blind or severely visually impaired due to trachoma and 63 million have active disease [1] . It is due to repeated conjunctival infection with the bacterium Chlamydia trachomatis , and the “SAFE” control strategy ( surgery , antibiotics , facial cleanliness and environmental improvement ) is recommended by the World Health Organization ( WHO ) [2] . The effects of control programs on community infection and active disease can be rapidly measured , but their effects on the severe sequelae ( trichiasis , corneal opacity and blindness ) will not be properly ascertained until decades after their implementation , so mathematical modeling provides an invaluable method for the prediction of program performance . Previous mathematical models of trachoma infection at the population level have primarily looked at the effects of treatment with antibiotics , and the rebound in the prevalence of active disease that follows treatment cessation [3]–[6] . However , no model has taken into account the important effects upon infection and disease of the apparent increase in the rate of bacterial clearance that is observed with age . On the one hand , this shortening of clearance time is thought to be attributable to an acquired , yet not protective , immune response that is enhanced with repeated exposure to C . trachomatis [3] , [7] . On the other hand , the mechanisms responsible for bacterial clearance may also be immunopathological , and they may lead to scarring and subsequent disease sequelae that are associated with trachoma [8] . Therefore , episodes of repeated infection and clearance of C . trachomatis may lead to the more severe complications of the disease: trachomatous scarring ( TS ) , trachomatous trichiasis ( TT ) , corneal opacity ( CO ) and , eventually , blindness . In order to predict the impact of treatment on scarring sequelae , which are the focus of the Alliance for the Global Elimination of blinding Trachoma by 2020 ( GET 2020 ) [9] , models need to incorporate contemporary understanding of the relationship between infection , disease , and disease progression . A mathematical model of ocular infection with C . trachomatis was developed , and its parameters were estimated through fitting the model to pre-intervention ocular chlamydial infection prevalence , rate of recovery from infection , and infection load data from three geographically-separate study sites in The Gambia and Tanzania , representing areas of low , moderate and high endemicity . Insights from the model help explain observed age-profile patterns of infection prevalence in these settings . Progression of individuals to greater numbers of infections , through repeated infection , is interpreted as leading to worsening scarring . Therefore , the model population that has progressed to various numbers of repeat infection represents the population suffering from each of the severe disease sequelae; the corresponding age-profiles of disease prevalence are then compared with clinico-epidemiological data . Finally , the degree to which the model captures the epidemiological patterns of infection and disease observed , and the possible causes for discrepancy are discussed .
The model developed here represents ocular infection with C . trachomatis in a community setting and is based upon a framework commonly used in the modeling of microparasitic infections [10] . In the model , susceptible ( S ) individuals become infected through contact with infected ( I ) individuals before recovering again to a susceptible state . Initial infection and reinfection occur through direct contact with other infected members of the community whereas indirect contact can occur through inanimate objects capable of carrying infection from an infected person to another person or through facial contact with flies carrying the bacteria [11] , [12] . In endemic settings , disease progression appears to occur through multiple reinfection [13] , [14] . Therefore , the model takes account of the importance of multiple reinfections on disease progression by keeping track of the number of infections an individual has experienced . Superinfection of an already-infected individual with a different strain of C . trachomatis , which does occur in endemic villages [15] , is ignored at this stage . Conceptually , the model represents a ‘ladder’ of infection , with each ‘rung’ of the infection ladder corresponding to an additional cumulative infection with C . trachomatis ( Figure 1 ) . Susceptible states are denoted by and infected states by , with subscript i denoting the number of previous infections experienced ( full details of the model are given in Text S1 ) . As individuals progress to the next state up the ladder of infection , a memory of the number of infections experienced is retained . Since birth and death rates are important when determining prevalence levels of the more severe disease sequelae , the demography of the population is included in the model . The disease sequelae are more prevalent at older ages and , once the population has had the antibiotic component of the SAFE strategy successfully implemented , we assume that the rate at which disease prevalence levels decline depends upon mortality among older individuals . Age-specific death rates and the crude birth rate for The Gambia and Tanzania were based on WHO life table estimates for the year 2001 [16] . It is the explicit inclusion of disease sequelae , age structure , differential infectivity and immunity considerations that distinguish this model from those that have been previously reported [5] , [6] . The equilibria of the model will provide a representation of the way in which each of these forces are balanced in the endemic state . Several studies have postulated that the sequelae of trachoma are caused by immunopathological processes that increase in severity with increasing age [17]–[19] . This idea is supported by work which shows that the duration of episodes of infection and active disease ( the latter encompassing trachomatous follicular and severe papillary conjunctivitis ) becomes markedly shorter with increasing age [3] , [7] . In this paper , it is assumed that adaptive immunity does not protect from acquiring infection but results in an increasing rate of recovery from infection as the number of previous infections increases; such a framework was chosen here due to the limited evidence for protective immunity against infection [20] and the preference for a parsimonious model . In Figure 1 the recovery rate from infection Ii is denoted by . This recovery rate approaches a limit at high numbers of infections . The parameter values determining: 1 ) the rate at which the curve rises with infection number; 2 ) the initial recovery rate; and 3 ) the recovery rate following a large number of infections , were estimated using maximum likelihood by fitting the model to data on the prevalence of infection as detailed in the Text S1 . In trachoma-endemic communities , bacterial infection load among individuals at young ages is higher than that at older ages [21]–[23] . In the model presented here , this decrease in infection load with age is ascribed to the acquired immune response to chlamydial infection that is developed through bacterial reinfection . Text S1 describes the decay function that was used to represent the average infection load for an individual who has experienced a given number of infections . The chlamydial load enters the model as a proxy for the infectivity of individuals; those who have experienced fewer infections have a higher infection load than those who have experienced many and are therefore more infectious . The model assumes that scarring worsens through repeat infection with C . trachomatis and that , as scarring ( TS ) becomes worse , the more severe disease sequelae ( TT and CO ) occur . However , due to the complex etiology of CO—reinfection is almost certainly not the only causal factor—only TS and TT are considered in the model , which is entirely based on the reinfection route . It is assumed that these conditions co-occur , so that an individual may have scarring , or scarring and trichiasis . In each case , it is assumed that where TT is present , TS will also be present ( Figure 2 ) . The model was fitted using maximum likelihood to pre-intervention prevalence and chlamydial load data collected in studies carried out in The Gambia and Tanzania , the details of which have been described by Burton et al . [22] , Solomon et al . [21] , [24] , and West et al . [25] ( Table 1 ) . In these studies community infection prevalence was determined by qualitative PCR and the individual infection load by quantitative PCR ( in the case of the hyperendemic Tanzanian community , a model fit was also obtained for which quantitative PCR was used to measure both prevalence and load; further details of these are available in Text S1 ) . These community data-sets illustrate the general differences between three distinct endemic levels while not necessarily being representative of all hyper- , meso- and hypoendemic communities . The model was first fitted using maximum likelihood to the hyperendemic data set for which three distinct data types were available: age-profiles of the prevalence of infection and the infection load from a community in Tanzania [25] , and the rate of recovery from infection based on a cohort study with frequent follow up in The Gambia [3] . A likelihood expression was formulated that combined all three data types and the prevalence , infection load , and duration of infection data were assumed to arise from binomial , Poisson and exponential distributions respectively . The Poisson distribution was selected in the absence of information regarding the distribution of the infection load for each infection category i; and an exponential distribution was used for the duration of infection data consistent with the model structure . The likelihood framework is outlined in Text S1 . Likelihood expressions were formulated for each of the data sets and the overall log-likelihood ( LL ) formed by summing the individual LLs—and assuming that measurements of prevalence and load were independent of one another ( Text S1 ) —so that: ( 1 ) The overall LL was maximized with respect to the six model parameters ( listed in Table 2 ) pertaining to the three data types . Parameter values and their 95% confidence intervals ( found using the profile-likelihood around the maximum LL value ) are provided in Table 2 . For each parameter , the profile likelihood was calculated by fixing the parameter and maximizing the LL with respect to all of the other parameters . Aside from the transmission parameter ( β ) , the parameter estimates thus obtained were then used in modeling the meso- and hypoendemic settings ( defined as those exhibiting an active disease prevalence lower than 10% and between 10 and 20% respectively ( Table 1 ) ) since they pertain to the biology and not to the transmission environment of the infection . Transmission parameters were then obtained separately for the hypo- and mesoendemic datasets based on maximum likelihood fitting to the prevalence data alone from each area . The pattern of population mixing among ages was assumed to lie between the extremes of entirely random and fully assortative [26] . The model was implemented in Matlab using the Euler integration method; the LL maximization was also performed using the Matlab package .
The curves shown in Figure 3A and 3B represent the model-generated age-profiles of the infection load and the rate of recovery from infection based on the maximum likelihood parameters obtained from the analysis of data from the hyperendemic setting ( i . e . as in Table 2 ) . These curves show the previously described reduction in infection duration and intensity with age and the parameters thus obtained were used for all subsequent modeling . Model fits to the data by endemicity level are shown in Figure 4 . The solid line in Figure 4A illustrates the model-generated infection prevalence curve corresponding to the fit to the hyperendemic data set published by West et al . [25] . The dotted line in Figure 4A is an illustrative fit obtained assuming a lower number of individuals in each age-group classified as positive for infection ( see Text S1 for details ) . The transmission parameter β estimated for this adjusted dataset is approximately two thirds the size of the value shown in Table 2 . Figure 4B and 4C illustrate the model-generated infection prevalence curves for the model fitted to the hypo- and mesoendemic datasets ( i . e . using the endemic-specific transmission parameter estimate but the infectivity and rate of recovery parameter estimates obtained from the analysis of the data from the hyperendemic setting ) . Figure 3B shows that the duration of chlamydial infection declines from its initial maximum value to its plateau very rapidly with age , and this is due to the rapid decrease in this value with infection number . The infection prevalence data come to a peak at young ages ( roughly 5 years ) in the hypo- and mesoendemic areas examined here , with model fits mirroring such peaks . Furthermore , the data also show some evidence for a peak shift [27] , characterized by the peak of infection being shifted towards younger ages as transmission levels increase ( Figure 4 ) . The threshold numbers of infections necessary for individuals to show signs of each of the sequelae were calculated for the hyperendemic setting . These thresholds were estimated by maximum likelihood using the published data of Munoz et al . [28] , for the age-dependent prevalence of each of the disease sequelae . The maximum likelihood estimate for the threshold number of infections required for TS was 102 and for TT it was 151 . In terms of the natural history of trachoma infection , disease , and disease sequelae , it is assumed here that these threshold values do not vary over the different endemicity levels but should be reached at different ages according to the intensity of transmission—individuals living in areas of different endemicity are assumed to show signs of each of the disease sequelae after having experienced the same number of infections , but they experience the sequelae earlier in their lives in those environments in which they are infected more frequently . The threshold infection numbers estimated here were used to generate the curves shown in Figure 5 .
The model presented in this paper reproduces many important aspects of trachoma epidemiology , namely: 1 ) the pattern of the prevalence of infection with age , which peaks at very young ages and then declines; 2 ) a peak shift towards younger ages in this prevalence in higher transmission settings; 3 ) a rise of the prevalence level with higher transmission; and 4 ) a rise with age of the prevalence of the severe disease sequelae . The model also allows estimates to be made of the threshold number of infections necessary for the appearance of the severe disease sequelae . Infection and disease profiles were obtained assuming long-term stability of prevalence levels in the model , and therefore represent the equilibrium , pre-intervention state in each of the endemicity settings . However , a limitation of the data used to estimate model parameters is that in some areas ( e . g . , The Gambia ) , there have been secular trends towards a decrease of trachoma incidence in the absence of control interventions . These trends are deemed much less significant in the hyperendemic setting . Following the fitting of the model to the data-sets from the three endemic settings , the prevalence curves generated show a close correspondence with the trend in the observed profiles of infection prevalence with age . While good visual fits to the data are encouraging , a full analysis of the uncertainty in the parameter estimates is essential to judge how well-determined the model fits are . In the younger ages , the prevalence peak is caused by the long duration of infection , high chlamydial loads and intense transmission that result from patterns of assortative mixing by age . Subsequently , the prevalence of infection drops at older ages , as a consequence of the age-associated increase in the recovery rate from infection and the drop in infectivity with age: an individual who experiences an increasing number of infections recovers faster from each infection , with accompanying reductions in chlamydial load and infectivity . ( In the model , the number of infections previously experienced tracks closely the age of an individual . ) There is also a peak shift of the maximum infection prevalence towards younger ages ( slightly greater than 5 years of age in the hypoendemic; slightly under 5 years in the mesoendemic , and very low ( under 1 year ) in the hyperendemic areas ) . For infectious diseases in general , this effect is usually due to acquired immunity and it occurs when individuals experiencing higher forces of infection either develop adaptive protective immunity earlier or clear their infection more rapidly at younger ages than they would in environments with lower force of infection . The observation of this phenomenon here lends further support to the importance of acquired immunity in trachoma [27] , [29] . The recovery rate from infection rises very rapidly with the number of prior infections and the 95% confidence interval of the rate of this rise includes extremely large values at the upper end . This rapidity suggests that the immune response to the first few infections is qualitatively different to that of the bulk of subsequent infections and therefore the maturation of trachoma immunity occurs after only few infections , a finding that may also be associated with limited variation in the pathogen population . Indeed , the possibility of extremely large values for the rate at which the duration of infection changes with the number of infections ( unbounded upper confidence limit of the rate , in Table 2 ) , suggests that the data are consistent with the development of immunity following a single initial infection . The data for the recovery rate from infection ( plotted in Figure 3B as its reciprocal , the average duration of infection ) used in the model were on average lower than the estimates reported by Bailey et al . [7]—who used a test for infection less sensitive than PCR-based testing and may have found longer durations ( i . e . lower recovery rates ) with a more sensitive test—and instead corresponded with a newer analysis of the same data [3] ( where the mean duration of infection was found to be around 5 months for young children ( <5 years old ) and under 3 months for older people ( >15 years old ) ) . If the rates of recovery in the model were not as low as those used in this work , the peak in infection prevalence , for settings with lower endemicity , would not occur at ages corresponding to those observed in the data . In the hyperendemic setting , the peak prevalence occurs at a very early age becoming barely perceptible due to its large transmission rate; this causes individuals rapidly to acquire infections from a very young age . Infection bacterial loads are explicitly included in the model; the chlamydial loads for those who have experienced few infections are typically higher than for those who have experienced many . The reason behind this difference is thought to be the development of acquired immunity through repeated exposure to the bacteria that , although it does not protect from incoming infection , may reduce its intensity . A model structure in which pathogen load is explicitly accounted for has been used extensively to model helminth infections , by assuming that acquired immunity to infection may be developed with cumulative infection experience and therefore with age , leading to peaked age-profiles of infection intensity and prevalence [29]–[31] . Age-specific changes in exposure are also likely to contribute to this pattern [32] , [33] , and indeed the peaked distributions observed in the model and the data for ocular chlamydial infection and disease are the result of both changes in the duration of infection and patterns of exposure to infection with age . The prevalence levels of the disease sequelae were modeled under the assumption that individuals who had experienced greater than or equal to a specific threshold number of infections would begin to show signs of the ocular sequelae . Threshold infection numbers were therefore estimated corresponding to each of TS and TT; these calculations were performed for the hyperendemic setting , because it would only be in communities where there has been no intervention ( at true endemic equilibrium ) that the transmission and repeat infection rates will give rise to current disease sequelae prevalence levels . The threshold infection numbers for TS and TT estimated in this paper are dependent upon the data we have used for the duration of infection and infection load; a higher duration , for example , would decrease these estimates and so these values are contingent upon future longitudinal studies . In those communities ( the hypo- and mesoendemic areas in this paper ) where there has been either some intervention or possibly a secular trend that has reduced transmission , the prevalence of those suffering from sequelae will , for some time , remain much higher than the current transmission level would suggest . Another explanation for differences here is the possibility that only a given fraction of the population progresses to each of the disease sequelae and this fraction may vary between populations due to factors such as the genetic predisposition to scarring of particular individuals in each population [34]–[41] . Although our working hypothesis is that repeat chlamydial infection is the main route to the severe disease sequelae , it may not be the only one . Some studies show that , once established , scarring complications may continue to progress , perhaps driven by factors other than Chlamydia spp . , such as non-chlamydial bacterial infection [42] . Trachomatous CO leading to blindness probably has a multi-factorial etiology . These effects will be examined in future work and may lead to lower threshold infection numbers than those calculated here . In summary , the balance between ocular exposure to C . trachomatis and acquired immunity , which is presumed to reduce the intensity and duration of infection , leads to the expected shape and magnitude of the age-profiles of infection prevalence observed in settings of variable endemicity . Data used in the model for the recovery rate from infection [3] led to lower corresponding parameter estimates than those previously reported [7] , closer to those used in other trachoma models [5] , although these models do not allow for an age-dependent recovery rate , nor do they investigate the relationship between infection and disease or incorporate chlamydial load . For the ( hyperendemic ) setting in which levels of current infection are those responsible for observed morbidity , the model captures well the progression of scarring with age and reproduces the observed age-profiles of ocular sequelae prevalence . Future work will investigate the effect of the ‘A’ component of the SAFE strategy ( mass administration of antibiotics ) on the age-profiles of infection and disease , and will present the implications of this model for trachoma control policy .
|
Trachoma is the worldwide leading infectious cause of blindness and is due to repeated conjunctival infection with Chlamydia trachomatis bacteria . The effects of control interventions on population levels of infection and active disease can be promptly measured , but the effects on severe ocular disease outcomes require long-term monitoring . We present a mathematical model of trachoma transmission and disease to predict the impact of interventions on blinding trachoma . The model is based on the concept of multiple re-infections leading to progressive scarring of the eye and the potentially blinding disease sequelae . It includes aspects of trachoma natural history such as an increasing rate of recovery from infection , and a decreasing chlamydial load with subsequent infections . The model reproduces key features of trachoma epidemiology such as the age-profile of infection prevalence; a shift in the prevalence peak toward younger ages in higher-transmission environments; and a rising profile of the prevalence of the severe sequelae ( scarring , trichiasis ) , as well as estimates of the number of infections experienced before these sequelae appear . The model can be used to examine the outcomes of various control strategies on infection and disease and can help to plan treatment interventions for different endemic settings .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"infectious",
"diseases/epidemiology",
"and",
"control",
"of",
"infectious",
"diseases"
] |
2009
|
The Development of an Age-Structured Model for Trachoma Transmission Dynamics, Pathogenesis and Control
|
Deleterious mutations inevitably emerge in any evolutionary process and are speculated to decisively influence the structure of the genome . Meiosis , which is thought to play a major role in handling mutations on the population level , recombines chromosomes via non-randomly distributed hot spots for meiotic recombination . In many genomes , various types of genetic elements are distributed in patterns that are currently not well understood . In particular , important ( essential ) genes are arranged in clusters , which often cannot be explained by a functional relationship of the involved genes . Here we show by computer simulation that essential gene ( EG ) clustering provides a fitness benefit in handling deleterious mutations in sexual populations with variable levels of inbreeding and outbreeding . We find that recessive lethal mutations enforce a selective pressure towards clustered genome architectures . Our simulations correctly predict ( i ) the evolution of non-random distributions of meiotic crossovers , ( ii ) the genome-wide anti-correlation of meiotic crossovers and EG clustering , ( iii ) the evolution of EG enrichment in pericentromeric regions and ( iv ) the associated absence of meiotic crossovers ( cold centromeres ) . Our results furthermore predict optimal crossover rates for yeast chromosomes , which match the experimentally determined rates . Using a Saccharomyces cerevisiae conditional mutator strain , we show that haploid lethal phenotypes result predominantly from mutation of single loci and generally do not impair mating , which leads to an accumulation of mutational load following meiosis and mating . We hypothesize that purging of deleterious mutations in essential genes constitutes an important factor driving meiotic crossover . Therefore , the increased robustness of populations to deleterious mutations , which arises from clustered genome architectures , may provide a significant selective force shaping crossover distribution . Our analysis reveals a new aspect of the evolution of genome architectures that complements insights about molecular constraints , such as the interference of pericentromeric crossovers with chromosome segregation .
Mating and meiosis are the masterpieces of an evolutionary invention thought to meet the challenges of changing environmental conditions that need to be solved by mutational inventions . Among the many hypotheses that govern the various benefits of mating and meiosis [1] , two main hypotheses stand out: enhanced purging of deleterious mutations [2] and the combination of beneficial alleles into one genome [3] . It remains a matter of discussion , however , which of these advantages constitutes the main reason for the evolution of sexual recombination and , furthermore , its continuing prevalence in most eukaryotic life forms [4] , [5] . Mutations take the form of DNA lesions that are caused by environmental factors , e . g . radiation , but they are also a natural byproduct of DNA replication . Genotypes that exhibit an elevated mutation rate are frequent in nature and can be induced in studies on experimental evolution . The complex interplay of factors that govern the adaptive significance of “mutator alleles” ( i . e . alleles that cause higher mutation rates ) has been studied in experimental and theoretical work in unicellular organisms [6]–[8] and during cancer progression [9] , [10] . In asexual yeast populations , a selective advantage of mutator alleles has been demonstrated , serving as a prerequisite for expanding the spectrum of mutations typically not accessible in non-mutator genotypes [11] . In this study with yeast , the mutator advantage was found to be more prominent in diploid rather than haploid cells , which can be explained by the presumed dominance of beneficial mutations [11] , [12] and by the recessive nature of most deleterious mutations in yeast [13] . Furthermore , the accumulation of recessive deleterious mutations in yeast may not significantly decrease the fitness of the genotype or population growth rates , as long as the diploid nuclear condition is maintained [14] . S . cerevisiae and many other yeast and fungal species seek to maintain the diploid state of the genome , if possible; after meiosis this usually happens by immediate mating of the gametes following germination . Mating occurs mostly between closely related spores , either among products of the same meiosis , or between spores from related cells according to population structure [15] . Outcrossing between unrelated strains [16] and even between closely related species of the sensu stricto yeast group does occur [17]; these events appear to be extremely rare , but they might be important for generating new persisting lineages . However , whether these rare events suffice to create a selective pressure towards maintaining a sexual cycle is doubtable . Upon inbreeding , high rates of homozygotisation occur at various loci . This is reduced for loci linked to the MAT locus , since mating type heterozygosity is the prerequisite for a mating event . A MAT-linkage to centromeres is frequently observed in yeast and other fungi [18] . In S . cerevisiae , the genetic distance between the two loci is in the range of 18–30 cM [19] , which is much smaller than expected from the physical length and caused by a region intervening these two loci that is cold for meiotic recombination [20] . In Neurospora tetrasperma MAT linkage to the centromere is enforced by crossover suppression in a long region of the chromosome , which correlates with an extensive unpaired region at pachytene [21] . Reasoning for such arrangements is provided by population genetic models that suggest a selective advantage arising from shielding of recurrent deleterious load via linkage to the MAT locus [22] . Population genetic modifier models that investigate alteration of the inbreeding frequency predict the evolution of high inbreeding rates , in particular for spores from the same tetrad ( i . e . automixis ) , and the linkage of load loci to the MAT locus [23] , [24] . These MAT linked chromosomal recombination abnormalities are believed to have initiated the evolution of sex chromosomes , which was then continued by expansion of the recombination suppressed region through the recruitment of other sex-related factors [25] , [26] . Non-random distribution of meiotic recombination throughout the genomes into cold and hot regions has been reported for many species [27] , but the molecular mechanisms as well as the selective forces that generate these patterns are still not fully understood . Similarly , the distribution of genes along chromosomes appears to be non-random , and in many species a significant clustering of essential genes ( “housekeeping genes” ) has been reported [28] , [29] . In budding yeast , a prominent genome-wide enrichment of essential genes has been observed in regions that are cold for meiotic recombination [30] . This finding is consistent with the observation of a slight enrichement of essential genes near the centromeres [31] , which are known to be cold for meiotic recombination [32] , [33] . Current models speculate about mechanisms of co-expression and reduction of gene expression noise as the driving force that shaped these patterns [29] , [30] , [34]–[36] . However , the selective advantage of such a scenario has not been demonstrated , and proof that single rearrangements associate with an advantage sufficient for their selection has not been provided . An alternative mechanism that may pool essential genes into clusters with a reduced probability of disruption by frequent crossovers could be a selection based on their common denominator . This appears to be their essential nature only , as no functional correlation of essential genes within the same cluster has been observed [30] . Meiotic recombination is favorable for purging deleterious load from populations . We hypothesize that clustering of essential genes may further enhance purging efficiency of lethal load from sexual populations , since non-uniform distributions of essential genes and crossover sites change the global genetic linkage relationship of all essential genes ( as compared to situations with uniform or random distributions ) . In the context of a scenario with more than one lethal mutation in the genome , this will influence the segregation of mutations during meiosis and subsequent mating . In order to address this question , we computationally studied the correlation between non-random distributions of essential genes and meiotic recombination with lethal load affecting essential genes . A Monte-Carlo-simulation of breeding diploid yeast populations and chromosome architectures , termed S . digitalis , allowed us to investigate the fitness of any genome architecture upon exposure to lethal mutations . We find that several hallmarks of yeast chromosome crossing over during meiosis are consistent with natural selection imposed by recessive lethal mutations affecting essential genes ( see Figure 1 for an overview of our approach and the results ) . Our simulations imply that lethal phenotypes are frequently caused by single essential gene inactivation . Alternatively , lethal phenotypes may arise from genetic interactions between only weakly deleterious mutations . We explored both possibilities using a conditional yeast mutator strain and analyzed the causes of the accumulation of haploid lethal phenotypes . By determining the global effect on germination and mating , we tested whether the associated load is being transmitted into the next round of diploid growth . Our combined results suggest an evolutionary history for yeast where sex and meiosis fulfilled a need for efficient purging of mutational load in important/essential genes .
We sought to assay the consequences of non-random distributions of essential genes and meiotic recombination hotspots for the fitness of populations upon frequent inactivation of essential genes by lethal mutations . This analysis requires a direct comparison of the fitness of yeast strains with the same genomic content but different arrangements of the genetic elements . Conducting an experiment such that this question can be addressed in isolation from the many other possible consequences of human-designed genomic architectures is far from trivial . Therefore , we developed a computer simulation of populations of digital genomes subjected to digital life cycles of mitosis , meiosis and mating , modeled according to a simple yeast life cycle ( Figure 2A ) . We used this simulation to study the relationship between genome architecture and the fitness of populations as well as the evolution of genome architectures upon exposure of populations to essential gene inactivating mutations ( Figure 2B ) . Haploid lethal mutations are frequently observed in yeast and constitute approximately 40% of all deleterious mutations [37] , [38] ( see also below ) . The remaining fraction of deleterious mutations has been reported to exhibit a weak impact on fitness [37] and both types of mutations are shielded well in heterozygous diploids [39] . Therefore , we decided to focus on heterozygous lethal mutations , which confer a lethal phenotype either on the level of haploids or upon homozygotisation of mutations in diploids ( Supplementary Figure 1A in Text S1 ) . S . digitalis simulates populations of diploid digital organisms with one chromosome , which consists of different building blocks: genes , intergenic elements , centromeres and mating type loci ( MATa and MATα ) . The digital genes are either non-essential or essential . Each gene of the latter category carries a unique identifier . Genes are separated by intergenic elements ( IE ) . IEs are either cold for crossing over ( = coldspots ) or hot ( = hotspots ) . Each feature is represented by an element in the matrix of the population genome , and mutations only affect elements that represent essential genes ( Figure 3A ) . Typical natural yeast chromosomes contain a few hundred genes , and so do our digital counterparts . Populations have a finite size and typically consist of a few hundred to ten thousand diploid individuals . Mitosis yields a copy of the original genome . It differs from the template genome by mutations , which are introduced at random and lead to the inactivation of essential genes ( Figure 3B ) . The statistical frequency of essential gene mutations per diploid genome and mitosis is given by the genomic recessive lethal mutation rate R [40] ( for example , R = 1 corresponds to an average of one essential gene inactivation per diploid genome and mitosis ) . Genomic rearrangements can be simulated . They manifest themselves either as positional swapping of genes and associated intergenic elements , or as segmental inversions ( Figure 3B ) . Genomic rearrangements occur in mitosis and always affect both homologous chromosomes . In meiosis , the genome duplicates and the homologous chromosomes undergo meiotic recombination ( crossing over ) . The distribution of crossover sites considers meiotic recombination hotspots and crossover interference based on a genetic distance definition ( hotspot distribution ) . The crossover frequency can be adjusted by the shape factor of the Erlang distribution that is used to describe crossover interference ( Figure 3C ) . The four meiotic haploid progenitor genomes constitute a tetrad . Mutations are allowed to occur in mitosis ( see Text S1 , section “Supplementary Results and Discussion” for an analysis of meiotic mutations ) . This implementation considers one single mitotic cycle between consecutive meiotic cycles . This mimics a situation in which many consecutive rounds of mitoses occur without exponential growth . This applies to scenarios where a high loss of individuals occurs ( e . g . many individuals eaten by predators or washed away into non-fertile grounds ) that keeps the size of a local population more or less constant . Under circumstances where deleterious mutations are recessive and do not influence the fitness of the cells ( as indicated by literature [39] ) , this would lead to the accumulation of mutations during the vegetative period of the life cycle . This simplified scenario should come close to a realistic description of natural S . cerevisiae that is consistent with the absence of reports of large natural cultures of S . cerevisiae ( outside of human-engineered fermentation processes ) . Moreover , this approximation allows us to simulate large numbers of complete life cycles , which would otherwise be inaccessible due to computational limitations . Upon germination , the haploid genomes directly engage in mating with other haploid genomes ( Figure 3D ) . Mating can occur between genomes from the same meiosis , which is called intratetrad mating and more generally referred to as automixis or inbreeding . Mating between haploid genomes from different tetrads can also occur , and is referred to as amphimixis or outbreeding . In this article , we use the terms inbreeding and outbreeding to distinguish between the two principal types of mating partner selection in the simulation: mating inside and outside the tetrad . Outbreeding events may nevertheless bring closely related genomes together , simply due to the finite size of the simulated populations . The total fraction of inbreeding matings per round of mating can be specified . Mating optionally considers mating types , of which two exist ( MATa and MATα ) . The MAT locus can be placed anywhere on the chromosome . In this case , the modeled chromosome can be considered to be the sex chromosome . Alternatively , the simulation can employ a virtual second chromosome that contains the MAT locus next to its centromere . In intratetrad mating , this causes a linkage of the MAT locus to the centromeric region of the investigated chromosome [41] . Diploid genomes with different gene order belong to different species . Individuals from different species are not able to mate with each other . These species represent sub-populations , which emerge in simulation scenarios with genomic rearrangements . Alternatively , different sub-populations can be specified at the beginning of the simulation , e . g . in order to compare the fitness of different genome architectures ( species ) in survival competition assays . Fitness of the individuals is assessed in the diploid stage before mitotic or meiotic cell division . We furthermore assumed that mating is not prevented by a haploid lethal mutation in an essential gene ( this assumption was experimentally tested; see below , section “Mating rescues genomes associated with lethal mutations” ) . We decided to use a simple fitness denominator for individuals: a 1 is assigned for diploid genomes that contain at least one functional copy of each essential gene , while a 0 is assigned for genomes , in which both copies of at least one essential gene are non-functional . Individuals with a fitness of 0 are removed . Hence , the only criterion underlying the loss of an individual due to mutations is the homozygotisation of a mutated essential gene . This can occur in two different ways: a new mutation inactivates the second wild type copy or a mating event brings together two chromosomes that both contain a mutated allele at the same position . Populations were limited in size according to a defined maximum ( the population size cap ) . Excess individuals are removed at random before the next round of mitotic or meiotic division . This simulates limited availability of nutrients . As a result , a selective pressure is introduced that has the potential of driving the evolution of species ( = different genome architectures ) that are better adapted to handle lethal mutations . Supplementary Figure 1A and 1B in Text S1 provides an overview of the mechanisms of the simulation . The simulation provides modules for different types of experiments , including mutational robustness benchmarks of populations with specific genomic architectures , selection advantage assessments with two or more isolated populations that compete for nutrients and evolution experiments of large in- and out-breeding populations that constantly undergo genomic rearrangements ( see Text S1 ) . Detailed descriptions of all simulation modules and the implementation of yeast and model chromosomes are provided in Text S1 . Using S . digitalis we first assessed the maximum mutation rate R populations with random distributions of genes and recombination hotspots can resist before becoming extinct ( the mutational robustness Rmax ) . We compared the results with the mutational robustness obtained for the S . cerevisiae chromosome IX architecture , which deviates significantly from a random arrangement of essential genes and meiotic recombination hotspots distribution [30] . This revealed a superior mutational robustness of the yeast chromosome architecture for the entire spectrum of inbreeding fractions ( Supplementary Figure 4A; Supplementary Figure 2A in Text S1 ) . We obtained the same result when allowing mutations to occur in meiosis only , or both in mitosis and meiosis ( Supplementary Figure 2B and 2C in Text S1 ) . Using a survival competition assay , we directly compared the persistence of populations with random chromosomes and of populations with yeast chromosome IX at different mutation rates and for different population sizes . The competition experiments revealed a clear selective advantage of the yeast chromosome IX architecture for most regions of the investigated parameter space ( Figure 4B ) . A stalemate situation was only observed at low mutation rates R<0 . 01 and for extreme inbreeding fractions ( i = 0 and i = 1 ) ( Figure 4B and 4C ) . We performed a control experiment to demonstrate that the quantitative outcome of the survival competition assay is unaffected by the choice , in which life cycle state mutations are simulated ( mitosis and/or meiosis ) ( Figure 2D in Text S1 ) . An analysis of the recently recorded distribution of 4 , 300 single crossover events in 50 meioses of yeast [42] indicated non-random distributions of crossovers and essential genes for the entire yeast genome . We found that the resulting average level of clustering is more than 2 . 5σ higher than the level expected for random distributions ( Table 1 in Text S1 ) . Using S . digitalis , we obtained comparative data by subjecting digitalized implementations of all chromosomes ( Text S1 ) to a survival competition against randomly generated chromosomes . The simulation outcome attests a superior fitness to almost all of the yeast chromosomes ( Figure 4D ) . Taken together , our data suggest that yeast-like chromosome architectures contain evolved features consistent with selection imposed by lethal mutations . The fitness advantage for the structure of chromosome IX relative to randomly structured chromosomes may result from essential gene clusters that are either centromere-linked or peripheral to the chromosome , or cumulatively from both . We designed synthetic chromosome architectures to discern the potential impact of these structural relationships . First , in simulations without MAT loci , the highest average Rmax were obtained for synthetic chromosomes , in which essential genes were distributed in a few large clusters ( Figure 5A ) . Moreover , the variability of Rmax as a function of the inbreeding ratio was larger in the case of the chromosomes with less essential gene clustering . The number of clusters providing the best performance depends on the population size: the larger the population the smaller the optimal number of clusters the pool of essential genes must be distributed to ( Figure 5B ) . For clustered architectures , the introduction of a MAT locus ( which constitutes an obligatory heterozygosity ) into one of the clusters increased the persistence compared to genotypes without a MAT locus ( Figure 5A , inset; and Supplementary Figure 3 in Text S1 ) . Using survival competition assays in genomes with MAT-linked clusters , we compared the fitness of genotypes with peripheral essential genes either in clusters or in random distributions . We obtained a fitness advantage of peripheral clustering for a wide range of inbreeding ratios and mutation rates ( Figure 5C ) . Thus , MAT-centromere-linked clusters and peripheral essential gene clusters provide cumulative fitness benefits . The tight physical linkage of all essential genes into a single , large cluster is not a very likely configuration for natural genomes . However , achiasmate meiosis ( the absence of meiotic crossovers ) results in a comparable situation , since it genetically links together all essential genes on a chromosome . In the following , we will use the word “achiasmate” to denote the absence of meiotic crossing over between all essential genes present on a chromosome . Meiosis without crossovers has been reported for several species [43] and was also suggested to occur in the hemiascomycete yeast Saccharomycodes ludwigii [44] , [45] . Using our survival competition assay we found that achiasmate meiosis exhibits a high mutational robustness Rmax ( Supplementary Figure 3 in Text S1 , genomes with one essential gene cluster and with mating types , +MAT ) and provides a particularly strong fitness advantage in the entire investigated parameter space ( considering the inbreeding fraction i and the mutation rate R ) when a MAT was present ( Figure 6A , top panel ) . Generally , no advantage of achiasmate meiosis would be expected for pure outbreeding . The advantage of achiasmate meiosis observed in our simulations in the pure outbreeding domain can be explained by the finite population size , which implies that all individuals are related to a certain degree . Without MAT linkage to the essential gene cluster , the advantage is reduced , but significant for mutation rates R<1 . 5 and non-extreme inbreeding fractions ( 0<i<1 ) ( Figure 6A , lower panel ) . Using direct competition , we found a strong advantage of achiasmate meiosis over random chromosomes for mutation rates R between 10−4 and 1 , which further increases with increasing population size ( Figure 6B and 6C ) . In achiasmate meiosis , linkage to the MAT locus either occurs physically ( on the chromosome where the MAT locus is located ) or via the centromeres ( for all other chromosomes due to intratetrad mating ) . This preserves heterozygosity at autosomal centromeres ( see Figure 6D ) . The resulting selection advantage may provide population-genetic reasoning for the secondary loss of meiotic crossing over in Saccharomycodes ludwigii . Our experiments demonstrated a fitness advantage of clustered chromosome architectures when exposed to deleterious mutations . This fitness advantage is the result of the cumulative effects arising from pre-existing essential gene clusters , but it does not allow us to deduce whether clustered genomes can evolve from unclustered or random architectures solely due to the exposure to lethal mutations . For example , alternative and potentially synergistic mechanisms are conceivable ( see Discussion ) . In order to constitute a driving force , the presence of deleterious load would have to cause the emergence of clusters in a self-organized manner , exclusively based on the rules that govern the evolutionary process of genomes in the context of unicellular breeding populations subjected to deleterious mutations . In our investigation of the in silico evolution of clustered genomes , we first dissected the complementary process: the maintenance of essential gene clusters in the context of chromosomal rearrangements and the linkage of the MAT locus to a gene cluster . We designed an initial genome that contained all essential genes in one large cluster and all meiotic recombination hotspots outside this cluster . Scenarios with and without MAT loci were considered in an inbreeding-only domain ( full intratetrad mating ) , which is least favorable to successful persistence of the lineage in the situation without a MAT ( see Figure 6A ) . For +MAT scenarios , the MAT locus was placed outside the essential gene cluster . Populations were evolved using different rearrangement rates ( r ) . For rearrangement rates r<10−4 , simulated architectures both with and without MAT loci preserved highly significant levels of essential gene clustering and anti-correlated meiotic crossover distributions over periods of at least 150 , 000 generations ( Figure 7A and 7B , Video S1 ) . Without MAT , high preservation of essential gene clustering was only observed in a relatively narrow range of mutation rates ( 0 . 6<R<1 . 4 ) . In the presence of MAT loci , however , clustering was well maintained over a significantly broader range of mutation rates ( R≥0 . 2 ) and also for higher rearrangement rates ( Figure 7A ) . We found that the mating type locus had always relocated to a position inside the cluster by stochastic rearrangement ( usually within the first 1 , 000 generations , n = 20 experiments ) and had remained in the cluster afterwards . This experiment demonstrates that a significant level of clustering can be preserved in the presence of lethal mutations . Less well-clustered architectures that arise in the presence of destructive forces ( genomic rearrangements ) are quenched due to a selective pressure towards the better-performing clustered architectures , as long as the rearrangement rate is not too high . In the next step , we investigated the de novo evolution of MAT-linked essential gene clusters in small model genomes containing five essential genes , five non-essential genes and four recombination hotspots ( Figure 7C ) . We found that genomes evolved MAT linked essential gene clusters over a broad range of mutation rates and rearrangement rates , which is consistent with the fact that inbreeding preserves MAT-linked heterozygosity ( Supplementary Figure 7D; and Supplementary Figure 4 in Text S1 ) . Architectures with a single cluster ( 2+1+1+1 , 3+1+1 , 4+1 or 5 genes ) were consistently favored over multi-cluster architectures ( 2+2+1 or 2+3 genes ) ( Figure 7C and 7D; grey rectangles indicate single-cluster architectures in Figure 7D ) . In order to investigate the evolution of large yeast-like chromosome architectures we switched to domains with 50% inbreeding , since no advantage of chromosome-peripheral clustering was apparent for the extreme breeding domains at i = 0 and i = 1 ( Figure 4 and Figure 5C ) . For this series of experiments , we implemented a species barrier in the simulation ( see also first section of Results , and Text S1 , section “Supplementary Results and Discussion” ) . Thereby , each genomic rearrangement leads to the formation of a new species . This scenario mimics reproductive isolation due to meiotically incompatible chromosomes . A new species might eventually dominate the population or become extinct depending on the reproductive success arising from the fitness ( dis- ) advantage of its particular genome architecture . We found that the simple life cycle of mitosis , meiosis and mating was sufficient to reproducibly evolve genomes with MAT-linked as well as non-random peripheral essential gene distributions ( Figure 8A and Video S2 ) . In order to obtain good statistics on this phenomenon , we parallelized the assay by using grid computing , which allowed us to simultaneously evolve many unrelated populations ( n = 3 , 000 experiments ) . On average after 15 , 000 generations , high-R +MAT populations reproducibly evolved a level of clustering 2σ above the mean level encountered in random architectures . The evolution of the same level of clustering in low-R +MAT populations and in −MAT populations required a 2–3 fold longer period . Importantly , all high-R +MAT populations and even a small fraction of the other populations eventually arrived at the level of clustering of the natural yeast chromosome IX ( emerging after 30 , 000–70 , 000 generations , statistics are provided in the legend of Figure 8A ) . Genomes with one MAT-linked cluster dominated at high mutation rates ( R = 1 ) , whereas genomes with several clusters , one of which associated with the MAT locus , typically evolved at lower mutation rates ( R = 0 . 1 ) . We must point out that the parameter space explored in the evolution of clustering constitutes a compromise enforced by limitations in computation time . The computation of all events during one generation requires approximately ten seconds of CPU time in a population of 4 , 000 individuals . In order to be able to perform a statistically meaningful number of experiments under different conditions , we applied relatively high rearrangement rates close to a “destructive regime” , in which any emerging cluster quickly became scrambled . This setting allowed us to simulate genomic restructuring as would quantitatively occur over long evolutionary timescales using reasonable amounts of computation time . However , up-scaling the rearrangement rate also necessitates up-scaling the mutation rate , in order to arrive at a selective pressure on par with the potential destructive force introduced by the random rearrangements . If computation time was unlimited , we would also expect a qualitatively comparable outcome for lower values of R in the context of lower rearrangement rates . We were able to qualitatively reproduce our results with respect to the evolution of essential gene clustering in three additional series of experiments . In these experiments , we provided the simulation framework with unclustered architectures assembled from the genetic building blocks of S . cerevisiae chromosomes VI , VII and X ( Supplementary Figure 5A in Text S1 and data not shown ) . Even in large chromosomes with chromosome VII- and X-like sizes , highly significant levels of essential gene clustering were reproducibly established . Moreover , similar results were obtained when using an externally linked mating type locus as well as when using lower rearrangement rates over longer evolution periods ( r = 10−5 for 200 , 000 generations , Supplementary Figure 5A in Text S1 ) . Taken together , these in silico experiments demonstrate that deleterious mutations inactivating important genes can provide a sufficient driving force to reproducibly evolve chromosome architectures resembling their natural counterparts with respect to essential gene distributions , meiotic recombination hotspot distributions and MAT-centromere linkage . Therefore , a recurrent exposure to lethal mutations can select for genome architectures in order to account for the associated load in the context of a sexual cycle . Any evolutionary process , when successful , should generate individuals that perform better under the conditions of their evolution than their ancestors . In order to determine the level of success of our evolutionary simulation , we performed survival competition experiments between the evolved genomes discussed above and random chromosomes or yeast chromosome IX . We observed a strong fitness advantage of +MAT genomes evolved at R = 1 . 0 when competing with random chromosomes and yeast chromosome IX ( Figure 8B ) . As expected from the presence of large MAT-associated essential gene clusters in these genomes , the fitness advantage results for a wide range of mutation rates ( R≥10−3 ) and for the entire inbreeding/outbreeding domain . The genomes of +MAT populations evolved at R = 0 . 1 also performed significantly better than random architectures , but only slightly better than the chromosome IX architecture , with the exception of mixed breeding ratios and high mutation rates ( R = 1 . 0 ) , for which chromosome IX performed better . In the statistical average , low-R +MAT populations exhibited almost the same overall performance as the chromosome IX architecture ( Figure 8C ) . To further expand this analysis , we also performed competitions of the chromosome X-like products of the evolution experiment shown in Supplementary Figure 5A in Text S1 with random architectures as well as with the actual S . cerevisiae chromosome X , using a chromosome description derived from [42] . The results of the competition experiments were qualitatively comparable to those obtained for chromosome IX ( Supplementary Figure 5B and 5C in Text S1 ) . The reproduction of our results in the context of chromosome X is particularly striking , since the digitalized chromosome X architecture exhibits the highest level of essential gene clustering of all sixteen yeast chromosomes ( see Table 1 in Text S1 ) and therefore constitutes a particularly challenging opponent for the evolution products in the survival competition . We conclude that , in the context of our reference life cycle , architectures with chromosome IX-like purging evolve in the regime characterized by the lower mutation rate . The high-R regime promotes the evolution of single large gene clusters , the extreme of which represents achiasmate meiosis . So far , we have investigated the correlation between lethal mutations in essential genes and the parameters that govern population fitness with respect to crossing over and population genetics via simulations of simple life cycles . There are also other processes that may influence the lethal load present in populations ( see also Discussion ) . In particular , mating type switching followed by mating of daughter cells ( termed “haplo-selfing” ) with their respective mothers is a prominent feature of the life cycle of S . cerevisiae ( but not of all yeasts , see Discussion ) . Mating type switching leads to a homozygous diploid , but only if it involves a haploid genome that is free of lethal mutations . Hence , if occurring at significant rates , haplo-selfing would be expected to decrease the lethal load in populations . However , if the mutagenic load in the population is too high , there is only a small probability of generating viable diploids by haplo-selfing . Comparing pre-loaded chromosomes with and without haplo-selfing revealed a sharp transition of the competitive advantage in favor of non-switching populations for a load higher than two lethal mutations per diploid ( Supplementary Figure 6 in Text S1 ) . In this regime , already a small percentage of diploidisation via haplo-selfing ( 2% ) constituted a strong disadvantage . To assay the impact of haplo-selfing on the fitness of different chromosome configurations , we performed several analyses . First , we compared the mutational robustness ( Rmax ) of chromosome IX and of random chromosomes for different levels of haplo-selfing . A high haplo-selfing rate of 50% leads to a significant increase in the mutational robustness Rmax , both for random architectures and for chromosome IX . At the lower rate of 10% haplo-selfing , the advantage of the chromosome IX architecture remained , but the mutational robustness decreased as compared to the situation without haplo-selfing ( see Supplementary Figure 2A and 2E in Text S1 ) . We further performed competitive advantage experiments of chromosome IX vs . random chromosomes at 10% haplo-selfing and for mutation rates between 10−3 and Rmax . In the absence of a mutagenic pre-load , we noticed the emergence of a phase transition in parameter space at high mutation rates , indicating a region that is dominated by random architectures ( Figure 9 ) . However , even in this extreme scenario the chromosome IX architectures , on average , still performed better than the random architectures ( 57% competition wins of chromosome IX vs . 43% competition wins of random architectures ) . Using S . digitalis , we determined the influence of the meiotic crossover rate of chromosomes on fitness and on the ability to purge mutational load . In yeasts , crossover rates vary considerably , from 0 ( achiasmate meiosis ) up to approximately 20–40 crossovers per chromosome in S . pombe . The genomic mean for all S . cerevisiae chromosomes is 5 . 6 crossovers per meiosis ( derived from the genetic map , www . yeastgenome . org ) [46] . This value was confirmed by the direct assessment of crossover frequency and distribution [42] . The chromosome-specific number of crossovers scales linearly with the number of genes ( R2 = 0 . 89 ) , in a range of 2 . 5–9 crossovers for individual chromosomes ( Supplementary Figure 7 in Text S1 ) . We used random genomes of different sizes ( 250–1 , 500 genes ) to assess the mutational robustness Rmax as a function of crossover frequency and inbreeding fraction . This revealed an increase of Rmax with increasing crossover rates , reaching saturation levels ( 95–99 . 5% ) in the range of 4 . 5–8 crossovers . The observed variability of Rmax as a function of the inbreeding fraction was minimal in the 95–99 . 5% saturation interval ( Figure 10A ) . A slight dependency on chromosome length was apparent ( Figure 10A , inset ) . Direct competition of chromosome IX populations subjected to the natural crossover rate with chromosome IX populations at modified crossover rates demonstrated that crossing over rates higher or lower than the naturally observed average lead to a decrease in the fitness advantage ( Figure 10B; see Supplementary Figure 8 in Text S1 for a plot of the average performance ) . Moreover , when assessing competition experiments of chromosome IX versus random architectures as a function of the crossover rate , chromosome IX performed best in a regime of yeast-like crossing over rates ( Figure 10C; see Supplementary Figure 8 in Text S1 for a plot of the average performance ) . This suggests that the natural rates of meiotic crossovers in S . cerevisiae are adapted to handle deleterious load . Haploid lethal phenotypes may be caused by mutational inactivation of one essential gene , or they may arise as a consequence of cumulative effects of non-lethal mutations in essential and non-essential genes . Dissecting synthetic lethal relationships demonstrated a 0 . 8–4% chance of lethality when two non-essential genes are deleted [47]; a simple extrapolation predicts a 50% chance for a lethal phenotype upon inactivation of approximately 7–14 non-essential genes , assuming scalability ( and ignoring the possibility of non-linear network properties , such as positive epistasis [48] ) . In this case , however , most of the cells would have died before reaching such a high load , due to inactivation of one of the 19% essential genes . This rather empirical analysis might indicate that many haploid lethal phenotypes are caused by inactivation of a single essential gene , as we hypothesized for the purpose of our investigation . In order to test this hypothesis directly , we generated a diploid strain containing MSH2 under control of the weak and fully repressible GalS promoter in order to perform mutation accumulation experiments . Deletion of the mismatch repair gene MSH2 leads to greatly elevated levels of point mutations [49] , [50] . Additionally , reduced sequence specificity for homologous recombination was observed [51] , [52] , leading to increased levels of recombination between similar sequences at separate loci ( ectopic recombination ) . Mismatch repair deficient strains accumulate high levels of lethal mutations , which are however not associated with an increase in gross chromosomal rearrangements [53] . Growth of this conditional mutator strain on glucose-containing medium led to depletion of Msh2 from the cells ( Figure 11A and 11B ) and to the accumulation of mutant phenotypes with reduced viability . In order to restrict the accumulation of mutations to vegetative growth , we shifted the cells to galactose-containing medium to induce MSH2 expression prior to sporulation ( Figure 11A ) , which also prevents aberrant post-meiotic segregation events and ectopic recombination associated with the MSH2 deletion [54] . When grown exclusively on galactose-containing medium , the GalS-MSH2 strain exhibited wild type spore viability ( Figure 11C ) . Upon mutation accumulation for approximately 30–36 generations ( three growth periods of one day each ) , tetrad analysis ( n = 400 ) revealed little 3-spore or 1-spore viability ( 8% and 15 . 5% ) . The majority of tetrads with unviable spores contained two viable spores ( 36% ) , which could be caused by single locus events leading to haploid lethality . Alternatively , some 2-spore viability may have resulted from meiosis I non-disjunction . We tested this option by measuring the linkage of the lethal load in 2-spore viable asci to the heterozygous leu2/LEU2 locus , which itself is centromere-linked ( 5 cM ) , and found that the lethal load exhibited on average only partial linkage to leu2 ( 27 cM ) . Since the load locus is different in each analyzed tetrad , only in approximately 25% of all cases crossing over between the centromere/leu2 locus is prevented ( see Text S1 , section “Supplementary Results and Discussion” ) , either due to tight centromere linkage of the load or as the consequence of homologous non-disjunction in meiosis I . Therefore the prominent fraction of tetrads with two viable spores is likely to be caused by freely segregating single mutagenic events affecting an essential function/gene . Importantly , the frequency of tetrads with only one dead spore was low , indicating low cumulative lethal effects of non-lethal mutations recombined into a haploid genome during meiosis . Taken together , these data suggest that losses of genomes associated with random mutagenic events are frequently associated with single events leading to a lethal haploid phenotype . In this scenario , any linked mutation providing an advantage is lost as soon as the lethal mutation becomes exposed , e . g . by haploid growth following meiosis or homozygotisation . However , the genome with the lethal mutation may be preserved via mating with a spore containing the wild type allele . In order to test whether haploid lethal mutations have a high or a low chance to render spores unable to mate , we investigated the rates of diploid colony formation of single dyads using fluorescence activated cell sorting ( FACS ) after different mutation accumulation periods ( Supplementary Figure 9 in Text S1 ) . The observed frequency of diploid colonies did not change significantly as a function of accumulated mutations , when compared to the wild type . If spores associated with lethal mutations were frequently unable to mate – either due to the lethal mutation itself or due to co-accumulated non-lethal mutations – a greater than two-fold decrease of diploid colonies during the course of the experiment would have been expected ( Figure 11C ) . Consistently , we found that only a very minor fraction of viable spores were impaired in mating and that dead spores would still germinate and often form microcolonies ( see Text S1 , section “Supplementary Results and Discussion”; and Supplementary Figure 10 in Text S1 for a histogram of the colony sizes ) . This latter result is also consistent with our observation that the majority of lethal phenotypes are not caused by meiosis I non-disjunction , as spores that lack entire chromosomes usually fail to germinate . Taken together , our results show that the majority of randomly occurring lethal mutations did not prevent the spores from mating .
The selective advantage of clustered genomes arises from multiple effects that cumulatively improve the fitness of such genomes . Initially , the average mutational load ( ) in a simulated population subjected to a mutation rate R increases until an architecture-dependent equilibrium is reached . In this equilibrium , the influx of new stochastic mutations is equal to the outflux associated with loss of individuals from the population ( Figure 2B ) . Upon elimination of an individual , all mutations associated with its genome are also removed from the population ( Supplementary Figure 1A in Text S1 ) . Within the simulation framework , the loss of individuals occurs via two factors . First , the random removal of excess genomes due to the population size cap is entirely neutral , since it affects all evolved sub-populations with the same probability . Second , and most importantly , genomes are lost due to homozygotisation of a mutant locus . The probability of a homozygotisation via a new mutation during the vegetative division is proportional to the mutational load present in a genome , irrespective of the genome architecture . A higher mutational load in a population therefore lowers its fitness in mitosis due to the increased sensitivity to new mutations , even if the heterozygous load is entirely neutral . Therefore , the only process whose outcome can be improved is meiosis and the loss of individuals due to homozygotisation after mating . Since homozygotisation of any single recessive lethal mutation removes the whole genome from the population , including the associated mutations , the effect of clustering can be explained generally by a reduction of the total number of individuals that are lost after mating . This maximization of mutation purging is caused by the evolutionary optimization of the genome in the context of a sexual lifestyle . Supplementary Figure 11 in Text S1 illustrates a simple scenario . Clustering increases the level of genetic load that can be maintained in the genome . In order to obtain a positive overall effect , clustering needs to compensate for increased mitotic losses , which in simulations of random chromosome architectures constitute approximately 20% of all homozygotisation losses during one life cycle . In those simulation scenarios , in which a large EG cluster evolved ( R = 1 . 0 ) , mitotic losses can reach the same magnitude as meiotic losses ( Supplementary Figure 12 in Text S1 ) . In conclusion , the advantage of clustered genomes must arise from an increase in reproductive fitness in mitosis and meiosis . This results from an optimized balancing of all factors that influence the frequency , with which homozygotisation of a lethal load locus occurs ( Supplementary Figure 1B in Text S1 and Supplementary Figure 11 in Text S1 ) . In mitosis , homozygotisation of lethal mutations is dependent on new mutations and the global load in the population , while in meiosis , homozygotisation results after mating . This is influenced by crossover distribution and frequencies , the breeding behavior and the genetic linkage relationship between crossover sites and essential gene loci . In several experiments , we applied mutation rates higher than the average deleterious mutation rate reported for some S . cerevisiae strains [37] , [55] , [56] . However , as noted in the description of the simulation ( see Results , section “Simulation of digital yeast genomes with S . digitalis” ) , we did not assume an exponential growth of our populations , but rather implemented a scenario where species compete for a limited pool of nutrients . In this scenario our implementation of only one round of mitosis between consecutive meioses mimics a scenario of many mitoses without population growth . Although this may not properly describe the growth of a population in a local niche for a short period of time , it certainly does recapitulate longer evolutionary periods that include many consecutive sexual cycles . It is reasonable to assume that natural yeast populations undergo more mitoses than meioses . Hence , if the mutation rate is to be compared to the natural mutation rate , it must be divided by the average number of consecutive mitoses . This number , however , is currently not known . Crossover suppression near centromeres may also be caused by the interference of pericentromeric crossovers with proper meiotic chromosome segregation [57] . Alternatively , it has been proposed that cold centromeres may be caused by a requirement to protect centromeric repeats [58] . These considerations relate to a general issue in evolutionary biology: frequently , several independent or interacting mechanisms exist or are at least conceivable to explain an observation . As a result , it is often difficult to precisely define the scenario in which one or the other mechanism is dominant and how different mechanisms interact with each other . The situation is further complicated , since there are usually exceptions to each explanation ( e . g . species , in which certain aspects are different ) : yeast does not exhibit centromeric repeats; there are species , in which crossovers occur preferentially in centromere-proximal regions [59] , etc . The fact that there are molecular mechanisms that provide support for different scenarios ( crossovers close to or further away from centromeres ) makes it difficult to deduce the causal connection . Exploring whether a specific mechanism is singularly able to provide a valid explanation constitutes one strategy to tackling this issue . In the case of S . cerevisiae , our evolutionary scenario is able to recapitulate the evolution of crossover suppression and essential gene enrichment in pericentromeric regions . Our simulation provides evidence that lethal mutations and chromosome evolution interact . But , of course , there is no record for the true historical events that enforced this constellation . The complete answer could be given by exploring the ( derived ) genomes of all ancestors of S . cerevisiae along with the evolution of the molecular mechanisms that govern centromeric crossover protection ( which is still subject to investigation ) . While we took advantage of recent data from Mancera et al . [42] in many experiments involving the digitalization of natural yeast chromosomes , our digital implementation of chromosome IX was derived using the hotspot mapping data from Gerton et al . [32] . In the meantime , additional hotspot data sets have been published [60] , [61] . Using data sets about double strand breaks to predict the distribution of crossovers is accompanied by the limitation that DSB frequencies do not translate linearly into crossover frequencies . In this sense the Gerton et al . data set ( which was the only data set available at time we began our analysis ) was very helpful [32] , since it provided a correct description of cold centromeres , as confirmed by the yeast genetic map . Two more recent studies [60] , [61] detected more DSBs near centromeres , which , however , do not convert into crossovers ( www . yeastgenome . org , see Text S1 , section “Supplementary Results and Discussion” for a quantitative assessment of the cM-to-kb relationship in pericentromeric regions ) . These studies do not deviate so much from the Gerton et al . study in the rest of the genome ( except for the subtelomeric regions , which are not so relevant in our context ) . Natural isolates of S . cerevisiae were reported to exhibit large diversity in terms of mutagenic load , yielding a significant fraction of isolates with low to extremely low spore viability . Some proportion of this load , which must have accumulated during mitotic growth and preserved during inbreeding [62] , [63] , may have been caused by mutator phenotypes . Widespread occurrence of natural variation that can give rise to mutator phenotypes has been reported [64] , [65]; this or a similar type of variation may be the reason for the formation of natural yeast strains with highly unstable karyotypes [66] . These may be caused by impaired recombinational repair , as indicated by a certain dependency on Rad52 [67] . Interestingly , these strains can give rise to meiotic offspring with stable karyotypes . Based on these considerations we speculate that genomic rearrangements , which are associated with the formation of new species or lineages , are frequently also associated with high mutation rates that occasionally – and in particular during the periods that shape the genome – exert a selective pressure by means of high levels of lethal mutations . The yeast evolutionary history governs several hundred million years and includes many species with ( at least nowadays ) predominantly diploid life cycles [19] , [68] , [69] . Although the average frequency of meioses is unknown ( from the general perspective as well as for individual species ) , a more than sufficient number of sexual cycles and genomic rearrangements must have taken place to allow for the maintenance and selection for particular genome architectures . In yeast , mating is a highly favored process that occurs whenever two cells of opposite mating type meet − in dense populations or on the level of the spores of a tetrad [31] , [70] . Even considering that S . cerevisiae sometimes aborts the formation of one to three spores during sporulation ( due to limited availability of nutrients ) , the overall formation of spores with a mating partner available from within the same tetrad is well above 80% for a broad range of conditions [31] . This circumstance as well as the possibility of mating of spores from different tetrads indicates that mating type switching is most likely not the dominant way of diploidisation in S . cerevisiae . Thus , it appears safe to assume that the relative effects we observed at a level of 10% haplo-selfing represents an overestimation of the actual situation , rather than an underestimation . Mating type switching was one of the key inventions of lifestyle variation that emerged around the time point of the whole genome duplication [68] , [71] . One may wonder about the reasons for this evolutionary invention in the first place , and what causes its recurrent secondary loss ( mating type switching deficient S . cerevisiae may constitute up to 10% of the strains isolated from natural wine fermentation , [72] ) . Lifestyles seem to evolve towards a preference for diploid stages [73] , in which heterozygosity can be maintained . An associated need for efficient purging of load may have influenced the evolution of mating type switching . Occasionally , when mutation rates become too high or meiotic cycles too infrequent , a resulting high lethal load may also be able to select against mating type switching . Diploidy in the context of lethal load can enforce the evolution of very high levels of intratetrad mating in unicellular eukaryotes [23] , as observed in the non-switching pre-WGD duplication yeast species Saccharomyces kluyveri [74] and Saccharomycodes ludwigii [45] , [75] . These considerations together do not exclude additional or other roles for mating type switching and intratetrad mating , but they suggest that lethal mutations and mutagenic load may frequently accompany diploid life cycles in unicellular organisms . Additional mechanisms exist that influence the lethal load in diploid populations . Loss of heterozygosity associated with mitosis acts in a position-specific manner with increased frequency further away from the centromere . It could therefore provide an additional reason for essential gene enrichment in pericentromeric regions . However , its frequency is low ( approximately 0 . 5 to 10 per 10 , 000 cell divisions in young cells and 50 to 500 in old mother cells [76] ) and it is therefore unlikely to significantly influence the global lethal load of the populations . In a race for beneficial mutations upon exposure to new conditions , essential genes would effectively reduce genetic drift , if their inactivation caused a significant fitness reduction in the diploid organism . Consequently , the chance of acquiring new mutations not immediately accessible to the original genome would decrease , e . g . reducing the frequency of beneficial mutations that are accompanied by a mutation that causes the loss of an essential function . The occurrence of deleterious mutations is predicted to select for the evolution of properties that increase robustness , even if the evolved genomes have a lower maximum fitness in the mutation-free environment . This is compensated by the higher mean fitness of the variants present in a mutant population , known as “survival of the flattest” [77] . Support is provided by studies of viruses and bacteria , but also by studies using digital organisms [78] . An important body of literature is summarized in Wilke and Adami ( 2003 ) [79] ) . In yeast , several mechanisms exist that account for buffering mutational load . One is global buffering or positive epistasis of the fitness reduction in combinations of non-essential gene deletions [48] . Essential genes exhibit less expression noise than non-essential genes [80] , [81] . Batada and Hurst [36] have proposed that the evolution of essential gene clustering was driven by their accumulation into chromosomal regions of low average nucleosome occupancy ( open chromatin ) , which are domains with lower expression noise and which coincide with the domains of low meiotic recombination [82] . Although this idea is intriguing , no support has been provided that this scenario results in a fitness advantage sufficient to select for the relocation of an essential gene into a cluster . S . cerevisiae grows predominantly as diploid . Since the individual intrinsic noise from one copy of a gene is uncorrelated to the noise from the other copy , a lower total noise level is present in diploids as compared to haploids [83] . Selection of low noise for essential genes would thus be less effective in diploids than in haploids . Only about 9% of heterozygous deletion strains of essential genes exhibit haplo-insufficiency in S . cerevisiae [84] , indicating that low noise may be particularly important in a diploid situation , where one copy of an essential gene is inactivated due to a mutation . This would minimize noise in the context of an overall reduced protein level . This result would suggest that a significant interaction may exist between the low noise model and purging of mutations from diploids , which could act synergistically towards improving the clustering of essential genes . Proliferation of tumor cells depends on the inactivation of tumor suppressor genes by subsequent multiple mutations . Due to early-acquired mutations in DNA mismatch repair and other pathways required for genetic stability [10] , [85]–[87] , mutator phenotypes have the potential of accelerating the progression of cancer development by enhancement of the variability upon which Darwinian selection can act [9] , [10] , [88] . In this scenario , evolved robustness towards haploid lethal load is an important factor that has implications for the understanding of genetic instability in cancer development . Purging of deleterious load depends on the tendency to accumulate load and on the presence of alleles that increase the rate at which mutations occur ( i . e . mutator alleles ) . The robustness inherent to the genome architecture of yeast therefore indicates that high mutation rates and genetic load have played an important role during evolution , when the ability to deal with deleterious load co-evolved with better suited genome architectures . Altogether there is increasing evidence that deleterious load is common to evolving yeast populations and that many aspects of the cellular physiology evolved to allow survival as fit but “flat” species .
A diploid strain carrying chromosomal MSH2 under control of the inducible GalS promoter was constructed [89] in the well-sporulating SK1 genetic background [90] . The cells were constantly maintained under GalS inducing conditions ( YP-galactose/raffinose ( YP-Gal ) , see also Text S1 , section “Supplementary Methods” ) and exhibited wild type spore viability . In order to allow random mutations to occur , the GalS promoter was repressed using glucose while cells were grown for 24 , 48 and 72 hours ( 24 hours≈10–12 generations ) using serial dilution . At each time point , an aliquot of the cells ( 5·107 cells ) was plated on YP-Gal for 16 hours . The cells were then washed off , plated on a sporulation plate ( 1% KAc , 0 . 02% raffinose , 0 . 02% galactose ) and incubated for 40 hours at room temperature . Ascus formation occurred in all cases with frequencies >99% . After disrupting the asci , FACS sorting was used to spot single spores and dyads on large YP-Gal plates ( 1536 spores or 384 dyads per plate ) . Sorted spores and dyads were analyzed for colony formation ( viability ) and mating type ( MATa , MATα and no mating type = diploid cells ) using mating type tester strains and a halo assay ( Text S1 , section “Supplementary Methods;” and Supplementary Figure 9 in Text S1 ) . The viability of the spores from 400 different tetrads was analyzed by tetrad dissection . The source code of the simulation S . digitalis is provided as Protocol S1 . A detailed description of the simulation is included in Text S1 , section “The Computer Simulation S . digitalis;” and in Table 2 in Text S1 .
|
Sexual life cycles constitute a costly alternative to vegetative modes of reproduction . Two categories of hypotheses seek to explain why sexual life cycles exist: those investigating the selective advantages that have driven the evolution of individual parts of this life cycle and those rationalizing the advantages sexual life cycles may offer as a whole , e . g . , in extant species . Sex and recombination can be understood as efficient ways to interact with mutations and their consequences . Mutations occur at random and are mostly either deleterious or neutral . A prominent hypothesis suggests that sex and recombination are advantageous since they enhance the purging of such deleterious mutations and create individuals with a lower than average deleterious load . Deleterious mutations should co-determine the parameters that govern recombination of genomes in meiosis . Using an evolutionary computer simulation of diploid , unicellular sexual populations , we show that recessive lethal mutations can drive the evolution of chromosome architectures , in which essential genes become genetically linked into clusters . Evolved architectures exhibit structural properties and fitness similar to digitized yeast chromosomes and provide mutational purging capabilities superior to those of randomly generated or unclustered architectures . Our study demonstrates the importance of sexual cycles in the context of lethal mutations .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"computational",
"biology/population",
"genetics",
"evolutionary",
"biology/evolutionary",
"and",
"comparative",
"genetics",
"microbiology/microbial",
"evolution",
"and",
"genomics",
"genetics",
"and",
"genomics/chromosome",
"biology",
"computational",
"biology/evolutionary",
"modeling",
"genetics",
"and",
"genomics/population",
"genetics"
] |
2009
|
Evolution of Mutational Robustness in the Yeast Genome: A Link to Essential Genes and Meiotic Recombination Hotspots
|
In summer 2014 , an autochthonous outbreak of dengue occurred in Tokyo , Japan , in which Yoyogi Park acted as the focal area of transmission . Recognizing the outbreak , concerted efforts were made to control viral spread , which included mosquito control , public announcement of the outbreak , and a total ban on entering the park . We sought to assess the effectiveness of these control measures . We used a mathematical model to describe the transmission dynamics . Using dates of exposure and illness onset , we categorized cases into three groups according to the availability of these datasets . The infection process was parametrically modeled by generation , and convolution of the infection process and the incubation period was fitted to the data . By estimating the effective reproduction number , we determined that the effect of dengue risk communication together with mosquito control from 28 August 2014 was insufficiently large to lower the reproduction number to below 1 . However , once Yoyogi Park was closed on 4 September , the value of the effective reproduction number began to fall below 1 , and the associated relative reduction in the effective reproduction number was estimated to be 20%–60% . The mean incubation period was an estimated 5 . 8 days . Regardless of the assumed number of generations of cases , the combined effect of mosquito control , risk communication , and park closure appeared to be successful in interrupting the chain of dengue transmission in Tokyo .
Dengue fever is a vector-borne viral disease , caused by dengue virus ( DENV ) and transmitted by Aedes aegypti and Aedes albopictus [1–4] . There are four antigenically related serotypes ( DENV1 to 4 ) , and first infection with one serotype is often self-limiting [5] , eliciting specific acquired immunity . However , following primary infection with one serotype , infected and recovered individuals remain prone to secondary infection with other serotypes that could induce a clinically severe form of infection , including dengue hemorrhagic fever and dengue shock syndrome [6 , 7] . DENV infections are seen mostly in tropical and subtropical countries , but nonendemic areas in temperate regions are also at risk [5 , 8–11] , owing partly to the changing ecological dynamics of vector abundance , perhaps induced by global warming [12] and an increased volume of international travel [13] . Although vaccination is partly underway in endemic countries , concrete plans for the prevention and specific treatment of dengue have yet to be fully established [7 , 14] . Moreover , the possible drawbacks of vaccination in low-endemicity settings remain a controversial issue [7] . Japan is in a temperate zone and dengue is not endemic in the country . However , Japan has experienced a steady increase in the number of imported cases , mainly from South and Southeast Asian countries [15–17] . Despite only sporadic abundance of Aedes aegypti , the species Aedes albopictus is widespread across Honshu Island and all western parts of Japan , theoretically allowing for chains of dengue transmission to exist [18–23] . Dengue was eliminated in the country by 1945 and transmission has not been observed for 70 years [24]; however , a German traveler visiting Japan in the summer of 2013 was later diagnosed with DENV2 infection upon returning to Germany [25] . In 2014 , autochthonous transmission was confirmed in metropolitan Tokyo [26–31] , resulting in a large outbreak involving a total of 160 confirmed cases , a shockingly high incidence for a previously dengue-free nation . It is now recognized that Japan is indeed at risk of dengue outbreaks during the summer season , indicating that a certain risk exists for the summer Olympic Games in 2020 . This threat requires concrete planning for possible countermeasures in the event of another outbreak . The 2014 dengue outbreak in central Tokyo was caused by a single serotype , DENV1 , which showed high homology with the predominant circulating serotype in Southeast Asia [26] . There were two notable characteristics of this outbreak . First , of the total 160 people with a confirmed dengue diagnosis , 129 had visited or worked near Yoyogi Park , a national park that belongs to Shibuya Ward . Shibuya is a special ward that is a major commercial and business center . Shibuya has one of the busiest railway stations in the city , Shibuya Station , about 1 km from the park . Dengue transmission was concentrated in the park , where relatively high vector competence ( mean biting rate 7 . 1 bites per person per 8 minutes ) was observed [32] . Second , a few local residents seemed to remain for extended periods in Yoyogi Park ( i . e . , perhaps homeless people ) , and these people appeared to have been infected at a higher frequency than other individuals [33]; however , the role of these individuals in amplifying transmission as primary cases has not been verified . Once the outbreak was recognized in late August 2014 , concerted efforts were made to contain spread of the virus , including mosquito control targeting both adults and larvae , disseminating news of the outbreak via mass media , communication of dengue risk by experts to raise public awareness , and even a total ban on entering the park . Descriptive documents with details of the outbreak are available but are mostly limited to Japanese language [26 , 28 , 30 , 33 , 34] . However , the effectiveness of interventions during the outbreak remains an important epidemiological question . Mathematical modeling techniques are powerful tools for retrospective assessment of disease outbreaks . These methods include objective measurement of transmission such as the effective reproduction number , i . e . , the actual average number of secondary cases generated by a single primary case , sometimes in the presence of interventions [35 , 36] . In the present study , we formulated a mathematical model and derived a likelihood function , aiming to estimate the effectiveness of interventions during the 2014 outbreak . Because several countermeasures were implemented on different dates during the outbreak , we calculated the effective reproduction numbers to assess the effectiveness of these interventions .
In Japan , DENV infection is categorized as a category IV disease , according to the Infectious Disease Law; thus , all physicians are required to notify diagnosed cases to the government via local health centers upon diagnosis [37] . The clinical characteristics of infection include ( i ) high-grade fever , which is typically biphasic , following an incubation period of 2–14 days [38 , 39]; ( ii ) headache , reddish face and/or conjunctivitis; ( iii ) general fatigue; ( iv ) muscle and joint pain , followed by ( v ) generalized rash that starts on the chest and abdomen . For patients with these characteristics , a physician must confirm the diagnosis via virus isolation , PCR method , detection of nonstructural protein 1 , elevated IgM antibodies against DENV , or plaque reduction neutralization testing . During the 2014 outbreak , notifications as well as details of the outbreak and interventions were summarized in an official report by the Tokyo metropolitan government [34] . In the present study , we retrieved the dates of exposure and illness onset from this report . The date of exposure was calculable because many cases were associated with exposure at Yoyogi Park . Individuals who did not have a history of visiting Yoyogi Park had a history of being bitten by mosquitoes in one of several other parks in the nearby Kanto region . However , the date of exposure was partly censored . Of the total 160 reported cases , four were excluded owing to the absence of information about illness onset ( i . e . , people who were serologically diagnosed , including local residents of the park ) . The remaining reported cases were statistically categorized into one of the following three groups ( Fig 1 ) : Individuals in Group 2 visited Yoyogi Park on two or more consecutive days . Accordingly , the reported cases were categorized into Groups 1 , 2 , and 3 ( S1 Fig ) . In this outbreak , the actual primary case , which must be an imported case , was not identified . We defined the initial calendar day as t0 , i . e . , day 0 of the epidemic , which was not empirically observed . We assumed that exposure among secondary and subsequent generations of human cases began to occur after t0 . The earliest observed date of exposure was 4 August 2014 , and we denoted d0 as the gap number of days between t0 and 4 August ( i . e . , d0 = 4 August minus t0 ) . Table 1 shows the timeline of the outbreak . The index case , i . e . , the first identified clinical case , had illness onset on 9 August and the confirmatory diagnosis was made on 26 August . Mosquito control and public notification of the outbreak started on 28 August; because infected adult Aedes mosquitos continued to be detected in Yoyogi Park , the government decided to close the park on 4 September . Here , we describe the transmission dynamics of DENV in Tokyo using a mathematical model . First , we decomposed the generation time of DENV infection into two parts , i . e . , ( i ) the time from illness onset in an infected human to secondary transmission in another human via a mosquito , denoted as the random variable , tTrans , and ( ii ) the time from infection in a human to their illness onset , again denoted by the random variable , tlP , corresponding to the intrinsic incubation period . Let wt and fs be the probability mass functions ( pmf ) to which random variables tTrans and tlP follow , respectively , and we assumed that both functions would be derived from the cumulative distribution functions of gamma distribution , G ( s ) , i . e . , fs = G ( s; μlP , σlP ) −G ( s−1; μlP , σlP ) for s>0 . The parameters μlP and σlP are the mean and standard deviation of the ( intrinsic ) incubation period in humans . Similarly , we assume that the parameters μTrans and σTrans would determine the mean and standard deviation of the pmf wt . Then , the pmf of the generation time , gt , defined as the time from infection in a human to infection in its secondary human case via a mosquito , can be modeled by convolution , gt=∑τ=0tft-τwτ . ( 1 ) As we did not know the parameters that govern fs and wt , we jointly estimated them using other epidemiological parameters ( see below ) . The abovementioned model does not explicitly account for the lifespan of the female Aedes species , which is considered to be 6 weeks or longer [40] . For simplicity , we ignored this matter , because the time scope of the Tokyo epidemic in 2014 was from 9 August to 7 October 2014 , consistent with the average lifespan; thus , the empirically estimated generation time was sufficiently shorter than the lifespan . Using gt , we devised a generation-dependent epidemiological model , which has been described elsewhere [41]; see S1 Text for the derivation of the generation-dependent model . In this model , we assumed that the epidemiological dynamics described by the generation-dependent model are what is expected in the absence of interventions . We defined the unobserved index case as generation 0 . The index case produces generation 1 , and the size of generation 1 is R0 cases with the relative timing of infection following gt ( i . e . , following the infection time of the index case , there would be R0 gt cases on day t ) . Subsequently , generation 1 produces R1 cases of generation 2 , where R1 is the reproduction number of generation 1 , and there would be R0R1 ( g * g ) t cases as a function of time since index case t , where * is the convolution operator . If there are only two generations ( excluding generation zero ) , the expected value of the incidence at t days since infection in the index case is R0gt + R0R1 ( g * g ) t . Continuing this procedure through generation 4 , and normalizing the quantity by the cumulative number of cases , we obtain the probability density function of infection , h ( t ) , as h ( t ) =R0 ( gt+R1 ( g*g ) t+R2R1 ( g*g*g ) t+R3R2R1 ( g*g*g*g ) t ) R0+R1R0+R2R1R0+R3R2R1R0 , ( 2 ) where Rm-1 denotes the reproduction number of generation m , describing the average number of secondary cases in generation m produced by a single primary case in generation ( m−1 ) , in the absence of interventions . R0 + R1R0 + R2R1R0 + R3R2R1R0 represents the total number of cases , considering up to the fourth generation . R0 is cancelled out and we have h ( t ) =gt+R1 ( g*g ) t+R2R1 ( g*g*g ) t+R3R2R1 ( g*g*g*g ) t1+R1+R2R1+R3R2R1 . ( 3 ) Eq ( 3 ) describes the epidemic curve of infection as the probability distribution in the absence of interventions . It should be noted that h ( t ) is a function of the time of infection , not illness onset . Though only arithmetically , R0 can be calculated by dividing the observed cumulative number of cases by ( 1 + R1 + R2R1 +R3R2R1 ) . Additionally , we incorporated the effectiveness of the interventions implemented during the 2014 outbreak . As Table 1 shows , mosquito control started on 28 August ( T1 ) ; subsequently , Yoyogi Park was closed on 4 September ( T2 ) . We wished to assess the effectiveness of these two interventions separately . We used relative reduction in the reproduction number , ε ( t ) , defined as follows: ε ( t ) ={1t<T1ε1T1≤t<T2ε1ε2T2≤t . . ( 4 ) Using h ( t ) ε ( t ) , we described the observed epidemic dynamics . Normalizing the product , h ( t ) ε ( t ) , we obtained the probability density of the epidemic curve of infection , u ( t ) =h ( t ) ε ( t ) ∑τ=0Th ( τ ) ε ( τ ) , ( 5 ) where T is the last date of the outbreak . Using the parameterized incidence function , u ( t ) , the effective ( or instantaneous ) reproduction number R ( t ) is calculated as the estimator of the renewal equation [42 , 43] , i . e . , R ( t ) =ut∑τ=0tut-τgτ . ( 6 ) Note that u ( t ) is now described on a daily basis; thus , ut represents the daily probability on day t . We did not know the first day of exposure , t0; therefore , we varied d0 within a plausible range , using 3 to 9 days as a theoretically possible range [34] . Assuming that the incubation period was independently and identically distributed , our mathematical model was formulated using the convolution of infection probability , ut , and distribution of the incubation period , fs . We let te be the date of exposure and ts be the date of illness onset . Unknown parameters θn = {μlP , σlP , μTrans , σTrans , R1 , R2 , … , Rn−1 , ε1 , ε2} were estimated using a maximum likelihood method . The exact number of generations was unknown; thus , we fit three different models with a variable number of generations ( i . e . , n = 2 , 3 , and 4 excluding generation 0 ) and later compared the Akaike information criterion values with a correction for small sample size ( AICc ) and mean squared error ( MSE ) . We did not consider additional generations , because more generations unrealistically required shorter duration of the extrinsic incubation period ( EIP ) . As a sensitivity analysis , we compared models with and without ε1 and ε2 , i . e . , models in which ( i ) two effectiveness measures are jointly estimated ( i . e . , ε1≠1 and ε2≠1 ) , ( ii ) only the park closure effect is estimated ( ε1 = 1 and ε2≠1 ) , ( iii ) only mosquito control and public awareness campaigns are factored in ( ε1≠1 and ε2 = 1 ) , and ( iv ) no effect of control measures are taken into account ( ε1 = 1 and ε2 = 1 ) . To this end , we fixed the generation time distribution , using parameters informed from the best model with ε1≠1 and ε2≠1 and jointly estimated only Ri and these effectiveness parameters , comparing AICc values across different models . For case i ∈ Group 1where the case has exact dates of exposure tie and symptom onset tis [44] , the likelihood of observation is Li1 ( θn;tie , tis , d0 ) =uz ( tie , d0 ) ftis-tie , ( 7 ) where z ( τ , d0 ) = τ − t* + d0 + 1 and t* represents the first exposure time ( 4 Aug 2014 ) . For case j in Group 2 , the case has an interval-censored observation for exposure date tje , ranging from the first date of visit EjL to the last date of visit EjR ( i . e . , tje∈ ( EjL , EjR] ) [44 , 45] . The likelihood function for case j in Group 2 is Lj2 ( θn;EjL , EjR , tjs , d0 ) =∑τ=EjLEjRuz ( τ , d0 ) ftjs-τ . ( 8 ) For case k in Group 3 , the exposure can take place from time 0 to the onset of symptoms ( i . e . , tke∈[t0 , tks] ) , Lk3 ( θn;tks , d0 ) =∑τ=t0tksuz ( τ , d0 ) ftks-τ . ( 9 ) In addition , cases in Group 1 were used to estimate the incubation period , i . e . , LIP ( μIP , σIP;te , ts ) =∏i=1n1ftis-tie , ( 10 ) where nm is the total number of cases in Group m = {1 , 2 , 3} . In Eq ( 10 ) , we did not consider interval-censored data because these are already reflected in Eq ( 8 ) with exposure distribution u ( t ) during the interval . The total likelihood function is L ( θn;te , ts , EL , ER , d0 ) =∏i=1n1Li1∏j=1n2Lj2∏k=1n3Lk3∏i=1n1LiIP . ( 11 ) For the estimation of unknown parameters θn , we used the method of maximum likelihood estimation to minimize the negative likelihood L ( θn; te , ts , EL , ER , d0 ) . We did not impose any constraints for the range of parameters . The 95% confidence interval ( CI ) of the effective reproduction number was computed with a parametric bootstrap method . We let H ( θ* ) be the Hessian matrix for estimated values θ* . The 100 sets of parametric bootstrap samples were generated from the multivariate normal distribution with the mean and covariance , the latter of which was obtained with diag ( H-1 ( θ* ) ) . Simulating 100 times , 2 . 5th and 97 . 5th percentile values of the resampled distribution were used to calculate the 95% CI . All statistical analyses were conducted using R 3 . 5 . 1 ( R Development Core Team [46] ) . In the present study , we analyzed data that are publicly available [34] . As such , the datasets used in this study were de-identified and fully anonymized in advance; the analysis of publicly available data with no identifying information does not require ethical approval . Individual datasets for the date of exposure and illness onset are available in S1 Table; these also served as the data source of the epidemic . The R programming code for the two-generation model has been made publicly available at https://github . com/Biomath-2019/Dengue .
Fig 2 shows the incubation period distribution , confirming that the observed and estimated frequencies agreed well . The mean and standard deviation of the incubation period were estimated at 5 . 8 ( 95% CI: 5 . 5 , 6 . 0 ) days and 1 . 8 ( 95% CI: 1 . 6 , 2 . 1 ) days , respectively . Fitting three different models with a different assumed number of generations , the observed temporal patterns were well captured overall ( Fig 3A ) . Identifying a few best-fit models with a different number of generations , the most likely first date of exposure was in the range 26–28 July 2014 , dating back 7–9 days from 4 August 2014 ( S2 Table ) . Regardless of the assumed number of generations , the first generation ( i . e . , generation 1 ) coincided with a small peak in the epidemic curve on 14 August; generation 2 was considered to be primarily responsible for the highest peak at the end of August ( Fig 3B–3D ) . Only when generation 4 was assumed to have existed , generation 3 was considered responsible for a small peak on about 7 September 2014 . Table 2 summarizes the parameter estimates derived from the identified best model , given the assumed number of generations . The mean generation time from DENV infection in a human to another DENV infection in a human via a mosquito bite ( i . e . , μIP + μTrans ) was estimated at 17 . 2 days , 16 . 1 days , and 12 . 4 days for the assumed number of generations 2 , 3 , and 4 , respectively . As the assumed number of generations increased , the estimated mean and standard deviation became significantly smaller , reflecting that the explainable transmission dynamics greatly varied by the assumed number of generations . The reproduction numbers in the absence of interventions were greater than the value of 1 by generation 2 ( i . e . , R1 ) for the three models , and by generation 3 ( i . e . , R2 ) if the assumed number of generations was 4 . It should be noted that if the number of generations was 2 , then the reproduction number of generation 3 ( i . e . , R2 ) was not estimated but must have been zero ( owing to the absence of generation 3 ) . It should be noted that the latest estimate of Ri was below the value of 1 and this was not surprising , because the epidemic came to an end in October 2014 . The mosquito control and public awareness campaigns from 28 August 2014 were considered to have definitely reduced transmission , with secondary transmission reduced by an estimated 30%–70% , with a wide confidence interval ( Table 2 ) . Using the AIC value as the weight , the ensemble estimates of ε1 and ε2 following model averaging over different assumed numbers of generations were 0 . 6 and 0 . 6 , respectively . Despite these reductions , it appeared that the estimated effective reproduction number did not decline below the threshold value of 1 ( Fig 4 ) . However , park closure in combination with the mosquito control and awareness campaigns successfully reduced the reproduction number . The relative reduction in the effective reproduction number owing to the closure of Yoyogi Park was estimated to be 20%–60% , again with a wide confidence interval ( Table 2 ) ; however , the combined effect was estimated to be as large as a 44%–88% reduction in the reproduction number . These findings were robust for the assumed number of generations . We conducted model comparisons to assess the importance of accounting for the effectiveness of the abovementioned control measures . The sensitivity results are summarized in S3 Table . It appeared that the latest estimate of the reproduction number was sensitive to the presence of effectiveness parameters ( ε1 and ε2 ) . Regardless of whether the assumed number of generations was 2 , 3 , or 4 , AICc values of the model with both ε1 and ε2 were minimal ( AICc = 1855 . 1 , 1856 . 0 , and 1855 . 2 for models with 2 , 3 , and 4 generations , respectively ) , outperforming models without one or both effectiveness parameters . S2 Fig compares the reproduction number for two-generation model over time , hypothetically examining counterfactual scenarios in which the relative reduction in the reproduction number during park closure , ε2 ( and mosquito control , measured by ε1 ) is assumed at 1 . If ε2 was 1 , the time at which the reproduction number declines would be delayed , indicating that the park closure has played an important role in reducing the transmissibility .
In the present study , we performed a retrospective epidemiological assessment of interventions to control the 2014 dengue outbreak in Tokyo , using a limited number of confirmed cases ( n = 160 ) . Because the anticipated number of generations of infection was limited , and also because we sought to conduct careful evaluation of the causal association between the timing of interventions and the effective reproduction number Rt , we did not use the structured compartmental modeling approach ( e . g . , using ordinary differential equations ) . Instead , we developed a novel method to directly parameterize the incidence of infection , by convoluting the incidence of infection with the incubation period , allowing us to precisely incorporate the timing of interventions and observe their effect on virus transmission dynamics on a given day . The proposed method suitable for application ( i ) when the generation structure is imaginable from a published estimate of the mean generation time or visually identifiable from the observed epidemic curve , ( ii ) when the time of infection needs to be modeled in relation to the timing of interventions , and ( iii ) when the observed data include doubly interval-censored data . As a consequence of our analyses , the effectiveness of mosquito control , dengue risk communication to elevate public awareness , and closure of the focal area of transmission were objectively evaluated in relation to the effective reproduction number of DENV infection . In practical terms , what the present study adds to the literature is that in the case of the 2014 dengue outbreak in Tokyo , all control measures that we explored ( i . e . , mosquito control , public awareness campaigns , and park closure ) acted as essential factors governing the observed patterns of the epidemic . This notion is supported by our model comparisons in which both ε1 and ε2 were required to act as free parameters , to better describe the observed epidemic dynamics . Of these interventions , mosquito control and raising public awareness were not sufficiently effective to break the chain of transmission , as they maintained Rt>1 , although the reproduction number exhibited a decreasing trend due to decrease in the observed incidence . Although mosquito control from 28 August 2014 was very intensive , DENV-positive Aedes were detected after these measures had been implemented [34] . To fully halt virus transmission , the combined effect of mosquito control , public awareness campaigns , and park closure was needed for a substantial reduction in Rt; this joint reduction effect was estimated to be 44%–88% . Therefore , according to our model results , we can conclude that control of the dengue outbreak at local level in Tokyo was essential to describe the empirically observed data and successful in reducing transmissions . It should be noted that the park closure effect includes not only the prevention of exposure among susceptible visitors but also removal of infectious hosts , including local residents of the park , from the focal area of transmission . In developing the generation-based modeling approach , the proposed system was quantified with estimated mean generation time from 12 to 17 days . Considering the published length of the EIP [38 , 39 , 47] , with summer temperatures in metropolitan Tokyo above 30 °C in August , a mean generation time of about 2 weeks is regarded as a reasonable length or slightly shorter than the temperature-dependent estimate . Our estimate was consistent with an empirical estimate by Siraj et al . [48] , indicating that a mean generation interval of 17 days occurs with the highest probability at 30 °C , although the estimate was obtained for Aedes aegypti and not for Aedes albopictus , the latter of which is abundant in Japan . The mean incubation period of 5 . 8 days is also consistent with the literature [37] , and we attained a finer estimation than that of Ishikawa et al . [30]; those authors used only known 67 intrinsic incubation periods for the estimation ( with an estimated mean of 6 . 3 days ) , potentially resulting in a biased estimate of the variance owing to small sample size . Although we imposed three different assumptions , i . e . , three different numbers of generations , the resulting AIC values were comparable , and we were unable to select the best model . However , the model with two generations alone indicated that the outbreak came to an end with the latest estimate of a generation-dependent reproduction number ( of generation 1 ) as large as 9 . 3; if this model result were true , the reproduction number of the subsequent generation ( generation 2 ) had to abruptly drop to 0 . As mentioned in the Results , regardless of the assumed number of generations , the latest estimate of Ri was below the value of 1 because the epidemic came to an end in October 2014 . However , using the model with two generations , this had to happen quite abruptly , with a drop from 9 . 3 to 0 . Using the model with four generations , the mean generation time had to be 12 days , allowing only about 6 to 7 days from illness onset in an infected person to EIP in a mosquito to biting a susceptible person . Thus , there was some interplay between the assumed number of generations and the resulting estimates , i . e . , as the number of generations increased , the generation time was estimated to be shorter . As long as we cannot specify the exact number of generations , the only approach to address relevant uncertainty is to use multiple models with different numbers of generations , to verify that our practical conclusions about the effectiveness of interventions would not change drastically by varying the number of generations . The present study was not free from limitations . First , this study rested on confirmed dengue cases in patients with symptomatic illness who undertook testing . Febrile patients who had visited Yoyogi Park were advised to seek medical attention during the outbreak; however , a substantial number of asymptomatic infections would have been missed [49] , and the present evaluation was made based only on diagnosed cases . Thus far , we have failed to explicitly estimate the ascertainment rate or number of asymptomatic infections using the observed empirical data . Second , Yoyogi Park was undoubtedly the focal area of transmission , but later transmission occurred in other parks in the Kanto region . In addition to spatial heterogeneity , the exposure behaviors in those parks were not rigorously traced , leading to substantial uncertainty in the empirical data ( i . e . , many cases belonging to Group 3 in our likelihood ) , thus making later epidemic data difficult to be captured by our simple model . Third , qualitatively , local residents of Yoyogi Park were suspected of being amplifiers of transmission . However , it was not possible to trace the behavior of these infected individuals , e . g . , when they left Yoyogi Park and where they went after leaving the park , including whether they moved to another park as a next destination , thus being responsible for causing subsequent cases . It must be remembered that our high estimate of the reproduction number in the early stage of the epidemic could have reflected the existence of superspreaders . Fourth , we ignored environmental and ecological factors in our model [50–52] . The temperature mostly remained stable during the course of the 2014 outbreak; thus , the EIP can be assumed to be approximately stable . However , we ignored rainfall , which could have altered the population dynamics of Aedes species . Fifth , we examined only a limited number of assumed generations; theoretically , there could be six , seven , or an even greater number of generations explaining the observed epidemiological dynamics . However , considering the published EIP , it was implausible that there were six or more generations during the course of the 2014 outbreak . Despite these limitations , we strongly believe that among the existing models , our model provides the most meticulous approach to account for the exact impact of the intervention start date in changing the dynamics of dengue infection over time , because we directly modeled the time of infection in relation to the timing of interventions while maximally using interval-censored data of exposure and illness onset among cases . We successfully captured that impact using convolution of the incidence and the incubation period . By combining mosquito control , public awareness campaigns , and park closure , dengue control during the 2014 outbreak in Tokyo was highly successful . Should a similar event happen in the future , concerted efforts including similarly combined interventions , accompanied by identification of the location of exposure , should be implemented .
|
Evaluating the interventions implemented during an outbreak of mosquito-borne disease is of utmost importance , offering lessons for future control strategies . By retrospectively analyzing data of the first autochthonous dengue epidemic of the 21st century in Tokyo , Japan , we assessed the effectiveness of the interventions . Once a dengue outbreak was confirmed in late August 2014 , the government of Japan took drastic mosquito control measures , targeting both adults and larvae . News of the outbreak was also widely disseminated via mass media along with experts’ recommendations as to how people could avoid the risks of dengue infection . As the outbreak was not immediately controlled , the focal area of transmission , Yoyogi Park , was closed on 4 September . Using a mathematical model , we assessed how well dengue virus transmission was intervened in relation to the start times of interventions . As we incorporated precise timing into the model , we directly modeled the time of infection and accounted for the time delay from infection to illness onset . Thus , we revealed that mosquito control and risk communication measures alone could not interrupt the chain of transmission; however , adding park closure to these interventions was substantially effective in reducing the number of transmissions .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"dengue",
"virus",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"japan",
"pathogens",
"geographical",
"locations",
"microbiology",
"mathematical",
"models",
"animals",
"viruses",
"probability",
"distribution",
"mathematics",
"rna",
"viruses",
"convolution",
"insect",
"vectors",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"mathematical",
"functions",
"aedes",
"aegypti",
"medical",
"microbiology",
"epidemiology",
"mathematical",
"and",
"statistical",
"techniques",
"microbial",
"pathogens",
"disease",
"vectors",
"insects",
"probability",
"theory",
"arthropoda",
"people",
"and",
"places",
"mosquitoes",
"eukaryota",
"asia",
"flaviviruses",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"physical",
"sciences",
"organisms"
] |
2019
|
Assessing dengue control in Tokyo, 2014
|
Processing of the Gag precursor protein by the viral protease during particle release triggers virion maturation , an essential step in the virus replication cycle . The first-in-class HIV-1 maturation inhibitor dimethylsuccinyl betulinic acid [PA-457 or bevirimat ( BVM ) ] blocks HIV-1 maturation by inhibiting the cleavage of the capsid-spacer peptide 1 ( CA-SP1 ) intermediate to mature CA . A structurally distinct molecule , PF-46396 , was recently reported to have a similar mode of action to that of BVM . Because of the structural dissimilarity between BVM and PF-46396 , we hypothesized that the two compounds might interact differentially with the putative maturation inhibitor-binding pocket in Gag . To test this hypothesis , PF-46396 resistance was selected for in vitro . Resistance mutations were identified in three regions of Gag: around the CA-SP1 cleavage site where BVM resistance maps , at CA amino acid 201 , and in the CA major homology region ( MHR ) . The MHR mutants are profoundly PF-46396-dependent in Gag assembly and release and virus replication . The severe defect exhibited by the inhibitor-dependent MHR mutants in the absence of the compound is also corrected by a second-site compensatory change far downstream in SP1 , suggesting structural and functional cross-talk between the HIV-1 CA MHR and SP1 . When PF-46396 and BVM were both present in infected cells they exhibited mutually antagonistic behavior . Together , these results identify Gag residues that line the maturation inhibitor-binding pocket and suggest that BVM and PF-46396 interact differentially with this putative pocket . These findings provide novel insights into the structure-function relationship between the CA MHR and SP1 , two domains of Gag that are critical to both assembly and maturation . The highly conserved nature of the MHR across all orthoretroviridae suggests that these findings will be broadly relevant to retroviral assembly . Finally , the results presented here provide a framework for increased structural understanding of HIV-1 maturation inhibitor activity .
Over 20 antiretroviral inhibitors have been approved for clinical use in HIV-1-infected patients . These drugs fall into several classes , mostly targeting the viral enzymes reverse transcriptase ( RT ) , protease ( PR ) , and integrase ( IN ) . A fusion inhibitor specific for the viral transmembrane envelope ( Env ) glycoprotein gp41 , and an entry inhibitor directed against the viral coreceptor CCR5 are also available . These antiretrovirals , administered in combinations referred to as highly active antiretroviral therapy ( HAART ) , are quite effective and have resulted in striking declines in AIDS-related mortality in treated patients [1] , [2] , [3] , [4] . However , resistance to these compounds , as well as a variety of drug tolerability and related compliance issues , have reduced their benefit in many patients . Given that resistance is likely to become an increasingly large problem over time , discovering new inhibitors that target distinct steps in the viral replication cycle remains a high priority [5] . In addition to the potential therapeutic benefit of such new antiretrovirals , drug discovery efforts are likely to provide novel and important insights into the molecular biology of HIV-1 replication . A promising but still underdeveloped class of anti-HIV-1 compounds are the maturation inhibitors [6] . Maturation is an essential step in the virus replication cycle that is triggered during or shortly after virus release from the infected cell by the PR-mediated processing of the Gag and GagPol polyprotein precursors . The Gag precursor , Pr55Gag , is cleaved into the following mature Gag proteins and intervening spacer peptides ( SPs ) : matrix ( MA ) , capsid ( CA ) , SP1 , nucleocapsid ( NC ) , SP2 , and p6 [7] , [8] . The GagPol precursor , Pr160GagPol , gives rise to the mature viral enzymes PR , RT , and IN . Because of the unique sequence and context of each of the processing sites in Pr55Gag and Pr160GagPol , PR cleaves each site at a different rate , leading to a step-wise proteolytic cascade [9] , [10] , [11] . Maturation involves a major conformational rearrangement of viral proteins within the virion . Most notably , CA reorganizes to form a conical shell that surrounds the viral RNA genome . This conical shell , which exhibits fullerene-like geometry , is constructed from a hexagonal lattice of CA monomers closed off at either end by the incorporation of a specific number of pentamers [8] , [12] , [13] . CA is composed of two structural domains: the N-terminal and the C-terminal domain ( NTD and CTD , respectively ) , which are connected by a short , flexible linker . Located just downstream of the interdomain linker at the N-terminus of the CA-CTD is a ∼20 amino acid sequence known as the major homology region ( MHR ) that is highly conserved across orthoretroviral capsid proteins [14] , [15] , [16] . The MHR adopts a strand/turn/helix fold that likely functions in forming the immature Gag lattice during virus assembly [17] . Previous mutational analyses have demonstrated that disruption of the MHR impairs assembly , maturation , and infectivity of a number of retroviruses including HIV-1 , Mason-Pfizer monkey virus , Rous sarcoma virus , and bovine leukemia virus [14] , [18] , [19] , [20] , [21] , [22] , [23] , [24] , [25] , [26] . Early X-ray crystallography data indicate that the HIV-1 MHR forms a network of hydrogen bonds involving CA residues R154 , Q155 , G156 , E159 , and R167 [17] . SP1 , located downstream of CA , has been proposed to adopt a helical conformation [27] . Based on cryo-electron tomography reconstructions , SP1 is suggested to form six-helix bundles that stabilize each CA hexamer [28] . Some mutations in SP1 disrupt virus assembly , consistent with this region of Gag playing a role in promoting Gag-Gag interactions during virus particle production [29] , [30] , [31] , [32] , [33] , [34] . We and others have characterized the first-in-class maturation inhibitor [3-O- ( 3′ , 3′-dimethylsuccinyl ) betulinic acid [known variously as PA-457 , bevirimat ( BVM ) , or DSB] [35] , [36] , [37] , [38] , [39] . Treatment of virus-producing cells with BVM leads to the accumulation of CA-SP1 , indicating that the compound blocks cleavage of this processing intermediate to mature CA . It is noteworthy that BVM disrupts but does not completely block CA-SP1 processing; some mature CA is generated even at high concentrations of the compound . Propagation of HIV-1 in the presence of BVM in culture gives rise to resistance mutations in the immediate vicinity of the CA-SP1 cleavage site [38] , [39] , [40] , [41] . It is currently hypothesized that BVM binds a pocket located near the CA-SP1 cleavage site , a region for which little high-resolution structural information is available . Direct binding of radiolabeled BVM to immature but not mature HIV-1 particles has been measured [42] , suggesting that the putative binding pocket is eliminated by PR cleavage of the Gag precursor . Some BVM resistance mutations reduced compound binding to immature particles , indicating that at least in some cases resistance is associated with disruption of the putative compound binding pocket [42] , [43] . Gag assembly is required for BVM binding; the compound is able to block PR-mediated CA-SP1 processing in the context of in vitro-assembled Gag but does not disrupt cleavage of non-assembled Gag [38] , [44] . A recent study using photoactivatable BVM derivatives observed direct cross-linking of the compound to residues in the vicinity of the CA-SP1 cleavage site , supporting the hypothesis that BVM interacts closely with this region of Gag [45] . Interestingly , some BVM cross-linking was also detected upstream in the MHR [45] . However , no resistance mutations arose in the MHR during extensive selection experiments [40] . BVM performed well in clinical trials , with significant reductions in viral loads in approximately half of treated patients [46] , [47] . Unfortunately , naturally occurring polymorphisms within SP1 downstream of the CA-SP1 cleavage site ( in the so-called “QVT motif” comprising SP1 residues 6–8 ) were associated with lack of response in other patients [48] , [49] , [50] . In vitro assays confirmed that SP1 polymorphisms , particularly SP1-V7A , reduced sensitivity of HIV-1 to BVM [51] , [52] . This lack of response in a significant percentage of treated patients led to discontinuation of efforts to develop BVM for clinical use . In 2009 , Blair and colleagues reported that a second compound , {1-[2- ( 4-tert-butylphenyl ) -2- ( 2 , 3-dihydro-1H-inden-2-ylamino ) ethyl]-3- ( trifluoromethyl ) pyridin-2 ( 1H ) -one} ( PF-46396 ) , blocked CA-SP1 processing [53] . A single PF-46396-resistance mutation was identified during selection experiments: CA-I201V , located 30 amino acids upstream of the CA-SP1 cleavage site . Intriguingly , BVM and PF-46396 are structurally distinct ( Fig . 1A ) , indicating that compounds from different chemical classes can interfere with CA-SP1 processing . A key question is therefore whether BVM and PF-46396 bind distinct pockets on the assembled Gag multimer or whether they occupy the same binding site . In this study , we sought a more in-depth understanding of the structure-function correlates of maturation inhibitor activity . We first examined whether combined treatment with BVM and PF-46396 would result in a more complete disruption of CA-SP1 processing than observed with either compound alone . Whether the activity of PF-46396 is reduced by polymorphisms in the SP1 QVT motif , as observed for BVM , was also investigated . Currently , no high-resolution structures of the CA-SP1 junction are available , and the precise binding site ( s ) of maturation inhibitors are not known , hence extensive selection experiments were performed to capture the full range of mutations that confer resistance to PF-46396 . In addition to resistance mutations near the CA-SP1 cleavage site that HIV-1 acquires during propagation in BVM , a cluster of mutations in the CA MHR far upstream of the CA-SP1 junction was identified . These MHR mutants exhibited a striking degree of compound dependence and could revert by acquiring second-site changes downstream in CA and in SP1 . The results of this study offer new insights into maturation inhibitor activity and the determinants of maturation inhibitor binding . The findings reported here also provide novel information about the CA-CTD and SP1 in HIV-1 assembly and maturation .
In 2009 , PF-46396 was described as a maturation inhibitor , which , like BVM , disrupts the cleavage of the CA-SP1 processing intermediate to mature CA [53] . To confirm this finding , and to compare the activity of the two compounds , 293T cells were transfected with the HIV-1 molecular clone pNL4-3 , and metabolically radiolabeled the transfected cells in the presence of compounds . The levels of CA-SP1 in virion fractions were determined by quantitative radioimmunoprecipitation ( radio-IP ) . As indicated in Fig . 1B and C , both BVM and PF-46396 treatment of virus-producing cells led to a dose-dependent accumulation of CA-SP1 in virions . Similar accumulation of CA-SP1 was also detected in cell-associated fractions ( data not shown ) . The shape of the dose-response curve ( Fig . 1C ) indicated that PF-46396 is less potent than BVM . For both compounds , CA-SP1 accumulation reached a plateau; this occurred at around 0 . 5 µM for BVM and 5 µM for PF-46396 . These results are consistent with prior calculation of the EC50 of BVM and PF-46396 [38] , [53] . We previously demonstrated that BVM treatment led to the production of virus particles with an aberrant morphology typified by the presence of an electron-dense crescent of Gag located just inside the viral envelope and an aggregate of electron density located in an acentric position in the virion [38] , [40] . This electron-dense crescent represents a remnant of the immature Gag lattice [54] . To investigate the effect of PF-46396 treatment on virion morphogenesis , we performed thin section transmission electron microscopy ( EM ) on pNL4-3-transfected cells treated with the compound . Untreated and BVM-treated samples were included as controls . Virions from untreated cells showed typical electron-dense conical cores ( Fig . S1 , panel a ) , whereas virus from BVM-treated cells displayed the previously described [38] , [40] morphology ( Fig . S1 , panel b ) . Some virions produced from PF-46396-treated cells showed a morphology similar to that of particles from BVM-treated cells ( e . g . , Fig . S1 , panel c , upper right ) . Interestingly , other particles produced from PF-46396 treated cells displayed a hybrid-type morphology , with a conical core and the electron-dense crescent ( blue arrow , Fig . S1 , panel c , lower left ) . Thus , particles from PF-46396-treated cells show greater morphological heterogeneity than those produced from BVM-treated cells ( Fig . S1 , panels c–g ) . As mentioned in the Introduction , we and others previously reported that polymorphisms in SP1 , specifically in the QVT sequence ( SP1 residues 6–8 ) , markedly reduce the sensitivity of HIV-1 to BVM . In particular , the SP1-V7A change almost completely abrogated the ability of BVM to block CA-SP1 processing . To test whether the SP1-V7A mutation similarly diminished the potency of PF-46396 , we compared the effect of BVM and PF-46396 on CA-SP1 processing for SP1-V7A ( Fig . 2 ) . As controls WT and SP1-T8A , a mutant that exhibited full susceptibility to BVM [5] were included . At compound concentrations of 1 µM , WT virus particles from BVM-treated cells showed a ∼70% accumulation of CA-SP1 , whereas WT particles from PF-46396-treated cells contained ∼35% CA-SP1 . Consistent with our earlier report [5] , under these conditions , markedly less ( ∼20% ) CA-SP1 was detected in V7A particles produced in the presence of BVM . In contrast , the V7A mutation had a much smaller , and statistically insignificant , effect on the ability of PF-46396 to block CA-SP1 processing . We also observed that PF-46396 significantly delayed the replication kinetics of V7A ( data not shown ) , consistent with the biochemical analysis demonstrating that V7A CA-SP1 processing is sensitive to PF-46396 ( Fig . 2 ) . As we reported previously [5] , the SP1-T8A mutation did not affect sensitivity to BVM; this mutation likewise did not alter the susceptibility to PF-46396 . Together , these results demonstrate that while PF-46396 is somewhat less potent than BVM , its activity is less affected by SP1 polymorphisms than that of BVM . A notable feature of both BVM and PF-46396 is that their activity plateaus at ∼70% CA-SP1 accumulation for BVM and ∼60% for PF-46396 ( e . g . , Fig . 1B and C ) . The structural dissimilarity between BVM and PF-46396 ( Fig . 1A ) raises the possibility that these two maturation inhibitors might occupy different pockets on the assembled Gag complex . If this were the case , the two compounds could potentially act in an additive or synergistic fashion , providing greater inhibition of CA-SP1 processing than achievable with either compound alone . To test this possibility , virus-producing cells were treated with 0 . 1–1 . 0 µM PF-46396 , 0 . 01–2 . 0 µM BVM , or simultaneously with a constant amount ( 0 . 5 µM ) of PF-46396 and an increasing concentration ( 0 . 01–2 . 0 µM ) of BVM . Consistent with the data in Fig . 1 , both PF-46396 and BVM inhibited CA-SP1 in a dose-dependent manner when either was added alone ( Fig . 3 ) . In contrast , the presence of 0 . 5 µM PF-46396 prevented escalating concentrations of BVM from increasing CA-SP1 accumulation . For example , at 0 . 05 µM BVM alone , an accumulation of ∼45% CA-SP1 was observed , whereas at the same concentration of BVM and 0 . 5 µM PF-46396 only ∼15% CA-SP1 accumulated . Similarly , treatment with 0 . 1 µM BVM resulted in ∼50% CA-SP1 accumulation , whereas this concentration of BVM in the presence of 0 . 5 µM PF-46396 again led to ∼16% CA-SP1 accumulation . The antagonistic behavior of PF-46396 towards BVM activity could be superseded at high concentrations of BVM , such that at 2 µM BVM , even in the presence of 0 . 5 µM PF-46396 , ∼60% CA-SP1 accumulation was observed ( Fig . 3 ) . To examine the effect of PF-46396 on the establishment of a spreading HIV-1 infection , the Jurkat T-cell line was transfected with pNL4-3 in the absence of the compound or at concentrations ranging from 0 . 1 to 10 µM . In the experiment shown in Fig . 4A , and in a number of repeat experiments , virus replication , as measured by reverse transcriptase ( RT ) activity in the culture medium , peaked approximately six days posttransfection in the absence of compound . However , at PF-46396 concentrations of 0 . 5 µM or higher , virus replication was significantly delayed , with peak RT levels occurring at around three weeks posttransfection in 5 ( Fig . 4A ) and 10 µM ( data not shown ) compound . Based on the reasoning that identification of PF-46396 resistance mutations would provide key insights into the residues that line the PF-46396 binding pocket , we determined whether the virus replicating with delayed kinetics had acquired PF-46396 resistance . Viral genomic DNA was purified at the peak of RT activity , amplified Gag and PR coding regions by PCR , and sequenced the PCR products . Five mutations were found: three ( CA-G225D , CA-H226Y , and SP1-A3V ) located near the CA-SP1 cleavage site , and two ( CA-G156E and CA-P157S ) within the CA MHR ( Fig . 4B ) . The CA-H226Y and SP1-A3V mutations were previously identified in our selections for BVM resistance [40] . Similar replication experiments were performed with pNL4-3 clones bearing common polymorphisms in residues 6–8 of SP1 [51] . Several of the mutations that emerged during propagation of WT NL4-3 were again observed in these selection experiments ( Table 1 ) . In addition , several other changes were found: CA-P160L , CA-I201V , and SP1-A3T . CA-I201V was previously identified in selection experiments for PF-46396 resistance [53] , and we previously selected for SP1-A3T during NL4-3 propagation in the presence of BVM [40] . Together , these results demonstrate that propagation in PF-46396 led to the selection of changes not only in the vicinity of the CA-SP1 cleavage site where BVM resistance maps , but also upstream in CA at residue 201 and in the MHR . The appearance of mutations in the MHR is particularly intriguing , given the essential nature of this domain in HIV-1 assembly and maturation ( see Introduction ) . We next sought to determine the replication capacity of the mutants shown in Fig . 4 and Table 1 in the presence and absence of PF-46396 . Specifically , we sought to confirm that these mutations confer resistance to the compound . We also wanted to determine the effect of one of our previously described BVM resistance mutations , SP1-A1V [38] , [40] , on susceptibility to PF-46396 . The mutations were each introduced independently into pNL4-3 , and the Jurkat T-cell line was transfected with WT and mutant molecular clones . Transfected cells were propagated in the presence or absence of PF-46396 at the indicated concentrations ( Fig . 5 ) . As shown earlier ( Fig . 4A ) , WT NL4-3 replicated with a peak at around 1 week posttransfection in the absence of compound , whereas the RT peak was delayed significantly in the presence of PF-46396 , with peaks at around 3 weeks posttransfection at higher compound concentrations . Mutants CA-I201V , CA-H226Y , and SP1-A1V replicated with similar kinetics in the presence and absence of PF-46396 , demonstrating that these mutants are compound-resistant . The CA MHR mutants ( G156E , P157S , and P160L ) and CA-G225D typically failed to replicate in the absence of PF-46396 or replicated with a significant delay relative to WT ( upon acquiring second-site mutations; see below ) . Remarkably , however , the replication of these mutants peaked at approximately 1 week , or soon thereafter , in the presence of high concentrations of the compound . These results demonstrate that the replication of the MHR mutants and CA-G225D is compound-dependent . In some cases ( e . g . , CA-P157S in Fig . 5A ) delayed replication in the absence of PF-46396 resulted from the acquisition of second-site compensatory mutations ( see below ) . The A3V and A3T mutants , which we previously showed were replication defective but enhanced by BVM [40] , also replicated in a PF-46396-dependent fashion ( Fig . 5 ) . As mentioned in the Introduction , the MHR has been shown to play a critical role in HIV-1 assembly and release . We thus sought to examine the assembly/release properties of the compound-dependent MHR mutants ( G156E , P157S , and P160L ) in the presence and absence of compound . The compound-dependent but non-MHR mutant CA-G225D and the PF-46396-resistant mutant CA-I201V were also included in this analysis . 293T cells were transfected with WT or mutant HIV-1 molecular clones , treated or not with PF-46396 , and metabolically radiolabeled for 2 h . Cell and viral lysates were prepared and immunoprecipitated with anti-HIV-1 antiserum ( HIV-Ig ) and viral proteins were visualized by SDS-PAGE and fluorography ( Fig . 6A ) . As shown earlier in Fig . 1B for the WT , PF-46396 treatment led to the accumulation of CA-SP1 in both cell and virion fractions ( quantified in Fig . 6C ) . PF-46396 did not have a significant effect on CA-SP1 processing for the I201V mutant , confirming the replication data indicating that this mutant is PF-46396-resistant ( Fig . 6A and C; [53] ) . Most interestingly , the CA-G156E , P157S , P160L , and G225D mutants showed a severe defect in particle production in the absence of PF-46396 but a substantial rescue of particle assembly and release in the presence of the compound ( Fig . 6A and B ) . For example , CA-G156E virus release efficiency was ∼4% that of the WT in the absence of compound , but ∼60% that of the WT in its presence . The defect in particle production observed for CA-G156E , P157S , P160L , G225D and SP1-A3T mutants was also associated with a modest accumulation of Pr55Gag in cell-associated fractions; this defect was also corrected by the compound ( Fig . 6D ) . CA-SP1 processing was not inhibited by PF-46396 in the context of these mutant Gags ( Fig . 6C ) confirming that they are PF-46396-resistant . The data in Fig . 6A and B demonstrate that PF-46396 rescues a severe defect in particle assembly and release imposed by the CA-G156E , P157S , and G225D mutations . To visualize this phenomenon at the particle morphogenesis level , we performed EM analysis of 293T cells transfected with WT or mutant molecular clones ( Fig . 7 ) . In the absence of PF-46396 , cells transfected with CA-G156E , P157S , and G225D showed the presence of electron-dense patches of Gag at the plasma membrane , but limited evidence of particle release . In contrast , in the presence of the compound , large numbers of particles , many with mature conical cores , were released . These biochemical and EM results , together with the replication data presented in Fig . 5 , establish the marked compound dependence of the MHR and CA-G225D mutations . The partially overlapping pattern of compound resistance observed with BVM and PF-46396 ( [40] and this study ) suggests that these two maturation inhibitors may share portions of a binding pocket . This , in turn , raises the question of whether BVM can rescue the virus assembly and replication defects observed with the PF-46396-dependent mutants ( e . g . , G156E , P157S , and G225D ) . To address this question , we first asked whether BVM can correct the assembly and release defects imposed by these mutations . Cells transfected with WT or mutant molecular clones were left untreated or were treated with 2 µM BVM; virus release efficiency and CA-SP1 processing were evaluated by radioimmunoprecipitation analysis as described above . In contrast to the 8 to 20-fold increase in virus release efficiency observed for these mutants in the presence of PF-46396 ( Fig . 6A and B ) , BVM had little or no effect on their release ( Fig . 8A ) . BVM also had no effect on CA-SP1 processing for these PF-46396-dependent mutants , likely reflecting a major change in the conformation of the compound binding pocket that prevented BVM binding ( Fig . 8A; see Discussion ) . As expected , the previously described [38] , [40] BVM-resistant mutant SP1-A1V was not affected by BVM , whereas CA-SP1 processing was disrupted for the PF-46396-resistant ( but non-dependent ) mutant CA-I201V ( Fig . 8A ) . The inability of BVM to correct the defects imposed by the CA-G156E , P157S and G225D mutations was confirmed in multiple-round replication assays ( Fig . 8B ) . The Jurkat T-cell line was transfected with WT or mutant HIV-1 molecular clones and cells were passaged in the absence of BVM or at 0 . 1 or 1 µM concentrations of the compound . No enhancement was observed in the ability of these mutants to establish a spreading infection . We also tested the PF-46396-resistant but non-dependent mutant I201V; this mutant showed a partially resistant phenotype; its replication was delayed to a lesser extent than that of WT , but it was also more sensitive to BVM than the previously described [38] , [40] BVM-resistant mutant SP1-A1V ( Fig . 8B ) . The MHR consists of a stretch of 20 amino acids highly conserved across diverse genera of Orthoretroviridae . Several residues are particularly well conserved: in HIV-1 CA , these correspond to residues 155Q , 156G , 159E , and 167R ( indicated with a double underline in Fig . 4B ) . Mutations at these positions in the MHR of HIV-1 or the analogous residues in the MHR of other retroviruses have been reported to cause severe defects in Gag processing , assembly , and release ( see Introduction ) . Because , as described above , MHR mutants emerged during selection in PF-46396 , and because these mutants showed a pronounced degree of compound dependence , we asked whether PF-46396 could also rescue the assembly defects imposed by other mutations in the MHR . To this end , we introduced the following mutations at highly conserved MHR positions: CA-Q155N , G156V , E159D , and E159Q , and tested the effects of these substitutions on particle production with and without PF-46396 . The compound-dependent mutant CA-G156E was included in these assays as a control . As expected , all mutations severely affected virus assembly and release ( Fig . 9 ) , with the CA-Q155N and G156V essentially abolishing detectable virus particle production and the G156E , E159D , E159Q mutations reducing virus production by ∼10–20 fold relative to that of the WT . Addition of 5 µM PF-46396 to the producer cells markedly enhanced particle production for CA-G156E ( as shown above ) but failed to rescue particle assembly in the case of the most severely affected mutants ( Q155N and G156V ) . A small and statistically insignificant increase in particle release was observed with the E159D mutant . In contrast , PF-46396 was able to increase particle production by ∼4-fold for the E159Q mutant ( Fig . 9 ) . As we observed for the MHR mutants selected during passaging in the compound , PF-46396 had little or no effect on CA-SP1 processing for these mutants relative to that observed for the WT . These results indicate that PF-46396 is able to rescue the assembly/release defects imposed by a subset of mutations in the CA MHR . The PF-46396-dependent mutants showed delayed replication in the absence or at low concentrations of PF-46396 ( Fig . 5 ) . To evaluate whether this delayed replication was associated with the acquisition of second-site compensatory mutations , we PCR amplified viral DNA from infected cultures at peak RT activity and performed DNA sequencing . Table 2 indicates the second-site changes that arose upon passaging of the CA-G156E , P157S , and G225D in the absence or presence of low concentrations of the compound . The CA-G156E mutant acquired a CA-N193H change . The CA-P157S mutant obtained a CA-G225S or an SP1-T8I change . The CA-G225D mutant also acquired an SP1-T8I substitution , or reverted back to an Asn at residue 225 ( CA-G225N ) . It is interesting to note that we previously reported that CA-G225S acts as a compensatory mutation for SP1-A3V [40] . To assess whether the second-site mutations could rescue the assembly and replication defects exhibited by the PF-46396-dependent mutants , we constructed G156E/N193H , P157S/T8I and G225D/T8I double mutants and carried out virus release and replication assays . We also evaluated the phenotype of the second-site mutants N193H and T8I alone . The CA-N193H mutation by itself did not affect virus release efficiency , but interestingly this mutant was more sensitive to PF-46396-mediated inhibition of CA-SP1 processing than the WT , with approximately 90% accumulation of CA-SP1 at 5 µM of the compound ( Fig . 10A ) . The G156E/N193H double mutant displayed a several-fold increase in virus release efficiency relative to that of the G156E single mutant , but was still released with an efficiency of only ∼30% that of the WT ( Fig . 10A; data not shown ) . Analysis of the T8I single and double mutants ( Fig . 10B ) demonstrated that T8I was able to markedly rescue the assembly/release defect observed with the CA-P157S and G225D single mutants . The double mutants remained insensitive to PF-46396 in terms of CA-SP1 processing . The T8I single mutant was released efficiently but , interestingly , showed a high level of CA-SP1 accumulation even in the absence of PF-46396 . In virus replication assays in Jurkat , the second-site mutations were able to rescue the replication defects imposed by the original PF-46396-dependent mutations ( Fig . 11 ) . As shown above ( Fig . 5A ) , G156E , P157S , and G225D were replication defective in the absence of PF-46396 but replicated in its presence ( Fig . 11 ) . The G156E/N193H , P157S/T8I and G225D/T8I double mutants were replication competent and compound-resistant . The N193H single mutant showed a ∼1 wk delay in peak replication relative to the WT and was sensitive to PF-46396 , consistent with the CA-SP1 processing data ( Fig . 10A ) . In most experiments , SP1-T8I was replication deficient , as one might predict from the high level of CA-SP1 accumulation seen with this mutant . In one experiment , T8I replication was observed in the absence of inhibitor , with a peak of RT activity around day 20 postinfection , presumably reflecting the acquisition of additional change ( s ) ( data not shown ) . The P157S/G225S double mutant replicated with near-WT kinetics and was PF-46396-resistant ( data not shown ) . Together , these data demonstrate that second-site mutations acquired during propagation of the PF-46396-dependent mutants are able to compensate for the replication defects observed for these mutants . The resulting viruses are both replication-competent and compound-resistant . It is significant that these compensatory changes emerged in the same three domains to which compound resistance mapped ( i . e . , the CA MHR , in the vicinity of CA residue 200 , and SP1 ) , suggesting structural and/or functional cross-talk between these three regions of Gag ( see Discussion ) . The above-described PF-46396-dependent MHR mutants display a severe defect in virus particle production that is rescued by second-site compensatory changes far downstream in Gag ( e . g . , SP1-T8I ) . To understand which step ( s ) in the assembly/release pathway are disrupted by these mutations , we examined in vivo Gag multimerization and Gag-membrane binding . Membrane flotation centrifugation assays were performed to evaluate the percentage of Gag that is associated with membrane in virus-producing 293T cells . As a control , we used the non-myristylated Gag mutant 1GA [55] . Glyceraldehyde 3-phosphate dehydrogenase ( GAPDH ) served as a non-membrane-associated control protein . As we reported previously [56] , [57] , [58] , we observed that approximately 50% of WT Gag is present in membrane fractions ( Fig . S2 ) . As expected , GAPDH was located exclusively in the bottom ( non-membrane ) fractions . The non-myristylated 1GA Gag mutant showed minimal membrane association ( 3% of Gag in membrane fractions ) . The CA-G156E and P157S mutants showed reductions in membrane association , but these reductions were not highly significant . As expected given their WT or near-WT virus release efficiency ( Fig . 10A ) the P157S/T8I and T8I mutants showed WT levels of Gag-membrane association ( data not shown ) . We next evaluated these mutants for their ability to undergo multimerization in cells , using our previously reported assay in which Gag multimerization is measured by comparing the amount of Gag immunoprecipitated without denaturation ( in which highly assembled Gag is recognized inefficiently by anti-Gag antibodies ) vs . the amount of Gag recognized after denaturation ( [59]; Materials and Methods ) . Again , non-myristylated 1GA Gag , which does not assemble into higher-order Gag multimers that undergo antibody epitope masking , was used as a control . Under the conditions of this analysis , approximately 60% of WT Gag had undergone higher-order Gag multimerization , whereas only ∼20% of 1GA Gag was epitope masked ( Fig . S3 ) . The CA-G156E and P157S showed defects in Gag multimerization that were comparable to those of 1GA . The assembly competent revertant , P157S/T8I , showed WT levels of Gag multimerization , whereas the T8I single mutant displayed a small but statistically significant increase in assembly . These results indicate that the primary defect in particle production for the MHR mutants is exerted at the level of Gag-Gag multimerization . The data presented in Fig . S3 demonstrate that the PF-46396-dependent mutants are defective for Gag oligomerization , presumably due to disruptions in Gag folding induced by the mutations . The presence of PF-46396 during the entire time course of the analysis ( from transfection through metabolic radiolabeling ) is able to rescue this assembly defect ( e . g . , Fig . 6 ) . To determine whether comparatively brief exposure of these compound-dependent mutants to PF-46396 could trigger particle assembly and release , we examined the kinetics with which this defect could be reversed by PF-46396 . Cells were transfected with WT , CA-G156E or CA-P157S molecular clones then pulse-labeled with [35S]Met/Cys . Cells were then chased for 0 , 30 , or 60 min in cold medium containing 0 or 5 µM PF-46396 . At the end of the chase time , cell and viral lysates were prepared and analyzed as described above . In WT-transfected cultures virus release efficiency did not differ significantly between PF-46396-treated and untreated cultures ( Fig . 12 ) . In contrast , even at the 30 min chase time point , significantly more virus was detected in PF-46396-treated vs . untreated cultures . The difference between treated and untreated cultures increased at the 60 min time point . These data demonstrate that addition of PF-46396 rapidly triggers the assembly and release of the assembly-deficient MHR mutants .
In this study , we demonstrate that resistance to PF-46396 maps to three domains in Gag: the CA-SP1 boundary region , CA residue 201 , and the CA MHR . Several of the resistance mutations in the CA-SP1 boundary region were observed in our previous studies with BVM; however , resistance mutations in the MHR or CA-I201 were not acquired during extensive selections in BVM . These results suggest that although BVM and PF-46396 likely share portions of a binding pocket , distinct upstream contacts are made by PF-46396 ( see structure model , Fig . 13 ) . The proximity of the MHR , CA residue 201 , and the CA-SP1 boundary region is suggested by structural models ( e . g . , [28] in which the MHR and residue 201 are located near the base of the CA CTD from which projects the putative helical bundle of SP1 peptides ( see Fig . 13 ) . The functional cross-talk between these three regions is demonstrated by our observation that second-site changes that rescue the assembly and replication defect of MHR mutants ( G156E and P157S ) map to residue 193 of CA ( CA-N193H ) or residue 8 of SP1 ( SP1-T8I ) . Although MHR mutations arose repeatedly during selection in PF-46396 but never during selection in BVM [40] , in experiments using a BVM analog with a photoaffinity label , some compound cross-linking was detected at the MHR [45] . These results suggest that BVM engages in contacts , albeit perhaps transient or low affinity , with the MHR . The inability of BVM to rescue the replication defects imposed by the MHR mutations could explain why MHR mutations were never observed during selections in BVM . It is interesting to note that in the avian retrovirus system ( Rous sarcoma virus ) , mutations in the spacer peptide downstream of the CA domain was reported to rescue defects imposed by MHR changes [18] . NMR analysis of a fragment of Rous sarcoma virus Gag spanning the CA CTD , the SP , and NC also provided evidence for an interaction between the MHR and SP [60] . These results , coupled with our findings , suggest that interactions between the MHR in CA and downstream putatively helical domains ( SP1 in the case of HIV-1 , SP in the case of Rous sarcoma virus ) may be a general feature of orthoretroviral Gag assembly . The ability of PF-46396 to block the activity of low concentrations of BVM is consistent with a partially shared binding pocket , though a mechanism of allosteric antagonism cannot be excluded at this time . As discussed in the Introduction , naturally occurring polymorphisms in SP1 , most notably SP1-V7A , significantly reduce the ability of BVM to disrupt CA-SP1 processing [48] , [50] , [51] . It was therefore of interest to test whether changes such as V7A likewise interfere with the activity of PF-46396 . Although its potency is lower than that of BVM , we observed that PF-46396 is less sensitive to the V7A change relative to BVM . A simple model to explain this finding is that SP1 residue 7 comprises part of the BVM but not PF-46396 binding site ( Fig . 13 ) . Alternatively , the SP1-V7A polymorphism could alter the conformation of the maturation inhibitor binding site such that the binding or activity of BVM but not PF-46396 is compromised . In any case , the relative lack of sensitivity of PF-46396 to SP1-V7A suggests that maturation inhibitors can be developed that are active against a range of strains containing polymorphisms in SP1 . However , it should be noted that although QVT polymorphisms appear to have little effect on sensitivity to PF-46396 , Blair and colleagues [53] observed that a number of primary clinical isolates were relatively refractory to inhibition by the compound . The genetic basis for this insensitivity awaits further study . A notable feature of several PF-46396-resistant mutants ( CA-G156E , P157S , P160L , and G225E ) is their high level of compound dependence . Most of these changes fall within the CA MHR , a 20-amino-acid sequence highly conserved among retroviral Gag proteins ( see Introduction ) . In the absence of PF-46396 , the MHR mutants that arose during PF-46396 selection are severely deficient for virus particle production . Biochemical assays demonstrate that these mutants are primarily deficient in Gag multimerization . We propose that binding of PF-46396 rescues these multimerization defects , possibly by correcting the impaired folding induced by the mutations . This rescue is specific to PF-46396 , as BVM exerts little or no effect on virus particle production for these MHR mutants and does not rescue their replication . We observe that defective Gag can be triggered to release following relatively short PF-46396 treatments . This compound-induced release may have useful applications in a variety of HIV-1 assembly analyses; for example , in imaging studies . Based on the model presented in Fig . 13 , it is possible that PF-46396 could make contacts with the CA-SP1 boundary region , residue 201 , and the MHR . Because structural studies suggest that these regions of Gag are close to each other in the assembled Gag complex ( e . g . , [28] , [61] ) , in our view this is the most straightforward model that is consistent with available data , including recent BVM binding results [45] . However , alternative models cannot be excluded at this time . For example , it is possible that the relationship between the MHR and maturation inhibitors is indirect . PF-46396 could stabilize the Gag multimer , thereby limiting the ability of PR to access and cleave the CA-SP1 junction . In such an allosteric model , mutations in the MHR would destabilize the Gag multimer , thereby reversing the effect of the compound on multimer stability . The opposing effect on Gag multimerization of MHR mutations and PF-46396 binding could explain the phenomenon of compound dependence . Support for the hypothesis that maturation inhibitors can stabilize the immature Gag lattice is provided by recent cryo-EM findings [54] . Additional details regarding the structure of the MHR and SP1 in the context of assembled Gag , and more information about the residues that engage in direct contacts with maturation inhibitors , will be required to better understand the molecular basis for maturation inhibitor activity and for the compound dependence observed in this study . Previous studies with BVM indicated that the ability of this compound to block CA-SP1 processing requires Gag assembly [38] , [44] . Furthermore , Zhou and colleagues demonstrated that BVM interacts with immature HIV-1 particles , but not with mature virions [42] . These observations are consistent with the idea that Gag assembly creates the maturation inhibitor binding site and that cleavage at the CA-SP1 junction , or at other Gag processing sites , destroys the binding pocket . Recent insights into the structure of the immature retroviral Gag lattice [61] allow one to propose that the putative maturation inhibitor-binding pocket may straddle adjacent subunits in the immature Gag lattice ( Fig . 13 ) . Such a model would explain why PR-mediated processing of monomeric Gag is not disrupted by maturation inhibitors , and why Gag processing during maturation prevents compound binding . Although the SP1-T8I mutation rescues the replication defect imposed by two highly defective , PF-46396-dependent mutants ( CA-P157S and CA-G225D ) by itself the SP1-T8I mutant is severely impaired . SP1-T8I displays a high level of CA-SP1 accumulation even in the absence of maturation inhibitor and is highly replication defective , suggesting that the T8I mutation mimics the effect of PF-46396 both in its ability to block CA-SP1 processing and to rescue the assembly defects elicited by upstream mutations in CA ( i . e . , CA-P157S and CA-G225D ) . The results presented here help to elucidate the structure-function relationship between the CA CTD and SP1 , two domains of Gag that are critical to both assembly and maturation . This study also provides insights into Gag residues that comprise the HIV-1 maturation inhibitor-binding pocket . This information should be useful in the rational design of maturation inhibitors with increased potency and breadth of activity .
BVM was prepared as described previously [35] and used at the concentrations indicated . Lyophilized PF-46396 ( Pfizer ) was suspended in DMSO to generate 10 or 20 mM stock solutions , stored at −20°C , and diluted in culture medium to the concentrations indicated . The Jurkat T-cell line was maintained in RPMI-1640 medium supplemented with 10% ( vol/vol ) fetal bovine serum ( FBS ) , L-glutamine , penicillin and streptomycin . 293T cells were maintained in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% ( vol/vol ) FBS and L-glutamine . Plasmid DNAs were purified with the Qiagen maxiprep kit . Jurkat and 293T cells were transfected with DEAE/dextran and linear polyethylenimine ( L-PEI ) , respectively [62] , [63] . PF-46396 resistance mutations were selected by multi-cycle replication assay using Jurkat T cells transfected with the WT HIV-1 molecular clone , pNL4-3 [64] , in the presence of 0 . 5 , 1 . 0 , 5 . 0 , and 10 µM PF-46396 . Virus replication was examined by RT activity as previously described [65] . To identify resistance mutations , cell pellets were collected on the days of peak RT activity . Genomic DNA was extracted by using the whole-blood DNA purification kit ( Qiagen ) , and the entire Gag-PR-coding regions was amplified by PCR using the primers: NL561F ( 5′-TGCCCGTCTGTTGTGTGACTC-3′ ) and NL2897R ( 5′-AAAATATGCATCGCCCATA-3′ ) [40] . The 2 . 3kb PCR products were purified by ExoSap-IT ( Affymetrix ) and sequenced using the primers: NL645F ( 5′-AACAGGGACTTGAAAGCGA-3′ ) , NL1155F ( 5′-AGGAAACAACAGCCAGgtc-3′ ) , NL1410F ( 5′-GGAAGCTGCAGAATGGGATA-3′ ) , and NL2135F ( 5′-TTCAGAGCAGACCAGAGCCAA-3′ ) . Selection for compensatory mutations to the PF-46396-dependent mutations was carried out as described above [40] . The 4 . 2 kbp Spe I-Sal I fragment from pNL4-3 ( nucleotide 1507 to 5785 ) was subcloned into Bluescript SK ( + ) [pBS ( NL ) ] and mutagenized to generate CA-G156E , CA-G156V , CA-G157S , CA-E159D , CA-E159Q , CA-P160L , CA-N193H , CA-I201V , CA-G225D , CA-G225N , and SP1-T8I mutations . To generate double mutants , CA-G156E , CA-P157S , SP1-T8I clones were subjected to a second round of mutagenesis using pBS ( NL ) CA-N193H , pBS ( NL ) SP1-T8I , pBS ( NL ) CA-G225D subclones to generate pBS ( NL ) CA-G156E/CA-N193H , pBS ( NL ) CA-P157S/SP1-T8I , pBS ( NL ) CA-G225D/SP1-T8I , respectively . To generate pBS ( NL ) CA-P157S/CA-G225S , the 0 . 5 kb SpeI-ApaI fragment from pNL4-3 CA-G225S [40] was subcloned to pBS ( NL ) and mutagenized to generate CA-P157S/CA-G225S . All mutagenesis was performed using the QuikChange site-directed mutagenesis kit ( Agilent Technologies , Santa Clara , CA ) . Following sequence confirmation , the SpeI-SbfI fragment was cloned back into the WT pNL4-3 to create the molecular clones containing the mutation ( s ) described above , which were reconfirmed by DNA sequencing . Radioimmunoprecipitation assays were carried out with some modification of the protocol described in detail previously [62] . Briefly , 293T cells were transfected with WT or mutant pNL4-3 by using L-PEI ( 4 µg L-PEI/µg DNA ) . At 24 hour posttransfection , HIV-1-expressing cells were starved in [35S]Met/Cys-free medium for 30 min and metabolically labeled with [35S]Met/Cys-Pro-mix ( Amersham ) for 2 h . Maturation inhibitors were maintained throughout the transfection and labeling . Viruses were collected by centrifugation at 99 , 000× g for 45–60 min . Cell and virus lysates were heated in the presence of Laemmli sample buffer ( LSB ) ( 5 . 7 µl 2× LSB/100 µl lysates ) , followed by pre-absorption for 2 h at 4°C , and immunoprecipitated with pooled immunoglobulin from HIV-1-infected patients ( HIV Ig ) obtained from the NIH AIDS Research and Reference Reagent Program . Immunoprecipitated proteins were separated on 13 . 5% acrylamide gels by SDS-PAGE , exposed to a phosphorimager plate ( Fuji ) and quantified by Quantity One software ( Bio-Rad ) . 293T cells were transfected with WT or mutant pNL4-3 in the absence or presence of 5 µM PF-46396 or 2 µM BVM . Fixation of cells , preparation of samples , and transmission EM were performed as previously described [62] . The Gag multimerization assay was performed with some modifications of the previously described methods [59] , [62] . Briefly , 293T cells were transfected with WT or mutant pNL4-3/PR− molecular clones . At 24 h posttransfection , cells were metabolically labeled with [35S]Met/Cys-Pro-mix for 2 h after 30 min starvation . Cells were lysed with 2× radioimmunoprecipitation assay ( RIPA ) buffer . Two 100 µl cell lysate aliquots were prepared: one for boiling with 2× LSB to denature Gag multimers , and the other without boiling , followed by pre-absorption , immunoprecipitation with HIV-Ig , and separation by SDS-PAGE . Gag bands were quantified by phosphorimager analysis . The extent of Gag multimerization was determined by calculating the ratio of Gag immunoprecipitated with and without sample boiling . 293T cells were transfected with WT or mutant pNL4-3 molecular clones . At 24 h posttransfection , cells were starved in Met/Cys-free medium for 30 min then metabolically labeled with [35S]Met/Cys-Pro-mix for 20 min . Cells were divided into two fractions , washed , and resuspended in 10% FBS/DMEM containing 0 or 5 µM PF-46396 . Each fraction was further divided into three aliquots and incubated at 37°C . Cells were collected at 0 , 30 , and 60 min chase time points . Cell lysates were immunoprecipitated with HIV-Ig and analyzed as described above . 293T cells were transfected with WT or mutant PR-defective ( pNL4-3/PR− ) molecular clones . Membrane flotation assays were performed as previously described [62] . Briefly , at 24 h posttransfection , cells were disrupted by sonication and post-nuclear supernatants were collected . Sonicated samples ( 160 µl ) were mixed with 0 . 8 ml of 85 . 5% sucrose in a centrifuge tube and overlayed with 2 . 4 ml of 65% sucrose and 0 . 8 ml of 10% sucrose . Samples were subjected to ultracentrifugation at ∼100 , 000× g for >16 h at 4°C . Ten fractions of 0 . 4 ml each were collected from the top of each tube , and mixed with 0 . 4 ml 2× RIPA buffer . Lysates were analyzed by western blotting using HIV-Ig and anti-glyceraldehyde 3-phosphate dehydrogenase ( GAPDH ) ( Santa Cruz Biotechnology ) . Gag bands were quantified by using a BioRad ChemiDoc XRS+ imaging system with Imaging Lab software .
|
Maturation of HIV-1 particles , which occurs as they bud off from the infected cell , is triggered by the step-wise cleavage of the major viral structural polyprotein , Pr55Gag , to individual , mature Gag proteins . The viral protease is the enzyme responsible for Gag polyprotein cleavage . Maturation inhibitors prevent the viral protease from processing Gag at one particular cleavage site , but how they accomplish this is not understood . In this study , the ability of HIV-1 to become resistant to the two structurally distinct maturation inhibitors that have thus far been reported was examined . We found that one of these compounds , PF-46396 , gives rise to resistance mutations that map to three domains in Gag , including a region known as the major homology region ( MHR ) . The MHR is highly conserved among retroviruses and is known to be very important for virus assembly and maturation . These MHR mutants were observed to replicate much better in the presence of PF-46396 than in its absence; i . e . , these mutants are compound-dependent . We were also able to select for second-site mutations in Gag that reversed the replication defects imposed by the MHR mutations . These results define residues in Gag that comprise the maturation inhibitor-binding pocket and also identify regions of Gag that structurally and functionally interact with the MHR .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"chemistry",
"biology"
] |
2012
|
Structural and Functional Insights into the HIV-1 Maturation Inhibitor Binding Pocket
|
Chromosome 9 of Trypanosoma brucei contains two closely spaced , very similar open reading frames for cyclic nucleotide specific phosphodiesterases TbrPDEB1 and TbrPDEB2 . They are separated by 2379 bp , and both code for phosphodiesterases with two GAF domains in their N-terminal moieties and a catalytic domain at the C-terminus . The current study reveals that in the Lister427 strain of T . brucei , these two genes have undergone gene conversion , replacing the coding region for the GAF-A domain of TbrPDEB2 by the corresponding region of the upstream gene TbrPDEB1 . As a consequence , these strains express two slightly different versions of TbrPDEB2 . TbrPDEB2a represents the wild-type phosphodiesterase , while TbrPDEB2b represents the product of the converted gene . Earlier work on the subcellular localization of TbrPDEB1 and TbrPDEB2 had demonstrated that TbrPDEB1 is predominantly located in the flagellum , whereas TbrPDEB2 partially locates to the flagellum but largely remains in the cell body . The current findings raised the question of whether this dual localization of TbrPDEB2 may reflect the two alleles . To resolve this , TbrPDEB2 of strain STIB247 that is homozygous for TbrPDEB2a was tagged in situ , and its intracellular localization was analyzed . The results obtained were very similar to those found earlier with Lister427 , indicating that the dual localization of TbrPDEB2 reflects its true function and is not simply due to the presence of the two different alleles . Notably , the gene conversion event is unique for the Lister427 strain and all its derivatives . Based on this finding , a convenient PCR test has been developed that allows the stringent discrimination between Lister-derived strains that are common in many laboratories and other isolates . The technique is likely very useful to resolve questions about potential mix-ups of precious field isolates with the ubiquitous Lister strain .
Cyclic nucleotide-specific phosphodiesterases ( PDEs ) are crucial players in cyclic nucleotide signalling of eukaryotic cells . Several of the human PDEs have become important drug targets , and all eleven human PDE families are under intense study for the development of new therapeutic PDE-inhibitors [1] . Over the last few years , the PDEs of protozoal parasites have also come into focus as potential targets for new and effective antiparasitic drugs [2]–[4] . The genomes of all kinetoplastid parasites , including Trypanosoma brucei , contain similar sets of genes that code for cyclic nucleotide-specific phosphodiesterases ( PDEs; [2] , [5] ) . Three of these , PDEA ( Tb10 . 389 . 0510 ) , PDEC ( Tb03 . 27C5 . 640 ) and PDED ( Tb03 . 3K10 . 420 ) are single copy genes located on different chromosomes , while two closely similar genes that code for the cAMP-specific phosphodiesterases PDEB1 and PDEB2 are tandemly clustered . In T . brucei , these two genes , TbrPDEB1 ( Tb09 . 160 . 3590 ) and TbrPDEB2 ( Tb09 . 160 . 3630 ) , are located on chromosome 9 , and their open reading frames are separated by 2379 bp . They code for very similar proteins of 930 and 925 amino acids , respectively . Both consist of a less-conserved N-terminal region of about 200 amino acids ( 44 . 8% amino acid sequence identity between TbrPDEB1 and TbrPDEB2 ) , followed by two highly conserved GAF domains [6] , GAF-A ( 94 . 4% identity ) and GAF-B ( 100% identity ) and a catalytic domain ( 90 . 8% identity ) [2] . The GAF-A domains of both proteins bind cAMP ( TbrPDEB1: A . Schmid , unpublished; TbrPDEB2: [7] ) and might function as allosteric regulators of enzyme activity . The precise function and potential ligand specificity of the GAF-B domains are currently unknown . Based on structural analyses with human PDEs that contain GAF domains [8] , they might be involved in dimer formation . Despite the extensive overall sequence conservation between PDEs TbrPDEB1 and TbrPDEB2 , their subcellular localization is distinct . TbrPDEB1 is located predominantly in the flagellum , with which it remains tightly associated after detergent extraction of the cells . In contrast , TbrPDEB2 is mainly located in the cytoplasm as a soluble enzyme , with only a small proportion also locating to the flagellum [3] . Considering the relatively low degree of sequence conservation between the N-terminal regions of TbrPDEB1 and TbrPDEB2 , these regions , and/or the GAF-A domains might contain the signals for intracellular localization . This study reports the occurrence of a gene conversion event between the two tandemly arranged genes TbrPDEB1 and TbrPDEB2 . Gene conversion in T . brucei has so far been mainly studied in the context of variable surface proteins [9]–[11] , where it is the predominant , though not the only mechanism that drives antigenic variation [12] . Homologous recombination and gene conversion are fundamental processes of genome biology that are involved in a broad range of cellular functions including DNA repair , telomere maintenance , DNA replication and meiotic chromosome segregation [13] . Thus , one might safely assume that they play similarly important roles in trypanosomes and are not restricted to the realm of antigenic variation . Depending on organism and cell division mode , the length of gene-conversion tracts varies considerably . In the yeast S . cerevisiae , mitotic gene conversion tracts are often larger than 4 kb , while the meiotic tracts are usually between 1 and 2 kb . In mammals , on the other hand , mitotic gene conversion tracts are usually between 200 and 1000 bp in length [13] . In T . brucei , gene conversion tracts of variable surface glycoprotein genes are usually around 3 kb [9] . Despite a long history of studying antigenic variation [14]–[18] , our understanding of the precise role and the mechanistic details of gene conversion in antigenic variation is still limited . Even less is known about these processes in regions of the genome that are not involved in antigenic variation . Recent data have shown that the efficiency of homologous recombination depends on target length and sequence conservation [19] and suggests that two distinct recombination mechanisms might be active in trypanosomes . Interestingly , T . brucei BRCA2 , a prominent player in homologous recombination , has acquired an unusually high number ( twelve ) of BRC repeats within its N-terminal domain [11] . The current study describes the occurrence of a gene conversion of several hundred bp within the coding region of the TbrPDEB2 gene by the corresponding region of the TbrPDEB1 gene . The gene conversion does not affect the intracellular localization of the TbrPDEB2 gene product . This event is unique for the Lister strain of T . brucei [20] and all its derivatives , but it is not found in other T . brucei strains . The presence of this particular gene conversion serves as a useful genetic marker to discriminate Lister derivatives from other T . brucei strains .
Procyclic trypanosomes were cultured in SDM-79 medium containing 5% FCS [21] , and bloodstream forms were grown in HMI-9 medium containing 10% FCS [22] . The following strains were used: the procyclic strain Lister427 [23] , the bloodstream form of Lister 427-2 ( strain 221; MiTat1 . 2; [24] ) , STIB247 , STIB345AD ( a derivative of EATRO1529 , which was isolated from Glossina pallipides in Kiboko , Kenya in 1969 and cryopreserved after six passages in mice . In 1973 , it was stabilated after five short passages in rats and renamed STIB345 ) , GVR35 ( isolated 1966 in the Serengeti ) , AnTat 1 . 1 [25] , and the 427-derived SM strain [26] . Genomic DNAs of strains 427 variant 3 and TREU927 [27] were generously supplied by Wendy Gibson ( University of Bristol , UK ) , and genomic DNA of the T . b . rhodesiense strain STIB900 was a gift of Barbara Nerima ( University of Bern ) . STIB900 was isolated as STIB704 in Ifakara , Tanzania in 1981 from a male patient . It was cloned and adapted to axenic culture . A detailed pedigree of many trypanosome isolates and derivatives can be found at http://tryps . rockefeller . edu/trypsru2_pedigrees . html as well as in a recent paper [20] . The following primers were used for PCR: TbrB2-for ( 28-mer; 53 . 6% GC , Tm 63°C , specific for TbrPDEB2 ) : C556ACGCCTCTACGATGCTTGAGTCATCAC TbrB1-for ( 26-mer , 57 . 7% GC , Tm 63°C , specific for TbrPDEB1; used for duplex PCR ) G258ATGGAGCACACAATGACGCACGGTG TbrGAFA1-rev ( 29-mer , 51 . 7% GC , Tm 63°C , specific for converted allele TbrPDEB2b , nucleotides 854–827 ) C854CTACAATGCCTGTTCCCTTGGGTATGGA TbrGAFA2-rev ( 27-mer , 55 . 6% GC , Tm 63°C , specific for wild-type allele TbrPDEB2a , nucleotides 852–827 ) G852GCAATACCTGCACCCCTAGGGATTGT As a template , genomic DNA was used in all reactions ( 100 ng/reaction ) . The annealing temperature of the primers was optimized using gradient PCR , and the final cycling protocol was as follows: an initial denaturation step of 4 min at 94°C , followed by 31 cycles consisting of 1 min at 94°C , 1 min at 63°C and 1 min at 72°C , followed by a final extension step of 10 min at 72°C . Typical reactions ( 20 µl ) with single primer pairs contained 100 ng genomic DNA , 300 nM of each primer , 500 µM of each NTP and 1 unit Taq polymerase . Reactions for duplex PCR contained 300 nM each of primers TbrB1-for and TbrB2-for , and 400 nM TbrGAFA1-rev . Procyclic forms of strain STIB247 as a representative strain that is homozygous for TbrPDEB2 , i . e . has not undergone gene conversion of one allele , were used to explore the intracellular localization of TbrPDEB2 . The cells were transfected with a construct that introduces a C-terminal triple c-Myc tag into one copy of TbrPDEB2 [3] , [28] . Transformation and selection of clones were done exactly as recently described [29] . Transfectants were selected with 25 µg/ml hygromycin , cloned and verified by Southern blotting , and protein expression was confirmed by Western blotting . Four independent clones were then used for immunofluorescence microscopy . Immunofluorescence microscopy was done as described previously [3] . Cells were fixed with 4% PFA in PBS and permeabilized with methanol for 10 min at −20°C . For the preparation of cytoskeletons , cells were extracted once with cold MME buffer ( 100 mM HEPES , pH 6 . 9 , 1 mM MgSO4 , 1 mM EGTA ) containing 0 . 5% Triton X-100 for 5 min , prior to the fixation with 4% PFA . The first antibody was a monoclonal mouse anti-c-Myc antibody ( 9E10 , Santa Cruz ) diluted 1∶200 in PBS+2 . 5% BSA ( w/v ) . The secondary antibody was Alexa Fluor 488 conjugated goat anti-mouse polyclonal antibody ( Molecular Probes ) diluted 1∶750 . Coverslips were mounted with Vectashield mounting medium containing DAPI ( Vector Laboratories ) , and slides were analyzed with a Leica DM6000B microscope . About 1×107 trypanosomes were washed once in PBS and then lysed on ice for 10 min in PBS/0 . 5% Triton X-100 , supplemented with protease inhibitor ( Roche Complete Mini , EDTA-free ) . The extracted cells were centrifuged for 10 min at 13 , 000 rpm at 4°C . Supernatants and pellets were analyzed by Western blotting . Gels were transferred to nitrocellulose filters and probed with mouse anti c-Myc 9E10 antibody ( Santa Cruz; diluted 1∶1000 ) . Control antibodies were polyclonal rabbit anti BiP ( endoplasmic reticulum staining; as a marker for Triton soluble proteins; 1∶50 , 000; gift of Jay Bangs , University of Wisconsin , Madison ) and polyclonal rat anti PFR ( as a marker for Triton insoluble proteins; 1∶30 , 000 ) .
In an effort to determine if the observed gene conversion represents an ancient event of the history of T . brucei , and thus can be found in many independent isolates , the genomic DNA of a number of independent strains with different histories were analyzed . The entire open reading frame of TbrPDEB2 was PCR amplified , followed by restriction enzyme analysis of the PCR products . As demonstrated in Fig . 3 , this procedure allowed the unambiguous discrimination between strains that had undergone the gene conversion event and those that did not ( Table 1 ) . The occurrence of gene conversion was initially detected in the procyclic strain 427 , our standard laboratory strain . This strain most likely is a derivative of the Lister427 strain ( http://tryps . rockefeller . edu/trypsru2_pedigrees . html ) , which in turn is probably derived from the Shinayaga III strain isolated in 1956 from cattle in Tanganyika . The other two Lister427 derivatives that were analyzed ( BS221 , the “NewYork single marker” strain derived in 1999 from BS221 [26] , and the ancestral 427 variant 3 strain documented and frozen down in 1964 at the Lister Institute ) were also heterozygous at the TbrPDEB2 locus , containing one allele for TbrPDEB2a and one for TbrPDEB2b . This indicated that the conversion , once it had occurred , was stably maintained . In contrast , a number of other strains and isolates ( AnTat1 . 1 ( isolated 1966 in Uganda ) , GVR35 ( isolated 1966 in the Serengeti ) , STIB247 ( isolated in 1971 in the Serengeti ) , STIB345AD ( isolated in 1969 in Kenya ) , TREU927 [27] and the T . b . rhodesiense strain STIB900 ) had not undergone gene conversion at this locus . In the past , the precise origin of trypanosomal strains has often caused much debate . In particular , accidental mixups between the ubiquitous laboratory strains of the 427 lineage and other strains have remained a lingering concern ( e . g . see http://tryps . rockefeller . edu/trypsru2_pedigrees . html ) . The presence of the gene conversion in TbrPDEB2 of the 427 lineage now provides a rapid and convenient means of strain identification . The presence or absence of the gene conversion can be monitored by using two sets of specific primers ( Fig . 4 ) . The primer set that is specific for the unconverted GAF-A region of TbrPDEB2a ( TbrB2-for and TbrGAFA2-rev ) produces an amplicon of 297 bp from genomic DNA of every T . brucei strain tested . In contrast , the primer set with specificity for the converted gene TbrPDEB2b ( TbrB2-for and TbrGAFA1-rev ) produces a PCR product ( 299 bp ) only with genomic DNA from 427 derivative strains . Thus , these two primer sets represent a convenient diagnostic tool for strain identification in cases where contamination or confusion of a strain with a 427 derivative is suspected . A duplex PCR setup using a combination of the three primers TbrB1-for , TbrB2for and TbrGAFA1-rev ( see Methods ) allows the simultaneous detection of the presence of TbrPDEB1 ( as a positive control ) and the presence or absence of the TbrPDEB2b allele ( Fig . 4B ) . Earlier work from this laboratory using strain 427 has demonstrated that TbrPDEB2 is located partly in the cytoplasm and partly in the flagellum [3] . A recent analysis demonstrated that in the strain used for these experiments , the tag had been integrated into the TbrPDEB2b allele . This raised the question if the observed intracellular localization was specific for the B2b allele , and might be different from the localization of the B2a gene product . The question was all the more pertinent as in many mammalian PDEs , the information for intracellular localization is contained in the N-terminal part of the PDE molecules [8] . To answer this question , one allele of the TbrPDEB2 gene of a strain that has not undergone gene conversion ( procyclic STIB247 , see above ) was tagged in-situ with a 3× c-Myc tag [3] , [28] . Expression of the protein and its intracellular localization were analyzed and compared with the earlier results obtained with the 427 strain ( Fig . 5 ) . Though cell shape and motility are rather different between strain Lister427 and STIB247 , no major differences in intracellular localization of the tagged TbrPDEB2 could be detected between the strains 427 ( containing the gene conversion ) and STIB247 ( no gene conversion ) . In a similar experiment with bloodstream forms of BS221 , either TbrPDEB2a or TbrPDEB2b were tagged with a c-Myc tag . Both gene products showed an identical subcellular localization . In conjunction , these observations indicate that the slight alterations in the amino acid sequence of the GAF-A domain of TbrPDEB2b do not affect the intracellular localization of the enzyme . Furthermore , they demonstrate that the presence of TbrPDEB2 both in the cell body and in the flagellum is not simply due to the presence of two slightly distinct alleles , but rather reflects a genuine property of the enzyme . The distribution of TbrPDEB2a and TbrPDEB2b were further analyzed using Triton X-100 solubilization . Earlier work had shown that TbrPDEB2 is mostly Triton soluble , with a minor fraction remaining Triton-insoluble [3] . To determine if the gene conversion might alter the Triton solubility of the affected TbrPDEB2 allele , the following cell lines were fractionated with Triton X-100: procyclic 247 wild type ( negative control ) , procyclic 247 ( homozygous for TbrPDEB2a , one allele tagged with c-Myc ) , procyclic 427 ( gene-converted TbrPDEB2b allele tagged with c-Myc ) , bloodstream form Lister427 ( gene-converted TbrPDEB2b allele tagged with c-Myc ) and bloodstream form Lister427 ( wild-type TbrPDEB2a allele tagged with c-Myc ) ( Fig . 6 ) . The blots were probed with anti c-Myc antibody , and subsequently with antibodies against BiP ( fully Triton soluble; [34] ) and PFR ( fully Triton insoluble; [35] ) as controls . In all strains , the TbrPDEB2 alleles behaved identically , the majority of the protein being Triton-soluble , with a minor portion remaining in the insoluble pellet . Triton extracted cytoskeletons were also analyzed by immunofluorescence ( Fig . 7 ) . Triton X-100 resistant TbrPDEB2a and TbrPDEB2b both localize along the flagellum . In conjunction , these data demonstrate that the gene conversion between TbrPDEB1 and one allele of TbrPDEB2 does not alter the subcellular localization of the gene product TbrPDEB2b that is produced from the converted allele .
The current study describes a gene conversion event that has occurred between the GAF domains of the tandemly arranged genes for phosphodiesterases TbrPDEB1 and TbrPDEB2 on chromosome 9 of T . brucei . It demonstrates that this gene conversion does not affect the subcellular localization of the gene product TbrPDEB2b . In addition , the study demonstrates that this gene conversion provides a sensitive marker to discriminate Lister427 derivatives from other trypanosome strains . In the T . brucei strain Lister427 , one allele of the TbrPDEB2 gene has undergone a gene conversion which replaces a stretch of the gene with the corresponding region of the upstream gene TbrPDEB1 . The tandemly arranged open reading frames of TbrPDEB1 and TbrPDEB2 are separated by a mere 2379 bp , and they share a high degree of sequence identity , in particular between the regions that code for the two GAF domains and for the catalytic domain . The conversion generated two non-identical alleles of TbrPDEB2 , TbrPDEB2a ( no gene conversion ) and TbrPDEB2b ( converted allele ) . Interestingly , the converted stretch covers precisely the GAF-A domain , so that the gene conversion effectively resulted in the replacement of the entire GAF-A domain of TbrPDEB2 by its homologue from TbrPDEB1 . The mechanism that converted TbrPDEB2 is likely to be gene conversion since the length of the converted tract is short , between 297 and 600 bp ( minimal and maximal converted tracts , respectively ) , and non-reciprocal . The size of the converted tract is considerably larger than the length of the minimal efficient processing segment determined for T . brucei , and well within the range determined for Leishmania , S . cerevisiae or mammals [19] . It is not possible to distinguish if TbrPDEB2 recombined with the upstream allele of TbrPDEB1 on the same or on the homologous chromosome . Also , it is not possible to decide if the altered allele TbrPDEB2b is the result of meiotic or mitotic gene conversion . A mitotic event should probably be favoured in view of the fact that meiosis still has not been demonstrated in T . brucei [36] , and might occur very rarely , if at all . Earlier studies from this laboratory had shown that , in the strain Lister427 , the enzymes TbrPDEB1 and TbrPDEB2 exhibit different subcellular localizations . TbrPDEB1 is predominantly located in the flagellum , where it remains tightly associated with a detergent-resistant structure . In contrast , TbrPDEB2 is partly located in the flagellum and partly in the cell body as a soluble enzyme [3] . Studies with human PDEs [8] , as well as data from our own laboratory with T . brucei ( Luginbuehl et al . , unpublished ) , indicate that the N-terminal regions of GAF-containing PDEs are involved in intracellular localization . The observation that in the 427 strain used for these localization studies [3] , the GAF-A domain of one allele of TbrPDEB2 had be replaced by the GAF-A domain of TbrPDEB1 raised the question if the observed subcellular localization of TbrPDEB2 might have been influenced by this gene conversion . To clarify this , one allele of TbrPDEB2 was C-terminally tagged in procyclic forms of strain STIB247 , a strain that has not undergone gene conversion . The subcellular localization of TbrPDEB2 in this strain was closely similar to that seen in the Lister427 strain . In conjunction , these data demonstrate that replacement of the GAF-A domain of TbrPDEB2b through gene conversion does not alter the intracellular localization of the enzyme . They indicate that the GAF-A domain may not contain the crucial signals for intracellular targeting . To explore the presence of the gene conversion in a number of trypanosome strains , suitable primer pairs were developed for an easy PCR characterization of strains . PCR reactions using individual primer pairs allow an easy detection of the two alleles . A duplex PCR assay using a combination of three primers simultaneously detects TbrPDEB1 ( as a positive control ) and the TbrPBEB2b allele . Both approaches were applied to genomic DNAs of a variety of strains with different histories . The unambiguous outcome of these experiments showed that the gene conversion is only present in the Lister427 strain and all its derivatives , but that it is absent from all other independent isolates tested , including a T . b . rhodesiense strain . This finding makes detection of the gene conversion in TbrPDEB2b a convenient and reliable marker for identifying trypanosome strains , separating Lister427 derivatives from all others . While this may not represent a diagnostic tool of immediate value for field work , the technique is very useful in the laboratory environment whenever a potential mixup of strains or strain origins are an issue . An example is given by the continuous debate if the strain designated s427 is in fact the origin of the Lister427 derivatives ( trypanosome pedigree section at http://tryps . rockefeller . edu/trypsru2_pedigrees . html ) . An other clear example for the potential usefulness of our analysis is presented in reference [37] ( Peacock et al . ) : The observation that a derivative of Lister427 can be transmitted through the fly violates the dogma that this is a monomorphic strain and not fly-transmissible . This immediately begs the question if their Lister427 variant 3 is really a Lister427 derivative . Our analysis has now made this unambiguously clear . This study has demonstrated the occurrence of gene conversion between two closely related and closely positioned PDE genes in T . brucei . The length of the converted fragment corresponds to those found in human gene conversions , and it corresponds precisely to the open reading frame of one protein domain , GAF-A . The conversion was found in only one trypanosome strain , Lister427 . All its derivatives that have been produced over the last 40 years stably maintain this conversion in the heterozygous state . The presence of the converted gene does not alter the intracellular localization of the gene product , TbrPDEB2 . In terms of the genetic mechanisms operating in T . brucei , the identification of a gene conversion event in a gene other than the much studied variant surface glycoprotein genes may provide further clues to the inner workings of the T . brucei DNA repair and recombination machinery .
|
Cyclic nucleotide specific phosphodiesterases are important regulators of cyclic nucleotide signalling in eukaryotes . In many organisms , including humans and trypanosomes , some of these enzymes contain specific domains ( GAF domains ) that bind cyclic nucleotides , and that are involved in the regulation of the catalytic domain . In the parasitic protozoon that causes human sleeping sickness , Trypanosoma brucei , two closely related phosphodiesterases each contain two such GAF domains , GAF-A and GAF-B . Their genes are tandemly located on chromosome 9 , spaced by only a few thousand nucleotides . We here show that a gene conversion event has exchanged the region that codes for the GAF-A domain of the downstream gene by the closely similar corresponding sequence of the upstream gene . This domain exchange has no effect on intracellular localization of the two enzymes . The gene conversion event has occurred in one particular strain of trypanosomes ( Lister427 ) and is found in all its derivatives , but not in any other strain or isolate . The presence or absence of this gene conversion represents a useful analytical marker for the stringent discrimination of Lister427 derivatives from other trypanosome strains .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"genetics",
"and",
"genomics/gene",
"function",
"cell",
"biology/cell",
"signaling"
] |
2009
|
Gene Conversion Transfers the GAF-A Domain of Phosphodiesterase TbrPDEB1 to One Allele of TbrPDEB2 of Trypanosoma brucei
|
Several proteins involved in the response to DNA double strand breaks ( DSB ) form microscopically visible nuclear domains , or foci , after exposure to ionizing radiation . Radiation-induced foci ( RIF ) are believed to be located where DNA damage occurs . To test this assumption , we analyzed the spatial distribution of 53BP1 , phosphorylated ATM , and γH2AX RIF in cells irradiated with high linear energy transfer ( LET ) radiation and low LET . Since energy is randomly deposited along high-LET particle paths , RIF along these paths should also be randomly distributed . The probability to induce DSB can be derived from DNA fragment data measured experimentally by pulsed-field gel electrophoresis . We used this probability in Monte Carlo simulations to predict DSB locations in synthetic nuclei geometrically described by a complete set of human chromosomes , taking into account microscope optics from real experiments . As expected , simulations produced DNA-weighted random ( Poisson ) distributions . In contrast , the distributions of RIF obtained as early as 5 min after exposure to high LET ( 1 GeV/amu Fe ) were non-random . This deviation from the expected DNA-weighted random pattern can be further characterized by “relative DNA image measurements . ” This novel imaging approach shows that RIF were located preferentially at the interface between high and low DNA density regions , and were more frequent than predicted in regions with lower DNA density . The same preferential nuclear location was also measured for RIF induced by 1 Gy of low-LET radiation . This deviation from random behavior was evident only 5 min after irradiation for phosphorylated ATM RIF , while γH2AX and 53BP1 RIF showed pronounced deviations up to 30 min after exposure . These data suggest that DNA damage–induced foci are restricted to certain regions of the nucleus of human epithelial cells . It is possible that DNA lesions are collected in these nuclear sub-domains for more efficient repair .
DNA damage induced by ionizing radiation ( IR ) elicits microscopically visible nuclear domains ( i . e . , foci ) marked by recruitment of certain proteins ( e . g . , 53BP1 ) or by particular modifications such as histone phosphorylation ( e . g . , γH2AX ) or as a result of both ( e . g . , phosphorylated ATM , ATMp ) [1–10] . Radiation-induced foci ( RIF ) are believed to form at or adjacent to sites of DNA damage . However , the use of RIF as an unequivocal indicator of double strand break ( DSB ) is problematic . The readout of RIF is complex as it is based on optical limitations during image acquisition ( e . g . , point-spread function ( PSF ) ) , non-homogeneity of the detector ( i . e . , nucleus ) , and biological kinetics . Our previous work and that of others have suggested that the detection of RIF reflects several factors: ( 1 ) the severity of the damage , ( 2 ) the efficiency of damage recognition , ( 3 ) repair capacity , and ( 4 ) the biological function of the specific RIF proteins [7 , 11–14] . Furthermore , some reports suggest that there are nuclear regions that are excluded from forming RIF . More specifically , in studies using densely ionizing particles that would lead to continuous DSB along their trajectories , nuclei showed discontinuous MRE11 RIF , with large gaps ( >1 μm ) in regions where DNA was present [15] . Finally , others have shown that some types of RIF are not necessarily associated with DSB [12] . In studying DNA damage responses using RIF , how can one interpret results if RIF are not necessarily related to DSB ? To sort out these discrepancies , one could compare the spatial distributions of RIF from different radiation qualities and relate them to the expected energy deposition described by physical attributes . We propose to compare γ-rays and high energy particles ( HZE ) , which lead to very distinct spatial distributions of energy deposition . HZE are high-LET radiation and deposit their energy in random clusters along a linear path [16 , 17] . Their complex physical interactions with cells have been well characterized and therefore can be modeled [18] . Cells exposed to HZE provide an excellent model in which to study the relationship between chromatin patterns and energy deposition since energy deposition , and therefore image analysis , is reduced to essentially 1-D linear profiles in a plane of the nucleus . In contrast , γ-rays are low-LET radiation that deposit energy uniformly in a small volume and thus induce single DSB randomly across the nucleus . While these events are easily modeled , characterizing low-LET RIF spatially is more complex since it requires 3-D image analysis of nuclei . We have previously used parameters determined by fitting data from DNA fragments sizes measured in pulsed-field gel electrophoresis ( PFGE ) experiments to show that radiation-induced DSB are generated as a stochastic process [19–21] . This assumption has led to computation models that simulate the production of DSB in hypothetical spatial geometries [22 , 23] . In this study , we further refine these models to include higher order nuclear territories such as euchromatin and heterochromatin . Artificial microscope images of RIF and nuclei can then be generated by including optical limitations of light microscopy . We then show that we can predict the DNA damage pattern for any given radiation by generalizing the theoretical model to an image-based model . The central point is to use the image-based model to test the controversial equivalence between RIF and DSB . If RIF are in fact DSB observed at a much lower resolution , then we would predict similar spatial distribution and frequencies . However , our results show that within 5 to 30 min following exposure to high-LET and low-LET radiation , RIF distributions deviate from the predicted DSB distribution . We further show that RIF non-randomly locate in specific regions of the nucleus . This suggests that nuclear organization modulates the DNA damage response of human epithelial cells , which has important implications for understanding DNA damage response and repair mechanisms .
In this approach , the 3-D space was divided into cubic pixels of the size equal to that of microscopy image ( i . e . , 0 . 16 μm pixels ) . DNA in the simulated nucleus was arranged into two types of intermittent bands: dense regions of DNA based on random-walk geometry ( heterochromatin ) , and low-density homogenous regions ( euchromatin ) . DNA double strand breaks ( DSB ) were simulated by Monte Carlo simulations for single traversal of 1 GeV/amu Fe ions or for exposure to 1 Gy of low-LET radiation [24] . Theoretical DSB are absolute , whereas the visualization and quantitation of RIF are subject to optical limitations during image acquisition . Therefore , to more closely approximate RIF from theoretical DSB , DSB locations were blurred by applying a Gaussian filter with σ = 0 . 16 μm , determined by the PSF of the microscope . The resulting images were similar to images collected experimentally as illustrated in Figure 1 . Applying the Gaussian blurring ( Gaussian convolution ) to the DSB frequency image produces images with foci-like objects ( Figure 1 ) , which we refer to as pseudo-foci ( pRIF ) . pRIF reflects the appearance of DSB at light microscope resolution should they bind enough antibody to emit sufficient fluorescence to be detected . Both high-LET ( 1 GeV/amu Fe ion tracks ) and low-LET simulated images are depicted and compared with real images labeled for the DNA damage marker γH2AX ( Figure 1A and 1B , respectively ) . One can appreciate from Figure 1A the fact that many close-by DSB along high-LET tracks cannot be resolved and end up appearing as large foci , a phenomenon that has been reported previously on experimental data [7 , 11] . To validate whether the frequency of theoretical pRIF occurring in synthetic nuclei are comparable to actual measurements , the frequency of RIF was measured for different DNA damage markers ( 53BP1 , γH2AX , and ATMp ) , and within the first hour following exposure to 1 Gy of either low-LET or high-LET radiation . The measured frequencies are shown in Table 1 . For high-LET radiation , we observed excellent agreement between pseudo- and measured RIF , leading to a maximum of 0 . 73 RIF/μm 4 . 5 min following exposure to 1 GeV/amu Fe . On the other hand , consistent with our previous findings and those of others [7 , 11–15] , the maximum measured frequencies for low-LET RIF in 3-D volume occurred 30–60 min after exposure and was 60% lower than predictions . Unfortunately , comparing pRIF to RIF at such late time points is difficult to interpret as DSB repair is significant over the first hour following irradiation . Comparing at earlier time points is not ideal either , as RIF frequencies at 4 . 5 min post-IR were even lower ( 70% lower than prediction , unpublished data ) . Note that the PFGE data used to predict DSB in our model do not distinguish DSB from heat-labile sites [25] . However , removing DSB from these sites would not be sufficient to have predictions that match measurements for low LET . Another point illustrated by these simulations is the effect of the optical PSF ( see Materials and Methods ) on visualizing DSB by light microscopy . For low-LET simulations , there is little variance between predicted DSB and pRIF . In this case , a 1:1 correspondence between DSB and pRIF is expected since such sparse damage events remain separate after optical blurring . On the other hand , for HZE , the frequency of pRIF and theoretical DSB , both generated in synthetic nuclear images , differ by 30% . The 30% loss between DSB and pRIF is due to clustered DSBs that are not resolvable by light microscopy . Since RIF frequencies match pRIF frequencies for HZE , experimental RIF in this case must represent clustered DSBs at a resolution lower than the original scale of DNA breaks ( nm versus μm for microscopy ) . Interestingly , this also suggests that if damages are more complex or span over a larger DNA range ( i . e . , cluster of DSBs ) , it will rapidly induce RIF ( within 5 min ) , whereas a single DSB may not always lead to RIF and/or may have a slower formation kinetic . pRIF predictions presented previously are based on DNA patterns from hypothetical nuclei , modeled at the nm scale . As described in methods , DSB are predicted to be dependent on DNA density . As such , any given nucleus has a unique DNA imaging pattern and a unique set of spatial probability for radiation-induced DSB . Therefore , we cannot directly predict the DSB patterns in real images and compare their spatial distributions to RIF using a theoretical model of the nucleus . To answer such need , we introduce here an imaging methodology that can predict DSB location for any given DNA nuclear pattern from real images . This methodology is based on the same Monte Carlo concepts described in Materials and Methods for the generation of DSBs in artificial nuclei: i . e . , the probability of a DSB at a given location is proportional to the DNA density at the same location . This statement is true at high resolution , if each pixel could contain a single DSB . However , this rule may not be true at the sub-micron resolution of a microscope , where each pixel encompasses large amounts of DNA , where neighboring pixels are slightly correlated [26] , and where individual DSBs are not always resolvable . We illustrate our imaging approach using high-LET radiation data , where RIF tracks demarcate the damaged nuclear regions . Within these tracks , spatial 1-D profiles of DNA density and the number of foci ( Nspot ) can be determined ( see Figure 2A ) . Assuming that the probability of a RIF at a given location along the track is proportional to the DNA density at that location , one can then compute the probability per unit pixel intensity to have a focus as follows: where DNA ( i ) is the DNA density at position i along the indexed pixel of the track . The probability PDSB of a DSB at any given pixel location along the track is then: If this probability is greater than a random value taken between 0 and 1 , then a focus is generated at this location . Applying this approach to all the pixels along the track , we can generate a set of new foci referred to as “reshuffled foci” ( see Figure 2B ) . The validity of this image methodology was tested on the set of Fe track simulations described previously in Table 1 . The damage pattern along the artificial tracks was characterized by measuring distances between consecutive pRIF ( Figure 3B and 3C ) . Reshuffling pRIF position led to spatial distributions similar to the original pRIF ( Figure 3A ) and thus confirmed that this image manipulation is an accurate way to predict damage distribution in a microscope image . On the other hand , even though the main shape of the pRIF distribution was conserved in the reshuffled pRIF distribution , the smaller variations were lost . These irregularities in the distribution probably reflect the lack of uniformity of DNA density at a much higher resolution since the same irregularities were also apparent in the DSB distribution predicted by modeling . To summarize , the probability of generating DSB remains proportional to DNA density at a lower resolution ( i . e . , sub-micron ) , and therefore DNA density measured by light microscopy along a track can be used as a DNA damage probability . As we validated in silico our imaging approach to predict DNA damage patterns along the HZE track , we next compared actual RIF with predicted RIF in situ for the same radiation quality . The nuclear dye 4′ , 6-diamidino-2-phenylindole ( DAPI ) binds to AT base pairs or to single strands by electrostatic interaction . Therefore , one can assume as a first-order approximation that the pixel intensity in a DAPI image is proportional to the DNA concentration at that location . We therefore used DAPI as an indicator of DNA densities in real images . Primary antibodies to ATMp , γH2AX , or 53BP1 and fluorescently labeled secondary antibodies were used to detect RIF . Our results show that distribution of foci along tracks in cells irradiated by 1 GeV/amu Fe deviate from truly random distribution . This deviation increases with time independently of the marker used for damage . This is illustrated in Figure 4 for γH2AX , where the distance distribution between consecutive foci is compared with the distribution of reshuffled γH2AX foci at 4 . 5 and 30 min following exposure to 1 GeV/amu Fe . As early as 4 . 5 min post-IR , only 60% of the RIF distribution correlated with the distribution of predicted damages ( i . e . , reshuffled RIF ) . We predicted that the majority of damages would be less than 1 μm apart . Instead , the majority of measured RIF were more than 1 μm apart . The exclusion of close-by foci in the experimental data also increased with time . Figure 5A summarizes these results for all the time points ( 4 . 5 , 11 . 5 , 31 . 5 , 61 . 5 min ) by plotting the average correlations measured between predicted ( i . e . , reshuffled foci ) and measured distance distribution for the different DNA damage RIF . All RIF show the same trend with an increasing deviation from random distribution over the first hour following exposure to 1 Gy of 1 GeV/amu Fe . Note that the predicted RIF distance distribution is based on reshuffling the exact same number of detected RIF for each individually analyzed track . Therefore , the loss of close-by foci cannot be attributed to their diminishing frequency ( shown in Figure 5B ) . These results show in fact an increase in organization suggesting DNA damage clustering into self-excluding sub-regions of the nucleus within an hour following exposure to HZE , as suggested by others [27] . Given the deviation of RIF from random distribution over time , we can then ask if foci relocate in regions of the nucleus with specific morphological features . To do so , we introduce a set of imaging parameters that ascertains the position of foci with respect to DNA density . This set of parameters can also be measured in any spatial dimension ( i . e . , line profiles , surfaces , or volumes ) . Figure 6 illustrates the approach on a given nucleus ( i . e . , center slice of a nucleus—DAPI stain ) . Using automatic spot detection ( see Materials and Methods ) , we consider the center of RIF as the brightest pixel in its vicinity . One can then compute the mean DNA density signal at the centers of all RIF in one nucleus and normalize it to the mean nuclear DNA density to get a relative DNA density value at these locations . We thus define the relative density of DNA at the foci locations as follows: where , I ( i ) is the intensity at pixel location i , Nfocus is the number of foci , and Nnucleus is the number of pixels in the nucleus ( note , I = focus refers to the brightest pixel in an identified focus ) . The DNA gradient is a good indicator of edges between high and low DNA density regions . Therefore , we can also compute the relative position of foci with respect to the surface of dense DNA regions by evaluating the DNA gradient value where foci are detected . This leads to the parameter Rgrad as described in Figure 6 . Similarly to Rdna definition , Rgrad is the mean DNA gradient at the foci location normalized to the mean gradient over the full nucleus , defined as follows: where ∇I ( i ) is the Euclidian norm of the gradient vector of I ( i ) at pixel location i . Note that the mask used to compute Rdna and Rgrad is based on a conservative segmentation of the nucleus . The contour of the nucleus is defined as a region well inside the nucleus ( i . e . , 0 . 48 μm inward of the nuclear boundary ) . This conservative segmentation is necessary to remove edge effect when computing the gradient of the nuclear DAPI image . If foci preferentially locate in bright regions of the DNA , Rdna will be greater than 1 . Similarly , if foci locate preferentially at the interface between bright and dim regions of the DNA , Rgrad will be greater than 1 . We can further test these concepts by measuring Rdna and Rgrad in the simulated data previously discussed where we know what to expect . For 1 GeV/amu Fe simulations , we logically find that Rdna values for pRIF are greater than one ( see Table 2 ) , reflecting the fact that generation of damage is proportional to the amount of DNA: i . e . , the more DNA , the more likely radiation will produce a break . On the other hand , Rgrad values are also larger than 1 . This result at first hand might look surprising as one would assume that there should be no preferential location of DSB with respect to nuclear interfaces . However , this result simply reflects the fact that gradient values are larger in denser regions of DNA . We can also compare the relative measurements between observed foci and reshuffled foci to verify that our image-based prediction of DNA damage leads to the same values . The ratio of Rdna ( R ) and Rgrad ( Rg ) between simulated pRIF ( indexed 1 ) and reshuffled pRIF ( indexed 2 ) are shown in Table 2 ( i . e . , R1/R2 and Rg1/Rg2 ) . These ratios are very close to 1 , indicating a foci pattern for pRIF that matches the expected random distribution of DNA damage . DNA damage induced by low-LET radiations are scattered throughout the nucleus . Therefore , in order to predict a DSB imaging pattern for low LET , we need to generalize the reshuffling approach in 3-D . This can be done in the same manner it was done for high-LET tracks using DNA density at any pixel in the nucleus as a probability to have damage at that location . Similarly to HZE , the validity of this approach was tested by comparing the Rdna and Rgrad values for pRIF and reshuffled pRIF in the low-LET simulations from Table 1 . Reshuffling pRIF over the full nucleus for low-LET data also led to similar Rdna and Rgrad values than for pRIF ( i . e . , R1/R2 of 1 . 05 ± 0 . 09 and Rg1/Rg2 of 0 . 96 ± 0 . 11 ) . Therefore , reshuffling foci in 3-D remains a valid approach to predict DSB patterns for low LET . The kinetics of normalized Rdna and Rgrad values for cells exposed to 1 GeV/amu Fe are shown in Figure 7A–7C . Rdna and Rgrad ratios along tracks confirm our previous finding that γH2AX and 53BP1 RIF spatial distributions deviate from the predicted nuclear locations of DNA damage . Rdna averages were less than the predicted Rdna ( ratio less than 1 ) at all time points , as early as 4 . 5 min post-IR . Similarly , the Rgrad averages were always greater than predicted . Thus , on average , RIF were located in lower DNA density regions than where DSBs were expected to occur , and RIF tended to be located at the interface between high and low DNA density regions ( see Figure 7D–7F for illustration of phenomenon ) . Interestingly , ATMp foci showed a slightly different dynamic . Although ATMp RIF localized to chromatin regions of slightly less DNA at the interface of high and low densities at the earliest time point , by 10 min post-IR the measured Rdna and Rgrad were 1 , i . e . , similar to that predicted for DSB . Analysis of low-LET data required specific imaging considerations as discussed previously for the 3-D generalization of our approach . In addition , due to the poor resolution of real conventional microscope images in the Z direction , Rdna and Rgrad computations were done only on the best focal plane of 3-D image stacks for each cell ( see Figure 8 ) . Our results showed that Rdna values were all slightly lower than the predicted values for all three RIF as summarized in Table 3 . Even though deviations from prediction were small , the robustness of these measurements was evident in the very small standard errors ( i . e . , 0 . 3%–2% ) , allowing the detection of very subtle differences . Such small errors led to statistical significance for γH2AX Rdna and Rgrad values early after exposure to radiation . All time points were also significant for the Rgrad values of 53BP1 . We monitored the relative amount of co-localization of γH2AX or ATMp with 53BP1 ( Figure 9 ) . As previously shown for other cell types and markers [6 , 10 , 28–30] , both markers show fast co-localization with 53BP1 RIF , independently of the radiation quality . Co-localization significantly increased from 44% to 64% for cells within the first 10 min following 1 Gy of 1 GeV/amu Fe . Representative images are shown in Figure 9B . It is important to note here that RIF frequencies following either high LET or low LET did not change appreciably during this time period ( i . e . , 2%–7% , all labels included ) . Therefore , it seems unlikely that a 20% increase of co-localization could be simply explained by more foci appearing in common regions of the nucleus . The above analysis suggests that either RIF occur at restricted locations or chromatin remodels as a result of DNA damage . Indeed , both 53BP1 and γH2AX are chromatin modifications . To test if the pattern of non-random distribution of RIF locations were due to global chromatin reorganization , we monitored chromatin in HeLa cells transfected with histone H1 . 2 fused to GFP . The chromatin pattern was monitored before and after 5 Gy of X-rays in the same cells . Representative time frames are shown in Figure 10 . Chromatin patterns were unaffected by irradiation .
We previously developed computation models that simulate the production of DSB in hypothetical spatial geometries [22 , 23] . In this work , we extended such models to predict the pattern of radiation-induced DNA damage detected by protein markers in biological images after taking into account the optical properties of the microscopy . Two distinct radiation qualities were considered , low LET ( i . e . , photons ) and high LET ( i . e . , 1 GeV/amu Fe ions ) using this model . DSBs or clusters of DSBs appeared as foci in simulated images or pRIF ) . We found that the frequency pRIF matched the frequency of RIF from three different markers measured shortly ( i . e . , 4 . 5 min ) after exposure to 1 GeV/amu Fe . Using these simulations , we noted that Fe tracks had many close-by DSBs that could not be resolved by optical microscopy , which led to larger and fewer foci along these pseudo tracks . This phenomenon was reported previously on experimental data for high-LET tracks [7 , 11] . The fact that frequencies were the same for pRIF and RIF in this case also suggests that when DNA damage extends over large regions in the nucleus it leads to a systematic and rapid formation of focal protein marks . Therefore , RIF are a good indication of DNA damage induced by HZE . In contrast , the frequencies measured for low-LET RIF from 4 . 5 to 60 min post-IR were reduced by more than 60% from what we predicted . RIF in this case is probably not a good DNA damage marker , suggesting that non-complex DSBs lead to a lower and/or slower RIF response . We then introduced an image-based model that could predict DNA damages for low LET or high LET in real nuclear images . This was done by randomly reshuffling the locations of detected foci , using DAPI intensity of each pixel as the probability to have damage at that pixel location . After validating this approach on the simulated nuclei previously analyzed for pRIF frequencies , we applied it to real data . We observed that the majority of RIF along tracks were spaced by gaps larger than 1 μm , whereas our image-based approach predicted the majority of damages to be less than 1 μm apart . This was observed as early as 4 . 5 min post-IR and the deviation from prediction was even greater at 30 min following exposure . To determine whether this result reflected preferential locations of foci in the nucleus , we introduced a new set of imaging parameters that quantify the location of RIF relative to the nuclear DAPI pattern . Using this tool , we showed for both high-LET and low-LET radiation that RIF were more frequently located in low DNA density regions than predicted by image-based modeling . We also showed that RIF occurred predominantly at the interface between high and low DNA density regions . Interestingly , ATMp did not show as strong a trend as the other markers , supporting its early , but mobile , role in the DNA damage response [31] . Finally , we measured a rapid increase in the co-localization between different DNA damage markers over the first 10 min following exposure to both radiation qualities . We conclude from these studies that nuclear organization plays an important role in the response to DNA damages . Specifically , foci locating preferentially in low DNA density regions suggest that damages occurring in condensed regions of the DNA are not always detected , leading to their lower proportion . Detection in condensed areas might then depend on either local decondensation of the chromatin as suggested by others [32 , 33] or movement of the damaged site to more open regions of the chromatin . We tend to be in favor of the DSB movement hypothesis for the following reasons . First , the rapid accumulation of RIF at the interface between high and low DNA density contradict what we know about the way radiation deposits its energy in tissue; second , imaging of HeLa cells transfected with histone H1 . 2 GFP exposed to a high radiation dose ( 5 Gy ) of X-rays did not show a change in chromatin pattern compared with unexposed cells . Furthermore , it has been shown that some genes become transcriptionally active only upon relocating into open regions of the nucleus [34] . In fact , whole parts of a chromosome have been reported to be able to move over a 1–5 μm path within a few minutes during transcription activation in mammalian cells [35] . The only way to resolve these hypotheses unequivocally is to use live cell imaging of GFP-fused proteins recruited at the site of DNA damage such as NBS1 or 53BP1 exposed to physiological doses of ionizing radiation . Another important aspect of DNA damage response to radiation provided in this work is the fact that damage appears to be spatially organized . Specifically , the lack of foci in close proximity along tracks suggests the existence of discrete self-excluding nuclear regions where DNA damages are clustered . This result supports the existence of “repairosomes” in mammalian cells , which has been suggested by Savage [36 , 37] . These nuclear domains would provide the necessary clamping and orientation for repair to take place but could also lead to translocations or chromosome aberrations when multiple breaks would be simultaneously processed . The existence of “repair centers” has already been shown in yeast Saccharomyces cerevisiae where Lisby et al . engineered a system of fluorescently marked DSB [38] . They showed in such a system that in 40% to 50% of the cases , two individual DSB relocalized into one common focus which was rich in Rad52 proteins . On the other hand , there is no direct evidence of repair centers in mammalian cells , although some reports have suggested their existence . For example , DiBiase et al . hypothesized that Ku proteins might recruit DSB to DNA-PKcs ( catalytic subunit of DNA-PK ) since it is fixed on the nuclear matrix , allowing fast DNA repair via nonhomologous end-joining [39] . More recently , it was also shown that MRN complex “tethers” damaged DNA to help activate ATM by increasing locally the concentration of DSBs [40] . Therefore , our work adds to this hypothesis by identifying for the first time in mammalian cells morphological features in the nucleus where protein markers of damage response preferentially locate . More work is needed to identify features that define these subnuclear regions .
Human mammary epithelial cells ( HMEC-184; 184v; passage 7–10 ) were cultured in serum-free medium as previously described [41] . HMEC-184 were irradiated with 1 Gy of ionizing radiation 2 d post-plating . Low-LET radiation exposures were conducted using a 5600 curie source of 137-Cs γ-radiation . The high-LET radiation was 1 GeV/amu Fe ions from the NASA Space Radiation Laboratory of Brookhaven National Laboratory . HeLa cells ( ATCC ) were cultured in DMEM with 10% FBS and exposed to a 160-kV X-ray source . In all three types of radiation , the same dose rate was used , i . e . , 1 Gy/min . Primary: mouse monoclonal anti phospho-histone H2AX ( Ser139 ) antibody ( lot 27505 , Upstate Cell Signaling Solutions , http://www . upstate . com/ ) used at 1 . 42 μg/ml; mouse monoclonal anti-phosphorylated ( pS1981 ) ATM protein kinase antibody ( lot 14354; Rockland , http://www . rockland-inc . com/ ) used at 2 . 15 μg/ml; rabbit polyclonal anti 53BP1 ( lot A300-272A , Bethyl Lab , http://www . bethyl . com/ ) used at 5 μg/ml . Secondary antibodies were used at 1:300 ( Dk anti-Rb Alexa 594 , lot 40247A , and Gt anti-Ms Alexa 488 , lot A11029 , from Molecular Probes , Invitrogen , http://www . invitrogen . com ) . The H1 . 2 GFP construct was a generous gift from Dr . Michael Hendzel , University of Alberta , Canada [42] . Cells were grown on tissue culture–treated LabTek eight-well chamber slides . Chambers were fixed at room temperature for 15 min using 2% paraformaldehyde followed by successive wash and permeabilization with 100% methanol for 20 min at 20 °C . Nonspecific sites were blocked using 1% BSA for 90 min . The cells were incubated 2 h at room temperature with primary antibodies in blocking buffer in a humidified chamber . Following washes , primary antibody binding was detected using species-appropriate fluorochrome-labeled secondary antibodies incubated for 1 h at room temperature . Nuclei were counterstained with DAPI ( 4′ , 6-Diamidino-2-Phenylindole ) using 0 . 5 μg/ml . Slides were mounted in Vectashield ( Vector Laboratories , http://www . vectorlabs . com/ ) and stored at −20 °C until evaluated . Cells were viewed and imaged using a Zeiss Axiovert epifluorescence microscope ( Carl Zeiss , http://www . zeiss . com/ ) equipped with a multiband pass filter and a differential wavelength filter wheel . Images were acquired using a Zeiss plan-apochromat 40× dry , with an NA of 0 . 95 and a scientific-grade 12-bit charged coupled device camera ( ORCA AG Hamamatsu , 6 . 45 × 6 . 45 μm2 pixels ) . The image pixel size was measured to be 0 . 16 μm , but based on the NA of the objective , the actual resolution of the image in the FITC channel was ∼0 . 5 × 0 . 488/NA = 0 . 26 μm . All images were captured with the same exposure time so that intensities were within the 12-bit linear range . All image manipulation and analysis were done with Matlab ( MathWorks , http://www . mathworks . com/ ) and DIPimage ( image processing toolbox for Matlab , Delft University of Technology , The Netherlands ) . For track analysis , Fe ion tracks were manually identified on the most-in-focus slice in a conventional image stack . Tracks were defined only if there were four or more foci within the nucleus , and if the cells on the stack slice had parallel tracks , reinforcing the assertion that the line in question was the result of a particle traversal and not a simple random alignment of points ( see Figure 2 ) . To keep track of the radial ( perpendicular to the track ) displacement of foci , a stripe of 0 . 8 μm width along each track was sampled for the maximum intensity in the direction perpendicular to the track . This led to a 1-D intensity curve with maximal intensities along the track . Herein this curve was called the “maximum intensity profile , ” or simply the “1-D profile . ” Using the intensity-inverted 1-D profile , foci were detected by searching for local minima along the track using watershed algorithms as shown in Figure 3 . For cells exposed to γ-rays , 3-D images had to be acquired since foci were not restricted along a line anymore . This made maxima detection more difficult . To address this issue , we used a method similar to previous work [43] where a tophat morphological filter was used to enhance the intensity of spherical spots in the image . The resulting image was then intensity-inverted , and watershed algorithms were applied to detect minima as was done for the 1-D profile . In both 1-D and 3-D cases , the center of the RIF was determined as a pixel with the maximum intensity as sampled over the identified RIF . These central pixels were taken as RIF coordinates and utilized for Rdna and Rgrad computations ( defined in the main text ) , and for co-localization analysis . More details on the methods of image processing and analysis are in Results . The Monte Carlo algorithm utilized here is based on the probability of a DNA DSB at a given location in the nucleus being proportional to the DNA density and the dose ( energy per unit mass ) deposited at that location . As shown previously [20 , 24] , the spatial DSB distribution is generated via a stochastic process given by where ψ is a probability to create a DSB at a monomer ( a small stretch of DNA containing 2 kbp of genomic information ) , D ( t ) is the local dose given by the track structure ( it can vary sharply with the distance t from the track center ) , and Q is the constant determined from model fits to PFGE data . The correctly determined Q would generate proper DSB yields , fragment-size distribution functions , average numbers of DSB per nucleus per track , and the spatial distributions of DSB for several ions , any E , and any dose [20 , 22] . In this approach , the frequency of DSB depends on the properties of a track given by D ( t ) [18 , 44] , but it will also depend on the DNA configuration given by a random walk model , as the probability ψ is applied to each monomer . In this model , a pixel can have a variable number of monomers corresponding to the density fluctuations of genetic material in the nucleus with high precision .
|
DNA damages are daily cellular events . If such events are left unchecked in an organism , they can lead to DNA mutations and possibly cancer over a long period of time . Consequently , cells have very efficient DNA repair machinery . Many studies have focused on the different molecular factors involved in the repair machinery , neglecting to consider the spatial context where damage occurs . Therefore , little is known about the role the nuclear architecture might have in the DNA damage response . In this study , we introduce computer modeling and image analysis tools in order to relate the position of DNA damage markers to morphologically distinct regions of the nucleus . Using these tools , we show that radiation-induced damages locate preferentially in non-condensed DNA regions or at the boundary of regions with condensed DNA . These results contradict the current dogma that the molecular response to randomly generated DNA damages is independent of their nuclear locations . Instead , this suggests the existence of repair centers in the nucleus . Overall , our approach shows that nuclear architecture plays a role in the DNA damage response , reminding us that the nucleus is not simply a soup of DNA and proteins .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biophysics",
"homo",
"(human)"
] |
2007
|
Image-Based Modeling Reveals Dynamic Redistribution of DNA Damage into Nuclear Sub-Domains
|
Animals acquire predictive values of sensory stimuli through reinforcement . In the brain of Drosophila melanogaster , activation of two types of dopamine neurons in the PAM and PPL1 clusters has been shown to induce aversive odor memory . Here , we identified the third cell type and characterized aversive memories induced by these dopamine neurons . These three dopamine pathways all project to the mushroom body but terminate in the spatially segregated subdomains . To understand the functional difference of these dopamine pathways in electric shock reinforcement , we blocked each one of them during memory acquisition . We found that all three pathways partially contribute to electric shock memory . Notably , the memories mediated by these neurons differed in temporal stability . Furthermore , combinatorial activation of two of these pathways revealed significant interaction of individual memory components rather than their simple summation . These results cast light on a cellular mechanism by which a noxious event induces different dopamine signals to a single brain structure to synthesize an aversive memory .
Mechanisms underlying memory can be as simple as a modulation of monosynaptic connection in the gill withdraw reflex of Aplysia [1] . Alternatively memory formation and storage may require dynamic interaction of distinct neuropiles in a brain [2] . In Drosophila melanogaster , a neuronal circuit centered on the mushroom body ( MB ) is important for the formation and storage of odor memory [3] , [4] , [5] , [6] . Signals of odor and shock are integrated in the MB for memory formation [7] , [8] , [9] . Identity of the odor is represented by a small subset of Kenyon cells [10] , [11] , [12] , which are major intrinsic neurons of the MB and categorized into the α/β , α'/β' and γ neurons ( Figure 1A ) [13] . According to their projection patterns in the lobes , these neurons can be further classified into 8 subtypes: α/βp , α/βs and α/βc ( also known as the pioneer α/β , the early α/β , or the late α/β respectively [14] ) for the posterior , surface or core layers of the α/β lobes; α'/β'a , α'/β'm and α'/β'p for the anterior , middle and posterior layers of α'/β' lobes; γd and γmain for the dorsal and main layer of the γ lobe [15] , [16] . The Kenyon cells receive a dopamine signal that mediates aversive reinforcement for odor memory formation [17] , [18] . Expression of DopR , a D1-like dopamine receptor also known as dDA1 , in the γ neurons is fully sufficient to rescue the mutant defect in aversive odor memory [18] . Activation of many types of dopamine neurons using TH-GAL4 can substitute for the aversive stimulus that induces odor memory [19] , [20] . In flies , dopamine neurons from protocerebral anterior median ( PAM ) , protocerebral posterior lateral 1 and 2ab ( PPL1 and PPL2ab ) clusters terminate in the entire MB ( Figure 1B–1C ) [21] , [22] , [23] . Individual neurons in these clusters terminate distinct subdomains along the longitudinal axis of the MB lobes . The application of noxious stimuli , such as electric shock , activates only a subset of dopamine neurons [21] , [24] , indicating that the response property greatly varies among individual cells within a cluster . Consistent with this observation , activation of specific subsets of these clusters , such as MB-M3 and MB-MP1 neurons , can induce aversive odor memory [25] . This dopamine input presumably modulates the pre-synaptic output of odor-representing Kenyon cells and drives memory formation [8] , [9] . Besides aversive reinforcement , dopamine is responsible for a wide range of physiological functions including appetitive memory , some of which are mapped in the MB [17] , [21] , [24] , [26] , [27] , [28] , [29] . This implies the functional differentiation of the MB subdivisions by different sets of dopamine neurons . While both MB-M3 and MB-MP1 neurons can induce aversive odor memory , these neurons terminate in spatially segregated subdomains of the MB . MB-M3 neurons primarily synapses in the medial tip of the β lobe , whereas MB-MP1 neurons terminates in the spur of the γ lobe ( γ1 ) and the peduncle of the α/β neurons [15] . Blocking the output of MB-M3 neurons during memory acquisition preferentially affects labile 2-hour memory [25] , indicating the partial contribution of MB-M3 neurons to aversive reinforcement of electric shock . This raises the question of how the other reinforcement pathways to the MB are coordinated to synthesize a complete aversive odor memory . Considering the selective phenotype of MB-M3 neurons and differential functions of the lobe systems [25] , [30] , [31] , [32] , [33] , it is likely that electric shock induces qualitatively different memory traces in parallel . In this study , we challenge this hypothesis by activating and inactivating individual dopamine pathways separately . By conducting further behavioral screening of GAL4 driver lines , we identify a third type of dopamine neurons that are capable to induce aversive odor memory . Using these tools to manipulate different dopamine pathways to the MB , we characterize each memory component in isolation and show how their interaction shapes the temporal stability of aversive memory .
To systematically identify reinforcement pathways to the MB , we elaborated our previous dTRPA1-based behavioral screening by testing 10 additional GAL4 lines for different subsets of dopamine neurons . Dopamine synthesis in MB-projecting neurons of these lines was confirmed with immunoreactive signals to Tyrosine Hydroxylase ( TH ) , the rate-limiting enzyme for dopamine synthesis . The pattern of TH signals coincides well with that of dopamine itself in the brain ( Figure S1 ) . This collection of GAL4 drivers covers most , if not all , dopamine neurons in the fly brain including those projecting to the MB ( Figure S2 ) [25] . For activation of target neurons , we expressed the thermo-sensitive cation channel dTRPA1 [34] . We elevated temperature to 30°C for 60 sec only during the presentation of an odor , as in the standard conditioning protocol ( Figure 1D ) . This temperature shift by itself had a slight aversive effect as seen in the tendency of conditioned odor avoidance of control groups [25] . Contingent activation with the odor presentation should induce significant conditioned avoidance of the paired odor , if a given GAL4 line drives dTrpA1 within neurons that signal aversive reinforcement . We found functional heterogeneity of dopamine neurons in inducing immediate aversive memory ( Figure 1E ) . Activation of particular cell types in the PAM and PPL1 clusters induced aversive memory , whereas drivers labeling other types in the same clusters did not ( Figure 1E; see below for the detailed description ) . However , most GAL4 drivers used in these experiments have expression outside the target dopamine neurons . To scrutinize whether the induced memory is due to the target cells , we suppressed dTRPA1 expression selectively in dopamine neurons using TH-GAL80 [35] and examined whether TH-GAL80 suppresses memory formation induced by thermo-activation . We found that aversive memory is indeed attributable to dopamine neurons in most cases . TH-GAL80 did not fully silence dTRPA1-dependent memory of MB-GAL80;c259 and NP2758 , which could be due to either an incomplete suppression of GAL4 in dopamine neurons or a potential contribution of non-dopaminergic cells in these drivers ( see below for MB-GAL80;c259 and [25] for NP2758 ) . To further narrow down the expression pattern to a single dopamine cell type , we combined Cha3 . 3kb-GAL80 with some drivers . The Cha3 . 3kb fragment drives strong expression not only in choline acetyltransferase-positive neurons but also non-cholinergic cells presumably due to ectopic expression [36] . Dopamine biosynthesis requires the enzymes TH and Dopa Decarboxylase ( DDC ) and these enzymes are expected to be in all dopamine neurons . TH-GAL4 and two versions of DDC-GAL4 ( HL8 and HL9 ) label the majority , but not all , dopamine neurons . TH-GAL4 labels all cells in the PPL1 and PPL2ab clusters and a small subset of the PAM cluster including MB-M3 neurons ( Figure 2A , 2E , 2I , 2M , 2Q ) . Activation of these neurons induced robust aversive memories even in the dTrpA1 mutant background ( Figure 2U , Figure S4 ) , suggesting that activation of dopamine neurons alone can substitute an aversive unconditioned stimulus . Cha3 . 3kb-GAL80 suppressed transgene expression in some of the PAM cluster cells and three cells in the PPL1 cluster of TH-GAL4 ( Figure 2B , 2F , 2J , 2N , 2R ) . The silenced cells include MB-M3 neurons and at least one MB-MP1 neuron [25] . TH-GAL4 Cha3 . 3kb-GAL80/UAS-dTrpA1 flies however showed a similar level of aversive memory as TH-GAL4/UAS-dTrpA1 flies ( Figure 2U ) . This result implies functional redundancy of Cha-GAL80-positive MB-M3 neurons and MB-MP1 neuron in immediate memory . HL8 and HL9 label the majority of PAM cluster cells , one or no PPL1 cells and several PPL2ab cells ( Figure 2C–2D , 2G–2H , 2K–2L , 2O–2P , 2S–2T ) . Although these drivers also label serotonergic neurons , some of which project to the MB [37] , [38] , terminals in the calyx are TH immunoreactive ( Figure S3 ) , suggesting that they belong to the PPL2ab cluster . With HL8 and HL9 , dTRPA1 activation did not induce significant aversive memory ( Figure 2V–2W ) , consistent with a previous report using light-dependent activation [19] . Because these drivers label many cell types , the roles of individual dopamine neurons might be obscured in final memory scores by antagonizing functions each other . As to the PPL1 cluster neurons , two drivers labeling MB-MP1 neuron ( i . e . MB-GAL80;NP0047 and MB-GAL80;c259 ) induced very robust aversive memories ( Figure 3 ) , which is fully consistent with the previous result using other drivers c061;MB-GAL80 and NP2758 [25] . TH-GAL80 silenced GAL4 activity selectively in the dopamine neurons in these drivers ( Figure 3B , 3F , 3J , 3D , 3H , 3L ) . Memory induced with these drivers was also significantly suppressed by TH-GAL80 , although it was not complete with MB-GAL80;c259 ( Figure 3R ) . Therefore , we used c061;MB-GAL80 , but not MB-GAL80;c259 , for the following experiment to manipulate MB-MP1 neuron . The GAL4 driver line 5htr1b-GAL4 labels one MB-V1 neuron and one MB-MV1 neuron in the PPL1 cluster ( Figure 4A , 4E , 4I , 4L ) . 5htr1b-GAL4 induced a slight but significant aversive memory ( Figure 4P ) . TH-GAL80 suppressed most of the GAL4-positive cells of 5htr1b-GAL4 ( Figure 4B , 4F , 4M ) and accordingly dTRPA1-induced memory ( Figure 4P ) . We combined Cha3 . 3kb-GAL80 with 5htr1b-GAL4 that preferentially silenced MB-MV1 neuron but did not abolish expression in MB-V1 neuron and other GAL4-positive cells ( Figure 4C , 4G , 4J , 4N ) . 5htr1b-GAL4 Cha3 . 3kb-GAL80 did not induce memory ( Figure 4Q ) , suggesting the critical role of MB-MV1 . Consistently , in NP7187 and MZ840 , which label one MB-V1 neuron , activation did not induce significant memory ( Figure 4D , 4H , 4K , 4O , 4R ) [25] . Thus , memory induced with 5htr1b-GAL4 is likely to be an effect of MB-MV1 neuron activation . In accordance with this interpretation , calcium imaging revealed that MB-MV1 neuron robustly responds to electric shock , whereas MB-V1 neuron preferentially responds to odors [21] . Considering the importance of the vertical lobes in olfactory learning [8] , [9] , [39] , [40] , [41] , [42] , MB-MV1 neuron might act together with MB-V1 neuron during memory acquisition . Thermo-activation with NP7323 , which labels a mixture of cell types in the PAM cluster ( MB-M2 ) , induced a slight but significant aversive memory ( Figure 5A , 5C , 5E , 5G , 5I ) . With MZ19;Cha3 . 3kb-GAL80 , which labels approximately ten PAM cluster cells , activation did not have significant effect ( Figure 5B , 5D , 5F , 5H , 5J ) . These results imply functional heterogeneity in memory formation of PAM cluster neurons . However , we could not test whether the induced memory in NP7323 can be attributed to non-dopaminergic GAL4-positive neurons , because TH-GAL80 does not suppress GAL4 expression in the majority of PAM cluster cells . These results , together with our previous report [25] , point to three distinct dopamine pathways to the MB that induce aversive memory; MB-M3 neurons in the PAM cluster and MB-MP1 neuron and MB-MV1/MB-V1 neurons in the PPL1 cluster . The dopamine neurons of the three distinct functional groups terminate in segregated MB subdivisions along the trajectory of Kenyon cell axon bundles: MB-M3 neurons to βsp2 and β'a2 , MB-MP1 neuron to γ1 and the core of the peduncle where the α/β neurons innervate , MB-MV1 neuron to γ2 and α'1 , MB-V1 neuron to α2 , α'2 and a part of α'1 ( Figure 1E; see [15] for the description of MB subdivisions ) . Because blocking all of these neurons with TH-GAL4 did not abolish shock-induced aversive memory , we reserve the possibility that additional types of neurons may contribute to aversive memory formation ( e . g . MB-M2 neurons in NP7323 and serotonin neurons [35] ) . To characterize the individual dopamine pathways to the MB , we determined cellular identity by counting TH-immunoreactive cells in single and combined drivers ( Figure 6A–6B ) . When two drivers label identical dopamine neurons , the number of cells in combined drivers should not differ from those in respective single drivers . Identical MB-V1 neuron is labeled in 5htr1b-GAL4 , MZ840 , NP7187 , NP0047 . 5htr1b-GAL4 and NP0047 label the same MB-MV1 neuron . c061;MB-GAL80 , NP0047 and NP2758 label the same MB-MP1 neuron , while c061;MB-GAL80 and MB-GAL80;c259 may label another MB-MP1 neuron in addition to that in NP2758 and NP0047 [25] . Expression of presynaptic markers in these dopamine neurons revealed that their arbors in the MB are presynaptic terminals ( Figure 7A–7C ) [25] . The processes of these dopamine neurons in the protocerebrum contained many fewer output sites , implying a dendritic nature ( Figure 7A–7C ) . To compare the distribution of the dendrites , we performed a non-rigid intensity-based transformation of brains [43] and 3D image analysis of the dopamine neurons . We used TH immunolabelling as landmarks to transform brains . Computational alignment of the MB-M3 neurons , MB-MV1/V1 neurons , and MB-MP1 neuron revealed that each cell type has a unique pattern of dendrite distribution in the superior and inferior protocerebral regions , and they are partially overlapping with each other ( Figure 7D , Video S1 ) . Furthermore , we aligned the neurons that did not induce significant aversive memory ( Figure 4 and Figure 5 ) and found that their processes outside the MB have different distribution and focus ( Figure 7E , Video S2 ) . When the memory-inducing and non-inducing neurons are separately pooled and superimposed in one brain , the dendrites of these groups are largely segregated , especially in the anterior inferior medial protocerebrum ( Figure 7F , Video S3 ) . These results imply that the dopamine neurons for aversive memory may partially share some neuronal input that is distinct from dopamine neurons that do not induce aversive memory . Cellular identification of presynaptic neurons requires more precise anatomical and functional analyses . Based on cell counting and the behavioral screen , we selected drivers NP5272 , 5htr1b-GAL4 , c061;MB-GAL80 to further characterize three MB-M3 neurons , one MB-MV1 neuron/one MB-V1 neuron , and one or two MB-MP1 neurons , respectively ( Figure 6C ) . We addressed whether these dopamine neurons function for mediating electric shock reinforcement . To test their necessity during shock conditioning , we transiently blocked corresponding neurons by expressing Shits1 [44] . Previous studies reported that blocking many dopamine neurons with TH-GAL4 ( including MB-M3 , MB-MP1 and MB-MV1/MB-V1 neurons ) severely impaired memory irrespective of retention times [19] , [25] , [45] , [46] . Notably , the transient Shits1 block of each dopamine pathway impaired specific temporal components of aversive memory of shock ( Figure 8 ) . The block with NP5272 preferentially impaired 2-hour memory , but not significantly affected 2-min or 9-hour memory ( Figure 8A ) [25] . We found that the block with 5htr1b-GAL4 with multiple copies of UAS-shits1 caused gradual memory impairment over time , leaving the immediate memory intact , although this fly had a significant phenotype at a permissive temperature ( Figure S5 ) . We reproduced the impairment of 9-hour memory with a single copy of UAS-shits1 ( Figure 8B ) . This impairment of 5htr1b-GAL4/UAS-shits1 in consolidated memory was due to the transient block during training , since the same inhibition during consolidation or retrieval did not significantly affect the performance ( Figure 9A–9D ) . In contrast to these defects in selective memory phases , the block with c061;MB-GAL80 significantly impaired all tested memory phases ( i . e . 2 min , 2 hours and 9 hours after training; Figure 8C ) , although a previous study found no significant impairment of 3-hour memory [26] . This is presumably due to the higher restrictive temperature in this study ( 33°C compared to 31°C ) , since the effect of Shits1 is sensitive to a small temperature difference [47] . The observed memory impairment with c061;MB-GAL80 should not be attributed to a defect in detecting odor or electric shock itself , as their reflexive avoidance was normal ( Figure S6D ) . Furthermore , blocking after training or experiments at permissive temperature did not cause this phenotype ( Figure 9E–9H ) . We also confirmed that blocking neurons outside our target dopamine neurons by combining GAL80 lines did not impair memory ( Figure S6A–S6C and S6E–S6F ) . Taken together , we propose that electric shock punishment recruits multiple types of dopamine neurons MB-MP1 , MB-M3 and MB-MV1/MB-V1 to form parallel memory traces with distinct temporal stability . Additional dopamine neurons might be recruited for signaling aversive reinforcement , since blocking all these dopamine pathways with TH-GAL4 leaves residual memory in contrast to complete abolishment of memory in dDA1 receptor mutants and neuron-specific TH mutants [17] , [18] , [25] , [45] , [46] . Contribution of each dopamine neuron to the synthesis of total memory depends on both the magnitude of initial memory and its stability over time . We found that the magnitude of dTRPA1-induced immediate memory depended on the activation temperature and cell type ( Figure 10 ) . This implies that the amount of dopamine input represents the strength of reinforcement . To better understand the role of the distinct reinforcement pathways , we characterized the retention of each memory component in isolation by activating the individual dopamine neurons . Memories induced by these drivers showed remarkable differences in decay dynamics ( Figure 11 ) . Initially robust memory induced with c061;MB-GAL80 decayed rapidly over 9 hours . In contrast , memory induced with 5htr1b-GAL4 was highly stable and still significantly present after 9 hours , although initial memory was moderate . This differential memory decay is not due to the magnitude of initial memory , because the equivalent initial memory with NP5272 or with milder activation using c061;MB-GAL80 disappeared completely within 3 hours . In contrast to the similar decay dynamics of memories induced with NP5272 and c061;MB-GAL80 ( Figure 11 ) , the requirement of these neurons for the retention of shock-induced memories is qualitatively different ( Figure 8 ) . The genetic background or the effect of dTRPA1 at permissive temperature is an unlikely cause of the faster/slower memory decay , because the memory retention of the dTRPA1-expressing flies was indistinguishable when trained with electric shock ( Figure S7 ) . Thus , these results suggest that each dopamine pathway to the MB establishes a memory component with unique temporal stability . The results of activation and inactivation experiments appear to be inconsistent: A simple sum of memory performances by activation of these pathways does not explain selective memory impairments when they are blocked ( Figure 8 and Figure 11 ) . For instance , the effect of the blocking MB-M3 neurons with NP5272 was most pronounced for 2-hour memory ( Figure 8A ) , whereas memory induced by activation of the same cells with NP5272 did not last 2 hours ( Figure 11 ) . The reason of this apparent discrepancy could be technical , such as the copy number of the effectors and temperature regimes . Alternatively it may suggest a synergistic interaction between memories induced by MB-M3 neurons and other dopamine neurons . We noticed another case of possible interaction that suggests functional redundancy; activation with NP5272 or 5htr1b-GAL4 induced immediate memory ( Figure 11 ) , whereas the block with these drivers did not show significant short-term defect ( Figure 8A–8B ) . To test the various forms of interaction between memory components , we first measured activation of MB-M3 neurons together with other reinforcement pathways . To activate MB-MP1 , MB-MV1 and MB-V1 neurons together , we took another driver MB-GAL80;NP0047 from the activation screening ( Figure 3Q ) , and measured the retention of dTRPA1-induced memory . In comparison to MB-GAL80;NP0047 alone , additional activation with NP5272 did not significantly improve the performance of immediate memory ( Figure 12A ) , supporting the redundancy of MB-M3 neurons and the others for immediate memory . This is in line with no significant requirement of MB-M3 neurons for immediate memory induced by electric shock ( Figure 8A ) . In contrast , combinatorial activation significantly improved performance in 2-hour retention ( Figure 12A ) , suggesting the synergistic contribution of MB-M3 neurons to 2-hour memory and recapitulating the selective impairment of 2-hour shock memory upon blocking with NP5272 ( Figure 8A ) . Also combinatorial activation of MB-M3 and MB-V1/MB-MV1 neurons with NP5272 and 5htr1b-GAL4 caused a similar pattern of interaction in immediate and 2-hour memories ( Figure 12B ) . In contrast , activation of MB-M3 and MB-V1 neurons without MB-MV1 neuron in NP5272 and MZ840 did not induce any 2-hour memory ( Figure 12C ) . Also NP5272 activation was not significantly additive across all tested retention times compared to c061;MB-GAL80 single activation ( Figure 12D ) . Thus , we propose that MB-MV1 , but not MB-MP1 or MB-V1 neurons , likely interacts with MB-M3 neurons for 2-hour memory . The enhancement of 2-hour memory was likely due to an increase of anesthesia-sensitive memory ( ASM ) , because the effect disappeared by 9 hour when the memory is primarily consisted of anesthesia resistant memory ( ARM ) . Indeed , MB-M3 neurons activation did not contribute to the increase of ARM ( Figure 13A , 13C ) . Memories induced by thermo-activation with 5htr1b-GAL4 and c061;MB-GAL80 were at least partially ARM ( Figure 13 ) . Combinatorial thermo-activation with 5htr1b-GAL4 and c061;MB-GAL80 was not significantly different from single activation by c061;MB-GAL80 ( Figure 12E ) . Milder activation at 28°C however revealed interaction ( Figure 12F ) that features gradual memory impairments upon blocking with 5htr1b-GAL4 ( Figure 8B ) . Ectopic expression of dTRPA1 or its effect at permissive temperature alone is an unlikely cause of the interaction , since 2-hour memory of all these genotypes was normal when they were trained with electric shock ( Figure 13E , 13F , 13G , 13H ) . Altogether , these results suggest that MB-MV1 and MB-M3 neurons are redundant for immediate memory when MB-MP1 neuron is activated , and different modulatory interactions of MB-MV1/MB-M3 neurons and MB-MV1/MB-MP1 neurons tune the stability of memory .
Using freely behaving animals , we demonstrated that at least three essential dopamine pathways to the MB can together synthesize aversive memory that shares similar temporal characteristics with shock-induced memory . They arborize in the different subdomains in the MB . Functional imaging of dopamine neurons revealed that response to electric shock significantly differ between cell types [21] . These results indicate shock reinforcement recruits a specific set of the dopamine pathways to the MBs and induces multiple memory traces in Kenyon cells through different receptors [18] , [30] , [32] . The Drosophila MB is subdivided into domains , that are defined by specific combinations of intrinsic and extrinsic neurons ( Figure 1A ) [15] . Each dopamine pathway for memory induction intersects the specific axonal compartment of Kenyon cells ( Figure 1E , Figure 7A–7B ) [15] , [21] , and we found that the memory components induced by the distinct dopamine neurons interact to tune the stability of collective memory ( Figure 12 ) . Therefore , multiple memory traces formed in spatially segregated synapses in the MB would interact with each other . One of the possible underlying mechanisms is intracellular interaction of the dopamine inputs to a single Kenyon cell . In the sensory neurons of Aplysia , simultaneous application of serotonin to different parts of a neuron - the cell body and presynaptic terminals of a sensory neuron - synergistically induces intermediate and long-term facilitation of the sensorimotor synapses [48] . Alternatively , subcellular memory traces may interact at the circuit level via interneurons , such as the DPM and APL neurons , for which previous studies showed their essential role in memory consolidation [6] , [38] , [49] , [50] . A recent behavioral and imaging study revealed that the spontaneous activity of MB-MV1 and MB-MP1 after olfactory conditioning is coordinated to control the stability of memory [51] . Such network level interaction can also be implemented as a neuronal integration of outputs from the multiple memories . Considering parallel formation and the interaction of distinct memory components by different Kenyon cell populations [7] , [30] , [31] , [32] , [33] , [52] , the latter scenario might be more likely . A combined neuroanatomical and computational analysis identified that the dendrites of MB-M3 , MB-MV1 and MB-MP1 neurons form a cloud cluster in the protocerebrum , and they project into the segregated domains of the MB ( Figure 7 ) . This neuronal configuration implies that a punishment signal undergoes parallel processing in the MB via distinct dopamine pathways to induce different memory traces . The response profile of MB-projecting dopamine neurons is indeed distinct , suggesting that they receive input from different neurons [21] . Characterization of presynaptic neurons innervating the dendritic regions of the dopamine neurons will help identifying a cellular mechanism of parallel processing of reinforcement . Recent studies including our results highlight distinct functions of single dopamine neurons . Especially , MB-MP1 neurons are required for: 1 ) Suppression of the retrieval of appetitive memory when flies are not starved [26]; 2 ) Formation of aversive odor memory by mediating electric shock ( this study ) ; 3 ) Regulation of long-term memory by synchronized spontaneous activity together with MB-MV1 after spaced training [51] . Although these functions seem incompatible at the first glance , we suggest context-dependent roles of single dopamine neurons . Functions of MB-MP1 neurons in suppression of conditioned odor approach and aversive reinforcement signaling may be reasonable , as it is not appropriate timing for an animal to follow appetitive memory when challenged by a noxious stimulus ( aversive reinforcement ) . Implementing these two functions to the same MB-extrinsic neuron can be an elegant design by evolution , since both appetitive and aversive memories are formed and stored in the MB . Furthermore , in the light of a recent study , the results of MB-MP1 neurons suggests that dopamine may play opposing roles in the formation and the degradation of ARM ( Figure 13 ) [51] . A key difference of the opposing actions of PPL1 neurons is that Kenyon cells are simultaneously activated by an odor . Therefore , the effect of dopamine release from MB-MP1/MB-MV1 neurons on Kenyon cell synapses might be dependent on the state of the cells . These distinct modes of dopamine action may be better characterized by physiological means , such as functional imaging or electrophysiology .
We generated flies for behavioral and anatomical studies using the GAL4/GAL80 lines and transgenes listed in Table S1 . 5htr1b-GAL4 ( II ) was generated with the enhancer fragment of 5htr1b gene . The fragment ( −542–+2606 ) was amplified using the primer pair GTC AAATTCGGTCTGGCATT and CTTGCCTATGATGGTGACG by PCR and cloned into the pCRII-TOPO® vector ( Invitrogen ) . After verifying the sequence , the fragment was cloned into the p221-4 GAL4 vector ( gift from E . Knust ) . UAS-UAS-Shits1 ×3 is a combination of P-element insertion on X and multiple insertions on third chromosome , which is identical to Shi2 in [44] , [47] . UAS-Shits1 ×1 is isolated from multiple insertions on third chromosome by recombination , obtained from T . Préat lab . For experiments with UAS-Shits1 and UAS-dTrpA1 , flies were raised at 18°C and 25°C , respectively , at 60% relative humidity . UAS-Shits1 and UAS-dTrpA1 flies were aged 8–14 and 7–12 days after eclosion , respectively , to allow sufficient accumulation of effecter proteins without age-related memory impairment . For anatomical studies , females of 5–10 days after eclosion at 25°C were analyzed . For olfactory conditioning , we used 4-methylcyclohexanol and 3-octanol diluted in the paraffin oil ( 1∶10 ) . One odor was presented for 1 min at elevated temperature or with 12 pulses of electric shocks ( 90 V ) ( Figure 1D ) . Subsequent to 1-min air flush , another odor was presented for 1 min . The reciprocal group of flies was trained by the protocol in which the identity of odors was altered . After a given retention time , the conditioned odor response was measured in a T-maze for two minutes . Then , a performance index ( PI ) was calculated by taking the mean preference of the two reciprocal groups [53] . The first odor was paired with reinforcement in a half of experiments and the second odor was paired with reinforcement in another half so that the effect of the order of reinforcement is canceled [53] . The protocol for conditioning with thermo-activation by dTRPA1 was essentially same as the standard protocol of olfactory conditioning using electric shock [45] , [53] , [54] , except that flies were transferred to the pre-warmed T-maze in the climate box only during the presentation of one of the two odorants ( 60 sec ) [25] . To minimize the noxious effect of heat itself , we used moderate temperature ( 30°C ) for activation . This temperature shift by itself scarcely induced a significant memory in control genotypes [25] . For measuring the ARM , trained flies were transferred into pre-cooled tube on ice for 60 sec at 100 min after training , then transferred back to the pre-warmed tube at 25°C . Statistical analyses were performed using Prism ( GraphPad Software ) . Most of the tested groups did not violate the assumption of the normal distribution and the homogeneity of variance . Therefore , mean performance indices were compared with t-test , or Dunnett's multiple comparison test or Bonferroni-post test for selected pairs following one-way or two-way ANOVA . For the groups that violated the assumption of parametric statistics , Mann-Whitney test ( Figure 13E ) or Dunn's multiple comparison test following Kruskal-Wallis-test was applied ( Figure 9D ) . The brain and thoracicoabdominal ganglion were prepared for immunolabeling and analyzed as previously described [16] , [25] , [54] . For dopamine staining ( Figure S1 ) , brains were fixed with 0 . 6% glutaraldehyde in PBS for 30 min; the unreacted aldehyde groups were subsequently reduced with 0 . 1% ( w/v ) sodium tetraborhydrate . Brains were embedded in 7% agarose and thick ( 150 um ) sections were obtained with Laica Vibratome and labelled with the polyclonal antibody against conjugated dopamine ( MoBiTech ) . Frontal optical sections of brain samples were taken with confocal microscopy , Olympus FV1000 or Leica SP2 . For evaluating the effect of GAL80 , brains to be compared were scanned with identical microscopy setting . Images of the confocal stacks were analyzed with the open-source software Image-J [55] . Intensity-based affine and non-rigid registration of whole brains were performed with the toolbox elastix [43] . Confocal images of entire brains of GAL4/UAS-mCD8::GFP were acquired at 1024×512 pixel resolution . All brains were registered to a representative brain using counterstaining with TH as a reference channel . TH signals are mainly composed of sparse bright landmarks and a detectable background , allowing affine registration . The transformations computed with the TH channel were then applied to the mCD8::GFP channel . For some samples , non-rigid registration based on BSpline interpolation has been applied following affine transformation . Critical parameters , such as grid spacing , were empirically optimized . The accuracy of the registration was evaluated manually based on the matching of anti-TH signals . To compare positions of dendritic arbors of dopamine neurons , we selected samples with a small registration error ( <20 µm ) and signal from the other GAL4 expressing cells was manually masked ( see Figure 4A , Figure 5B , and [25] for the GAL4 expression patterns ) , and mCD8::GFP channels of different drivers were blended as different colors using ImageJ . The final images in Figure 7D–7F and Videos S1 , S2 , S3 show only regions surrounding the MB . Employed GAL4 drivers: NP5272 , NP2758 , 5htr1b-GAL4 ( Figure 7D and Video S1 ) ; MZ840 , NP6510 and MZ19;Cha3 . 3kb-GAL4 ( Figure 7E and Video S2 ) ; NP5272 , NP2758 , 5htr1b-GAL4 , NP6510 and MZ19;Cha3 . 3kb-GAL4 ( Figure 7F and Video S3 ) .
|
Punishment not only repels animals but also drives the formation of aversive memory of contiguous stimuli . Guided by the memory , animals can later avoid the cues that predict negative outcome . How is a punishing event represented in the brain ? We have found that at least three types of dopamine neurons in the Drosophila brain contribute to memory formation . Genetic activation of these neurons temporally paired with an odor presentation induced aversive odor memory , raising a question about the functional distinction of these neurons . Here we characterized aversive memories induced by these dopamine neurons . The magnitude of immediate memory and following memory decay differ greatly among the three cell types . Interestingly , combinatorial activation of two cell types revealed that induced memory is not a simple sum of the two memories , but rather the result of non-linear interaction specific for different retention times . Taken together , we propose that a punishing event induces aversive memory with unique temporal dynamics by tuning the activation of selective dopamine neurons .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"behavioral",
"neuroscience",
"animal",
"genetics",
"physiology",
"genetics",
"integrative",
"physiology",
"biology",
"anatomy",
"and",
"physiology",
"neuroscience",
"learning",
"and",
"memory",
"genetics",
"and",
"genomics"
] |
2012
|
Three Dopamine Pathways Induce Aversive Odor Memories with Different Stability
|
Schistosomiasis is a neglected infection affecting millions of people , mostly living in sub-Saharan Africa . Morbidity and mortality due to chronic infection are relevant , although schistosomiasis is often clinically silent . Different diagnostic tests have been implemented in order to improve screening and diagnosis , that traditionally rely on parasitological tests with low sensitivity . Aim of this study was to evaluate the accuracy of different tests for the screening of schistosomiasis in African migrants , in a non endemic setting . A retrospective study was conducted on 373 patients screened at the Centre for Tropical Diseases ( CTD ) in Negrar , Verona , Italy . Biological samples were tested with: stool/urine microscopy , Circulating Cathodic Antigen ( CCA ) dipstick test , ELISA , Western blot , immune-chromatographic test ( ICT ) . Test accuracy and predictive values of the immunological tests were assessed primarily on the basis of the results of microscopy ( primary reference standard ) : ICT and WB resulted the test with highest sensitivity ( 94% and 92% , respectively ) , with a high NPV ( 98% ) . CCA showed the highest specificity ( 93% ) , but low sensitivity ( 48% ) . The analysis was conducted also using a composite reference standard , CRS ( patients classified as infected in case of positive microscopy and/or at least 2 concordant positive immunological tests ) and Latent Class Analysis ( LCA ) . The latter two models demonstrated excellent agreement ( Cohen’s kappa: 0 . 92 ) for the classification of the results . In fact , they both confirmed ICT as the test with the highest sensitivity ( 96% ) and NPV ( 97% ) , moreover PPV was reasonably good ( 78% and 72% according to CRS and LCA , respectively ) . ELISA resulted the most specific immunological test ( over 99% ) . The ICT appears to be a suitable screening test , even when used alone . The rapid test ICT was the most sensitive test , with the potential of being used as a single screening test for African migrants .
Schistosomiasis is a fluke worm infection affecting at least 78 countries and more than 200 million people worldwide , and causing an estimated burden of 3 . 3 million disability-adjusted life-years ( DALYs ) [1 , 2] . Ninety-three percent of the global cases occur in sub-Saharan Africa , mostly caused by the species Schistosoma mansoni and S . haematobium . In this area , approximately 300 , 000 deaths due to schistosomiasis are estimated annually [3–6] . A huge wave of migrants , in particular asylum seekers , has reached Italy from sub-Saharan Africa in the past few years [7] . The health screening of newly arrived migrants is usually limited to detect potentially transmissible diseases such as tuberculosis and scabies . Schistosomiasis , often clinically silent , is not routinely screened . Given the severe complications related to both S . mansoni and S . haematobium [8–12] , and the availability of an effective and relatively inexpensive treatment [1 , 13 , 14] , adequate protocols for screening and treatment of migrants are required . Further , the recent outbreak of urinary schistosomiasis in Corsica [15] , originated by a West African strain [16 , 17] , is a useful reminder that local transmission of schistosomiasis in Europe is still possible [18 , 19] . The estimated prevalence of schistosomiasis in African migrants , reported from a scarce European literature , ranged from 9–15% if based on use of microscopy [20 , 21] to 5 . 8–24 . 7% if based on ELISA serology [22 , 23] . Clearly , all data obtained from reference centers are to be taken with caution as they may not be representative of the general population . These prevalence estimates are based on relatively insensitive tests , which under represent infections with low parasite loads . Traditionally , the gold standard for the diagnosis of schistosomiasis has been considered to be the urine and stool microscopy of several specimens [24] . Microscopy is 100% specific , but the sensitivity ( 40–60% ) varies with the intensity of infection , the number of specimens collected , and the circadian and day-to-day variation of egg counts in stool and/or urine [25–27] . The diagnosis of schistosomiasis by detection of specific antibodies ( with tests generally based on crude antigens of S . mansoni ) is more sensitive than microscopy , particularly in light infections [26 , 28] . However , commercial serologic tests for schistosomiasis have a sub-optimal sensitivity too , in particular for S . haematobium infections ( ranging from 21 . 4% to 71 . 4% ) [29] . The combinations of two or more serologic tests , markedly increased the sensitivity of serology to almost 80% [29 , 30] . Other diagnostic techniques have been implemented more recently [24 , 31 , 32] . A Western Blot ( WB ) containing S . haematobium and S . mansoni soluble antigens was used on a limited number of subjects exposed to the recent outbreak of urinary schistosomiasis in Corsica , where a hybrid form between human S . haematobium and animal S . bovis was identified [18 , 33] . However , the accuracy of this test has never been formally evaluated . Previously , an immunoblot produced by the same manufacturer , using S . mansoni antigens only , had showed for this species a good accuracy ( sensitivity 89 . 5% and specificity 100% ) [34 , 35] . The urine Circulating Cathodic Antigen ( CCA ) is a rapid test , based on the identification in the urine of the genus-specific proteoglycan antigen of the schistosomal gut epithelium , regurgitated by the live adult worm only [36] . The CCA test has some limitations in identifying S . haematobium infection . In a Cochrane meta-analysis of studies conducted in endemic areas , the CCA test , compared to microscopy , showed an average sensitivity and specificity of 39% and 78% , respectively , for S . haematobium infection , and of 89% and 55% , respectively , for S . mansoni . This test has been utilized only recently with immigrants in Europe [37 , 38] . A rapid diagnostic test ( RDT ) incorporating antigens of adult S . mansoni cercarial transformation fluid for detection of antibodies in blood showed a sensitivity and specificity of 100% and 39 . 5% , respectively , in a schistosome-endemic area [39] . However , the reference standard used in that study was a fecal examination using Kato-Katz concentration technique , admittedly less sensitive than the index test . No studies have been carried out in non-endemic countries . The main purpose of our study was to identify the more accurate screening strategy for schistosomiasis in a non endemic area . The target condition was schistosomiasis ( caused by S . mansoni , S . haematobium , or both ) rather than the infection by either species . The rationale is that , with the exception of microscopy ( and potentially PCR , that was not targeted by this study ) , the other techniques are not able to discriminate between the two species . Furthermore , the treatment of both infections is with the same drug praziquantel . The main objective was to assess the accuracy of a series of diagnostic tests for the screening of schistosomiasis , including: a ) the new , commercially available RDT ( Schistosoma ICT IgG-IgM ) , and the Western Blot ( Schisto II Western Blot IgG ) ( LD-BIO Diagnostics , Lyon France ) , both aimed at detecting S . mansoni and S . haematobium antibodies; the former test has been made commercially available recently ( October 2015 ) and is based on the principle of the homogeneous sandwich ( immunological reaction of 2 identical antigenic epitopes with the two binding sites of a bivalent antibody ) . b ) the tests routinely used at CTD , namely: the circulating cathodic antigen ( CCA ) urine dipstick test for S . mansoni ( NADAL CCA Bilharzia test , nal von minden , Germany ) , an enzyme-linked immunosorbent assay ( ELISA Bordier Affinity Products , Crissier , Switzerland ) , and the microscopic examination of stool and urine .
Ethical clearance from Comitato Etico Provinciale di Verona e Rovigo: protocol n 33909 of July 13th , 2016 . At the Centre for Tropical Diseases , all subjects submitted to any serological exam are asked to sign an informed consent for the anonymous storage of a serum specimen for any future research purpose . Parents’ or legal guardians’ consent is obtained for individuals of less than 18 years of age . The assessment of accuracy of all tests for schistosomiasis is hampered by the lack of a gold standard . In particular , microscopy ( of stools and urine ) is virtually 100% specific but lacks sensitivity [25–27] . On the contrary , serologic tests are known to be more sensitive but may provide false-positive results [24 , 29 , 30] . Therefore , the test accuracy was evaluated using three different standards: From one to three urine samples were obtained ( from 10 a . m . to 12 a . m ) over consecutive days for each patient . The urine were subjected to CCA dipstick test and to filtration method for S . haematobium egg count on the day of the sample collection [36 , 43] . Up to 10 ml of venous blood was collected from each patient in silica-coated tubes without anticoagulant . The serum was separated by centrifugation at 3000 rpm for 5 min . Bordier ELISA was executed on 10 μl of serum , and remaining aliquots were stored at -80°C . Both Schisto II Western Blot IgG and Schistosoma ICT IgG-IgM were performed using the latter serum . Cutoffs for each test were pre-determined prior to testing .
The proportion of positive microscopic results in the population under study was 65/373 ( 17 . 4% ) . In particular , S . mansoni eggs were found in the stools of 32/373 subjects ( 8 . 6% ) and S . haematobium eggs in the urine of 40/373 subjects ( 10 . 7% ) . Seven subjects ( 1 . 9% ) had both the infections . The results of the 4 index tests according to microscopy are reported in Table 1 . Patients were classified as microscopically positive if ova of at least one of the two species were identified in the stools or urine . ICT and WB presented the highest sensitivity ( 94% and 92% , respectively ) and negative predictive values ( NPV ) : 98% , both tests . CCA demonstrated a low sensitivity ( 48% ) , while its specificity was the highest ( 93% ) . Considering S . mansoni only , CCA sensitivity increased to 72% ( 23/32 ) , while that of the other tests resulted similar , in particular: ELISA , 84% ( 27/32 ) ; WB , 91% ( 23/32 ) ; and ICT , 94% ( 29/32 ) . Considering samples with positive microscopy for S . haematobium and negative for S . mansoni , CCA sensitivity dropped to 24% ( 8/33 ) , while that of the other tests , again , resulted similar , in particular: ELISA , 79% ( 26/33 ) ; WB , 94% ( 31/333 ) ; and ICT , 94% ( 31/33 ) . We also assessed if the presence of haematuria significantly influenced the proportion of CCA positivity . This data was available for about half of the study subject ( 185/373 or 49 . 6% ) . Considering subjects who resulted microscopically negative to both species , CCA resulted positive in 6/43 subjects with haematuria ( 14% , CI 5–28 ) and in 9/73 subjects without haematuria ( 11% , CI 5–20 ) . The proportion of positive results according to the CRS was 38 . 6% ( 144/373 ) . According to LCA modeling , the proportion of subjects classified in Class 1 was 35 . 6% . The results are detailed in Table 2 . ICT confirmed the highest sensitivity ( 96% for both models ) , followed by WB . CCA was again the test with the lowest sensitivity , that resulted very low ( 29% , both models ) . On the other hand , although the specificity of CCA was confirmed to be higher than 90% , both CRF and LCA identified ELISA as the indirect test with the highest specificity , being over 99% . The accuracy and predictive values , according to the three reference standards used , is summarized in Fig 2 . Subjects with at least 2 concordant positive results of the index tests OR with a positive microscopy ( irrespective of the other results ) had 91 . 1% probability of being classified as cases by LCA , while all the others ( negative microscopy and <2 positive index tests ) had 99 . 5% probability of being classified as non-cases by LCA . The concordance between the two methods of classification was excellent ( Cohen’s kappa = 0 . 92 , indicating near-perfect agreement ) . Assuming using ICT as a screening test ( then with a second , confirmatory test in case of ICT positivity ) , we assessed , using the LCA model , the positive and negative predictive values of a combination of a positive ICT ( the test with the highest sensitivity ) with a positive or a negative second test , respectively . The results are reported in Table 3 .
Schistosomiasis prevalence was very important in the study population , composed for the vast majority of asylum seekers . Even when considering microscopy only , that is not sufficiently sensitive , the proportion of positives was higher than 17% . According to CRS and LCA , this proportion was very similar , 38 . 6% and 35 . 6% , respectively . The urinary , rapid antigen test CCA was unsatisfactory: its sensitivity was poor , and this , with any reference standard used , makes it inadequate for screening purpose . The higher sensitivity found by other studies [37] probably reflects higher , average parasitic loads and a different reference standard based on microscopy only . Therefore , CCA is possibly a useful test ( for S . mansoni ) when the goal is to identify infections in an endemic country , but , according to our results , its sensitivity in non-endemic settings should be better assessed , before considering it as a potential screening tool , considering that the goal is to find and treat any infection , including chronic infections with a low egg output . WB showed an excellent combination of sensitivity and specificity , although the latter was not as high as one might expect by this technique . According to the manufacturer , a single band is sufficient to define a positive result . It is possible that a more conservative definition ( i . e . requiring at least two positive bands ) would bring the test specificity closer to 100% , though , expectedly , at the expense of some loss of sensitivity . We plan to investigate this with the next , longitudinal study ( see below ) . ELISA was the test providing the most surprising results , as it was found to be virtually 100% specific according to both CRS and LCA . This result was not expected , as serologic tests in general are held to give a variable proportion of false positive results , due to cross-reaction with other pathogens and to other reasons [24] . We tried to figure out if a possible selection bias could explain this unexpected result , but we couldn’t find any , although the retrospective design of the study suggests that the results should be interpreted with some caution . According to both CRS and LCA , the sensitivity of microscopy was poor , thus confirming that finding schistosome eggs in the stools or urine is diagnostic , but their absence does not by any mean rule out the infection . All the main results , both of the infection prevalence in the study population and of the accuracy of the different tests , were very similar between the two models , indicating a satisfactory robustness of the ( independent ) methods used . The agreement between the two models in classifying the study subjects was nearly perfect ( Cohen’s kappa test = 0 . 92 ) . Should a screening program be planned , it should rely on a sensitive approach , that minimizes the risk of leaving infected subjects without the necessary treatment . Regardless the reference standard used , the ICT resulted the most sensitive test , with the potential of being used as a single screening test ( NPV >97% according to the primary and composite reference standard , and to LCA ) . The PPV of ICT was low when using the primary reference standard ( reflecting the low specificity when a poorly sensitive test is used as reference ) , but reasonably high according to both the composite reference standard and the LCA , thus probably justifying a treatment , in a screening context , even without a confirmation test . Would a second test necessary/useful ? The answer to this question depend on the context . Newly arrived immigrants/asylum seekers are often a very mobile group , and it is crucial to be able to screen ( and treat the positives ) at the first visit . The ICT appears to be sufficiently accurate for this purpose . On the other hand , in a clinical context , a confirmatory test would be obviously indicated . A positive microscopy would be the only test providing the certainty of infection , but a negative result would by no mean rule out the infection ( Table 3 ) . ELISA is a good confirmer but a weak excluder , too . ( Table 3 ) . Only the combination of more , concordantly negative tests would safely exclude the infection in a subject with a positive ICT . In summary , ICT is the ideal test for screening purpose . It is a simple test , not even requiring a laboratory or any special equipment , can be read in a few minutes and is the only test that virtually rules out the infection , if negative . This test would be even more suitable for screening purpose if validated for use on whole blood obtained through a finger prick . Unfortunately the producer has not validated this procedure . In a screening context , this test can be used alone as a tool to decide the treatment , the only reasonable alternative being , in a high prevalence population such as our study population , a presumptive treatment of all immigrants coming from high prevalence countries , considering the potential severity of the chronic infection and the relatively harmless and highly effective treatment available ( praziquantel ) . The cost-effectiveness of either approach was beyond the scope of this research and should be planned as a further study . The new , rapid diagnostic test ICT is a suitable screening tool for schistosomiasis . A positive result should ideally be confirmed by a second test . ELISA and of course microscopy are the best confirmers of a positive ICT . Nevertheless , neither second test , if negative , is sufficient to rule out the infection . In a screening context , the ICT should probably used alone as a decision tool about treating or not the infection , the only practical alternative being a presumptive treatment of the population at risk .
|
Schistosomiasis is probably the most important of the neglected tropical diseases ( NTD ) caused by helminthes ( worms ) . It is acquired bathing in freshwater in endemic areas . The life cycle is complex and involves freshwater snails . Schistosomiasis , caused by Schistosoma mansoni , S . haematobium and less frequently by other species , affects more than 200 million people , mostly in Africa , and may chronically cause irreversible damage of the liver ( S . mansoni ) or of the kidneys and the urinary tract , including cancer of the bladder ( S . haematobium ) . As in chronic infections eggs of both species are often missed by microscopy of faeces and urine , with this retrospective study we evaluate the accuracy of different , alternative diagnostic tests , for the screening of schistosomiasis in African migrants and asylum seekers , of whom many thousands reach the Italian coast every year proceding from the most endemic areas . The most interesting finding of our study is that a rapid diagnostic test for antibody detection in blood , easy to use as a point-of-care tool , resulted the most sensitive of the five tests evaluated , and thus is very promising as a screening tool even when used without any additional test .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"schistosoma",
"invertebrates",
"schistosoma",
"mansoni",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"helminths",
"tropical",
"diseases",
"parasitic",
"diseases",
"animals",
"urine",
"neglected",
"tropical",
"diseases",
"immunologic",
"techniques",
"research",
"and",
"analysis",
"methods",
"serology",
"schistosoma",
"haematobium",
"immunoassays",
"helminth",
"infections",
"schistosomiasis",
"diagnostic",
"medicine",
"anatomy",
"physiology",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2017
|
Accuracy of parasitological and immunological tests for the screening of human schistosomiasis in immigrants and refugees from African countries: An approach with Latent Class Analysis
|
The World Health Organization ( WHO ) recommends oral cholera vaccines ( OCVs ) as a supplementary tool to conventional prevention of cholera . Dukoral , a killed whole-cell two-dose OCV , was used in a mass vaccination campaign in 2009 in Zanzibar . Public and private costs of illness ( COI ) due to endemic cholera and costs of the mass vaccination campaign were estimated to assess the cost-effectiveness of OCV for this particular campaign from both the health care provider and the societal perspective . Public and private COI were obtained from interviews with local experts , with patients from three outbreaks and from reports and record review . Cost data for the vaccination campaign were collected based on actual expenditure and planned budget data . A static cohort of 50 , 000 individuals was examined , including herd protection . Primary outcome measures were incremental cost-effectiveness ratios ( ICER ) per death , per case and per disability-adjusted life-year ( DALY ) averted . One-way sensitivity and threshold analyses were conducted . The ICER was evaluated with regard to WHO criteria for cost-effectiveness . Base-case ICERs were USD 750 , 000 per death averted , USD 6 , 000 per case averted and USD 30 , 000 per DALY averted , without differences between the health care provider and the societal perspective . Threshold analyses using Shanchol and assuming high incidence and case-fatality rate indicated that the purchase price per course would have to be as low as USD 1 . 2 to render the mass vaccination campaign cost-effective from a health care provider perspective ( societal perspective: USD 1 . 3 ) . Based on empirical and site-specific cost and effectiveness data from Zanzibar , the 2009 mass vaccination campaign was cost-ineffective mainly due to the relatively high OCV purchase price and a relatively low incidence . However , mass vaccination campaigns in Zanzibar to control endemic cholera may meet criteria for cost-effectiveness under certain circumstances , especially in high-incidence areas and at OCV prices below USD 1 . 3 .
Despite efforts to improve water supply and sanitation , cholera still represents a serious public health burden in low- and middle-income countries . In 2009 , more than 220 , 000 cases and almost 5 , 000 deaths were reported to the World Health Organization ( WHO ) [1] . Due to underreporting and difficulties with surveillance , however , the true burden is likely in the range of 3 million cases and 100 , 000 deaths per year [2] , [3] . A recent review of official cholera-related morbidity and mortality data from the WHO Africa region also indicated a potential economic burden of cholera for families and the health sector [4] . Cholera is an enteric bacterial disease caused by Vibrio cholerae serogroup O1 or O139 that usually occurs in sudden epidemics . Main features include acute , profuse watery diarrhea and vomiting that may lead to dehydration with concurrent electrolyte loss and eventually death if timely treatment is unavailable . Even though case-fatality rates ( CFRs ) may reach 50% , a rate below 1% has been achieved with proper case management [3] , [5] . Treatment is based on prompt rehydration with oral rehydration solution ( ORS ) for mild to moderate cases and intravenous ( IV ) fluids for severe cases [3] . Antibiotics are recommended for severe , and also moderate cases , to reduce the duration of episodes and shedding of infectious V . cholerae [3] , [6] . Traditionally , cholera control has been based on prevention ( i . e . , adequate water supply , improved sanitation and health education , and timely treatment ) . The role of vaccination for cholera control has recently received increased attention from public health officials; the WHO recommends oral cholera vaccines ( OCVs ) as a supplementary public health tool to traditional prevention and treatment in endemic and epidemic settings [7] . A series of research studies , done as part of the Diseases of the Most Impoverished ( DOMI ) project coordinated by the International Vaccine Institute ( IVI ) , evaluated the use of OCVs in Asia and Africa for control of endemic cholera . Private demand for cholera vaccines was examined through willingness-to-pay studies [8]–[11] , costs of illness ( COI ) and mass vaccination data were collected [12]–[14] , and cost-effectiveness and cost-benefit analyses were performed [15] , [16] . Besides the recent article by Poulos et al . [13] , published information about COI due to cholera is lacking even though patient-level data is needed for economic evaluations to improve local planning of cholera control . A joint initiative between the WHO , the IVI and the Ministry of Health of Zanzibar ( MoH ) implemented a mass vaccination campaign with an OCV in two selected cholera-endemic areas of Zanzibar in 2009 . This intervention-cum-research project provided the opportunity to assess costs of immunization in an endemic setting . Public COI—defined as fixed and variable costs borne by the health care provider for setting up and running cholera treatment centers ( CTCs ) —were estimated from three outbreaks that happened in 2009 outside the mass vaccination target communities . Private direct COI—defined as medical and non-medical expenses related to patient treatment , and indirect COI—defined as loss of income borne by patients and their families—were elicited from a sample of patients admitted to CTCs during these outbreaks . This study aims to estimate ( i ) public and private COI due to cholera , ( ii ) costs of an oral cholera mass vaccination campaign , and ( iii ) the cost-effectiveness of using OCVs in endemic regions of Zanzibar from a health care provider and a societal perspective .
Written informed consent was obtained from all study participants interviewed for private costs of illness . Patients aged 18 years or older were directly interviewed while caregivers were interviewed if the patient was younger than 18 years . No incentives were provided to them . The protocol of this study was cleared by the WHO Research Ethics Review Committee and the MoH Ethics Committee . All data were handled confidentially and made anonymous before analysis . Zanzibar consists of two major islands , Unguja ( also named Zanzibar ) and Pemba , which are situated in the Indian Ocean about 40–60 km off the coast of Tanzania . Zanzibar , a semiautonomous polity within the United Republic of Tanzania , consists of five regions , which are subdivided into ten districts , 50 constituencies and 296 Shehias , the latter being the smallest administrative unit . The main islands cover ∼2 , 557 km2 ( Unguja: ∼1 , 651 km2 , Pemba ∼906 km2 ) . The archipelago is inhabited by a fast-growing population of ∼1 . 2 million Kiswahili-speaking Muslim people . Monthly mean per capita expenditure for all goods and services was TZS 21 , 000 ( ∼USD 18 ) in 2004/5 with a 2 . 1% share for health-related expenditures [17] . Life expectancy at birth has risen from 47 years in 1988 to 57 years in 2002 [18] . The economy of the islands depends on agriculture ( primarily cloves , coconuts/copra and seaweed ) , fishing and tourism . The public health care delivery structure in Zanzibar comprises two zones , Unguja and Pemba , each with three levels: the primary , the secondary and the tertiary level . Each zone is headed by a zonal medical officer . Most of the health care services are provided at the primary level through Primary Health Care Units ( PHCU ) ( n = 124 ) . The majority of these units is open during the day to outpatients and provides basic services . Primary Health Care Centers ( PHCC ) ( n = 4 ) are additional facilities on the primary level; they operate on a 24-hours basis and can admit up to 30 patients . At the secondary level , three district hospitals ( only in Pemba ) are operational while the country's only tertiary level hospital ( Mnazi Mmoja ) is located in the capital Stonetown in Unguja . The top causes of primary- and secondary-level outpatient visits in 2008 were upper respiratory tract infections ( 23% ) , pneumonia ( 10% ) , malaria ( 10% ) and diarrhea ( 9% ) [19] . In recent times , the first cholera outbreak with 411 cases and 51 deaths was reported in 1978 from two fishermen villages in Zanzibar [20] . More than a dozen outbreaks followed since then with almost annual episodes since the year 2000 . Reyburn et al . reported an annual incidence of 0 . 5 cases per 1 , 000 population based on a review of routine surveillance data for the years 1997 to 2007 [21] , although the true incidence was likely higher due to underreporting . A seasonal pattern can be observed that follows the rainy seasons ( usually from March to June and from October to December ) during which widespread flooding occurs . Such deteriorating environmental conditions subsequently expose the majority of inhabitants on both islands to an increased risk of waterborne diseases due to the scarcity of safe drinking water supplies and a generally poor or lacking sanitation infrastructure in periurban and rural areas . Based on a consideration of areas of recent cholera activity , three Shehias per island , adjacent to each other , were selected as sites for the mass vaccination campaign . In Unguja , the Shehias of Chumbuni and Karakana in Urban district and Mtopepo in West district were targeted for the campaign; in Pemba , the Shehias of Kengeja , Mwambe and Shamiani , all located in the rural southeastern Mkoani district , were chosen . Dukoral , the only OCV that was pre-qualified by the WHO in 2009 , was used in the mass vaccination campaign . Dukoral is a V . cholerae serogroup O1 whole-cell , killed vaccine containing recombinant cholera toxin ( CT ) B subunit protein; it has to be administered in two doses at least one week apart and requires a cold chain ( 2–8°C ) [22] . This OCV was originally designed for immunologically naïve travelers from the north to tropical countries; it is licensed for use from two years of age and above and was shown to be 60–90% protective for up to three years [23]–[25] . One three-ml vial of Dukoral contains 1×1011 killed V . cholerae O1 ( biotype classical and El Tor ) and 1 mg of the CT B subunit protein in a suspension . Because the CT B subunit protein is not gastric acid-fast , the suspension has to be mixed with 1 . 5 dl of drinking water and a buffer sachet containing effervescent granules of sodium bicarbonate before ingestion . Recipients need to fast one hour before and after ingestion . Table 1 describes cost components and sources of data collected for this study . Estimates for public COI were obtained from interviews with local experts and unvaccinated patients and from reports and record review . Cost data for the mass vaccination campaign were collected based on actual expenditure and planned budget data . Private direct and indirect costs were collected through interviews done with unvaccinated patients on Pemba . All costs are reported in 2009 USD from an economic perspective , based on mid-2009 exchange rates obtained from http://www . oanda . com/currency/converter/ . Based on a previous study for Bangladesh [29] , a model was developed in Microsoft Excel to estimate the costs and health effects of a mass vaccination campaign program compared to standard treatment in CTCs in Zanzibar . A static cohort of 50 , 000 individuals , reflecting the target population of the 2009 mass vaccination campaign in Zanzibar , was examined from a health care provider and a societal perspective . Input parameters for inclusion in the model were related to vaccine characteristics and vaccination costs , burden and impact of cholera , and public and private COI . Private providers were not considered since the majority of patients would visit public facilities in case of an outbreak [26] . Indirect effects due to herd protection were also included in the model since they may play a considerable role in the overall impact of cholera vaccination [30] and were shown to make community-based programs in three Asian and one African setting cost-effective [15] . The base-case model considered costs and effects of a one-time vaccination program over the duration of protection ( i . e . , three years ) . The annual number of cases without vaccination was obtained by multiplying the population size times the annualized cholera incidence obtained from surveillance of diarrhea cases with laboratory confirmation for cholera in the study area [31] . The annual number of cases under the vaccination program was derived from adding up direct and indirect effects of the vaccination program: direct effects were calculated by multiplying the annual incidence of cases without vaccination with ( 1 – protective efficacy among vaccinated people [PE] ) , coverage , and population size; indirect effects were calculated by multiplying the annual incidence of cases without vaccination with ( 1 – protective efficacy among unvaccinated people [PEU] ) , ( 1 – coverage ) , and population size . The variable PEU was derived using the concepts and a formula from Longini et al . [30] ( p . 1778 ) . It calculates what they refer to as “indirect vaccine effectiveness” = 1− ( r01/r02 ) , where r01 is the cholera incidence among unvaccinated people within a vaccinated sub-region and r02 is the cholera incidence in an unvaccinated sub-region . In the absence of incidence data among unvaccinated people from the mass vaccination campaign area , the incidence rate of 2 . 34 cases per 1 , 000 population ( after annualizing , Khatib et al . [31] ) calculated from people that resided in the lowest quintile of surrounding coverage ( i . e . , <39% ) in a cluster with a radius of 400 m around vaccinated households was used as proxy for r02 . Khatib et al . showed that herd protection effects mainly existed within that radius . A longer distance from the household of the vaccinated person was considered to dilute the benefit of herd protection . The incidence of 1 . 29 cases per 1 , 000 population ( after annualizing ) among all unvaccinated people was used as an approximation for r01 [31] . This leads to a base-case estimate of PEU = 45% . The number of annual deaths without a vaccination program was calculated by multiplying the CFR with the annual number of cases without vaccination . The number of deaths with a vaccination program was calculated by using the CFR times the annual number of cases under the vaccination program . Incremental cost-effectiveness ratios ( ICER ) calculated as incremental costs per death , per case and per disability-adjusted life-year ( DALY ) averted were used as outcome measures . Incremental costs were calculated as the difference between costs of the vaccination program and public COI saved due to the vaccination from the health care provider perspective . Private direct COI saved were added in the base-case model adopting the societal perspective . Private indirect COI saved were not included in the base-case model [32] . The number of deaths , cases or DALYs averted was equal to the difference in numbers with and without the vaccination program . DALYs , which are an aggregate measure combining morbidity ( i . e . , years of life lived with disability ) and mortality ( years of life lost ) , were calculated according to Jeuland et al . [15] , assuming no age weighing . Since no disability weights are available for cholera , the disability weight of 0 . 11 for diarrheal diseases [33] was used . Life expectancy at the average age of onset of 18 years based on patient data was obtained from WHO life tables for Tanzania [34] . The vaccine was directly purchased from the manufacturer at a UN price . Future effects were discounted at a rate of 3 . 0% for the base case . Campaign costs were not discounted since the mass campaign happened over one single year . Cost-effectiveness was examined according to widely-used WHO criteria that define an intervention as ‘cost-effective’ if the ICER is less than three times per capita gross domestic product ( GDP ) per DALY averted and as ‘highly cost-effective’ if the ICER is less than per capita GDP per DALY averted [35] . One-way sensitivity analyses were done to estimate the influence of changes in potentially influential input parameters on model outcomes . Such key parameters included vaccine purchase price and delivery costs , protective efficacy ( PE , PEU ) , duration of protection , incidence , CFR and so forth [15] . Plausible ranges were based on public health considerations ( for vaccine purchase price and delivery costs , incidence ) , guidelines ( discount rate ) and variation for local data ( PE , PEU , CFR , number of ill days , public and private COI ) . Base-case values and plausible ranges are presented in Table 2 . Threshold analyses examined at which vaccine purchase price the intervention would become cost-effective .
Table 3 presents the fixed and variable mean public COI at the three CTC sites . Fixed costs of USD 51 accounted for 85% of public COI , with mean fixed costs ranging from USD 21 to USD 88 . Direct and indirect human resources costs accounted for the majority of fixed costs; they were highest in Kiuyu Minungwini ( 86% ) , medium in Micheweni ( 85% ) and lowest in Chumbuni ( 80% ) . The remaining fixed costs were used for setting up and running the centers . Health care personnel working in Unguja received higher top up payments than in Pemba , but the latter were given food to cater for themselves while on shift . Variable costs of USD 9 . 2 were mainly driven by IV fluid use as patients were administered on average 8 . 8 liters , which cost USD 7 . 1 . Further details on public variable costs for treatment can be found as supporting information in Table S1 . A total of 95 individuals were interviewed . All but one of the interviewed patients had been admitted at the CTC at Micheweni PHCC . Total direct and indirect mean private COI amounted to USD 43 , with almost three-fourth ( USD 32 ) being indirect costs ( i . e . , productivity losses to the patient or caregiver and other household members ) ( Table 4 ) . Among direct costs , which amounted to USD 11 , feeding the patient at the CTC accounted for the biggest share ( USD 8 . 3 , 19% of total costs ) . Other direct costs , incurred for treatment ( mainly for plastic sheets needed to cover cots ) , transport and communication , were reported each by less than 3% . Total mass vaccination campaign costs amounted to USD 760 , 000 , with USD 510 , 000 ( 68% ) spent on vaccine purchase and USD 240 , 000 ( 32% ) on delivery ( Table 5 ) . The vaccine was purchased from SBL Vaccin AB , Sweden , at a price of USD 10 per course ( 2 doses ) . Delivery costs comprised transport of the vaccine from Stockholm to Zanzibar and procurement of cups and water required for the buffer solution ( 6 . 0% of campaign costs ) , the work of two experienced international consultants ( 14% ) , training of locally recruited implementers ( 1 . 3% ) and the implementation ( social mobilization and vaccination ) itself ( 10% ) . More details on delivery costs are presented as supporting information in Table S2 . At a vaccine purchase price of USD 10 per course , the estimated total costs per fully immunized individual amounted to USD 30 , with mean costs per vaccine course of USD 21 and mean costs for delivery of USD 9 . 7 . The latter amounted to USD 5 . 3 after exclusion of services from international consultants , lowering the overall total costs per fully immunized individual to USD 26 . Mean costs were adjusted for actual coverage of 50% , relating to 23 , 921 fully immunized individuals out of a population denominator of 48 , 178 used in the analysis by Khatib et al . [31] .
This study has several limitations . First , due to limited data availability , this cost-effectiveness analysis assessed the value for money of a population-wide OCV campaign and not of a targeted approach for high-risk or specific age groups , which might make the intervention cost-effective as shown by Jeuland et al . for school-based programs in Kolkata and Beira [15] . However , threshold analyses using Shanchol indicated that scenarios targeting high-risk groups may become cost-effective in Zanzibar if the OCV was procured at a price below USD 1 . 3 , a level acceptable by many public health policy makers in Asia [36] . Second , it may be argued that the assumption of a preference for health facilities during a cholera outbreak may not necessarily reflect actual behavior as patients could also be negatively influenced by accessibility problems . Local observation and informal interviews , however , support this assumption , and the dense primary health care system reduces transport issues ( according to the 2004/5 household budget survey , mean distance to a health care center in the urban and rural area was 0 . 4 and 1 . 7 km , respectively [17] ) and because treatment for diarrhea is free . Third , non-diarrhea patients were usually not treated or admitted by their local public health care facility during the time it operated as a CTC . People seeking treatment for non-diarrheal diseases ( e . g . , for malaria ) during an ongoing cholera outbreak will have to bear extra direct and indirect costs related to additional travel or potential serious complications due to delayed treatment . These additional costs have not been included in the cost-effectiveness analysis due to a lack of relevant data; future studies in the area are advised to collect estimates on the costs of patients who are not able to get treatment at their usual center to assess the relevance of this ‘crowding out’ effect on cost-effectiveness . Fourth , the ICER might have been overestimated because waning has not been included in the estimate for PE . Jeuland et al . have adjusted their estimate in year 3 down by 17% [15] . However , since this represents a limited effect , and since sensitivity analysis included a minimum PE of 47% , omission of waning as input parameter in the model is likely to have only a limited effect . Fifth , threshold analysis for Shanchol did not consider potential savings due to the probably easier and faster administration of this new vaccine by oral syringe; this may have resulted in more favorable cost-effectiveness , but any beneficial effect will likely be limited because delivery costs influenced the ICER only to a small extent . Sixth , even though uncertainty in input parameters was considered in one-way sensitivity analyses , no full probabilistic uncertainty and sensitivity analysis was conducted which would provide a more complete picture of the distribution of possible outcomes and may find that some combinations of assumptions lead to greater cost-effectiveness than identified in the one-way sensitivity analysis [38] . Finally , a cost-benefit analysis may provide more useful information to local policy makers than a cost-effectiveness analysis because it explicitly characterizes the monetary value of prevented disease . However , willingness-to-pay data were not available for Zanzibar and contingent valuation exercises were beyond the scope of this study . Also , since campaign coverage with the free OCV in Zanzibar was merely 50% overall [31] and the community demanded a free vaccine [51] , not making the OCV available for free in future campaigns would further jeopardize vaccine effectiveness and thus make such a program even less economical . The analysis presented here suggests that costs averted by a mass vaccination campaign with an OCV in endemic areas of Zanzibar were negligible when compared to standard treatment in decentralized cholera treatment centers . Mass vaccination was not cost-effective based on empirical data and the stated assumptions , mainly due to the relatively high purchase price and the relatively low cholera incidence in Zanzibar . However , mass vaccination campaigns in Zanzibar for endemic cholera control may meet WHO criteria for cost-effectiveness under certain circumstances , especially in high-incidence areas and when OCV prices are reduced to levels below USD 1 . 3 .
|
Despite efforts to improve water supply and sanitation , cholera still represents a serious burden in developing countries . Use of oral cholera vaccines ( OCVs ) in endemic and epidemic situations has recently shown a promising potential to mitigate this burden . To provide local decision-makers with specific information on OCV use for cholera control , we assessed the costs and benefits of a mass vaccination campaign that was conducted in 2009 in selected endemic areas of Zanzibar . We estimated the cost-effectiveness of OCVs by collecting health care provider and household costs of illness from cholera outbreaks and costs of the mass vaccination campaign that used the two-dose OCV Dukoral . Cost-effectiveness was expressed as the incremental costs of the one-off vaccination program per case , per death and per disability-adjusted life-year averted , over a three-year time period . Our model showed that the 2009 mass vaccination campaign in Zanzibar was not cost-effective , mainly due to the high OCV price ( USD 10 ) and the relatively low incidence . Threshold analyses with Shanchol , the second OCV that is recommended by the WHO , indicated that mass vaccination in Zanzibar to control endemic cholera may become cost-effective if done in higher incidence areas and when OCV prices are reduced to levels below USD 1 . 3 .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"public",
"health",
"and",
"epidemiology",
"social",
"and",
"behavioral",
"sciences",
"cost-effectiveness",
"analysis",
"non-clinical",
"medicine",
"neglected",
"tropical",
"diseases",
"cost",
"effectiveness",
"immunizations",
"infectious",
"diseases",
"cholera",
"health",
"economics",
"public",
"health",
"economics",
"socioeconomic",
"aspects",
"of",
"health"
] |
2012
|
Costs of Illness Due to Cholera, Costs of Immunization and Cost-Effectiveness of an Oral Cholera Mass Vaccination Campaign in Zanzibar
|
The mitochondrial genome of Trypanosoma brucei contains many cryptogenes that must be extensively edited following transcription . The RNA editing process is directed by guide RNAs ( gRNAs ) that encode the information for the specific insertion and deletion of uridylates required to generate translatable mRNAs . We have deep sequenced the gRNA transcriptome from the bloodstream form of the EATRO 164 cell line . Using conventionally accepted fully edited mRNA sequences , ~1 million gRNAs were identified . In contrast , over 3 million reads were identified in our insect stage gRNA transcriptome . A comparison of the two life cycle transcriptomes show an overall ratio of procyclic to bloodstream gRNA reads of 3 . 5:1 . This ratio varies significantly by gene and by gRNA populations within genes . The variation in the abundance of the initiating gRNAs for each gene , however , displays a trend that correlates with the developmental pattern of edited gene expression . A comparison of related major classes from each transcriptome revealed a median value of ten single nucleotide variations per gRNA . Nucleotide variations were much less likely to occur in the consecutive Watson-Crick anchor region , indicating a very strong bias against G:U base pairs in this region . This work indicates that gRNAs are expressed during both life cycle stages , and that differential editing patterns observed for the different mitochondrial mRNA transcripts are not due to the presence or absence of gRNAs . However , the abundance of certain gRNAs may be important in the developmental regulation of RNA editing .
The life cycle of Trypanosoma brucei involves two distinct environments , the animal host and the insect vector . These environments are distinct in temperature and nutrient composition , providing a unique challenge to T . brucei as it cycles between hosts . In the bloodstream , trypanosomes exist in two forms , the actively dividing slender form and the non-dividing stumpy form . The slender form is optimized to utilize its glucose rich environment , using glycolysis to generate energy [1] . The stumpy form appears to be transitional , activating mitochondrial genes in preparation for uptake in a blood meal by its tsetse fly vector and subsequent transfer to a harsher environment [1] . Once inside the tsetse fly , the parasite utilizes proline to drive oxidative phosphorylation and ATP production in the mitochondrion [2] . While the activity of the mitochondrion is relatively low during the bloodstream stage ( BS ) , expression of the mitochondrial genome is still essential [3 , 4] . In T . brucei , the mitochondrial genome consists of two types of DNA molecules , maxicircles and minicircles . Maxicircles are 22kb circular DNA that contain the genes for two ribosomal RNAs , 12S and 9S , and eighteen mRNA genes [5] . While some of the protein-coding genes do not require RNA editing prior to translation , most require extensive editing before they can be translated [for review see 6 , 7] . This process involves the insertion of hundreds of uridylates ( U ) s and less frequently deletion of Us , often doubling the size of the transcript . The sequence changes are guided by small complementary RNA molecules ( the guide RNAs ) that are encoded on the minicircles [8] . Minicircles make up the bulk of the kinetoplastid network ( anywhere from 5 , 000–10 , 000 present in each network ) with each minicircle encoding 3–5 gRNAs . In T . brucei , there are more than 200 different minicircle sequence classes ( ~1200 gRNAs ) [8] . Distinct differences in mitochondrial transcript abundance , polyadenylation and the extent of RNA editing are observed during the complex life cycle ( Table 1 ) . The pattern of differential RNA editing observed is especially interesting . For example , the cytochrome b ( CYb ) and cytochrome oxidase II ( COII ) mRNAs are edited during the insect stage , but are primarily unedited in bloodstream forms [9 , 10] . In contrast , editing of the NADH dehydrogenase subunit transcripts ( ND3 , ND7 , ND8 and ND9 ) and editing of the ribosomal protein subunit 12 transcript ( RPS12 ) appears to occur preferentially in bloodstream forms [5 , 11–15] . Other transcripts , cytochrome oxidase III ( COIII ) and ATPase subunit 6 ( A6 ) are edited in both life cycle stages [16 , 17] . Early studies using both Northern blot and primer extension analyses on a limited number of gRNAs indicate that gRNAs are present in both insect and bloodstream forms , suggesting that the regulation of RNA editing is not at the level of gRNA availability [13 , 18 , 19] . Our lab has previously published deep sequencing results of the gRNA transcriptome of the T . brucei EATRO 164 procyclic form [20] . Here we present the deep sequencing data for the gRNA transcriptome of a bloodstream form of EATRO 164 . A total of 211 populations of gRNAs were identified . We define a population as a group of gRNAs that may vary in sequence , but direct the editing of the same or near same region of the mRNA . Because kinetoplastid RNA editing allows G:U base pairing , most populations contain multiple sequence classes that can guide the generation of the same mRNA sequence . While the number of populations identified was similar to the number identified in the procyclic gRNA transcriptome ( 214 populations ) , the total number of gRNAs identified was much reduced and the coverage was less complete; full complements of gRNAs were only identified for COIII and CYb . In spite of the reduced number of gRNAs , an interesting correlation was found that suggests a relationship between the relative abundance of initiating gRNAs between stages and the developmental pattern of mRNA editing .
T . brucei brucei clone IsTar from stock EATRO 164 were grown in rats and isolated as previously described [37] . Bloodstream forms were virtually all long-slender forms isolated after 4 days of infection . Parasites were used immediately for isolation of mitochondria using differential centrifugation as previously described or stored frozen at -80°C until RNA extraction [20] . Both total RNA from whole parasites and mitochondrial RNA ( mtRNA ) from purified mitochondria were isolated by the acid guanidinium-phenol-chloroform method [38] . Rats were raised according to the animal husbandry guidelines established by Michigan State University . All vertebrate animal use procedures were approved by MSU’s Institutional Animal Care and Use Committee ( Application 03/11-051-00 ) . MSU has filed with the Office of Laboratory Animal Welfare ( OLAW ) an assurance document that commits the university to compliance with NIH policy and the Guide for the Care and Use of laboratory Animals . Samples of mtRNA and total RNA were both treated with DNAse RQI and size fractioned on a polyacrylamide gel as previously described [20] . Guide RNAs were extracted from the gel and prepped for sequencing using the Illumina ‘Small RNA’ protocol as previously described [20] . Libraries from both mtRNA and total RNA samples were deep sequenced on Illumina GAIIx . Reads were then processed and trimmed as previously described [20] . Data with two or more Ns , shorter than 20nts after trimming or with an overall mean Q-score < 25 were discarded . Redundant reads were then removed , while maintaining the number of redundant reads and reads containing fewer than 4 consecutive Ts were removed . To identify gRNAs , each transcript read was aligned to the conventionally edited mRNAs based on known base pairing rules ( canonical Watson-Crick base pairs and the G-U base pair ) . In the initial screen , no gaps were allowed in the alignment , allowing the formulation of the gRNA-mRNA alignment as an extended longest common substring ( LCS ) problem as previously described [10] . Matched gRNAs were then scored ( two points for G:C and A:U base pairs and one point for G:U base pairs ) . gRNAs with scores >45 were identified as guiding a specific region based on the identified mRNA fully edited sequence . Additional searches with reduced stringency ( scores >30 ) were performed on regions with low gRNA coverage . The matched gRNAs were sorted into populations based on their guiding positions , and the populations analyzed and sorted into major sequence classes .
In order to determine if the BS gRNA transcriptome contained a full complement of guide RNAs , the gRNA populations were aligned to the fully edited mRNAs ( S2 Fig ) . We note , that for an mRNA to be fully edited , not only must all editing sites on the mRNA be covered by a gRNA , the downstream gRNA must generate the anchor binding site for the subsequent gRNA . Therefore , adjacent gRNAs must overlap . Overall , there was an average of 17 nts of overlap between adjacent gRNAs , with the average overlap varying slightly by gene ( Table 3 ) . As the median Watson-Crick Anchor is 11 nts , in most cases , the overlap extends beyond the Watson-Crick anchor of the subsequent gRNA . However , we did observe a number of regions where the overlap is minimal . Currently , there is no data that stipulates the minimum anchor needed for efficient editing . However , we postulate that similar to microRNAs , for an anchoring sequence to be sufficiently specific , it should be at least six nucleotides [40] . Indeed , when examining the overlaps between most gRNAs , there are only ten ( four procyclic and six BS ) that are less than six nucleotides ( Fig 5 ) . We therefore used six nucleotides as a cut off to identify regions with potential missing guide RNAs for both life cycle stage transcriptomes . In contrast to the procyclic data , where full complements of gRNAs were identified for five of the mRNA transcripts ( A6 , COIII , CR4 , CYb , and RPS12 ) , in the BS transcriptome , a full complement of gRNAs was only identified for COIII and CYb . Overall , there are 12 edited regions where no gRNAs were identified , and five regions with weak gRNA overlaps in the BS data ( Table 5 ) . Of these 17 regions , seven belong to ND7 alone . Interestingly , nine of the 17 missing populations are in very low abundance in the procyclic data , having 100 or fewer reads . Because the number of reads in the BS data is ~3 . 5 fold less abundant , this could account for some of these regions of poor coverage . There are six regions that lack gRNA coverage in both data sets . These are found in CR3 , MurfII , ND3 and ND7 ( Table 5 ) . Interestingly , three of these regions are close to the 3’ end of their respective genes . Regions of weak overlap ( ND9 ( 238–242 ) , ND9 ( 609–612 ) ) and regions without gRNA coverage ( CR3 ( 278–292 ) , ND8 ( 541–553 ) ) that are unique to the procyclic transcriptome were also observed . Interestingly , the regions of poor procyclic coverage are found in CR3 , ND8 and ND9 , all transcripts that are preferentially edited in the BS form [5 , 13 , 14 , 34] . While the number of reads in the BS data is less abundant than the procyclic data in general , there are 87 gRNA populations with more identified reads than in the procyclic data set ( Table 6 ) . These populations are found in every gene except CYb and MurfII , but most of the populations belong to one of the NADH dehydrogenase subunits , particularly , ND7 or ND9 . Another trend worth noting is that the all of the genes with fully edited transcripts that are more abundant in the BS stage , ( CR3 , CR4 , ND3 , ND7 , ND8 , and ND9 , with the exception of RPS12 ) , have a higher percentage of classes that are more abundant in the BS ( Table 6 ) .
Surprisingly , complete gRNA coverage was observed only for the pan-edited COIII and for CYb , where editing is limited to the 5’ end . The identification of the CYb gRNAs was expected , as it has been previously reported that the gRNAs are present in both life cycle stages even though editing of CYb is limited to the procyclic stage [8 , 9] . The full coverage of COIII was also not surprising , as COIII was shown to be fully edited and equally abundant in both stages [17] . However , we expected to see complete coverage of ATPase 6 and RPS12 as both of these transcripts have been shown to be essential in both life cycle stages [3 , 44 , 45 , 50] . For ATPase 6 , we did identify a total of 29 gRNA populations that do cover all of the editing sites . However , one of the gRNAs ( bsA6 ( 643–667 ) ) has a single nucleotide mismatch ( C:U ) and one would introduce a frameshift ( bsA6 ( 520–553 ) ) . The C:U mismatch occurs near the middle of the gRNA , placing the C:U mismatch in a region that is unusually high in Gs and Cs ( S2A Fig ) . It may be that the G:C basepairs immediately upstream of the mismatch stabilize the gRNA/mRNA interaction , allowing it to be tolerated . The frameshift gRNA is also interesting , as it occurs just upstream ( 1 editing site ) of another site where we had previously observed a frameshift sequence anomaly . Both frameshifts ( the BS 4U and the Procyclic 11U ) generate a predicted protein with nearly the same amino acid sequence . As the frameshifts occur downstream of the highly conserved amino acid region involved in proton translocation [16] , it may be that this different carboxyl terminus is tolerated . Near full coverage is also observed for RPS12 . For this transcript , one BS identified gRNA ( bsRPS12 ( 96–121 ) ) has an A-nt insertion that disrupts the gRNA complementarity . Surprisingly , the other mRNA transcript found with near complete coverage was ND9 ( one gRNA has a single nt mismatch ) . All of the other mitochondrially encoded Complex I members did have substantial gaps in coverage . Currently , there is considerable debate on the necessity of Complex I subunits for either stage of the trypanosome life cycle . Studies using RNAi and knockout cell lines of nuclear-encoded members of Complex I have shown that the complex is unnecessary for survival in either life cycle stage [51 , 52] . However , the nuclear-encoded Complex I member genes are maintained [29] , and while we not did identify full coverage for the ND transcripts , a vast majority of the gRNAs were found in both life cycle stages . This study used high-throughput sequencing to characterize the gRNA transcriptome during the bloodstream stage of the trypanosome life cycle . This work suggests that gRNAs are expressed during both life cycle stages , and that differential editing patterns observed for the different mitochondrial mRNA transcripts are not due to the presence or absence of gRNAs . SAMN04302078 , SAMN04302079 , SAMN04302080 , and SAMN04302081 NCBI’s Sequence Read Archive .
|
Trypanosoma brucei is the causative agent of African sleeping sickness , a disease that threatens millions of people in sub-Saharan Africa . During its life cycle , Trypanosoma brucei lives in either its mammalian host or its insect vector . These environments are very different , and the transition between these environments is accompanied by changes in parasite energy metabolism , including distinct changes in mitochondrial gene expression . In trypanosomes , mitochondrial gene expression involves a unique RNA editing process , where U-residues are inserted or deleted to generate the mRNA’s protein code . The editing process is directed by a set of small RNAs called guide RNAs . Our lab has previously deep sequenced the gRNA transcriptome of the insect stage of T . brucei . In this paper , we present the gRNA transcriptome of the bloodstream stage . Our comparison of these two transcriptomes indicates that most gRNAs are present in both life cycle stages , even though utilization of the gRNAs differs greatly during the two life-cycle stages . These data provide unique insight into how RNA systems may allow for rapid adaptation to different environments and energy utilization requirements .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"sequencing",
"techniques",
"messenger",
"rna",
"parasitic",
"protozoans",
"parasitology",
"developmental",
"biology",
"trypanosoma",
"brucei",
"gambiense",
"protozoans",
"genome",
"analysis",
"energy-producing",
"organelles",
"mitochondria",
"bioenergetics",
"molecular",
"biology",
"techniques",
"rna",
"sequencing",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"life",
"cycles",
"molecular",
"biology",
"biochemistry",
"rna",
"trypanosoma",
"cell",
"biology",
"nucleic",
"acids",
"rna",
"editing",
"transcriptome",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"computational",
"biology",
"organisms",
"parasitic",
"life",
"cycles"
] |
2016
|
Analysis of the Trypanosoma brucei EATRO 164 Bloodstream Guide RNA Transcriptome
|
Due to their error-prone replication , RNA viruses typically exist as a diverse population of closely related genomes , which is considered critical for their fitness and adaptive potential . Intra-host demographic fluctuations that stochastically reduce the effective size of viral populations are a challenge to maintaining genetic diversity during systemic host infection . Arthropod-borne viruses ( arboviruses ) traverse several anatomical barriers during infection of their arthropod vectors that are believed to impose population bottlenecks . These anatomical barriers have been associated with both maintenance of arboviral genetic diversity and alteration of the variant repertoire . Whether these patterns result from stochastic sampling ( genetic drift ) rather than natural selection , and/or from the influence of vector genetic heterogeneity has not been elucidated . Here , we used deep sequencing of full-length viral genomes to monitor the intra-host evolution of a wild-type dengue virus isolate during infection of several mosquito genetic backgrounds . We estimated a bottleneck size ranging from 5 to 42 founding viral genomes at initial midgut infection , irrespective of mosquito genotype , resulting in stochastic reshuffling of the variant repertoire . The observed level of genetic diversity increased following initial midgut infection but significantly differed between mosquito genetic backgrounds despite a similar initial bottleneck size . Natural selection was predominantly negative ( purifying ) during viral population expansion . Taken together , our results indicate that dengue virus intra-host genetic diversity in the mosquito vector is shaped by genetic drift and purifying selection , and point to a novel role for vector genetic factors in the genetic breadth of virus populations during infection . Identifying the evolutionary forces acting on arboviral populations within their arthropod vector provides novel insights into arbovirus evolution .
Due to the low fidelity of their RNA-dependent RNA polymerase , rapid replication kinetics and large population size , RNA viruses consist of a heterogeneous intra-host population of related mutants , sometimes referred to as a quasispecies [1] . This mutant swarm as a whole defines the properties of the viral population , and is considered critical for the fitness and adaptive potential of RNA viruses [1] . For example , high fidelity poliovirus mutants are attenuated in mice in vivo , demonstrating the functional importance of intra-host genetic diversity for pathogenesis [2] . Arthropod-borne viruses ( arboviruses ) are maintained by transmission between vertebrate hosts and blood-feeding arthropods such as mosquitoes that serve as vectors . Although arboviruses span a wide range of viral taxa in the Togaviridae , Flaviviridae , Bunyaviridae , Rhabdoviridae and Orthomyxoviridae families , the vast majority are RNA viruses , with the single known exception of a DNA arbovirus ( African swine fever virus ) . The genetic plasticity of an RNA genome may confer arboviruses the remarkable ability to alternate between two fundamentally different hosts , and to quickly adapt to novel hosts [3] . Like other RNA viruses , high levels of intra-host genetic diversity are critical for arboviral fitness , as demonstrated in both host types for chikungunya virus [4 , 5] and West Nile virus [6–8] . Arboviruses usually rely on horizontal transmission between vertebrate hosts and arthropod vectors , although vertical transmission from an infected female arthropod to her offspring may occur [9 , 10] . After being ingested in a blood meal taken from a viremic vertebrate , arboviruses initially establish infection in the midgut epithelial cells of the arthropod vector . Transmission to another vertebrate host occurs after an extrinsic incubation period during which the arthropod develops a systemic infection that results in the release of viral particles in saliva . During infection of the arthropod vector , arboviruses are confronted with several anatomical barriers that are believed to impose severe population bottlenecks on viral populations [11] . Bottlenecks are dramatic reductions in population size , resulting in stochastic sampling of a small number of viral genomes from the mutant swarm . Population bottlenecks can significantly reduce the fitness of RNA viruses through accumulation of deleterious mutations that cannot be efficiently removed by purifying selection [12] . Initial infection and traversal of midgut cells , followed by virus dissemination and invasion of the salivary glands are expected to result in strong population drops that represents a challenge to maintaining arboviral genetic diversity during systemic vector infection [13] . Despite such population bottlenecks , arboviruses typically maintain high levels of genetic diversity during transmission by their arthropod vectors [11] . For example , analyses of West Nile virus populations in the midgut , hemolymph and saliva of Culex mosquitoes failed to document reductions in genetic diversity [14] . However , the authors of this study did not determine whether a large effective population size was maintained , or if viral genetic diversity was quickly replenished by mutation and demographic expansion following population bottlenecks . In a recent study of dengue virus ( DENV ) , genetic diversity was maintained during human-to-mosquito transmission but the variant repertoire changed substantially between venous blood and different organs of Aedes mosquitoes that became infected by feeding on the person [15] . Over 90% of DENV genetic variants were lost upon transition from venous blood to mosquito abdomen , as well as from abdomen to salivary glands , which led the authors to estimate that about a hundred viral genomes initially established a productive midgut infection [15] . However , this number could have been underestimated because the calculation assumed that the observed change in variant frequency was due to chance alone ( i . e . , it did not account for the effect of natural selection ) . Genetic drift and purifying selection , for example , can result in a similar loss of genetic diversity . The relative strength of natural selection and genetic drift is informed by the effective population size ( Ne ) , defined as the size of an idealized population that would drift at the same rate as the observed population [16] . Ne indicates whether the evolution of a population is better described as a deterministic ( selection ) or stochastic ( drift ) process . When Ne is large , competition between variants occurs with little interference of random processes . When Ne is small , stochastic sampling of variants counteracts selection and hinders adaptation . For example , genetic drift plays a limited role during systemic infection of the plant host by cauliflower mosaic virus , as viral populations maintain an effective size of several hundreds of viral genomes [17] . Understanding the relative role of genetic drift and natural selection is critical to evaluate the risk of arboviral emergence through adaptive processes [3] . For example , limited epidemic potential of an Asian lineage of chikungunya virus was associated with fixation of a deleterious deletion likely due to a founder effect [18] . In the present study , we investigated the intra-host evolution of DENV in the main mosquito vector Aedes aegypti by deep sequencing the full genome of viral populations at different time points of infection . Importantly , we accounted for the potential role of mosquito genetic variation on DENV intra-host genetic diversity . DENV intra-host genetic diversity has attracted considerable interest since the confirmation of its quasispecies nature [19] . Until now , however , most of this research has focused on viral genetic diversity in humans [20–23] . A few studies examined DENV intra-host genetic diversity in the mosquito vector [15 , 24 , 25] , but these studies did not account for vector genetic heterogeneity . There is substantial evidence for genetic variation in Ae . aegypti vector competence for DENV [26–32] , as well as specific interactions between Ae . aegypti genotypes and DENV genetic variants [33–37] . Our objectives were three-fold: ( i ) measure the bottleneck size during initial midgut infection of Ae . aegypti mosquitoes by DENV; ( ii ) monitor DENV intra-host genetic diversity during population expansion and systemic infection; and ( iii ) determine the influence of the vector genotype on bottleneck size and intra-host DENV genetic diversity .
The Institut Pasteur animal facility has received accreditation from the French Ministry of Agriculture to perform experiments on live animals in compliance with the French and European regulations on care and protection of laboratory animals . This study was approved by the Institutional Animal Care and Use Committee at Institut Pasteur . This study used a wild-type DENV-1 isolate ( KDH0026A ) that was originally recovered from the serum of a clinically ill dengue patient attending Kamphaeng Phet Provincial Hospital , Thailand as previously described [36] . Informed consent of the patient was not necessary because the virus was isolated in laboratory cell culture for diagnostic purposes ( unrelated to this study ) and , therefore , was no longer a human sample . The isolate was passaged three times in Aedes albopictus C6/36 cells prior to its use in this study . The full-length consensus genome sequence of the isolate is available from GenBank under accession number HG316481 . Aedes aegypti females used in this study belonged to the 16th generation of four isofemales lines ( referred to as A , B , C , and D thereafter ) derived from wild Ae . aegypti specimens collected in Kamphaeng Phet Province , Thailand . The lines were initiated by single mating pairs of field-caught males and females as previously described [36] . One male and one female from different collection sites ( subdistricts ) of the Muang district , Kamphaeng Phet Province , were randomly paired . The mothers of lines A and B , and the father of line C were collected in Thep Na Korn . The fathers of lines A , B and D were collected in Mae Na Ree . The mothers of lines C and D were collected in Nhong Ping Kai . They were maintained in the laboratory by mass sib-mating and collective oviposition at each subsequent generation . Quantification of genetic variation within and between the four isofemale lines was conducted as part of this study ( see below ) . To initiate the experiment , eggs were hatched in filtered tap water . Larvae were reared in 24×34×9 cm plastic trays at a density of about 200 larvae per tray . Adults were maintained in 30×30×30 cm screened cages under controlled insectary conditions ( 28±1°C , 75±5% relative humidity , 12:12 hour light-dark cycle ) . They were provided with cotton soaked in a 10% ( m/v ) sucrose solution ad libitum and allowed to mate for 6–7 days before the experimental infection . Genetic characterization of the Ae . aegypti isofemale lines used single nucleotide polymorphism ( SNP ) markers identified and genotyped by Restriction-site Associated DNA ( RAD ) sequencing [38] . Ten females from the 16th generation of each isofemale line ( i . e . , from the same generation that was used in the experimental infection ) and 10 females from the 1st generation of an outbred population collected in 2013 in Thep Na Korn , Kamphaeng Phet Province , Thailand ( i . e . , the region where the isofemale lines originated ) were genotyped using RAD sequencing . Mosquito genomic DNA was purified using the procedure developed by Pat Roman's laboratory at the University of Toronto [39] . DNA concentration was measured with Qubit fluorometer and Quant-iT dsDNA Assay kit ( Life Technologies , Paisley , UK ) . A modified version of the original double-digest Restriction-site Associated DNA ( ddRAD ) sequencing protocol [40] was used as previously described [41] with minor additional modifications . Briefly , 350 ng of genomic DNA from each mosquito were digested in a 50-μl reaction containing 50 units each of NlaIII and MluCI restriction enzymes ( New England Biolabs , Herts , UK ) , 1× CutSmart Buffer and water for 3 hours at 37°C , without a heat-kill step . Digestion products were cleaned with 1 . 5× volume of Ampure XP paramagnetic beads ( Beckman Coulter , Brea , CA , USA ) and ligated to the modified Illumina P1 and P2 adapters with overhangs complementary to NlaIII and MluCI cutting sites , respectively . Each mosquito was uniquely labeled with a combination of P1 and P2 barcodes of variable lengths to increase library diversity at 5’ and 3’ ends ( S1 Table ) . Ligation reactions were set up in a 45-μl volume with 2 μl of 4 μM P1 and 12 μM P2 adapters , 1 , 000 units of T4 ligase and 1× T4 buffer ( New England Biolabs ) and were incubated at 16°C overnight . Ligations were heat-inactivated at 65°C for 10 minutes and cooled down to room temperature ( 20–25°C ) in a thermocycler at a rate of 1 . 5°C per 2 minutes . Adapter-ligated DNA fragments from all individuals were pooled and cleaned with 1 . 5× bead solution . Fragments from 350 to 440 base pairs ( bp ) were selected using a Pippin-Prep 2% gel cassette ( Sage Sciences , Beverly , MA , USA ) . Finally , 1 μl of the size-selected DNA was used as a template in a 10-μl PCR reaction with 5 μl of Phusion High Fidelity 2× Master mix ( New England Biolabs ) and 1 μl of 50 μM P1 and P2 primers ( S1 Table ) . To reduce bias due to PCR duplicates , 8 PCR reactions were run in parallel , pooled , and cleaned with a 0 . 8× bead solution to make the final library . At this step , final libraries were quantified by quantitative PCR using the QPCR NGS Library Quantification Kit ( Agilent Technologies , Palo Alto , CA , USA ) . Libraries containing multiplexed DNA fragments from 50 mosquitoes were sequenced on an Illumina NextSeq platform using a NextSeq 500 High Output 300 cycles v2 kit ( Illumina , San Diego , CA , USA ) to obtain 150-bp paired-end reads . An optimized final library concentration of 1 . 1 pM , spiked with 15% PhiX , was loaded onto the flow cell . Raw sequences were deposited in the NCBI Sequence Read Archive under accession number SRP075401 . A previously developed bash script [41] was used to process raw sequencing reads with minor modifications . Briefly , the DDemux program was used for demultiplexing fastq files according to the P1 and P2 barcodes combinations . Sequence quality scores were automatically converted into Sanger format . Sequences were filtered with FASTX-Toolkit . The first 5 bp ( i . e . , the restriction enzyme cutting site ) and last 11 bp of P1 and P2 reads were trimmed . All reads with Phred scores <25 were discarded . P1 and P2 reads were then matched and unpaired reads were sorted as orphans . Paired reads were aligned to the reference Ae . aegypti genome ( AaegL3 , February 2016 ) [42] using Bowtie version 0 . 12 . 7 [43] . Parameters for the ungapped alignment included a maximum of three mismatches allowed in the seed , suppression of alignments if more than one reportable alignment existed , and a “try-hard” option to find valid alignments . Orphans were combined with all unaligned paired reads and single-end alignment was attempted . All aligned Bowtie output files were merged per individual and were imported into the Stacks pipeline . A catalog of RAD loci used for SNP discovery was created using the ref_map . pl pipeline in Stacks version 1 . 37 [44 , 45] . First , sequences aligned to the same genomic location were stacked together and merged to form loci using Pstacks . Only loci with a sequencing depth ≥3X per individual were retained . Cstacks was used to create a catalog of consensus loci , merging alleles together and Sstacks was used to match all identified loci . The Stacks pipeline identified a total of 899 , 892 RAD loci . The “populations” module was used to export markers with a sequencing depth ≥10X that were present in ≥98% of samples . The mosquito phylogenetic analysis was performed with the resulting subset of 2 , 321 SNPs , which belonged to 1 , 319 distinct RAD loci ( 0 . 15% ) . Phylogenetic trees were constructed using a Bayesian Markov Chain Monte Carlo ( MCMC ) algorithm , implemented in the BEAST 1 . 8 . 3 package [46] . Inferences were produced under the coalescent model ( constant size ) , and under the GTR+G ( global time reversible with gamma distribution and no invariable sites ) and the HKY+G ( Hasegawa-Kishino-Yano ) substitution models . Heterozygote positions were considered in calculations by enabling the use of IUPAC code and associated degeneracy within the substitution model . The length of MCMC was set at 3x107 states to obtain Effective Sampling Size ( ESS ) values >200 . Six- to seven-day-old Ae . aegypti females were deprived of water and sucrose for 24h prior to the infectious blood meal . The virus stock was diluted in cell culture medium ( Leibovitz’s L-15 medium + 10% heat-inactivated fetal calf serum + non-essential amino-acids + 0 . 1% penicillin/streptomycin + 1% sodium bicarbonate ) to reach an expected infectious titer of 3×106 focus-forming units ( FFU ) per mL . One volume of virus suspension was mixed with two volumes of freshly drawn rabbit erythrocytes washed in distilled phosphate-buffered saline ( DPBS ) . After gentle mixing , 2 . 5 mL of the infectious blood meal was placed in each of several Hemotek membrane feeders ( Hemotek Ltd , Blackburn , UK ) maintained at 37°C and covered with a piece of desalted porcine intestine as a membrane . Sixty μL of 0 . 5 M ATP were added to each feeder as a phagostimulant . Each isofemale line was allowed to feed during two rounds of 15 min on different feeders to ensure randomization of a potential feeder effect . Actual virus titer in the blood meal was measured by standard focus-forming assay in C6/36 cells [33] . After feeding , mosquitoes were cold anesthetized on ice and fully engorged females were transferred to 1-pint cardboard cups . They were incubated under controlled conditions ( 28±1°C , 75±5% relative humidity , 12:12 hour light-dark cycle ) in a climatic chamber . At 4 , 7 and 14 days post exposure ( dpe ) , the midgut of 8–12 individuals from each isofemale line ( i . e . , biological replicates ) were dissected . Midguts were homogenized individually in 140 μL of DPBS + 560 μL of QIAamp Viral RNA Mini Kit lysis buffer ( Qiagen , Hilden , Germany ) during two rounds of 30 sec at 5 , 000 rpm in a mixer mill ( Precellys 24 , Bertin Technologies , Montigny le Bretonneux , France ) . At 7 and 14 dpe , the legs of midgut-dissected mosquitoes were removed and homogenized as described above . At 14 dpe , the salivary glands of the midgut- and leg-less individuals were harvested and processed as above . Total RNA was extracted from mosquito homogenates using QIAamp Viral RNA Mini Kit ( Qiagen ) and reverse transcribed using Transcriptor High Fidelity cDNA Synthesis Kit ( Roche Applied Science , Penzberg , Germany ) and a specific reverse primer located at the 3’ end of the viral genome ( S1 Table ) . Presence and amount of viral cDNA was assessed by quantitative PCR using the LightCycler DNA Master SyberGreen I kit ( Roche Applied Science ) and custom primer pairs ( S1 Table ) . Absolute quantification used a standard curve generated with serial dilutions of PCR amplicons of known concentration . Selected samples were amplified by 40 cycles of PCR in 10 overlapping amplicons with Q5 High Fidelity DNA polymerase ( New England Biolabs ) and custom primer pairs ( S1 Table ) . PCR products were purified with Agencourt AMPure XP magnetic beads ( Beckman Coulter ) and their concentration was measured by Quant-iT PicoGreen dsDNA fluorometric quantification ( Invitrogen ) . Equal amounts of each amplicon were pooled by sample and brought to a final concentration of 0 . 2 ng/μL . Multiplexed libraries were prepared using Nextera XT DNA Library Preparation Kit ( Illumina ) and single-end sequenced on an Illumina NextSeq 500 platform using a high-output 75 cycles v1 kit ( Illumina ) . Sequencing reads were demultiplexed using bcl2fastq v2 . 15 . 0 ( Illumina ) . Raw sequences were deposited in the NCBI Sequence Read Archive under accession number SRP075335 . After demultiplexing , reads were trimmed to remove Illumina adaptor sequences using Trimmomatic v0 . 33 [47] and amplification primers if matching sequences were found on either the 5’ or 3’ end of the reads using Cutadapt v . 1 . 8 . 3 [48] . Reads shorter than 32 bp were discarded and remaining reads were then mapped to the reference DENV genome sequence using Bowtie2 v2 . 1 . 0 [49] . The alignment file was converted , sorted and indexed using Samtools v0 . 1 . 19 [50] . Coverage and sequencing depth were assessed using bedtools v2 . 17 . 0 [51] . Single nucleotide variants ( SNVs ) and their proportion among all reads were called using LoFreq* v2 . 1 . 1 [52] and their effect at the amino-acid level assessed by SNPdat v . 1 . 0 . 5 [53] . Two sets of SNV markers were used for analyses of genetic diversity and natural selection . The ‘full’ marker set excluded all nucleotide positions in a given sample that had ( i ) a sequencing depth <500X or ( ii ) where potential sequencing or library preparation artifacts [54] were detected . The ‘conservative’ marker set excluded all nucleotide positions in all samples that had ( i ) a sequencing depth <500X or ( ii ) where potential sequencing or library preparation artifacts [54] were detected in a least one sample . The conservative marker set minimized the potential bias owing to the unique mutational profile of each nucleotide position . However , because some of the overlapping fragments covering the viral genome could not be successfully amplified in a few samples ( S1 Fig ) , the conservative marker set failed to cover large fractions of the viral genome ( S2A Fig ) . The full marker set , conversely , minimized the potential bias owing to distinct evolutionary properties of the different regions of the viral genome . Genetic complexity of the viral population was estimated using normalized Shannon entropy ( Sn ) for each nucleotide site [55]: Sn = - ( p ln ( p ) ) + ( ( 1-p ) ×ln ( 1-p ) ) ln ( 4 ) where p is the SNV minor allele frequency at the considered position , and ln ( 4 ) corresponds to maximum complexity ( i . e . , four possible nucleotides at each position ) . For individual SNVs , Sn values range from 0 to 1 . For diallelic SNVs , Sn values range from 0 ( no diversity ) to 0 . 5 ( maximum complexity , when the two alternative nucleotides are present at equal frequency ) . For each sample , Sn was averaged over all nucleotide sites included in either the full or the conservative set of SNV markers ( i . e . , total genome length minus number of excluded positions ) . Genetic diversity of the viral population was also estimated using nucleotide diversity at each nucleotide site [56]: π = DD-1 × ( 1- ( p2+ ( p-1 ) 2 ) where D is the sequencing depth at the considered position and p is the SNV minor allele frequency . Like for Sn , π values for a diallelic SNV range from 0 ( no polymorphism ) to 0 . 5 ( when the two alternative nucleotides are present at equal frequency ) . For each sample , π was averaged over all nucleotide sites included in either the full or the conservative set of markers . This index of genetic diversity is less sensitive to low-frequency variants than Sn , due to the lack of log-transformation of the frequencies . The magnitude and direction of natural selection were assessed using the dN/dS ratio , which is the ratio between the number of non-synonymous substitutions per non-synonymous site ( dN ) over the number of synonymous substitutions per synonymous site ( dS ) of a coding sequence , assuming synonymous substitutions are selectively neutral: dS = -3 ×ln ( 1-4 ×SdSs3 ) 4 and dN = -3 ×ln ( 1-4 ×NdNs3 ) 4 where Sd is the number of synonymous substitutions in the sequence , Ss is the number of synonymous sites , Nd is the number of non-synonymous substitutions in the sequence and Ns is the number of non-synonymous sites [57] . A dN/dS ratio >1 means that there is an excess of normalized number of non-synonymous substitutions relative to the normalized number of synonymous substitutions and is interpreted as evidence for positive selection ( i . e . , driving change ) . A dN/dS ratio <1 means that there is an excess of normalized number of synonymous substitutions relative to the normalized number of non-synonymous substitutions and is interpreted as evidence for negative selection ( i . e . , acting against change ) . A dN/dS ratio equal to 1 is interpreted as evidence for the absence of natural selection ( i . e , neutral evolution ) . The dN/dS ratio was computed using the Nei-Gojobori method [57] with suggested modifications for high-throughput sequencing data [58] . Briefly , Nd and Sd were calculated for each sample as the sum of SNV frequencies . Mean Nd and Sd were computed for each isofemale line at each time point and used for dN and dS calculation , respectively . Numbers of synonymous and non-synonymous sites from the initial population consensus sequence were estimated using MEGA v . 7 . 0 . 16 [59] by computing the number of 0- , 2- , 3- and 4-fold degenerate sites following the Nei-Gojobori method [57] . The full marker set had a variable number of synonymous and non-synonymous sites depending of the number of nucleotide sites retained or excluded for each sample . The conservative marker set had 328 . 67 synonymous and 1 , 481 . 33 non-synonymous sites for all samples . Statistical analyses were performed in the statistical environment R , version 3 . 2 . 0 ( http://www . r-project . org/ ) using the packages car [60] , plyr [61] , ggplot2 [62] , stringr [63] , reshape2 [64] , gridExtra [65] , fitdistrplus [66] and boot [67] . In all analyses , the individual mosquito sample was considered a biological unit of replication . Infection prevalence and cDNA copy numbers were compared among isofemale lines at each time point by pairwise Pearson χ2 tests and pairwise Wilcoxon signed-rank tests , respectively , followed by a Holm correction of p-values for multiple testing . The proportion of SNVs per position , mean Sn and mean π estimates were compared between the input and later time points using pairwise Wilcoxon signed-rank tests and a Holm p-value adjustment . The proportion of SNVs per position , Sn , π and dN/dS estimates in midgut samples were analyzed between 4 and 7 dpe as a function of the combined effects of time point and mosquito genotype using a linear model with an identity link function and a normal error distribution . Model validity was verified with quantile-quantile ( Q-Q ) plots of residuals and by computing Cook’s distance to assess influence of observations . Statistically significant effects ( p<0 . 05 ) of time point , mosquito genotype and their interactions were determined using type-II analysis of variance . Statistically insignificant interactions were removed from the model , subsequently repeating model validation . Statistical testing of pairwise differences between isofemale lines used the linear regression coefficients . Estimated regression coefficients were extracted and their 95% confidence intervals and p-values were calculated based on their standard errors compared to a reference level . Isofemale line A was arbitrarily chosen as the reference level . Following a published method [17] , bottleneck size at initial midgut infection was estimated by analyzing the change in frequency distribution of neutral markers between blood meal ( initial ) and midgut ( final ) samples . Under the assumption of neutrality ( i . e . , absence of natural selection ) , the idealized number of founding genomes ( N ) initiating the midgut infection can be computed as: N = p ( 1-p ) Var ( p′ ) -Var ( p ) where p is the marker allele frequency in the initial population and p′ is the marker allele frequency in the final population [17] . This method considers that changes in the genetic variance between sequential samples result exclusively from genetic drift and therefore requires neutral or quasi-neutral markers . SNVs that were presumably neutral were selected based on the following set of criteria: ( i ) synonymous change at the third codon position , ( ii ) no significant change in mean frequency between sampling time points , ( iii ) SNV detected in ≥80% of the five viral input replicates ( viral stock and blood meal samples ) , and ( iv ) mean frequency >0 . 02 in the input population . Confidence intervals of N estimates were computed using a bootstrapping procedure as described in [17] . Briefly , for each bootstrap all individuals were sampled with replacement to calculate N . This was repeated 1 , 000 times to generate a distribution of N values and derive 95% confidence intervals corresponding to the 2 . 5 and 97 . 5 percentiles of the distribution . The effect of the initial midgut infection bottleneck on viral diversity indices was simulated in R based on 100 sampling events from an initial viral population containing 100 independent SNVs . SNV minor allele frequency was randomly drawn from an exponential distribution ( λ = 100 ) . Initial viral population size ( equivalent to the infectious dose ingested in the blood meal ) was drawn from a normal distribution ( mean = 2 , 000; standard deviation = 200 ) . Bottleneck size was drawn from a normal distribution ( mean = 28; standard deviation = 5 ) . Mean Sn and mean π were computed for all samples in the presence or the absence of a detection threshold arbitrarily set at an SNV minor allele frequency of 0 . 01 .
A genome-wide set of 2 , 321 SNPs generated by RAD sequencing was used to genetically characterize the four Ae . aegypti isofemale lines ( A , B , C , and D ) . These markers had a sequencing depth ≥10X per sample and were missing in <2% of samples . An outbred Ae . aegypti population from the same geographic location where the lines were created was also genotyped to provide a phylogenetic background . Phylogenetic relationships among individuals from the four isofemale lines and the outbred population were determined with a Bayesian method ( Fig 1 ) . As expected , the outbred mosquito population was paraphyletic , reflecting its genetic diversity . Mosquitoes from isofemale lines A and B clustered independently with strong statistical support , confirming their distinct genetic identity . Unexpectedly , mosquitoes from isofemale line C grouped with mosquitoes from isofemale line D within the same clade . This could be the result of relatedness between the parents randomly chosen to initiate the lines , as the mothers of lines C and D came from the same collection site and may have been siblings . Two different substitution models for the phylogenetic reconstruction gave similar clustering patterns . Similar results were also obtained when testing a variable number of markers ( allowing from 0% to 30% of missing genotypes ) with the same method . Because isofemale lines C and D were not unambiguously assigned to different monophyletic groups , they could not be considered as distinct genotypes and were thus combined in all subsequent analyses . They are jointly referred to as line CD hereafter . Mosquitoes from the three different genotypes ( A , B , and CD ) were exposed to DENV through an artificial blood meal at a final titer of 1 . 52×106 focus-forming units ( FFU ) /mL . Assuming a blood meal size of approximately 2 μL , the infectious dose ingested by each mosquito was about 3 , 000 infectious viral particles . The proportion of mosquitoes that acquired a midgut infection ranged from 75 to 100% and did not differ significantly between time points or isofemale lines ( Fig 2A ) . The proportion of mosquitoes with a DENV infection that disseminated to their legs increased from 10–40% at 7 days post exposure ( dpe ) to 60–100% at 14 dpe , but the rate of virus dissemination to the legs did not differ significantly between isofemale lines ( Fig 2A ) . However , the proportion of mosquitoes with a disseminated infection in the salivary glands was significantly higher for line CD ( 87 . 5% ) than for line A ( 37 . 5% ) and line B ( 41 . 7% ) at 14 dpe ( line A vs . line CD , p = 0 . 037; line B vs . line CD , p = 0 . 037 ) . Among infected mosquitoes , viral load ranged from 8 . 9×102 to 2 . 8×106 DENV genome copies per sample with no significant difference between isofemale lines at any of the time points , with the exception of lines B and CD that had significantly different viral loads ( p = 0 . 037 ) in their salivary glands at 14 dpe ( Fig 2B ) . Deep sequencing of viral genomes was performed for a subset of 78 infected samples at selected time points ( Fig 2B ) that were processed individually and treated as biological replicates . Some samples were excluded because their low concentration of viral RNA resulted in unsuccessful RT-PCR amplification . A total of 4 , 7 and 13 infected midguts at 4 dpe and 7 , 11 and 21 infected midguts at 7 dpe were analyzed for lines A , B , and CD , respectively . Ten infected salivary glands at 14 dpe were analyzed in line CD . In addition , DENV genomes were deep sequenced in the initial viral stock and in four replicates of the infectious blood meal . On average , 3 , 615 , 466 sequencing reads per sample aligned to the reference DENV genome . Mean DENV genome coverage with a sequencing depth >500X was 10 , 594 nucleotides per sample , which represents 98 . 8% of the 10 , 718 nucleotides of the total genome length . Mean sequencing depth was 24 , 212X per sample ( S1 Fig ) . The full set of SNV markers retained for population genetic analyses included an average of 5 , 843 nucleotide sites across the DENV genome , whereas a more conservative set ( see Materials and Methods ) was restricted to 1 , 810 nucleotides ( S2A Fig ) . SNVs of the full marker set were randomly distributed across the genome without obvious mutation hot or cold spot ( S3 Fig ) . A new variant reached consensus level ( frequency >0 . 5 ) in one midgut sample at 4 dpe and one midgut sample at 7 dpe , but the SNV was different in each case . In salivary glands collected at 14 dpe , new variants reached consensus level at 11 positions , none of which was shared among individuals within or between isofemale lines ( S3 Fig ) . In the more restricted conservative set of markers , no variant reached consensus level at any time point ( S2B Fig ) . To determine the effect of initial midgut infection on DENV genetic diversity , a first series of analyses compared viral genetic diversity observed in the input samples with genetic diversity observed at any of the later time points . In the full marker set , initial infection of the midgut was associated with an increase in viral genetic diversity relative to the input ( Fig 3 ) both when measured with normalized Shannon entropy Sn ( 0 dpe vs . 4 dpe , p = 0 . 0001; 0 dpe vs . 7 dpe , p = 0 . 002; 0 dpe vs . 14 dpe , p = 0 . 003 ) ( Fig 3A ) and when measured with nucleotide diversity π ( 0 dpe vs . 4 dpe , p = 0 . 0001; 0 dpe vs . 7 dpe , p = 0 . 0004; 0 dpe vs . 14 dpe , p = 0 . 003 ) ( Fig 3B ) . Viral diversity was also significantly higher in the salivary glands at 14 dpe than in the midgut at 7 dpe ( Sn: p = 0 . 012; π: p = 0 . 012 ) . The proportion of variable sites detected also increased following initial midgut infection ( S4 Fig ) although differences were only statistically significant between 0 dpe and 4 dpe ( p = 0 . 0029 ) and between 0 dpe and 14 dpe ( p = 0 . 0067 ) . Similarly , in the conservative set of markers , mosquito infection was associated with a relative increase in viral genetic diversity following initial midgut infection , albeit more modestly due to the smaller number of markers , both when measured with normalized Shannon entropy Sn ( 0 dpe vs . 4 dpe , p = 0 . 046; 0 dpe vs . 14 dpe , p = 0 . 008 ) ( S5B Fig ) and when measured with nucleotide diversity π ( 0 dpe vs . 14 dpe , p = 0 . 008 ) ( S5C Fig ) . The proportion of variable sites detected , however , did not differ statistically between time points ( S5A Fig ) . To evaluate the dynamics of DENV genetic diversity during viral population expansion in the midgut , a second series of analyses compared viral genetic diversity between 4 and 7 dpe , accounting for potential differences between mosquito genotypes . In the full set of markers , both the time point ( proportion of variable sites: p = 0 . 03; Sn: p = 0 . 04; π: p = 0 . 04 ) and the isofemale line ( proportion of variable sites: p = 0 . 0035; Sn: p = 0 . 0002; π: p = 0 . 0005 ) significantly influenced viral genetic diversity . Overall , DENV genetic diversity slightly decreased between 4 dpe and 7 dpe . Isofemale line A displayed significantly higher viral genetic diversity than lines B and CD , for all three indices: proportion of variable sites ( p = 0 . 012 and p = 0 . 0008 , respectively ) , Sn ( p = 0 . 006 and p = 0 . 0005 respectively ) and π ( p = 0 . 017 and p = 0 . 0001 , respectively ) . Similar results were obtained with the conservative set of markers . Both the time point ( proportion of variable sites: p = 0 . 01; Sn: p = 0 . 013; π: p = 0 . 015 ) and the isofemale line ( proportion of variable sites: p = 0 . 0011; Sn: p = 0 . 0006; π: p = 0 . 0029 ) significantly influenced viral genetic diversity . Overall , DENV genetic diversity slightly decreased between 4 dpe and 7 dpe . Isofemale line A displayed significantly higher viral genetic diversity than lines B and CD , for all three indices: proportion of variable sites ( p = 0 . 017 and p = 0 . 0002 , respectively ) , Sn ( p = 0 . 0047 and p = 0 . 0001 , respectively ) and π ( p = 0 . 012 and p = 0 . 0007 , respectively ) . Based on the full set of SNV markers , dN/dS ratios were predominantly negative indicating purifying selection ( Fig 3C ) . There was no statistically significant difference in dN/dS ratios between time points or mosquito isofemale lines . Computing dN/dS ratios was not possible with the conservative set of markers because the smaller number of SNVs resulted in a large proportion of samples with dS = 0 . Analysis of dN/dS ratios calculated per isofemale line , however , provided results consistent with predominantly purifying selection using the conservative set of markers . Average dN/dS ratios were remarkably similar among lines and time points around 0 . 2218 ( S1 Table ) . Three SNVs that complied with criteria of quasi-neutral evolution were selected to estimate the idealized number of founding viral genomes ( N ) initiating the midgut infection based on changes in the variance of their frequency between input and midgut samples ( Table 1 ) . Based on the three markers , initial midgut infection was founded by 23–34 genomes when estimated at 4 dpe ( Fig 4A ) and 5–42 genomes when estimated at 7 dpe ( Fig 4B ) . Collectively , 95% confidence intervals ranged from 2 to 161 founding genomes . N estimates and their confidence intervals were consistent across time points , especially for marker at position 1556 . For this marker , 4 dpe and 7 dpe data were pooled to compute isofemale line-specific N estimates . There were no statistically significant differences among lines in the estimated bottleneck size ( Fig 4C ) , ranging from 83 ( 95% confidence interval: 52–396 ) founding genomes for line A , to 23 ( 9–220 ) for line B and 33 ( 16–108 ) for line CD . Simulations were performed to model the effect of population bottlenecks on DENV intra-host genetic diversity . The simulation randomly assigned SNV minor allele frequency , initial viral population size and bottleneck size to explore whether a minimum threshold for SNV detection would alter the observed genetic diversity following a population bottleneck compared to the true genetic diversity . When 100 SNVs were present in the input viral population and no minimum detection threshold was set , mean Sn and π estimated in 100 replicate samples decreased following the bottleneck ( Fig 5A ) . However , when only SNVs with a minor allele frequency >1% were detected , mean Sn and π estimates increased after the bottleneck ( Fig 5B ) .
We investigated the evolutionary forces acting on DENV populations within their arthropod vector . Specifically , we evaluated the relative effects of natural selection and genetic drift on DENV intra-host evolution in the midgut of Ae . aegypti . In addition , we assessed the influence of vector genetic heterogeneity on intra-host viral genetic diversity . Our results show that DENV intra-host genetic diversity in Ae . aegypti is shaped by the combined effects of genetic drift , purifying selection and vector genotype . Reshuffling of the variant repertoire during initial infection of the midgut was associated with a bottleneck size ranging from 5 to 42 founding viral genomes , irrespective of the mosquito genotype . DENV genetic diversity increased significantly following initial infection , but was restricted by strong purifying selection during DENV population expansion in the midgut . Observed levels of DENV genetic diversity in the midgut differed significantly between mosquito isofemale lines despite a similar bottleneck size at initial infection . Arboviruses typically maintain high levels of genetic diversity during transmission by their arthropod vectors despite anatomical barriers that often result in severe population drops [11] . Such population bottlenecks have been documented for several arboviruses in their vectors using artificial mutant swarms [68] , marked viral clones [13] , viral replicons [69] or stochastic simulations based on observed changes in variant frequencies [15] . Although the overall level of arboviral genetic diversity is usually maintained during vector infection [14] , the viral variant repertoire can be significantly altered [15 , 68 , 70] . Presumably , viral genetic diversity is quickly replenished by mutation and demographic expansion following population bottlenecks [11] . However , whether changes in the viral variant repertoire are due to stochastic sampling ( i . e . , genetic drift ) , purifying selection ( i . e . , removal of variants with lower fitness ) , or vector genetic heterogeneity combined with specific interactions between vector and virus genotypes [33–37] has remained largely unresolved . Our analysis used neutral or quasi-neutral genetic markers to estimate the effective DENV population size during initial infection of the Ae . aegypti midgut . This approach rules out natural selection and only measures the effect of genetic drift due to random sampling . It is worth noting , however , that true neutral mutation may not exist because even synonymous mutations can have a fitness effect , especially in RNA viruses [71] . Deviation from our assumption of neutrality or quasi-neutrality of the chosen markers may have overestimated the bottleneck size . Indeed , both positive selection and negative selection would likely act to decrease the variance of marker frequency and therefore result in a larger estimate of Ne with our method . Therefore , our conclusion that DENV populations undergo a strong population bottleneck during initial midgut infection should be robust to any undetected departure from neutrality . Moreover , we chose markers whose average frequency was similar before and after the bottleneck , supporting the assumption that they were not under directional selection . Our estimation that initial midgut infection is founded by only a few tens of DENV genomes is consistent with previous estimations for DENV based on stochastic simulations using empirical data [15] . We went one step further by demonstrating that genetic drift , rather than natural selection , is the main evolutionary force underlying this population bottleneck . Although our estimated bottleneck size is larger than for other RNA viruses during host-to-host transmission [72] , it is expected to substantially reduce the genetic breadth of the viral quasispecies [73] . This finding has important implications for DENV evolution in general . A small effective population size means that natural selection will be relatively inefficient during human-to-mosquito transmission . It will prevent adaptive evolution especially if beneficial SNVs are present at low frequencies in the mutant swarm transmitted from the human host [74] . On the other hand , the population bottleneck associated with initial midgut infection may not be small enough to prevent the long-term maintenance of defective viral genomes through complementation by co-infection of host cells with functional viruses . Such a phenomenon was previously documented in the case of a stop-codon mutation that became widespread in DENV populations sampled in Myanmar in 2001 [75] . The frequency of the stop-codon mutation was likely high enough to overcome the effect of population bottlenecks during multiple host-to-host transmission events . During DENV population expansion following initial midgut infection , natural selection was predominantly negative ( i . e . , acting against change ) . Accordingly , the consensus sequence remained unchanged in most of the midgut samples . Only in the salivary glands did several SNVs reach consensus level ( frequency >0 . 5 ) , but with no evidence of evolutionary convergence . As was observed for West Nile virus [68] , DENV intra-host genetic diversity in midguts slightly decreased between 4 and 7 dpe . Importantly , we found that overall levels of DENV intra-host genetic diversity differed significantly between distinct mosquito genetic backgrounds . Both the initial bottleneck size and the census size of the viral population did not significantly vary among mosquito genotypes , and thus are unlikely to explain this difference . The mechanistic basis of this finding remains to be determined , but we speculate that viral populations may undergo different selective constraints in different mosquito genotypes . Mosquito genotypes could vary in the intensity of purifying selection ( i . e . , variation in the efficiency of mechanisms that remove deleterious de novo mutations ) , but this was not supported by our data . Likewise , the overall lack of positive selection that we observed indicates that it is unlikely to be the underlying mechanism . Alternatively , mosquito genotypes may differ in the level of balancing selection ( i . e . , mechanisms that act to promote genetic diversity such as negative frequency-dependent selection or spatiotemporal fluctuations in the strength and direction of selection ) . The antiviral RNA interference ( RNAi ) pathway of mosquitoes was suggested to play a role in viral genetic ‘diversification’ [76 , 77] , by promoting escape to complementary base-pairing required for RNAi-mediated cleavage . Variation in host factors could also result in differences in viral intra-host genetic diversity through subtle changes in viral RNA-dependent RNA polymerase fidelity [78] . Mutation rates of RNA viruses are not only determined by virus-encoded factors , by also by host-dependent processes . Replicase fidelity of a plant RNA virus was found to differ according to the host type [79] . Replication fidelity in retroviruses can be affected by intracellular dNTP imbalance [80 , 81] . Viral mutation rate can also be influenced by the expression of host genes , such as cellular deaminases that promote hypermutation in RNA viruses [82–84] . Interestingly , the isofemale line that displayed the lowest level of DENV genetic diversity in the midgut ( i . e . , line CD ) was also associated with the highest prevalence and highest viral load in salivary glands . Because we did not examine viral populations in saliva samples , whether this translates in differences of virus transmission potential is unknown . It is tempting to speculate that the vector competence phenotype relates to the level of viral genetic diversity . Unfortunately , we could not compare DENV intra-host diversity in salivary glands between mosquito isofemale lines because DENV amplification was unsuccessful in two out of three lines due to low template concentration . A recent study found differences in the intra-host genetic diversity of West Nile virus among different species of Culex mosquitoes [85] . Here , we provided evidence that such differences exist at the intra-specific level . The potential relationship between viral intra-host genetic diversity and vector competence variation among mosquito genotypes deserves further investigation . It will be interesting to determine in future experiments whether the effect of the vector genotype varies according to the mosquito generation , the virus strain , and/or the specific combinations of mosquito lines and virus strains . Finally , we introduced a non-exclusive , alternative scenario to the ‘diversification’ hypothesis that may contribute to explain why the level of arboviral genetic diversity increases despite a population bottleneck . Our proposed scenario is based on the counter-intuitive idea that a strong initial population bottleneck may actually result in a higher observed level of genetic diversity if low-frequency SNVs go undetected for methodological reasons . We used a model based on stochastic simulations to illustrate the effect of a minimum detection threshold of low-frequency SNVs on observed genetic diversity . When all SNVs present were detected regardless of their frequency ( i . e . , no detection threshold ) , mean viral genetic diversity indices decreased following a simulated population bottleneck . Conversely , mean genetic diversity indices increased when only SNVs present at a frequency >1% were successfully detected . In our empirical data , it was not possible to ascertain whether SNVs newly detected after the initial population bottleneck resulted from de novo mutations or were already present prior to the bottleneck at frequencies lower than the detection threshold . However , our model indicated that a change in the SNV frequency spectrum following the population bottleneck combined with a minimum detection threshold is a potential explanation to the observed increased genetic diversity following the initial bottleneck . Taken together , our results show that DENV intra-host genetic diversity in the mosquito vector is shaped by stochastic events during initial midgut infection due to a sharp reduction in population size , followed by predominantly purifying selection during population expansion and diversification in the midgut . Differential diversification between mosquito isofemale lines indicates a genetic foundation , but the lack of convergent SNVs does not support the existence of mosquito genotype-specific directional selection . We conclude that the evolution of DENV intra-host genetic diversity in mosquitoes is not only driven by genetic drift and purifying selection , but is also modulated by vector genetic factors . Characterizing the evolutionary forces that govern arboviral genetic diversity contributes to understanding their unique biology and adaptive potential .
|
During infection of their arthropod vectors , arthropod-borne viruses ( arboviruses ) such as dengue viruses traverse several anatomical barriers that are believed to cause dramatic reductions in population size . Such population bottlenecks challenge the maintenance of viral genetic diversity , which is considered critical for fitness and adaptability of arboviruses . Anatomical barriers in the vector were previously associated with both maintenance of arboviral genetic diversity and alteration of the variant repertoire . However , the relative role of random processes and natural selection , and the influence of vector genetic heterogeneity have not been elucidated . In this study , we used high-throughput sequencing to monitor dengue virus genetic diversity during infection of several genetic backgrounds of their mosquito vector . Our results show that initial infection of the vector is randomly founded by only a few tens of individual virus genomes . The overall level of viral genetic diversity generated during infection was predominantly under purifying selection but differed significantly between mosquito genetic backgrounds . Thus , in addition to random evolutionary forces and the purging of deleterious mutations that shape dengue virus genetic diversity during vector infection , our results also point to a novel role for vector genetic factors in the genetic breadth of virus populations .
|
[
"Abstract",
"Introduction",
"Material",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"ecology",
"and",
"environmental",
"sciences",
"viral",
"transmission",
"and",
"infection",
"population",
"genetics",
"conservation",
"genetics",
"microbiology",
"animals",
"viral",
"vectors",
"population",
"biology",
"insect",
"vectors",
"microbial",
"genetics",
"conservation",
"biology",
"infectious",
"diseases",
"epidemiology",
"arboviral",
"infections",
"disease",
"vectors",
"insects",
"genetic",
"drift",
"arthropoda",
"conservation",
"science",
"mosquitoes",
"natural",
"selection",
"virology",
"genetics",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"evolutionary",
"biology",
"evolutionary",
"processes",
"organisms"
] |
2016
|
Genetic Drift, Purifying Selection and Vector Genotype Shape Dengue Virus Intra-host Genetic Diversity in Mosquitoes
|
The Caenorhabditis elegans dauer larva is a facultative state of diapause . Mutations affecting dauer signal transduction and morphogenesis have been reported . Of these , most that result in constitutive formation of dauer larvae are temperature-sensitive ( ts ) . The daf-31 mutant was isolated in genetic screens looking for novel and underrepresented classes of mutants that form dauer and dauer-like larvae non-conditionally . Dauer-like larvae are arrested in development and have some , but not all , of the normal dauer characteristics . We show here that daf-31 mutants form dauer-like larvae under starvation conditions but are sensitive to SDS treatment . Moreover , metabolism is shifted to fat accumulation in daf-31 mutants . We cloned the daf-31 gene and it encodes an ortholog of the arrest-defective-1 protein ( ARD1 ) that is the catalytic subunit of the major N alpha-acetyltransferase ( NatA ) . A daf-31 promoter::GFP reporter gene indicates daf-31 is expressed in multiple tissues including neurons , pharynx , intestine and hypodermal cells . Interestingly , overexpression of daf-31 enhances the longevity phenotype of daf-2 mutants , which is dependent on the forkhead transcription factor ( FOXO ) DAF-16 . We demonstrate that overexpression of daf-31 stimulates the transcriptional activity of DAF-16 without influencing its subcellular localization . These data reveal an essential role of NatA in controlling C . elegans life history and also a novel interaction between ARD1 and FOXO transcription factors , which may contribute to understanding the function of ARD1 in mammals .
Animal development is a complex process that involves hierarchical gene regulatory networks and is influenced by environmental conditions . When food is abundant , the post-embryonic development of C . elegans consists of four larval stages ( L1–L4 ) and the adult . During the L1 stage , environmental factors determine whether C . elegans molts to an L2 larva or a pre-dauer L2d larva [1] . At least three environmental cues have been defined: food supply , temperature , and a constitutively secreted dauer-inducing pheromone that signals population density [2] . The L2 larva is developmentally committed to continued growth , whereas the L2d larva can molt to a dauer larva if food is scarce and the animals are overcrowded , or to an L3 larva should conditions improve . Mutations affecting dauer larval development include dauer-defective ( daf-d ) mutations that prevent entry into the dauer stage , and dauer-constitutive ( daf-c ) mutations that mandate entry into the dauer stage [2] . Based on epistatic relationships between daf-c and daf-d mutations , more than twenty genes controlling dauer formation have been ordered in a genetic pathway [2] representing generation of the pheromone signal [3] , response by chemosensory neurons [4] , [5] and transduction of the signal to other cells . Three functionally overlapping neural pathways control the developmental response to environmental cues . They involve DAF-7/TGF-ß [6] , [7] , DAF-11/cyclic GMP [8] , and DAF-2/insulin-like [9] , [10] pathways , which relay the environmental signals to a nuclear hormone receptor , DAF-12 [11] , to control dauer versus non-dauer morphogenesis . Mutations in two genes , daf-9 and daf-15 , lead to non-conditional formation of detergent-sensitive dauer-like larvae [12] . These mutants form dauer larvae constitutively and display some characteristics of dauer larvae formed under starvation , such as a high density of intestinal and hypodermal storage granules . daf-9 encodes a cytochrome P450 related to those involved in the biosynthesis of steroid hormones in mammals [13] , [14]; it was found to specify a step in the biosynthetic pathway for a DAF-12 steroid ligand called dafachronic acid [15]–[17] . daf-15 encodes the C . elegans ortholog of Raptor [18] that is proposed to interact with C . elegans target-of-rapamycin kinase ( LET-363/CeTOR ) to control C . elegans larval development [18] . Both daf-9 and daf-15 also regulate fat metabolism and adult lifespan [13] , [14] , [18] . The dauer-like mutants represent a mutant class distinct from the previously defined daf-c and daf-d mutants . Unlike most daf-c mutants , the dauer-like mutants are not ts , and they do not complete dauer morphogenesis . The daf-d genes such as daf-12 have non-conditional alleles and fail to respond to pheromone [1] , but unlike the dauer-like mutants they can execute non-dauer development . The dauer-like mutants define a third class of mutants , one in which the animals are incapable of executing either complete dauer or non-dauer development . The daf-31 mutant was isolated in genetic screens to identify genes similar to daf-9 and daf-15 [17] . The overall aim of the present study was to clone the daf-31 gene and characterize the DAF-31 function . Our genetic epistasis analysis suggests daf-31 functions downstream of or in parallel to daf-3 , daf-12 and daf-16 dauer-defective genes , and acts upstream of or in parallel to daf-15/raptor . We cloned the daf-31 gene by positional cloning and showed that it encodes an ortholog of arrest-defective-1 protein ( ARD1 ) , the catalytic subunit of the major N alpha-acetyltransferase ( NatA ) . Moreover , our data reveal that daf-31 has an essential role in controlling C . elegans larval development , metabolism and adult longevity .
Entry into the dauer stage is determined by the pheromone/food ratio , with high pheromone and low food supply favoring dauer formation [2] . Dauer larva is considered as an alternative L3 larval stage . Compared to L3 larva ( Figure 1A ) , the dauer larva has a constricted pharynx ( Figure 1B ) and a special cuticle with dauer alae ( Figure 1C ) . In the presence of dauer-inducing pheromones , daf-31 mutants cannot form SDS-resistant dauer larvae [17] . In order to determine whether daf-31 mutants enter the dauer stage in response to starvation , we examined the progeny of strain unc-24daf-31/nT1 under starvation conditions and observed uncoordinated ( Unc ) dauer larvae . These dauer larvae showed normal dauer features , such as a dark body , fully constricted pharynx ( Figure 1D ) , and a cuticle with dauer alae ( Figure 1E ) . However , daf-31 dauer larvae were not SDS-resistant like normal dauer larvae . Furthermore , daf-31 dauer larvae could not resume development when food was provided , dying shortly thereafter . Therefore , daf-31 mutants could not complete dauer morphogenesis under starved conditions , and those incomplete dauer larvae could not finish reproductive development after food was provided . Fat accumulation is one characteristic of C . elegans dauer larvae . We examined fat accumulation in daf-31 homozygous mutants using Sudan Black B staining . As shown in Figure 1F , daf-2 mutant dauer larvae accumulate fat as described previously [9] . The daf-31 mutant worms also accumulate more fat droplets than wild-type worms and fat droplets in the daf-31 mutants are larger than those in wild-type worms ( Figure 1G and 1H ) . To confirm this phenotype , Nile red was used to stain fixed worms; this approach has been reported to reliably detect fat droplets in C . elegans [19] . Similar to Sudan Black staining , Nile red also detected fat accumulation in daf-31 mutant worms ( Figure S1 ) . Therefore , daf-31 mutants shift metabolism to fat accumulation . To position daf-31 in the dauer formation pathway , we examined the epistatic relationship between daf-31 and daf-d genes including daf-3 , daf-12 and daf-16 . The daf-31 mutation is epistatic to all three daf-d mutations as judged by the ratio of progeny ( 1∶2∶1 ) ( Table 1 ) . For the epistasis analysis with daf-16 , the ratio of progeny is 1∶2 because nT1 homozygous animals are lethal . We repeated the epistasis analysis of daf-31 and daf-12 by using the daf-12 ( rh61rh411 ) null allele [11] and obtained a similar result ( Table 1 ) . These epistatic relationships suggest that daf-31 functions downstream of or in parallel to daf-3 , daf-12 and daf-16 in dauer formation . To examine the epistatic relationship of daf-31 and daf-15 , we injected dsRNA of daf-15 into unc-24daf-31/nT1 young adult worms . Wild-type ( N2 ) animals were treated equally and used as controls . We examined the phenotype of progeny reproduced at various time periods after injection . The progeny reproduced between seven and eighteen hours after injection arrested development at a dauer-like stage three days after egg lay ( Table 2 ) . These dauer-like animals have a similar phenotype to daf-15 mutants . Thus , regarding dauer entry , it appears that the daf-15 mutation is epistatic to the daf-31 mutation . We scored the recovery of both N2 and unc-24daf-31 dauer-like animals two days after dauer-like arrest . 48% of N2 dauer-like worms remained at dauer-like stage and the rest of the animals recovered and grew to L4 larval or adult size ( Table 2 ) . By contrast , 100% of unc-24daf-31 animals stayed at dauer-like stage ( Table 2 ) . For unc-24daf-31 dauer-like larvae without daf-15 RNAi treatment , the majority of these animals died within five days . However , surviving animals all grew to L4 larval or adult size ( n = 52 ) . Taken together , these data indicate that daf-15 is epistatic to daf-31 as unc-24daf-31 mutants treated by daf-15 RNAi form dauer-like larvae similar to daf-15 mutants . Moreover , these two mutants have a synergistic effect on C . elegans development because no unc-24daf-31 dauer-like larvae treated by daf-15 RNAi recovered . Thus , these two genes may function in the same pathway and daf-31 is upstream of daf-15 . However , the possibility that these two genes act in parallel cannot be excluded . A positional cloning strategy was used to identify the daf-31 gene on chromosome IV between unc-24 and fem-3 ( Figure S2A ) . daf-31 was found to lie between the physical SNP markers T09A12 and F17E9 ( Figure S2A ) . A genomic fragment corresponding to the K07H8 . 3 open reading frame fully rescued the genetic daf-31 null mutant [daf-31 ( m655 ) ] phenotype , i . e . the transgenic animals did not form dauer-like larvae , but grew to fertile adults . Sequence analysis of the daf-31 gene in the mutant revealed a 393 bp deletion which removed 151 bp of promoter upstream of the ATG start codon and 242 bp of daf-31 coding region downstream of the ATG start codon , which may completely block daf-31 transcription as both the essential promoter region and the N-terminal portion of the gene were deleted ( Figure S2B ) . Primers were designed to flank the deletion region and PCR analysis of mutant worms' genomic DNA detected a 1 , 449 bp band ( 393 bp smaller than the wild-type band ) in both homozygous and heterozygous daf-31 ( m655 ) mutant worms ( Figure S2C ) . The daf-31 gene encodes the ortholog of ARD1 with a predicted molecular weight of 21 . 2 kDa . ARD1 is the catalytic subunit of NatA that catalyzes the acetylation of proteins beginning with Met-Ser , Met-Gly and Met-Ala [20] . Amino acid sequence alignment showed that DAF-31 shares 75% identity with human ARD1 , 77% identity with mouse ARD1 , 72% identity with Drosophila melanogaster ARD1 and 46% identity with yeast ARD1 ( Figure S2D ) . Given that there is only a single mutant allele of daf-31 , we used RNAi to inhibit daf-31 in the N2 background to confirm the daf-31 mutant phenotype . Inhibition of daf-31 by feeding animals with E . coli that express daf-31 dsRNA did not induce a dauer-like phenotype . From our previous work , dsRNA injection can create a stronger mutant phenotype similar to that of a genetic null mutant [18] . In vitro synthesized daf-31 dsRNA was injected into gonads of N2 young adult worms and the progeny displayed a dauer-like phenotype similar to the daf-31 ( m655 ) mutants . The starvation-induced dauer morphology of daf-31 ( m655 ) mutants , such as dauer alae formation and contrasted pharynx ( described in Figure 1 ) could not be examined using this RNAi method . Therefore , we examined fat accumulation in daf-31 RNAi-treated animals . Similar to the daf-31 ( m655 ) mutant , daf-31 RNAi-treated animals accumulated fat as detected by both Sudan Black and Nile red staining of fixed animals ( Figure S3 ) . Based on these results , we conclude that the daf-31 mutant phenotypes described in this study most likely resulted from daf-31 mutation instead of secondary mutations in the background . In order to characterize the daf-31 expression pattern , we constructed a daf-31 promoter::gfp reporter construct . In N2 animals , GFP expression was detected from L1 to the adult stages in multiple tissues including the hypodermis , pharynx , intestine , and neurons ( Figure 2 ) . To confirm the GFP expression pattern of daf-31 promoter fusion reporter , we constructed daf-31 translational fusion reporter genes in which the GFP open reading frame was fused to the full-length daf-31 genomic DNA in frame either at the N-terminus or at the C-terminus . Both translation fusion reporter genes fully rescued the dauer-like phenotypes of daf-31 mutants . However , we did not observe GFP expression in the rescued daf-31 mutant worms , a phenomenon previously reported with other GFP fusion gene mutant rescues [21] . Thus , our observations of daf-31 expression pattern were limited to the daf-31 promoter fusion , which may not represent the endogenous expression pattern of the entire daf-31 gene if enhancer elements are present in introns or in 3′ untranslated sequences . Increased adult longevity is a phenotype associated with many dauer mutants . Since daf-31 homozygous mutants arrest development at L4 stage , we inhibited the daf-31 gene by feeding RNAi . The RNAi treatment successfully reduced daf-31 mRNA level as measured by qRT-PCR ( Figure S4 ) . However , daf-31 RNAi treatment had no obvious effect on the lifespan of wild-type worms as the daf-31 RNAi-treated worms had similar mean and maximum lifespans as control vector RNAi-treated worms ( Figure 3A and Table S1 ) . When RNAi-sensitive rrf-3 ( pk1426 ) mutants were treated with daf-31 RNAi , their lifespans were significantly decreased ( p<0 . 0001 , log-rank test ) ( Figure 3B and Table S1 ) . Compared to controls , the mean lifespan of rrf-3 mutants treated with daf-31 RNAi was shortened by four days ( Figure 3B and Table S1 ) . To test if daf-31 mutations influence the longevity phenotype of daf-2 mutants , we treated the rrf-3 ( pk1426 ) ;daf-2 ( e1370 ) mutant with daf-31 RNAi . The mean lifespan of daf-31 RNAi-treated rrf-3;daf-2 mutants was five days shorter compared to that of control animals ( p = 0 . 0005 , log-rank test ) ( Figure 3C and Table S1 ) . Thus , inhibition of daf-31 partially suppressed the longevity phenotype of daf-2 mutants . Based on this result , we postulate that overexpression of daf-31 may further increase the lifespan of daf-2 mutants . To test this , daf-2 ( e1370 ) and daf-16 ( mgDf47 ) ;daf-2 ( e1370 ) mutants overexpressing daf-31 were constructed . The overexpression of daf-31 was confirmed by qRT-PCR ( Figure S4 ) . As shown in Figure 3D , daf-31 overexpression increased the lifespan of daf-2 mutant worms ( p<0 . 0001 , log-rank test ) ( Figure 3D and Table S1 ) . The mean lifespan was increased by eight days and the maximum lifespan was extended by seven days ( Figure 3D and Table S1 ) . This increased lifespan was due to daf-31 overexpression as daf-31 RNAi treatment completely abrogated it ( Figure 3E and Table S1 ) . Interestingly , daf-31 overexpression failed to extend the lifespan of daf-16;daf-2 double mutants ( Figure 3F and Table S1 ) , indicating that DAF-16 is required for daf-31 overexpression to enhance the daf-2 longevity phenotype . We also measured the lifespan of N2 animals overexpressing the daf-31 gene . As shown in Figure S5 and Table S1 , daf-31 overexpression did not extend the lifespan of N2 worms . In fact , it slightly decreased the lifespan of N2 worms . Finally , to confirm daf-31 functions through daf-16 in C . elegans lifespan regulation , we tested if daf-31 RNAi can further decrease the lifespan of RNAi-sensitive daf-16 mutants ( daf-16;rrf-3 ) . We found daf-31 RNAi had no obvious effect on the lifespan of daf-16;rrf-3 mutants ( Figure S6 and Table S1 ) . The forkhead transcription factor DAF-16/FOXO controls the transcription of an array of genes essential for lifespan extension and oxidative stress resistance including the antioxidant enzyme superoxide dismutase ( sod ) -3 gene and beta-carotene 15 , 15′-monooxygenase gene ( bcmo-2 ) [22] . We used qRT-PCR to measure the expression level of sod-3 and bcmo-2 in daf-31 overexpression strains and control animals . As shown in Figure 4A , the expression level of sod-3 was significantly increased when daf-31 was overexpressed in the N2 background and in daf-2 mutants . Similar to sod-3 , the expression of bcmo-2 is also significantly increased when daf-31 is overexpressed in the daf-2 mutant background ( Figure 4B ) . Taken together , these data indicate overexpression of daf-31 stimulates the transcriptional activity of DAF-16 . Reduction of daf-2 insulin-like signaling activity increases C . elegans lifespan by promoting nuclear localization of DAF-16 [23] , [24] . We crossed the DAF-16::GFP reporter gene into daf-31 overexpressing animals to examine if daf-31 overexpression influences the subcellular localization of DAF-16 . We found the percentage of animals showing DAF-16 nuclear localization was not significantly different between daf-31 overexpressing animals and control worms ( Figure 4C and D ) . daf-2 mutants are resistant to environmental stresses such as high temperature [25] , [26] . We examined if daf-31 overexpression could enhance the thermotolerance of daf-2 mutants . As reported previously [25] , the survival rate of daf-2 mutants at 35° is doubled compared to N2 worms ( Figure S7 and Table S2 ) . However , the mean survival for N2 and N2 overexpressing daf-31 was similar ( 9 . 8 hours vs . 10 hours ) ( p = 0 . 2420 , log-rank test ) ( Figure S7 and Table S2 ) . Similarly , daf-31 overexpression did not increase the survival of daf-2 mutants at 35° ( p = 0 . 4623 , log-rank test ) ( Figure S7 and Table S2 ) . We tested the influence of daf-31 overexpression on the reproduction of N2 and daf-2 mutant animals . daf-31 overexpression increased the brood size of N2 animals and daf-2 mutants by about 23% and 30% , respectively ( Figure S8 ) . While daf-31 overexpression increased the daf-2 mutant lifespan in a daf-16 dependent manner , daf-31 overexpression increased the reproduction of daf-2 mutants significantly in a daf-16 independent way . Overexpression of daf-31 increased the brood size of daf-16;daf-2 mutants by 40% ( P<0 . 01 , t-test ) ( Figure S8 ) .
We demonstrated that daf-31 mutants form dauer-like larvae that share some characteristics of wild-type dauer larvae such as fat accumulation . Many daf genes , especially those from the insulin-like signaling pathway , are involved in the regulation of lifespan [27] . Mutations in daf-2 , which encodes an insulin/IGF-1 receptor [9] , convey a temperature-sensitive Daf-c phenotype , and the adults live twice as long as wild-type animals [9] , [28] , [29] . Mutations in some genes downstream of daf-2 , such as age-1 and pdk-1 , also extend lifespan [30] , [31] . Conversely , mutations in other downstream genes , including daf-18 and daf-16 , shorten lifespan [32]–[34] . We examined whether daf-31 is also involved in aging and found that daf-31 partially mediates the effect of reduced daf-2/IGF signaling pathway on C . elegans lifespan . Moreover , overexpression of daf-31 enhances the longevity phenotype of daf-2 mutants depending on the activity of DAF-16 . Supporting this lifespan data , the expression levels of sod-3 and bcmo-2 , the transcriptional targets of the DAF-16 FOXO3 transcription factor , are up-regulated in the daf-31 overexpression strains . Thus , it is reasonable to argue that DAF-31 regulates C . elegans lifespan by influencing DAF-16 transcriptional activity and daf-31 overexpression stimulates DAF-16 activity . Indeed both DAF-31 and DAF-16 are expressed in neurons and intestine , two major tissues essential for regulation of C . elegans lifespan by the DAF-2/IGF signaling pathway [35]–[37] . However , overexpression of daf-31 has no influence on the subcellular localization of DAF-16 . It is consistent with the lifespan data that overexpression of daf-31 does not increase the lifespan of N2 animals . Thus , overexpression of DAF-31 only extends C . elegans lifespan in the daf-2 mutant background in which DAF-16 has entered the nucleus due to inhibition of the IGF signaling . Previous studies show that the stress-resistance phenotype can be uncoupled from the longevity phenotype [38] . Indeed , although daf-31 overexpression further increases the long-lived lifespan of daf-2 mutants , it has no effect on the thermotolerance of daf-2 mutants . Interestingly , daf-31 overexpression increases the reproduction of both wild-type animals and daf-2 mutants , which is not dependent on DAF-16 , suggesting DAF-31 functions through DAF-16 for lifespan regulation but not for reproduction . Since DAF-16 is required for the stress resistance of daf-2 mutants , it is likely that daf-31 overexpression extends the daf-2 mutant lifespan through DAF-16-dependent mechanisms other than increasing stress-resistance . DAF-31 is found in multiple tissues including neurons . It is known that many C . elegans neurons are refractory to RNAi treatment in wild-type background [39] . It is possible that neuronal DAF-31 activity is more important for lifespan regulation because daf-31 RNAi treatment only shows influence on lifespan of RNAi-sensitive mutants . Supporting this assumption , it has been reported that daf-16/FOXO activity in neurons accounted for only 5–20% of the lifespan extension seen in daf-2 mutants [37] . Since DAF-31 may only influence DAF-16 activity in neurons , its overexpression only increases the daf-2 mutant lifespan modestly . We cloned the daf-31 gene and sequence analysis indicates DAF-31 is a worm ortholog of ARD1 that was first identified in yeast [40] . ARD1 is the catalytic subunit of the major NatA that transfers an acetyl group from acetyl coenzyme A to the N-terminal of nascent polypeptides . Yeast ARD1 mutants fail to enter stationary phase and sporulate during nitrogen deprivation [40] . The yeast stationary phase is comparable to C . elegans dauer stage and is essential for survival when nutrients are limited . C . elegans enters dauer stage during starvation or under high concentration of pheromone . Our data show daf-31 mutants could not complete dauer morphogenesis in response to pheromone and starvation , which indicates daf-31 is required for dauer formation . Thus , both yeast ARD1 and worm DAF-31 play an important role in the developmental switch in response to the environmental nutrient limitation . Additionally , daf-31 mutants could not grow to fertile adults in an environment with abundant food suggesting its essential role in normal development . Similar to our observation , it has been reported that loss of Ard1 is lethal for D . melanogaster and affects cell survival or proliferation , indicating ARD1 is required for D . melanogaster development [41] . In addition to developmental arrest , the daf-31 mutants shift metabolism to fat accumulation . Interestingly , yeast ARD1 mutants not only fail to enter stationary phase but also do not accumulate as much carbohydrates as wild-type yeast strains [40] . Thus , the function of ARD1 in regulating development and metabolism appears conserved from yeast to C . elegans . N-terminal acetylation is one of the most common posttranslational protein modifications . It is estimated to occur on 50% of yeast proteins [20] , 71% of D . melanogaster cytosolic proteins [20] and 84% of human proteins [42] . NatA plays the most prominent role in N-terminal acetylation . It would be interesting to know whether the pleiotropic phenotypes of ard1 mutants result from global changes of protein N-acetylation or from acetylation status of specific protein substrates . It has been reported that human ARD1 directly acetylates β-catenin and enhances its transcriptional activity [43] . We show that overexpression of DAF-31 stimulates the transcriptional activity of DAF-16 . It would be interesting to examine if DAF-31 overexpression acetylates DAF-16 . Alternatively , a suppressor screening of daf-31 mutants may help to identify the essential substrates of the DAF-31 acetyltransferase and contribute to understanding the mechanisms by which ARD1 influences development , metabolism and aging . Moreover , emerging evidence has revealed that abnormal regulation of ARD1 is associated with tumorigenesis and ARD1 represents a novel cancer drug target [44] , [45] . Identification of DAF-31 substrate proteins may uncover new therapeutic targets of cancer diseases .
All strains were grown on NG agar plates seeded with E . coli strain OP50 [46] . Mutations are listed by linkage groups as follows: LG I: daf-16 ( mgDf47 ) ; LG II: rrf-3 ( pk1426 ) ; LG III: daf-2 ( e1370 ) ; LG IV: unc-24 ( e138 ) , daf-31 ( m655 ) ; LG X: daf-3 ( e1376 ) , daf-12 ( m20 ) , daf-12 ( rh61rh411 ) . All mutants are derived from the wild-type Bristol N2 strain . To make the daf-16 ( mgDf47 ) I; daf-2 ( e1370 ) III double mutant , daf-2 ( e1370 ) males were mated with daf-16 ( mgDf47 ) hermaphrodites . Ten F1 adults ( daf-16/+; daf-2/+ ) were incubated at 25°C . F2 dauer larvae ( either +/+; daf-2 or daf-16/+; daf-2 ) were transferred to a fresh plate at 15°C for recovery . Then adults were shifted to 25°C . Since daf-16 ( mgDf47 ) can suppress the daf-2 ( e1370 ) Daf-c phenotype , non-dauer adults from the next generation were daf-16 ( mgDf47 ) ; daf-2 ( e1370 ) double mutants . Double mutants were constructed for epistatic tests between daf-31 and daf-d mutants at 20° . However , daf-16 RNAi was used for epistasis analysis between daf-31 and daf-16 . daf-16 RNAi fully suppressed dauer formation of daf-2 ( e1370 ) mutants . daf-12 ( m20 ) and daf-12 ( rh61rh411 ) mutations were used to construct the strain+daf-31 ( m655 ) /unc-24 ( e138 ) +; daf-12 ( m20 ) and +daf-31 ( m655 ) /unc-24 ( e138 ) +; daf-12 ( rh61rh411 ) using standard genetic methods . The daf-12 ( rh61rh411 ) mutation was confirmed by sequencing . +daf-31 ( m655 ) /unc-24 ( e138 ) +; daf-3 ( e1376 ) was constructed to determine the epistatic relationship between daf-31 and daf-3 . As the daf-31 ( m655 ) mutation was not marked by a genetic mutation in daf-31;daf-3 and daf-31;daf-12 mutant worms , the daf-31 ( m655 ) deletion mutations were confirmed by using single worm PCR . Representative gel pictures are shown in Figure S9 . Injection of RNAi was used to inhibit the daf-15 gene in unc-24 ( e138 ) daf-31 ( m655 ) /nT1 to examine the epistatic relationship between daf-15 and daf-31 . To construct the daf-31 overexpressing strains , the full-length daf-31 genomic DNA , including its native 760-bp promoter and 3′-UTR was cloned into pGEM-T ( Promega ) ; this construct successfully rescued the daf-31 dauer-like mutant phenotype . Multiple copies of the construct were integrated into chromosomal DNA by γ-irradiation to make an N2 transgenic line overexpressing daf-31 . Then the daf-31 overexpressing chromosome was introduced into both daf-2 ( e1370 ) and daf-16 ( mgDf47 ) ;daf-2 ( e1370 ) mutants by genetic crosses . To make a daf-31 promoter-GFP transcriptional fusion , the 760-bp daf-31 promoter was inserted into the GFP vector pPD95 . 70 ( a gift from Dr . Andrew Fire at Stanford University ) between the PstI and BamHI sites . The construct was injected into N2 adults at a concentration of 100 µg/ml . pRF4 , which encodes a mutant collagen and induces a dominant roller ( Rol ) phenotype , was co-injected at the same concentration as a transformation marker . Rol animals were selected from the F2 generation and used to establish stable transgenic lines . To evaluate the subcellular location of DAF-16 , TJ356 ( zIs356 [daf-16p::daf-16a/b::GFP+rol-6] ) males were crossed to daf-31 overexpressing hermaphrodites . The roller progeny were mounted on 2% agar pads to examine DAF-16::GFP subcellular localization . To stain fat using Sudan Black B , N2 , daf-2 ( e1370 ) and daf-31 ( m655 ) IV/nT1[unc- ? ( n754 ) let- ? ] ( IV;V ) synchronized L1 larvae were placed on NG agar plates , incubated at 20°C until they entered L3 or L4 stages , collected and washed two to three times with M9 buffer . Paraformaldehyde stock solution ( 10% ) was added to a final concentration of 1% . The samples were frozen in dry ice/ethanol and then thawed under a stream of warm water . After a total of three freeze-thaw cycles , the worms were stained with Sudan Black B as described by Kimura et al . [9] . Nile red staining of fixed worms was performed as described by Pino et al . [47] . Worm samples were collected and washed twice with M9 buffer . After the final wash , worms were fixed in 40% isopropanol at room temperature for three minutes . The fixed worms were stained in Nile red/isopropanol solution for 30 minutes at room temperature with gentle rocking . The stained worms were washed once with 1 ml M9 buffer and mounted on a 2% agarose pad for microscopy under the fluorescence channel . In order to compare the fat content in different strains , the pictures were taken at the same camera setting under 20× magnification . Three-factor-mapping with SNP markers and cosmid rescue were performed as previously described [18] . To determine the mutation in daf-31 ( m655 ) , the K07H8 . 3 gene ( GenBank accession # NM_068991 . 4 ) was amplified using primers 5′- GTG AGT CGA AAC CCA TTT TG -3′ and 5′- GAA TGA ACC AGT TGG AAA AGG -3′ from both N2 and daf-31 ( m655 ) mutant homozygotes . PCR products were cloned into the pGEM-T vector ( Promega ) following the manufacturer's instructions . T7 primer 5′- GTA ATA CGA CTC ACT ATA GGG -3′ and SP6 primer 5′- TAC GAT TTA GGT GAC ACT ATA G -3′ were used in DNA sequencing reactions . Part of the daf-31 coding region was amplified from C . elegans genomic DNA using primers 5′- CGG GAT CCA TTC GTT GTG CTC GCG TG -3′ and 5′- CCC AAG CTT GCA GTG GTA TAG GCC TC -3′ . The PCR products were then purified and cloned into the feeding RNAi vector L4440 ( Addgene ) between the BamHI and HindIII sites . The RNAi construct was transformed into E . coli HT115 ( DE3 ) and RNAi feeding was performed as previously described [48] . To inhibit daf-15 and daf-31 genes by injection of RNAi , a 1 kb daf-15 cDNA fragment and the full-length daf-31 cDNA were cloned into pGEM-T vector ( Promega ) , respectively . The gene identity was confirmed by sequencing . The Riboprobe Combination System-SP6/T7 ( Promega ) was used to transcribe RNA in vitro according to the manufacturer's protocol . Double-stranded RNA was synthesized and injected as described by Fire et al . [49] . Synchronized N2 L1 larvae were treated with either control ( empty ) vector or daf-31 RNAi by feeding as previously described [48] . Day 1 adult animals were collected for total RNA extraction using the Trizol kit ( Zymo ) . Synchronized L1 larvae of daf-31 overexpressing strains were allowed to grow on OP50 food plates . Day 1 adult animals were collected for total RNA extraction using the Trizol reagent ( Zymo ) . The first strand cDNA was synthesized using the ImProm-II reverse transcription system ( Promega ) . SYBR green dye ( Quanta ) was used for qRT-PCR to measure the expression level of daf-31 , sod-3 and bcmo-2 in corresponding worm samples . Reactions were performed in triplicate on an ABI Prism 7000 real-time PCR machine ( Applied Biosystems ) . Relative-fold changes were calculated using the 2−ΔΔCT method . The primers used for qRT-PCR were: daf-31 , 5′- GAA GAT CAC AAG GGA AAT GTT G -3′ and 5′- CTC TTG CGG TCT GAT CCA TC -3′; act-1 , 5′- CAA TCC AAG AGA GGT ATC CTT ACC CTC -3′ and 5′- GAG GAG GAC TGG GTG CTC TTC -3′; bcmo-2 , 5′- GCC GAT TTA GAG AAC GGA GAT CAC -3′ and 5′- TGA GAA TTC CGT CAT CTT CCC GA -3′; sod-3 , 5′- GGA ATC TAA AAG AAG CAA TTG CTC -3′ and 5′- CGC GCT TAA TAG TGT CCA TCA G -3′ . About 120–150 L4 larvae raised at 20°C were transferred to ten NG agar plates ( twelve to fifteen animals per plate spread with either OP50 or RNAi food ) and incubated at 25°C . The first day of adulthood is day 1 in the survival curves . During the reproductive period , adult animals were transferred daily to fresh plates . Thereafter , animals were transferred every ten days ( OP50 food ) or every six days ( RNAi food ) . Animals were scored as alive , dead , or lost every other day . Animals that do not move in response to touching were scored as dead . Animals that died from causes other than aging , such as sticking to the plate walls , internal hatching or bursting in the vulval region , were scored as lost . GraphPad Prism was used for statistical analysis and generation of survival curves . For the thermotolerance experiment , day 1 adult animals were incubated at 35°C and survival was scored as described above . To measure reproduction of worms , L4 larvae growing at 20°C were transferred daily to fresh plates and the progeny were counted .
|
The development of a living organism is influenced by the environmental conditions such as nutrient availability . Under starvation conditions , the C . elegans larvae will enter a special developmental stage called dauer larva . An insulin-like signaling pathway controls dauer formation as well as adult lifespan by inhibiting the activity of FOXO transcription factor DAF-16 that regulates expression of stress-resistant genes . Here we isolate a new gene called daf-31; this gene encodes a protein that regulates C . elegans larval development , metabolism and adult lifespan . This protein has been found in other species to be part of an enzyme that functions to modify other proteins . We show that overexpression of our newly discovered protein stimulates the transcriptional activity of DAF-16 . Interestingly , abnormal regulation of human proteins similar to DAF-31 results in tumor formation . It is known that human FOXO proteins prevent tumorigenesis . Therefore , it is possible that abnormal DAF-31 activity may lead to tumor growth by reducing DAF-16 activity . Thus , the present study may not only contribute to understanding the role of a universal enzyme in controlling development , metabolism and lifespan in other organisms besides worms but may also shed light on the mechanisms of tumorigenesis in humans .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology",
"and",
"life",
"sciences"
] |
2014
|
daf-31 Encodes the Catalytic Subunit of N Alpha-Acetyltransferase that Regulates Caenorhabditis elegans Development, Metabolism and Adult Lifespan
|
Trachoma , caused by Chlamydia trachomatis ( Ct ) , is the leading infectious cause of blindness worldwide . Yearly azithromycin mass drug administration ( MDA ) plays a central role in efforts to eliminate blinding trachoma as a public health problem . Programmatic decision-making is currently based on the prevalence of the clinical sign “trachomatous inflammation-follicular” ( TF ) in children . We sought to test alternative tools for trachoma surveillance based on serology in the 12-year cohort of Kahe Mpya , Rombo District , Tanzania , where ocular chlamydial infection was eliminated with azithromycin MDA by 2005 . The present study was a community-based cross-sectional survey in Kahe Mpya . Of 989 residents , 571 people aged 6 months to 87 years were enrolled: 58% of the total population and 73% of 1–9 year olds , the key WHO indicator age group . Participants were examined for TF , had conjunctival swabs collected for nucleic acid amplification test ( NAAT ) -based detection of Ct , and blood collected for analysis of antibodies to the Ct antigens pgp3 and CT694 by multiplex bead-based immunoassay . Seroconversion rate was used to estimate changes in the force of infection in a reversible catalytic model . No conjunctival swabs tested positive for Ct infection by NAAT . Among 1–9 year olds , TF prevalence was 6 . 5% , whereas only 3 . 5% were seropositive . Force of infection modelling indicated a 10-fold decrease in seroconversion rate at a time corresponding to MDA commencement . Without baseline serological data , the inferences we can make about antibody status before MDA and the longevity of the antibody response are limited , though our use of catalytic modelling overcomes some of these limitations . Serologic tests support NAAT findings of very low to zero prevalence of ocular Ct in this community and have potential to provide objective measures of transmission and useful surveillance tools for trachoma elimination programs .
Trachoma , caused by the bacterium Chlamydia trachomatis ( Ct ) , is the leading infectious cause of blindness worldwide [1] . Infection can manifest clinically in a number of ways , including follicular conjunctivitis , classified as “trachomatous inflammation-follicular” ( TF ) in the WHO simplified grading system [2] if five or more follicles are present in the central upper tarsal conjunctiva; and/or inflammatory thickening , classified as “trachomatous inflammation-intense” ( TI ) if more than half of the deep tarsal vessels are obscured . Repeated infections can lead to conjunctival scarring ( TS ) and trichiasis ( TT ) , in which in-turned eyelashes rub against the globe and may result in visual impairment or blindness caused by corneal opacity ( CO ) [3] . Azithromycin mass drug administration ( MDA ) , recommended where the prevalence of TF is ≥10% in children aged 1–9 years , is a critical component of the strategy for Global Elimination of Trachoma by 2020 ( GET2020 ) [4] . The current WHO endpoint for cessation of community-based antibiotic treatment is a TF prevalence in 1–9 year-olds of <5% . Prevalence surveys illustrate that signs of active trachoma , TF and TI , exceed Ct infection rates . Follicular or intense conjunctivitis may be caused by non-chlamydial bacteria , with the relative importance of this phenomenon probably increasing after populations begin to receive azithromycin MDA [5] . Furthermore , the examination process can be difficult to standardize [6–9]; inter-observer agreement is often sub-optimal . The poor correspondence between signs and infection—seen at both individual and community level—is problematic , given that field grading is the basis of public health decision-making [5 , 10] . As trachoma elimination efforts are intensified globally and interventions move populations towards trachoma elimination goals , the availability of a post-elimination surveillance methodology with greater reliability than clinical examination will become increasingly important to allow programs to identify and respond to recrudescent infection . Recent efforts to evaluate serology as a viable option for post-MDA surveillance identified tests using two previously-described chlamydial antigens , pgp3 and CT694 , as having high sensitivity to detect current ocular infection , and high specificity using non-endemic controls [11] . The age-specific prevalence of serological responses to Ct antigens at community level could provide an informative proxy measure of intensity of transmission and an early indicator of transmission recrudescence . This study therefore examined the use of serological tools for monitoring and evaluation in a post-MDA setting by assessing the age-specific prevalence of signs of trachoma and Ct-specific antibody responses within a community in which MDA ceased in 2002 and ocular Ct infection was subsequently found to have been eliminated in 2005 [12] .
This study was conducted in the Tanzanian community of Kahe Mpya , Rombo District . Kahe Mpya consists of approximately 250 households , with a population ( in July 2012 ) of 989 . A Kilimanjaro Christian Medical College ( KCMC ) /London School of Hygiene & Tropical Medicine ( LSHTM ) /Huruma Hospital collaboration has been conducting trachoma research in this community since 2000 [13 , 14] . High coverage azithromycin MDA was delivered in 2000 and 2002 , and topical tetracycline ointment treatment was given , at intervals between 2000 and 2005 , to individuals with active trachoma; elimination of ocular Ct infection by 2005 was previously documented [12] . Ethical approval to carry out this research was obtained from the ethics committees at LSHTM ( UK ) , Centers for Disease Control and Prevention ( USA ) , KCMC / Tumaini University , and the National Institute for Medical Research ( TZ ) . All adults provided written informed consent , and for children under 18 , the consent of a parent or guardian was obtained . All Kahe Mpya residents were invited by village leaders to a series of central locations , where those consenting to the study underwent examination of both eyes by a trained , highly experienced ophthalmic nurse known to the community , using binocular loupes ( magnification ×2·5 ) and a torch . Signs of trachoma were graded according to the WHO simplified grading system [2] . After examination , swabs were collected from the everted upper eyelid of the right eye using a sterile polyester-tipped-swab by passing the swab across the conjunctiva four times . Swabs were placed into sterile polypropylene tubes and kept at 4°C until frozen ( -20°C ) . Individuals with signs of active trachoma were given a tube of 1% tetracycline eye ointment free of charge and instructed to apply it daily to both eyes for six weeks . Fingerprick blood was collected by Tanzanian registered physicians onto filter paper with six circular extensions , calibrated so that each extension absorbed 10μl of whole blood ( TropBio Pty Ltd , Townsville , Queensland , Australia ) . Each filter paper was air-dried then individually placed in a zip-lock bag and frozen ( -20°C ) . Each sample was affixed with a pre-printed bar-coded label that linked all samples from an individual but had no other patient identifier . Dried blood spots were shipped to the Centers for Disease Control and Prevention in Atlanta GA , USA , for detection of IgG antibodies against the previously described chlamydial proteins pgp3 and CT694 , on the Luminex platform , using previously defined cut-offs for positivity [11] . Briefly , serum eluted from dried blood spots was incubated with microbeads coupled to the antigens of interest , then excess serum washed off and bound antibody detected with an anti-human IgG and anti-human IgG4 biotinylated detection antibody , and finally detected using streptavidin-conjugated to phycoerythrin ( PE ) . The fluorescent signal emitted by bound PE was converted to a median fluorescence intensity ( MFI ) with background from the blank subtracted out ( MFI-BG ) . For pgp3 , a MFI-BG value of 1024 was established as the low-limit value for positivity , with an indeterminate range of 1024 to 5998 . For CT694 , a MFI-BG value of 232 was established as the low-limit value for positivity , with an indeterminate range of 232 to 1982 [11] . To examine the change in transmission following MDA , we used seroconversion rate ( SCR ) to estimate the force of infection by fitting a simple reversible catalytic model to the measured seroprevalence , stratified into yearly age-groups , using maximum likelihood methods [15] . For these models only individuals aged one year and over were included to remove the effect of maternally derived antibodies in infants . Evidence for temporal changes in SCR was explored by fitting models in which the SCR was allowed to change at a single time-point . The significance of the change was identified using likelihood ratio tests against models with no change , and profile likelihoods were plotted to determine confidence intervals for the estimated time of the change . Samples were processed at the LSHTM and tested in pools of five using the Roche CT/NG Amplicor kit ( Roche Molecular Systems , Pleasanton , CA , USA ) , with the intention of re-testing positive pools as individual samples [16–18] . Manufacturer’s instructions were followed except for sample extraction where a previously published protocol was used[14] . Two Ct positive and two Ct negative processing controls were run with each batch of specimens . According to the manufacturer’s directions , the Amplicor test was positive if the optical density read at 450 nm was ≥0·8 , negative if the signal was <0·2 , and equivocal if in-between . All equivocal tests were re-tested in duplicate , and only graded positive if at least one test was positive . All samples were analysed in anonymous fashion through the use of non-sequential sample codes linked only to patient records through the data collection sheet . Statistical analysis was carried out using STATA 12 and GraphPad Prism ( version 6 . 0 ) .
The population and study population structure of Kahe Mpya sub-village is summarized in Table 1 , based on census data collected in July 2012 for this study . From the total 989 residents of Kahe Mpya sub-village , 575 ( 58 . 1% coverage ) people aged 0 . 2–87 . 6 years ( median age 12 . 6 , Table 1 ) participated in the study . The overall prevalence of active trachoma ( TF , TI or both ) in the examined population ( n = 571; four individuals refused clinical exams ) was 4 . 6% , with 21 . 5% exhibiting signs of scarring trachoma ( TS/TT/CO , Table 2 ) . There were no WHO simplified grading scheme signs of trachoma in 76·6% of the study group . The prevalence of TF amongst the WHO index age group ( ages 1–9 years ) was 6·5% ( Table 2 ) . Only one individual ≥ 10 years had TF . TS was absent in those <10 years , but was observed in all age groups >10 years ( Table 2 ) . TT was present only in individuals >20 years of age , with an overall population prevalence of 1% ( Table 2 ) . CO was only diagnosed in 2 individuals ( 0·4% of study participants ) , both of whom were over 70 years old ( Table 2 ) . Overall , 33 . 8% of participants were seropositive against at least one antigen ( Fig . 1A ) . Seropositivity increased with age . By age 40 , over 90% of participants tested positive to at least one antigen ( pgp3 alone , CT694 alone , or both pgp3 and CT694 , black squares , Fig . 1A ) , and over 60% tested positive to both antigens ( Fig . 1A , red squares ) ; this trend continued to the oldest age groups ( Fig . 1A ) . The MFI also increased with age ( Fig . 1B ) . Of 200 children aged 1–9 , seven ( 3 . 5% ) had antibody responses to one antigen , whereas only two ( 1% ) had antibody responses to both antigens ( Fig . 1C ) . Five of the seven samples with pgp3 reactivity fell into the indeterminate range , as did both of the CT694-reactive samples ( Fig . 1C ) . Samples from six of the seven 1–9 year olds testing positive by serology were re-tested with separate pgp3 and CT694 bead sets and data replicated the original results ( S1 Table ) . None of the ocular swabs tested positive by NAAT . When a seroconversion model , which allowed for a single change in SCR , was fitted to the data , the best fit was provided by a change in transmission between 10–15 years previously , consistent with the timing of MDA in the years 2000 and 2002 ( Fig . 2A for antibody responses to either antigen; responses to individual antigens gave similar profiles ) . We chose a model in which SCR changed 10 years previously , which had a better fit than the model that assumed the SCR had remained constant ( Fig . 2B ) . The change in SCR before and after this change point is approximately a 10-fold reduction , from a pre-MDA SCR of 0 . 0448 ( 95%CI 0 . 0373–0 . 0537 ) to a post-MDA SCR of 0 . 004 ( 95%CI 0 . 0024–0 . 0093 ) ] .
Global efforts toward the elimination of blinding trachoma are being rapidly intensified , thanks to strong donor interest . As programs reduce the prevalence of disease and infection , robust surveillance systems will become crucial to detect any recrudescence in populations living in post-elimination settings . In this study , we examined the use of serological tools for trachoma in a post-MDA setting . The virtual absence of antibody responses in children born after MDA-precipitated elimination of ocular Ct infection reflects the lack of Ct transmission ( as suggested by NAAT ) and provides the first evidence that serological monitoring of antibody responses could be viable for informing programmatic decisions in the surveillance phase . Force of infection modelling shown here strongly supports the hypothesis that reductions in transmission in this community coincident with the commencement of azithromycin MDA were reflected in changes in Ct seroconversion rate . This suggests that serology could have a very useful programmatic role even in the absence of complete transmission interruption . Several factors could contribute to the presence of signs of active trachoma in a community with low or no transmission of conjunctival Ct . First , the WHO simplified grading system employs strict criteria for diagnosis , but was designed for simplicity rather than specificity . Our grader was , however , well trained , highly experienced , and internationally certified , and we are confident of the accuracy of his judgements about the presence or absence of TF . Second , the natural histories of infection and disease differ , with signs arising weeks after infection has been acquired and persisting for weeks or months after infection clears . At the population level , the prevalence of infection declines more rapidly than the prevalence of TF following MDA [12 , 19 , 20] with some studies showing that TF persists at levels >10% within the population for months or years after infection has subsided [21 , 22] . Finally , evidence suggests that , in low-trachoma-prevalence settings , the majority of TF is associated with conjunctival infection with non-chlamydial bacteria , including S . pneumoniae and H . influenzae [5 , 23] . Non-bacterial causes of conjunctivitis such as adenovirus [24] may also contribute to TF clinical diagnoses in low-trachoma-prevalence settings . The use of photographs to validate field exams is becoming increasingly common but we have not found it to be reliable [25] and did not incorporate it into this study . Antibodies against the Ct antigens among 1–9 years old in this study were present at very low prevalence and in general at very low densities . This is in stark contrast to areas of active transmission in which seropositivity exceeds rates of clinical disease , as would be expected from long-lived antibody responses . [26] and has high sensitivity for ocular infection , [11 , 26] In the present study , antibody responses in 1–9 year olds may be Ct-specific , resulting from ocular or respiratory Ct infection acquired at birth from a mother with genital tract infection [27] , or from ocular infection acquired outside the village or in the village itself . Because the target for trachoma programs is not the complete interruption of transmission , it would not be an indication of programmatic failure to find ongoing low-level transmission in a community . However , it should also be noted that the previously determined specificity limits of this serological assay were 96–98% , such that the 3·5% of 1–9 year old samples testing positive may be false positives[11] . Because data were collected from a single community and enrolment was lower than anticipated ( primarily due to lack of availability of participants at the time of enrolment , as many adults were working outside of the community at the time of the study ) , additional studies in post-MDA settings will be needed to confirm the generalizability of our data . While the overall study enrolment was 58 . 1% of the total population , enrolment of 1–9 year olds , the key WHO indicator age group , was approximately 72 . 9% ( extrapolated from 2010 census data ) . Without baseline serology data , the inferences we can make about antibody status before MDA , the longevity of antibodies , and how antibody titers change over time in relation to one another are restricted , although our application of catalytic modelling overcomes some of these limitations . While comparing baseline to post-MDA antibody levels would be optimal , programs using serological tests as monitoring tools for intervention impact would need to do so in populations from whom baseline serological data will be absent . Antibody responses will therefore be most useful as surveillance tools by focusing analyses on children born after initiation or cessation of interventions . The data presented in the current study show the power of antibody-based surveillance in children born after cessation of an MDA program , data supported by the historical documentation of interruption of ocular Ct transmission in this community . Antibody responses represent exposure to infection and , when integrated with age , represent exposure over time; this can be done simply by applying a catalytic conversion model . SCR has been used widely in a range of infectious diseases [28 , 29] , most recently and extensively for malaria , for which SCR has been shown to correlate with the force of infection [15 , 30 , 31] . Fitting models with two SCRs enabled the measurement of changes in force of infection . SCR suggests a 10-fold decrease in the force of infection from approximately 5% seroconversion in the population per year prior to MDA , to approximately 0 . 5% after MDA , which closely approximates the 0% ocular infection prevalence seen in this study . Catalytic models can be refined by using serological data from multiple settings , pre- and post-MDA , to further validate the use of serological testing for programs . Serological tests for measuring antibodies in children may represent the best option for monitoring transmission because of the potential for greater sensitivity as population-based markers of exposure . Additionally , they provide an objective marker , relatively free of observer bias ( unlike examination for clinical signs ) , and are likely to be lower in cost than NAATs and provide data on cumulative exposure to the bacterium . Programmatically , such an assay could be used in the same way that antigen detection assays are used in surveillance for lymphatic filariasis elimination programs , and seroprevalence has been proposed for malaria control and elimination programs [30 , 32]; that is , to document reductions in the force of transmission . With the recent increased emphasis on a more horizontal approach to disease control , given similarities in control methods ( particularly periodic MDA ) and the geographical overlap between trachoma and other NTDs , integration across NTD programs is the next step [33] . This will provide economic and pragmatic benefits , as a multiplexed serological tool has the potential to map , monitor and evaluate several diseases simultaneously , facilitating efforts to achieve long-term elimination goals .
|
Trachoma is the leading infectious cause of blindness . The infectious agent , Chlamydia trachomatis , can be treated with a single oral dose of azithromycin . Donated drug is a cornerstone of programs dedicated to the elimination of trachoma as a public health problem . Azithromycin is given to the entire district for 3–5 years when 10% or more of 1–9 year-olds in the district have signs of a defined follicular conjunctivitis in one or both eyes . However , follicles can be difficult to reliably diagnose and can be caused by other pathogens , especially in settings with low trachoma prevalence . More sensitive and specific ways to assess communities for trachoma transmission at program endpoints are needed . Herein we examined antibody responses in children living in a community in Tanzania born after stopping drug treatment 10 years previously . Low antibody levels ( 3 . 5% in 1–9 year-olds ) reflected the lack of ocular chlamydial infection in these children . We also modelled the data to show that changes in age-specific antibody prevalence occurred when the mass drug treatment stopped . These data suggest that the age-specific prevalence of antibody responses may be of use to programs seeking to demonstrate the impact of interventions against trachoma .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Serology for Trachoma Surveillance after Cessation of Mass Drug Administration
|
Apicomplexan parasites are responsible for a myriad of diseases in humans and livestock; yet despite intensive effort , development of effective sub-unit vaccines remains a long-term goal . Antigenic complexity and our inability to identify protective antigens from the pool that induce response are serious challenges in the development of new vaccines . Using a combination of parasite genetics and selective barriers with population-based genetic fingerprinting , we have identified that immunity against the most important apicomplexan parasite of livestock ( Eimeria spp . ) was targeted against a few discrete regions of the genome . Herein we report the identification of six genomic regions and , within two of those loci , the identification of true protective antigens that confer immunity as sub-unit vaccines . The first of these is an Eimeria maxima homologue of apical membrane antigen-1 ( AMA-1 ) and the second is a previously uncharacterised gene that we have termed ‘immune mapped protein-1’ ( IMP-1 ) . Significantly , homologues of the AMA-1 antigen are protective with a range of apicomplexan parasites including Plasmodium spp . , which suggest that there may be some characteristic ( s ) of protective antigens shared across this diverse group of parasites . Interestingly , homologues of the IMP-1 antigen , which is protective against E . maxima infection , can be identified in Toxoplasma gondii and Neospora caninum . Overall , this study documents the discovery of novel protective antigens using a population-based genetic mapping approach allied with a protection-based screen of candidate genes . The identification of AMA-1 and IMP-1 represents a substantial step towards development of an effective anti-eimerian sub-unit vaccine and raises the possibility of identification of novel antigens for other apicomplexan parasites . Moreover , validation of the parasite genetics approach to identify effective antigens supports its adoption in other parasite systems where legitimate protective antigen identification is difficult .
The protozoan phylum Apicomplexa contains pathogens of substantial medical and veterinary importance including Plasmodium , Toxoplasma , Cryptosporidium , Eimeria , Neospora and Theileria species . Despite several decades of effort vaccines protective against these and other related parasites are scarce . Empiric approaches to identify genuinely immunoprotective antigens as vaccine candidates have achieved mixed results ( e . g . [1] , [2] ) and problems differentiating immunogenicity from real immune protection persist . Here we report the culmination of our efforts to develop a new approach to candidate antigen identification , one based upon using immunity as a selective barrier allied with pathogen genetics and mapping to identify true immune-targeted loci . Having identified candidate genomic regions we used a variety of strategies to locate the antigen responsible for protection . This approach has yielded new protective antigens for Eimeria maxima , one of the most important apicomplexan parasites to afflict livestock and provides insight into the nature of protective antigens in a broader context . As sustainable food security gains in importance pathogens which impact upon poultry production are re-emerging as serious threats to global food supply and human poverty [3] , [4] . Eimeria species parasites have a globally enzootic distribution and can cause severe enteric disease in all livestock , most notably poultry , where the annual cost is estimated to exceed £2 billion worldwide [5] . Current control is dominated by prophylactic application of anticoccidial drugs but drug resistance , political/consumer concerns over residues and the lack of new pipeline products renders this an unsustainable approach . Alternatives are limited by cost and/or efficacy and new solutions are urgently required . Eimeria are highly immunogenic parasites . Infection with as few as five E . maxima oocysts can induce complete protective immunity against subsequent homologous challenge [5] . Conversely , different strains of E . maxima can be antigenically diverse such that infection by one strain can induce little or no protection against challenge by a different strain [6] . Through the combination of in vivo selection imposed by strain-specific immunity and/or anti-parasitic medication with a population-based mapping strategy developed for use with apicomplexan parasites , we and others have shown that loci affecting strain-specific immunity can be mapped genetically [6] , [7] , [8] . We previously identified a panel of genetic markers within the E . maxima Weybridge ( W ) genome whose inheritance correlated absolutely with susceptibility to strain-specific immune killing [6] . Here we identify that the genetic markers associated with immunity map to just six regions representing less than 0 . 8% of the genome . Using a combination of sequencing , fine mapping and vaccination screens we have identified antigens responsible for protection in two of these loci . Homologues of both can be found in multiple apicomplexan parasites and one is known to be protective in diverse parasites of this phylum including Plasmodium spp . [1] . Hence , the novel antigens and remaining immune-mapped loci found within this study will have direct impact on eimerian vaccine development and have potential to impact on development of new vaccines with other apicomplexan parasites .
The E . maxima Houghton ( H ) and W strains are genetically and phenotypically distinct [6] . The H strain is characterised by sensitivity to dietary robenidine at 66 ppm and complete escape from W strain-specific immune killing during passage in inbred Line C White Leghorn chickens ( induced by previous host exposure to the W strain ) . In contrast , the W strain is resistant to 66 ppm robenidine but completely susceptible to W strain-specific immune killing . Genetic characterisation of the E . maxima H and W strains using amplified fragment length polymorphism ( AFLP ) with five different enzyme combinations to minimise restriction-associated bias generated a total of 3 , 230 genetic markers ( Table S1 ) . Comparison between the two strains revealed 1 , 122 markers to be polymorphic ( 34 . 7%; considerably higher than described previously for two Eimeria tenella strains [9] ) . Inheritance patterns of these strain-specific ( SS ) markers by a parasite mapping panel ( Figure 1; Table S2 ) , which included the uncloned progeny of eight independent H/W strain crosses before and after concurrent immune/robenidine selection , revealed the absolute correlation of 36 and 2 markers with the immune and drug barriers respectively ( Figure 2 ) . Importantly , all of the 36 immune correlated markers were subject to strong negative selection and were completely lost in all independent parasite lineages . Some other AFLP fragments were found to be less intensely amplified from immune selected parasites , suggesting more distant linkage to a mapped locus or linkage to loci containing genes contributing subtly to the complex biology of protection , either as modifiers , regulators , or minor antigens . Our analysis has focussed on the regions marked by AFLP fragments under the strongest selection by immunity . Serial in vivo passage of four hybrid populations under double barrier selection for up to five generations did not change the marker inheritance profile ( as [6] ) . Sequencing the 36 SS markers whose inheritance correlated with susceptibility to strain-specific immune selection identified 32 suitable for use as hybridisation probes ( EMBL FN813211-8 and unpublished ) . When radiolabeled with 32P and used to probe an E . maxima W strain BAC library representing ∼7 . 5-fold genome coverage a total of nine BACs were highlighted ( Table 1 ) . Each BAC was identified by at least two independent markers . BAC end sequencing and subsequent sequence-specific PCR facilitated the assembly of these nine BACs into six clusters , representing six distinct loci ( Table 1 ) . BAC insert release by Not I digestion and resolution by pulsed field gel electrophoresis ( PFGE ) provided approximate cluster sizes which were later refined following BAC insert sequencing and assembly ( Table 1; FN813242–4 and unpublished ) . Marker hybridisation to Southern blotted full PFGE-resolved karyotypes identified the chromosomal location of five of the six clusters ( Table 1 ) . Confirmation of the potent immunogenicity of the early stages of the E . maxima lifecycle supported the development of an in vivo BAC-screening system to begin fine mapping for each locus ( Table S3 , trial 1 ) . Utilising recent advances in transfection protocols for Eimeria [10] purified E . maxima H strain sporozoites were transiently transfected with each individual W strain BAC identified by hybridisation to an immune-correlated SS marker or a randomly selected control BAC ( Table S3 , trial 2 ) . When Line C chickens were immunised by infection with 1×106 test- , control- or non-BAC transiently transfected H strain sporozoites , drug-cleared by exposure to 66 ppm dietary robenidine to avoid clinical coccidiosis and challenged three weeks later with the W strain the capacity to induce cross-protective immunity was conferred by BACs derived from five of the six mapped loci ( Table 1 ) . In three independent experiments BACs representing loci 2 and 5 consistently induced the highest levels of immune protection , followed by BACs from loci 1 , 3 and 4 ( Table 1 ) . No cross-protective phenotype was observed using the sixth locus , suggesting either a mapping error , a requirement for a strain-specific partner molecule ( encoded elsewhere in the genome ) or stage-specific antigen expression towards the latter stages of the E . maxima lifecycle . The transient nature of our transfection approach is such that that if a protective antigen or other protection-affecting element was not expressed before loss of the host BAC then protection would not be detected . The six loci under intense immune selection ( i . e . essential loci ) are dispersed across the E . maxima genome ( Table 1 ) . Importantly , BAC-transfection based vaccination with five of the loci induced strain-specific immunity , indicating that each can act independently . Having identified relevant polymorphisms we can use this information to perform much wider studies on the genetic basis of protection including the population dynamics of the protective-antigen encoding genomic regions in laboratory and field populations . Preliminary co-transfection trials using two BAC clones representing loci 2 and 3 induced a higher level of immune protection ( 83 . 2±7 . 7% referenced to control BAC transfected parasites ) than either individual BAC ( 43 . 1% and 17 . 1% protection respectively ) . Indeed , locus 2/3 co-transfection immunisation gave greater protection than the additive effects of the individual loci; future studies will explore the combinatorial nature of protection in more detail . The parasite mapping panel was supplemented by backcrossing and re-selecting one double barrier-selected population to the immune-targeted parental W strain on six successive occasions . Genotyping each selected backcross generation maintained linkage with 11/32 SS markers , covering all six loci ( e . g . Figure 3A ) . Backcross genotyping fine-mapped locus 1 to ∼64 Kb of genomic sequence contained within BAC EmaxBAC8f18 . Typing additional SS PCR markers produced by targeted sequencing from the H strain across locus 1 provided further focus , mapping the region of interest to ∼50 Kb ( Figure 3B , Tables S4 and S5 ) . Bioinformatic examination of fine-mapped locus 1 identified three predicted coding regions , annotated as encoding ( i ) a sulphate transporter , ( ii ) an apical membrane antigen-1 ( AMA-1 ) homologue and ( iii ) a transcription elongation factor ( Figure 3C ) . Sequencing from the H strain identified amino acid polymorphism in the first two candidates but not the third ( FN813219–24 ) . In vivo infection by H strain sporozoites transiently transfected with purified genomic Long Distance ( LD ) PCR amplicons covering each predicted coding sequence and flanking regions induced a cross-protective immune phenotype only when using EmAMA-1 ( Figure 4 ) . AMA-1 has been widely proposed as an anti-apicomplexan vaccine candidate [11] , [12] and the E . tenella homologue has recently been found to be similarly protective ( Tomley , Billington et al , manuscript in preparation ) . Immunisation using EmAMA-1 as a DNA vaccine in the eukaryotic expression vector pcDNA3 . 1 ( + ) ( Invitrogen ) or as a bacterially-expressed recombinant protein induced significant immune protection against subsequent challenge by the W strain ( Figure 4 ) . The level of immune protection observed following immunisation using EmAMA-1 as a DNA vaccine was similar when challenged with 250 or 2 , 000 sporulated oocysts ( data not shown ) . Comparison between the H and W EmAMA-1 coding sequences revealed four nucleotide polymorphisms , two of which yield non-synonymous changes , one in the likely pro-domain and one in domain 1 [13] . Interestingly , AMA-1 domain 1 is also polymorphic among Plasmodium falciparum strains and polymorphism in this region confers the greatest level of escape from inhibitory antibodies [14] . Immunity induced by exposure to Eimeria spp . is highly species specific and although the E . maxima and E . tenella AMA-1 molecules share many structural characteristics the primary amino acid sequence is considerably different ( 57% aa identity ) . This level of divergence between eimerian AMA-1 is similar to that observed with Plasmodium AMA-1 from different species ( e . g . P . falciparum compared with Plasmodium vivax , 60% aa identity ) . Whole BAC transient transfection identified locus 5 , contained within BAC EmaxBAC2k08 , as capable of inducing the strongest protective immune response as judged by a reduction of oocyst production during challenge infection by 59 . 7% ( Table 1 ) . Backcross genotyping failed to provide any further focus , yielding a genomic region of interest spanning ∼90Kb ( Figure 5A–B ) . The difference in locus size may reflect variation in genome-wide recombination rates , as have been reported for other Apicomplexa [15] . Two parallel approaches were used to focus the search for protective antigens , one based upon targeted disruption of regions predicted to contain open-reading frames and the second based upon defined fragments of the BAC purified from restriction digests . Targeted disruption by homologous recombination ( BAC recombineering [16]; Table S6 ) at eight predicted coding regions ( identified by similarity to other annotated genes or EST sequences and clusters of candidate open reading frames; Figure 5C ) created a panel of eight otherwise unaltered EmaxBAC2k08 versions ( Figure S1 and S2 ) . The capacity of each daughter BAC to confer W-strain-specific protective immunity was tested by immunisation using the BAC transfection immunisation route followed by challenge with W strain parasites . For seven of the eight disrupted BACs no significant change in oocyst output was obtained compared with immunisation with the unmanipulated parent BAC . When region 7 was disrupted the protective effect was reduced to zero indicating that the important antigen or controlling element was associated with this region ( Figure 5D ) . When the regions flanking region 7 ( 6 and 8 ) were disrupted a small , but non-significant reduction in oocyst output was observed , possibly resulting from disruption of associated controlling or stabilising sequences . In parallel we analysed purified BAC sub-sections that resulted from Not I/Sfi I digestion of EmaxBAC2k08 , which yielded ∼63 . 7 and ∼22 . 4 Kb fragments of the insert separate from the vector , the smaller fragment including the candidate disrupted region . Immunisation using H strain sporozoites transiently transfected with either BAC fragment confirmed the induction of a cross-protective immune phenotype associated with the smaller but not the larger fragment ( 40% and −16% protection respectively compared to a randomly selected BAC control ) . Similarity-led gene annotation suggested the presence of two coding sequences within the ∼22 . 4 Kb locus . More detailed scrutiny using a locus-wide NimbleGen tiling array with cDNA derived from E . maxima W strain sporozoites , merozoites ( harvested 67 hours post infection , hpi ) and chicken intestine centred on Meckel's diverticulum without infection or 6 and 16 hpi with 1×106 W strain oocysts , revealed four sequences transcribed by the stages tested ( Figure 5E–G ) . Although eimerian lifecycles are relatively complex [5] , the choice of lifecycle stages to be sampled was informed by quantitative PCR-based enumeration of E . maxima replication in naive and immunised Line C chickens , which revealed the first 24 hpi to cover the period of most immune killing during homologous challenge ( ∼88% reduction at 20 hpi; Figure S3 ) . Examination of the array-identified transcribed sequences revealed ( i ) a putative non-coding RNA , ( ii ) an unknown coding sequence , ( iii ) a putative cyclophilin-RNA interacting protein and ( iv ) a SCY kinase-related protein ( FN813225–8 ) . LD PCR amplicons covering each predicted W strain coding sequence and flanking regions , including a section of repeats not found to be transcribed but identified as a cross-reactive feature by the array , were used in the BAC transfection immunisation assay . Protective immunity against W strain challenge was only evident with the unknown coding sequence ( Figure 6 ) . To confirm the protective capacity of this antigen , which we now term ‘immune mapped protein-1’ ( IMP-1 ) , we produced and purified bacterially-expressed protein and vaccinated Line C chickens . E . maxima IMP-1 recombinant protein induced 45% immune protection against challenge by the W strain as judged by reduction in oocyst output ( compared with the thioredoxin protein control , 50% compared with the unimmunised control; Figure S4 ) . Comparison between the H and W IMP-1 coding sequences revealed five nucleotide polymorphisms , two of which yield non-synonymous changes in amino acid sequence . Interrogation of the IMP-1 sequence using identification/prediction platforms including Phobius [17] , SignalP [18] , and SMART [19] suggest the absence of a classical signal peptide or recognisable domains . Nonetheless , predicted homologues can be identified within the genomes of other coccidial parasites including E . tenella ( 2e-87; FN813229; NCBI BLASTp2seq [20] ) , Toxoplasma gondii ( XM_002370108; 2e-36 ) and Neospora caninum ( GeneDB NCLIV_000430; 4e-37 ) , all of which share a common intron/exon structure ( Figure S5 ) . Further work on the biology of IMP-1 and the eimerian AMA-1 may reveal characteristics common to molecules that are capable of inducing strong protective immunity . Overall , our finding that just six regions of the genome were affected by strong immune selection is important since it suggests that protective immunity is focussed on a limited repertoire of parasite antigens . The genome of Eimeria spp . is estimated to be between 55 and 60 Mbp in size , encoding 8 , 000–9 , 000 genes ( http://www . genedb . org/Homepage/Etenella ) , and the adaptive immune system recognises a large number of antigens . Our data indicate , in this precisely controlled genetic context , that natural infection-induced protective immunity is focussed on the recognition of only a small subset of the antigenic repertoire expressed by the parasite . This feature of antigenically complex pathogens has troubled those involved in vaccine development for many years . Our finding that strain-specific immunity against E . maxima is absolutely targeted against just six loci is comparable with the number identified with murine malaria [21] and highlights the concept that antigenically complex pathogens may only express small numbers of protective antigens . This feature of immunity may explain the lack of success in developing effective sub-unit vaccines against antigenically complex pathogens despite decades of effort directed at using immunodominant antigens . Indeed , although measurable responses will be directed against “protective antigens” , responses against “non-protective” or “poorly protective” antigens obscures effective antigen selection . Using response as the major selection criterion in antigen discovery pipelines confers high rates of false positive leads . Protection is a much more discriminatory tool that can be interrogated using a technically straightforward genetic mapping approach , focussing discovery on “protective antigens” and importantly supporting simultaneous consideration of all elements of the pathogen , identifying “sets of antigens” responsible for strong protective immunity . Understanding the basis for discrimination of antigenic molecules that stimulate ineffective responses from those that stimulate protective responses has the potential for impact far beyond the scope of this project . In this report we document the application of a genetic approach to discover two protective antigens , one of which encodes an eimerian homologue of AMA-1 and the other a new vaccine candidate , IMP-1 . The former raises an interesting possibility that there are features of certain molecules that confer sensitivity to protective immune responses across a wide range of Apicomplexa . Interestingly , homologues of the IMP-1 gene can be readily identified in non-eimerian apicomplexan parasites and these may also be candidate protective antigens . Three of the other four loci contain elements that confer strain specific protective immunity ( by BAC transfection-immunisation studies ) and it is likely that these also contain protective antigens . One alternative is that these regions may exert their effects indirectly ( for example by regulation of other non-polymorphic loci ) although at present this seems unlikely . The nature of the protective effects encoded by the remaining loci is the focus for ongoing studies . We propose that our strategy will contribute to development of new anti-eimerian vaccines and may have much broader impact on the development of vaccines against some of the most devastating parasitic diseases of humans and livestock .
This study was carried out in strict accordance with the Animals ( Scientific Procedures ) Act 1986 , an Act of Parliament of the United Kingdom . All animal studies and protocols were approved by the Institute for Animal Health Ethical Review Committee and the United Kingdom Government Home Office under the project licences 30/2047 and 30/2545 . The E . maxima Houghton ( H , sensitive to dietary robenidine ) and Weybridge ( W , resistant to 66 ppm dietary robenidine ) strains were used as the parental parasites in these studies . Oocysts were propagated and genetic crosses were carried out in vivo as described previously [6] . Uncloned populations of hybrid parasites were derived from eight independent crosses between the H and W E . maxima strains in Line C White Leghorn chickens as described previously ( [6]; oocysts recovered from between two and ten birds and pooled for each population ) . Hybrid sub-populations were derived from each cross following in vivo passage under a double selective barrier comprising W strain-specific immune ( induced by previous infection with 100 oocysts of the pure W strain ) and H strain-specific drug ( 66 ppm robenidine ) selection [6] . All eight first generation selected parasite populations , together with four serially-selected populations ( three to five rounds of selection , two to ten birds pooled per population per generation ) , were used to prepare the parasite mapping panel ( Table S2 ) . One selected parasite population was backcrossed six times under double barrier selection , each backcross generation derived after in vivo phases of ( i ) cross fertilisation: in vivo passage using 100 sporulated hybrid oocysts with 400 sporulated W ( immune-targeted ) parental strain oocysts and ( ii ) selection: passage of 10 , 000 sporulated recovered parasites per bird under double barrier selection . Parasites recovered from ten birds were pooled for each backcross stage . Experiments to measure immune protection induced by ( i ) immunisation through previous parasite exposure , ( ii ) recombinant protein or ( iii ) DNA vaccination followed standardised protocols . All treatment groups comprised six individually caged specific pathogen free Line C White Leghorn chickens . Total daily oocyst excretion per bird was determined following daily faecal collection from days 6–7 , 7–8 and 8–9 post infection by flotation in saturated salt solution using a modified McMaster protocol as described previously [6] . All experiments included a non-immunised control group . cDNA sequences corresponding to the predicted AMA-1 ectodomain and IMP-1 proteins ( loci 1 and 5 respectively: AMA-1 coding nucleotides 79–1 , 347 , IMP-1 161–1 , 275 ) were amplified from E . maxima W strain sporozoite cDNA , cloned into the expression vector pET32b ( Novagen ) using Bam HI/Hind III and Nco I/Eco RI restriction sites respectively and sub-cloned in E . coli BL21 ( DE3 ) pLysS ( Novagen ) . Recombinant proteins were expressed and purified using HisTrap FF purification columns ( GE Healthcare ) as described by the manufacturer , dialysed overnight against PBS and finally mixed with an equal volume of adjuvant shortly before use . Thioredoxin expressed in the same manner using the unmodified pET32b vector was purified and used as a negative control . cDNA sequences corresponding to the predicted AMA-1 ectodomain and sulphate transporter proteins ( locus 1: AMA-1 coding nucleotides 79–1 , 347 , sulphate transporter 1–2 , 988 ) were amplified from E . maxima W strain cDNA , cloned into the eukaryotic expression vector pcDNA3 . 1 ( + ) ( Invitrogen ) using Hind III/Bam HI restriction sites and sub-cloned in E . coli XL1-Blue MRF ( Stratagene ) . Plasmid DNA was purified using the Qiagen EndoFree Plasmid Maxi kit as recommended by the manufacturer ( Qiagen ) , precipitated and re-suspended in endotoxin-free TE at 250 µg/ml . Genomic DNA was extracted from oocysts as described previously using a physical smashing step followed by phenol/chloroform extraction [22] and from chicken intestinal tissue samples using a Qiagen DNeasy tissue kit as described by the manufacturer ( Qiagen ) followed by RNase A treatment [23] . Total RNA was purified from E . maxima sporozoite , merozoite ( harvested 67 hours post infection ) , infected and uninfected chicken intestinal tissue using a Qiagen RNeasy kit as described by the manufacturer [24] . A BAC library was constructed for the E . maxima W strain in the pBACe3 . 6 vector following the protocol of Osoegawa et al [25] based upon chromosomal DNA prepared from E . maxima sporozoites as described elsewhere [26] . BAC DNA was prepared using the Qiagen Large-Construct kit as described by the manufacturer ( Qiagen ) . Standard PCR amplification was completed using BIO-X-ACT Short DNA Polymerase ( Bioline Ltd . ) . Each PCR reaction contained 5 ng template DNA , 20 pmol of relevant forward and reverse primers , 0 . 5 U Taq polymerase , 10 mM Tris–HCl , 1 . 5 mM MgCl2 , 50 mM KCl and 0 . 2 mM dNTPs . Standard cycle parameters were 1× ( 5 min at 94°C ) , 30× ( 1 min at 94°C , 1 min at 54–58°C and 1–5 min at 72°C ) and 1× ( 10 min at 72°C ) . For LD PCR BIO-X-ACT Long DNA Polymerase ( Bioline Ltd . ) was used as recommended by the manufacturer . Where required cDNA was prepared using Invitrogen Superscript II reverse transcriptase and oligo dT as described by the manufacturer ( Invitrogen Ltd . ) . PCR fragments were cloned using pGEM-T Easy ( Promega ) in XL1-Blue Escherichia coli ( Stratagene ) , miniprepped ( Qiagen ) and sequenced ( Beckman CEQ 8000 genetic analysis system ) as described by the respective manufacturers . Sequence assembly , annotation and interrogation were undertaken using VectorNTI v11 . 0 ( Invitrogen ) except where stated . Groups of four inbred SPF Line C White Leghorn chickens were either left naive or were immunised by infection with 100 sporulated E . maxima W strain oocysts at three weeks of age . All birds were subsequently challenged by infection with 1 . 0×106 E . maxima W strain oocysts at six weeks of age ( homologous challenge ) . Unimmunised and immunised groups were culled at 0 , 2 , 4 , 6 , 8 , 12 , 16 , 20 , 24 , 32 , 40 , 48 and 72 hours post challenge , when an 8 cm length of intestine centred on Meckel's diverticulum was recovered post-mortem from each test bird and homogenised in sterile TE using a Qiagen TissueRuptor ( 230 V , 50/60 Hz ) . Total genomic DNA was extracted from three 25 µl aliquots of each sample using a Qiagen AllPrep DNA/RNA Mini kit as described by the manufacturer ( Qiagen ) . The total number of E . maxima genomes per host genome was determined from each sample using TaqMan quantitative PCR assays specific for the E . maxima microneme protein 1 ( MIC1 ) and chicken glyceraldehyde 3-phosphate dehydrogenase ( GAPDH ) loci in duplex with the 7500 Fast Real-Time PCR System ( Applied Biosystems ) [23] . TaqMan probes were 5′ labeled with FAM ( MIC1 ) or Yakima Yellow ( GAPDH ) and 3′ quenched with Eclipse Dark Quencher ( Eurogentec ) . TaqMan conditions and cycle parameters were modified from the standard Applied Biosystems Fast protocol ( 1×95°C , 20 s; 40×95°C , 15 s and 60°C , 30 s ) . Quantitative calculations were facilitated and validated by comparison with known concentrations of the relevant genomic DNA template . AFLP was used to generate the majority of the genetic markers used during these studies as described elsewhere [27] . Approximately 50 ng total E . maxima genomic DNA was digested using one of five restriction enzyme combinations ( Table S1; New England Biolabs ) prior to ligation to adapters derived from those described by Vos et al , adapted for the respective restriction enzyme [27] . Primer pairs ( MWG Biotech ( UK ) Ltd . ) were based on the adaptor sequences and provided 0 and 1 ( primary amplification ) or 1 and 2 ( secondary amplification ) selective bases , respectively . Markers of interest were gel excised , re-amplified , cloned and sequenced as described previously [6] . Marker specific-primers were designed using Primer3 to amplify marker-associated DNA fragments [28] . Additional genetic markers were developed by sequencing 600–750 bp sections of E . maxima H strain genomic DNA corresponding to targets distributed across each locus mapped in the W strain ( Table S4 for the primers used , FN813230–41 ) . Strain-specific primer pairs were developed following sequence alignment and SNP identification ( ClustalX [29]; Table S5 ) . Chromosomal karyotypes were resolved by PFGE . Briefly , E . maxima W strain chromosomal DNA in ∼4 mm×3 mm×2 mm sections cut from an agarose block [26] was separated in a 0 . 8% SeaKem HGT agarose gel ( Lonza ) prepared in 0 . 5×Tris–borate EDTA ( TBE ) buffer and subjected to PFGE in a 21 cm×14 cm gel using a CHEF DRII system ( Bio-Rad ) in 2 L of 0 . 5×TBE running buffer at 14°C . PFGE conditions were ( i ) 216 h at 1 . 3 volts/cm with a switch time of 3 , 000–3 , 500 s followed by ( ii ) 120 h at 1 . 3 volts/cm with a switch time of 3 , 300–3 , 600 s finishing with ( iii ) 48 h at 1 . 2 volts/cm with a switch time of 3 , 200–3 , 400 s . Gels were stained in a 0 . 5 µg/ml aqueous solution of ethidium bromide for 30 min , destained in water and then photographed . Hansenula wingei and Schizosaccharomyces pombe DNA plugs ( Bio-Rad ) were included as size markers . Individual BAC clone insert size was determined by Not I ( New England Biolabs ) digestion followed by PFGE in 1% Bio-Rad pulsed field certified agarose prepared in 0 . 5×TBE as above . PFGE conditions were 18 h at 4 volts/cm with a switch time of 1–6 s . Low range PFG size markers ( New England Biolabs ) were included as size markers . PFGE-resolved chromosomal DNA was transferred to Hybond-N+ membrane ( Amersham Biosciences ) as recommended by the manufacturers . A total of 3072-BAC transformed E . coli DH10B were robotically gridded onto replicated filter arrays providing ∼7 . 5-fold coverage of the E . maxima W strain genome . PCR products derived from AFLP markers of interest were labelled with 32P using a Prime-It II random priming kit ( Stratagene ) and hybridised to filters for 16–24 h at 65°C as described by Amersham Biosciences . Filters were washed three times at 65°C in 0 . 1×SSC before exposure to X-ray film ( Kodak BioMax MS ) at −80°C against intensifying screens . BAC DNA was prepared from clones identified by hybridisation to AFLP markers of interest using the Qiagen Large-Construct kit as recommended by the manufacturer . Small insert libraries ( 2–4 Kb ) were prepared by shearing the DNA by sonication , blunt ending and size selecting by agarose gel electrophoresis prior to sub-cloning into Sma I digested dephosphorylated pUC18 for use in a whole-BAC shotgun sequencing strategy . ABI PRISM BigDye Terminator ( Applied Biosystems ) forward and reverse plasmid end sequences generated using an ABI3730 capillary sequencer were assembled using the Staden-based PHRAP ( P . Green , unpublished ) . Contig assembly was based upon LD PCR . Preliminary annotation of each assembled BAC sequence was achieved using tBLASTx [20] interrogation of all publically accessible sequences through the National Center for Biotechnology Information ( NCBI , http://www . ncbi . nlm . nih . gov/ ) , supplemented by Eimeria species EST data from the Eimeria ORESTES and E . maxima EST sequencing projects ( Gruber and Madeira , unpublished; Wan and Blake , unpublished , respectively ) . All sequences produced in this study have been submitted to EMBL , where they are available under the accession numbers FN813211–44 . The E . coli DY380 strain [16] ( kindly supplied by the National Cancer Institute , Frederick , USA ) was initially transformed with E . maxima W strain BAC EmaxBAC2K8 as described elsewhere [30] . Subsequently , the DY380/EmaxBAC2K8 line was transformed with each of eight PCR products representing ( i ) unique EmaxBAC2K8 sequences for targeted homologous recombination and ( ii ) the β-lactamase coding sequence amplified from pGEM-T Easy ( Promega; primers as shown in Table S6; PCR as above , purified using the Qiagen Gel Extraction kit as described by the manufacturer ) [30] . The resulting bacterial strains were sub-cloned and tested for evidence of correctly targeted insertion in a pure clonal line and the absence of widespread BAC disruption by ( i ) positive PCR between insert and flanking BAC sequences ( Figure S1 ) , ( ii ) negative PCR between target and flanking BAC sequences ( Figure S1 , primers shown in Table S7 ) and ( iii ) unchanged BAC PFGE profile following Not I/Sfi I digestion ( Figure S2; enzymes New England Biolabs ) . E . maxima W strain genomic DNA , presented as whole , recombineered or partial BAC-encoded templates or LD PCR amplicons , was used to transiently transfect the E . maxima H strain . For whole and recombineered BAC transfection purified plasmid DNA was re-suspended at 50 µg/10 µl TE . BAC EmaxBAC2k08 was subdivided by Not I/Sfi I digestion ( New England Biolabs ) and subsequent PFGE ( as above ) , yielding ∼63 . 7 and ∼22 . 4 Kb fragments of the insert as well as the vector . The 63 . 7 and 22 . 4 Kb fragments were gel excised following large-scale PFGE , electroeluted into dialysis bags in TE , precipitated and re-suspended at 37 and 13 µg/10 µl respectively in TE as described elsewhere [31] . LD PCR amplicons representing three EmaxBAC8f18 candidate regions and four EmaxBAC2k08 regions were amplified in triplicate , ( primers shown in Table S8 , designed to lie>1 Kb outside any BLAST hit with an E value of 1e-05 or below or to yield an amplicon>7 Kb in size across the predicted coding sequence , whichever was the greater ) . All triplicates were electrophoresed to check for purity and target size , identified by test secondary PCR ( Table S8 ) , pooled once validated , precipitated and re-suspended in 10 µl TE [31] . Transient transfection was accomplished following a protocol modified slightly from that described previously [10] . Briefly , freshly excysted and purified E . maxima H strain sporozoites were washed in incomplete cytomix and re-suspended in AMAXA Basic Parasite Nucleofector Solution 2 at 3 . 0×107/ml immediately prior to nucleofection . Subsequently , 100 µl sporozoite suspension was mixed with 10 µl DNA at room temperature , transferred to a cuvette and nucleofected as described by the manufacturer using a Nucleofector II with program U-033 ( Lonza ) . Post-nucleofection all sporozoites were immediately re-suspended in 3 ml PBS+1% glucose ( w/v ) and left to rest for 20 mins at room temperature . Sporozoites were then either used as an oral dose to initiate in vivo infection or incubated overnight at 41°C in a 5% CO2 incubator . For oral dosing the output from two nucleofections were pooled and dosed using 1 . 0×106 sporozoites ( counted pre-nucleofection ) per bird . Total RNA was extracted from incubated sporozoites using the Qiagen RNeasy mini kit for cDNA preparation and PCR of one or more transfected DNA-specific transgenes to confirm transfection ( data not shown ) . A custom-made NimbleGen genome tiling array containing 79 , 955 50–75 bp probes ( presented as forward and reverse strands in duplicate ) was designed and produced using BAC sequences obtained during these studies ( Roche ) . E . maxima W strain genomic DNA covered by BACs EmaxBAC8f18 and EmaxBAC2k08 were represented by 9 , 798 and 7 , 387 probes respectively ( EmaxBAC8f18: 8 , 321 unique , 1 , 477 twice , 84 . 5% coverage; EmaxBAC2k08: 7 , 315 unique , 72 twice , 95 . 5% coverage ) . Total RNA extracted from purified E . maxima W strain sporozoites , merozoites ( harvested 67 hpi ) , uninfected chicken intestinal tissue sampled at Meckel's diverticulum and chicken intestinal tissue sampled 6 and 16 hpi with 1×106 E . maxima W strain oocysts were processed to produce labelled cDNA and hybridised to the array as described by the manufacturer ( Roche ) . Arrays were scanned using a GenePix 4000B scanner and the images were visualised using GenePix Pro software ( Axon ) . Array design was viewed using NimbleGen SignalMap v1 . 9 software . Arrays were read using NimbleGen NimbleScan v2 . 5 software with alignment and uniformity cut off points of 0 . 15 and 0 . 3 respectively . Arithmetic mean and associated standard error of the mean ( SEM ) for each sample or group were calculated using Excel ( Microsoft Excel 2002 , Microsoft Corporation , 2001 ) . Statistical analyses were performed using the t-test , ANOVA , Chi2 or Kruskal-Wallis tests in Minitab ( Minitab Release 14 , Minitab Inc . , 2003 ) , complimented by post hoc analysis using the Tukey's test . Differences were deemed significant with a p value<0 . 05 .
|
Protozoan parasites are responsible for serious diseases in humans and livestock species . Vaccination is a declared intervention of choice with these infections , but even after many years of effort few effective vaccines are available . Identification of the right antigens for inclusion in sub-unit vaccines is a particular problem with complex pathogens . Moreover , the host response does not discriminate between protective and non-protective antigens , confounding development of effective screening systems . This study represents the culmination of work using parasite genetics and immunity as a selective barrier to find parts of the parasite genome targeted by immunity . The pathogen used in these studies ( Eimeria maxima ) is very important in livestock and related to a number of human pathogens including those responsible for malaria . Our studies indicate that just six regions in the genome were targeted by immunity and two of these have now been interrogated to determine the protective antigen encoding gene . Interestingly , one of these ( called AMA-1 ) has homologues known to be protective with other apicomplexan parasites . This raises the intriguing possibility that a set of homologous antigens may be protective across the apicomplexan parasites and that protective antigen discovery in one parasite may generate new leads in other vaccine programmes .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/gene",
"discovery",
"microbiology/immunity",
"to",
"infections",
"genetics",
"and",
"genomics/complex",
"traits",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"microbiology/parasitology",
"infectious",
"diseases/protozoal",
"infections",
"immunology/immunity",
"to",
"infections"
] |
2011
|
Genetic Mapping Identifies Novel Highly Protective Antigens for an Apicomplexan Parasite
|
Yeast prions are heritable amyloid aggregates of functional yeast proteins; their propagation to subsequent cell generations is dependent upon fragmentation of prion protein aggregates by molecular chaperone proteins . Mounting evidence indicates the J-protein Sis1 may act as an amyloid specificity factor , recognizing prion and other amyloid aggregates and enabling Ssa and Hsp104 to act in prion fragmentation . Chaperone interactions with prions , however , can be affected by variations in amyloid-core structure resulting in distinct prion variants or ‘strains’ . Our genetic analysis revealed that Sis1 domain requirements by distinct variants of [PSI+] are strongly dependent upon overall variant stability . Notably , multiple strong [PSI+] variants can be maintained by a minimal construct of Sis1 consisting of only the J-domain and glycine/phenylalanine-rich ( G/F ) region that was previously shown to be sufficient for cell viability and [RNQ+] prion propagation . In contrast , weak [PSI+] variants are lost under the same conditions but maintained by the expression of an Sis1 construct that lacks only the G/F region and cannot support [RNQ+] propagation , revealing mutually exclusive requirements for Sis1 function between these two prions . Prion loss is not due to [PSI+]-dependent toxicity or dependent upon a particular yeast genetic background . These observations necessitate that Sis1 must have at least two distinct functional roles that individual prions differentially require for propagation and which are localized to the glycine-rich domains of the Sis1 . Based on these distinctions , Sis1 plasmid-shuffling in a [PSI+]/[RNQ+] strain permitted J-protein-dependent prion selection for either prion . We also found that , despite an initial report to the contrary , the human homolog of Sis1 , Hdj1 , is capable of [PSI+] prion propagation in place of Sis1 . This conservation of function is also prion-variant dependent , indicating that only one of the two Sis1-prion functions may have been maintained in eukaryotic chaperone evolution .
Yeast prions are amyloid aggregates of functional yeast proteins that are both self-templating and heritable to daughter cells [1] , [2] , [3] , [4] . The best studied yeast prion , called [PSI+] , is the aggregated form of the yeast translation termination factor Sup35 [5] , [6] . [PSI+] cells have a distinct phenotype characterized by enhanced nonsense suppression causing increased read-through of stop codons by translating ribosomes [7] , [8] , [9] . The phenotype arises from Sup35 sequestration in prion aggregates and the strength of the nonsense suppression correlates to the lack of soluble Sup35 [10] . Another yeast prion , first identified as the genetic element [Pin] for Psi inducibility , was later shown to be the aggregated form of the Rnq1 protein , hereafter called [RNQ+] for the high Asn ( N ) and Gln ( Q ) content of its prion forming domain [3] , [11] , [12] . Rnq1 is a cytosolic yeast protein of unknown function [4] . Neither the deletion nor overexpression of Rnq1 , nor its aggregation in [RNQ+] cells results in any distinguishable phenotype beyond the tendency of [RNQ+] cells to spontaneously become [PSI+] at a greatly accelerated rate , hence the original denotation [PIN+] [3] , [12] , [13] . While both Rnq1 and Sup35 are cytosolic and have prion-forming domains , they do not significantly intermix in aggregates , and so [RNQ+] and [PSI+] form independent and stable structures in vivo [14] . Yeast prion propagation is dependent on the formation of heritable protein aggregates , often called ‘seeds’ or ‘propagons’ , that can be passed on to daughter cells during cell division [15] , [16] . Creation of yeast prion propagons results from the remodeling , and ultimately fragmentation , of amyloid aggregates by a specific set of cellular chaperone proteins minimally composed of Hsp70 , Hsp104 , and the J-protein Sis1 , the focus of this study [17] , [18] , [19] , [20] , [21] . Sis1 functions as a co-chaperone protein with the Hsp70 Ssa in the fragmentation of at least four yeast prions ( [PSI+] , [RNQ+] , [URE3] , and [SWI+] ) and probably others [21] , [22] , [23] , [24] , [25] . Hsp70s , like the yeast protein Ssa , have two distinct domains for client-peptide binding and ATP hydrolysis [26] . The hydrolysis of ATP in one domain triggers a structural change in the other which enhances client peptide binding [27] . J-proteins , like Sis1 , bind to the Hsp70 ATPase domain and catalyze ATP turnover , thus stimulating the association of Hsp70s with client peptides [28] . Because some J-proteins also bind client peptides themselves , they can also act as targeting factors , bringing Hsp70 to various cellular targets [29] [26] . Current models suggest that the J-protein Sis1 may act as a targeting factor which brings Ssa to prion aggregates , but the details of these interactions are unclear [21] , [24] , [26] . Hsp104 , a disaggregase , is essential for the propagation of all known yeast prions [30] , [31] , [32] . The current model for chaperone-dependent prion fragmentation posits that following Sis1/Ssa intervention , which most likely results in a partial unfolding of a prion protein , the chaperone Hsp104 binds a free end or loop of the prion protein and feeds the full protein through its central cavity [33] [21] , [23] . Multiple rounds of Hsp104-dependent monomer unfolding likely destabilize the aggregates , resulting in their eventual fragmentation [19] . Like other yeast J-proteins , Sis1 also has a domain-type architecture ( Figure 1 ) including an N-terminal J-domain , which is critical for stimulation of Hsp70's ATPase activity , and C-terminal domains which are known to bind model peptides and are homologous to other known J-protein peptide-binding sites [34] [26] , [35] . Separating the J-domain and C-terminal domains are two glycine-rich regions known as G/F ( Gly and Phe rich ) and G/M ( Gly and Met rich ) . Sis1 has recently been implicated in spatial protein quality control pathways involving target-protein ubiquitylation and protein sorting [36] [37] , [38] , [39] . While its specific cellular functions are still unclear , Sis1 is also an essential yeast protein; yeast cells are inviable unless the J-domain of Sis1 is expressed in cis with at least one of these two glycine-rich regions , underscoring their biological importance [40] . However , despite being essential , Sis1's expression can be greatly reduced , or the protein may be truncated , resulting in the support of cell viability but not prion propagation [18] [23] , [24] , [25] , [41] , [42] , [43] . For example , the C-terminal peptide-binding domain is generally dispensable for both viability and prion propagation [40] [41] , [43] . In contrast , when the G/F region is absent ( Sis1-ΔG/F , Figure 1 ) , [PSI+] is maintained but [RNQ+] is no longer supported [41] [24] . A construct of Sis1 which supports viability but not [PSI+] propagation has never before been identified . Due to a distinctive color phenotype , high mitotic stability , and a plethora of knowledge gleaned from years of investigations , the prion [PSI+] has become the best understood and arguably , the model yeast prion against which others are often compared [44] , but the behavior of a yeast prion in vivo can depend on multiple factors . Distinct prion ‘variants’ can result from the prion-forming protein's assumption of multiple amyloid conformations , which have distinct heritable characteristics [5] [45] , [46] , [47] , [48] . Prion variants are typically classified as “strong” or “weak” based on phenotypic strength and mitotic stability , and often such variants have distinct chaperone requirements [42] , [49] [47] , [50] . Indeed , multiple variants of [PSI+] are well-studied and described in the literature [5] [45] , [46] , [49] . Likewise , yeast genetic background can also affect the results of prion-chaperone experiments [18] [42] . Here we explore the requirement of Sis1 domains in the maintenance of [PSI+] , with special care to explore the potential effects of yeast genetic background and yeast prion structural variability . In a previous investigation , one of us ( JKH ) found that both the weak and strong [PSI+] variants [PSI+]Sc4 and [PSI+]Sc37 could be maintained by Sis1-ΔG/F , demonstrating that [PSI+] does not share the absolute requirement for Sis1's G/F domain with [RNQ+] [24] . In a separate investigation , Kirkland et al . found that the regions of Sis1 known to be necessary to support cell viability ( minimally Sis1-121 ) are also sufficient to support the propagation of a strong [PSI+] variant [43] . To further investigate the ability of Sis1 to support [PSI+] propagation , and to determine the influence of prion-variant structure and yeast genetic background on this model system , we first investigated the Sis1 requirements of well-characterized strong and weak [PSI+] variants using yeast plasmid shuffling . Utilizing two distinct yeast genetic backgrounds , this approach , in combination with biochemical assays , revealed that Sis1 requirements are consistent between yeast genetic backgrounds and consistent among prion variants of similar mitotic stability ( ‘strength’ ) but differ greatly between weak and strong variants . Sis1-ΔG/F , which cannot maintain the prion [RNQ+] , maintains all variants of [PSI+] whereas Sis1-121 , which supports [RNQ+] propagation , cannot support weak [PSI+] variants . Likewise the human homolog of Sis1 , Hdj1 , which supports [RNQ+] , is here shown to be capable of strong but not weak [PSI+] variant propagation . This mutually exclusive set of chaperone requirements for by [RNQ+] and weak [PSI+] cannot be rectified by positing that the two prions simply have different levels of stringency for a singular Sis1 function . Rather , these data indicate that Sis1 must have at least two distinct functions in yeast prion maintenance which are prion specific and allow for J-protein-dependent prion selection .
To begin to examine the impact of amyloid variation on Sis1-domain requirements , we first determined the ability of two commonly used Sis1 protein constructs lacking key domains ( Sis1-121 and Sis1-ΔG/F ) to propagate the well-studied strong [PSI+] variant [PSI+]Sc4 by yeast plasmid shuffling [49] [46] . The plasmid shuffling procedure was the same as previously described ( see Materials and Methods ) [42] [25] . All strains remained [PSI+] as indicated by white/pink colony color as compared to the [PSI+] parent-strain ( pink ) and cured [psi−] strain controls ( red ) ( Figure 2A ) . These results confirm the combined observations of Higurashi et al . and Kirkland et al . that strong [PSI+] can be maintained by either Sis1-121 or Sis1-ΔG/F [24] [43] . To next determine if Sis1 domain requirements are altered when the [PSI+] prion is an alternate amyloid structure , i . e . , an alternate prion variant , we next examined the Sis1 requirements of [PSI+]Sc37 , a well-characterized weak variant ( Figure 2B ) [49] [46] . [PSI+]Sc37 was maintained by the Sis1-ΔG/F construct , but , in contrast to the stronger variant , the prion was lost when Sis1-121 is the only version of Sis1 expressed in the cell . These results are surprising for two reasons: first , they are in direct opposition to the requirements for Sis1 by [RNQ+] , that is , [RNQ+] is maintained by Sis1-121 but not Sis1-ΔG/F and second , because [PSI+] has been previously thought to less-stringently require Sis1 activity for prion propagation when compared to [RNQ+] and other prions [18] [25] , [41] , [42] . Previous rationalizations regarding the distinctions between [PSI+] and [RNQ+] in terms of Sis1 requirement based on Sis1-repression experiments have posited that [PSI+] requires less Sis1 activity than [RNQ+] [24] . The ability of Sis1-121 , but not Sis1-ΔG/F , to maintain [RNQ+] implies that the former construct is somehow more active than the latter . If valid , these new observations regarding [PSI+]Sc37 would negate this model , as the two prions , [RNQ+] and [PSI+]Sc37 , appear to have mutually exclusive requirements of Sis1 . To confirm that prion loss is due to the inability of Sis1-121 to support the prion , rather than due to a low level of protein expression , we next subjected these strains to immunoblotting using a Sis1 antibody which recognizes all three constructs ( Figure 2C ) [42] . Sis1-121 protein is expressed to level which is equal to or greater than the wild-type protein , indicating that the loss of [PSI+]Sc37 in these cells was not due to abnormally low Sis1-121 protein expression . Sis1 expression is subject to tight auto-regulation , even when expressed from an exogenous promoter , making it difficult to significantly over-express in yeast [42] [51] . In an effort to produce a higher level of Sis1 expression , we included in our experiment a 2μ plasmid expressing Sis1-121 ( p324SIS1-sis1-121 ) that produces slightly higher levels of expression than our CEN plasmid ( p314SIS1-sis1-121 ) [42] . Results for this plasmid were the same , Sis1-121 supported the strong [PSI+] variant [PSI+]Sc4 but not the weak variant [PSI+]Sc37 ( Figure 2A and 2B ) despite being expressed at a modestly higher level when compared to the loading control ( Figure 2C ) . To confirm that colony color is accurately reporting prion maintenance or loss in our strains , we next verified our results using an additional biochemical assay , semi-denaturing detergent agarose gel electrophoresis ( SDDAGE ) , in which large detergent-resistant aggregates may be resolved using an agarose gel and then visualized by immunoblotting [52] . In all cases SDDAGE analysis confirmed our colony color observations: Sis1-ΔG/F maintained both variants while only the [PSI+]Sc37 variant was lost when Sis1-121 is the sole form of Sis1 expressed ( Figure 2D ) . Additionally , no drastic aggregate-size shifts were apparent in these samples . We next considered whether this unusual pattern of Sis1-domain requirement was specific to the individual weak [PSI+] variant examined . In a prior investigation in which wild-type Sis1 expression was chemically repressed , large variations in curing rates were found between weak [PSI+] variants , including [PSI+]Sc37 , in the W303 genetic background , indicating that there are differences between weak variants with regard to their interactions with Sis1 [42] . To determine whether the Sis1 requirements of [PSI+]Sc37 are specific to this variant only , or are shared by other weak [PSI+] variants , we expanded our investigation to include additional strong and weak variants of [PSI+] in the W303 genetic background using the same plasmid shuffling approach described above . Specifically , the Sis1 domain requirements for three additional strong variants ( [PSI+]STR , [PSI+]VH , and [PSI+]93S ) and one additional weak variant ( [PSI+]VL ) were examined [42] . Colony color assays ( Figure 3A ) as well as SDDAGE analyses ( Figure 3B ) indicated that all three strong variants were maintained by all Sis1 constructs examined , whereas [PSI+]VL was lost specifically when Sis1-121 was the sole Sis1 construct expressed . These results support those already obtained here for strong and weak variants [PSI+]Sc4 and [PSI+]Sc37 , respectively , and indicate that the minimal requirements for Sis1 function of strong variant [PSI+]Sc4 , and the unusual requirement for Sis1 of the weak variant [PSI+]Sc37 , are not peculiarities of these two particular variants , nor are they dictated by subtle difference in variants . Rather , they appear to be primarily determined by prion-variant strength , a property that arises from gross differences in amyloid core structure . Finally , we also considered whether the maintenance of weak [PSI+] variants by Sis1-ΔG/F here could be due to the unexpected expression of wild-type Sis1 in these strains , as homologous recombination rates are high S . cerevisiae and cross-over may occasionally occur during plasmid-shuffling during the period when both full-length and variant copies of Sis1 are present within a given cell . Immunoblotting with a Sis1 antibody confirmed that only Sis1-ΔG/F , not full-length Sis1 , is expressed in these samples ( Figure S1 ) . As noted in the introduction above , a common limitation of investigations in S . cerevisiae is that , for practical purposes , observations are rarely confirmed in more than one yeast genetic background , leaving open the possibility that polymorphisms of a particular yeast strain may affect the experimental outcomes and interpretations [42] [18] , [44] . Indeed , incongruencies in observations of prion-chaperone interactions have been attributable to yeast strain variations in the past [18] [42] . To directly address this issue , we took advantage of a series of yeast strains used in a previous investigation which uncovered peculiar distinctions in the behavior of weak prion variants upon Sis1 repression between two different genetic backgrounds , W303 and 74D-694 . These distinctions indicate that some still unidentified factors which differ between these two genetic backgrounds affect prion behavior in vivo [42] . To ensure that any results are not due to a peculiarity of the W303 yeast genetic background we reconstructed all of our [PSI+] Sis1-plasmid shuffling strains in the 74D-694 genetic background and reexamined the Sis1 domain requirements for all of the variants described above . The results were summarily consistent with those obtained in the W303 background ( Figure S2 ) , reaffirming our initial observations and confirming that the unusual Sis1 requirements exhibited by [PSI+]Sc37 and other weak variants are not due to any specific factor of the W303 genetic background , but are rather determined primarily by the strength of the [PSI+] variant . [PSI+] cells exhibit a slow growth phenotype when Sis1 expression is chemically repressed ( unpublished observations ) or when Sis1 is ectopically expressed in the form of C-terminal truncation mutants , a phenomenon which has been interpreted as a [PSI+]-dependent toxicity against which Sis1 protects cells [43] . As such , the appearance of [psi−] cells in the experiments described herein could be explained either as the inability of a particular Sis1 construct to support prion propagation , or as the result of an induced selection for [psi−] cells as [PSI+] cells become sick . In a previous investigation , others demonstrated that cytotoxicity is not due to reductions in Sup35 , Sup45 or decreased Sis1 levels in the soluble fraction and suggested that toxicity is correlated to prion propagon number because toxicity diminishes as propagon number is decreased during Hsp104 inhibition by GdnHCl [43] . Indeed , the phenotype appears to be specific to [PSI+] , the yeast prion with the highest known number of heritable prion propagons per cell [44] [24] , [25] . If true , then the patterns of prion curing that we have observed in this investigation are inconsistent with a [psi−] cell selection model , since prion curing has occurred only in strains bearing weak [PSI+] variants that are known to have fewer prion propagons than strong variants [5] [24] , [46] . To further explore this phenomenon and to clarify our interpretations , we investigated whether strong or weak [PSI+] variants which are known to differ in prion propagon number exhibit differential cytotoxicity upon Sis1 repression . To do this , we first set out to estimate the relative propagon numbers of the variants used in this study by conducting a propagon counting assay [53] . Each variant was tested in quadruplicate in the W303 genetic background . All four strong [PSI+] variants produced curing data which was overlapping and fit a model with approximately 300 propagons/cell ( Figure S3 ) ; as such , we were unable to distinguish between these variants on the basis of these data , consistent with previous observations that distinct strong [PSI+] variants are cured with virtually identical kinetics upon Sis1 repression indicating that strong [PSI+] variants may have converged on a similar , and perhaps optimum , amyloid structure for stable propagation in yeast [42] . In contrast , [PSI+]Sc37 and [PSI+]VL curing data produced estimates of 90 and 75 propagons/cell , respectively . These numbers are in general agreement with previous estimates made for [PSI+]STR , and [PSI+]Sc37 using the same methods and genetic background [24] , although it is worth noting that this method , while useful for drawing comparisons among prions , likely systematically underestimates the actual number of heritable propagons [54] . Next , we utilized a set of tetracycline-repressible strains which have been used previously to study Sis1•[PSI+] genetic interactions [24] [42] . These strains have SIS1 under the control of the tetracycline repressible ( TETr ) promoter ( sis1-Δ::LEU2 [TETrSIS1] ) . Following the addition of the tetracycline analog doxycycline , Sis1 expression is reduced , leading to eventual prion curing [42] [24] . Four W303 strains , each bearing a different [PSI+] variant ( [PSI+]Sc4 , [PSI+]Sc37 , [PSI+]VH , or [PSI+]VL ) were cultured in log phase in rich media with similar growth rates , measured as averages of periods of approximately 8–10 generations , of 1 . 4–1 . 7 hrs/gen . Following the addition of doxycycline , all four strains experienced a slowing of growth rate , consistent with both reduced Sis1 levels and [PSI+]-dependent toxicity . However , as predicted , the decline in growth rate was more dramatic for both of the strong [PSI+] bearing strains , declining to 4 . 5 hrs . /gen . for [PSI+]Sc4 and 3 . 1 hrs . /gen . for [PSI+]VH before beginning to recover slightly , than for the weak variant-bearing strains ( slowest observed rates were 2 . 1 hrs . /gen . and 2 . 2 hrs . /gen . for [PSI+]Sc37 and [PSI+]VL , respectively ) . Cured versions were also examined as a control and exhibited no differences in growth rate among the four strains , indicating that these discrepancies are indeed caused by differences between strong and weak prion variants . These observations confirm the hypothesis forwarded by Kirkland et al . that [PSI+] induced toxicity correlates with propagon number [43] . They also support the conclusion that the curing of weak [PSI+] variants but not strong variants in this study is not due to cell selection by [PSI+]-dependent cytotoxicity , but rather , by an inability of these chaperone constructs to support the propagation of the prion . The human protein Hdj1 ( DNAJB1 ) shares >36% of residue identities with Sis1 ( ALIGN ) , indicating that it is a Sis1 ortholog and that the two proteins likely share similar overall folds [55] . Indeed , Hdj1 can substitute for Sis1 in maintaining yeast cell viability with no obvious phenotypic differences , and can support the maintenance of the [RNQ+] prion in cells otherwise lacking Sis1 expression [41] . Another investigation found that Hdj1 was incapable of supporting [PSI+] , though this investigation examined only one strong variant and experiments were conducted in a different yeast genetic background than utilized previously for [RNQ+] [43] . In order to more broadly examine the ability of Hdj1 to maintain prions in yeast , we tested Hdj1 expression under the control of the GPD promoter from a 2μ plasmid in all of our aforementioned Sis1-plasmid shuffling strains . Surprisingly , we found that Hdj1 was able to support all strong variants of [PSI+] examined , in both the W303 or 74D-694 genetic backgrounds , as confirmed by both colony color and SDDAGE ( Figure 4A ) . In contrast , Hdj1 was unable to maintain any weak variant of [PSI+] , similar to the pattern of [PSI+] prion maintenance exhibited by Sis1-121 . Notably , some size shifts in aggregate bands were apparent for strong [PSI+] variants ( Figure 4A ) . Prion aggregates maintained by Sis1 migrated further into the gel than those maintained by Hdj1 , indicating that Hdj1 expressing cells have a distribution of larger prion aggregates as compared to the control strain . These size-shifts are consistent with a reduction in prion fragmentation in the Hdj1 samples , indicating that while Hdj1 is sufficient to replace Sis1 , it may be less efficient than Sis1 in accomplishing this function . Notably , similar shifts are observable in some samples expressing Sis1-121 in Figures 3B and S2 . To assure that the maintenance of these strong [PSI+] variants is due to Hdj1 expression only , we also examined these cells for Sis1 expression ( Figure S4 ) . A strong band for Sis1 is apparent in the control strain , but no comparable bands appear in any of the Hdj1-expressing strains maintaining strong [PSI+] variants . To confirm that the loss of weak [PSI+] variants in Hdj1-expressing cells is not due to differential protein expression in these strains , we next examined Hdj1 expression directly using a commercially available Hdj1 antibody . No band of the appropriate size was found in control cells lacking the Hdj1 expression plasmid , indicating that the antibody specifically recognizes Hdj1 and not Sis1 ( Figure 4B , lane 3 ) Using this antibody , immunoblots of strains which lost the weak [PSI+] variants and similar strains maintaining strong [PSI+] variants indicated that Hdj1 expression levels are similar across all the strains examined ( Figure 4B ) . Taken together , these results demonstrate that Hdj1 is capable of substituting for Sis1 in the maintenance of [PSI+] in a prion-variant dependent manner that preferentially maintains strong , but not weak , [PSI+] variants . The observation that weak [PSI+] variants are maintained by Sis1-ΔG/F but not Sis1-121 or Hdj1 are of particular interest because they contradict the known Sis1 domain requirements of [RNQ+] , a prion which , until now , was considered to be more sensitive to Sis1 expression than either weak or strong [PSI+] [24] , [43] . An alternate explanation for the discrepancy between the Sis1 domain requirements between [RNQ+] and weak [PSI+] variants is that despite being examined in cells of the same genetic background , these two prions are not being examined within the same yeast cells , leaving open the possibility that an unanticipated polymorphism between our [RNQ+] and [PSI+]Sc37 tester strains , or an unrecognized difference in experimental conditions , is responsible for the disparate results . To address this issue , and to directly and unambiguously compare the requirements for these two prions , we mated our [PSI+]Sc37 strain to an otherwise isogenic [RNQ+] strain of opposite mating type in the W303 genetic background to create a new diploid strain that possessed both prions . Following sporulation and selection for the sis1::LEU2 allele and the [SIS1-Sis1 , URA3] plasmid , we isolated a [RNQ+]/[PSI+]Sc37 haploid strain again suitable for Sis1 plasmid shuffling . Following transformation of this new strain with our Sis1 and Hdj1 expression constructs and subsequent loss of the URA3-marked plasmid , confirmed again by both uracil auxotrophy and immunoblotting , the continued presence of [PSI+]Sc37 was again monitored by color assay while the presence of [RNQ+] was determined by fluorescence microscopy following a second transformation by a plasmid expressing an Rnq1-GFP chimera . [RNQ+] cells expressing Rnq1-GFP exhibit heterogeneous ( punctate ) fluorescence patterns as the fluorescent chimera is recruited into preexisting prion aggregates [23] . In a [rnq−] cell the fluorescence is homogenously distributed about the cytoplasm ( diffuse fluorescence ) [23] . [PSI+]Sc37 was again maintained by Sis1-ΔG/F but lost in the presence of either Sis1-121 or Hdj1 while results for [RNQ+] matched those previously reported in the literature as expected: the prion was maintained by Sis1-121 and Hdj1 but lost in the presence of only Sis1-ΔG/F ( Figure 5A ) [41] . We again employed SDDAGE to confirm that both our color assay and GFP assay accurately report the aggregation states of the respective prion proteins and to examine any changes in aggregate size . Notably , SDDAGE analysis revealed that [RNQ+] aggregates are larger in cells expressing only Sis1-121 or Hdj1 , similar to the results observed for some strong [PSI+] variants . More important however , for both prions , SDDAGE analyses unambiguously confirmed that the prions are lost or maintained in a reciprocal manner even when assayed in the same yeast cells ( Figure 5B ) demonstrating that these two prions have mutually exclusive requirements for Sis1 functions . The results described above reaffirm that [RNQ+] and strong [PSI+] variants can be maintained minimally by co-expression ( in cis ) of only Sis1's J-domain and G/F regions ( Sis1-121 ) , yet this construct is insufficient for weak [PSI+] propagation . Notably , this is the first time that a construct of Sis1 has been identified that separates the maintenance of cell viability from [PSI+] prion propagation . The observation that Sis1-ΔG/F supports all variants of [PSI+] examined herein raises the question: What is the minimum Sis1 construct necessary for weak [PSI+] propagation ? Because of the apparent importance of Sis1's two glycine-rich regions and because Sis1-ΔG/F retains only G/M region , we speculated that this region may be of particular importance in the maintenance of weak [PSI+] variants . To test this hypothesis , we first constructed a new expression construct for Sis1 which lacks the G/F region and ends with residue 171 ( Sis1-171ΔG/F ) ; this minimal construct consisted of only the J-domain of Sis1 and the G/M region ( Figure 1 ) . Although both Sis1-121 and Sis1-ΔG/F have been used as alternative minimal constructs for both prion maintenance and cell viability in the past , the minimal regions of Sis1 that are sufficient to support cell viability in the absence of the G/F region are also not known , so we first needed to confirm that this construct supports yeast cell viability independent of prion propagation . To avoid potential complications with prion-associated cytotoxicity , we first utilized cured ( [psi−] ) versions of our plasmid shuffling strains . Despite successful isolation of transformants and repeated attempts at plasmid shuffling , no colonies formed on 5-FOA media indicating that this construct is unable to maintain cell viability as the sole form of Sis1 . Because minimal J-domain expression constructs have been found to be active when expressed with a short , C-terminal trailer sequence like an HA-tag [56] , we also investigated whether constructs expressing a few amino acids following the G/M region might be more active . We next constructed two new expression constructs for Sis1 , again lacking the G/F region , but ending at either residue 171 or 206 and containing a random seven amino acid trailer sequence ( Sis1-206ΔG/F* and Sis1-171ΔG/F* , Figure 1 ) . Again , to avoid potential complications with prion-associated cytotoxicity , we first checked whether these constructs could support cell viability in cured versions of our plasmid shuffling strains . Following transformation , small numbers of slow-growing colonies formed on 5-FOA media indicating a successful replacement of the wild-type Sis1 plasmid and that both of the new Sis1 truncation constructs Sis1-206ΔG/F* and Sis1-171ΔG/F* maintain cell viability . These observations demonstrated that the critical functions of Sis1 may be accomplished by solely the expression of the J-domain in cis with the G/M region , as long as this region is not at the extreme C-terminus of the polypeptide . This result also confirms previous observations that the G/F and G/M domains have redundant functionality that is required for cell viability [40] [41] , but now clarify that this function does not require co-expression of the C-terminal peptide binding or dimerization domains . Notably , we did not isolate many colonies on 5-FOA especially for Sis1-171ΔG/F* , indicating that , as one might expect , these minimal constructs are lacking when compared to full-length Sis1 and so it is difficult to isolate cells which preferentially lose the plasmid bearing the full-length construct . We next examined whether these same constructs could support [PSI+] prion propagation by retransforming all of our prion tester strains . Interestingly , despite repeated attempts , we were able to isolate only a few viable colonies from 5-FOA media for any strains bearing strong [PSI+] variants; however , SDSPAGE and immunoblotting confirmed that these colonies were still expressing full-length Sis1 , despite the loss of the URA3-marked plasmid , indicating that a homologous recombination event had occurred during plasmid shuffling . In contrast , we were successful in isolating strains bearing the weak [PSI+] variant [PSI+]Sc37 , though similar challenges with low viability and frequent cross-over events were also observed . Figure 6 illustrates the results for these strains: both the Sis1-206ΔG/F* and Sis1-171ΔG/F* constructs maintained the pink colony color phenotype indicative of [PSI+] as compared to parental-strain and cured-strain controls but with a notable change in color ( Figure 6A ) . Likewise , SDDAGE analysis indicated prion maintenance , albeit with a severe shift in aggregate size toward higher molecular weights ( Figure 6B ) . Immunoblotting with a Sis1 antibody confirmed that , unlike many other cells we isolated , these cells lacked full-length Sis1 expression ( Figure 6C ) . Aggregate size and colony color can be causally related , that is , weaker [PSI+] variants which are more difficult to fragment , tend to exhibit darker colony colors and larger prion aggregates than strong variants due to the decreased ability of large aggregates to sequester soluble Sup35 . As such , both the dark color of the colonies and the high molecular weight bands observed here indicate that [PSI+]Sc37 prion fragmentation is likely impaired in Sis1-206ΔG/F* or Sis1-171ΔG/F* expressing cells . Despite these apparent deficiencies in function , the observation that [PSI+]Sc37 can be maintained at all by Sis1-171ΔG/F* in particular demonstrates that only the J-domain and G/M regions of Sis1 are fundamentally necessary for Sis1's function in weak [PSI+] prion propagation and establish Sis1-171ΔG/F* as the new minimum construct for weak [PSI+] maintenance for future experimentation .
In our analyses of Sis1 domain requirements by multiple [PSI+] variants , we found a tremendous amount of consistency both between yeast genetic backgrounds and among prion variants of similar phenotypic strength . In contrast , Sis1 requirements differed greatly between so called ‘weak’ and ‘strong’ variants of [PSI+] , raising the question as to why these variants might exhibit such distinctions . Yeast prion variants are typically defined by phenotypic differences that arise from differences in the structure of the amyloid core [5] [45] , [46] , [57] . For two of the prions used here , [PSI+]Sc4 and [PSI+]Sc37 , the relationship between phenotypic strength and amyloid structure is well-understood . [PSI+]Sc37 , the weaker variant , has a more extensive amyloid core with a greater number of residues involved and therefore a greater number of hydrogen bonds stabilizing the cross-beta structure [58] . As a result , [PSI+]Sc37 aggregates are less easily fragmented in vitro , a characteristic that results in fewer heritable propagons per cell and a weaker prion phenotype in vivo [49] [46] , [58] . It is reasonable to suggest , then , on the basis of our observations , that weak variants of [PSI+] may require additional Sis1 intervention in order to allow prion fragmentation to keep up with cell division . Stein and True have recently shown that [RNQ+] variants differ in Sis1 association suggesting that distinct prion variants may expose different regions of the protein to solvent for chaperone interactions [59] . Indeed , recent structural analyses by Frederick et al . have revealed that [PSI+]Sc4 and [PSI+]Sc37 fibers also differ in their interaction with Hsp104 , possibly by virtue of a difference in the mobility of residues in the middle ( M ) domain of Sup35 . That weak fibers bind better to Hsp104 is seemingly paradoxical , however as suggested in that investigation , this binding maybe non-functional and thereby obscure functional Hsp104 binding sites . It is reasonable to suggest , on the basis of our data , that these non-functional interactions may rather obscure Sis1 binding , which normally leads to the recruitment of Ssa and Hsp104 at partially unfolded sections of the protein . A combination of differences in structure of the prion aggregates , mobility of residues , and crowding by Hsp104 then would give rise to distinct binding interactions between Sis1 and weak [PSI+] variants which might require specific portions of Sis1 . Determining whether the removal of various domains impairs Sis1-binding per se , or the enabling of fragmentation after binding , will require additional co-aggregation experiments to distinguish . Strong [PSI+] , on the other hand , appears to require very little from Sis1 by comparison . Despite now two direct investigations into the Sis1 domain requirements of strong [PSI+] , no single construct of Sis1 has been identified which can separate Sis1's role in maintaining cell viability from its role in strong [PSI+] maintenance . These observations alone suggest that either Sis1 has more than one role in prion fragmentation , or that the distinctions between weak and strong variants are simply a matter of stringency in the requirement of a singular Sis1 function . Indeed , on the basis of prior data , an identical argument has been made to explain the discrepancies between the requirements for Sis1 between the prions [PSI+] and [RNQ+] , however , as our data now suggest , the additional activities required by weak [PSI+] variants and [RNQ+] must be somehow distinct from one another . As noted in the Results section , the patterns of prion loss that we observed are inconsistent with those that would be predicted if cytotoxicity was the driving factor in prion curing since curing occurred only in weak [PSI+] variants with low propagon numbers . Likewise , the low propagon number of weak [PSI+] variants is also insufficient alone to explain our results because another prion [SWI+] , which has even fewer propagons/cell than [PSI+]Sc37 , was found to be maintained by the same Sis1-121 expressing plasmid and in the same 74D-694 yeast strain during a previous investigation [25] . Additionally , the curing of [PSI+] by Sis1 depletions was previously shown to be independent of Hsp104 overexpression [24] . Rather , the data presented here support the idea that Sis1's glycine-rich domains impart at least two distinct functionalities to the protein which prions require differentially . Specifically , the prions [RNQ+] and [PSI+]Sc37 can be selectively supported or lost in a reciprocal manner by replacement of wild-type Sis1 with a construct expressing only the J-domain and either the G/F region ( Sis1-121 ) or the G/M region ( Sis1-171ΔG/F* ) , respectively . Do these regions impart prion-specific functions to Sis1 ? In the case of [RNQ+] and the G/F region , the answer appears to be ‘yes’ , that is , to date no construct of Sis1 which lacks the G/F region has been found to support [RNQ+] indicating that the G/F region is both necessary and sufficient in combination with the J-domain . However , with respect to weak [PSI+] maintenance and the G/M region , the situation is complicated by functional overlap between the two regions that was first revealed in the context of cell viability; viability may be maintained by expression of only the J-domain and either glycine-rich region . While a construct consisting of only the J-domain and G/M region is sufficient to maintain weak [PSI+] , the construct Sis1-ΔG/M is also able to maintain all variants of [PSI+] examined , indicating that there is also some functional overlap between the G/F and G/M domains in the context of prion maintenance but that the function ( s ) of these regions depends upon the context of the glycine-rich domain within the polypeptide that is expressed . It is interesting to note that this functional overlap appears to hierarchical , at least in this one instance; that is , the G/F region , in the context of the Sis1-ΔG/M construct , can substitute for the G/M region in weak [PSI+] maintenance , but the reverse is not true for [RNQ+] , that is , G/M region is unable to substitute and support [RNQ+] in the context of the Sis1-ΔG/F construct . Notably , both minimal constructs produced noticeable increases in aggregate size in the respective prions they support , observations which would be consistent with either construct creating a small but noticeable defect in the efficiency of prion fragmentation . However , these size shifts were absent in cells bearing the longer constructs Sis1-ΔG/F and Sis1-ΔG/M which differ from the minimal constructs only by the addition of Sis1's C-terminal domains ( CTD1/2 and the dimerization domain ( DD ) , Figure 1 ) . This observation is interesting because , giving the ability of shorter constructs to maintain prions , the C-terminal domains of Sis1 are generally regarded as unimportant for prion maintenance . These observations indicate that the addition of the C-terminal domains to each respective minimal construct creates an observable change in aggregate size consistent with an increase the overall fragmentation of the prion aggregates which is similar to the full-length protein . This effect is not reproduced when only the first 35 residues of CTD1 are added back as in the Sis1-206ΔG/F* construct , demonstrating that is not simply a function of having additional amino acids at the C-terminal end of the glycine-rich domains . Nor does this effect appear to be due to expression issues , as at least for Sis1-121 , protein levels are at least as great as the wild-type protein ( Figure 2C ) . Sis1 is known to cycle in and out of the nucleus as part of spatial protein quality control and cytosolic misfolded protein-targeting for degradation by nuclear proteasomes [37] [36] . It is possible that our observations are largely affected by alterations in Sis1 localization then , particularly if Sis1 is sequestered to the nucleus when our minimal constructs are expressed . Intriguingly , one investigation found that movement of Sis1 into the nucleus was fully dependent upon its interaction with the sorting factors Btn2 and Cur1 , which required the expression of Sis1 dimerization domain ( DD , Figure 1 ) , but not a functional J-domain or expression of CTD1/2 [37] . Considering that our minimal constructs ( Sis1-121 and Sis1-171ΔG/F* ) lack the dimerization domain while longer constructs ( Sis1-ΔG/F and Sis1-G/M ) maintain it , is highly unlikely that either the defects in fragmentation activity , or the distinctions between prions , revealed here are due to sequestration of Sis1 minimal constructs into the nucleus . Indeed , our experimental observations that single constructs have reciprocal effects on two prions expressed in the same cells further supports that assertion . Future Sis1 localization and co-aggregation experiments will help to further clarify not only this issue , but will also address the unanswered question of whether various mutant Sis1 constructs are deficient in prion-aggregate binding , or are competent for binding but fail to stimulate prion fragmentation . Additionally , sucrose gradient sedimentation may reveal unresolved changes in native aggregate size which may differ from changes in SDS-resistant aggregate size , and could lead to deficiencies in prion transmission to daughter cells . Regardless , these observations taken together indicate that while not essential , the C-terminal domains of Sis1 do contribute significantly to the ability of Sis1 to facilitate prion fragmentation , for both [RNQ+] and [PSI+] . Our finding that the human homolog of Sis1 , Hdj1 , supports strong [PSI+] strains conflicts with the observations of Kirkland et al . who found that a strong [PSI+] variant was lost when Hdj1 was expressed in the absence of Sis1 [43] . These contradictory observations could be due to a difference in the yeast strain used , the specific prion variant examined , or due to a difference in amount of Hdj1 expression achieved in the experimental setup . The congruency of our other observations regarding strong [PSI+] variants , both between variants in this study and with the observations made in that study make it unlikely that a difference in prion variant is to blame . The most likely reason for the discrepancy is that Hdj1 expression was driven by an exogenous promoter from a multicopy plasmid in our experiments , however we cannot , at present , rule out the possibility of an uncharacterized polymorphism between lab strains that may exist which effects prion-chaperone experimental results as we have observed similar phenomena in the past [42] . These conflicting observations underscore the benefit of confirming findings , when possible , in more than one genetic background and/or prion variant to control for strain- or variant-specific phenomena as well as protein expression levels . One possible explanation for our observations regarding Hdj1's inability to support weak [PSI+] is that perhaps Hdj1 is less active than Sis1 in prion fragmentation and the weak [PSI+] variants examined here are simply more generally sensitive to reductions in Sis1 activity than [RNQ+] . One means of testing this hypothesis would be examine the curing rates of these variants and [RNQ+] upon Sis1 repression . Serendipitously , curing of both weak [PSI+] variants examined here and [RNQ+] have been examined within the same yeast genetic background ( 74D-694 ) under identical conditions during two previous investigations and exhibit typical sigmoidal curing curves upon Sis1 repression which may be compared by estimating the mid-point of curing , that is , the number of generations at which ∼50% of the cell population has been cured [42] [25] . In the case of the two weak [PSI+] variants examined here , [PSI+]Sc37 and [PSI+]VL , ∼50% curing was attained at 17 and 22 generations , respectively [42] . Under identical conditions , the ∼50% curing mark for [RNQ+] occurred at only 13 generations [25] , a curing rate for [RNQ+] which is consistent with previous estimations made in W303 background from GFP-counting data [24] [23] . Taken together , these observations indicate that [RNQ+] is at least comparably sensitive to general reductions in Sis1 activity as either [PSI+]VL and [PSI+]Sc37 , if not more sensitive than both . Therefore , it is unlikely that Hdj1 maintains [RNQ+] but not [PSI+]Sc37 and [PSI+]VL simply due to differences in sensitivity to generic Sis1 activity , but rather suggest that Hdj1 lacks a distinct functionality of Sis1 that is specifically required by these weak variants . Finally , we found that Hdj1 behaved similarly in our assays to the construct bearing only the J-domain and G/F region of Sis1 ( Sis1-121 ) , indicating that perhaps these regions are better conserved in the human protein . We have also observed that Hdj1 is incapable of substituting for Sis1 in the curing of [PSI+] by Hsp104 overexpression , a phenomenon which also requires an unknown Sis1 function ( Hines J . K . , unpublished observations ) . It is plausible that the inability of Hdj1 to fully substitute for Sis1 in some biological functions stems from an inability to correctly partner with the yeast Hsp70s . Additional experiments co-expressing both Hdj1 and human Hsp70 will be necessary to further clarify the interpretation of these findings . Are Sis1 and Hdj1 amyloid recognition factors ? Several lines of evidence support this hypothesis . Sis1 is required for the propagation of all four yeast prions for which there is data ( [PSI+] , [RNQ+] , [URE3] , and [SWI+] ) , and because it likely acts upstream of both Hsp70 and Hsp104 , it is positioned to potentially be the first responding protein to direct chaperone activity toward amyloids [23] [21] , [24] , [25] . Sis1 is also found directly associated with other Q/N-rich proteins and polyglutamine aggregates in addition to yeast prions [60] [18] , [36] , [41] , [61] , and , perhaps most telling , a recent report revealed that Sis1 alone can ‘direct’ bacterial chaperones to maintain yeast prions [62] . If Sis1's role is indeed to ‘recognize’ amyloids in vivo , then understanding this functionality at the biochemical level would be of great interest . Perhaps even more intriguing would be to understand how one or the other of two short glycine-rich regions not only imparts Sis1's J-domain with the ability to maintain prions but imparts prion-specific maintenance . Additional work utilizing these new minimal constructs in combination with new insights about the human homolog will likely shed new light on this protein mystery in the near future .
Haploid Saccharomyces cerevisiae W303 and 74D-694 derived strains were used throughout . To create [PSI+] strains competent for Sis1-plasmid shuffling , yeast strains bearing distinct [PSI+] variants from both backgrounds ( [PSI+] [rnq−] [p414-TETr-Sis1] sis1::LEU2 ade1-14 ura3-52 leu2-3 , 112 trp1-289 his3-200 ) that were utilized in previous investigations for Sis1 repression experiments were transformed by a URA3 marked plasmid expressing wild-type Sis1 ( p316-SIS1-Sis1 ) [42] . Transformants were selected on synthetic media lacking uracil and then passaged on synthetic complete media containing 5-fluoroanthranilic acid ( 5-FAA ) which counter-selects against the original TRP1-marked Sis1 expression plasmid ( p414-TETr-Sis1 ) . Strains were tested for [PSI+] maintenance as well as uracil prototrophy and tryptophan auxotrophy prior to plasmid shuffling experiments . Additional W303 and 74D-694 strains bearing the strong variant [PSI+]93S were constructed by yeast lysate transformation in which recipient [psi−] spheroplasts were co-transformed with cell extracts of the donor strain SL1293 ( a gift from Susan Liebman ) and the URA3-bearing Sis1 expression plasmid ( p316-SIS1-Sis1 ) . Transformants were then selected on media lacking uracil and candidate strains patched onto rich media to analyze prion status based on colony color . Prion status was verified by curability with GdnHCl and SDDAGE analysis as described in a subsequent section below . To create a [PSI+]/[RNQ+] strain competent for Sis1-plasmid shuffling , the W303 [PSI+]Sc37 plasmid shuffling strain described above ( [PSI+]Sc37 [p316-SIS1-Sis1] sis1::LEU2 ade1-14 ADE2 ) was mated to strain EAC Y639 ( [RNQ+] ADE1 ade2-1 ) . Diploids were selected by adenine prototrophy and sporulated . Following tetrad dissection , pairs of haploids forming pink colonies from parental ditype tetrads were identified as sis1::LEU2 , ade1-14 , ADE2 , [PSI+] first by colony color and leucine prototrophy and then by the ability to convert to red-forming colonies upon GdnHCl treatment . Candidate haploids were then transformed by the plasmid p413CUP1-RNQ1-GFP and examined by fluorescence microscopy following selection on media lacking uracil . Strains exhibiting punctate fluorescence patterns , characteristic of [RNQ+] , which could be converted to a diffuse pattern upon GdnHCl treatment , were selected . Finally , the presence of both prions simultaneously was confirmed by semi-denaturing detergent agarose gel electrophoresis ( SDDAGE ) as described below in a later section . Plasmids used in this study are based on the pRS series [63] . The gene fragments encoding Sis1-171ΔG/F ( residues 1-70 and 122-171 ) or Sis1-206ΔG/F ( residues 1-70 and 122-206 ) were amplified by polymerase chain reaction ( PCR ) , introducing a 5′ BamHI site and 3′ Sal1 site using plasmid p313-SIS1-Sis1-ΔG/F as the template . Linear insert was then digested with BamHI and Sal1 and ligated ( T4 DNA ligase ) into pre-digested p414-GPD . Introduction of a random C-terminal seven amino acid tag ( VDLESCN ) was accomplished by site-directed mutagenesis PCR ( Quikchange ) . Plasmid p424-GPD-sis1-171ΔG/F was likewise created by PCR amplification of p313-SIS1-sis1-ΔG/F to introduce sites for EcoRI and SpeI , upstream and downstream , respectively , followed by digestion with these enzymes and ligation into precut p424-GPD . All other plasmids , listed in Table 1 , have been described elsewhere . Total protein extracts for SDS-PAGE were prepared by harvesting yeast cells in mid-log phase followed by vortexing in 1 M NaOH at 25°C . Cells were then spun at 13 , 500 rpm on a table-top centrifuge at 25°C and the supernatant was removed . Pellets were resuspended in sample buffer containing SDS and boiled for five minutes before resolving in a 12 . 5% polyacrylamide gel . The protein was transferred to nitrocellulose membrane at 1 A for 1 hour at 25°C in a tris-glycine/methanol buffer and probed with polyclonal antibodies specific to either Sis1 ( a gift from the Craig lab ) or Hdj1/DNAJB1 ( Cayman Chemicals ) . Western ladder from New England Biolabs was used as a marker to detect relative protein sizes . To conduct plasmid shuffling experiments , sis1-Δ [PRION+] cells expressing Sis1 from a URA3-marked plasmid were transformed by plasmids expressing either wild-type Sis1 or a Sis1-mutant protein and ∼10 transformants selected by growth on solid selective media . The action of the gene product of URA3 , the enzyme orotidine-5′-phosphate decarboxylase , converts harmless 5-fluoro-orotic acid ( 5-FOA ) into 5-flourouracil ( 5-FU ) , a chemotherapeutic agent which is toxic to dividing cells through its potent inhibition of thymidylate synthase . Subsequent growth on synthetic media containing 5-FOA counter-selects against the URA3-marked plasmid; only cells that stochastically lose the URA3-marked plasmid form colonies . Complete loss of the URA3-marked plasmid was then further confirmed by uracil auxotrophy . Following an additional passage on selective media to allow additional time for potential prion-loss , shuffled cells ( 6–10 transformants in each experiment ) were examined for the maintenance of the prion by colony color on rich glucose media . Propagon counting assays using GdnHCl and time course experiments utilizing SIS1 under the control of the tetracycline-repressible promoter ( TETr-SIS1 ) were conducted as previously described [24] [53] , [64] . The presence or absence of [PSI+] was confirmed by observation of colony color on glucose-based rich media YEPD ( Teknova ) where [PSI+]-mediated aggregation of Sup35 , a translation termination factor , causes read-through of the premature nonsense codon in the ade1-14 mutant allele [65] , [66] . Strains which are otherwise wild-type for adenine production appear pink or white in the presence of [PSI+] or dark red in the absence of [PSI+] due to the accumulation of a red intermediate when adenine production is blocked [67] . Cells were grown at 22°C for 3–6 days to allow color development prior to imaging . [RNQ+] aggregates in cells were observed directly following transformation by a plasmid expressing Rnq1 fused to green-fluorescent protein ( p416CUP1-RNQ1-GFP ) . [RNQ+] cells can be easily distinguished from [rnq−] cells when examined under a microscope by characteristic punctuate or diffuse fluorescence patterns , respectively [23] . To create [prion−] control strains , prion bearing cells were treated with the Hsp104 inhibitor GdnHCl ( final concentration 4 mM ) and grown in liquid culture with agitation for two days at 30°C to allow adequate cell divisions for prion curing . Semi-denaturing detergent agarose gel electrophoresis ( SDD-AGE ) , a method for resolving detergent resistant aggregates , was used to confirm the presence or absence of both [RNQ+] and [PSI+] and to determine relative aggregate size distributions [24] , [68] . Briefly , cells were lysed using sterile glass beads by vortexing at 4°C . Following centrifugation at 4°C , cleared lysates were mixed with SDS loading buffer and incubated at 25°C for 7 minutes . Aggregates were resolved in a 1 . 5% ( w/v ) Tris-glycine ( 0 . 1% SDS ) agarose gel ( SeaKem Gold PFGE agarose ) and protein was transferred to a nitrocellulose membrane at 1A for 1 hr at 22°C in a tris-glycine/methanol buffer . To visualize aggregates , membranes were blocked with 5% ( w/v ) milk and probed with antibodies specific for either Rnq1 or Sup35 ( gifts from the Craig and Tuite labs , respectively ) .
|
Multiple neurodegenerative disorders such as Alzheimer's , Parkinson's and Creutzfeldt-Jakob disease are associated with the accumulation of fibrous protein aggregates collectively termed ‘amyloid . ’ In the baker's yeast Saccharomyces cerevisiae , multiple proteins form intracellular amyloid aggregates known as yeast prions . Yeast prions minimally require a core set of chaperone proteins for stable propagation in yeast , including the J-protein Sis1 , which appears to be required for the propagation of all yeast prions and functioning similarly in each case . Here we present evidence which challenges the notion of a universal function for Sis1 in prion propagation and asserts instead that Sis1's function in the maintenance of at least two prions , [RNQ+] and [PSI+] , is distinct and mutually exclusive for some prion variants . We also find that the human homolog of Sis1 , called Hdj1 , has retained the ability to support some , but not all yeast prions , indicating a partial conservation of function . Because yeast chaperones have the ability to both bind and fragment amyloids in vivo , further investigations into these prion-specific properties of Sis1 and Hdj1 will likely lead to new insights into the biological management of protein misfolding .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"biochemistry",
"cell",
"biology",
"proteins",
"prions",
"fungal",
"genetics",
"chaperone",
"proteins",
"genetics",
"genetic",
"elements",
"biology",
"and",
"life",
"sciences",
"epigenetics",
"molecular",
"genetics"
] |
2014
|
Functional Diversification of Hsp40: Distinct J-Protein Functional Requirements for Two Prions Allow for Chaperone-Dependent Prion Selection
|
Sapovirus , a member of the Caliciviridae family , is an important cause of acute gastroenteritis in humans and pigs . Currently , the porcine sapovirus ( PSaV ) Cowden strain remains the only cultivable member of the Sapovirus genus . While some caliciviruses are known to utilize carbohydrate receptors for entry and infection , a functional receptor for sapovirus is unknown . To characterize the functional receptor of the Cowden strain of PSaV , we undertook a comprehensive series of protein-ligand biochemical assays in mock and PSaV-infected cell culture and/or piglet intestinal tissue sections . PSaV revealed neither hemagglutination activity with red blood cells from any species nor binding activity to synthetic histo-blood group antigens , indicating that PSaV does not use histo-blood group antigens as receptors . Attachment and infection of PSaV were markedly blocked by sialic acid and Vibrio cholerae neuraminidase ( NA ) , suggesting a role for α2 , 3-linked , α2 , 6-linked or α2 , 8-linked sialic acid in virus attachment . However , viral attachment and infection were only partially inhibited by treatment of cells with sialidase S ( SS ) or Maackia amurensis lectin ( MAL ) , both specific for α2 , 3-linked sialic acid , or Sambucus nigra lectin ( SNL ) , specific for α2 , 6-linked sialic acid . These results indicated that PSaV recognizes both α2 , 3- and α2 , 6-linked sialic acids for viral attachment and infection . Treatment of cells with proteases or with benzyl 4-O-β-D-galactopyranosyl-β-D-glucopyranoside ( benzylGalNAc ) , which inhibits O-linked glycosylation , also reduced virus binding and infection , whereas inhibition of glycolipd synthesis or N-linked glycosylation had no such effect on virus binding or infection . These data suggest PSaV binds to cellular receptors that consist of α2 , 3- and α2 , 6-linked sialic acids on glycoproteins attached via O-linked glycosylation .
Caliciviruses ( family Caliciviridae ) are small ( 27–40 nm ) , non-enveloped , icosahedral viruses that possess a single-strand , plus-sense genomic RNA of 7–8 kb [1] . Caliciviruses are important veterinary and human pathogens which are associated with a broad spectrum of diseases in their respective hosts . A member of the genus Lagovirus , rabbit hemorrhagic disease virus ( RHDV ) , is associated with a fatal liver disease in rabbits [2] . Feline calicivirus ( FCV ) , a member of the genus Vesivirus , causes respiratory disease in cats [3] , [4] . Caliciviruses in the genera Norovirus and Sapovirus are important acute gastroenteritis pathogens in humans and animals [5] , [6] . Each year , human noroviruses cause at least 1 . 1 million episodes and 218 , 000 deaths in developing nations as well as approximately 900 , 000 cases of pediatric gastroenteritis in industrialized nations [7] . Sapoviruses have also been associated with gastroenteritis outbreaks and with disease in pediatric patients [1] . The genus Sapovirus can be divided into five genogroups ( GI–GV ) , among which GI , GII , GIV and GV are known to infect humans , whereas GIII infects porcine species [8] . No fully permissive cell culture system currently exists for the enteric caliciviruses associated with gastroenteritis in humans , hampering the study of viral pathogenesis and immunity of these ubiquitous pathogens [1] . The initial events in a viral infection are induced by binding of the virus to the surface of the host cell , followed by penetration or release of the virus particle into the cytoplasm of the cell . Binding occurs through interactions between the virion and receptors on the plasma membrane of the target cell , and consequently receptors are important determinants of viral tissue tropism and pathogenesis [1] . Among the members of the Calicivirdae family , an attachment factor for RHDV was identified as H-type 2 histo-blood group antigen ( HBGA ) , and this led to further studies identifying factors involved in the attachment of the other members of the family [9] . HBGAs function as the attachment factor of both human and bovine noroviruses [5] , [10] , while sialic acid linked with gangliosides acts as at least part of the murine norovirus ( MNV ) receptor [11] . In addition , Tulane virus , the newly discovered rhesus monkey calicivirus , uses HBGA as a receptor [12] . FCV is reported to recognize terminal sialic acid on an N-linked glycoprotein for attachment [4] , with junctional adhesion molecule-1 ( JAM-1 ) functioning as a receptor to facilitate FCV penetration and infection of host cells [5] , [13] . Although these reports strongly suggest that the recognition of a carbohydrate receptor may be a common feature of many caliciviruses , it is also clear that different caliciviruses recognize different carbohydrate receptors [5] . Importantly , however , virus-like particles of human sapovirus genogroups GI and GV strains are known not to bind salivary HBGAs or synthetic carbohydrates [14] . The close genetic relationship of noroviruses and sapoviruses found in humans and animals has led to concern over the possibility of zoonotic transmission of these viruses [15] . Although animal sapoviruses and noroviruses have not yet been isolated from humans , the detection of antibodies against bovine GIII norovirus and canine GVI norovirus in human serum samples [16] , [17] , and the detection of human-like GII . 4 norovirus in porcine and bovine fecal samples [18] suggest the possible zoonotic transmission of noroviruses . Furthermore , it has been demonstrated that human norovirus GII4-HS66 strain can induce diarrhea and intestinal lesions in the proximal small intestine of gnotobiotic piglets or calves [19] . There is also evidence to suggest that highly conserved receptors across host species can be shared by different viruses , even among different genera or families [10] , [20] , [21] . Such concerns have prompted us to investigate the ability of porcine sapovirus ( PSaV ) to recognize carbohydrates present on cultured cells of porcine origin and small intestinal epithelial cells of piglets . In this study , we demonstrate that PSaV binds to both α2 , 3- and α2 , 6-linked sialic acids present on an O-linked glycoprotein .
Although carbohydrate moieties are known to act as receptors or attachment factors for various caliciviruses [5] , their role as receptors or attachment factors for members of the Sapovirus genus remains unknown . To determine if PSaV Cowden strain requires carbohydrate moieties for binding and infection , we removed the carbohydrate moieties from permissive porcine LLC-PK cells by treatment with sodium periodate ( NaIO4 ) , which is known to cleave carbohydrate groups without altering proteins or membranes [4] , [22] , [23] . Pretreatment of LLC-PK cells with 1 mM or 5 mM NaIO4 markedly reduced the binding of Alexa 594-labeled PSaV Cowden strain compared to mock treated control ( Fig . 1A ) . To quantify the effect of NaIO4 treatment more accurately , LLC-PK cells were pretreated in a similar manner , and were incubated with radio-labeled PSaV Cowden strain . Cells were washed thoroughly , and virus binding was determined by liquid scintillation counting . Binding of PSaV Cowden strain was reduced to 12% of the levels observed in mock treated cells with 1 mM NaIO4 , and to 2% in cells treated with 5 mM NaIO4 ( Fig . 1B ) . The infection rate , as determined by staining cells for the viral antigen VPg , was also significantly reduced; infection rates of 17% and 3% were observed for 1 mM and 5 mM NaIO4 , respectively , when compared with mock-treated cells ( Fig . 1C and 1D ) . A similar degree of inhibition of binding and infection was observed in FCV F9 strain-infected Crandall-Reese feline kidney ( CRFK ) cells that were pretreated with NaIO4 ( Fig . 1B and 1D ) . However , binding and infection of coxsackievirus B3 ( CVB3 ) Nancy strain , which is known to us decay-accelerating factor as a receptor [24] , was not influenced by the pretreatment of HeLa cells with 1 mM or 5 mM NaIO4 ( Fig . 1B and 1D , and Fig . S1 ) . In addition , pretreatment of LLC-PK cells with 5 mM NaIO4 had no effect on binding of MNV-1 strain CW1 or vesicular stomatitis virus glycoprotein ( VSV-G protein ) pseudotyped lentiviruses ( Fig . S1 ) . These data strongly indicated that like FCV F9 strain , PSaV Cowden strain utilizes carbohydrate moieties for binding and infection . HBGAs are complex carbohydrates present on the surfaces of RBCs , and mucosal epithelia , mucin of respiratory , genitourinary , and digestive tract [25] . Among caliciviruses , human and bovine noroviruses , RHDV , and Tulane virus are known to use HBGAs as receptors , resulting in the agglutination of RBCs [5] , [12] . To determine if PSaV Cowden strain agglutinated RBCs , hemagglutination assay ( HA ) was performed by using RBCs that originated from various animal species including pigs , rats , chickens , and humans , which was further classified into ABO and Lewis types . PSaV Cowden strain displayed no hemagglutination activity with RBCs from any species at 4°C or 20°C incubation ( Fig . 2 ) . In contrast , influenza A virus PR8 ( H1N1 ) strain agglutinated human , rat , pig and chicken RBCs . P particles of human norovirus VA387 strain and a VP8* of human rotavirus DS-1 strain confirmed binding to corresponding HBGAs [26] ( Fig . 2 ) . Whilst positive controls showed differing degrees of HA activities against RBCs from various species , the absence of HA activity by PSaV indicated that PSaV Cowden strain may not recognize HBGAs as receptors for its binding and infection of host cells . In order to confirm whether PSaV recognizes HBGAs as receptors , a synthetic HBGA binding assay was conducted [26] , [27] . The descriptions and structures of all the oligosaccharides tested are provided in Table 1 . Consistent with the HA assay , PSaV Cowden strain did not bind to immobilized synthetic oligosaccharides , including the A and H types , both of which are known to be expressed in pigs ( Fig . 3 ) [28] . However , recombinant proteins consisting of the P particles of human norovirus VA387 bound to synthetic oligosaccharides of A type , B type and H type , the P particles of human norovirus VA207 bound to synthetic oligosaccharides Lewis types and H type , and a VP8* of human rotavirus DS-1 bound to synthetic oligosaccharides B type and αGal , respectively ( Fig . 3 ) [5] , [26] . Collectively , these results demonstrated that PSaV Cowden strain does not utilize HBGAs as receptors . Sialic acid is an abundant carbohydrate moiety on the cell surface [25] , which acts as a functional receptor for many viruses , including caliciviruses [4] , [11] . To determine if sialic acid is a functional receptor for the PSaV , PSaV Cowden strain was incubated with various concentrations ( 20–160 mM ) of the sialic acid containing molecule , N-acetyl neuraminic acid ( NANA ) , and was then inoculated with LLC-PK cells . NANA at 20 mM significantly reduced the binding activity of Alexa 594-labeled PSaV Cowden strain , and almost completely inhibited binding at 80 mM ( Fig . 4A ) . Binding of radio-labeled PSaV Cowden strain or FCV F9 strain was reduced by NANA in a dose-dependent manner , and was almost completely abolished at 80 mM NANA for PSaV Cowden strain and 40 mM NANA for FCV F9 strain ( Fig . 4B ) . Infection of cells with PSaV Cowden strain was also reduced by incubation with NANA in a dose-dependent manner , and was almost completely inhibited at 80 mM NANA ( Fig . 4C and 4D ) . Similar observations were found with FCV at 40 mM NANA ( Fig . 4D ) . In addition , when plaque reduction assays were performed , PSaV infection was blocked at 80 mM NANA ( Fig . S2 ) . Among other monosaccharides and oligosaccharides , N-glycolyl neuraminic acid also inhibited PSaV binding and infection of LLC-PK cells in a dose-dependent manner , but no inhibitory effect of galactose or sialyllactose was found regardless of the concentration ( data not shown ) . PSaV infects small intestinal epithelial cells , leading to villous atrophy [29] . To confirm whether PSaV binding in small intestinal epithelial cells is also dependent on sialic acid and blocked by NANA , PSaV was incubated with 160 mM of NANA , and was then incubated with porcine small intestinal tissue sections . Duodenal , jejunal and ileal tissue sections incubated with PSaV Cowden strain alone showed a positive signal for PSaV antigens on the villous epithelial cells . The very weak signal observed in the presence of NANA alone is seen only in the duodenal and ileal tissue sections , probably due to increased non-specific binding of the anti-PSaV antibody ( Fig . 5 ) . In contrast to these results , intestinal tissue sections incubated with a mixture of PSaV and 160 mM NANA showed markedly decrease of PSaV antigen intensity ( Fig . 5 ) . In addition , pretreatment of small intestinal sections with 1 mM NaIO4 ( data not shown ) or 10 mM NaIO4 markedly reduced the binding activity of PSaV Cowden strain ( Fig . 5 ) . To rule out any non-specific effects of NANA addition on the integrity of carbohydrate moieties on the tissue sections the effect of NANA pre-incubation on binding of the P domain of human norovirus VA387 , which recognizes HBGAs but not sialic acids as attachment factor , was examined [30] . As expected , a mixture of P domain of VA387 strain and 160 mM NANA had no influence on binding to intestinal epithelial cells ( Fig . S3 ) . Moreover , binding of the P domain to intestinal epithelial cells was markedly reduced by the pretreatment of 10 mM NaIO4 but not by 1 mM NaIO4 ( Fig . S3 ) . These results fit with previous observations that 1 mM NaIO4 eliminates terminal sialic acids only but 10 mM NaIO4 removes terminal sialic acids as well as HBGAs on carbohydrate moieties [10] . Taken all together , these data strongly indicated that PSaV binds to sialic acid on the cell surface . Sialic acid is attached to glycans via α2 , 3- , α2 , 6- or α2 , 8-linkages . To determine if PSaV requires these linkages for PSaV binding and infection , LLC-PK cells were pretreated with 200 mU V . cholerae neuraminidase ( NA ) ml−1 , which cleaves α2 , 3-linked , α2 , 6-linked and α2 , 8-linked sialic acids from the underlying glycans [4] . Treated cells were then incubated with either Alexa 594- or radio-labeled PSaV Cowden strain . Pretreatment with NA markedly reduced the binding of Alexa 594-labeled PSaV ( Fig . 6A ) , and radio-labeled PSaV binding was reduced to 2% of the levels observed in mock treated cells ( Fig . 6B ) . Infection assays , performed by indirect immunofluorescence for the viral antigen VPg , showed almost complete inhibition of PSaV infection at 200 mU NA ml−1 ( Fig . 6C and 6D ) . A similar degree of reduction in binding and infection was observed in the cells infected with α2 , 6-linked sialic acid-dependent FCV F9 strain ( Fig . 6B and 6D ) [4] and α2 , 3-linked sialic acid-dependent influenza virus Kr96 strain ( H9N2 ) ( Fig . 6B and 6D , and Fig . S4A and S4B ) [31] after pretreatment with NA . In addition , pretreatment of NA had no effect on binding of P domain of human norovirus VA387 strain ( Fig . S5 ) . To further identify which specific linkage is used for PSaV binding and infection , LLC-PK cells were initially pretreated with sialidase S ( SS ) from Streptococcus pneumoniae , which exclusively cleaves α2 , 3-linked sialic acid from complex carbohydrates and glycoproteins [4] . Pretreatment of LLC-PK cells with 40 mU SS ml−1 reduced PSaV binding to 49% ( Fig . 6A and 6B ) and decreased PSaV infection to 44% ( Fig . 6C and 6D ) . Importantly , complete inhibition of binding or infection was not achieved , even with increasing doses of SS ( data not shown ) . As expected , pretreatment of 40 mU SS ml−1 significantly blocked binding and infection of α2 , 3-linked sialic acid-dependent influenza virus Kr96 strain ( H9N2 ) [31] ( Fig . 6B and 6D , and Fig . S4A and S4B ) . In contrast , pretreatment of 40 mU SS ml−1 had no effect on α2 , 6-linked sialic acid-dependent FCV F9 binding or infection ( Fig . 6B and 6D ) [4] as well as binding of P domain of HBGAs-dependent human norovirus VA387 strain ( Fig . S5 ) [30] . These data suggested that PSaV Cowden strain may use not only α2 , 3-linked but also α2 , 6-linked sialic acids . In order to further identify the sialic acid linkages required for PSaV attachment and infection , LLC-PK cells were pretreated with specific lectins to block each specific isoform of sialic acid: 1 ) 400 µg Maackia amurensis lectin ( MAL ) ml−1 , which binds preferentially to α2 , 3-linked sialic acid , and 2 ) 400 µg Sambucus nigra lectin ( SNL ) ml−1 , which binds preferentially to α2 , 6-linked sialic acid [4] . Individual pretreatment of both lectins reduced , but did not completely block PSaV binding ( Fig . 7A ) . Quantitation of PSaV binding by using radio-labeled virus also demonstrated reduced binding , with 66% and 62% binding observed , after pretreatment with MAL and SNL , respectively ( Fig . 7B ) . Likewise , infection of LLC-PK cells by PSaV Cowden strain was decreased to 64% by 400 µg MAL ml−1 and 61% by 400 µg ml−1 SNL ( Fig . 7C and 7D ) . In contrast , MAL had no effect on FCV , but SNL significantly reduced binding and infection ( Fig . 7B and 7D ) . To establish whether PSaV recognizes both α2 , 3-linked and α2 , 6-linked sialic acids , LLC-PK cells were pretreated with mixtures of both MAL and SNL , and were then infected with PSaV Cowden strain . PSaV binding and infection were decreased , and complete reduction was observed by the treatment of a mixture at 400 µg MAL ml−1 and 400 µg SNL ml−1 ( Fig . 7A–7D ) . Taken together , all findings above confirm that PSaV Cowden strain uses both α2 , 3-linked and α2 , 6-linked sialic acids for binding and infection . Sialic acids are typically found at the terminal position of N- and O-linked glycans attached to the cell surface and to secreted glycoproteins or glycosphingolipids [32] . To test whether sialic acid moieties used for PSaV binding attach to a glycoprotein , LLC-PK cells were pretreated with either trypsin or chymotrypsin , and were then inoculated with Alexa 594-labeled or radio-labeled PSaV Cowden strain . As shown in Fig . 8A and 8B , pretreatment of 10 µg trypsin ml−1 and 10 µg chymotrypsin ml−1 reduced PSaV attachments to 35% and 41% , respectively , compared to mock treated and PSaV-inoculated control . Furthermore , individual pretreatment of cells with proteases reduced PSaV infection to 25% by trypsin and to 40% by chymotrypsin ( Fig . 8C and 8D ) . Comparatively , FCV binding and infection was reduced by trypsin or chymotrypsin pretreatments , to as low as 12% or 24% , respectively ( Fig . 8B and 8D ) . This inhibition to PSaV binding and infection by trypsin or chymotrypsin pretreatments suggested that like FCV , PSaV Cowden strain binds to sialic acid attached to a glycoprotein . To investigate if PSaV also utilizes glycolipid containing sialic acid moieties , LLC-PK cells were pretreated with 50 µM DL-Threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol ( PDMP ) , a well-known inhibitor of glucosylceramide synthase , and were then infected with PSaV Cowden strain . Regardless of PDMP pretreatment , PSaV bound to and replicated in cells to levels which were identical to those observed in mock treated cells ( Fig . 8A–D ) . Likewise , FCV F9 strain bound to and replicated in CRFK cells despite pretreatment of cells with 50 µM PDMP ( Fig . 8B and 8D ) . As expected , binding of MNV-1 CW1 strain which uses glycolipid as a receptor [11] was significantly inhibited by pretreatment of 50 µM PDMP ( Fig . 8 and Fig . S6 and S7 ) . In addition , binding of a VSV-G protein pseudotyped lentivirus , binding of which is also known to at least partially involve glycolipids [33] , was also affected by PDMP treatment of porcine LLC-PK cells ( Fig . S7 ) . These data confirm that PDMP treatment of LLC-PK cells reduced glycolipid synthesis , therefore confirming that PSaV Cowden strain attaches and infects cells via a glycoprotein containing sialic acid moieties . As sialic acid moieties may by attached to glycoproteins via both O or N-linkages , the ability of PSaV to bind to LLC-PK pretreated with benzyl 4-O-β-D-galactopyranosyl-β-D-glucopyranoside ( benzylGalNAc ) , an inhibitor of O-linked glycosylation , tunicamycin , an inhibitor of N-linked glycosylation , or PNGase F , which removes N-linked glycans , was examined . Pretreatment of cells with 3 mM benzylGalNAc reduced PSaV attachment to 2% of the levels observed in mock treated cells , whereas 3 µg tunicamycin ml−1 and 200 U PNGase F ml−1 pretreatment had no effect ( Fig . 9A and 9B ) . Infection of cells by PSaV Cowden strain was also reduced by ∼97% by benzylGalNAc treatment , but was unaffected by tunicamycin or PNGase F ( Fig . 9C and 9D ) . As a control , the FCV F9 strain , known to use N-linked sialic acid , was examined . As expected , FCV attachment and infection of CRFK cells was significantly inhibited by pretreatment with tunicamycin and PNGases , but not by benzylGalNAc ( Fig . 9B and 9D ) . These data indicated that PSaV attachment occurs via sialic acid linked by O-linked glycosylation .
The lack of an efficient cell culture system of human noroviruses and sapoviruses has hampered the study of virus entry and the molecular mechanisms of virus replication . Among enteric caliciviruses , PSaV Cowden strain is the only cultivable enteric sapovirus and has been shown to replicate in a continuous cell line ( LLC-PK ) , but only in the presence of intestinal contents from gnotobiotic pigs or bile acids as a medium supplement [6] . Although virus-host cell receptor interactions are the first step in the initiation of virus infection , the exact nature of the receptors which are recognized by the Sapovirus genus has not been determined . In the present study , therefore , we used the cell culture adapted PSaV Cowden strain as a model to investigate the entry strategy of an enteric sapovirus in vitro . Three glycoconjugates with relevance as calicivirus receptors have been described so far; human and bovine noroviruses and Tulane virus utilize HBGAs; MNV uses sialic acid linked to ganglioside or protein in a strain-dependent manner; FCV recognizes terminal sialic acid on an N-linked glycoprotein [4] , [5] , [10] , [11] , [12] . Our data would indicate that PSaV does not utilize HBGAs for attachment in agreement with previous work on human sapovirus GI and GV strains that also do not appear to bind HBGAs [14] . Instead , our data Indicate that PSaV , like FCV and some MNV isolates , utilizes sialic acid as a receptor for its binding and infection [4] , [11] , [34] . Most sialic acids which are recognized as receptors are terminal sialic acids attached to a penultimate galactose by either α2 , 3-linkage or α2 , 6-linkage . Previous results have shown that the FCV F9 strain utilizes α2 , 6-linked sialic acid as a receptor [4] , whereas MNV-1 strain CW3 recognizes both α2 , 3- and α2 , 6-linked sialic acids for binding and infection [11] . Our data would indicate that similarly to the MNV-1 CW3 strain [11] , PSaV Cowden strain recognizes both α2 , 3- and α2 , 6-linked sialic acids which are attached to glycans as receptors . The inability of PSaV to form stable agglutinates of RBCs from numerous species , despite the presence of α2 , 3 and α2 , 6-linked sialic acid , would suggest that PSaV also forms stabilizing interactions with the specific glycoproteins to which the sialic acids are linked . This is in agreement with our observations that protease treatment reduces virus binding ( Fig . 8B ) and suggests that this glycoprotein ( or glycoproteins ) are not expressed on the surface of RBCs . In addition , the conditions under which stable agglutination occurs , i . e . , pH , ionic conditions etc . , often varies from virus to virus and is dependent on the species of RBCs used [35] . Therefore , it is also possible that the lack of stable agglutination was due to suboptimal conditions used in the HA assay . The identification of specific glycoproteins involved in PSaV binding and the optimal conditions for HA activity of PSaV forms the basis of ongoing work . The different tissue distribution of α2 , 3- and α2 , 6-linked sialic acids can strongly influence viral tissue tropism and pathogenesis [11] , [25] , [34] , [36] . In pigs , both α2 , 3- and α2 , 6-linked sialic acids are expressed along the epithelial border as well as in goblet cells of the small and large intestines [37] . In addition , both α2 , 3- and α2 , 6-linked sialic acid receptors are distributed extensively in the major organs of pigs , including the trachea , lungs , liver , kidney , spleen , heart , skeletal muscle , and cerebrum [37] . In our previous experiments , we found that pigs orally or intra-venously infected with wild-type virulent PSaV exhibited intestinal pathology as well as viremia [29] . However , systemic infections caused by this strain were not observed , and virus was not isolated from every organ , where both α2 , 3- and α2 , 6-linked sialic acid receptors are well expressed [29] . One explanation for this that the concentration of bile acids which supports the replication of PSaV in cell culture is much higher in the proximal intestine than in the blood and extraintestinal organs [6] , [29] . Therefore , it is plausible that the restriction of growth of PSaV within the small intestine is at least partially due to the requirement of a high concentration of bile acids [29] . However , we cannot rule out the possibility that some other potential co-receptors may not be present at extraintestinal sites . It has been suggested that the initial attachment of a virus to a primary receptor enriches the virus at the cell surface and primes the attached virus for interaction with secondary receptor ( s ) at the cell surface , which is ( are ) necessary for virus uptake , subsequent uncoating of the nucleic acid , and infection of the target cell [25] . One well characterized example would be the multistep entry of rotavirus into cells , where various cell receptors , including the terminal sialic acid or HBGAs , integrins , and heat shock protein Hsc70 are utilized by the outer most proteins , VP4 and VP7 , of rotaviruses [38] . Among caliciviruses , FCV is known to utilize not only sialic acid , but also JAM-1 for virus entry into cells; the latter presumably aiding FCV penetration into the host cells as a co-receptor [4] , [13] . In agreement with this hypothesis , cyro-EM reconstruction and biochemical studies of the FCV capsid with JAM-1 indicated that JAM-1 binding results in significant conformational changes in the capsid [39] , [40] , [41] . Glycosylation produces different types of glycans which are typically attached to cellular proteins and lipids [32] , [37] . Different members of caliciviruses seem to utilize different linkage of glycans for their binding and entry , i . e . MNV use sialic acid bearing gangliosides ( CW3 like strains ) or proteins ( CR3 strain ) [11] , [34] , while FCV recognizes sialic acid bearing protein components [4] . In the present study , pretreatment of PDMP had no influence on the binding and infection of PSaV , suggesting that like FCV [4] , PSaV does not utilize sialic acid bearing lipids . Protein glycosylation encompasses N-glycans , O-glycans , and glycosaminoglycans [37] . Different types of glycans which are attached to cellular proteins are utilized by the different viruses , even within the same genus . For example , adeno-associated virus type 5 ( AAV5 ) interacts with sialic acid on N-linked carbohydrates , whereas AAV4 interacts with sialic acid on O-linked carbohydrates , and both viruses require 2 , 3-linked sialic acid for binding [42] . We found that the sialic acid-bearing glycans used for PSaV binding are attached to cell surface proteins in a similar manner to FCV [4] . Unlike FCV [4] , however , carbohydrate moieties linked to terminal sialic acid as a receptor for PSaV are present on an O-linked glycoprotein . In conclusion , we have demonstrated that unlike noroviruses , PSaV infects cells via both α2 , 3- and α2 , 6-linked sialic acids attached to O-linked glycoproteins . This work has provided new insights into the mechanisms of sapovirus entry , and may provide additional information relevant to the identification of inhibitors of sapovirus pathogenesis .
LLC-PK cells , Caco2 cells and HeLa cells obtained from the American Type Culture Collection ( ATCC , USA ) were maintained in Eagle's minimal essential medium ( EMEM ) containing 10% fetal bovine serum ( FBS ) , 100 U/ml penicillin , and 100 µg/ml streptomycin . CRFK cells and MDCK cells from ATCC was grown in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 5% FBS , 100 U/ml penicillin , and 100 µg/ml streptomycin . MA-104 cells from ATTCC was grown in α-MEM supplemented with 5% FBS , 100 U/ml penicillin , and 100 µg/ml streptomycin . RAW264 . 7 cells were grown in RPMI1640 supplemented with 5% FBS , 100 U/ml penicillin , and 100 µg/ml streptomycin . The tissue culture-adapted PSaV Cowden strain was recovered from the full-length infectious clone pCV4A , and was propagated in LLC-PK cells with the supplement of bile acid [6] . The FCV F9 strain , human influenza A virus A/Puerto Rico/8/34 ( H1N1 ) ( PR8 virus ) strain , human rotavirus DS-1 strain , and CVB3 Nancy strain were purchased from the ATCC . Chicken influenza A virus A/Korea/96/96 ( H9N2 ) ( Kr96 virus ) strain was obtained from Animal and Plant Quarantine Agency , Korea and MNV-1 CW1 strain was a kind gift of Dr . H . W . Virgin , Washington University School of Medicine , USA . These viruses were propagated in CRFK cells , MA-104 cells , MDCK cells , HeLa cells , or RAW264 . 7 cells , respectively [4] , [11] , [43] , [44] , [45] . Sodium periodate ( Sigma-Aldrich ) , trypsin ( Sigma-Aldrich ) , chymotrypsin ( Sigma-Aldrich ) , NANA ( Fluka , USA ) , sialyllactose ( GeneChem , Korea ) , galactose ( Sigma-Aldrich ) , MAL ( Sigma-Aldrich ) , and SNL ( Sigma-Aldrich ) were dissolved in PBS pH 7 . 2 . Tunicamycin ( Sigma-Aldrich ) and Alexa 594 ( Invitrogen , USA ) were dissolved in DMSO . PDMP ( Calbiochem , USA ) and benzylGalNAc ( Sigma-Aldrich ) were dissolved in ethanol . Other reagents included 35[S]methionine/cysteine ( PerkinElmer , USA ) , NA ( Sigma-Aldrich ) , SS ( Prozyme , USA ) , PNGase F ( NEB , UK ) , anti-PSaV capsid monoclonal antibody [46] , anti-PSaV VPg polyclonal antibody , anti-FCV capsid monoclonal antibody ( Santa Cruz , USA ) , anti-CVB3 capsid monoclonal antibody ( Millipore , IL , USA ) , anti-Hu/NoV/GII . 4/HS194 virus-like particles ( VLPs ) polyclonal antibody ( kind gift of Dr . L . Saif , The Ohio State University , Ohio , USA ) , anti-influenza virus nucleoprotein monoclonal antibody ( Median Diagnostic , ChunCheon , Korea ) , anti-rabbit IgG-FITC antibody ( Jackson Immuno Research Lab , USA ) , and anti-mouse IgG-FITC antibody ( Santa Cruz ) . VP8* of human DS-1 rotavirus ( RV ) strain ( P[4] genotype ) and P particles of VA387 ( GII-4 ) and VA207 ( GII-9 ) norovirus ( NV ) strains were cloned , expressed and purified as described previously [26] , [27] . Briefly , the cDNAs encoding RV DS-1 strain VP8* and encoding NV VA387 or VA207 P particles with cysteine peptide were cloned into the expression vector pGEX- 4T-1 ( glutathione S-transferase [GST]–gene fusion system; GE Healthcare Life Sciences , Piscataway , NJ ) . After sequence confirmation , the recombinant GST-VP8 and GST-P fusion proteins were expressed in Escherichia coli strain BL21 as described previously ( Table 2 ) [26] , [27] . Expression of protein was induced by IPTG ( isopropyl-β-D-thiogalactopyranoside; 0 . 2 mM ) at room temperature ( 22°C ) overnight . RV VP8-GST and NV GST-P fusion proteins were purified using the Pierce GST spin purification kit ( Pierce , IL , USA ) according to the manufacturer's protocol . The P particles of VA387 and VA207 were released from GST by thrombin ( Sigma-Aldrich ) digestion at room temperature overnight . The concentration of the purified RV VP8* and NV P particles were determined by measuring the absorbance at 280 nm . Hemagglutination ( HA ) assay was performed using RBCs from animals and humans . Blood from pig , cow , chicken , and rat were obtained from the College of Veterinary Medicine , Chonnam National University , and human blood was provided by volunteer donors ( human ABO , Lewis Oa+b- , Lewis Oa−b+ , and Lewis Oa−b− types ) at Chonnam National University Hospital . RBCs were packed in PBS pH 7 . 2 without Ca2+ , and were centrifuged at 500×g for 10 min . PSaV Cowden strain ( 10 µg/ml ) , human influenza virus PR8 ( H1N1 ) ( 5 µg/ml starting dilutions ) , P particles ( 10 µg/ml ) of norovirus VA387 strain , and VP8* of rotavirus DS-1 strain ( 10 µg/ml ) were diluted serially 2 fold in PBS ( 0 . 01 M sodium phosphate , 0 . 15 M NaCl , pH 5 . 5 ) using V-shaped 96-well plates . The HA activity was tested by mixing equal volume of RBCs 1% to the prepared viruses or protein dilutions . The reactions were allowed to proceed for 1 h at 4°C or 20°C , and the agglutination of RBCs was observed and recorded [10] . The HA titer was the reciprocal of the highest virus dilution that allowed sedimentation of the RBCs compared to control wells . The synthetic oligosaccharide-based histo-blood group antigen binding assay was carried out as described elsewhere [47] , [48] . Briefly , 96-microtiter plates were coated with PSaV Cowden strain ( 10 µg/ml ) , P particles of NV VA207 ( 10 µg/ml ) and VA387 ( 10 µg/ml ) strains , human influenza virus PR8 strain ( H1N1 ) ( 10 µg/ml ) , or VP8* of RV DS-1 strain ( 10 µg/ml ) at 4°C overnight . Coated plates were blocked with 5% bovine serum albumin ( BSA ) for 1 h at room temperature , and each synthetic oligosaccharide-polyacrylamide ( PAA ) -biotin conjugate ( 10 µg/ml ) was then added and further incubated at 4°C overnight . Oligosaccharides used in this study included Lewis antigens ( Lea , Leb , Lex , and Ley ) , H type 1 , H type 2 , H type 3 , type A disaccharide , type B disaccharide , type A trisaccharide , type B trisaccharide , sLea and sLex tetrasaccharides , which were conjugated with biotin ( GlycoTech Co , USA; Table 1 ) . Bound oligosaccharides were detected using HRP-conjugated-streptavidin ( Jackson Immuno Research Lab , USA ) . The signal was visualized by TMB ( Komabiotech , Korea ) followed by measurement at 450 nm . In each step , the plates were incubated for 1 h at 37°C , after which they were washed five times with PBS-Tween 20 . Labeling of PSaV Cowden , FCV F9 , CVB3 Nancy , MNV-1 CW1 , and influenza virus Kr96 strains with 35[S]methionine/cysteine were carried out as described previously [4] . Briefly , confluent monolayers of permissible cells for above viruses in five 175 cm3 flasks were infected with above each strain , respectively , at a multiplicity of infection ( MOI ) of 0 . 1 for 4 h at 37°C . The medium was replaced with RPMI 1640 lacking methionine and cysteine ( Sigma-Aldrich ) for 2 h . The medium was then supplemented with 1 Mbq 35[S] methionine/cysteine . After 72 h ( PSaV Cowden strain ) , 16 h ( FCV F9 strain ) , 16 h ( CVB3 Nancy strain ) , 16 h ( MNV-1 CW1 strain ) , and 72 h ( influenza virus Kr96 strain ) post virus inoculation , the cultured virus was frozen and thawed three times . Each virus was pelleted by ultracentrifugation for 10 h at 104 , 000×g in a Hitachi P28S rotor , and then was purified by sucrose gradient density ultracentrifugation for 10 h at 104 , 000×g using a Hitachi P28S rotor . Labeling of PSaV Cowden , FCV F9 , CVB3 Nancy , MNV-1 CW1 , P domain of VA387 strain and influenza virus Kr96 strains with Alexa 594 ( Invitrogen ) was performed following the manufacturer's instruction . Briefly , 1 part of Alexa 594 solution was mixed with 9 parts of a solution containing 1×108 pfu/ml of above each strain or 9 parts of a solution containing 500 µg/ml of P domain of VA387 strain . Each reaction was mixed thoroughly for 30 sec and was then incubated for 1 h at room temperature . Attachment assays with 35[S]methionine/cysteine-labeled PSaV Cowden , FCV F9 , CVB3 Nancy , MNV-1 CW1 or influenza virus Kr96 strains to LLC-PK , CRFK , HeLa , RAW264 . 7 or MDCK cell lines were performed as described previously with slight modifications [4] . Briefly , subconfluent monolayers of permissible cells grown on 6-well plates were mixed with purified 35[S]methionine/cysteine-labeled above each strain ( 50 , 000 c . p . m . ) , and were then incubated for 45 min on ice . Cells were washed three times with ice-cold PBS , after which they were lysed with 0 . 1% SDS and 0 . 1 M NaOH . Total radioactivity in the cell lysate was determined by liquid scintillation counting [4] . To visualize virus attachment to subconfluent monolayers of cells grown on the confocal dish were pretreated with or without inhibitors or enzymes , as described above . Mock- or reagent-treated cells were then inoculated with PSaV Cowden , FCV F9 , CVB3 Nancy , MNV-1 CW1 , P domain of VA387 strain or influenza virus Kr96 strains labeled with Alexa 594 dye or Alexa 594 dye alone , after which they were incubated for 5 min on ice . Cells were washed 1 time with cold PBS 1% containing fetal bovine serum ( PBS-FBS ) , fixed with 4% formaldehyde in cold PBS for 10 min , and washed 3 times with cold PBS . The cells which were incubated with Alexa 594 dye-labeled viruses or P domain of VA387 strain were stained with 300 nM 4′ , 6-diamidino-2-phenylindole ( DAPI ) solution for nucleus staining , mounted with using SlowFade Gold antifade reagent ( Invitrogen ) , and examined using an EZ-C1 confocal microscope and EZ-C1 software ( Nikon , Japan ) . Laser and microscope settings were adjusted according to the manufacturer's instructions . Cells infected with viruses unlabeled with Alexa dye were analyzed by immunofluorescence , as described below . Attachment assays by RT-qPCR with PSaV Cowden , MNV CW1 or VSV-G protein pesudotyped lentivirus ( LV ) strains to LLC-PK cell line was performed as described above using Alexa of radio-labeled virus . Briefly , subconfluent monolayers of LLC-PK cells on 24-well plate were pretreated with 50 µM PDMP or 5 mM NaIO4 . PDMP-treated cells were inoculated with above each strain ( 3 TCID50 of PSaV and MNV; 1 . 25 transducing units per cell of LV ) , and were then incubated for 45 min on ice . Cells were washed three times with ice-cold PBS , Viral RNA was immediately extracted and analyzed by RT-qPCR with primer specific to each virus . RNA was extracted using GenElute Mammalian Total RNA Miniprep Kit ( Sigma ) . One hundred nanograms were subsequently reverse transcribed using random hexamers . Fragments of ∼200 bp were amplified using the following gene-specific primers: PSaV: 5′-CAACAATGGCACAACAACG-3′ ( forward ) and 5′-ACAAGCTTCTTCACCCCACA-3′ ( reverse ) ; MNV-1: 5′-TGGACAACGTGGTGAAGGAT-3′ ( forward ) and 5′-CAAACATCTTTCCCTTGTTC-3′ ( reverse ) , and LV-WPRE: 5′-TCGGCCCTCAATCCAGCGGA-3′ ( forward ) and 5′-TCGTCTGAGGGCGAAGGCGA-3′ ( reverse ) . Standard curves were generated for all genes quantified . Additional non-template and non-reverse transcriptase samples were routinely analyzed as negative controls . Data were collected using a ViiA 7 Real-time PCR System ( Applied Biosystems ) . Infectivity assays of PSaV Cowden , FCV F9 , CVB3 Nancy or influenza virus Kr96 strains in the permissible cells , respectively , were carried out as described previously with slight modifications [4] . Briefly , confluent monolayers of each permissible cell were treated with various inhibitors or enzymes as described below . Either mock or treated cells were infected with above each strain at an MOI of 0 . 1 pfu/cell , and were then incubated at 37°C for 1 h . Cells were washed three times with PBS , and were then replaced with maintenance medium . The cells were incubated for 72 h ( PSaV Cowden ) , 8 h ( FCV F9 ) , 9 h ( CVB3 Nancy ) or 48 h ( influenza virus Kr96 ) at 37°C prior to being fixed with 4% formaldehyde in PBS , and were analyzed by immunofluorescence assay as described below . Immunofluorescence assays was performed as previously described [4] . Briefly , fixed cells in 8-well chamber slides were permeabilized by the addition of 0 . 2% Triton X-100 , incubated for 5 min at room temperature , and then washed with PBS containing 0 . 1% newborn calf serum ( PBS-NCS ) . Anti-PSaV VPg polyclonal antibody , anti-FCV capsid monoclonal antibody , anti-CVB3 capsid monoclonal antibody or anti-influenza virus nucleoprotein monoclonal antibody was then added at a dilution rate of 1∶100 , 1∶200 or 1∶500 , respectively . Chamber slides were incubated at 4°C overnight . Cells were washed 3 times with PBS-NCS , and FITC-conjugated secondary antibodies ( diluted to 1∶100 ) were then added . Nuclei were stained with propidium iodide ( PI ) , and cells were examined using confocal microscopy . A total of 700 cells , as indicated by PI or DAPI staining , were counted per condition , and were scored for PSaV VPg protein expression . After image analysis with Metamorph Premier v6 . 3 software ( Molecular Devices , PA ) , the infected cells were counted as positive for viral antigen if they had a fluorescent intensity which was at least three times the fluorescent intensity of the uninfected controls . The percentage of positive cells was then normalized to that of the untreated control . 3-day-old piglets obtained from sows by hysterectomy were used to attain the small intestinal segments including the duodenum , jejunum and ileum . Segments were excised after sacrifice , immediately immersed in 10% buffered formalin , processed routinely for paraffin embedment , sectioned , and stained with hematoxylin for histology . For immunohistochemical studies , paraffin-embedded sections were deparaffinized , and were then rehydrated through graded alcohols into 0 . 1 M PBS , after which they were treated with 0 . 3% H2O2 to quench endogenous peroxidase . Sections were either incubated in PBS ( 200 µl ) , PSaV Cowden strain ( 1×106 pfu/ml ) , P domain of VA 387 strain ( 10 µg/ml ) , NANA ( 160 mM , pH 7 ) , PSaV ( 1×106 pfu/ml ) and NANA ( 160 mM , pH 7 ) mixture , P domain of VA 387 strain ( 10 µg/ml ) and NANA ( 160 mM , pH 7 ) mixture with or without 1 mM or 10 mM NaIO4 pretreatment for 1 h at room temperature . Pretreated sections with NaIO4 were incubated for 30 min prior to the addition of virus . After washing with PBS , a monoclonal antibody against PSaV Cowden strain capsid protein or a polyclonal antibody against Hu/NoV/GII . 4/HS194 VLPs was incubated with the sections at 4°C overnight . Sections were then rinsed 3 times with PBS , and incubated with biotinylated secondary antibody ( Dako , USA ) and peroxidase-conjugated streptavidin ( Dako , USA ) . Reactions were developed with 3-amino-9-ethylcarbazol ( AEC; Vector laboratories , USA ) , and followed by Mayer's hemalum solution ( Merck , Germany ) for counterstaining . Confluent monolayers of LLC-PK cells on 6-well plates were inoculated with PBS ( 200 µl ) , PSaV Cowden strain ( 1×105 pfu/ml ) , or mixtures of PSaV ( 1×105 plaque forming unit/ml ) and various concentrations of NANA ( 20–160 mM , pH 7 ) , after which they were incubated in a 5% CO2 incubator . After 2 h of virus adsorption , PBS and virus inocula were thoroughly discarded and washed 3 times with PBS . Overlay medium containing a 1× concentrated MEM , 10% FBS , 1 . 2% ( w/v ) avicel ( FMC BioPolymer , Belgium ) , and 200 µM GCDCA was added to each well . Plates were incubated for 96 h in a 5% CO2 incubator . After incubation , inoculated cells were fixed with 20% trichloroacetic acid , and the avicel was then removed . Plaques were visualized by staining with 1% ( w/v ) crystal violet solution [49] , [50] . Subconfluent monolayers of LLC-PK , CRFK , HeLa , Caco2 , RAW264 . 7 or MDCK cells grown on 6-well plates or 8-well chamber slides for confocal analysis were treated with chemicals , metabolic inhibitors , or enzymes optimized at the specific concentrations , incubation times and temperatures described in Table 3 . After pretreatment , cells were washed three times with PBS , and binding and infectivity assays were carried out as described above . Mock and control treatments were performed at the same time . Statistical analysis was performed using SPSS version 11 . 5 . 1 for window ( SPSS , USA ) . One way analysis of variance ( ANOVA ) test were used . A P-value <0 . 05 was considered statistically significant . All animals were handled in strict accordance with good animal practices , as described in the NIH Guide for the Care and Use of Laboratory Animals ( NIH Publication No . 85–23 , 1985 , revised 1996 ) . The protocol was approved by the Committee on Ethics of Animal Experiments , CNU with permit number ( CNU No . 2012-87 ) . The human blood samples collected with written consent from the patients were handled in strict accordance with human subjects , as described in the Guidance for the Care and Use of Human Samples of the CNU adhered from the WMA Declaration of Helsinki ( Ethical Principles for Medical Research Involving Human Subjects ) . The protocol was approved by the Committee for Research Ethics Concerning Human Subjects , CNU with permit number ( CNU IBR No . 1040198-130807-BR-002-01 ) .
|
Although enteropathogenic sapoviruses and noroviruses are leading causes of acute gastroenteritis in both humans and animals , the study of viral pathogenesis and immunity of these ubiquitous pathogens has been hampered due to the lack of a fully permissive cell culture system . Porcine sapovirus Cowden strain provides a suitable system that can be used to identify the molecular mechanisms of viral pathogenesis . Previous studies have shown that carbohydrates and glycolipids play important roles in the attachment of members of the Caliciviridae; histo-blood group antigens ( HBGAs ) are used by Norovirus genogroups I to IV , as well as members of the Lagovirus , and Recovirus genera , whereas terminal sialic acid is recognized as a receptor for feline calicivirus and murine norovirus . To date , however , the role of carbohydrates in the life cycle of sapoviruses has remained largely unknown . We found that porcine sapovirus binds to susceptible host cells through both α2 , 3- and α2 , 6-linked terminal sialic acids which are attached to O-linked glycoproteins . These efforts , findings and insights will significantly contribute to a better understanding of the sapovirus life cycle .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"veterinary",
"microbiology",
"biochemistry",
"biology",
"and",
"life",
"sciences",
"veterinary",
"science"
] |
2014
|
Both α2,3- and α2,6-Linked Sialic Acids on O-Linked Glycoproteins Act as Functional Receptors for Porcine Sapovirus
|
The envelope glycoproteins ( Envs ) of HIV-1 continuously evolve in the host by random mutations and recombination events . The resulting diversity of Env variants circulating in the population and their continuing diversification process limit the efficacy of AIDS vaccines . We examined the historic changes in Env sequence and structural features ( measured by integrity of epitopes on the Env trimer ) in a geographically defined population in the United States . As expected , many Env features were relatively conserved during the 1980s . From this state , some features diversified whereas others remained conserved across the years . We sought to identify “clues” to predict the observed historic diversification patterns . Comparison of viruses that cocirculate in patients at any given time revealed that each feature of Env ( sequence or structural ) exists at a defined level of variance . The in-host variance of each feature is highly conserved among individuals but can vary between different HIV-1 clades . We designate this property “volatility” and apply it to model evolution of features as a linear diffusion process that progresses with increasing genetic distance . Volatilities of different features are highly correlated with their divergence in longitudinally monitored patients . Volatilities of features also correlate highly with their population-level diversification . Using volatility indices measured from a small number of patient samples , we accurately predict the population diversity that developed for each feature over the course of 30 years . Amino acid variants that evolved at key antigenic sites are also predicted well . Therefore , small “fluctuations” in feature values measured in isolated patient samples accurately describe their potential for population-level diversification . These tools will likely contribute to the design of population-targeted AIDS vaccines by effectively capturing the diversity of currently circulating strains and addressing properties of variants expected to appear in the future .
HIV-1 is the primary etiologic agent of the global AIDS pandemic . Soon after identification of HIV-1 in the early 1980s , the tremendous sequence diversity of circulating strains was appreciated [1 , 2] . The genetic diversity of HIV-1 has posed a major obstacle to development of an efficacious vaccine . Several factors contribute to the sequence heterogeneity of this virus . Mutations are frequently introduced in the viral genome during replication by the error-prone reverse transcriptase enzyme [3–7] . In addition , HIV-1 has a high propensity for recombination during coinfection of a cell by two different isolates [8–12] . The high rate of viral replication ( 1010–1012 new virions can be generated daily ) increases the appearance of sequence variants [13 , 14] . Persistence of the newly formed variants in the host is determined by the selective pressures exerted on the different virus components . Of the proteins encoded by HIV-1 , the envelope glycoproteins ( Envs ) show the greatest degree of in-host and between-host diversity [15] . The HIV-1 Envs are contained on the surface of the viral particle and function as a membrane-fusing machine that mediates entry into host cells [16 , 17] . Env is composed of a surface subunit ( gp120 ) and a transmembrane subunit ( gp41 ) [18] . HIV-1 infection of the host is most frequently initiated by a single virus [19–22] . From this founder state , the virus replicates to form multiple quasispecies , which elicit formation of antibodies that can bind to Env and neutralize infectivity of virions [23 , 24] . However , frequent mutations in Env allow emergence of escape variants that contain changes in the antibody-targeted epitopes . Such variants can then persist in the infected individual . Thus , antibody pressure applied by the host defines properties of circulating Envs [14 , 25 , 26] . A second type of pressure applied on Env is the requirement to effectively fuse with host cells . This pressure is dynamic , since availability of susceptible cells can alter during the course of infection and in different body compartments [27–35] . As a consequence of selection , there is continuous replacement of circulating viral lineages in the host , associated with increased divergence of viral quasispecies [29 , 36–42] . The balance between forces that increase sequence diversity ( i . e . , random mutations and recombinations ) and the selective forces that act to contain diversity determines the pattern of virus evolution in the infected host [29 , 40 , 43 , 44] . Multiple host factors affect the population-level evolution of HIV-1 , including the humoral and cellular immune responses [45–47] and antiretroviral treatment regimens [48 , 49] . Several studies have examined historic population-level trends in genomic properties of HIV-1 [50–53] . In addition , changes in sensitivity of circulating strains to the humoral and cellular immune responses have been studied [54–56] . Changes are often attributed to population-level adaptation of the virus to the host immune response and to the fitness pressures applied . Nevertheless , for most properties of Env ( sequence and structural ) , the basis for the observed changes is not understood . Why are certain properties altered whereas others remain conserved over the years ? Are these trends sufficiently stable to allow us to predict the range of variants that will evolve in the future ? The ability to capture the current diversity of phenotypes in the population and to predict properties of future variants will likely improve the efficacy of population-targeted immunogens . To achieve the above goals , we conducted a combined cross-sectional ( population-level ) and longitudinal study in a defined geographic location in the United States . Two types of properties ( features ) were examined: ( i ) integrity of Env epitopes recognized by broadly neutralizing antibodies , and ( ii ) sequence characteristics ( e . g . , identity of individual amino acids at key antigenic sites or the length and charge of Env segments ) . We found that each Env feature is maintained at a defined level of variance among strains cocirculating in the same host at any given time , which is conserved among different individuals . We designate this property of each feature as its “volatility index . ” Based on this parameter , we modeled the changes in sequence and structural features of Env in the patient and population as a linear diffusion process , which progresses with increasing genetic diversity . In-host volatility is highly correlated with the longitudinal divergence of features in patients over time . A strong relationship also exists between the volatility of each feature and its diversity in the population . Based on volatility indices measured in a small set of patient samples collected during the 1980s , we accurately predict the diversity of features that developed during the next three decades . Therefore , volatility and its translation patterns into population diversity explain many of the historic changes in Env features during the course of the epidemic . The ability to predict clade-specific patterns of change through limited patient sampling will likely contribute to the tailoring of AIDS vaccines to structural properties of Envs circulating within specific populations and the changes expected to occur during defined timeframes in the future .
Several studies indicate that population-level changes have occurred in HIV-1 properties over the course of the AIDS pandemic [54–56] . To identify “clues” that could help us predict future population-level changes in sequence and structural properties of Env , we conducted a comprehensive study in a geographically-defined region of the US . Plasma samples provided to the University of Iowa HIV Clinic between 1985 and 2012 were used for isolation of the env gene from circulating viruses . A total of 371 Envs from 113 Iowa City patients were examined . For 101 patients , one plasma sample was available ( these samples are designated below as cross-sectional ) . In addition , 12 patients provided longitudinal samples , collected over the course of 2–11 y . We also isolated 177 Envs from longitudinal plasma samples of 14 patients from the University of Washington Center for AIDS Research ( CFAR ) repository in Seattle ( designated as UW samples ) . From each plasma sample , we amplified env genes of individual viruses by the single genome amplification ( SGA ) method [20 , 57] . Amplification products were cloned into a vector that allows expression of the Env protein . To focus our studies on Envs of potentially transmissible viruses , we measured the ability of each Env to mediate entry into cells ( see Materials and methods section ) . Further analyses were performed only for fusion-competent Envs . To avoid direct effects of antiretroviral therapy ( ART ) on Env structure or function , none of the patients were treated by entry inhibitors during or prior to plasma sample collection . Phylogenetic relationships between Envs isolated for this work are shown in S1 Fig . All primary data , GenBank accession numbers , and the amino acid sequence alignment are provided in S1 , S2 and S3 Data . All Envs from Iowa City and Seattle belong to clade B viruses , except Envs from Iowa City patients IC . 798 and IC . 999 , which are from clades AD and A , respectively . For each Env , we examined the integrity of defined epitopes by measuring recognition by specific probes using a cell-based ELISA system [58 , 59] . The assay involves expression of full-length trimeric cleaved Envs on the surface of human osteosarcoma ( HOS ) cells and measurement of probe binding [60] . A panel of 11 broadly neutralizing monoclonal antibodies ( mAbs ) that recognize well-defined epitopes in antigenically dominant regions of Env was selected as structural probes . Such antibodies constitute the primary specificities in sera with broadly neutralizing activity [61 , 62] . Distinct but overlapping epitopes were chosen for improved characterization of the structural layout of these regions . Carbohydrate binding antibodies PGT121 and PGT126 share a critical glycosylation site at position 332 [63] and partly compete with mAb 2G12 that targets mannose glycans on gp120 [64 , 65] . This glycosylation site is highly accessible on the trimer and is considered a “supersite of vulnerability” for HIV-1 neutralization [66] . The mAbs PG9 and PG16 target overlapping , trimer-dependent epitopes [67] . The mAbs 2F5 and 10E8 target the N- and C-terminal domains of the gp41 membrane-proximal ectodomain region ( MPER ) , respectively [68–71] . The 10E8 epitope is conserved among diverse HIV-1 clades whereas the 2F5 epitope shows some intra- and inter-clade diversity . Several probes that target the conserved , antigenically dominant CD4-binding site were tested , including mAbs b12 , VRC03 , and CD4-Ig , which contains two copies of CD4 linked to the Fc region of human IgG1 [58] . The mAb 39F binds to the V3 variable loop of gp120 [72] , which shows clade-specific levels of solvent exposure [73] that we sought to characterize . Binding of probes describes the integrity of their target epitopes and is expressed as a percent of probe binding to the control AD8 Env , which contains all epitopes tested in this study . Data are normalized for the level of Env expression using saturating concentrations of CD4-Ig [58] . The output of the binding assay spans a range of approximately five orders of magnitude . Since changes in epitope integrity are measured by fold-changes in probe binding efficiency , we minimized the confounding effects of the very high and low values by correcting the log10-transformed data with a logistic function ( see Materials and methods section and S2 Fig ) . We examined historic changes in integrity of the above epitopes in viruses from Iowa City samples collected between 1985 and 2012 . The epitopes tested were present in most Envs from the early part of the epidemic ( 1985–1991 , designated herein Period1 ) . However , during subsequent years , multiple isolates appeared with lower binding efficiencies , representing loss of epitope integrity ( see Fig 1A and complete set of mAbs in S3 Fig ) . To quantify the historic changes in these features , we compared their population diversity during Period1 and Period3 ( 2005–2012 ) using Levene’s test for equality of variances ( see details of statistical tools in the Materials and methods section ) . Different patterns of historic changes were observed for the epitopes . For example , PGT126 , b12 , and 10E8 show relatively low ( and similar ) levels of diversity during the early part of the epidemic . However , PGT126 and b12 gradually increased in their diversity from Period1 to Period3 whereas limited changes occurred in 10E8 ( compare p-values of Levene’s test , labeled Pvar in Fig 1A ) . Quantification of the percentage of Envs with very low or no binding of each probe showed that integrity of the 10E8 , b12 , and PGT126 epitopes was similar in the population during the 1980s ( labeled by green circles in Fig 1A ) . However , the PGT126 epitope was gradually lost over the years whereas 10E8 and b12 were less affected . The epitopes of mAbs VRC03 , 39F , and 2F5 were also similarly distributed in the population during the 1980s . However , VRC03 was then mostly eliminated ( ~70% of isolates from the latest time period do not contain this epitope ) whereas the epitopes of 39F and 2F5 were retained in most circulating isolates . We also examined historic changes in sequence features of the Env variable loops , including length , charge density ( i . e . , total charge per amino acid loop length ) , density of potential N-linked glycosylation sites ( PNGSs ) , and mean loop hydropathy score ( based on the Black and Mould scale [74] ) . For definition of Env features and segment boundaries , see Materials and methods section . Macroarchitectural ( segmental ) properties of Env describe the context in which epitopes are contained and are often indicative of important biological phenotypes , such as coreceptor tropism [75–78] or formation of some epitope groups [79 , 80] . As expected [34 , 81] , the variable loops showed different patterns of historic changes in population diversity ( see length of loops V1–V5 in Fig 1B and all feature types in S4 Fig ) . The V1 and V2 loops , which are located at the trimer apex and are mostly solvent-exposed [82–84] , show increasing diversity in their lengths from Period1 to Period3 . By contrast , the V3 loop , which is relatively cryptic on the Env trimer , shows little change over the past three decades . The V4 loop , which is solvent exposed but indispensable for trimer integrity and function [85] , demonstrated a level of diversity similar to the V1 and V2 loops during Period1 but then diversified minimally over subsequent years . Therefore , although diversity of Env segmental features has generally increased from Period1 to Period3 ( S5 Fig ) , different loops demonstrate different patterns of change . In summary , sequence and antigenic features of Env show different patterns of change in the population over the past 30 y . We sought to examine the basis for such patterns , which could allow us to predict future changes in properties of circulating strains . For this purpose , we analyzed the spread of Env features from patient to population by studying the relationships between the following: ( i ) variance among strains cocirculating in the host at any time point , ( ii ) longitudinal divergence patterns in patients , and ( iii ) diversification of features in the population . We first measured for each feature the level of variance among functional , cocirculating Envs . The coefficient of variation ( CoV ) of feature values among Envs isolated from the same plasma sample was calculated . Such measurements were performed for 60 cross-sectional samples . We found that some features demonstrate higher in-host variance than others ( compare different columns in Fig 2A ) . For example , the CoV of PG9 was generally high in most patients ( i . e . , many hosts contained cocirculating viruses that had low and high binding efficiencies to PG9 ) . For other features ( e . g . , 10E8 or b12 ) , the variance among cocirculating strains was minimal ( i . e . , either high or low values in all Envs isolated from the same plasma sample ) . Consequently , the mean CoV values of each feature ( averaged for all patients in each column ) varied among probes ( Fig 2B ) . The in-host variance pattern of each feature appeared to be conserved across different patients ( see standard error bars in Fig 2B ) . Therefore , different structural features of Env appear to have different propensities for in-host variance . We also examined the in-host variance for segmental features of the five variable loops of gp120 . As expected , the V3 loop demonstrated relatively high conservation of length , charge , hydropathy score , and PNGS in each plasma sample ( see mean CoVs in Fig 2C and the complete dataset in S6 Fig ) . The limited variation in the V3 loop likely reflects the restricted range of states this segment can assume and still maintain Env functionality [33 , 86 , 87] . Other variable loops , which show greater degrees of diversity in the population than V3 ( Fig 1B and S4 Fig ) , also demonstrate higher in-host variance . We emphasize that the above-described in-host CoV does not aim to quantify the absolute level of variance that may exist for each feature; such a value cannot be accurately approximated by the limited sampling we employ ( 2–8 Envs per sample ) . Instead , it serves as a relative measure of the propensity of features for variance in the infected individual at any given time point . Broad sampling ( 60 patients ) allows us to identify such relative propensities with a good degree of confidence . Many of the features that demonstrate increased in-host variance also show high levels of diversity in the population ( e . g . , epitopes of mAbs 2G12 and PG16 ) . This suggested a possible association between in-host variance and the potential of features for diversification between hosts . We therefore sought to generate a more precise measure of the propensity of each feature for in-host variance , which could be used for quantitative comparison with its patterns of longitudinal divergence and population-level diversification . A small set of env genes randomly selected from circulating strains can show different genetic relationships; in some plasma samples , Envs differ by a single amino acid , whereas in other samples Envs can differ in ~10% of their amino acid content . To account for such differences , we corrected the phenotypic distance between Envs for the genetic distance that separates them ( see schematic in Fig 3A ) . Pairwise phenotypic distances ( e . g . , differences in binding of a probe ) between all Envs in a plasma sample were measured . Similarly , the genetic distances ( based on amino acid sequence ) between all Env pairs in a sample were determined ( see Sequence analysis in the Materials and methods section and S7 Fig ) . The ratio between the squared phenotypic distance and genetic distance was calculated and averaged for all Env pairs in that plasma sample . This measure , which we designate the volatility index , describes the propensity of features for variance within the host ( at a given time point ) per genetic distance unit: Vk=2n ( n−1 ) Σi=1n Σj=1i−1 ( Δijk ) 2γij k=1 , 2 , … , m ( 1 ) where Vk is the volatility index of the kth feature of Env , m is the total number of features , n is the number of Envs isolated from each plasma sample , Δijk is the difference between the values of the kth feature for Envs i and j , and γij is the genetic distance between amino acid sequences of Envs i and j . The volatility index is thus regarded as a constant property of each feature . It describes the propensity of the feature for variance per genetic distance unit ( rather than the level of phenotypic variance ) . Accordingly , the few Envs we isolated with identical sequences ( 16 of 523 ) were disallowed . The volatility index of each feature was calculated in each of the 60 cross-sectional patients and log-transformed to reduce the effects of extreme values on averages ( Fig 3B ) . We found that the mean volatilities of epitopes differed significantly; the epitopes of mAbs 10E8 , 2F5 , and b12 demonstrated low values relative to the epitopes of mAbs 2G12 , PG9 , and VRC03 . A high correlation was observed between volatility indices measured in samples collected in Iowa City and Seattle ( p-value of 0 . 00013 in a Spearman correlation test , Fig 3C ) . We also examined the volatility indices of segmental features of gp120 ( see Fig 3D and volatilities of all gp120 and gp41 segments in S8A Fig ) . Similar to the antigenicity features , volatility indices measured using plasma samples from Iowa City and Seattle correlated well ( p-value < 10−6 in a Spearman correlation test , Fig 3E ) . Volatility indices were also measured using a third panel of plasma samples from 20 clade B–infected individuals who enrolled in the MOTIVATE trial , which examined efficacy of the CCR5 inhibitor Maraviroc [89 , 90] ( see alignment of these Envs in S4 Data ) . Only plasma samples collected prior to initiation of Maraviroc treatment were studied . The correlation between volatility indices measured in the Seattle and MOTIVATE panels was high ( Fig 3F ) . Interestingly , the indices measured for the MOTIVATE samples were generally higher than those of the Seattle samples . We hypothesized that the small differences in volatility between panels may result from differential sampling of Envs in each group; the average number of Envs isolated from each plasma sample in the Iowa City , Seattle , and MOTIVATE cohorts was 2 . 83 , 3 . 04 , and 9 , respectively . We therefore compared the variance between volatility indices of features measured in Iowa City patients with two , three , or more than three Envs per plasma sample . As expected , the larger the number of Envs in each sample , the greater the similarity in volatility values among patients ( see Wilcoxon signed-rank test comparing standard deviations of hydropathy volatilities among patients in each group , Fig 3G ) . Similar results were obtained for the antigenicity features , whereby variance in volatility values among patients was greater when only two Envs were isolated from a plasma sample relative to three or more than three Envs ( p-values of 0 . 005 and 0 . 001 , respectively ) . Therefore , the measured volatility index is not independent of sample size . Sampling of four to nine Envs per patient appears to generate a volatility value that is conserved among different individuals . The relationship between the number of patients studied and the cumulative mean volatility of V1 loop hydropathy is graphically demonstrated in Fig 3H . Greater sampling of Envs from each patient results in reduced variance and a mean volatility value closer to that of the MOTIVATE panel using samples from less patients . In summary , the volatility index is a measure of the in-host propensity of features for variance at any given time point ( rather than absolute variance in their values ) . This property of each feature is highly conserved in different patient populations , at least in the context of viruses from the same clade . The conserved nature of volatility indices in different patients suggested that they may translate into defined longitudinal divergence patterns . We thus compared volatility with the mean divergence of Env feature values in a group of longitudinally monitored patients ( see primary data and sequences in S1 and S3 Data ) . The panel includes 18 patients; for each , we examined two to five plasma samples separated by up to 11-y intervals ( two to seven Envs were isolated from each sample ) . We note that 15 of the 18 patients were chronically infected at the time of first plasma sample collection . Only 3 patients ( UW . 1313 , UW . 1842 , and UW . 1406 ) may have been in the acute phase at the time of first sample collection ( 34 , 74 , and 159 d from first HIV+ test , respectively ) . Therefore , we view these analyses as generally representative of transitions between chronic states . To measure longitudinal feature divergence from the mixed state that exists in these patients , we used each of the isolates from the ( chronologically ) first plasma sample as a reference ( designated herein as the reference Env ( s ) ) . Phenotypic and genetic distances between each reference and all other Envs were then measured ( see schematic in Fig 4A ) . Variable loop features demonstrated gradual divergence with increasing genetic distance from the reference Env ( s ) ( see loops V1 , V3 , and V5 in Fig 4B and all loops in S9A Fig ) . As genetic distance increased , features diverged to different extents . The V1 and V5 loops exhibited the greatest degree of divergence for most feature types . By contrast , the V3 loop maintained conserved feature values , likely reflecting the limited range of lengths , PNGS , and charge values that allow effective Env trimer packing and interaction with a coreceptor on target cells [86 , 87] . To quantify the patient-averaged divergence of features , we dissected the range of genetic distances into sections of 0 . 01 units ( see vertical dashed lines in Fig 4B ) and examined the progression of variance . For some features , the trend of gradual divergence from the initial state is disrupted at genetic distances greater than 0 . 08 , likely because of the relative paucity of in-host Env pairs with such genetic distance separation . These highly divergent Envs are mainly derived from two patients that contain major recombination events ( UW . 1140 and UW . 1393 ) . The feature-specific patterns of divergence suggested that we may apply the volatility indices ( measured using the cross-sectional patient samples ) to predict the range of feature values that developed in the longitudinal patients . Based on the dispersion patterns of values ( Fig 4B ) , we modeled phenotypic changes as a linear diffusion process that progresses with genetic distance ( γ ) from the reference state . As such , we expect it to satisfy the following stochastic differential equation: dXk ( γ ) =μkdγ+σkdW ( γ ) ( 2 ) where Xk is the value of the kth feature , μk is a constant and represents the mean rate of change ( drift ) parameter , σk represents the “diffusion” index of the feature ( i . e . , the tendency for dispersion of feature value ) , and dW ( γ ) describes the incremental contribution of a random variable that is normally distributed with mean 0 and variance d ( γ ) . Traditionally , contribution of the stochastic component ( i . e . , random variable ) to evolution of features is described by a Wiener process denoted by W ( t ) [91] . To describe evolution of Env features , changes are indexed by genetic distance ( γ ) rather than time ( t ) . At this stage , we assume the absence of a deterministic component ( μk ) to the changes in feature value ( i . e . , the absence of a constant phenotypic drift ) . Thus , for a process that is solely driven by the stochastic component , variance of the increments to feature Xk is described by: Var ( ΔXk ) =σk2⋅Δγ ( 3 ) where Δγ represents the genetic distance between the reference and tested isolates . We examined whether we can substitute the volatility index measured using the cross-sectional samples ( Eq 1 ) for σk2 to predict the variance that developed in the group of longitudinally monitored patients at each genetic distance section from the ( mixed ) reference state . The measured sectional variance value was compared with the predicted value , calculated as the product of the volatility index and the genetic distance of the section analyzed . Strong relationships were observed between predicted and measured variance for the length and hydropathy of the five variable loops ( Fig 4C ) . Predictions of charge and PNGS divergence were also good , although some differences were observed between variable loops in the relationships between volatility and divergence ( compare with uniform slopes of length and hydropathy ) . Therefore , translation of volatility into longitudinal divergence is relatively similar for some feature types whereas other features exhibit more complex patterns of translation . As an alternative to the sectional approach , we also analyzed the overall pattern of divergence in each patient ( fit to a single linear regression model ) , followed by averaging of all patients . Divergence of each feature calculated using a simple regression model correlated well with the volatility index ( S9B Fig ) . By applying a sectional approach , we minimize the effects of changes that occur at greater genetic distances on the overall divergence pattern in each patient ( e . g . , by recombination events ) and can thus analyze the progression of divergence more accurately . We also examined the longitudinal divergence of individual amino acid positions of Env and the association of this process with their volatility indices . The V3 loop of HIV-1 often evolves during the course of infection and allows a switch from utilization of the CCR5 coreceptor to CXCR4 [33 , 35] . Longitudinal analyses of V3 loop features show it is highly conserved in length and PNGS whereas charge and hydropathy can alter over the course of infection ( Fig 4B ) . We examined the longitudinal evolution of hydropathy of each amino acid position of the V3 loop crown . Different propensities for longitudinal divergence were observed , as measured by the sectional variance values ( Fig 4D ) . We also measured the hydropathy volatility of each amino acid position by the same approach applied to measure segmental hydropathy volatilities , using Env panels from Iowa City and the MOTIVATE trial . Since it was previously suggested that exposure of V3 loop crown residues differs in clades B and C viruses [73] , we also measured volatility in two independent panels of Envs from plasma of clade C–infected individuals [92 , 93] ( S5 Data ) . Indeed , notable differences were observed between volatility of V3 loop tip residues in the panels of Envs from clades B and C , specifically for the sequence His-Ile-Gly-Pro-Gly-Arg at positions 308–315 ( Fig 4D ) . Volatility in this region was higher for clade B Envs . Interestingly , the same positions also exhibit increased Shannon entropy values in clade B relative to clade C viruses [73] . A strong relationship was observed between the measured sectional divergence of each amino acid hydropathy and the predicted divergence ( calculated by the product of hydropathy volatility and sectional genetic distance , Fig 4E ) . Divergence of antigenicity features was examined in the longitudinal patients . In accordance with their patterns of population diversity ( Fig 1A ) and in-host variance ( Fig 2B ) , the epitopes of mAbs 2G12 and PG9 showed significant propensities for longitudinal divergence ( see Fig 5A and entire dataset in S10 Fig ) . By contrast , divergence of the CD4-binding site probes b12 and CD4-Ig and the MPER-targeting probes 10E8 and 2F5 was limited . A strong relationship was observed between the predicted and measured sectional divergence values ( Fig 5B ) . In some cases , ( e . g . , PG9 and 2F5 ) , the gradual divergence from the initial state becomes less apparent at genetic distances greater than 0 . 08 . We attribute this change in pattern to the relative paucity of Env pairs with such large genetic distance separation . Nevertheless , overall the selective forces that act in the individual appear to be sufficiently stable over time to allow application of the volatility index to predict the propensity for divergence of each feature . The lengths of the Env variable loops are not normally distributed in the population ( see histograms and results of D'Agostino–Pearson Omnibus test in S11 Fig ) . The observed patterns suggest existence of constraints that may limit the appearance of Envs with loop lengths below or above certain values . Such constraints may affect the longitudinal divergence from some reference states . For example , analysis of changes in the length of V5 from a reference state that contains only 9 amino acids show mainly positive increments relative to the more symmetric changes from a reference state with 11 amino acids ( Fig 5C ) . Thus , restraints imposed by biological properties of Env can affect feature evolution . However , such bounds are often not absolute . For example , a V5 loop that contains 15 amino acids is less preferred; longitudinal analyses reveal that such Envs will normally “drift” back to a state that contains a shorter V5 loop ( Fig 5C ) . Nevertheless , two patients in the Iowa City panel have viruses that contain even larger ( 17 amino acid ) V5 loops . Therefore , the restraints imposed on V5 length are better described by preferences for specific states . Restraints can also be applied by inherent properties of the features themselves . For example , acquisition of an epitope is ( in many cases ) less likely than loss of the epitope ( e . g . , by a single point mutation ) . Diffusion from a state that lacks an epitope is thus restrained ( Fig 5D ) . Therefore , the short-range diffusion of feature values ( i . e . , for small genetic distance increments ) is controlled by the initial state of the feature and by the volatility index . Accordingly , rather than treating evolution as a symmetric diffusion process that is only determined by volatility , we can introduce a state-specific drift component ( μk ( Xk ( γ ) ) ) , which is composed of the restraining forces imposed by properties of the molecule or the feature . For a process affected by volatility and such a drift the incremental change in feature value can be described by: dXk ( γ ) =μk ( Xk ( γ ) ) dγ+Vk⋅dγ dW ( γ ) ( 4 ) We assume that over short genetic distances , the drift component is a constant that depends on the reference state . For longer-range paths , the expression should accommodate the dynamic changes that occur over increasing genetic distances ( see Discussion section ) . In summary , the volatility index of each Env feature provides an accurate measure of the mean degree of longitudinal divergence expected to occur in patients . Although different pressures are likely applied in different patients , the “noise” measured in the individual at any moment is highly correlated with the propensity for change over the course of time . A few features demonstrate imperfect correlations between the predicted and measured divergence ( e . g . , PG9 , 10E8 , and 2F5 ) . Such variations may result from the following: ( i ) limited sampling of some genetic distance sections ( i . e . , paucity of Env pairs at greater genetic distances ) , ( ii ) different pressures applied in different individuals , ( iii ) low volatility and divergence values ( e . g . , for mAbs 10E8 and 2F5 ) , and ( iv ) lack of complete diffusion symmetry from all states ( i . e . , dependence on initial feature values ) . Nevertheless , overall the volatilities of individual amino acids , Env segments , and epitopes are highly conserved among patients and allow accurate predictions of the mean divergence expected in a group of longitudinally monitored individuals ( Figs 4C , 4E and 5B ) . Diversification of Env features in the population is affected by their propensity for divergence within patients and the selective forces that act during transmission [49 , 94 , 95] . We hypothesized that if longitudinal patterns of change are sufficiently conserved among different individuals and if selective pressures that act during virus transmission have comparable magnitude , then volatility could be applied to predict population-level changes in each feature . We thus examined whether the diversification patterns of Env features over the past three decades ( Fig 1 ) can be explained by differences in feature volatilities . We tested this relationship for the antigenicity features; volatility was measured using patient samples from Period1 or Period3 and compared with the diversity of the features in the population during Period3 or Period1 , respectively . A linear relationship between volatility and diversity was observed ( Fig 6A ) . Interestingly , volatility in Period1 samples served as a better predictor of feature diversity during Period3 than vice versa . Such a pattern could result from limited diversity of these features during Period1 or from changes in volatility from Period1 to Period3 ( see S5 , S8B and S8C Figs ) . Nevertheless , whether in-host volatility is stable or dynamic over time , this inherent property of each feature is translated in a defined and predictable manner into its population-level diversity . We also examined the relationship between volatility and historic changes in segmental features of gp120 . Volatility measured using the 20 patient samples of the MOTIVATE trial was compared with the historic diversification of each feature in Iowa City ( measured by the difference in diversity between Period3 and Period1 ) . For both segmental length and hydropathy , we observed a clear linear relationship between in-host volatility and the changes that occurred in feature diversity between the two periods ( Fig 6B ) . By contrast , charge and PNGS showed a nonuniform association pattern . Comparison between volatility and P3 diversity of charge and PNGS also showed that comparable levels of population diversity may exist despite significant differences in volatility ( S12 Fig ) . Therefore , all antigenic features we tested and some segmental feature types show direct “translation” of their in-host volatility into population-level diversity . For other feature types , translation is not identical for all segments , suggesting potential involvement of additional factors . The above-described changes in segmental features describe evolution of Env structure at “low resolution” ( i . e . , the context in which epitopes are expressed ) . We sought to examine the ability of the diffusion-based model to predict changes in amino acid sequence of antigenically significant regions of Env . We examined whether application of the volatilities of individual positions of Env would allow us to predict the specific amino acid variants that appeared in the Iowa City population over the course of three decades . We first examined the V3 loop crown . Similar to the segmental features , hydropathy volatility of each amino acid ( calculated using the 20 plasma samples of the MOTIVATE trial ) correlated well with its diversity in the Iowa City population during Period3 ( Fig 7A ) . Positions that show limited or no variance in the infected individual at any time point also show minimal longitudinal changes ( Fig 4D ) and were unaltered in the population over the course of three decades . We hypothesized that if the sequence of Env was sufficiently conserved during Period1 then we could predict both the level of diversity that developed and potentially the specific amino acid variants that appeared in the population . For this purpose , we applied the volatility of three features of each amino acid ( charge , molecular weight , and hydropathy ) in a joint probability density function . The consensus of Period1 Envs in Iowa City was used as the reference ( “ancestral” ) sequence . For each position , we then calculated the likelihood of each amino acid variant ( k ) to evolve from the ancestral state ( α ) based on the propensity for change in that feature ( i . e . , volatility ) . We thus treat the changes that occur in properties of amino acids at each position as a one-dimensional random walk; for each position , we measure the likelihood of the feature value to “diffuse” away from its ancestral state to any other amino acid value . A schematic example of the approach is provided in S13 Fig . We assume that the above feature types are normally distributed and apply a probability density function: Pki=f ( xk|αk , Vk ) =12πVke− ( xki−αk ) 22Vkk=1 , 2 , … , m; i=1 , 2 , … , n ( 5 ) where Pki is the likelihood of obtaining the ith variant of an amino acid based on volatility of the kth feature type ( Vk ) , xki is the value of the ith variant of feature k , and αk is the value of the reference state for feature k . This calculation is repeated for each feature type ( hydropathy , molecular weight , and charge ) , and the combined likelihood ( PTotali ) of obtaining the ith variant of an amino acid based on volatilities of all feature types is defined by: PTotali=∏k=1m Pki ( 6 ) We compared the consensus sequence of the V3 loop crown in Iowa City during Period1 ( similar to that of the clade B ancestor ) with the distribution of sequence variants in the Iowa City population during Period3 and the predicted range of variants ( see sequence logo representation in Fig 7B ) . We found that the expression performed well; positions that were predicted to diversify minimally ( based on volatility of the three features ) also showed limited or no diversity in the population . Positions that showed high volatility in patients also demonstrated greater diversity in the population . Thus , using measurements from 20 plasma samples ( and the ancestral sequence ) we can predict for many positions the diversity that developed over the course of 30 y . We emphasize that the approach does not take into account the likelihood of each amino acid substitution; incorporation of substitution likelihoods further improves the predictive capacity of this basic model . We also note that the model treats all amino acid residues as independent variables and does not acknowledge the well-defined networks of association that exist within the V3 loop [89 , 96 , 97] . Inclusion of such considerations is expected to improve performance of the model still further . The model was also tested for the ability to predict changes in the MPER of gp41 . Several epitopes of broadly neutralizing antibodies map to this Env region [68 , 98 , 99] . Similar to the V3 loop crown , hydropathy volatility of each amino acid position correlated well with its diversity in the population during Period3 ( Fig 7C ) . The N-terminal portion of the MPER , which contains the 2F5 epitope , was relatively more volatile than the C-terminal part , which contains the 10E8 epitope . Such a pattern is expected of the conserved C-terminus , which interacts with cholesterol [100–102] and can regulate global sensitivity of Env to antibodies [103] . These data also correlate well with antigenicity results , which show conserved integrity of the 10E8 epitopes but some diversification ( albeit limited ) of the 2F5 epitope over the past three decades ( Fig 1A ) . Interestingly , the highest volatilities in the MPER were measured at positions 662 and 674 . Position 662 is associated with changes that regulate coreceptor tropism , from CCR5 to CXCR4 [104] . Position 674 is associated with regulating the global responsiveness of Env to inhibitory agents ( such as antibodies ) and to Env-activating molecules ( such as the coreceptors ) [59] . Changes at this position allow transition from a state of increased fusogenicity and sensitivity to antibodies ( advantageous in vitro ) to a state of reduced fusogenicity but also reduced sensitivity to antibodies ( advantageous in vivo ) [59 , 105 , 106] . The relatively frequent “fluctuations” at these positions may allow the virus to achieve such phenotypic switches more effectively and thus to rapidly adapt to the environment . An interesting discrepancy was observed for position 681 between the intermediate-level volatility measured in the MOTIVATE trial samples and the complete conservation of this residue ( Tyr ) in the Iowa City population and among all group M , N , O , and P strains of HIV-1 [107] . Indeed , Envs containing mutations at this position are often fusion-competent [108] . That variants at position 681 are found among cocirculating strains ( in highly sampled individuals ) but do not appear in the general population suggests the involvement of selective pressures applied on this site over time in the individual or during virus transmission . Discrepancies between the levels of in-host variance , longitudinal divergence , and population diversity allow us to identify bottlenecks that restrict the continuity of heterogeneity across time and different patients and permit preferential spread of only selected forms of the virus . Similar to the V3 loop crown , application of the joint probability density function allowed us to predict well the positions that remained unchanged and often the variants that evolved from the ancestral state and currently circulate in the population ( Fig 7D ) . Therefore , through limited sampling of the population ( i . e . , using volatility indices measured from 20 patient samples ) we can predict the diversity that developed at many Env positions and approximate well the nature of the specific amino acids .
For all Env features , variance gradually increases with genetic distance from the initial reference state ( s ) , both within the infected individual and in the population . Different features demonstrate different ( but conserved ) progression “rates” of variance per genetic distance unit . Such propensities for variance can be clade-specific and potentially account for observed diversification patterns of features in the different clades . We thus model the evolution of feature values as a linear diffusion process that is controlled by volatility . Accordingly , the value of feature Xk at a given genetic distance γ from the reference state Xk ( 0 ) can be described by: Xk ( γ ) =Xk ( 0 ) +μk ( Xk ( γ ) ) ⋅γ+ Vk⋅γ⋅W ( γ ) ( 7 ) Therefore , Xk ( γ ) is normally distributed , with an expected value ( mean ) of Xk ( 0 ) + μk ( Xk ( γ ) ) · γ and variance Vk · γ . The expression distinguishes between contributions of the stochastic and deterministic components . For diffusion across short genetic distances , we can assume that the drift component μk ( Xk ( γ ) ) is still controlled by the reference state value . However , at greater genetic distances from the reference state , the drift is less well defined . In this work , which primarily focuses on the population-level spread of features , we acknowledge the presence of a state-specific deterministic drift but primarily focus on the stochastic component . By presenting our results in the context of the complete model , we aim to demonstrate how evolution of phenotypes can be captured by a diffusion process and the possibilities this approach offers . Changes in antigenicity features of Env ( in patients and population ) are treated as a diffusion process—a Markov process indexed by genetic distance ( rather than time ) with sample paths that are almost surely continuous . Sample paths for binding efficiencies of probes are indeed clearly continuous . We apply a similar approach for analyzing the chemical properties of amino acids ( charge , hydropathy , and molecular weight ) , which are approximated here to be normally distributed . Accordingly , the process of changes in chemical properties of amino acids from a defined reference state represents a diffusion approximation . Evolution of Env features differs from other systems typically described by random walk-like changes ( e . g . , displacement of colloidal particles or evolution of stock option prices ) in several ways: ( i ) progression is indexed by genetic distance , ( ii ) features are characterized by partial sampling from a mixed state ( characterization of all states present is not viewed as a feasible option ) , and ( iii ) increments cannot be assumed to be completely independent ( although they are treated as such in this study ) . Specific characteristics of this form of feature evolution , which we expect can be applied to describe other biological systems , are discussed below . Stochastic differential equations are used in many scientific disciplines to model systems that contain a dominant “uncertainty” component [117–124] . In many of these studies , a generalized Wiener process is used to describe the random function . It is a normally distributed , continuous-time stochastic process with independent increments . Here , we define a stochastic process we designate Wγ as the single “generator” of random variables ( i . e . , as the randomness-introducing function ) . Its contribution is controlled by the volatility of the feature . Since we preselect all Envs for functionality , the increments represent the combined effect of diversity-increasing forces ( mutations and recombination events ) and diversity-decreasing forces ( immune- and fitness-selective forces applied on the molecule ) . Therefore , Wγ is likely better represented not by a single continuous process . Indeed , two types of events introduce genetic ( and thus phenotypic ) change; point mutations and recombinations [8 , 10 , 11] . In 2 of the 18 longitudinal patients , we observed major recombination events ( as predicted by the RDP4 software [125] ) . Both patients showed significant genetic distance leaps ( see extreme genetic diversities in Figs 4B and 5A and heavy-tailed distribution in S14 Fig ) . Such genetic “leaps” can potentially account for “leaps” in virus phenotypes . Therefore , although we currently model increments as part of a single , normally distributed process that we designate Wγ , the data could be represented more accurately by several random processes that introduce different increments , such as those described by the Merton jump-diffusion model of option pricing [126 , 127] . Future studies will compare the effects of single-site mutations and recombination events on evolution of feature values . The volatility index describes the propensity of each feature for variance with increasing genetic distance . Conservation of volatility across multiple patients from different geographic regions illustrates its robustness in the context of within-clade analyses . Whereas volatility is treated in this study as a constant , it is clear that , similar to the drift component , this parameter may show state-conditional effects and potentially non-Markov properties [128 , 129] . Volatility can be divided into two “tiers . ” Preselection volatility describes the range of feature values that can appear in the infected individual in the absence of any restraining selective pressures . This theoretical range of variants includes all potential progeny isolates , functional and nonfunctional . The measured ( postselection ) volatility accounts for the effect of immune and fitness constraints and describes only the functional variants that circulate in patients . Assuming a constant rate of mutations and a homogenous distribution of the mutations across the env gene , preselection volatility is controlled only by the “complexity” of the feature . Complexity describes sensitivity of the feature value to a random change in amino acid sequence ( i . e . , the number of residues that affect the feature ) . In our analyses , we rendered complexity comparable for segmental features of the variable loops by normalizing each for loop length ( i . e . , volatility is calculated per amino acid ) . Therefore , differences between measured volatilities of segmental features reflect the effects of selective forces rather than feature complexity . By contrast , for the antigenicity features , different epitopes involve different numbers of residues in formation or maintenance of their integrity ( i . e . , varying degrees of complexity ) . Whereas most Env epitopes recognized by neutralizing antibodies are discontinuous , we included two probes that recognize linear epitopes; 10E8 and 2F5 . Their epitopes thus likely have relatively low complexity . Indeed , the volatility of both epitopes was low . In particular , volatility of 10E8 was also low since it is associated with regulating function and global antibody sensitivity of Env [103 , 130] and is therefore under greater selective pressure . In a similar manner , volatility of individual amino acids is controlled by immune and fitness pressures applied on each position . The effect of such pressures is associated with the degree of solvent exposure of the residue [131 , 132] . As expected , our data suggest that a cryptic state is associated with lower volatility ( Fig 4D ) . Our analysis of phenotypically mixed states through partial sampling does not seek to capture the entire diversity in the system; it is appreciated that many states remain unsampled . Envs that we isolated and tested likely represent the more abundant quasispecies circulating in the individual and thus ( potentially ) are more transmissible relative to other variants and quiescent forms [36 , 49 , 95 , 133] . The basic model described here does not ( yet ) account for the potentially dynamic nature of the selective forces applied; all Envs from the longitudinal patients are treated as part of one compartment and indexed by genetic diversity from the reference isolate ( s ) . Accordingly , the factors that can alter with time in the infected individual ( e . g . , replication rate and selective pressures ) are treated as uniform for all samples . Similarly , other than excluding all patients treated by entry inhibitors , the model does not yet account for the potential effects of antiretroviral therapy on virus diversity at each time point . For increased accuracy , the models will be adapted to describe the state of such dynamic variables of the virus and immune system in each unique environment as well as functional features of Env [114 , 115] . Vaccine immunogens are selected according to the clades that circulate in the target population . Our results suggest that in addition to phylogenetic considerations , vaccines should likely also address the dynamic nature of Env structure in the population . For example , the epitopes of mAbs VRC03 and 2F5 were similarly distributed in HIV-infected individuals during the 1980s ( Fig 1A ) . However , the VRC03 epitope is currently present only in ~30% of the population ( and will likely continue to disappear ) whereas the 2F5 epitope is found in ~90% of circulating strains . When such historic information is available ( by extensive sampling of infected individuals over the course of decades ) , predictions of future states can be generated based on past patterns of change . However , such information is not available for the many HIV-1 clades and recombinant forms that infect individuals worldwide . Therefore , reliable predictors of changes expected to occur in Env structure are required . The volatility index provides the clues necessary to achieve this goal , based on minimal sampling of patients . Knowledge of the population longevity of epitopes will likely be beneficial to immunogen design . The basic tools described here allow accurate predictions of changes in feature diversity and ( for some amino acid positions ) the nature of variants expected to evolve at key antigenic sites . By taking into account the network of associations that exists within Env and the coevolutionary patterns of its components [96 , 97] , the accuracy of these predictions can be further improved . The flexibility imparted by the model , which is based on the stochastic component of change but also combines the effects of deterministic drifts , will likely facilitate analyses of changes in other dynamic biological systems .
This study involved the use of peripheral blood samples from HIV-infected adult subjects who gave informed consent under clinical protocols approved by the participating institutions' human use review boards , including those at the University of Washington at Seattle Center for AIDS Research ( CFAR ) and University of Iowa ( IRB numbers 8807313 , 200010008 , and 2010101730 ) . This study did not involve animal research ( vertebrate animals , embryos , or tissues ) . This is not a field of study , nor did it involve collection of plant , animal , or other materials from a natural setting . Blood was collected from patients in tubes containing acid citrate dextrose . Plasma was then separated , divided into aliquots and stored at −80°C until use . Viral load in all plasma samples was measured; samples that contained a measurable viral load ( generally greater than 1 , 000 copies per ml ) were further processed for isolation of the viral env genes , as described below . To isolate RNA , plasma samples were centrifuged twice at 5 , 500 g for 5 min to remove cell debris . Virus particles were then pelleted by centrifugation at 21 , 000 g for 3 h at 10°C , and the supernatant was removed and stored at –80°C for future use . Pellets were then resuspended in 140 μl of 150 mM NaCl , and RNA was extracted using QIAamp MinElute Virus Spin kit ( Qiagen ) . Extraction was performed as recommended by the manufacturer except that instead of using carrier RNA , the AL buffer was supplemented with 40 μg/ml acrylamide ( Sigma ) to allow RNA sequencing of the samples . RNA was recovered from spin columns in 30 μl water and then either frozen at −80°C or immediately used to synthesize cDNA . SuperScript III ( Invitrogen ) was used for cDNA synthesis with 0 . 5 μM of the reverse primer Env3out ( 5′-TTGCTACTTGTGATTGCTCCATGT-3′; nucleotides 8913 to 8936 according to HXB2 numbering [88] ) , as previously described [134] . The cDNA from the reverse transcription reaction served as template for PCR amplification of the 3-kb fragment containing the env and rev genes . cDNA was serially diluted in water in replicates of three PCR wells and subjected to a nested PCR reaction . The high-fidelity polymerase PrimeStar Max ( Takara ) was used for first- and second-round PCR reactions . First-round PCR was performed with 1 mM MgCl2 , 0 . 2 mM of each dNTP , 0 . 4 μM of forward primer Env5out ( 5′-TAGAGCCCTGGAAGCATCCAGGAAG-3′; nucleotides 5853 to 5877 ) , and 0 . 4 μM of the reverse primer Env3out in an 8-μl reaction mixture . PCR conditions were 94°C for 30 s , followed by 38 cycles of 94°C for 15 s , 55°C for 30 s , and 68°C for 3 min , with a final extension period of 8 min at 68°C . Second-round PCR was performed by diluting the product of the first round 10-fold in water and then transferring 1 μl of the diluted first-round product into a final volume of 8 μl of a reaction mixture containing 0 . 2 mM of each dNTP , 0 . 4 μM of forward primer Env5in ( 5′- GGCATCTCCTATGGCAGGAAGAAG-3′; nucleotides 5960 to 5983 ) , and 0 . 4 μM of reverse primer Env3in ( 5′-GTCTCGAGATACTGCTCCCACCC-3′; nucleotides 8904 to 8882 ) . PCR conditions were identical to the first-round PCR . Dilutions that yielded approximately one of three PCR-positive wells were then retested in 12 replicates to identify a dilution in which <30% of wells were positive for amplification products . The amplified fragments from cDNA dilutions in which <30% of wells were positive were then purified from agarose gels using PureLink gel extraction kit ( Invitrogen ) . Purified fragments were cloned into the pSVIIIenv vector [135] using Infusion cloning system ( Clontech ) . All cloned env genes were then screened for functionality of their protein product by generating recombinant viruses that contain each Env and testing their infection of CD4+CCR5+ and CD4+CXCR4+ cells , as detailed below . Envs that mediated infection of either cell type were further analyzed whereas the remaining Envs were archived for future tests . From each plasma sample , we isolated in the above manner one to eight functional env genes . All functional Envs were fully sequenced ( see list of accession numbers in S1 Data ) . Envs that were identical in amino acid sequence to an already existing Env ( 16 of 523 functional Envs isolated thus far ) were discarded . Similarly , for all clade B and C sequence datasets used to calculate volatility indices , identical protein sequences were excluded . Protein sequences were aligned using a Hidden Markov Model with the HMMER3 software [136] . Since automated algorithms cannot perfectly align Env sequences ( mainly because of insertions and deletions ) , we edited the HMMER3 alignment product manually [137] . For calculation of genetic pairwise distances , gapped sites were not counted in the distance calculations unless present in 95% of the aligned sequences . This cutoff was applied to minimize the false similarities between sequences introduced by gapped sites when using sequences that are highly divergent . The use of multiple sequence alignment for measuring genetic distances between isolates allows rapid calculations . This tool is thus well suited for very large datasets . Nevertheless , genetic distances calculated by this approach are affected by insertions and deletions relative to pairwise alignment tools . We also calculated the volatility indices using a basic pairwise alignment tool ( ClustalW ) . Comparison of the mean volatility indices from 22 patient samples calculated using genetic distances from the multiple sequence and pairwise alignment tools showed limited differences ( S7 Fig ) . The greater precision of pairwise alignments , which allow customization using HIV-specific scoring matrices [138] , contributes primarily to analyses that include a small number of patient samples . Phylogenetic trees were reconstructed from protein sequences using the maximum likelihood method with an HIVb ( between patient ) amino acid substitution model using PhyML3 . Sequence features of Env segments , including length , charge , and number of PNGSs were examined using in-house generated macros for Excel and the “Variable Region Characteristics” tool of the Los Alamos website ( hiv . lanl . gov ) . Boundaries of Env segments ( shown in Fig 3D and S8A Fig ) conform to the standard segmentation of the env gene specified in the above online tool and are based on standard HXBc2 numbering of env [88] . PNGSs were defined by presence of the sequence Asn-X-Ser/Thr , in which X can be any amino acid except Pro . The mean hydropathy score for each loop was calculated based on the Black and Mould scale [74] , which places values on a scale of 0 to 1 . Sequence analysis of the cloned envs suggests that the high-fidelity polymerase used in both rounds of the nested PCR ( error rate 7 . 6 × 10−7 by sequencing ) introduced minimal errors and therefore had a limited effect on the observed variance in segmental and antigenic features . For the amplification protocol we applied ( and given the above error rate ) , we would expect that ~17% of the 3-kb products would contain an error . Indeed , 16 of the functional Envs we isolated were identical in amino acid sequence to other Envs isolated from the same plasma sample but in different amplification reactions ( these Envs were discarded ) . Twelve additional Envs contain a single amino acid difference from another isolate amplified separately . Therefore , although we do not exclude that some changes may have occurred because of mutations during in vitro amplification , the effect of such changes on sequence and antigenic features is likely minimal . Single-round , recombinant HIV-1 viruses that express the luciferase gene were generated by transfection of human embryonic kidney 293T cells ( obtained from the American Type Culture Collection , ATCC ) using JetPrime transfection reagent ( Polyplus ) . Briefly , cells were seeded in six-well plate wells ( 8 . 5 × 105 cells per well ) and transfected the next day with 0 . 4 μg of HIV-1 packaging construct pCMVΔP1ΔenvpA , 1 . 2 μg of firefly luciferase–expressing construct pHIvec2 . luc , 0 . 4 μg of a plasmid-expressing HIV-1 Env , and 0 . 2 μg of a plasmid-expressing HIV-1 Rev . On the next day , transfection medium was changed to culture medium ( DMEM/10% FBS ) . Virus-containing supernatants were collected on the following day , cleared of cell debris by low-speed centrifugation , and filtered through 0 . 45-μm filters . Viruses were then used for infection or snap frozen on dry ice , immersed in ethanol for 15 min , and stored at –80°C until use . Canis familiaris thymus normal ( Cf2Th ) cells ( obtained from the NIH AIDS Reagent Program ) expressing CD4 and CCR5 ( Cf2Th CD4+CCR5+ ) or CD4 and CXCR4 ( Cf2Th CD4+CXCR4+ ) were used as target cells for measuring infection . Approximately 5 h before infection , cells were detached from culture plates using PBS supplemented with 7 . 5 mM EDTA and were seeded in 96- or 384-well , luminometer-compatible plates ( at a density of 14 or 4 . 5 × 103 cells per well , respectively ) . Viruses were then added to the cells and further incubated for 3 d , at which time the medium was removed; cells were lysed with passive lysis buffer ( Promega ) and subjected to three freeze–thaw cycles . To measure luciferase activity , 100 μl of luciferin buffer ( 15 mM MgSO4 , 15 mM KPO4 [pH 7 . 8] , 1 mM ATP , and 1 mM dithiothreitol ) and 50 μl of 1 mM D-luciferin potassium salt ( Syd Labs , MA ) were added to each sample in 96-well plates ( 30 μl and 15 μl , respectively , for samples in 384-well plates ) . Luminescence was recorded using a Synergy H1 microplate reader ( BioTek Instruments ) . The monoclonal antibodies ( mAbs ) indicated below were obtained through the NIH AIDS Reagent Program , Division of AIDS , NIAID , NIH . The mAb 39F that targets the V3 loop of Env was contributed by James Robinson [72 , 139 , 140] . The mAb 10E8 that targets the MPER of gp41 was contributed by Mark Connors [68] . Hermann Katinger provided the MPER-targeting mAb 2F5 [70] and mAb 2G12 , which targets a carbohydrate-dependent gp120 epitope [65] . The mAb IgG1 b12 , which recognizes the CD4-binding site of gp120 [141 , 142] , was a kind gift from Dennis Burton . The CD4-binding site mAb VRC03 was provided by John Mascola [143] . The International AIDS Vaccine Initiative ( IAVI ) Neutralizing Antibody Consortium kindly provided mAbs PG9 and PG16 that target overlapping , trimer-dependent epitopes [67] and mAbs PGT121 and PGT126 , which target partially overlapping epitopes on gp120 that contain glycan and protein components [63] . The CD4-Ig fusion protein is composed of the Fc region of human IgG1 linked to two copies of the two N-terminal domains of the CD4 molecule . The CD4-Ig protein was produced and purified as previously described [58 , 144] . Binding of probes to HIV-1 Env trimers expressed on human osteosarcoma ( HOS ) cells ( obtained from the ATCC ) was measured using a modified protocol of the cell-based enzyme-linked immunosorbent assay ( ELISA ) described previously [58 , 59] . Briefly , HOS cells were seeded in 96-well plates ( 1 . 2 × 104 cells per well ) and transfected after 6 h with 55 ng of a plasmid-expressing Env and 12 ng of a Tat-expressing plasmid per well using 0 . 18 μl per well of JetPrime ( Polyplus Inc . ) transfection reagent . For experiments performed in 384-well plates , each well contained 4 . 5 × 103 cells , which were transfected with 26 ng of an Env-expressing plasmid and 5 . 5 ng of a Tat-expressing plasmid using 0 . 08 μl JetPrime reagent . In all experiments , a negative control plasmid was used that contains a stop mutation in amino acid position 46 of Env ( according to standard HXBc2 numbering [88] ) to determine background binding of each probe to the cells . Three days after transfection , cells were washed twice with blocking buffer ( 20 mg/ml BSA , 1 . 8 mM CaCl2 , 1 mM MgCl2 , 25 mM Tris [pH 7 . 5] , and 140 mM NaCl ) and incubated with the indicated probes in blocking buffer for 45 min . Unless indicated otherwise , all mAbs were added at 0 . 5 μg/ml whereas CD4-Ig was added at 2 μg/ml . Binding of each probe to the Envs is normalized for the level of Env expression using this saturating concentration of CD4-Ig [58] , which binds to the highly conserved CD4-binding site on Env . Relative to other methods of normalization for the level of expression ( e . g . , by polyclonal sera from multiple patients ) , the use of CD4-Ig is less affected by the variable antigenicity of the different isolates [111] . All samples were then washed six times with blocking buffer and incubated with a horseradish peroxidase ( HRP ) -conjugated goat antihuman IgG polyclonal antibody preparation for 45 min . Cells were subsequently washed six times with blocking buffer and six times with washing buffer ( 140 mM NaCl , 1 . 8 mM CaCl2 , 1 mM MgCl2 , and 20 mM Tris [pH 7 . 5] ) . HRP enzyme activity was determined after addition of 35 μl per well of a 1:1 mix of SuperSignal West Pico Chemiluminescent peroxide and luminol enhancer solutions ( Thermo Fisher Scientific ) supplemented with 150 mM NaCl . To samples in 384-well plates we added 25 μl of the reagent mix . Light emission was measured with a Synergy H1 reader . All software for the Data Processing , Archiving and Exploration Platform was custom developed in-house with the software company Bio::Neos ( Coralville , IA ) . Cell-based ELISA measurements ( exported from the luminometer and expressed as relative light units [RLUs] ) are processed by the software . Reliability indices are assigned to each set based on the expression level of the Env , the quality of the reads for the negative and positive controls , and the measured variance between the three replicates tested . Binding values are then associated with each Env and stored in a MySQL database . All features of each Env and experimental results can be queried and exported using a graphical user interface ( GUI ) . The amino acid sequence of each Env is also archived in the database , allowing querying of different sequence and segmental features . Analyses of antigenic and segmental data were performed using in-house-designed Excel VBA macros . The K2 Omnibus statistic of the D'Agostino and Pearson test was calculated using GraphPad Prism version 6 . 00 for Windows ( Graphpad Software ) . All other statistics , including Levene’s test and Generalized Estimation Equations ( GEE ) , were performed using R Studio Version 2 . 11 . 1 with the car and geepack software packages , respectively . GEE was performed by defining the feature value as the dependent variable , which is approximated by the Period ( i . e . , Period1 or Period3 ) . Calculations were performed by using an identity vector to cluster unique patients and a Gaussian function for link and variance , with an exchangeable correlation structure . The numerical output of the cell-based ELISA spans a range of five orders of magnitude . Data are normalized for the cell-surface expression levels of each Env using the CD4-Ig probe and are expressed as percent binding of the probe to the control AD8 Env [111] . Our previous work has shown that the biological relevance of a given fold-change in binding efficiency is not identical throughout the 5-log dynamic range ( e . g . , the interval between 10% and 100% is not equivalent to the interval between 0 . 01 and 0 . 1% ) [59 , 145] . Based on these studies , we determined the parameters for a logistic function that “trims” lower and upper extreme values: xc= ( 41−e−k ( x−2 ) ) ( 8 ) where xc is the logistic function–corrected value , k is the slope ( calculated as 0 . 6 ) , and x is the log-transformed binding value . The comparison between distribution of the raw ( log-transformed ) and logistic function–corrected data is shown in S2 Fig .
|
HIV-1 is the causative agent of the global AIDS pandemic . The envelope glycoproteins ( Envs ) of HIV-1 constitute a primary target for antibody-based vaccines . However , the diversity of Envs in the population limits the potential efficacy of this approach . Accurate estimates of the range of variants that currently infect patients and those expected to appear in the future will likely contribute to the design of population-targeted immunogens . We found that different properties ( features ) of Env have different propensities for small “fluctuations” in their values among viruses that infect patients at any given time point . This propensity of each feature for in-host variance , which we designate “volatility” , is conserved among patients . We apply this parameter to model the evolution of features ( in patients and population ) as a diffusion process driven by their “diffusion coefficients” ( volatilities ) . Using volatilities measured from a few patient samples from the 1980s , we accurately predict properties of viruses that evolved in the population over the course of 30 years . The diffusion-based model described here efficiently captures evolution of phenotypes in biological systems controlled by a dominant random component .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"sequencing",
"techniques",
"united",
"states",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"evolutionary",
"biology",
"pathogens",
"split-decomposition",
"method",
"immunology",
"geographical",
"locations",
"microbiology",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"north",
"america",
"multiple",
"alignment",
"calculation",
"rna",
"viruses",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"immune",
"system",
"proteins",
"sequence",
"alignment",
"bioinformatics",
"proteins",
"medical",
"microbiology",
"hiv",
"antigens",
"microbial",
"pathogens",
"hiv-1",
"molecular",
"biology",
"evolutionary",
"genetics",
"people",
"and",
"places",
"biochemistry",
"computational",
"techniques",
"viral",
"pathogens",
"database",
"and",
"informatics",
"methods",
"genetics",
"physiology",
"biology",
"and",
"life",
"sciences",
"iowa",
"human",
"genetics",
"lentivirus",
"protein",
"sequencing",
"organisms"
] |
2017
|
Accurate predictions of population-level changes in sequence and structural properties of HIV-1 Env using a volatility-controlled diffusion model
|
The neglected human diseases caused by trypanosomatids are currently treated with toxic therapy with limited efficacy . In search for novel anti-trypanosomatid agents , we showed previously that the Crotalus viridis viridis ( Cvv ) snake venom was active against infective forms of Trypanosoma cruzi . Here , we describe the purification of crovirin , a cysteine-rich secretory protein ( CRISP ) from Cvv venom with promising activity against trypanosomes and Leishmania . Crude venom extract was loaded onto a reverse phase analytical ( C8 ) column using a high performance liquid chromatographer . A linear gradient of water/acetonitrile with 0 . 1% trifluoroacetic acid was used . The peak containing the isolated protein ( confirmed by SDS-PAGE and mass spectrometry ) was collected and its protein content was measured . T . cruzi trypomastigotes and amastigotes , L . amazonensis promastigotes and amastigotes and T . brucei rhodesiense procyclic and bloodstream trypomastigotes were challenged with crovirin , whose toxicity was tested against LLC-MK2 cells , peritoneal macrophages and isolated murine extensor digitorum longus muscle . We purified a single protein from Cvv venom corresponding , according to Nano-LC MS/MS sequencing , to a CRISP of 24 , 893 . 64 Da , henceforth referred to as crovirin . Human infective trypanosomatid forms , including intracellular amastigotes , were sensitive to crovirin , with low IC50 or LD50 values ( 1 . 10–2 . 38 µg/ml ) . A considerably higher concentration ( 20 µg/ml ) of crovirin was required to elicit only limited toxicity on mammalian cells . This is the first report of CRISP anti-protozoal activity , and suggests that other members of this family might have potential as drugs or drug leads for the development of novel agents against trypanosomatid-borne neglected diseases .
The pathogenic trypanosomatids from the genera Leishmania and Trypanosoma infect over 20 million people worldwide , with an annual incidence of ∼3 million new infections in at least 88 countries . An additional 400 million people are at risk of infection by exposure to insect vectors harboring parasites [1]–[3] . Leishmania and trypanosome infections predominate in poorer nations , and are considered neglected diseases that have “fallen below the radar of modern drug discovery” [4] . Leishmania parasites cause five different disease forms – cutaneous ( CL ) , mucocutaneous ( MCL ) , diffuse cutaneous leishmaniasis ( DCL ) , post-kala-azar dermal leishmaniasis ( PKDL ) and visceral leishmaniasis ( VL , also known as ‘black fever’ or ‘kala-azar’ in India ) [5] . VL is the most severe and debilitating form of leishmaniasis , and can be fatal if left untreated . First-line treatment for leishmaniasis is based on pentavalent antimonials such as meglumine antimoniate ( Glucantime ) and sodium stibogluconate ( Pentostan ) . Amphotericin B and pentamidine are used as second-line drugs in patients resistant to first-line therapy [1] , [6] . Recently , miltefosine has been used in India as part of combination therapy regimens to treat VL , and the largest increase in miltefosine activity was seen in combination with amphotericin B [7] , [8] . There are two forms of HAT ( also known as sleeping sickness ) , caused by two subspecies of T . brucei parasites ( T . b . gambiense or T . b . rhodesiense ) . Both HAT forms culminate in parasite invasion of the central nervous system , with gradual nervous system damage if untreated . The currently used anti-HAT drugs - melarsoprol , eflornithine , pentamidine , and suramin - are highly toxic and have lost efficacy in several regions . Also , treatment is difficult to administer in resource-limiting conditions , and often unsuccessful [9] , [10] . Chagas' disease , caused by T . cruzi , affects the cardiovascular , gastrointestinal , and nervous systems of human hosts and has become , in recent decades , a worldwide public health problem due to travelers and migratory flow [2] , [11] . Chagas' disease chemotherapy is based on the use of nifurtimox and benznidazole , two very toxic nitroheterocyclic compounds with modest efficacy ( especially against late stage chronic disease ) , and ‘plagued’ by the emergence of drug resistance [12] . Given the high toxicity and limited efficacy of current treatments for leishmaniasis , Chagas' disease and HAT , the development of novel chemotherapeutics against these neglected diseases is essential . Animal venoms and poisons are natural libraries of bioactive compounds with potential to yield novel drugs or drug leads for pharmacotherapeutics [13] . In particular , snake venoms have proven to be interesting sources of potential novel agents against neglected diseases , including Chagas' disease [14]– and leishmaniasis [18]–[23] . Cysteine-rich secretory proteins ( CRISPs ) are single chain bioactive polypeptides with molecular masses of ∼20–30 kDa found in snake venom , reptilian venom ducts [24]–[26] and also in the salivary glands , pancreatic tissues , reproductive tracts [27]–[31] . In mammals , CRISPs are also expressed at low levels in non-reproductive tissues and organs , including skeletal muscle , spleen and thymus [32] . CRISPs belong to the CAP ( Crisp , antigen 5 , and pathogenesis-related ) superfamily of proteins [33] . CRISP amino acid sequences have high degree of sequence identity and similarity , and include a highly conserved pattern of 16 cysteine residues which form 8 disulfide bonds [34] . Ten of these cysteine residues form an integral part of a well-conserved cysteine-rich domain at the C-terminus , although CRISP N-terminal sequences are overall more conserved than other regions of these proteins [33]–[35] . Snake venom CRISPs belong to the CRISP-3 subfamily [36] , one of four subgroups of CRISPs , according to amino acid sequence homology . Most biological targets of snake venom CRISPs described to date are ion channels [37]–[43] , although the functions and the molecular targets of most snake venom CRISPs remain to be determined . Some snake venom CRISPs had their biological activities tested on crickets and cockroaches [35] . Snake venom CRISPs have been shown to block the activity of L-type Ca2+ and/or K+-channels and also of cyclic nucleotide-gated ( CNG ) ion channels , thereby preventing the contraction of smooth muscle cells [26] , [37] , [40]–[43] . The CRISPs catrin , piscivorin and ophanin , from the snake Crotalus atrox , caused moderate blockage of L-type calcium channels , partially inhibiting the contraction of smooth fibers from mouse caudal arteries [26] . The Philodryas patagoniensis ( green snake ) CRISP patagonin was capable of generating myotoxicity when injected into the gastrocnemius muscle , but did not induce edema formation , haemorrhage or inhibition on platelet aggregation [44] . Despite their myotoxicity , there are no reports of CRISP protein lethality to mice , in concentrations of up to 4 . 5 mg/kg [35] , [45] , and patagonin did not induce systemic alterations in mice , or histological changes in tissues from the cerebellum , brain , heart , liver and spleen [44] . In a previous publication , we showed that crude venom from the rattlesnake Crotalus viridis viridis had anti-parasitic activity against all forms of T . cruzi , and could be a valuable source of molecules for the development of new drugs against Chagas' disease [46] . In search for the molecular source of the anti-parasitic activity found in Cvv crude venom , we purified a Cvv CRISP that will be henceforth referred to as ‘crovirin’ . Here , we describe the purification , biochemical characterization and biological activity of crovirin against pathogenic trypanosomatids parasites and mammalian cells , showing that crovirin is active against infective developmental forms of trypanosomes and Leishmania , at doses that elicit no or minimal toxic effects on human cells .
Crude venom from the rattlesnake Crotalus viridis viridis ( Cvv ) and adjuvants such as parasites growth media , were purchased from Sigma–Aldrich Chemical Co ( St . Louis , MO , USA ) . Benznidazole ( Bz ) ( Laboratório Farmacêutico do Estado de Pernambuco [LAFEPE] , Brazil ) , diminazene aceturate ( Berenil [Ber] , Hoechst Veterinãr GmbH , München , Germany ) , and Amphotericin B ( Amp-B ) ( Sigma ) were used as a reference drugs for Chagas disease , sleeping sickness and leishmaniasis treatment , respectively . The material and reagents used in SDS-PAGE were from Bio-Rad Laboratories , Inc . Molecular weigh markers LMW were from Fermentas Life Sciences . Mass spectrometry grade Trypsin Gold was from Promega . All other reagents and chemicals were from Merck ( Darmstadt , Germany ) , Tedia Company and Eurofarma Laboratórios SA . Lyophilized Cvv venom ( 10 mg ) was dissolved in 1 ml of 20 mM Tris–HCl , 150 mM NaCl , pH 8 . 8 and centrifuged at 5 , 000 g for 2 min . The supernatant was applied onto a reverse phase analytical C8 column ( 5 µm , 250×4 . 6 mm ) ( Kromasil , Sweeden ) , previously equilibrated with the same buffer . Venom proteins were separated by reverse phase HPLC ( Shimadzu , Japan ) . Fractions ( 0 . 7 ml/tube ) were collected at a 1 ml/h flowrate . A linear gradient of water/acetonitrile containing 0 . 1% trifluoroacetic acid ( TFA ) was used . The elution profile was monitored by absorption at 280 nm , and the molecular homogeneity of the relevant fractions was verified by SDS-PAGE . Fractions containing protein peaks were dried in a Speed-Vac ( Savant , Thermo Scientific , USA ) and resuspended in distilled water prior to protein quantification by the Bradford method . Molecular mass determination was performed by MALDI-TOF and by electrospray ionization ( ESI ) mass spectrometry using a Voyager-DE Pro and a QTrap 2000 ( both from Applied Biosystems ) , respectively . Protein bands were excised from Coomassie Brilliant Blue-stained SDS-PAGE gels and cut into smaller pieces , which were destained with 25 mM NH4HCO3 in 50% acetonitrile for 12 h . The pieces obtained from the non-reducing gels were reduced in a solution of 10 mM dithiothreitol and 25 mM NH4HCO3 for 1 h at 56°C , and then alkylated in a solution of 55 mM iodoacetamide and 25 mM NH4HCO3 , for 45 min in the dark . The solution was removed , the gel pieces were washed with 25 mM NH4HCO3 in 50% acetonitrile , and then dehydrated in 100% acetonitrile . Finally , all pieces from reducing and non-reducing gels were air-dried , rehydrated in a solution of 25 mM NH4HCO3 containing 100 ng of trypsin , and digested overnight at 37°C . Tryptic peptides were then recovered in 10 µl of 0 . 1% TFA in 50% acetonitrile . The peptides extracted from gel pieces were loaded into a Waters Nano Acquity system ( Waters , MA , USA ) and desalted on-line using a Waters Symmetry C18 180 µm×20 mm , 5 µm trap column . The typical sample injection volume was 7 . 5 µl , and liquid chromatography ( LC ) was performed by using a BEH 130 C18 100 µm×100 mm , 1 . 7 µm column ( Waters , MA , USA ) and eluting ( 0 . 5 µl/min ) with a linear gradient of 10–40% acetonitrile , containing 0 . 1% formic acid . Electrospray tandem mass spectra were performed in a Q-Tof quadrupole/orthogonal acceleration time-of-flight spectrometer ( Waters , Milford , MA ) linked to a nano ACQUITY system ( Waters ) capillary chromatograph . The ESI voltage was set at 3300 V , the source temperature was 80°C and the cone voltage was 30 V . The instrument control and data acquisition were conducted by a MassLynx data system ( Version 4 . 1 , Waters ) , and experiments were performed by scanning from a mass-to-charge ratio ( m/z ) of 400–2000 using a scan time of 1 s , applied during the whole chromatographic process . The mass spectra corresponding to each signal from the total ion current ( TIC ) chromatogram were averaged , allowing for accurate molecular mass measurements . The exact mass was determined automatically using Q-Tof's LockSpray ( Waters , MA , USA ) . Data-dependent MS/MS acquisitions were performed on precursors with charge states of 2 , 3 or 4 over a range of 50–2000 m/z , and under a 2 m/z window . A maximum of three ions were selected for MS/MS from a single MS survey . Collision-induced dissociation ( CID ) MS/MS spectra were obtained using argon as the collision gas at a pressure of 40 psi , and the collision voltage varied between 18 and 90 V , depending on the mass and charge of the precursor . The scan rate was 1 scan/s . All data were processed using the ProteinLynx Global server ( version 2 . 5 , Waters ) . The processing automatically lock mass calibrated the m/z scale of both the MS and the MS/MS data utilizing a lock spray reference ion . The MS/MS data were also charge-state deconvoluted and deisotoped with the maximum entropy algorithm MaxEnt 3 ( Waters , MA , USA ) . Proteins corresponding to the tryptic peptides from peak 3 were identified by correlation of tandem mass spectra and the NCBInr database of proteins ( Version 050623 ) , using the MASCOT software ( Matrix Science , version 2 . 1 ) . Settings allowed for one missed cleavage per peptide , and an initial mass tolerance of 0 . 2 Da was used in all searches . Cysteines were assumed to be carbamidomethylated , and a variable modification of methionine ( oxidation ) was allowed . Identification was considered positive when at least two peptides matched the protein sequence with a mass accuracy of less than 0 . 2 Da . T . cruzi tissue culture trypomastigotes ( CL-Brener clone ) were obtained from the supernatants of 5 to 6-day-old infected LLC-MK2 cells maintained in RPMI-1640 medium ( Sigma ) supplemented with 2% FCS for 5–6 days at 37°C in a humidified 5% CO2 . Theses trypomastigotes were also used to obtain intracellular amastigotes in macrophage cultures . The MHOM/BR/75/Josefa strain of L . amazonensis , isolated from a patient with DCL by C . A . Cuba-Cuba ( Universidade de Brasilia , Brazil ) , was used in the present study . Amastigote forms were maintained by hamster footpad inoculation , while promastigotes were cultured axenically in Warren's medium [47] supplemented with 10% fetal bovine serum ( FBS ) at 25°C . Infective promastigotes were used to obtain intracellular amastigotes in macrophage cultures , as described previously [48] . Bloodstream form ( BSF ) T . brucei rhodesiense ( strain IL1852 ) were cultivated in HMI-9 medium ( Invitrogen ) supplemented with 10% inactivated FBS ( Biosera-South America ) and 10% of serum plus supplement ( SAFC Bioscience , USA ) , at 37°C in a humidified 5% CO2 incubator [49] . Procyclic-form ( PCF ) T . brucei rhodesiense ( strain 457 ) were grown in SDM-79 medium ( LGC Biotecnologia ) supplemented with 10% heat-inactivated FBS , at 28°C [50] . In this study , we used 5-week-old female CF1 mice as sources of peritoneal macrophages and of muscle sample for ex vivo assays ( described below ) . All animal experimentation protocols received the approval by the Commission to Evaluate the Use of Research Animals ( CAUAP , from the Carlos Chagas Filho Biophysics Institute - IBCCF ) , and by the Ethics Committee for Animal Experimentation ( Health Sciences Center , Federal University of Rio de Janeiro – UFRJ ) ( Protocol no . IBCCF 096/097/106 ) , in agreement with Brazilian federal law ( 11 . 794/2008 , Decreto n° 6 . 899/2009 ) . We followed institutional guidelines on animal manipulation , adhering to the “Principles of Laboratory Animal Care” ( National Society for Medical Research , USA ) and the “Guide for the Care and Use of Laboratory Animals” ( National Academy of Sciences , USA ) . Crovirin was purified as described above and stored at −20°C , in 3 . 6 mg/ml stock solutions prepared in PBS ( pH 7 . 2 ) . All experiments were carried out in triplicates . Stock solutions of Bz ( 14 mg/ml ) and Amp-B ( 10 mg/ml ) were prepared in dimethyl sulfoxide ( DMSO ) , and the final concentration of the solvent never exceeded 0 . 5% , which is not toxic for parasites and mammalian cells . Ber stock solution ( 0 . 188 mg/ml ) was prepared in pyrogen-free water . Axenically grown parasite forms were treated with crovirin for up to 72 h in the same culture conditions used for growth ( described above ) . The following crovirin concentrations were used to treat axenic forms: 1 . 2–4 . 8 µg/ml ( L . amazonensis promastigotes ) and 0 . 6–4 . 8 µg/ml crovirin ( T . brucei rhodesiense BSF and PCF ) . IC50 values were calculated based on daily counting of formalin-fixed parasites using a hemocytometer . Positive controls were run in parallel with 4 . 7 µg/ml Amp-B [51] and 39 . 8 ng/ml Ber [52] , respectively . T . cruzi tissue culture trypomastigotes were treated with crovirin ( 0 . 45–4 . 8 µg/ml ) at a density of 1×106 cells/ml , for 24 h at 37°C ( in RPMI media containing 10% FCS ) . LD50 ( 50% trypomastigote lysis ) values were determined based on direct counting of formalin-fixed parasites using a hemocytometer . Bz was used as reference drug , in a 3 . 39 µg/ml concentration [53] . To evaluate the effects of crovirin on T . cruzi and L . amazonensis intracellular amastigotes , peritoneal macrophages from CF1 mice were harvested by washing with RPMI medium ( Sigma ) , and plated in 24-well tissue culture chamber slides , allowing them to adhere to the slides for 24 h at 37°C in 5% CO2 . Adherent macrophages were infected with tissue culture T . cruzi trypomastigotes ( at 37°C ) or L . amazonensis metacyclic promastigotes ( at 35°C ) at a macrophage-to-parasite ratio of 1∶10 , for 2 h . After this period , non-internalized parasites were removed by washing , cultures were incubated for 24 h in RPMI with 10% FCS , and fresh medium with crovirin ( 0 . 45–3 . 6 µg/ml for T . cruzi , and 0 . 6–9 . 6 µg/ml for L . amazonensis ) was added daily for 72 h . At different time-points ( 24 , 48 and 72 h ) cultures were fixed with 4% paraformaldehyde in PBS ( pH 7 . 2 ) and stained with Giemsa for 15 min . The percentage of infected cells and the number of parasites per 100 cells were determined by light microscopy examination . Positive controls of T . cruzi and L . amazonensis amastigotes infected cells were run in parallel with cultures treated with 0 . 73 µg/ml Bz [53] and 0 . 07 µg/ml Amp-B [54] , respectively . LLC-MK2 cells were maintained in RPMI medium supplemented with 10% FCS . Prior to treatment with crovirin , cells were seeded in 24-well plates containing glass coverslips and incubated in RPMI medium supplemented with 10% FCS for 24 h at 37°C . Cells were then treated with 4 . 8 , 10 and 20 µg/ml crovirin at 37°C for 72 h . LC50 values ( concentrations that reduces by 50% the cellular viability ) for crovirin were calculated from daily counts of the number of viable cells , using trypan blue as an exclusion dye . At least 500 cells were examined per well , on a Zeiss Axiovert light microscope ( Oberkochen , Germany ) . In addition , mouse peritoneal macrophages were seeded on 96-well plates , incubated in RPMI medium with 10% FCS for 24 h at 37°C and treated with 4 . 8 , 10 and 20 µg/ml crovirin at 37°C , for 72 h . After this period , cells were washed with PBS ( pH 7 . 2 ) , and the wells were filled with RPMI medium without phenol red containing 10 mM glucose and 20 µl of a solution of 2 mg/ml MTS ( 3- ( 4 , 5-dimethylthiazol-2-yl ) -5- ( 3-carboxymethoxyphenyl ) -2- ( 4-sulfophenyl ) -2H-tetrazolium salt ) and 0 . 92 mg/ml PMS ( phenazine methosulfate ) , prepared according to the manufacturer's instructions ( Promega , Madison , WI , USA ) . Following 3 h of incubation at 37°C , formation of a soluble formazan product by viable cells was measured using a plate reader , by absorbance at 490 nm . All cytotoxicity experiments were carried out in triplicates . The myotoxicity of crovirin was studied ex vivo using a muscle creatin kinase ( CK ) activity assay [55] . The analysis consisted of monitoring the rate of CK release from isolated mouse extensor digitorum longus ( EDL ) muscle bathed in a solution containing crovirin ( 10 µg/ml ) . Adult male and female Swiss mice ( 25 . 0±5 . 0 g ) were anesthetized with ethyl ether and killed by cervical dislocation . EDL muscles were collected , freed from fat and tendons , dried and weighed . Muscle samples were then homogenized in 2 ml saline/0 . 1% albumin and their CK content was determined using a commercial diagnostic kit ( Bioclin , Brazil ) . Four EDL muscles were mounted vertically on a cylindrical chamber and superfused continuously with Ringer's solution equilibrated with 95% O2/5% CO2 . At 30 to 90-min intervals , the perfusing solution was collected and replaced with fresh solution . The collected EDL samples were used for the measurement of CK activity as described above . Muscles were weighed at the end of the experiment ( 2 h later ) . Enzyme activity is reported as international units corrected for muscle mass . Mean value comparisons between control and treated groups were performed using the Kruskal-Wallis test in the BioEstat 2 . 0 program for Windows . Differences with p≤0 . 05 were considered statistically significant .
In a previous study , we showed that the Cvv venom had anti-parasitic activity against T . cruzi [46] . Preliminary analysis of Cvv venom fractions by reverse-phase chromatography ( not shown ) indicated that the activity eluted with fractions containing peak 3 of the chromatographic profile ( Fig . 1A ) . Thus , we analyzed the main chromatographic fraction corresponding to peak 3 by SDS-PAGE and MALDI-TOF mass spectrometry ( Fig . 1B–C ) . SDS-PAGE analysis of peak 3 showed a single polypeptide , with a relative molecular mass of 24 kDa ( Fig . 1B ) and 28 kDa ( data not shown ) , under reducing and non-reducing conditions , respectively . We will refer to this protein henceforth as crovirin . MALDI-TOF analysis of the intact protein showed a molecular mass of 24 , 893 . 64 Da ( Fig . 1C ) . The peaks of 12 , 424 . 36 and 12 , 477 . 62 Da in the MS profile correspond to doubly-charged ( z = 2 ) cationic forms of the protein . The amino acid sequence of tryptic crovirin peptides ( produced by Nano LC-MS/MS mass spectrometry analysis ) is nearly identical to a partial sequence of a Cvv CRISP ( GenBank gi:190195319 ) ( Fig . 2 ) . The MS/MS-derived sequences are also nearly identical to those of a CRISP protein from Calloselasma rhodostoma ( GenBank gi:190195317 ) and have high degree of sequence similarity to several other snake venom CRISPs , including ablomin ( Fig . 2 ) . The MS/MS spectrum of the fragmented peptide ions was matched by MASCOT displayed a coverage of 48% of identical peptides , with a p≥355 indicating extensive homology to the CRISP from C . rhodostoma . The MS results strongly suggested that a CRISP from Cvv snake venom had been purified , and corresponded to crovirin . First of all , we investigated crovirin citotoxicity over mammalian host cells before proceeding with our analysis of the anti-parasitic activity of this venom protein . LLC-MK2 cells were treated with crovirin for 72 h and examined for viability using a trypan blue exclusion assay ( Fig . 3A ) . None of the tested crovirin concentrations ( 4 . 8 , 10 or 20 µg/ml ) were capable of inducing significant loss of cell viability , even after 72 h of treatment . In addition , we tested the activity of crovirin against murine peritoneal macrophages to investigate its cytotoxicity towards primary host cells . Treated cells were examined using an MTS assay , and no significant toxicity ( p≤0 . 05 ) was observed in any treatment conditions ( Fig . 3B ) . Creatine kinase ( CK ) activity was measured before and two hours after extensor digitorum longus ( EDL ) muscle exposure to 10 µg/ml crovirin . We did not observed significant CK release from treated EDL muscles compared to control ( saline ) after 2 hours of incubation with crovirin , indicating that this protein did not generated appreciable myotoxicity at the concentration tested . After establishing that crovirin had only minimal cytotoxic effects towards mammalian cells at concentrations of up to 20 µg/ml , we tested the anti-parasitic activity of purified crovirin against relevant developmental forms of three different species of pathogenic trypanosomatid parasites , namely L . amazonensis , T . cruzi and T . brucei rhodesiense . We tested crovirin activity against the two infective T . cruzi forms , trypomastigotes and amastigotes . Trypomastigote forms do not multiply and do not remain viable after several days in culture media at 37°C . Therefore , the effect of crovirin towards T . cruzi trypomastigotes was evaluated as the ability of the protein to lyse cells after 24 h of treatment ( Fig . 4A ) . The calculated LD50 of crovirin for trypomastigotes was 1 . 10±0 . 13 µg/ml ( Table 1 ) . This concentration displayed the second higher selectivity index ( 18 . 2 ) ( Table 1 ) among all crovirin treatments . The treatment with 3 . 39 µg/ml Bz exhibited a 65 . 8% of parasites lysis at same conditions . T . cruzi amastigotes multiply in the intracellular environment . Crovirin inhibited the growth of amastigotes inside peritoneal macrophages in a dose-dependent manner ( Fig . 4B ) , with an IC50 of 1 . 84±0 . 53 µg/ml when cells were treated with crovirin for 72 h ( Table 1 ) . Crovirin presented a discret superior trypanocidal activity against the intracellular forms as compared with Bz ( Fig . 4B ) . Crovirin activity was also tested against infective promastigote and amastigote forms of L . amazonensis , one of the species responsible for CL . None of the crovirin concentrations tested inhibited significantly the proliferation of L . amazonensis promastigotes in axenic media , unlike Amp-B treatment , which resulted in a reduction of a little over 80% in the number of parasites after 72 h of treatment . In contrast , crovirin inhibited the proliferation of intracellular amastigotes of L . amazonensis in a concentration-dependent manner ( Fig . 4C–D ) . The effect of crovirin on amastigote proliferation was evident as early as 24 h after the start of treatment , and the IC50 for crovirin after 72 h of treatment was 1 . 21±0 . 89 µg/ml ( Table 1 ) . After 48 h incubation , the IC50 of 1 . 05 µg/ml also resulted in the highest selectivity index ( 19 . 1 ) , being less toxic treatment to mammalian host cells . However , no tested concentration of crovirin had superior leishmanicidal activity against amastigotes forms as compared with Amp-B ( Fig . 4D ) . Both developmental forms of T . brucei rhodesiense tested here ( PCF and BSF ) were sensitive to crovirin treatment . A different profile of growth inhibition in the presence of crovirin was observed for PCF ( Fig . 4E ) and BSF ( Fig . 4F ) parasites , with IC50 values of 1 . 13±0 . 31 and 2 . 06±0 . 12 µg/ml , respectively , after 72 h of treatment . The 39 . 8 ng/ml Ber treatment resulted in a remarkable growth inhibition of both BCF and PCF than crovirin treatment ( Fig . 4E–F ) .
There is an urgent need for the development of novel compounds for the treatment of trypanosomatid-borne diseases , currently treated with ‘dated’ chemotherapeutic agents with high toxicity and limited efficacy , partly due to the emergence of drug resistance . Animal venoms and toxins , including snake venoms , can provide compounds directly useful as drugs , or with potential as drug leads for the synthesis of novel therapeutic agents [22] . Previously , our group showed that Cvv crude venom displayed anti-parasitic activity against different T . cruzi developmental forms [46] . We have now extended this research with the purification of crovirin , a CRISP from Cvv venom with promising activity against key infective stages of the life cycle of T . cruzi , T . brucei rhodesiense and L . amazonensis . Furthermore , we show that crovirin has low toxicity towards host cells and mouse muscle , in agreement with the low or absent toxicity reported for most CRISPs proteins [35] , [44]–[45] . CRISPs proteins are often given names that refer to the organism from which they were isolated . The first CRISP described in reptiles was isolated from the skin secretion of the lizard Heloderma horridum , and was named helodermin [56] . Examples of proteins isolated from snake venoms are patagonin , isolated from Philodryas patagonensis [44] , latisemin , isolated from sea snake Laticauda semifasciata , tigrin isolated from Rhabdophis tigrinus tigrinus [41] , and ablomin , isolated from Gloydius blomhoffi [41] . CRISPs sequences have also been identified in transcriptome analysis of venom glands [57]–[58] or are deposited at databanks but were not purified or studied . A partial CRISP sequence from Crotalus viridis viridis ( GenBank gi:190195319 ) likely corresponding to central and C-terminal regions of crovirin was identified by transcriptome analysis of venom gland tissue . However , this is the first report on the purification and study of crovirin . One of the most important findings of the present study was the activity of crovirin against the intracellular proliferation of trypanosomatids . Amastigotes are key developmental forms during the development and maintenance of infections by Leishmania and T . cruzi , representing the replicative intracellular stages of these protozoan parasites . Substantial inhibition of both T . cruzi and L . amazonensis intracellular amastigote proliferation was observed at crovirin concentrations significantly lower than those required to cause damage to host cells , including mouse EDL muscles . These results are particularly important because the currently available drugs to treat leishmaniasis and Chagas' disease are known to have lower anti-amastigote activity [1] , [6] . The effects of crovirin over both the procyclic and the bloodstream form of T . brucei rhodesiense are also encouraging , suggesting that crovirin might be useful in the development of new anti-HAT chemotherapeutics . In conclusion , our results demonstrate that crovirin has promising trypanocidal and leishmanicidal effects , and represents a potential avenue for drug development against leishmaniasis , Chagas' disease and HAT , since its anti-parasitic effects are matched by low toxicity to host cells and muscles . Further studies are now required to extend our knowledge on the potential use of crovirin as an alternative compound to improve the effectiveness of treatment of trypanosomatid-borne neglected diseases .
|
The pathogenic trypanosomatid parasites of the genera Leishmania and Trypanosoma infect over 20 million people worldwide , with an annual incidence of ∼3 million new infections . An additional 400 million people are at risk of infection by exposure to parasite-infected insects which act as disease vectors . Trypanosomatid-borne diseases predominant in poorer nation and are considered neglected , having failed to attract the attention of the pharmaceutical industry . However , novel therapy is sorely needed for Trypanosoma and Leishmania infections , currently treated with ‘dated’ drugs that are often difficult to administer in resource-limiting conditions , have high toxicity and are by no means always successful , partly due to the emergence of drug resistance . The last few decades have witnessed a growing interest in examining the potential of bioactive toxins and poisons as drugs or drug leads , as well as for diagnostic applications . In this context , we isolated and purified crovirin , a protein from the Crotalus viridis viridis ( Cvv ) snake venom capable to inhibiting and/or lysing infective forms of trypanosomatid parasites , at concentrations that are not toxic to host cells . This feature makes crovirin a promising candidate protein for the development of novel therapy against neglected diseases caused by trypanosomatid pathogens .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"biochemistry",
"medicine",
"and",
"health",
"sciences",
"peptides",
"proteins",
"biology",
"and",
"life",
"sciences",
"pharmacology",
"proteomics",
"tropical",
"diseases",
"parasitic",
"diseases"
] |
2014
|
Crovirin, a Snake Venom Cysteine-Rich Secretory Protein (CRISP) with Promising Activity against Trypanosomes and Leishmania
|
An effective surveillance system is critical for the elimination of canine rabies in Latin America . Brazil has made substantial progress towards canine rabies elimination , but outbreaks still occurred in the last decade in two states . Brazil uses a health information system ( SINAN ) to record patients seeking post-exposure prophylaxis ( PEP ) following contact with an animal suspected of having rabies . This study evaluated: ( i ) whether SINAN can be reliably used for rabies surveillance; ( ii ) if patients in Brazil are receiving appropriate PEP and ( iii ) the benefits of implementing the latest World Health Organization ( WHO ) recommendations on PEP . Analysing SINAN records from 2008 to 2017 reveals an average of 506 , 148 bite-injury patients/year [range: 437k-545k] in the country , equivalent to an incidence of 255 bite-injuries/100 , 000 people/year [range: 231–280] . The number of reports of bites from suspect rabid dogs generally increased over time . In most states , records from SINAN indicating a suspect rabid dog do not correlate with confirmed dog rabies cases reported to the Regional Information System for Epidemiological Surveillance of Rabies ( SIRVERA ) maintained by the Pan American Health Organization ( PAHO ) . Analyses showed that in 2017 , only 45% of patients received appropriate PEP as indicated by the Brazilian Ministry of Health guidance . Implementation of the latest WHO guidance using an abridged intradermal post-exposure vaccination regimen including one precautionary dose for dog bites prior to observation would reduce the volume of vaccine required by up to 64% , with potential for annual savings of over USD 6 million from reduced vaccine use . Our results highlight the need to improve the implementation of SINAN , including training of health workers responsible for delivering PEP using an Integrated Bite Case Management approach so that SINAN can serve as a reliable surveillance tool for canine rabies elimination .
Dog-mediated rabies in humans has been reduced by over 99% in Latin America , but 106 cases in dogs were still reported in 2014 [1] . Dog vaccination and prompt administration of post-exposure prophylaxis ( PEP ) for bite victims are essential to the successful reduction of human cases of rabies across Latin America [2 , 3] . However , detecting the final cases of rabies and interrupting transmission is difficult and therefore continued precautionary use of PEP is required for persons bitten by dogs that show signs suspicious of rabies [4] . In Latin America , PEP is broadly available and 0 . 2% of people are referred to health services annually after potential exposure to the risk of rabies ( mainly because of bites by dogs ) , with around 25% of patients receiving free PEP [3] . Control of rabies in dogs is expected to reduce bites by rabid dogs but not necessarily bites by healthy dogs for which patients do not require PEP . Unnecessary and indiscriminate use of PEP strains local and national healthcare budgets , and is a missed opportunity for improving the detection of rabies cases [5 , 6] . Eliminating dog-mediated rabies requires an efficient and sensitive surveillance system that can detect new , but rare , outbreaks of dog rabies . However , outbreaks of dog rabies are difficult to detect without using targeted methods and community surveillance because of the millions of dogs in Latin America and because rabies incidence is low even when endemic ( less than 1% of dogs become infected per annum ) [7] . Passive surveillance of dog bite victims seeking help within the health care system can support active surveillance in detecting animal rabies cases [8–10] . Dog rabies can be detected by focusing on bite patients and evaluating the dogs responsible for bites , because dog bites are frequent and unprovoked aggression is one the main clinical signs of this disease . Surveillance of bites can also reveal secondary cases caused by a rabid dog and has potential to reduce unnecessary PEP use [5 , 8 , 10 , 11] . For example , surveillance of dog bites in Haiti identified areas of rabies risk and led to considerable reductions in PEP use [8 , 11] . Improved surveillance of dog-bite patients can also help to detect the rare but increasing cases in dogs caused by bat rabies in countries such as Brazil [12] . Brazil has one of the largest estimated populations of dogs worldwide and the second largest of the continent after the United States , with around 30–50 million dogs [13 , 14] . The country has made substantial progress towards the elimination of dog-mediated rabies over the last decades , but cases persist in the state of Maranhão , and a large outbreak recently occurred in Mato Grosso do Sul as a result of incursions from bordering Bolivia [15 , 16] . Thus , Brazil requires an effective surveillance system to support current dog vaccination campaigns to eliminate rabies from remaining endemic foci . Brazil’s National Health System ( SUS ) provides universal health coverage to most Brazilians using a decentralized network of public hospitals and health facilities . Since 1998 , Brazil has used a national health ‘Information System on Diseases of Compulsory Declaration’ ( SINAN ) to record all patients seeking medical care in public health facilities following contact with an animal that can transmit rabies . SINAN is a relatively inexpensive tool that records information on patients , the animal contacted and PEP use . The SINAN database records several variables that can be used for rabies surveillance including the severity of the bite , whether the biting animal is suspected for rabies , whether the animal is alive and can be observed for a 10-day period , and the PEP regimen administered to the patient [17 , 18] . Previous studies have performed descriptive analysis to summarize the SINAN database at national and state level [18–21] , but the accuracy of the SINAN database as a tool for the surveillance of rabies and for the appropriate use of PEP has not been evaluated . In Brazil , dog bite incidence was high between 2002 and 2009 with 425 000 bites estimated per year [19 , 22] . The reduction of canine rabies cases over the last decade in Brazil should cause a concommittant decline in bites from rabid dogs and potential savings in PEP administration , but this has not been realized . Here we examine ( i ) whether SINAN can be reliably used for rabies surveillance; ( ii ) if patients in Brazil are receiving appropriate PEP according to Brazilian Ministry of Health ( MoH ) guidelines and ( iii ) the benefits of implementing the latest World Health Organization ( WHO ) recommendations on PEP .
The incidence of patients seeking heath care following a bite ( noted here as ‘bite incidence’ ) in Brazil remained relatively stable from 2008 to 2016 , with an average of 257 patients/100 000 population ( range: 231–280 ) . Within that range , an increase was observed from 2008 ( 230 patients/100 000 ) to 2012 ( 279 patients/100 000 ) , followed by a decline until 2016 ( 252 patients/100 000 ) ( Fig 2 ) . Annual bite incidence is , however , very variable per state with a maximum of 544 bites/100 000 population in the state of Roraima and a minimum of 97 bites/100 000 population in Sergipe in 2016 ( Fig 2 ) . There was no correlation between bite incidence in a state and either the state’s average income ( Spearman’s correlation test , Rho = 0 . 20 , p-value = 0 . 31 ) in 2016 , the state’s number of houses with at least one dog ( Spearman’s correlation test , Rho = -0 . 18 , p-value = 0 . 37 ) in 2013 , nor the state’s percentage of houses with at least one ( Spearman’s correlation test , Rho = 0 . 19 , p-value = 0 . 33 ) . The percentage of patients bitten by dogs classified as suspect by health workers ( including categories ‘rabies suspicious’ , ‘rabid’ or ‘dead/disappeared’ in the SINAN form ) , which are expected to result in the use of PEP , increased from 17% in 2008 to 25% in 2017 ( Fig 3A ) . This increase was mainly due to an increase in the number of dogs that were categorized as ‘non-observable’ , which more than doubled ( 3% to 7% ) during this period . Among dog bites , the absolute number and the relative proportion due to suspect rabid dogs varied between states ( Fig 3B ) and over time from 2008 until 2017 ( Fig 3C ) . For example , the maximum proportion of bites due to suspect dogs was observed in Roraima state ( 35% ) and the minimum proportion in Alagoas ( 15% ) . Likewise , the difference in the percentage of bites from suspect dogs between 2008 and 2017 was the highest in Roraima , which increased by 24% , whereas the only reductions were observed in the states of Santa Catarina ( -6% ) and Mato Grosso do Sul ( -0 . 4% ) . The total number of dogs reported as ‘laboratory confirmed rabies’ in SINAN was much higher ( 2482 dogs ) than in the SIRVERA database ( 269 dogs ) for the period of 2008–2017 . More than 2000 reports in the SINAN data could therefore be considered as ‘false positives’ , assuming these were not verified for inclusion in SIRVERA . Out of the 27 states in Brazil , 26 reported at least one lab-positive rabid dog in SINAN whereas over the same period only 14 states reported confirmed rabid dogs in SIRVERA ( Fig 4A and 4B ) . States reporting the most positive rabid dogs also differed between SINAN and SIRVERA . In SINAN , larger more populous states such as Sao Paulo and Minas Gerais reported the most positive dogs with 481 and 296 respectively ( Fig 4A ) , while from SIRVERA the states of Maranhão and Mato Grosso do Sul reported the most positive dogs with 116 and 83 respectively ( Fig 4B ) . On a monthly basis , trends in rabies-positive dogs were similar in SINAN and SIRVERA for states with known foci of rabies transmission , such as Maranhão ( 2011–2014 ) and Mato Grosso do Sul ( 2015 ) ( Fig 4C ) . However , in states with very few or no rabies cases identified in SIRVERA , such as Bahia , Minas Gerais , Sao Paulo and Rio Grande do Sul , monthly data on positive dogs from SINAN showed very different temporal trends , with peaks in cases that were not evident in SIRVERA ( Fig 4C ) . Based on SINAN , there was an increase in the percentage of patients requiring PEP ( either vaccines or vaccine and serum ) from 2008 ( 67% ) to 2015 ( 77% ) . In the last two years , however , there has been a reduction with 59% of patients receiving PEP in 2017 , which coincides with a vaccination shortage experienced by Brazil in 2015 ( Fig 5A ) . This reduction was observed in 19 out of 27 states . Based on the PEP guidelines from the Ministry of Health used from 2008–2017 , we estimated the percentage of patients that received appropriate PEP . The percentage of patients receiving appropriate PEP remained relatively stable at approximately 56% from 2008 ( 56% ) to 2015 ( 54% ) . However , appropriate PEP administration declined in the last two years to 45% in 2017 ( Fig 5A ) . Most of the incorrect PEP use in 2017 was due to either an under-use of vaccines or serum in high-risk bites that required PEP ( 37% of total PEP ) or an unnecessary use of vaccines and serum in low-risk bites that did not require PEP ( 8% of total PEP ) . Under-use following a national vaccine shortage in 2015 increased from 25% to 37% ( 2017 ) of total PEP . We used patient presentations in 2017 as a baseline from which to estimate vaccine requirements under different hypothetical PEP regimens . We estimated that 1 , 152 , 057 doses of vaccines were needed for appropriate PEP administration using the 5 dose IM Essen regimen that was recommended up to 2017 across Brazil . In comparison , we estimated that 1 , 040 , 444 doses would be needed under the 4 dose IM updated Essen regimen recommended in the updated MoH guidelines in 2017 and the WHO ( Fig 5C ) . If Brazil implemented the latest WHO recommendations ( e . g . 3 ID doses for complete prophylaxis ) , the number of required doses without vial sharing would be reduced to 631 , 835 ( 39% reduction from the 4 dose IM regime ) , and with vial sharing to 379 , 101 ( 64% reduction ) . Considering a price of 10 USD for each 0 . 5 mL vaccine vial , the WHO recommendations without vial sharing ( 6 . 31M USD ) would save 4 . 08 M USD in vaccine costs compared to the current 4 IM guidelines ( 10 . 40M USD ) . Under the hypothetical scenario with complete vial sharing ( 3 . 79M USD ) , savings of up to 6 . 61M USD in vaccine costs could be made . In practice considerable vial sharing would be expected , especially in urban clinics with high throughput of patients , and so savings of between 4 . 08 M USD and 6 . 61M USD would be expected .
An effective national surveillance system is essential to detect remaining cases of dog-mediated rabies and to ensure that patients bitten by dogs that pose a risk of rabies transmission are given appropriate PEP . Control programmes in Brazil have substantially reduced dog-mediated rabies transmission and human cases of dog-mediated rabies in the last decade . Despite this reduction in dog rabies , our results show that the number of patients presenting to health facilities due to bites by dogs have remained relatively constant in the country over the last decade . However , the number of bites by high-risk ‘suspect’ dogs indicating the need PEP have increased . This mismatch between rabies cases and suspicion of a rabid dog suggests a ‘fear’ of rabies by health workers that has contributed to a risk averse increase in PEP use over the last decade , with only half of patients being administered PEP appropriately . Following the human rabies vaccine shortage experienced in 2015 , the percentage of patients receiving appropriate PEP declined . We do not know of any human rabies deaths from dog-mediated rabies as a result of this apparent misuse of PEP and suggest that more judicious use of PEP could be safely undertaken if health workers are appropriately trained in risk assessments of bitten patients and use available epidemiological data to evaluate the risk of dog rabies in their area . Our analysis also showed that switching to ID administration of rabies vaccines according to the latest WHO position could save millions of dollars to the Brazilian health system related to vaccine volume , although the implementation of such a switch requires further investigation . The overall incidence of patients in Brazil seeking care due to dog bites ( 255 bites/ 100 , 000 people ) was similar to estimates from Kenya ( 289 bites ) [32] but higher than incidence in patients reported in Tanzania ( 12–120 ) [33] , Uganda ( 40 ) [34 , 35] and Switzerland ( 180 ) [36] . Community-based surveys focusing on dog bites , not on patients seeking care , show a much larger bite incidence in India ( 1700–2500 ) [37] and the USA ( 300–1000 ) [38] . Although bite incidence has remained relatively stable over the last decade in Brazil , bite incidence varies more than 6-fold between states , resulting in an uneven burden to state health centers that assess and treat patients . There was also a small increase up to 2012 followed by a decrease . Although the causes of this temporal variation remain unknown , changing dog populations , dog ownership patterns and access to health care could all have contributed to the observed decline after 2012 . For example , large densely populated states such as São Paulo have a relatively low bite incidence but a high absolute number of patient presentations compared to less populous states such as Roraima that have a much higher incidence of bite patients . Because most of these bites come from healthy dogs , investigation of the factors underpinning these differences is warranted . In our study , we did not find any correlation between bite incidence and each state’s average income nor the number ( or percentage of houses ) with at least one dog , suggesting that bite incidence was not necessarily explained by socio-economic status nor the overall dog population . A more detailed analysis at the municipality level could reveal local drivers associated with bite incidence such as the dog population ( and human:dog ratio ) , the percentage of free-roaming dogs or people’s attitudes to dogs and their care . For example , some states might have a disproportionately higher population of free-roaming dogs or dog breeds that are more prone to bite as well as a higher proportion of at-risk populations ( e . g . young males under 20 years old [26 , 39] . A particularly high bite incidence was observed in the state of Roraima , which neighbours with Venezuela . Attention should be given to improve measures in this state , since bites pose a considerable public health burden even in the absence of rabies and information on rabies in Venezuela is currently limited . In Roraima , no rabies cases have been reported to SIRVERA in the last decade but there is little active surveillance , which could contribute to a slow public health response if rabies were to emerge . Despite a substantial reduction in dog-positive cases in the last decade , our analysis shows that the percentage of bites from dogs that are considered ‘suspect’ for rabies ( i . e . showing symptoms , death or disappear ) has increased . Most of the increase is due to dogs that were reported to not be observable . This increase is also very variable across states , with some states such as Roraima showing an increase of over 30% . An increase in the percentage of non-observable dogs could arise from an increase in bites from free-roaming dogs or a decreased in bites from owned ( observable ) dogs due to more responsible pet ownership . Alternatively , this could be due to increased unwillingness of health workers to perform dog follow up and the recommended 10-day observation and to directly prescribe PEP to avoid risks related to PEP underuse . Current data cannot resolve either of these hypotheses which requires further research . Similarly , about a third of reports from suspect dogs do not have data entered on whether the dog is observable or not , which highlights the need for better training of health workers in this crucial step that should determine PEP administration . If dogs are not properly observed and assessed following a bite , this will increase the chance that genuinely rabid dogs will not be identified , which could result in further dog-to-dog transmission , or even failure to detect bat-to-dog transmission without any emergency response or prevention measures . In states with endemic transmission of rabies in the last five years , such as Maranhão and Mato Grosso do Sul , data from SINAN matches reasonably well with the positive cases reported to SIRVERA , suggesting that SINAN could be used effectively to enhance rabies surveillance . However , hundreds of ‘false positives’ were observed in other states such as São Paulo , Bahia or Minas Gerais . ‘False positives’ could arise if health workers accidentally mis-record data on SINAN forms and this data is not further checked . Alternatively , mistakes could be made to justify the use of PEP if health workers are not confident in recommending appropriate PEP . Discriminating between these hypotheses will require further work on the factors contributing to the recording of the SINAN data by health workers . We hypothesize that in states with endemic rabies healthcare workers may be more aware of risks because they have received specific and recent training on rabies by the Ministry of Health and PANAFTOSA following outbreaks and thus take more care to record data . In contrast , in states where rabies has been absent for many years health care workers may become lax in recording because the repercussions of over- or under-prescription may not have immediately apparent ramifications . States such as São Paulo or Minas Gerais have also the largest human populations of Brazil and thus , even if the percentage of mistakes on recording SINAN data is the same as in other smaller states , it results in a larger number of ‘false positives’ . However , maintaining vigilance in rabies-free states is important when the risk of importation from endemic areas remains . Overall , these results highlight the need to both improve and update health care worker’s knowledge on the signs of rabies in dogs by strengthening a ‘one health’ link with veterinarians . For example , a user-friendly mobile app could i ) put in contact state’s health workers with veterinarians to improve dog assessments , ii ) improve dog follow ups , and ( iii ) rapidly detect a ‘false positive’ or priority areas for educational campaigns on dog rabies and vaccination . ‘One health’ workshops performed annually in each state could also help in updating knowledge of health and veterinarian professionals on the current rabies situation , dog assessment , and best surveillance practices . Inadequate training of health workers may also limit the appropriate administration of PEP according to the guidelines of the Brazilian MoH . We estimated that PEP use increased over the last decade to almost 80% of patients seeking healthcare in 2015 , although only around 58% of patients were administered PEP appropriately according to guidelines . Despite an increase in overall PEP use until 2015 , the main cause of inappropriate PEP was ‘under-use’ of vaccines or serum when required for high-risk bites . The vaccine shortage experienced in 2015 reduced PEP use for the following two years , but also reduced the appropriate administration to 45% in 2017 , and ‘under-use’ increased by 12% from 2015 to 2017 . This suggests that vaccine shortages resulted in inappropriate PEP use rather than a more rationalized use . Given that no human rabies cases of dog-mediated rabies were observed as a result of inappropriate PEP use as defined by current guidelines , this indicates that further improvements can be made in identifying the risk that a biting animal does have rabies . Updating health workers knowledge on both national and state guidelines for PEP administration is urgently required , particularly in northern states such as Roraima , where the turnover of health professionals is high . One approach to reduce PEP costs in Brazil would be adoption of the latest WHO recommendations [29] . We estimate that in 2017 , the use of the 1-week abridged ID regimen could have saved at least 4 . 08 M USD only on vaccine volume . This reduction is much higher than changing to the 4 dose IM regimen recommended by the Brazilian MoH and the WHO . Reductions in vaccine volume related to adopting WHO recommendations will come with a cost of training health workers to use this technique and replace the current vial/transport system of PEP . Cost reduction of vaccines could outweigh possible obstacles associated with changing to ID , but a cost-benefit analysis should evaluate whether cost reductions could outweigh costs associated with training . Several countries ( Philippines , Sri Lanka , Bhutan , several states in India , Bangladesh , and Madagascar ) with a high number of people attending health care after a bite have successfully switched from IM to ID vaccination [40] . Furthermore , ID vaccination is already undertaken in several Brazilian states as pre-exposure prophylaxis of professionals working in at-risk laboratories or activities . Many health workers therefore already have experience of ID vaccination with the BCG vaccine for tuberculosis disease and more recently with fractionated dosing of yellow fever vaccine . A more detailed analysis of SINAN data at the municipality level could provide evidence of which health centers would have the highest levels of dose sharing based on patient throughput and identify small health centers where vial sharing is not possible due to a low number of patients seeking health care each day . The economic savings associated with a change in PEP administration could be used , for example , to improve activities of rabies surveillance and prophylaxis in a large population of Brazil , including people exposed to vampire bat bites . Our analysis of the SINAN data allows several conclusions to be drawn about potential for improvements to surveillance and PEP administration at the state and national level , however we acknowledge several limitations . First , the SINAN dataset includes a small but unknown number of ‘duplicate’ records , corresponding to patients accessing more than one health centre and generating more than one SINAN form . We expect this percentage to be low and not to bias results at the state level . Second , errors on data recording , which are likely generating many of the rabies-positive dogs in the database , can mislead the assessment of whether PEP was appropriately administered . For example , health workers can mis-record the ‘rabid dog’ field in the form manually or when transcribing data from handwriting to digital format . These errors , adding ‘false positives’ to the dataset , will prevail if not checked by local or national authorities responsible for SINAN . In our analysis , although we reported and excluded missing data and attempted to minimize the influence of data inconsistencies , inaccuracies will undoubtedly still be present . Third , SINAN does not capture patients that seek private health care following a dog bite or patients who are bitten that do not seek health care . Therefore , our estimates of bite incidence , although high , are likely an underestimate of the true burden of dog bites , although more than 70% of patients seek emergency healthcare within the public SUS system [41 , 42] . Fourth , reports of rabid dogs to SIRVERA likely underestimated the actual number of rabid dogs in Brazil . This is mainly due to a non-automatic transmission of SINAN data to SIRVERA and a lack of data uploading from states or national levels . We stress the need to improve partnerships between national and regional authorities for a more coordinated effort to eliminate canine rabies in the continent . However , if rabies is circulating in dogs within a state , we expect that at least occasionally cases will be reported to SIRVERA because canine rabies is a mandatory notifiable disease in Brazil , dog rabies cases are routinely monitored and other species infected with rabies ( e . g . bats ) are detected through passive surveillance [1 , 43 , 44] . Detection capacity will ultimately depend on laboratory capacity , which remains variable between states and countries of Latin America [1] . Fifth , we did not obtain data from the last three months of 2017 from all states because the data , requested in 2018 , was not yet entered to the national database . This shows that months of delays can limit timely control measures based on data from SINAN . Finally , data on the number of doses administered , although in the SINAN form , is not filled electronically to the main national system , which reduces our ability to estimate both appropriate PEP and potential for cost savings . Previous studies at the state level in Brazil have shown that 8–55% of patients do not complete the PEP prescribed and 70% of PEP courses are interrupted due to withdrawal of patients [45 , 46] . Overall , these limitations reinforce the need for improving the accuracy of SINAN so that it can be used as an efficient passive surveillance tool for dog rabies . In summary , this study demonstrated an increasing and uneven burden related to the prevention of dog-mediated rabies to the Brazilian health care system , despite a major reduction in disease transmission throughout the country . Our results highlight that improving the training of health workers , automating the data recording system to identify and require correction of errors , and changing PEP administration protocols could substantially reduce healthcare costs . Focusing on Brazil as an example , this study calls for improving passive surveillance and strengthening ‘One health’ links between health workers and veterinarians in order to move forward towards the elimination of dog-mediated rabies across Latin America . Better use of PEP and enhanced surveillance could enable more efficient and effective allocation of resources towards other emerging public health problems such as PrEP and PEP for vampire bat rabies prevention .
|
Dog-mediated rabies has declined to only a few cases in Latin America over the last decade . Brazil has the largest human and dog population of Latin America . Despite the decline of canine rabies , the country’s public health system still spends millions of dollars annually on half a million patients seeking health care for dog bites . In this study , we analysed a decade of national surveillance data on dog bites . These data suggest that health workers report dog rabies in many states where the disease is likely to be absent , with false positive cases frequently reported into the surveillance system . In addition , only half of patients appear to receive the appropriate rabies post-exposure prophylaxis as recommended by the Ministry of Health . We estimated that Brazil could save up to USD 6 million per annum on vaccine by reducing the number of doses administered during prophylaxis and adopting the intradermal vaccine delivery technique following the latest WHO recommendations . Our study highlights an urgent need for updating health care workers on canine rabies knowledge , prophylaxis and assessment of dog bites to improve prophylaxis provision and surveillance of dog rabies .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"post-exposure",
"prophylaxis",
"immunology",
"tropical",
"diseases",
"geographical",
"locations",
"vertebrates",
"dogs",
"animals",
"mammals",
"vaccines",
"preventive",
"medicine",
"rabies",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"vaccination",
"and",
"immunization",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"zoonoses",
"south",
"america",
"epidemiology",
"brazil",
"people",
"and",
"places",
"infectious",
"disease",
"surveillance",
"eukaryota",
"prophylaxis",
"disease",
"surveillance",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"amniotes",
"organisms"
] |
2019
|
An evaluation of Brazil’s surveillance and prophylaxis of canine rabies between 2008 and 2017
|
The phenomenology and cellular mechanisms of cortical synaptic plasticity are becoming known in increasing detail , but the computational principles by which cortical plasticity enables the development of sensory representations are unclear . Here we describe a framework for cortical synaptic plasticity termed the “Convallis rule” , mathematically derived from a principle of unsupervised learning via constrained optimization . Implementation of the rule caused a recurrent cortex-like network of simulated spiking neurons to develop rate representations of real-world speech stimuli , enabling classification by a downstream linear decoder . Applied to spike patterns used in in vitro plasticity experiments , the rule reproduced multiple results including and beyond STDP . However STDP alone produced poorer learning performance . The mathematical form of the rule is consistent with a dual coincidence detector mechanism that has been suggested by experiments in several synaptic classes of juvenile neocortex . Based on this confluence of normative , phenomenological , and mechanistic evidence , we suggest that the rule may approximate a fundamental computational principle of the neocortex .
Animal learning is believed to occur primarily through changes in synaptic strengths . Experimental work has revealed an increasingly detailed picture of synaptic plasticity [1] , [2] , at the level of both phenomenology and cellular mechanisms . However an understanding of synaptic plasticity's computational role in cortical circuits lags far behind this experimental knowledge . While spike timing dependent plasticity ( STDP ) has gained much attention , the STDP rule is simply a description of how synapses respond to one particular paradigm of temporally offset spike pairings , and is neither a complete description of synaptic behaviour , nor a computational principle that explains how learning could occur in cortex [3]–[6] . It therefore seems likely that STDP is just an approximation to a more fundamental computational principle that explains the form and function of cortical synaptic plasticity . Such a principle would not only have to be consistent with experimental results on the phenomena and mechanisms of synaptic plasticity , but also explain why it provides a computational benefit . A strong test of the latter is whether simulated cortex-like circuits employing the same principle can learn to perform real-world information processing tasks . The nature and mechanisms of synaptic plasticity differ between brain regions , developmental stages , and cell types , likely indicating different computational roles of synaptic plasticity in different contexts . In the sensory cortex , synaptic plasticity is strongest at early ages [7] , and is believed to play an important role in the development of sensory representations . The juvenile cortex learns to form representations of sensory stimuli even in the absence of any required behavior or reward: the acquisition of native language sounds , for example , begins through passive exposure to speech before infants can themselves speak [8] . The outcome of such learning is not simply a more faithful representation of the learned stimuli — which are already faithfully represented by sensory receptors themselves — but a transformation of this representation into a form where relevant information can be more easily read out by downstream structures [9] . This problem of forming easily-decoded representations of a data set , without reward or training signals , is called “unsupervised learning” [10] , [11] . Unsupervised learning has long been proposed as a primary function of the sensory cortex [12] , [13] . An intriguing connection between cortical plasticity and artificial algorithms for unsupervised learning arises from work of Bienenstock , Cooper , and Munro ( BCM ) [14] . A key feature of the BCM rule is that inputs occurring when the postsynaptic firing rate is below a “plasticity threshold” will be weakened , whereas inputs firing when postsynaptic firing rate exceeds the plasticity threshold will be strengthened; the rule is made stable by allowing the plasticity threshold to “slide” as a function of mean postsynaptic activity . The BCM rule operates at the level of firing rate neurons , and at this level has been successful in modelling a number of experimental results such as the development of visual receptive fields [15] . Theoretical analysis [16] has shown that this scheme allows simplified neuron models to implement an unsupervised learning algorithm similar to projection pursuit [17] or independent component analysis ( ICA ) [18] , [19] , extracting non-Gaussian features of their inputs which are a priori more likely than Gaussian features to correspond to signals of interest . Although the BCM theory was originally defined at the level of firing rates , more recent modeling work [20]–[24] has reproduced a dependence of the direction of synaptic plasticity on postsynaptic firing rate in spike-based neurons . In cortical neurons synaptic plasticity depends not only of postsynaptic firing rates , but also shows a similar dependence on subthreshold depolarization , with presynaptic spikes during strong postsynaptic depolarizations leading to potentiation , and during weak postsynaptic depolarization leading to depression [25] , [26] . Computational models incorporating such behavior have successfully matched several experimental findings of in vitro plasticity [23] . In the present work , we present a framework for unsupervised learning in cortical networks . The rule is derived as an optimization of the skewness of a cell's postsynaptic membrane potential distribution under a constraint of constant firing rate , and leads to a voltage-dependence similar to that observed experimentally [25] . We term the resulting framework the Convallis rule after the Latin word for “valley” , in reference to the shape of the voltage objective function . We show that the Convallis rule causes simulated recurrent spiking networks to perform unsupervised learning of speech sounds , forming representations that enable a downstream linear classifier to accurately identify spoken words from the spike counts of the simulated neurons . When presented with paired pre- and postsynaptic spikes or other paradigms used in vitro , predictions of the Convallis rule more accurately match experimental results than the predictions of STDP alone . Furthermore , simulation of STDP alone ( or of previously published plasticity rules [21] , [23] ) produced poorer performance on speech learning than the full Convallis rule , indicating that STDP may be just one signature of a cortical plasticity principle similar to Convallis . The mathematical form of the Convallis rule suggests implementation by a dual coincidence detector mechanism , consistent with experimental data from juvenile sensory cortex [6] , [27]–[33] .
We derived the Convallis rule from two principles , analogous to those underlying artificial unsupervised learning algorithms such as ICA . The first principle is that synaptic changes should tend to increase the skewness of a neuron's subthreshold membrane potential distribution . Because the physical processes that produce structure in real-world data sets often show substantial higher-order moments , whereas random and uninformative combinations follow a Gaussian distribution , projections with non-Gaussian distribution are a priori more likely to extract useful information from many real-world data sets [19] . The second principle is that despite synaptic plasticity , neurons should maintain a constant average firing rate . This principle is required for stable operation of the rule , and is again analogous to a step of the ICA algorithm ( see below ) . To derive the rule , we first defined an objective function that measures the non-Gaussianity of the subthreshold distribution . The function has the valley-shaped form shown in Figure 1B . Optimization of this objective function ensures that the postsynaptic neuron spends as much time as possible close to either resting potential or spiking threshold , but as little time as possible in a zone of intermediate membrane potential , i . e . exhibiting a skewed , non-Gaussian subthreshold distribution . The form of used in simulations is described in the Materials & Methods , although our results did not depend critically on this precise formula ( data not shown ) . To implement the first principle of skewness optimization , we first compute the derivative of this objective function with respect to the neuron's input weights . Making certain assumptions ( see Materials and Methods for a full derivation ) we obtain: ( 1 ) where is the reversal potential of synapse , is the rest voltage of the neuron , are the times of action potentials incoming onto synapse , and is the shape of a postsynaptic potential elicited by synapse . When a presynaptic input fires shortly before the neuron is close to spiking threshold , the integrand is positive leading to an increase in synaptic weight , but when a presynaptic neuron fires shortly prior to a potential only just above rest the integrand is negative leading to a decrease in synaptic weight . This voltage dependence is similar to that observed experimentally in cortical neurons [25] and also employed in previous phenomenological models [23] . We note that a direct computation of this integral would be computationally prohibitive , as it would require numerical solution of a differential equation for every synapse and at every time step of the simulation . Tractable simulation of this rule was however made possible by a trick that enabled solution of only a single differential equation per neuron ( see Materials and Methods ) . In our simulations , voltage was reset to a level of −55 mV after action potential firing , followed by an afterdepolarization simulating the effects of active dendritic conductances [34] ( see Materials and Methods ) . This reset mechanism , rather than the reset to rest commonly employed in integrate-and-fire simulations , was necessary in order to produce voltage traces similar to those seen in experimental recordings of cortical pyramidal cells ( see Figure S1 ) , and also played an important role in matching in vitro plasticity results ( see below ) . While equation 1 is sufficient to implement our first principle of skewness optimization , we found that better learning performance , as well as a closer match to physiological data , could be obtained with an additional feature modeled after the statistical technique of shrinkage [11] . Specifically , the integrand of equation 1 was not used directly to modify weights , but first convolved with a decaying exponential to yield a function , and then passed through a nonlinear shrinkage function to ensure plasticity only occurs in response to multiple coincidences: ( [22] , [24] , [35]; see Materials and Methods for more details ) . This ensures that weight changes occur only due to reliable and repeated relationships between presynaptic activity and postsynaptic membrane potentials , rather than random occurrence of single spikes . An illustration of how pre- and post-synaptic activity lead to weight changes under this rule is shown in Figure 1C . Physiologically , such an integration mechanism could be instantiated via self-exciting kinases as suggested previously [22] . The second principle underlying the Convallis rule is a constraint on the mean firing rate of each neuron to a target value . Analogous principles are also often found in machine learning algorithms: in ICA , for example , the root-mean-square activity of each unit is fixed at a constant value by a constraint on the weight vector norm together with sphering of inputs [19] . Such constraints are typically implemented in one of two ways: by including a penalty term in the objective function , whose gradient is then added to the learning rule resulting in “weight decay”; or by repeated projection of the system parameters onto a subspace satisfying the constraint [19] . In our simulations , we found that simple gradient ascent was not effective at enforcing stability , and therefore used a projection method . This was implemented by a mechanism which responded to deviations from the target firing rate by linearly scaling all excitatory synaptic weights up or down [36] , and suppressing activity-dependent plasticity until the rate constraint was restored ( Figure 1D; see Materials and Methods for details ) . Physiologically , the “metaplasticity” [37] , [38] required for suppression of synaptic changes until rate homeostasis is restored , could be instantiated via one of the many molecular pathways gating induction and expression of synaptic plasticity . To study the rule's effects , we first considered the behaviour of an individual neuron implementing the rule on a simple artificial data set . The parameters used in the learning rule were fixed in this and all subsequent simulations ( see Materials and Methods for more details ) . For this first artificial task , inputs consisted of a population of 1000 excitatory sources ( see Figure 2A ) . The simulated postsynaptic neuron received plastic excitatory synapses from these sources , as well as constant inhibitory background with input at 10 Hz through 250 synapses which were not subject to plasticity . We first considered a simple case where inputs fired as Poisson spike trains with rates determined as spatial Gaussian profiles whose centre changed location every 100 ms ( Figure 1A; see Materials and Methods ) [21] , [22] , [39] . When weights evolved according to the rate constraint only , no structure was seen in the weight patterns . With the Convallis rule , postsynaptic neurons developed strong weights from groups of closely-spaced and thus correlated inputs , but zero weights from neurons uncorrelated with this primary group . When weights instead evolved by classical all-to-all STDP augmented by the rate constraint ( called rcSTDP , see Materials and Methods for details ) , the firing rate was kept at the desired value of 10 Hz , and weights became more selective , but in a manner less closely related to the input statistics . Examination of post-synaptic voltage traces showed that after learning with the Convallis rule , but not after rate constraint alone , the membrane potential spent considerably longer close to resting potential ( Figure 2C ) , corresponding to an increased skewness of the membrane potential histogram , ( Figure 2D; , t-test ) . This in turn reflected the development of selectivity of the neurons to particular stimuli ( Figure 2E ) ( , t-test ) . Application of rcSTDP caused an increase in skewness tuning intermediate between rate constraint alone and the Convallis rule , even after optimizing by parameter search ( , t-test; see Figure S2 ) . This confirms that the Convallis rule is able to perform unsupervised learning in a simple artificial task , causing neurons to select inputs from groups of coactive neurons; STDP produces a poorer approximation to the same behavior . We next asked whether the Convallis rule would enable individual simulated neurons to perform unsupervised learning in a real-world problem . Because we are interested in the development of cortical representations of sensory stimuli , we asked whether the Convallis rule could promote unsupervised formation of representations of speech sounds . Spike train inputs were generated from the TIDIGITS database of spoken digits [40] , by pre-processing with a cochlear model filter bank [41] , followed by transformation into inhomogeneous Poisson spike trains that contacted the simulated neuron with a range of synaptic delays ( Figure 3A; see Materials and Methods ) . Figure 3B ( top row ) shows a representation of the output of the cochleogram for utterances of the digits “four” , and “five” . To the right is a pseudocolor representation of the excitatory weights developed by neurons initialized to random weights and trained on 326 utterances of all digits by the rate constraint mechanism alone , by the Convallis rule , or by rcSTDP . Each digit was repeated ten times . Figure 3B ( lower three rows ) shows the response of these three neurons to a test set consisting of previously unheard utterances of the same digits by different speakers . The neuron trained by Convallis responds selectively to “four” while the response to “five” is largely eliminated , whereas the neuron trained by rate constraint alone responds equally to both . Thus , the Convallis rule has enabled the neuron to develop a differential response to the presented digits , which has generalized to utterances of the same digits spoken by new speakers . To verify that this behaviour holds in general , we performed five thousand independent simulations of the Convallis rule in single neurons , with excitatory and inhibitory inputs drawn from the simulated cochlear cells , each trained by 10 presentations of the TIDIGITS training set , which we found sufficient to ensure convergence of all learning rules ( Figure S3 ) . Each simulation began from a different random weight configuration . The mean firing rate constraint was fixed to 1 . 5 Hz for all cells . As previously seen with artificial inputs , the membrane distribution produced in response to this real-world input was more skewed after training with the Convallis rule ( Figure 4A for the example cell shown in Figure 3 , Figure 4B for population summary ) . On average , over 1000 independent runs , there was a significant difference in skewness between Convallis and rate constraint alone , with rcSTDP producing an intermediate increase in skewness ( ) . We measured the selectivity of the simulated neurons using an F-statistic that measured differences in spike count between different digits ( see Materials and Methods ) . The Convallis rule caused neurons to become more selective ( , t-test ) , whereas application of rate constraint alone or rcSTDP led to output neurons that were actually less selective than the raw cochleogram input ( Figures 4C for the same example cell shown in Figure 3 , Figure 4D for population average ) . Similar results were found when comparing Convallis to multiple implementations of the STDP rule as well as for other plasticity rules described in the modelling literature [21] , [23] ( see Figure S4 ) . The aim of unsupervised learning is to generate representations of input data that enable downstream neurons to easily form associations with them . Although complete information about the stimulus is of course present in the raw input , a downstream cell may not be able to extract this information unless it is represented in a suitable form . We next asked whether the representation generated by the Convallis rule allowed improved classification by a linear downstream readout in which spike timing information was discarded; this choice was motivated by results indicating that information in higher sensory cortices can be progressively more easily read out in such a format [9] . Specifically , we used a linear support vector machine to predict which digit was uttered , from the spike counts of a population of simulated cells arranged in a feedforward configuration ( Figure 4E; see Materials and Methods; note that while the SVM was trained with a biologically unrealistic quadratic programming algorithm , the same solution would be found by a large-margin perceptron [42] ) . Figure 4F shows the generalization performance of the classifier ( measured on the TIDIGITS test set ) as a function of population size . Performing the classification from a layer of neurons that used rate constraint alone produced an improvement over prediction directly from the cochleogram . The size of this improvement increased with the number of cells used , consistent with reports that large numbers of random projections can provide useful data representations [43] , [44] . Applying the Convallis rule produced a substantially improved representation over the rate constraint alone ( 18% vs 29 . 9% errors; , t-test ) , whereas rcSTDP produced an intermediate improvement ( 25 . 9% error; , t-test ) . Evaluation of performance with time-reversed digit stimuli indicated that the neurons had learned specific temporal features of the input rather than simply frequency content ( Figure S3 ) . Evaluation of several other proposed learning rules for spiking neurons taken from the literature , such as rcNN-STDP ( STDP with interactions only between neighbouring pairs of spikes , and the rate constraint ) , triplet STDP [21] with rate constraint , or phenomenological rules also based on post-synaptic voltages [23] ( see Materials and Methods for details ) also confirmed that their performance did not match those of the Convallis rule ( 25 . 0% , 27% and 25 . 9% vs 18 . 0% errors; see Figure S4 ) . The above analysis showed that the Convallis rule caused individual neurons to develop selective representations of the digit stimuli , which when arranged together in a feedforward configuration formed a population code that enabled the spoken digit to be decoded with 82% accuracy . The cortex , however , is a recurrent rather than a feedforward network , and we next asked whether a recurrent architecture would lead to further improved classification performance ( Figure 5A ) . Recurrent spiking network models can exhibit multiple global patterns of population activity , of which the asynchronous irregular state provides the closest match to in vivo cortical activity in alert animals [45]–[47] . We set the initial conductances ( prior to training ) to obtain asynchronous irregular activity at a mean spontaneous activity at 1 . 5 Hz , and with the coefficient of variation of inter-spike intervals ( CV ISI ) equal to 1 . 1 ( Figure 5C; see Materials and Methods ) . When a sound input was presented to the network , mean firing rates increased from 1 . 5 Hz to 15 Hz ( Figure 5B ) , while remaining in the asynchronous irregular regime . To measure the ability of the Convallis rule to produce unsupervised learning in recurrent spiking networks , we trained the network with 10 iterations of the TIDIGITS training set , which were again sufficient for convergence ( see Figure S5 ) . All recurrent excitatory connections in the network were plastic , while inhibitory and input connections were fixed . Running the learning rule did not disrupt the asynchronous irregular dynamics of the network , as indicated by the ISI CV , mean firing rate distribution , and mean spontaneous correlation values ( Figure 5B and Figure 5C , D , E ) . As in the feed-forward case , the network's constituent neurons showed increased tuning and membrane potential skewness after training ( Figure 5F , G ) . The ability to perform unsupervised learning in a recurrent network was again measured by ability to identify the spoken digits using a linear classifier trained on the spike counts of the network's excitatory neurons ( Figure 5H ) . We note that even prior to training , as in the feed-forward case , the representation generated by the recurrent network allowed higher classification performance than the raw cochleogram input ( 5 . 8% error ) , consistent with previous reports that randomly connected “liquid-state” networks can compute useful representations of spatiotemporal input patterns [48]–[50] . Training with the Convallis rule significantly boosted performance to reach 3 . 3% error ( Figure 5H ) . As in the feedforward case , application of rcSTDP produced error rates more than 50% higher than those of the full Convallis rule ( Figure 5H ) ( 5 . 1% error; ) . Thus , the Convallis rule enables spiking neurons to perform unsupervised learning on real-world problems , arranged either in a feedforward or in a recurrent configuration . As in the feed-forward scenario , performance with time-reversed digit stimuli indicated that the neurons had learned specific temporal features of the input rather than simply frequency content ( Figure S5 ) . Once again , we were unable to produce comparable results with rules previously published in the literature , which resulted in error rates more than 50% higher than those produced by Convallis ( 5 . 2% and 5 . 3% errors for rcNN-STDP and rcTriplet , respectively; see Figure S6 ) . The Convallis rule was derived mathematically from an optimization principle , rather than by fitting to experimentally measured parameters . Before suggesting that an analogous process might occur in the cortex , it is thus important to check how a neuron employing this rule would behave in paradigms that have been used to experimentally probe cortical synaptic plasticity . Although we found simulation of rcSTDP alone produced poorer learning than Convallis , STDP is a robustly observed experimental result that the Convallis rule must reproduce if a similar rule does occur in cortical neurons . To test this , we applied a spike-pairing paradigm to two simulated cells , using the same parameters as in the previous speech-classification simulations . Figure 6A shows a close-up view of the Convallis rule in operation for three spike pairings . The green trace shows a pre-post interval of 10 ms . Here , the period immediately after the presynaptic spike ( where is positive ) contains an action potential , leading to a high value of , and synaptic potentiation . The black trace shows a post-pre pairing of −10 ms . In this case , the period immediately following the presynaptic spike occurs during the postsynaptic afterdepolarization , a moderately depolarized voltage range for which is negative . The gray trace shows a pre-post interval of 30 ms , longer than the duration of the kernel . Now , the postsynaptic potential during the entire period while is very close to rest , leading to a value of close to zero , and neither potentiation nor depression . Figure 6B shows the results of similar simulations for a range of pre-post intervals , applying 60 spike pairings performed at 1 Hz . The Convallis rule reproduces a STDP curve similar to bi-exponential form found in many computational models [51] . STDP does not fully summarize the nature of cortical synaptic plasticity , which cannot be explained by linear superposition of effects caused by individual spike pairs . Various in vitro pairing protocols , in hippocampus [52] or in cortex [26] , [53] , [54] showed that LTP and LTD pathways can not be reduced to additive interactions of nearby spikes . Therefore , we next asked whether the Convallis rule would also be able to predict additional experimental results beyond STDP . As one of the pieces of evidence in favor of the original BCM theory is the dependence of the sign of plasticity on the rate of tetanic stimulation , we asked if the Convallis rule could produce a similar result . To simulate extracellular stimulation in vitro , we synchronously simulated multiple excitatory and inhibitory presynaptic synapses at a range of frequencies ranging from 0 . 1 Hz to 100 Hz , and investigated the amount of plasticity produced in a downstream neuron . Consistent with experimental data in cortical [55] as well as hippocampal [56] slices in vitro , low frequencies resulted in depression while higher frequencies resulted in potentiation ( Figure 6C ) . As a second example , we considered spike triplets in paired recordings ( see Materials and Methods ) . Linear superposition of STDP would predict that presentation of post-pre-post spike triplets should cause no synaptic change; experimentally however , this causes robust potentiation ( although pre-post-pre triplets do not ) [52] . The Convallis rule is able to reproduce this finding ( Figure 6D ) . A third example of nonlinear plasticity effects concerns the spike pairing repetition frequency . In cortical slices , post-pre pairings at low repetition rates cause synaptic depression , but this converts to potentiation for fast enough repetition rates , a non-linear effect that likely reflects subthreshold phenomena [26] . The Convallis rule produces a similar effect ( Figure 6E , top ) . For pre-post pairings , potentiation is not seen experimentally at low ( 0 . 1 Hz ) repetition rates in L5 of juvenile cortex [26] . The Convallis rule also replicated this finding ( Figure 6E , bottom ) ; for this , the shrinkage mechanism was critical ( data not shown ) . Finally , we asked whether network-level plasticity using the Convallis rule left traces similar to those seen experimentally in vivo . Specifically , we assessed whether simulated neurons with similar receptive fields would exhibit higher connection probabilities , as has been reported in mouse visual cortex [57] , [58] . This was indeed the case ( Figure 6F ) , strongly for Convallis ( , t-test ) , weakly for rcSTDP ( , t-test ) , but not for rate constraint alone . We therefore conclude that the Convallis rule is consistent with a wide range of plasticity phenomena described in vitro and in vivo , supporting the possibility that a similar process occurs in cortex . If cortical neurons do indeed implement a rule similar to Convallis , what cellular mechanisms might underlie it ? Plasticity in the developing neocortex appears to involve different cellular mechanisms to those of the well-studied hippocampal Schaffer collateral synapse . One of the leading mechanistic models of hippocampal synaptic plasticity is the calcium concentration hypothesis [59]–[61] . In this model , both LTP and LTD are triggered by calcium influx through NMDA receptors , with LTP triggered by high Ca2+ concentrations , and LTD triggered by low concentrations ( see Figure 7A ) . This model has a similarity with Convallis in that weak activation causes LTD and strong activation LTP . Nevertheless , the functional form of the Convallis rule ( Eqn . 1 ) has a critical difference to the calcium hypothesis . In the Convallis rule , the nonlinear function that determines the sign of synaptic plasticity operates directly on the membrane potential prior to coincidence detection with presynaptic input , whereas in the calcium rule this nonlinearity happens after coincidence detection . This leads to a diverging experimental predictions , with the calcium model predicting a triphasic STDP curve [60] ( but see also [61] ) . This has been reported in some hippocampal experiments [62] , [63] , but not in the neocortex ( Figure 7B ) . A substantial body of experimental evidence suggests that in juvenile neocortical neurons , the potentiation and depression components of STDP are produced by different cellular mechanisms [27]–[33] . While these data are obtained from different sensory cortices ( visual , somatosensory ) , and for different cortical synapse types ( typically L4→L2/3 or L5→L5 ) , they suggest a hypothesis for a common mechanism underlying STDP in at least some neocortical synapses [6] . In these systems , LTP appears of the conventional type , dependent on postsynaptic NMDA activation caused by coincident glutamate release and release of magnesium block by postsynaptic depolarization . For LTD however , induction is independent of postsynaptic NMDA receptors , and instead appears to be induced by a separate mechanism in which postsynaptic phospholipase Cβ acts as a coincidence detector for the activation of group I metabotropic glutamate receptors , and postsynaptic depolarization detected by voltage-sensitive calcium channels ( VSCCs ) , leading to presynaptic expression of LTD via retrograde endocannabinoid signaling . Importantly , the VSCCs implicated are of the low-threshold T-type [27] , [30] . Together , these results suggest a hypothesis that in the developing sensory cortex , there exist two separate molecular coincidence detectors for LTP and LTD , and that the coincidence detector for LTD has a lower voltage threshold ( Figure 7C; [6] , [32] . The mathematical form of the Convallis rule is consistent with just such a mechanism . The function can be expressed as a difference of two non-negative functions , both sigmoidal in shape , but with having a lower threshold . The rule can then be expressed as a sum of two termsThis equation has a natural mechanistic interpretation , as the result of two coincidence detectors . The first , corresponding to , is activated when the membrane is strongly depolarized after a presynaptic spike fires , and leads to synaptic potentiation . The second , corresponding to , is activated when the membrane is moderately depolarized after presynaptic firing , and leads to synaptic depression . Linear addition of and would be expected due to their implementation by separate coincidence detectors , triggered by spatially separated calcium sources [64] . The mathematical form of the Convallis rule therefore bears a striking resemblance to a leading hypothesis for the mechanisms synaptic plasticity in the juvenile sensory cortex .
We derived a synaptic plasticity rule for unsupervised learning in spiking neurons , based on an optimization principle that increases the skewness of subthreshold membrane potential distributions , under the constraint of a fixed mean firing rate . Applying this rule to a speech recognition task caused individual neurons to develop skewed membrane potential distributions and selective receptive fields both in a feedforward configuration and within a recurrent network . The spike count outputs of the recurrent network were sufficient to allow good readout by a linear classifier , suggesting that this unsupervised rule had enabled the network to form an easily-decoded representation of the key spatiotemporal features of the input that distinguished the spoken digits . Simulation of paradigms used to study synaptic plasticity in vitro produced similar behaviour to that found experimentally . Furthermore the form of the rule is consistent with a dual-sensor mechanism that has been suggested experimentally for cortical neurons . The phenomenon of spike-timing dependent plasticity has been robustly observed in a large number of neuronal systems ( see for example [65] for review ) . It is important to remember however that STDP is not a fundamental description of synaptic plasticity , but simply an experimental observation that describes how synapses respond to one particular stimulus of temporally offset spike pairings [3]–[6] . We found that the Convallis rule , when presented with paired spikes , reproduced a biphasic STDP curve . However , implementation of all-to-all STDP alone produced both a worse fit to experimental plasticity paradigms , and poorer unsupervised learning of speech sounds than the full Convallis rule . Implementation of other learning rules described in the literature which match more experimental observations than STDP alone [21] , [23] also produced poorer results . The higher performance of Convallis compared to rules based on spike timing alone may reflect the fact that the subthreshold potential conveys additional information that is useful to guide synaptic plasticity . We note however that better unsupervised learning was also obtained compared to a previous phenomenological rule [23] that exhibited a similar voltage dependence , but was derived primarily to match experimental observations , rather than derived from an optimality principle . Other than the similar voltage dependence , this rule was different in many details to Convallis , for example with regard to the precise temporal relationship of presynaptic activity and postsynaptic voltage required for potentiation or depression . The derivation of these relationships from an optimality principle might underlie Convallis' better performance . Additionally or alternatively , the difference might reflect a difference in the stabilizing mechanism between the two rules . For Convallis , we found that a penalty-based weight decay term could not provide optimal stability , and much better performance was obtained with a hard constraint on firing rate with plasticity inhibited until the constraint was satisfied . In our simulations of the framework of [23] , we were similarly unable to obtain robust stabilization of firing rates , which may have contributed to poorer learning performance . Although unsupervised learning has long been proposed as a primary function of the sensory cortex [12] , [13] , the circuit mechanisms underlying it are still unknown . One influential class of models holds that unsupervised learning occurs through the coordinated plasticity of top-down and bottom-up projections , leading to the development of “generative models” by which the brain learns to form compressed representations of sensory stimuli [66]–[68] . Although these models have produced good performance in real-world tasks such as optical character recognition , the mapping between these abstract models and concrete experimental results on cortical circuitry and plasticity is as yet unclear , and their implementation in spiking neuron models has yet to be demonstrated . Here we describe an alternative scheme for unsupervised learning in cortex , in which every neuron acts essentially independently , using a plasticity rule to form an unsupervised representation of its own synaptic inputs . Despite the simplicity of this approach , it could be applied in recurrent spiking networks to produce good unsupervised learning . We hypothesize that incorporating other mechanisms to coordinate plasticity at the network level [69] may further improve network performance . In psychophysical experiments , perceptual learning is typically studied by repeated practice at sensory discrimination tasks . In such cases , learning might be boosted by attention directed to the stimuli to be learned , or rewards delivered after a correct response . Nevertheless , purely unsupervised perceptual learning can also occur in humans , both in development [8] and adulthood [70] . The Convallis rule as simulated here is a purely unsupervised rule that operates continuously . The effects of attention , reward and task-relevance could be captured in the same framework by a modulation of learning rates by neuromodulatory tone [71] , [72] . This would allow cortical networks to devote their limited resources to representing those stimulus features most likely to require behavioural associations . Models of synaptic plasticity typically fall into three classes: phenomenological models , which aim to quantitatively summarize the ever-growing body of experimental data [21]–[23]; mechanistic models , which aim to explain how these phenomena are produced by underlying biophysical processes [60] , [73]; and normative models , which aim to explain the information-processing benefit that synaptic plasticity achieves within the brain [74]–[79] . The Convallis rule bridges all three levels of analysis . Being mathematically derived from an optimization principle , it belongs in the normative class , and the fact that it can organize recurrent spiking networks to perform unsupervised learning in a real-world task supports the idea that a similar principle could enhance cortical information processing . The rule is consistent with a number of experimental findings on cortical plasticity , including but not limited to STDP , suggesting that a similar principle may indeed operate in cortical cells . Finally , the functional form of the Convallis rule has a direct mechanistic interpretation in terms of a dual coincidence-detector model , for which substantial evidence exists in neocortical synapses [27]–[32] , [32 , 33] . Based on this confluence of normative , phenomenological , and mechanistic evidence , we suggest that the Convallis rule may approximate a fundamental computational principle of the neocortex .
Simulations of the spiking neurons were performed using a custom version of the NEST simulator [80] and the PyNN interface [81] , with a fixed time step of 0 . 1 ms . In all simulations , we used an integrate-and-fire neuron model with a membrane time constant , a leak conductance of , and a resting membrane potential . Spikes were generated when the membrane potential reaches the threshold . To model the shape of the action potential , the voltage was set to 20 mV after threshold crossing , and then decayed linearly during a refractory period of time to a reset value of , following which an exponentially decaying after-depolarizing current of initial magnitude 50 pA and time constant was applied . We used this scheme with a high reset voltage and ADP , rather than the more common low reset value , as it provided a better match to intracellular recordings in vitro and in vivo ( see supplementary Figure S1 ) . Synaptic connections were modelled as transient conductance changes with instantaneous rise followed by exponential decay . Synaptic time constants were chosen to be and for excitation and inhibition respectively , and reversal potentials were and . The complete set of equations describing the dynamics of a neuron is thus given by ( 2 ) where , are the incoming synaptic spike trains represented as sums of delta functions . In the Convallis rule , a neuron adapts its synapses in order to optimize an objective function depending on its membrane potential : ( 3 ) To enforce skewness of the distribution of postsynaptic potentials , we chose an objective function that penalized intermediate membrane potential values , but rewarded membrane potentials close to either resting potential or spike threshold . Because the neuron spent considerably less time depolarized than hyperpolarized , the objective function was chosen to reward potentials close to spike threshold more strongly than potentials close to rest . For all simulations in the present paper , we used a sum of a logistic function and of its integral . More precisely: ( 4 ) Parameters values were taken as V0 = −55 mV , V1 = −52 mV , σ0 = 4 mV , σ1 = 2 mv and , and the same parameters were used for both the speech processing application and simulation of in vitro experiments . The shape of was therefore constant in all the simulations of the paper , and its exact form did not appear to be crucial ( as long as a clear valley-shaped function was used ) , since similar results were achieved with a variety of functions ( not shown ) . To derive the Convallis rule , we used a gradient ascent method . Differentiating with respect to incoming synaptic weights gives ( 5 ) To compute , we considered the variable . Equation 2 can be rewritten as ( 6 ) Where is the total synaptic conductance and the synaptic current . Specifically , if are the times at which a particular synapse of weight is active , and if ( if ) is the kernel function representing the conductance time course , ( 7 ) where is the reversal potential of synapse . Inspecting equation 6 , we see that for a conductance-based neuron , integrates with an effective time constant . Approximating by a constant equal to where denotes a running average of the synaptic conductance [82] , we can approximate by the following equation: ( 8 ) where ( 9 ) Note that this approximation holds as long as we ignore the reset mechanism and non-linearity due to the spike , an approximation that will be more accurate when using a “soft” reset mechanism as described here . Substituting in equation 5 , we obtain the following equation for the gradient: ( 10 ) This generic form is similar to previous supervised learning rules that were also based onto the post-synaptic , such as the Tempotron [82] , [83] or Chronotron [84] . As noted by [85] , is used here as a proxy for the input current flowing into the cells , which is the only relevant quantity at the cell level to measure the correlation between incoming pre and post-synaptic activity . To prevent plastic changes for spurious single pairings , plasticity changes are accumulated through the convolution of a slowly decaying exponential , and then expressed at the synapse level only if the accumulated value crosses thresholds and for respectively potentiation and depression . Specifically , we define ( 11 ) The time constant of the slowly decaying exponential is taken to be 1 second throughout the paper . The final weight changes are then given by ( 12 ) where the shrinkage function is defined as ( 13 ) Throughout the paper , we fixed the values of to −10 and 50 respectively . A graph of can be seen in Figure 1C . Note that the weights are clipped to hard bounds values nS and nS . The Convallis rule has therefore have 3 parameters in addition of the shape of : the time at which the changes are accumulated , and those two thresholds for the shrinkage function . Direct calculation of the above integrals would be prohibitive in large-scale simulations , as it would require computing the products , for all synapses and for each time step , resulting in a complexity scaling in , where is the number of synapses , the time step , and the simulation length . To speed up implementation of the algorithm , we write: ( 14 ) where . We can implement the rule much faster by first computing and storing the history for neuron , and computing weight changes as a sum over all input spikes for all synapse , which is of order . To compute , we note that is the convolution of and a filter which is a difference of decaying exponentials ( see Equation 9 ) . By defining , we can write . Integrating by parts , we obtainTherefore , we have a differential equation that can be used to compute look-up tables of for all neurons during this period , by running backwards in time from starting values . Weight changes are then calculated by summing over spikes . We note that this method of running backward in time is simply a trick to speed up execution time , and is equivalent to the original deterministic algorithm . In practice , we perform this by stopping the simulation after the presentation of each input pattern ( T = 1 s ) . This implementation does not impact the results when the frequency of the updates is changed ( data not shown ) , as long as the assumption is valid , which will hold provided the support of the filter is shorter than . Run in isolation , the above rule is unstable , as the response of the neuron tends to accumulate either above or below the plasticity threshold , leading to either explosive increases in synaptic weights or convergence of all weights to zero . In the BCM theory , this problem was solved by a sliding plasticity threshold , computed as a long-running average of the firing history of the post-synaptic neuron . For the Convallis rule we found that a sliding threshold was not necessary , provided a mechanism was in place to constrain the neurons firing rate to a fixed value . We implemented this via “synaptic scaling” [86] , using an approach analogous to the projected subgradient method for constrained optimization . In the projected subgradient method , gradient-following steps are allowed to temporarily break the constraint , but are followed by a projection onto the constraint subspace . Because direct projection onto the subspace of synaptic weights corresponding to the targeted mean firing rate would not be computationally tractable or biologically realistic , we instead used a Proportional-Integral ( PI ) controller [87] to enforce the constraint , and suppress gradient learning until the constraint was re-established . Specifically , we define to be the deviation from target mean firing rate , where is a cell's firing rate computed as a running average over its past-history with a time constant T ( 10 s in our simulations ) and is the targeted mean rate . The output of the PI controller iswhere is a coefficient regulating the contribution of the integral term . The value of balances speed of convergence against the possibility of oscillation; in all simulations , we fixed . To suppress gradient descent until the constraint was satisfied , we scaled the synaptic plasticity rule by a term that was small if either or was not close to zero , leading to a final form of ( 15 ) The parameters were set to and , respectively . We found this latter feature was essential for stable operation of the Convallis rule . In simulations of artificial data ( Figure 2 ) , 1000 excitatory and 250 inhibitory inputs were connected to a single post-synaptic neuron . Only excitatory connections were plastic . Initial values of the weights were drawn from Gaussian distribution with . The values were and , and the target output rate was fixed to 10 Hz . Pre-synaptic neurons were stimulated with wrapped Gaussian profiles of rates spikes/sec , the centre being shifted randomly every 100 ms over all possible positions and with . The tuning index used in Figure 2 was computed as a directional statistic: for each cell , the distance between neuron 0 and 1000 was mapped into an angle , and if is the average firing rate for this particular angle , the tuning was defined as . The closer the tuning is to 1 , the more the neuron is responding only to one particular angle . To test the ability of the rule to perform unsupervised learning in a real-world context , we applied it to a problem of speech recognition , using the TIDIGITS database [40] . This data consists of recordings of eleven English digits ( “zero” to “nine” plus “oh” ) , spoken twice each by 326 speakers of various ages and genders ( man , woman , boy , girl ) , at a sampling rate of 20 KHz . The TIDIGITS database was separated into its standard training and test sets of 167 speakers each . The raw recorded waveforms were pre-processed into spike trains using the Lyon model [41] , to produce a simulated cochleogram of 93 frequency channels . The cochleogram output for each digit was centered in a one second epoch , sampled at 500 Hz , and normalized to equalize the summed activity of all frequencies for all digit utterances . Input spike trains were generated as inhomogeneous Poisson spike trains with intensity function given by the cochleogram output , at an average frequency of 5 Hz . For feedforward simulations ( Figure 3 ) , each target neuron received plastic excitatory projections from 50% of randomly chosen cochleogram cells with initial conductances uniformly drawn in [0 , 10 nS] and synaptic delays uniformly drawn from [0 . 1 ms , 5 ms] , while also receiving static inhibitory projections from all cells in the cochleogram with conductances uniformly drawn in [0 , 40 nS] . For recurrent network simulations , 4500 neurons were simulated with an excitatory/inhibitory neuron ratio of 4∶1 on a square sheet with periodic boundary conditions . Every neuron was sparsely connected with the rest of the network with a connection probability of 5% . Synaptic delays were drawn randomly from a uniform distribution between 0 . 1 and 5 ms . Initial synaptic conductances were taken randomly from Gaussian distributions with means and , and standard deviations equal to a third of their means . To sustain spontaneous activity , each neuron also received an independent Poisson spike train at a frequency of 300 Hz , through an excitatory synapse of weight . Although recurrent connections were uniform , input connections were arranged in a tonotopic manner , with each cochleogram cell projecting with excitatory synapses to a fraction of of neurons in the network , with a probability following a Gaussian profile ( being the distance between the source and a target neuron within the network , and being equal to 0 . 2 unit ) . The mean conductances of the external connections were equal to the recurrent ones , i . e , and all external inputs were fixed rather than plastic . To measure the selectivity of a neuron to the digit stimuli , we used the F-statistic , commonly used in one-way analysis of variance ( ANOVA ) . Specifically , to measure the difference between mean spike counts of each digit , relative to within-digit variance , we computed ( 16 ) where is the spike count the neuron produces on the presentation of digit , is the mean response to digit , the overall mean response , the number of digits , and the total number of stimulus presentations . To quantify the efficacy of unsupervised learning , we evaluated the ability of a downstream linear classifier to identify the digit spoken from the spike counts of each simulated neuron . This approach therefore evaluates the network's ability to form a linearly separable representation of the digit inputs that can be read out without requiring temporal analysis . Specifically , if is a matrix of size containing the mean firing rate of all cells to each of the digit utterances in the training set , and if is an “answer” matrix of size with each row consisting of all zeros except a single 1 indicating the presented digit during this trial , we used multi-class linear support vector machine [88] to find a matrix of size to predict B from A . Performance was evaluated by computing on the test set , and classifying each utterance according to the highest value . The cost parameter used for the support vector machine was set to 0 . 01 . We note that while the SVM was for efficiency trained with a ( biologically unrealistic ) quadratic programming algorithm , the same solution would be found by the perceptron rule [42] . Ridge regression learning was also tried ( data not shown ) , leading to qualitatively similar results . Throughout the paper , the rcSTDP rule is implemented as a normal additive STDP rule combined with the PI mechanism described for the Convallis rule ( Equation 15 ) , in order to ensure that the same output firing rate is achieved . Optimization of this rule's parameters is described in Figure S2 . To compare the Convallis rule with NN-STDP ( STDP with interactions only between neighbouring pairs of spikes [20] ) or triplet STDP [21] , we again combined these rules with a PI mechanism to make sure that they were stable and had the same rate constraint . For the rule of [23] , we did not add the firing rate constraint , as it already contains a homeostatic mechanism . In all cases , we used the parameter values in the originally published manuscripts; in the case of the triplet rule , we used the data obtained from the fit to visual cortex data . For all in vitro simulations ( except Figure 6C ) , we considered only two neurons with a single connection between them . The parameters used for the learning rules were the same as in the learning applications . The initial synaptic strength of the connection , if not specified elsewhere or varied , was taken to be 2 nS . All parameters had the same values as in the network simulations , but since it is assumed that these in vitro protocols are taking place over a short time scale , the rate constraint mechanism of the model was turned off . For Figure 6C , we considered a group of 20 excitatory and 5 inhibitory synapses , connected onto a single post-synaptic neuron . For each stimulation of the simulated afferent fibers , every synapse had 50% chance of being active . The fibers were stimulated with 100 presynaptic pulses at varying frequencies , as in in vitro experiments [55] . To reproduce the triplet experiment [21] , [52] , we use a stimulation protocol of 60 triplet of spikes repeated at 1 Hz . Each triplet consists of two pre and one post synaptic spikes or two post and one pre-synaptic spikes , as can be seen in the inset of Figure 6D ( see references for more details ) . To reproduce the dependance on frequency [26] , we used a protocol as in the original paper: interdigitated burst of 5 spikes paired with a given and frequency repeated 15 times at a 0 . 1 Hz frequency , thus leading to 75 spikes in total .
|
The circuits of the sensory cortex are able to extract useful information from sensory inputs because of their exquisitely organized synaptic connections . These connections are wired largely through experience-dependent synaptic plasticity . Although many details of both the phenomena and cellular mechanisms of cortical synaptic plasticity are now known , an understanding of the computational principles by which synaptic plasticity wires cortical networks lags far behind this experimental data . In this study , we provide a theoretical framework for cortical plasticity termed the “Convallis rule” . The computational power of this rule is demonstrated by its ability to cause simulated cortical networks to learn representations of real-world speech data . Application of the rule to paradigms used to probe synaptic plasticity in vitro reproduced a large number of experimental findings , and the mathematical form of the rule is consistent with a dual coincidence detector mechanism that has been suggested experimentally in juvenile neocortex . Based on this confluence of normative , phenomenological , and mechanistic evidence , we suggest that the rule may approximate a fundamental computational principle of the neocortex .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
The Convallis Rule for Unsupervised Learning in Cortical Networks
|
Insulin and related peptides play important and conserved functions in growth and metabolism . Although Drosophila has proved useful for the genetic analysis of insulin functions , little is known about the transcription factors and cell lineages involved in insulin production . Within the embryonic central nervous system , the MP2 neuroblast divides once to generate a dMP2 neuron that initially functions as a pioneer , guiding the axons of other later-born embryonic neurons . Later during development , dMP2 neurons in anterior segments undergo apoptosis but their posterior counterparts persist . We show here that surviving posterior dMP2 neurons no longer function in axonal scaffolding but differentiate into neuroendocrine cells that express insulin-like peptide 7 ( Ilp7 ) and innervate the hindgut . We find that the postmitotic transition from pioneer to insulin-producing neuron is a multistep process requiring retrograde bone morphogenetic protein ( BMP ) signalling and four transcription factors: Abdominal-B , Hb9 , Fork Head , and Dimmed . These five inputs contribute in a partially overlapping manner to combinatorial codes for dMP2 apoptosis , survival , and insulinergic differentiation . Ectopic reconstitution of this code is sufficient to activate Ilp7 expression in other postmitotic neurons . These studies reveal striking similarities between the transcription factors regulating insulin expression in insect neurons and mammalian pancreatic β-cells .
Insulin , a key regulator of carbohydrate and lipid metabolism , is secreted by the ß-cells of the pancreas [1] . Deficiencies in insulin signalling underlie type I diabetes and also the more widespread type II diabetes often associated with obesity [2 , 3] . Considerable effort , therefore , has been focused on understanding the development and differentiation of ß-cells [4–6] , resulting in the identification of a number of mammalian regulators of insulinergic identity . However , the precise lineage relationships between progenitor cells and each of the five different postmitotic cell types of the endocrine pancreas , including the ß-cells , remains unclear . This , together with uncertainty about whether some key regulators act in progenitors or postmitotic ß-cells , or in both , obscures the sequence of regulatory events converting pancreatic progenitors into fully differentiated , insulin-expressing ß-cells . Insulin-like peptides ( Ilps ) have been identified in many vertebrate and invertebrate species , suggesting that they evolved prior to a specialised endodermally derived pancreas [7 , 8] . In Drosophila , seven differentially expressed Ilps have been identified as orthologues of mammalian insulins and/or insulin-like growth factors [9] . Four of the Ilps ( Ilp1 , Ilp2 , Ilp3 , Ilp5 ) are coexpressed in a small population of median neurosecretory cells ( mNSCs ) in the brain that share a common progenitor lineage [10] . This provides a source of Ilps that can be secreted into the circulating haemolymph to promote insulin signalling in many different target tissues . Several elegant studies have harnessed the power of Drosophila genetics to demonstrate that insulin signalling plays important functions in regulating organismal growth , cell differentiation , glucose homeostasis , lipid metabolism , lifespan , and fertility [11–16] . Drosophila also promises to provide a powerful system for investigating how insulin production is regulated . Progress in this area , however , remains limited by the lack of information on the key regulators that generate insulin-producing cells from their progenitors . During central nervous system ( CNS ) development , early differentiating neurons extend pioneer tracts that prefigure the major axonal tracts and guide subsequent nerve growth [17–19] . The presence of these pioneer ( or primary ) neuron populations is a conserved feature of CNS development and , remarkably , the early axonal scaffolds of the insect ventral nerve cord ( VNC ) and the vertebrate spinal cord are very similar [17 , 20–22] . The importance of pioneer tracts is well established: targeted ablation or blocked differentiation of pioneer neurons results in defective fasciculation and misrouting of follower axons [23–25] . As pioneer neurons have been observed to undergo apoptosis in the embryo , it has been proposed that once they have established the axonal scaffold , they have no further function [24 , 26 , 27] . In the embryonic Drosophila VNC , however , identified dMP2 pioneer neurons only die in anterior segments , leaving a surviving posterior subpopulation . Cell death in posterior dMP2 neurons is normally prevented by the Hox protein Abdominal-B ( Abd-B ) , which represses the pro-apoptotic genes reaper ( rpr ) and grim [28] . The persistence of some dMP2 neurons raises the possibility that , either there is a longer-lasting requirement for axonal scaffolding functions in posterior than in anterior segments , or that surviving dMP2 neurons perform additional , as yet uncharacterised , roles . Here , we show that surviving dMP2 neurons express the insulin-like peptide Ilp7 at late stages , thus identifying a posterior neural source of insulin distinct from the brain mNSCs . We use axonal tracing and genetic cell ablations to show that dMP2 neurons mature and differentiate into insulin-producing visceral neurons after they are no longer required for axonal scaffolding . Then , making use of the well-defined dMP2 cell lineage and postmitotic mutant rescue experiments , we identify five regulators of the late neuronal programmes for cell survival and Ilp7 expression . This study reveals , for the first time to our knowledge , similarities between the insulinergic genetic programmes of the Drosophila CNS and the mammalian pancreas .
To investigate potential late roles for the surviving posterior dMP2 pioneer neurons , we searched the literature for candidate factors with restricted expression in a single neuron in posterior hemisegments . A previous study noted that Ilp7 has a similar segmentally restricted pattern of mRNA expression in the VNC [9] . After confirming the predicted Ilp7 mRNA sequence by reverse transcriptase-PCR ( see Text S1 for details ) , we generated antibodies recognizing either the A or the B chain of the Ilp7 peptide ( Figure S1D ) . These both show that the spatial expression pattern of Ilp7 peptide is similar to that of Ilp7 RNA ( Figure S1A–S1C ) . Henceforth , the antibody against the Ilp7 B chain is used in this study . Ilp7 peptide is first detected just prior to larval hatching , in late stage 17 embryos with air-filled tracheae ( AFT ) . A dMP2-specific GAL4 driver [28] was used to reveal that Ilp7 is strongly activated at this stage in the single dMP2 neuron in each of the A6-A9 hemisegments ( Figure 1A ) . Thus , in addition to weakly expressing the widely distributed neuropeptide proctolin [28] , dMP2 neurons specifically and strongly express Ilp7 peptide . We hereafter refer to these Ilp7-producing neurons as insulinergic . Once activated , Ilp7 is strongly and stably expressed in postmitotic dMP2 neurons throughout the three larval instars ( L1–L3 ) and also into adulthood ( Figure 1B–1F ) . Ilp7 is also weakly and dynamically expressed in a few other neurons located anterior to A6 ( Figures 1B–1F ) . In contrast to four other Ilps , it is not expressed in the brain mNSCs . However , Ilp7 neurites , likely originating from a single dorsal pair of Ilp7-expressing neurons ( DP , Figure 1E ) , project close to the insulin-expressing mNSCs ( Figure S2 ) . The dMP2 neuron is produced from the MP2 precursor , an atypical neuroblast that divides once and only once to generate a two-cell lineage in which the differential activation of Notch signalling determines sibling identity [29–32] . Combining this previous lineage information with our Ilp7 analysis indicates that the MP2 precursor division in A6–A9 generates an Ilp7+ neuron ( dMP2 ) , in which Numb suppresses Notch signalling , and an Ilp7− Numb− sibling neuron ( vMP2 ) with active Notch signalling . In addition to the neural Ilp7 expression , we observed strong Ilp7 immunoreactivity in the larval and adult hindgut . Costaining with dMP2 drivers indicates that the axons of posterior dMP2 neurons exit the VNC in the posterior nerve at embryonic stage 16 and innervate the hindgut , forming two fascicles that extend on opposite sides of the hindgut in an aboral to oral direction during larval stages ( Figure 1E and 1G ) . Ilp7-positive nerve swellings are apparent throughout the length of these dMP2 fibres , consistent with en passant release sites over the circular muscles of the hindgut . Pre-incubation of Ilp7 antibodies with dissected hindguts prior to fixation reveals extracellular expression of mature Ilp7 peptide at dMP2 terminals ( Figure S1E–S1F' ) . By the adult stage , hindgut Ilp7-positive fibres have become arborized with branches innervating both the anterior intestine and the rectum ( Figure 1H and 1I ) . Besides the CNS and hindgut , we did not detect Ilp7 immunoreactivity in any other larval tissues . Our analysis thus far reveals that , following a segment-specific burst of cell death , surviving posterior dMP2 neurons undergo a transition to become insulin-producing visceral neurons that innervate the larval and adult hindgut ( Figure 1J ) . We first detect Ilp7 expression in surviving posterior dMP2 neurons more than 5 h after their anterior counterparts have undergone apoptosis [23 , 28] . To determine whether Ilp7-expressing dMP2 neurons are retained because they are required for some late pioneer-like role in posterior fascicle maintenance , we used dMP2-GAL4 to express the cell death activators hid and rpr ( Figure 2A and 2B ) . Ablation of posterior dMP2 neurons at stage 17 did not disrupt the integrity of the major Fasciclin II-positive larval longitudinal connectives ( Figure 2E and 2F ) . Nevertheless , as different pioneer neurons can function synergistically in longitudinal tract formation [23] , we next ablated two types of pioneer neuron simultaneously . Using odd-GAL4 [28 , 33] to express hid and rpr , we were able to co-ablate dMP2 and MP1 pioneer neurons during embryonic stages 15–17 ( Figure 2C and 2D ) . Despite loss of all dMP2 and MP1 pioneer neurons from stage 17 onwards , the Fasciclin II-positive longitudinal tracts remained intact at late L1 stages ( Figure 2G ) . Together , the ablation experiments reveal that after posterior dMP2 neurons have fasciculated with later-differentiating neurons to form mature axon bundles and have activated Ilp7 expression , they are no longer required for fascicle maintenance . Therefore , pioneer function and insulinergic activity correspond to distinct early and late steps during the postmitotic maturation of the dMP2 neuron . We then explored whether cell extrinsic factors might regulate the dMP2 insulinergic programme . Expression of the neuropeptide FMRFamide ( FMRFa ) requires a retrograde signal from Glass-bottom boat ( Gbb ) , a bone morphogenetic protein ( BMP ) family growth factor [34 , 35] . This activates the Wishful-thinking ( Wit ) receptor in neurons , leading to the phosphorylation and consequent activation of Mothers against dpp ( Mad ) . Although pMad activation is an absolute requirement for expressing FMRFa , no other Drosophila neuropeptides or neurotransmitters have yet been shown to be regulated in this manner . As Mad is also activated in posterior dMP2 neurons around the time when Ilp7 is first expressed ( Figure 3G ) [28] , we examined Ilp7 expression in BMP pathway mutants . Reduced or absent Ilp7 expression was observed in gbb mutants ( Figure 3B ) , Mad mutants ( Figure S3E and S3F ) , saxophone ( sax ) mutants ( Figure S4 ) , and upon cell-autonomous interference with retrograde axonal transport using the dominant-negative dynactin , P150/Glued ( UAS-GluedDN , Figure 3E ) . In wit mutants , we observed that Ilp7 expression is consistently reduced or absent in late embryos and early L1 larvae ( Figure 3A , 3C , and 3H , p < 0 . 001 ) . Even by the end of L1 , Ilp7 intensity levels remain significantly lower than those of stage-matched wild-type larvae ( Figure 3H , p < 0 . 01 ) . Ilp7 expression is likely to be regulated at the transcriptional level by wit , gbb , and Mad as loss-of-function mutants show altered expression of Ilp7 RNA ( Figure S3A–S3F ) . Cell-autonomous reintroduction of wit in dMP2 neurons rescues Mad activation and Ilp7 expression , thus demonstrating that the insulinergic programme requires BMP signal transduction in a cell-autonomous and postmitotic manner ( Figure 3F and 3H , p < 0 . 001 ) . In contrast to Ilp7 expression , dMP2 early specification and segment-specific survival are not affected in P150/Glued or BMP pathway mutants ( Figures 3A–3E , insets , S5 , and S6 ) . Hence , retrograde BMP signalling does not provide a general dMP2 specification or survival signal but is required , at late postmitotic stages , to promote Ilp7 expression . We next investigated the identity of the cell-intrinsic transcription factors specifying the late postmitotic programme for insulinergic activity . The basic helix-loop-helix factor Dimmed ( Dimm ) , the Drosophila orthologue of vertebrate Mist1 , is an important regulator of neuroendocrine cell differentiation [36] . Dimm is first expressed in dMP2 neurons when apoptosis is underway in anterior segments , several hours before the onset of detectable Ilp7 expression ( Figure 3G ) [37] . Although many aspects of dMP2 differentiation are normal in dimm mutants , including segment-specific death ( Figures 3D , inset , and S5C ) , Ilp7 peptide expression is significantly downregulated ( Figure 3A , 3D , and 3H , p < 0 . 01 ) . Dimm regulation of Ilp7 is not confined to the level of peptide processing as Ilp7 RNA is also downregulated in dimm mutants ( Figure S3G and S3H ) . Ilp7 levels remain abnormally low in dimm mutants throughout development , and this phenotype can be reversed ( to higher Ilp7 levels than those of wild-type larvae ) by postmitotic , cell-autonomous expression of dimm in dMP2 neurons using dMP2-GAL4 ( Figure 3H , p < 0 . 0001 ) . Thus , Dimm expression not only precedes but is also required for high-level Ilp7 production . In Drosophila , the homeodomain transcription factor Hb9 is expressed in a subset of postmitotic motor neurons , where it regulates axonal pathfinding [38 , 39] . Interestingly , embryos lacking the cell death activators hid , grim , and rpr contain increased numbers of Hb9-positive neurons , suggesting that Hb9 might also promote neuronal cell death [40] . Both anterior and posterior dMP2 neurons express Hb9 from embryonic stage 10–11 , soon after they become postmitotic [38] . Hb9 expression in dMP2 neurons is maintained throughout embryogenesis but , by first-instar stage , it is only apparent in dMP2 neurons of A8 and A9 ( Figure 4A and 4B ) . Examination of dMP2 specification in hb9 mutants revealed two distinct phenotypes: 68 . 3% of anterior dMP2 neurons fail to undergo apoptosis ( Figure 4C , 4D , and 4M ) , and 40 . 8% of posterior dMP2 neurons ( as well as 100% of ectopic anterior counterparts ) fail to express Ilp7 ( Figures 4C , 4D , 4M , and S3J ) . Reintroduction of hb9 from embryonic stage 16 onwards using dMP2-GAL4 rescued Ilp7 expression ( 96 . 7% , Figure 4E and 4M ) but failed to rescue anterior apoptosis ( 66 . 7% of anterior dMP2 neurons persist in early first-instar larvae , Figure 4E and 4M ) . This experiment uncouples the dual roles of Hb9 in cell death and Ilp7 expression and indicates that late cell-autonomous activity of Hb9 is sufficient to activate Ilp7 . We then reintroduced Hb9 in dMP2 neurons using odd-GAL4 , which is expressed in dMP2 neurons from embryonic stage 10 onwards , much earlier than dMP2-GAL4 . This efficiently rescued both anterior apoptosis ( 96 . 9% of anterior dMP2 neurons underwent apoptosis by early first-instar ) and Ilp7 expression ( 94 . 1% of posterior dMP2 neurons were Ilp7-positive , Figure 4F and 4M ) . We thus conclude that the pro-apoptotic activity of hb9 is required at an earlier stage of postmitotic dMP2 maturation than its function in Ilp7 activation . To establish whether properties of dMP2 other than apoptosis and Ilp7 expression are also regulated by Hb9 , we examined a number of markers that define different stages of dMP2 specification . Early dMP2 markers such as Odd-skipped ( Odd ) and Fork Head ( Fkh , see next section ) remain expressed in hb9 mutants in both anterior and posterior dMP2 neurons ( Figure 4G , 4H , and 4M ) . At stage 16 , the posterior dMP2-specific marker Abd-B is also retained ( Figure 4I and 4M ) . By stage 17 , dMP2 neurons in hb9 mutants are still able to express Dimm and activated Mad , project axons posteriorly along the most medial fascicle , and exit the VNC to innervate the hindgut correctly ( Figure 4J–4N ) . Thus , Hb9 is not a general postmitotic regulator of all aspects of dMP2 differentiation . Instead , it promotes apoptosis and Ilp7 expression in a cell-autonomous and sequential manner ( Figure S8 ) . We next examined the expression pattern of the winged-helix/forkhead box transcription factor Fkh , which has not been well characterised in the Drosophila CNS . We find that Fkh is expressed in segmentally repeated clusters of midline neurons , including dMP2 , vMP2 , MP1 neurons , and the VUM interneurons ( Figure 5A and unpublished data ) . Within the MP2 lineage , Fkh is first expressed in the MP2 neuroblast at stage 9–10 and continues to be expressed in both the dMP2 and vMP2 daughters throughout embryonic and larval stages ( Figure 5B ) . In fkh mutants , 95% of anterior dMP2 neurons fail to undergo apoptosis , and 95 . 3% of posterior dMP2 neurons ( and 100% of ectopic anterior counterparts ) fail to express Ilp7 ( Figures 5C , 5D , 5L , and S3I ) . Both of these dramatic phenotypes could be rescued to near wild-type levels by reintroducing Fkh under odd-GAL4 regulation , indicating a cell-autonomous requirement for promoting dMP2 apoptosis and Ilp7 expression ( Figure 5E and 5L ) . The Fkh requirement does not extend to all features of dMP2 specification as fkh mutant dMP2 neurons still retain expression of some early dMP2 markers such as Odd , Hb9 , and Abd-B at stage 15–16 ( Figure 5F–5H and 5L ) . However , later events such as Mad activation and Dimm expression are delayed or blocked ( Figure 5I , 5J , and 5L ) . Also , at stage 17 , dMP2 axons often exit the VNC prematurely or , occasionally , cross the midline ( Figure 5K and 5L ) . These phenotypes could all be rescued by reintroducing Fkh in dMP2 neurons in a cell-autonomous manner ( Figure 5L–5N ) . Thus , dMP2 neurons are initially specified with the correct early postmitotic identity in an Hb9- and Fkh-independent manner . Their late differentiation , however , requires inputs from both transcription factors . Removing either factor blocks segment-specific survival and subsequent Ilp7 expression . However , whereas Hb9 specifically affects Ilp7 expression , Fkh acts as a postmitotic progression factor required for most aspects of late dMP2 identity ( Figure S8 ) . Members of the conserved Hox gene family regulate segment-specific aspects of cell fate ( reviewed in [41 , 42] ) . The Hox protein Abd-B is known to promote posterior dMP2 survival [28] . We next addressed whether Abd-B or other Hox proteins might also regulate aspects of dMP2 differentiation . At early stage 17 , prior to the onset of Ilp7 expression , posterior dMP2 neurons in A6–A7 coexpress the Hox proteins Abdominal-A ( Abd-A ) and Abd-B ( and , occasionally in A6 , Ultrabithorax [Ubx] ) , whereas in A8–A9 they express Abd-B only ( Figure 6A ) . To test the involvement of Hox proteins in dMP2 differentation , we generated triple mutants simultaneously lacking the activity of Abd-Bm ( the Abd-B isoform expressed in A6–A8 dMP2s ) and the pro-apoptotic genes rpr and sickle ( skl ) . In these triple mutants , posterior dMP2 neurons survive even though Abd-Bm activity is absent from A6–A8 . In A6 and A7 , dMP2 neurons still express Abd-A , A8 is Hox-free , and A9 retains expression of the Abd-Br isoform ( Figure 6C ) [28] . We examined Ilp7 peptide in these mutants and found it to be expressed in A9 but absent from A6–A8 ( Figure 6D ) . Ilp7 RNA was also absent from A8 and reduced or absent from A6/A7 ( Figure S3K ) . This indicates that Abd-B activity is required for Ilp7 expression and suggests that the other Hox proteins expressed in dMP2 neurons do not possess this function . To test this more directly , we reintroduced several Hox proteins in Abd-B , rpr , skl mutant dMP2 neurons postmitotically and cell-autonomously using dMP2-GAL4 . In this context , Abd-B expression is able to rescue Ilp7 expression ( 87% , Figure 6K ) but neither Abd-A ( Figure 6L ) nor Ubx ( Figure 6M ) are able to substitute for this function . Thus , Abd-B is required cell-autonomously for Ilp7 expression in posterior dMP2 neurons . We then assessed whether Abd-B has a positive input into Ilp7 activation or whether it functions negatively in dMP2 neurons to repress more anterior Hox proteins [43] . We therefore examined the development of dMP2 neurons in A8: the Hox “ground-state” segment in Abd-B , rpr , skl mutants that lacks expression of any Hox genes . dMP2 generation and early specification in A8 occur normally , as revealed by their expression of Odd , Fkh , and Hb9 ( Figure 6E–6G , respectively ) . Nevertheless , acquisition of late identity at stage 17 was partially disrupted as 27% of A8 dMP2 neurons fail to express Dimm ( Figure 6I ) and 37% of them fail to activate BMP signalling ( Figure 6H ) . These low-expressivity effects appear unlikely to account for the 100% absence of Ilp7 as restoration of Dimm and BMP signalling was unable to rescue Ilp7 expression in triple-mutant dMP2 neurons in A8 ( Figure 6N ) . Although we could not assess pathfinding in an A8 segment-specific manner , no ectopic dMP2 projections were observed ( Figure 6J ) . In summary , while early dMP2 specification is largely Hox-independent , Ilp7 expression is under postmitotic Hox control at two levels: Abd-B promotes posterior dMP2 survival and subsequently plays a specific role in promoting Ilp7 expression ( Figure S8 ) . To test the sufficiency of the transcription factors identified in this study , we attempted to reconstitute an ectopic “Ilp7 code” in other postmitotic neurons . To bypass embryonic lethality , we confined our misexpression to a subset of postmitotic neurons . OK371-GAL4 was used to drive expression in about 30 postmitotic motor neurons per hemisegment [44] . We reasoned that , as motor neurons possess active BMP signalling and often express Hb9 [35 , 38 , 39] , misexpression of Dimm , Abd-B , and Fkh in this neuronal type might be sufficient to reconstitute the Ilp7 code described here . When either Abd-B or Fkh were misexpressed in this motor neuron domain , no ectopic Ilp7 activation was observed ( Figure 7A , 7B , and 7E ) . Misexpression of Dimm alone led to occasional , weak activation of Ilp7 in one cell per hemisegment ( Figure 7C and 7E ) . In contrast , combinatorial misexpression of Dimm , Fkh , and Abd-B consistently triggered strong ectopic Ilp7 activation in one neuron per hemisegment in A1–A7 together with weaker ectopic expression in several other neurons ( Figure 7D and 7E ) . This triple misexpression resulted in more Ilp7-positive ectopic cells than in any double misexpression ( Figure 7E , p < 0 . 001 ) , indicating that all three Ilp7 regulators contribute in a combinatorial manner to activate Ilp7 . Interestingly , combinatorial misexpression of Dimm , Fkh , and Abd-B in motor neurons did not lead to any detectable ectopic expression of Ilp2 , an insulin-like peptide known to be expressed in the brain mNSCs ( Figure 7F ) , nor did we observe ectopic activation of other neuropeptides/peptide hormones such as FMRFa ( Figure S7 ) . Thus , the insulinergic code that we have identified appears to be specific for Ilp7 expression .
The observed death of some Drosophila pioneer neurons has been used to argue that their function is transient [24 , 27] , but persistence in other cases suggested that , either they continue to play an axonal-scaffolding role , or that they adopt some other identity [27 , 45] . Our findings resolve this long-standing issue by clearly demonstrating that , for dMP2 neurons , the axonal scaffolding function is only transient . After this role is no longer required , surviving dMP2 neurons become insulinergic and innervate the hindgut . The other known innervation of the Drosophila gut occurs much more anteriorly , in the foregut and anterior midgut , from neuronal cell bodies located in the peripheral ganglia of the stomatogastric nervous system [46] . Unlike dMP2 neurons , however , the individual identities of the stomatogastric neurons and their cell lineages remain to be clearly defined . Thus , dMP2 neurons may provide a simple and well-characterised system for studies of the guidance cues involved in enteric innervation . Future studies , however , will be needed to determine the functions of Ilp7 in dMP2 neurons . It will be important to distinguish if this posterior neural source of insulin acts humorally to promote growth , like the more anterior brain mNSCs [15 , 16] , or if it has more local effects in abdominal tissues . In this regard , the presence of Ilp7-expressing neurites in close proximity to the Ilp2-producing mNSCs is intriguing . The transition from pioneer to neuroendocrine neuron is not unique to dMP2 neurons , as Drosophila MP1 pioneer neurons also become neuropeptidergic at larval stages [47] . In the grasshopper , segment-specific survival of pioneer neurons has also been reported [45] , raising the possibility that they too may become neuroendocrine . Studies in other species , including vertebrates , will be needed to reveal the extent to which the linkage between pioneer and neuroendocrine functions is conserved . Identifying pioneer neurons with an “ancestral” neuroendocrine identity [48] in other phyla would lend further support to the proposal that pioneer neurons are highly conserved in evolution [17 , 20 , 49] . Apoptosis of postmitotic neurons is a widespread feature of normal VNC development [28 , 40] , but few developmental regulators of core pro-apoptotic genes such as grim , hid , and rpr have been identified . This study uncovers roles for Fkh and Hb9 . Hb9 , at least , appears linked to cell death in neurons other than dMP2: in Df ( 3L ) H99 mutant embryos , where apoptosis is blocked , ectopic Hb9-positive RP motor neurons are observed in segments A7–A8 [40] . Hb9 is an important regulator of motor neuron identity in both Drosophila and vertebrates [50 , 51] . Our finding of a pro-apoptotic function for Hb9 in Drosophila , together with the neurotrophic requirement for motor neuron survival in vertebrates , raises the possibility that the same genetic programmes specifying the identities of motor neurons also sensitise them for postmitotic editing via apoptosis . Hb9 and Fkh expression in many neurons that do not die suggests a combinatorial mechanism for the control of developmental apoptosis . One possibility is that several transcription factors function in combination to activate the core pro-apoptotic genes . Given the proposed role for Foxa proteins in chromatin accessibility [52] , Fkh expression in dMP2 neurons may render the promoters of core pro-apoptotic genes responsive to activation by Hb9 . An alternative but not mutually exclusive mechanism involves individual transcription factors activating different pro-apoptotic genes such that a combination of these would then be required to trigger neuronal death . For example , Hb9 could be required for rpr/skl but not grim expression . Some support for this idea comes from the observation that loss of hb9 activity blocks rpr/skl-mediated death of dMP2 neurons but not the largely grim-dependent apoptosis of anterior MP1 neurons ( unpublished data ) . An important conclusion from this study is that the combinatorial transcription factor code controlling apoptosis partially overlaps with that regulating insulinergic identity ( see next section ) . Thus , Fkh and Hb9 are both essential components of the codes for anterior apoptosis and also Ilp7 expression , illustrating that these transcription factors play surprising dual roles as pro-apoptotic and pro-differentiation factors within the same neuronal subtype . Importantly , our results also show that the segment-specific Hox protein Abd-B acts as a postmitotic switch , converting the pro-apoptotic Fkh+ Hb9+ code into an insulinergic Fkh+ Hb9+ Abd-B+ code . Three Ilp7 regulators ( Hb9 , Abd-B , and Fkh ) are expressed at least 12 h before Ilp7 is first activated: from the time when the MP2 neuroblast exits the cell cycle . In the case of Hb9 , we were able to uncouple two temporally separable functions . Early postmitotic expression of Hb9 is important for its death-activating function , whereas later expression suffices for activating Ilp7 . Similarly , the Hox protein Abd-B generates a segment-specific neuropeptide pattern via postmitotic regulation of posterior dMP2 survival and also Ilp7 activation . As vertebrate neuropeptides are also expressed in restricted neuronal populations within specific rostrocaudal domains [53 , 54] , they may be similarly regulated by Hox survival/neuroendocrine inputs . In the case of Fkh , it is required for many different aspects of the progression from the early to the late postmitotic dMP2 fate . Fkh expression is restricted to VNC midline neurons and its vertebrate orthologue Foxa2 functions in the differentiation of the floor plate and ventral dopaminergic and serotonergic neurons [55–57] . Thus , in both the Drosophila midline and its vertebrate counterpart , the floor plate [20 , 58] , Fkh proteins play a conserved role in the differentiation of ventral neuronal subtypes . The other two dMP2 regulators identified in this study , Dimm and the BMP pathway , are switched on shortly before the onset of Ilp7 expression . The timing of onset of these two broad neuroendocrine regulators is likely to specify when Ilp7 is first activated , whereas the earlier factors Fkh , Hb9 , and Abd-B may contribute more specifically to insulinergic identity . Together , the genetic and expression analyses in this study demonstrate that the combinatorial code of genetic inputs required for Ilp7 expression is assembled in a step-wise manner during postmitotic maturation ( Figure 8 ) . Importantly , this allows a subset of the components to be shared ( such as Fkh and Hb9 ) between sequential neuronal programmes ( survival and Ilp7 expression ) without losing output specificity . Two observations from this study indicate that insulinergic combinatorial codes can vary from cell-to-cell and also from one Ilp to another . First , none of the regulators of Ilp7 in dMP2 neurons appear to regulate it in DP neurons . Second , the dMP2 insulinergic code is sufficient to trigger ectopic expression of Ilp7 but not Ilp2 or other neuropeptides such as FMRFa . These findings suggest the existence of additional , as yet unidentified , insulinergic factors in DP neurons and also in the brain mNSCs where Ilp2 is expressed . Recent identification of the neural progenitor for these mNSCs [10] should facilitate characterization of the Ilp1/Ilp2/Ilp3/Ilp5 combinatorial codes and thus clarify the extent to which different insulinergic transcriptional programmes overlap . Our finding that an Ilp7-expressing neuron derives from the MP2 lineage reveals that at least some insulinergic regulators are similar in insects and mammals . Three apparent similarities may not be very insulin-specific but reflect more general processes shared by neural and endocrine programmes in many species . First , Notch signalling singles out the MP2 neuroblast and distinguishes its two progeny neurons , while in mammals , it limits pancreatic expression of the “proneural” gene Ngn3 to prospective endocrine cells [6] . Second , the survival and pro-Ilp7 functions mediated by Abd-B in the dMP2 neuron could also have their postmitotic counterparts in ß-cells , either mediated by related Hox genes [59] or via another homeobox gene , Pdx-1 , following its early input into pancreatic induction [6 , 60] . And third , Nerfin-1 is required for dMP2 pioneer function [61] , while its mammalian orthologue Insm1/IA1 is important for pancreatic ß-cell specification [62] . Several more specific regulatory similarities exist between the insulinergic differentiation factors active in postmitotic dMP2 neurons ( Figure 8 ) and pancreatic ß-cells [4 , 6] . For example , the role of fkh in dMP2 neurosecretory differentiation described in this study is similar to the functions of HNF3b/Foxa2 in islet maturation and insulin secretion [63 , 64] . In addition , mammalian Nkx2 . 2 is important for pancreatic ß-cell specification [65] and is known to activate transcription of the insulin regulator Nkx6 . 1: an important late event in ß-cell differentiation [6] . Intriguingly , the Drosophila orthologue of Nkx2 . 2 , Vnd , is required for dMP2 formation [61 , 66] . Drosophila Nkx6 . 1 , the orthologue of mammalian Nkx6 ( FlyBase name HGTX ) , is expressed by postmitotic dMP2 neurons [67] , and it will be interesting to determine whether it too functions downstream of Vnd during Ilp7 regulation . Most strikingly , mammalian equivalents of two of the insulinergic inputs identified in this study , Hb9 and BMP signalling , are also required for several aspects of late ß-cell differentiation including the expression of Nkx6 . 1 and insulin [68–70] . Together , these insect-mammalian comparisons provide evidence that , although the cell types involved look very different , some of the genetic circuitry regulating insulin is conserved between arthropods and chordates . This suggests that the power of fly genetics can now be harnessed to identify additional mammalian regulators of neuroendocrine cell fates and insulin expression .
The following fly stocks were used: dMP2-GAL4 , UAS-nls-myc-EGFP ( referred to as UAS-nmEGFP ) , Abd-BM5 , UAS-Antennapedia , UAS-Ubx , UAS-abd-A , UAS-Abd-Bm ( referred to as UAS-Abd-B ) , Df ( 3L ) XR38 and Df ( 3L ) H99 deficiencies ( referred to as XR38 and H99 , respectively ) [28]; witA12 , witB11 , UAS-tkvA , UAS-saxA , UAS-GluedDN , gbb1 [34]; dimmrev4 , dimmP1 , UAS-dimm [36]; hb9KK30 , hb9JJ154 [38]; UAS-hb9 [39]; fkh6 [71]; Mad10 , Mad12 [72]; CY27-GAL4 [73]; UAS-fkh [74]; oddGAL4 , UAS-GFP [33]; UAS-CD8-GFP [75]; UAS-hid , UAS-rpr [76]; OK371-GAL4 [44]; and Ilp2-GAL4 [16] . Mutants were maintained over CyO , Act-GFP or TM3 , Ser , Act-GFP balancer chromosomes . y w or w1118 or heterozygous mutant stocks were typically used as control . Crosses were maintained at 25 °C , except for OK371 crosses , which were kept at 21 °C . Immunolabelling was carried out as previously described [34] . Antibodies used were: rabbit α-Ilp7 ( 1:5 , 000 , see below for details ) , rabbit α-Odd ( 1:1 , 000 ) [31] , guinea pig α-Fkh ( 1:800 ) [77] , rabbit α-Hb9 ( 1:2 , 000 ) [38] , guinea pig α-Dimm ( 1:1 , 000 ) [78] , rabbit α-pMad ( 1:2 , 000 ) [79] , mouse α-Ubx mAb FP3 . 38 ( 1:20 ) [80] , rat α-abd-A ( 1:400 ) [81] , rat α-Ilp2 ( gift of P . Leopold , 1:400 ) , mouse and rabbit α-green fluorescent protein ( GFP ) ( Molecular Probes , 1:1 , 000 and 1:2 , 000 , respectively ) , rat α-mouse CD8a ( Caltag , 1:100 ) , mouse mAb 1D4 α-fascilin II ( 1:50 ) , mouse α-Antennapedia mAb 4C3 ( 1:50 ) , mouse α-Abd-B mAb 1A2E9 ( 1:20 ) ( all from Developmental Studies Hybridoma Bank ) . Phalloidin-TX ( Molecular Probes ) was used at 1:1 , 000 . FITC- , Rhodamine-Red-X- , and Cy5-conjugated secondary antibodies were obtained from Jackson Immunolabs and used at 1:200 ( 1:100 for the Cy5-conjugated antibody ) , and a Vectastain ABC kit was used for the detection of biotinylated anti-mouse antibody ( Vector Laboratories , 1:1 , 000 ) . An antibody to Ilp7 was generated in rabbits against the synthetic peptide RERSQSDWENVWHQET . This sequence is within the predicted B chain and corresponds to amino acids 41–57 of the predicted Ilp7 protein . Peptide synthesis , antibody generation , and affinity purification were done by Eurogentec . Details of a second antibody directed against the predicted A chain are available in Text S1 . Confocal images were obtained using a Leica SP5 confocal microscope . Where immunolabels were analysed for levels of expression , wild-type and mutant tissues were stained on the same slide . The intensity index used to quantify Ilp7 expression levels in dimm , wit , gbb mutants , and rescues ( Figure 3H ) was obtained as previously described [36] . Briefly , mutant and wild-type CNS were dissected , fixed , and processed together on polyLysine slides . They were imaged with exposure settings adjusted to optimize detection without saturating the signal , and identical settings were used for all preparations and genotypes . Mean pixel luminosity for the area covering the soma ( S ) was measured for each neuron using Adobe Photoshop . An adjacent area was sampled to measure the background signal ( B ) . The intensity index was calculated as ( S-B ) /B . Statistical significance was calculated using two-tailed Student t-tests , assuming equal sample variance . Where appropriate , images were false-coloured for clarity , and red was converted to magenta for the benefit of colour-blind readers .
|
Genetic studies using invertebrate model organisms such as Drosophila have provided many new insights into the functions of insulin and related peptides . It has , however , been more difficult to use Drosophila to study the regulation of insulin , at least in part because the relevant insulinergic cell lineages were not well characterised . Here , we have identified a cell lineage that generates a single Drosophila insulin-producing neuron . This neuron first functions as a pioneer , guiding the axons of other neurons within the central nervous system of the embryo . It then develops long axons that exit the central nervous system to innervate the gut and also begins to express an insulin-like peptide . Genetic analysis identifies four transcription factors and one extrinsic signal that instruct the pioneer neuron to become an insulin-producing neuron . The analysis also reveals similarities between the genetic programmes specifying insulin production by Drosophila neurons and mammalian pancreatic ß-cells . This suggests that Drosophila may , in the future , prove a useful model system for identifying new regulators of human insulin production .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology",
"diabetes",
"and",
"endocrinology",
"neuroscience"
] |
2008
|
Postmitotic Specification of Drosophila Insulinergic Neurons from Pioneer Neurons
|
In 2010 , 2012 , 2013 and 2014 dengue outbreaks have been reported in Dar es Salaam , Tanzania . However , there is no comprehensive data on the risk of transmission of dengue in the country . The objective of this study was to assess the risk of transmission of dengue in Dar es Salaam during the 2014 epidemic . This cross-sectional study was conducted in Dar es Salaam , Tanzania during the dengue outbreak of 2014 . The study involved Ilala , Kinondoni and Temeke districts . Adult mosquitoes were collected using carbon dioxide-propane powered Mosquito Magnet Liberty Plus traps . In each household compound , water-holding containers were examined for mosquito larvae and pupae . Dengue virus infection of mosquitoes was determined using real-time reverse transcription polymerase chain reaction ( qRT-PCR ) . Partial amplification and sequencing of dengue virus genome in infected mosquitoes was performed . A total of 1 , 000 adult mosquitoes were collected . Over half ( 59 . 9% ) of the adult mosquitoes were collected in Kinondoni . Aedes aegypti accounted for 17 . 2% of the mosquitoes of which 90 . 6% were from Kinondoni . Of a total of 796 houses inspected , 38 . 3% had water-holding containers in their premises . Kinondoni had the largest proportion of water-holding containers ( 57 . 7% ) , followed by Temeke ( 31 . 4% ) and Ilala ( 23 . 4% ) . The most common breeding containers for the Aedes mosquitoes were discarded plastic containers and tires . High Aedes infestation indices were observed for all districts and sites , with a house index of 18 . 1% in Ilala , 25 . 5% in Temeke and 35 . 3% in Kinondoni . The respective container indices were 77 . 4% , 65 . 2% and 80 . 2% . Of the reared larvae and pupae , 5 , 250 adult mosquitoes emerged , of which 61 . 9% were Ae . aegypti . Overall , 27 ( 8 . 18 ) of the 330 pools of Ae . aegypti were positive for dengue virus . On average , the overall maximum likelihood estimate ( MLE ) indicates pooled infection rate of 8 . 49 per 1 , 000 mosquitoes ( 95%CI = 5 . 72–12 . 16 ) . There was no significant difference in pooled infection rates between the districts . Dengue viruses in the tested mosquitoes clustered into serotype 2 cosmopolitan genotype . Ae . aegypti is the main vector of dengue in Dar es Salaam and breeds mainly in medium size plastic containers and tires . The Aedes house indices were high , indicating that the three districts were at high risk of dengue transmission . The 2014 dengue outbreak was caused by Dengue virus serotype 2 . The high mosquito larval and pupal indices in the area require intensification of vector surveillance along with source reduction and health education .
Dengue is one of the most important mosquito-borne viral diseases in the tropics and subtropics . Although the exact global burden of dengue cases is unknown , recent estimates indicate that 390 million dengue infections occur annually . Of these , 96 million cases manifests clinically [1] with about half a million cases of dengue haemorrhagic fever requiring hospitalization [2] . Dengue is prevalent in Africa , though rarely reported . Statistics indicate that the disease has increased dramatically since 1980 , with epidemics occurring in both eastern and western Africa [3–5] . The World Health Organization statistics indicate that 2 . 4% of the global dengue haemorrhagic fever cases occur in Africa and one-fifth of the population in the continent is at risk [6] . Although very little is known of the disease in Tanzania , dengue has an historical relationship with this East African country . The name dengue is believed to have been derived Kiswahili words “ki denga pepo” , meaning cramp-like seizure caused by an evil spirit . This condition was first described by Spanish sailors visiting the southern coast of Tanzania during the 15th Century [4 , 7 , 8] . In the 1823 and 1870 , dengue was reported on Zanzibar Islands [3] . Since then , three epidemiological surveys have indicated that different areas of Tanzania including Zanzibar , Iringa and Mbeya have reported varying prevalence of antibodies against dengue virus ( DENV ) in humans [9–11] . In 2010 , 2012 , 2013 and 2014 , dengue outbreaks were reported in Dar es Salaam , with the worst outbreak in 2014 ( Ministry of Health & Social Welfare , unpubl . ) . From January 2014 until end of May 2014 , the cumulative number of confirmed and suspected dengue cases was 961 and 1 , 969 , respectively ( Ministry of Health and Social Welfare , unpubl . ;http://promedmail . chip . org/pipermail/promed-eafr/2014-May/001515 . html ) . The most important vectors of dengue world-wide are Aedes aegypti and Ae . albopictus . Of the two mosquito species , Ae . aegypti is the most important vector because of being highly domesticated , strongly anthropophilic , a nervous feeder and a discordant species [12] . Despite the recent dengue outbreaks in Tanzania , there is no comprehensive data on the risk of transmission in the country . This study was therefore carried out to assess the risk of transmission of dengue in Dar es Salaam during an outbreak in 2014 . Specifically , the study aimed to determine the ( i ) distribution and abundance of dengue vectors in Dar es Salaam; ( ii ) pattern of Ae . aegypti infestation and container productivity; and ( iii ) dengue virus infection in the mosquito vectors .
This study was carried out in Dar es Salaam region in eastern Tanzania and involved Kinondoni , Ilala and Temeke Districts ( Fig 1 ) . Dar es Salaam is the largest commercial City in Tanzania with an area of 1 , 339 km2 . The region has a population of 4 , 364 , 541 ( Ilala = 1 , 220 , 611; Kinondoni = 1 , 775 , 049; Temeke = 1 , 368 , 881 ) , with 3 , 313 people per square kilometre and an annual growth rate of 5 . 6% [13] . The climate of Dar es Salaam is generally hot and humid with small seasonal and daily variations in temperature . The mean daily temperature is about 26°C , with the lowest and highest temperatures in July-August and February-March , respectively . Dar es Salaam has two dry and two rainy seasons . The main dry season lasts from June to September . The rainy seasons are from October-December and March-May . The total annual rainfall averages 1100 mm . Relative humidity is generally high , reaching 100% almost every night throughout the year , but falling to 60% during the day [14] . This cross-sectional study involved three sites ( wards ) in each district . The sites were selected based on the ecological and demographic characteristics . In Kinondoni the study sites were Msasani , Sinza and Kwembe . In Ilala , the sites were Kivukoni , Jangwani and Tabata . In Temeke the sites were Kigamboni , Miburani and Chamazi . Msasani , Kivukoni , and Kigamboni are high-income areas located along the Indian Ocean and characterized by sparsely populated neighbourhood . The sites have regular blocks with high standard dwellings , high vegetation coverage , piped water supply and regular garbage collection . The houses generally have large and shaded peridomestic environments . Sinza , Jangwani and Miburani are middle-class residential areas , with high density population and almost at the central location in the respective districts . The areas are characterized by poor garbage storage , collection and disposal . In these areas , shortage of piped water is common . Kwembe , Tabata and Chamazi are rural and located at the peripheral areas of the respective districts and characterized low population density . The survey was conducted during the dengue outbreak in May-June 2014 . A survey frame of households was obtained using a simple random sampling ( without replacement ) technique that was applied to select 100 households for the study in an identified block within each study site . A hand-held Geographical Positioning Systems ( Magellan Meridian ) was used to determine the geo-coordinates of the study sites . Adult mosquitoes were collected using carbon dioxide-propane powered Mosquito Magnet Liberty Plus traps ( American Biophysics Corporation , Rhode Island , USA ) , fixed at nine sentinel sites ( three sites in each district ) . Nine traps were set in each site and allowed to operate for three days consecutively . The traps were set in the morning and operated throughout the day and night . The mosquito catches were collected the following day and stored for further analysis . Each household compound was examined for water holding containers and the presence of larvae and pupae of Aedes mosquitoes . Upon the entrance of the household premise , the yard and the interior of the household were thoroughly inspected for water-holding containers . All water-holding containers found were examined for presence of mosquito larvae and pupae . In each household compound , information on type of container , water type/state , approximate container volume , and water volume within the container and the number of larvae and pupae was recorded . All larvae and pupae were collected using large-mouthed pipettes . In large containers a standard plastic dipper was used for larvae and pupae collection . Larvae and pupae were transported in lidded plastic containers and reared to adult stage in an insectary at Muhimbili University of Health Sciences . Pupae were kept in small water-filled plastic vials and placed into a netted cage until emergence of adults for accurate species identification . Adult mosquitoes were identified to genus or species using morphological identification keys [15] . The hatched mosquitoes were immediately killed by freezing and stored in liquid nitrogen for later screening of dengue virus . The data collected were entered in Epidata Version 3 . 1 software to develop a database which was later migrated to STATA ( Stata Corps , 2007 ) for further analysis . Data were later cleaned and verified for quality by assessing a 10% of random selection of original forms that were compared with the entered data . From the database variables of interest that were calculated include the numbers of containers , infested containers , larvae and pupae collected . This was done for each type of container , and of infested containers found for every category of the variables measured . For populations larvae or pupae , three main methods were used to assess dengue transmission levels: ( i ) House index as the percentage of houses infested with larvae and / or pupae; ( ii ) Container index as the percentage of water-holding containers infested with larvae and/or pupae; and ( iii ) Breteaux index calculated as the number of positive containers per 100 houses inspected in a specific location . For adult mosquitoes , the proportion of Aedes mosquito out of those collected was assessed and all mosquitoes ( including those hatched from pupae/larvae ) were preserved in liquid nitrogen and later screened for dengue virus using qRT-PCR . Maps showing spatial variation of important indicators for dengue were produced . Infection rate was analysed using a PooledInfRate software version 4 . 0 to compute infection rates from pooled data [19] . The study protocol was approved by the Tanzania Medical Research Coordinating Committee of the National Institute for Medical Research . Permission to conduct the study was sought from all administrative levels including regions , districts and wards . Verbal informed consent was obtained from the heads of households before the entrance and inspection of the house premises and/or installation of mosquito traps .
A total of 1 , 000 adult mosquitoes were collected using Mosquito Magnet Liberty Plus traps . The largest proportion of adult mosquitoes ( all species ) was collected from Kinondoni ( 59 . 9% ) . Ae . aegypti accounted for 17 . 0% of the total adult mosquitoes collected; with the largest proportion ( N = 154; 90 . 6% ) been collected from Kinondoni ( Table 1 ) . A total of 796 houses were inspected for presence of potential Aedes mosquito breeding sites . Of these , 305 ( 38 . 3% ) houses had water-holding containers in their premises . Kinondoni had the largest proportion ( 52 . 8% ) of households with water-holding containers , followed by Temeke ( 29 . 8% ) and Ilala ( 17 . 4% ) . A total of 219 ( 71 . 8% ) of the households had at least one water-holding container with mosquito larvae and/or pupae . Water holding containers in 140 houses were found to harbour mosquito pupae . The most common breeding containers for the Aedes mosquitoes were medium-sized plastic containers and tires ( Table 2 ) . In terms of productivity , tires , plastic containers and flower pots were the most productive containers for both larvae and pupae in Dar es Salaam ( Fig 2 ) . Most of the flower pots ( 71 . 4% ) were found in Kinondoni , followed by Temeke ( 21 . 4% ) and Ilala ( 7 . 1% ) . Msasani and Sinza were the areas with the largest proportion of flower pots infested with larvae and/or pupae . The overall house larvae , container and Breteaux indices for Dar es Salaam were 27 . 5% , 71 . 8% and 8 . 2 , respectively . House Index ( HI ) was highest in Kinondoni ( 35 . 3% ) followed by Temeke ( 25 . 5% ) and Ilala ( 18 . 1% ) . The Container Index ( CI ) was highest in Temeke ( 80 . 2% ) followed by Ilala ( 77 . 4% ) and Kinondoni ( 65 . 2% ) . This indicates that , though HI was higher in Kinondoni , the CI was higher in Temeke . Breteaux index was highest in Ilala ( 30 . 6 ) , followed by Temeke ( 25 . 3 ) and was lowest in Kinondoni ( 20 . 8 ) ( Table 3 ) . Similarly , the highest pupal index was found in Temeke ( 52 . 9% ) , followed by Ilala ( 49 . 1% ) and Kinondoni ( 41 . 0% ) . Analysis by site showed that Kivukoni had the highest CI while Sinza had the highest HI . The lowest HIs were observed in Chamazi , Jangwani , and Tabata . BIs were highest in Kivukoni and Tabata and lowest in Jangwani and Kwembe ( Table 3 ) . Of the hatched 5 , 250 adult mosquitoes , 3 , 250 Ae . aegypti , 800 Ae . simpsoni and 1 , 200 Culex quinquefasciatus were preserved in liquid nitrogen for later screening of dengue virus . A total of 368 mosquito pools , each containing up to 10 Ae . aegypti were processed to extract RNA . Of these , 330 pools were subjected to qRT-PCR for dengue virus detection . Overall , 27 ( 8 . 18 ) of the 330 pools of Ae . aegypti were positive for dengue virus . Of the 27 positive pools , 25 were from larval collections and two were from adult mosquito collected from Sinza . On average , the overall maximum likelihood estimate indicates pooled infection rate of 8 . 49 per 1 , 000 mosquitoes ( 95%CI = 5 . 72–12 . 16 ) . There was no significant difference in pooled infection rates between the districts ( Table 4 ) . None of the mosquitoes collected in Kwembe and Chamazi was infected with dengue virus . Dengue viruses in the tested mosquitoes clustered into serotype 2 cosmopolitan genotype . Partial nucleotide sequence of the dengue virus genome ( CprM region ) in mosquitoes collected were highly identical to and clustered with serotype 2 cosmopolitan genotype dengue viruses collected in Asia ( Fig 3 ) . These serotype 2 Asian dengue viruses were collected from human patients in 2007 , 2008 and 2010 in Makassar ( Indonesia ) , 2010 and 2014 in Guangdong ( China ) and 2005 and 2009 in Singapore .
In this study , Aedes aegypti accounted for about one fifth of the man-biting mosquitoes caught in Dar es Salaam . These findings provide the most up-date of data on Ae . aegypti in the region since 1980s . Over one-third of the inspected house premises had water-holding containers positive for larvae or pupae of Aedes mosquitoes . The most common breeding containers for Ae . aegypti in Dar es Salaam were discarded plastic containers and tires . Kinondoni district accounted for over half of the Aedes positive water holding containers . Most of these water holding containers are man-made , with only a few natural breeding sites identified . Previous studies in Tanzania have shown that in most places Ae . aegypti breeds in both artificial and natural sites [14 , 20 , 21] . Similar to our findings , a study by Trpis [14] reported that tires , wrecked motor cars , water-pots , coconut shells , snail shells and tins were the most productive containers in some areas of the City during the early 1970s . It has already been shown that a part of the Ae . aegypti population in East Africa is maintained in some biotypes such as automobile dumps and coral rock holes by continuous breeding [14 , 20] . Discarded tires and animal watering pans have been reported to be the two most common larval breeding sites elsewhere [22] . Ae . aegypti has been reported as a common outdoor breeding mosquito even in rural areas of Tanzania . In the rural areas of Kilombero and Kilosa districts , Trpis [23] observed that more than a quarter of the water containers outside houses were harbouring Ae . aegypti larvae , while there was no breeding in containers indoors . In the same areas , clay pots , buckets , tins , drums used for storing water and tree holes were the common breeding sites . Similar to our findings , breeding of Ae . aegypti in banana flower bracts has been reported in southern-east Tanzania [23] . This study indicated that medium-sized containers and tires that are often susceptible to the rainfall regime are the most productive containers for Ae . aegypti . Similar findings have been observed in Trinidad [24] and Brazil [25] . In Brazil , the four most productive containers were found to generate up to 90% of total pupae . It has been reported that container productivity varies according to seasons and urbanization degree [25] . In this study container productivity varied by site and district and according to urbanization degree . The findings that the peri-urban areas of Dar es Salaam had low larval infestation levels suggest a low probability of dengue transmission in the areas . In a concurrent study the incidence of dengue was found to be relatively lower in the peri-urban areas ( ( F . Vairo et al . , unpubl . ) . The overall house and container index for Dar es Salaam was 27 . 5% and 71 . 8% , respectively . Previous studies elsewhere in Tanzania , have reported relatively lower larvae indices than those observed in our study [20] . In this current study , the higher Aedes mosquito HI , CI and BI were not surprising as the study was carried out during an epidemic that had already lasted for five months . The Breteaux indices in the three districts of Dar es Salaam ( 20 . 8–30 . 6% ) were high . Higher Aedes indices , BI in particular , provides indication of geographical areas at high risk for dengue transmission [26] . Similar higher HIs have been reported in Singapore [27] . In Brazil , high infestation indices have been observed for both urban and sub-urban localities [28] . In a study in northern Ghana , similar higher HI , CI and BI have been reported in a recent study; with generally higher infestation rates during the dry season [29] . The spatial heterogeneities in Ae . aegypti larval density , HI , CI and BI found in this study have also been reported elsewhere in the world [30–34] . Identification of such heterogeneity is important in identifying areas for intervention . The infection rates found in this study are relatively higher compared to many other reported elsewhere . The mosquito infection rates with dengue virus are of the order of 1% even in areas where transmission is ongoing [35–38] . Similar higher infection rates have been reported in a study in Merida , Mexico [39] and Odisha in India [40] . Although the mosquito infection rate in the current study was highest in Ilala followed by Temeke and Kinondoni , the incidence of dengue cases during the outbreak was highest in Kinondoni followed by Temeke and Ilala ( Ministry of Health and Social Welfare , unpubl . ) . This is not surprising for the fact that , such a scenario is likely to be determined by dispersal of vector [37 , 41] which itself can vary over time [42] , and is influenced by house density [43] and by human movement within and beyond the infection cluster [44] . In this study , the largest adult mosquito density per trap was observed in Kinondoni . Moreover , Human movements have been described to potentially confound dengue vector data that derive from residential areas alone as increasingly , evidence indicates that only a proportion of dengue infection are acquired in the individual’s own home , with the majority resulting from bites by virus-infected mosquitoes at their schools , workplaces or numerous locations far from the home [35 , 39 , 44 , 45] . In a recent study in rural Thailand [37] , it was found that human and mosquito infections are positively associated with each other at small geographic and temporal scales . Nucleotide sequencing and phylogenetic analysis of the CprM region of dengue virus found in mosquitoes in Dar es Salaam had highest identity and clustered with serotype 2 cosmopolitan genotype . Highest identity was with dengue serotype 2 strains collected from China , Indonesia and Singapore . These results indicate the widespread nature of dengue viruses and possible link to travel between Asia and Tanzania [46] . The recent outbreak of Dengue in Dar es Salaam occurred during the long rainy season ( March-June , 2014 ) . The high indices of Ae . aegypti in Dar es Salaam during the long rainy season were likely to be responsible for dengue outbreak . In a previous study of Ae . aegypti larval population in Dar es Salaam the density of the mosquito was high in April because of heavy rains and declined towards the end of June and remained at a very low level until the beginning of August [14] . This trend correlated with the decline of dengue cases observed in a parallel study ( F . Vairo et al . , unpubl . ) . However , a few studies have also shown dengue occurrence during the dry season . It is likely that in coastal areas of East Africa there are always some short scattered showers during the dry season that prevent these foci from completely drying out [20] . In conclusion , findings of this study indicate that the 2014 dengue outbreak in Dar es Salaam was caused by Dengue virus serotype 2 and most likely a result of the introduction from south-east Asia . Ae . aegypti is the only vector of Dengue in the region and breeds mostly in medium-sized plastic containers and tires . The high overall and site specific house indices and mosquito infection rates indicate that all three districts were at high risk of dengue transmission . With the high larval and pupal indices observed in the area there is need to intensify vector surveillance activities along with source reduction and public health education . The introduction of detailed systematic vector surveillance in Dar es Salaam and elsewhere in Tanzania , before , during , and after any dengue epidemic will offer an opportunity to analyse entomological information at different geographic units .
|
Until 2010 , little was known about Dengue in Tanzania . Since then , four outbreaks have been reported in Dar es Salaam City . This study was therefore carried out to assess the risk of transmission of dengue in Dar es Salaam during an outbreak in 2014 . In this study adult mosquitoes were collected using carbon dioxide-propane powered traps . In addition , household compounds were visited and all water-holding containers examined for presence of mosquito larvae and pupae . Mosquito virus infection was determined using real-time reverse transcription polymerase chain reaction ( qRT-PCR ) . Of the total of 1 , 000 adult mosquitoes collected , Aedes aegypti accounted for 17 . 2% . A total of 796 houses were inspected and 38 . 3% had water-holding containers in their premises . The most common breeding containers for the Aedes mosquitoes were discarded plastic containers and tires . High Aedes infestation indices were observed for all districts and sites , with a house and container indices ranging from 18 . 1–25 . 5% and 65 . 2–80 . 2% , respectively . The Breteaux indices were 30 . 6 , 20 . 8 and 25 . 3 in Ilala , Kinondoni and Temeke , respectively . An overall 8 . 18% of mosquito pools were infected with dengue virus serotype 2 . The overall maximum likelihood estimate of pooled infection rate of 8 . 49 per 1 , 000 mosquitoes was observed . This information is useful for the design of appropriate vector surveillance and control strategies in the City of Dar es Salaam .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2016
|
The Risk of Dengue Virus Transmission in Dar es Salaam, Tanzania during an Epidemic Period of 2014
|
The chromatin remodeler BRAHMA ( BRM ) is a Trithorax Group ( TrxG ) protein that antagonizes the functions of Polycomb Group ( PcG ) proteins in fly and mammals . Recent studies also implicate such a role for Arabidopsis ( Arabidopsis thaliana ) BRM but the molecular mechanisms underlying the antagonism are unclear . To understand the interplay between BRM and PcG during plant development , we performed a genome-wide analysis of trimethylated histone H3 lysine 27 ( H3K27me3 ) in brm mutant seedlings by chromatin immunoprecipitation followed by next generation sequencing ( ChIP-seq ) . Increased H3K27me3 deposition at several hundred genes was observed in brm mutants and this increase was partially supressed by removal of the H3K27 methyltransferase CURLY LEAF ( CLF ) or SWINGER ( SWN ) . ChIP experiments demonstrated that BRM directly binds to a subset of the genes and prevents the inappropriate association and/or activity of PcG proteins at these loci . Together , these results indicate a crucial role of BRM in restricting the inappropriate activity of PcG during plant development . The key flowering repressor gene SHORT VEGETATIVE PHASE ( SVP ) is such a BRM target . In brm mutants , elevated PcG occupancy at SVP accompanies a dramatic increase in H3K27me3 levels at this locus and a concomitant reduction of SVP expression . Further , our gain- and loss-of-function genetic evidence establishes that BRM controls flowering time by directly activating SVP expression . This work reveals a genome-wide functional interplay between BRM and PcG and provides new insights into the impacts of these proteins in plant growth and development .
Plant development takes place in distinct phases , each of which is characterized by the activation of a particular set of genes and the repression of others . Precise control of gene expression in each phase is crucial for proper growth and development . The transition from the vegetative to the reproductive phase is controlled precisely by multiple genetic pathways in response to environmental and endogenous signals [1–4] . In Arabidopsis , a repressor complex that consists of two MADS box transcription factors , FLOWERING LOCUS C ( FLC ) and SVP , serves as a negative regulator of flowering time by directly repressing the expression of the floral pathway integrators FLOWERING LOCUS T ( FT ) and SUPPRESSOR OF OVEREXPRESSION OF CO 1 ( SOC1 ) [1 , 5 , 6] . SVP is highly expressed during the vegetative phase [5 , 7] , but is down-regulated during the floral transition by the autonomous and gibberellin ( GA ) pathways [5] , which results in the de-repression of FT and SOC1 to promote flowering . Despite its key role in controlling flowering time , the mechanisms by which SVP expression is regulated are still unknown . Particularly , no positive regulator ( s ) of SVP expression in the vegetative phase have been identified . Polycomb Group ( PcG ) proteins are epigenetic repressors that maintain the repressed state of genes in cells where the genes should be inactive [8–11] . PcG proteins repress genes through combined activities of at least two multi-protein complexes known as Polycomb Repressive Complex 1 ( PRC1 ) and PRC2 [8] . PRC2 is involved in the establishment and maintenance of the repressed chromatin state , by introducing the H3K27me3 mark . Subsequently , PRC1 binds to the H3K27me3 mark and compacts the chromatin , resulting in the repressed state of PcG target genes . In Arabidopsis , at least three forms of PRC2 complexes exist and each of them acts at specific developmental phases [12–15] . CLF and SWINGER ( SWN ) are two putative H3K27 methyltransferases and act redundantly during the vegetative and reproductive stages [16] . Several thousands of Arabidopsis genes were reported to carry the H3K27me3 mark in young seedlings [17–19] . A fraction of PcG target genes was found to carry the H3K27me3 mark specifically in either the shoot apical meristem or leaf cells [18] , suggesting dynamic regulation of H3K27me3 deposition . Studies have been carried out to address how PcG proteins deposit H3K27me3 to target genes [12 , 13 , 20] . It is less known , however , about the mechanisms by which PcG activities are prevented from targeting certain genes to keep these genes on at particular developmental phases . SWI/SNF-type chromatin-remodeling protein complexes are thought to utilize energy from ATP hydrolysis to mobilize , disrupt or change nucleosomes to create an open chromatin structure for the access of transcriptional factors or other regulators [21 , 22] . The SWI2/SNF2 ATPase in Drosophila , BRM , was initially classified as a Trithorax group ( TrxG ) protein since it activates the transcription of homeotic genes and thus antagonizes the function of PcG during fly development [23 , 24] . However , recent studies indicate that it can either activate or repress target gene expression , through increasing or decreasing the accessibility of the target DNA [24–26] , yet its role in the regulation of gene expression is not well understood . Although the biochemical activities of plant SWI/SNF complexes have not been examined , progress has been made to identify the plant SWI/SNF complexes through genetic and molecular analysis [27–30] . In Arabidopsis , four SWI2/SNF2 ATPases including BRM and SPLAYED ( SYD ) , four SWI3 proteins ( SWI3A to SWI3D ) , two SWI/SNF ASSOCIATED PROTEINS 73 ( SWP73A and SWP73B ) , two ACTIN RELATED PROTEINS ( ARP4 and ARP7 ) , and a single SNF5 subunit termed BUSHY ( BSH ) were predicted subunits of SWI/SNF complexes [27] . Previous in vitro protein-protein interaction data [28 , 31] and a recent effort in protein complex purification followed by peptide sequencing [32] suggest that these proteins could form several SWI/SNF complexes . Subunits of Arabidopsis SWI/SNF complex ( es ) play crucial roles in many aspects of plant development [26 , 27 , 33–36] . The SWP73B ( also called BAF60 ) subunit has been shown to participate in the control of flowering time [37] . The SWI3C protein is involved in gibberellin ( GA ) responses [38] . brm mutants show pleiotropic phenotypes , such as reduced plant size [28 , 39] , downward curling of leaves [28 , 33] , mild floral homeotic defects [28 , 34] , hypersensitivity to abscisic acid [26] and early flowering [33 , 39 , 40] . Efforts have been made to understand the reason why brm mutants show an early flowering phenotype [40] , but the precise role of BRM in flowering time control is still not clear . Although the roles of PcG proteins and BRM during plant development have been investigated individually , how their activities are coordinated is poorly understood . Interestingly , a recent report in Arabidopsis showed that loss of BRM activity led to the increased H3K27me3 at two floral homeotic genes [34] , suggesting the antagonistic relationship between BRM and PcG . However , the current model is solely based on the characterization of a few identified targets of BRM , and it remains unknown to what extent BRM is required for antagonizing PcG function in plant . Also the precise mechanism by which BRM antagonizes PcG activity during plant development remains unclear . Finally , whether or not plant BRM might work synergistically with PcG proteins is completely unknown . To address these questions , we have performed a genome-wide analysis of H3K27me3 in brm mutant seedlings by chromatin immunoprecipitation followed by next generation sequencing ( ChIP-seq ) . We identify several hundred genes that show increased levels of H3K27me3 upon loss of BRM activity , demonstrating the critical role of BRM in preventing genes from H3K27me3-mediated repression in plant cells . We further show that there is inappropriate invasion of PcG proteins . Finally , by taking advantage of our genome-wide data , we uncover a role for BRM in repressing flowering by activating directly the expression of SVP , thus providing an explanation for the early flowering phenotype of brm mutants . Together , our results demonstrate that BRM is essential for proper H3K27me3 distribution in the genome and thus plant development .
To examine whether BRM affects the patterns of H3K27me3 deposition and distribution in a genome-wide scale , we performed ChIP-seq with anti-H3K27me3 antibodies in wild-type Col and brm-1 , a null allele with a T-DNA insertion [28] . Two independent biological DNA samples were generated and used for sequencing . We mapped the reads to the Arabidopsis genome and identified H3K27me3-enriched regions in both wild-type and brm mutants . Only H3K27me3-enriched regions identified in both biological replicates were chosen for further data analysis . In 14-day-old wild-type Col seedlings , we identified 5 , 591 regions , corresponding to 7 , 230 genes , which were marked by H3K27me3 ( S1 Dataset ) . H3K27me3 target genes identified in our study cover more than 95% ( 6 , 322 out of 6634 ) of those reported in a previous ChIP-seq analysis [19] . Furthermore , in both Col and the brm-1 mutant , the patterns of H3K27me3 at several well-characterized H3K27me3 target genes , such as AGAMOUS ( AG ) , APETALA3 ( AP3 ) , FLC and FT , are very similar to those reported by Lu et al [19] ( Fig . 1A ) . In contrast , no H3K27me3 signals at two highly expressed genes , ACTIN2/7 and TUB2 , were observed ( Fig . 1B ) , demonstrating the quality and reliability of our ChIP-seq data . Compared to wild-type , we identified 276 genes at which H3K27me3 levels changed more than twofold in the brm-1 mutant ( see the Materials and Methods section for details ) . Out of the 276 genes , 258 ( 93 . 5% ) genes showed more than a twofold increase in H3K27me3 in brm-1 , while only 18 ( 6 . 5% ) genes showed more than a twofold reduction in H3K27me3 in brm-1 ( S2 Dataset ) . Our genome-wide data show that BRM mainly acts to antagonize PRC2 activity during vegetative development , which is consistent with its expected role as a TrxG protein . However , the decreased H3K27me3 at a smaller set of genes in brm mutant suggests that BRM could also promote PcG activity at certain loci . We performed a Gene Ontology ( GO ) analysis for the genes showing increased H3K27me3 deposition using the BINGO software [41] . In the classification of biological processes , these genes were highly enriched in “regulation of metabolic process” ( P = 9 . 69E-4 ) and “regulation of gene expression” ( P = 3 . 8E-4; Fig . 1C ) , and in terms of molecular function , the most enriched category observed was “transcription regulator activity” ( P = 1 . 57E-4 ) . Thus , BRM is involved in a wide spectrum of cellular processes such as gene expression regulation and metabolism through preventing PcG proteins from H3K27me3 deposition . To validate our ChIP-seq data , we randomly chose a subset of genes and performed ChIP followed by quantitative PCR ( ChIP-qPCR ) using independent chromatin samples . We confirmed the changes in H3K27me3 levels at all 10 selected genes in brm-1 ( Fig . 1D and 1E ) . We did not detect any marked changes at the PcG non-targets ACTIN2/7 or the PcG target AG ( Fig . 1E ) . Next , we asked whether the elevated H3K27me3 levels in the brm-1 mutant caused down-regulation of the corresponding genes . We measured the expression levels of several selected genes that showed increased H3K27me3 levels in brm-1 by quantitative Reverse Transcription-PCR ( qRT-PCR ) and observed decreased expression for most but not all of them in brm-1 ( Fig . 1F ) . Interestingly , we also found increased expression of WRKY23 ( Fig . 1F ) , a gene with decreased H3K27me3 levels in brm-1 ( Fig . 1D and 1E ) . These data indicate that there is a positive correlation between increased H3K27me3 levels and decreased gene expression in brm-1 and also suggest that increased H3K27me3 deposition alone in brm might not be sufficient for gene repression at some target loci . In Arabidopsis , CLF is thought to be a major H3K27 methyltransferase responsible for the deposition of H3K27me3 in tissues other than seeds [42 , 43] . LIKE HETEROCHROMATIN PROTEIN 1/TERMINAL FLOWER 2 ( LHP1/TFL2 ) associates with regions with H3K27me3 across the Arabidopsis genome and was proposed to be a key component of a plant PRC1 complex [44 , 45] . Both clf and tfl2 single mutants showed up-ward curling of leaves ( Fig . 2A ) [42 , 46] . We reasoned that CLF might be required for the increased H3K27me3 levels at some genes in the brm-1 mutant . To test this , we first generated a brm clf double mutant to examine the genetic relationship between CLF and BRM . clf single mutants display up-wardly curled leaves while brm single mutants show down-ward curling of leaves [28] . Up-ward leaf curling in clf mutants can be caused by ectopic expression of floral homeotic genes such as AG , AP1 , and AP3 [16 , 42] . In the brm clf double mutants , the up-ward curling of leaves was weaker than that in clf single mutants ( Fig . 2A and S1 Fig . ) , suggesting that brm can partially suppresses clf . We also generated brm tfl2 double mutants . The leaves of the brm tfl2 double mutants showed down-ward curling as those in brm single mutants ( Fig . 2A and S1 Fig . ) suggesting that brm suppresses tfl2’s phenotype of up-wardly curled leaves . These genetic data support a notion that BRM antagonizes PcG function during vegetative development . Consistent with the partially rescued up-ward leaf curling in brm clf double mutants , we found decreased ectopic expression of several floral homeotic genes such as AG , AP1 , and AP3 in brm clf double mutant leaves compared to clf single mutants ( S2 Fig . ) . Interestingly , the brm clf double mutants were also smaller in terms of overall size than either single mutant , suggesting the additive effect of the two mutations in plant development . Supporting this observation , we noticed that there were more genes mis-regulated in brm clf double mutants than either single mutant ( S3 Fig . ) . To test if CLF is required for the increased H3K27me3 levels at some genes in the brm-1 mutant , we measured genome-wide H3K27me3 levels in brm clf double mutants by ChIP-seq and compared them with those in brm single mutants . We found that removal of CLF activity led to a marked reduction of H3K27me3 levels at approximately half of the genes with increased H3K27me3 levels in brm-1 ( 133 out of 258; Fig . 2B; S3 Dataset ) , indicating the requirement for CLF activity for the increased H3K27me3 levels at some of the genes in brm mutants . We validated these results by ChIP-qPCR at selected genes ( Fig . 2C ) . It is worth noting , however , that there was no drastic loss in H3K27me3 levels at the rest of the genes in the brm clf double mutant relative to the brm single mutant ( Fig . 2B and 2C ) , which might be explained by the redundant SWN activity at these loci . To examine the contribution of SWN , we generated the brm-1 swn-4 double mutant . We found that H3K27me3 levels at the majority of the selected loci in the brm swn double mutant were lower than those in the brm single mutant ( Fig . 2C ) . Furthermore , we scanned the SVP locus and included a region from the neighboring gene At2g22560 , for H3K27me3 distribution in all five mutant backgrounds . As shown in Fig . 2D , the results were consistent with those in Fig . 2B and 2C , suggesting a redundant role of CLF and SWN at SVP . These observations are consistent with a scenario in which BRM acts to protect some gene loci from PcG activity in developing seedlings so that these genes stay active . To determine if the increase in H3K27me3 levels in brm mutants was due to increased CLF/SWN presence at the loci , we first measured CLF occupancy at the loci in the brm mutant relative to wild-type using a GFP-tagged CLF line [43] . As shown in Fig . 3A , CLF occupancy was increased at all the selected genes , suggesting that , in the absence of BRM , CLF is allowed to access some inappropriate genomic regions , resulting in increased H3K27me3 levels . We then examined the involvement of SWN . For that , we generated an YFP-tagged SWN line and performed ChIP with anti-YFP antibodies to measure SWN occupancy in the brm mutant relative to wild type . We found that the occupancy of SWN was also increased at the majority of the loci when BRM was absent ( Fig . 3B ) . Furthermore , we also scanned the SVP locus , including the region in the neighboring gene , to compare patterns of CLF/SWN occupancy between wild type and brm-1 . As shown in Fig . 3C and 3D , the two proteins were markedly enriched in brm-1 across the SVP locus with a strong bias towards the transcription start site ( TSS ) , implicating a redundant action of CLF and SWN at SVP . These observations suggest that increased CLF/SWN occupancy could contribute to the elevated levels of H3K27me3 in brm mutants . On the other hand , by comparing H3K27me3 levels and CLF/SWN occupancy at SVP relative to the control loci such as At2g22560 and ACTIN , low but significant levels of H3K27me3 ( Fig . 2D ) and CLF/SWN ( Fig . 3C and 3D ) at SVP were found in wild-type plants . This suggested that BRM prevents high levels of H3K27me3 and CLF/SWN occupancy rather than excluding them . Alternatively , it could also function to keep PcG in an inactive state . At the WRKY23 locus , CLF/SWN occupancy was reduced in the brm mutant ( Fig . 3A and 3B ) , consistent with the decreased H3K27me3 levels observed at this locus ( Fig . 1D and 1E ) . Next , we asked how BRM antagonizes PcG function during vegetative growth , i . e . , whether it does so directly or indirectly . One of the possibilities that could explain the increased H3K27me3 deposition and PcG protein occupancy on chromatin in brm is the elevated expression level of genes encoding PcG subunits . To address this issue , we examined the expression levels of genes encoding PRC2 components , including CLF , SWN , EMBRYONIC FLOWER2 ( EMF2 ) , VERNALIZATION2 ( VRN2 ) , FERTILIZATION-INDEPENDENT ENDOSPERM ( FIE ) and FERTILIZATION INDEPENDENT SEED2 ( FIS2 ) [12] in brm mutants . The expression of these genes was not increased markedly in brm-1 compared to that in wild-type ( S4 Fig . ) , indicating that BRM does not antagonize PcG through repressing the expression of PcG-encoding genes . We also measured histone H3 levels at selected genes , and found a slight increase in brm-1 ( S5 Fig . ) . However , the change in H3 levels was very small and thus could not fully account for the change in H3K27me3 levels . These observations point to the possibility that BRM acts directly at the target loci to antagonize PcG proteins . We then tested whether BRM acts directly on the affected genes by physically binding to these genes . We performed ChIP-qPCR experiments to examine BRM occupancy at the affected genes . For the ChIP assay , we used a transgenic Arabidopsis line expressing a GFP-tagged BRM transgene under the control of the BRM native promoter ( ProBRM:BRM-GFP ) [47] . The transgene could fully rescue the morphological defects of the brm-1 null mutant ( Fig . 4A ) , suggesting that it is functional in vivo . ChIP was performed with anti-GFP antibodies and Pro35S:GFP plants were used as the negative control . The ChIP DNA was analyzed by qPCR to examine the enrichment of BRM at target genes . Genomic regions around the transcription start site at all examined genes were significantly enriched in the BRM-GFP ChIP ( Fig . 4B ) . Furthermore , we scanned the SVP locus , including the negative control region in the neighboring gene , for BRM occupancy . As shown in Fig . 4C , the BRM was found to be significantly enriched at the SVP locus , particularly near the TSS . The physical association of BRM with these selected genes , in combination with increased H3K27me3 deposition and decreased expression of the genes in brm mutants , suggests that BRM acts directly at these target loci , to keep the PRC2 activity off and thus promote gene activity . Loss of BRM activity allows the access to these loci by PRC2 , which turns off or decreases gene expression . In the sections below , we present our observations to demonstrate that SVP is a main target of BRM in the control of flowering . SVP is a key negative regulator of flowering in Arabidopsis , and loss-of-function of SVP results in early flowering [5 , 7] . Consistent with its role in maintaining the duration of the vegetative phase , SVP is highly expressed in seedlings , but is barely detectable in inflorescence tissues [7] . We noticed initially from our ChIP-seq and ChIP-qPCR data ( Fig . 1D and 1E ) that H3K27me3 levels drastically increased at the SVP locus in brm-1 compared with wild-type . These data suggest that the SVP locus becomes a PRC2 target in the absence of BRM activity . The increase in H3K27me3 levels at the SVP locus in brm raises the possibility that BRM may act to keep SVP on by antagonizing PcG activity during vegetative growth . To test this hypothesis , we first extended the single time point expression analysis of SVP in brm-1 as presented in Fig . 1F by examining the expression of SVP in the brm-1 mutant spanning several developmental time points . Indeed , the expression of SVP in the brm-1 mutant was consistently lower than that in wild-type plants over a time course spanning 8 to 14 days after germination ( DAG , Fig . 5A ) , suggesting that BRM activity is continually required for the high levels of SVP expression in seedlings . The decreased expression of SVP was unlikely due to the accelerated floral transition of brm-1 plants , since the expression of AP1 , a marker gene for the vegetative-to-floral developmental transition [48 , 49] , remained low throughout the time course ( S6 Fig . ) . To confirm that BRM activates the expression of SVP , we generated XVE::aMIRBRM transgenic lines that harbor an inducible artificial microRNA ( amiRNA ) targeting BRM ( Fig . 5B ) . As shown in Fig . 5C , BRM transcript levels in 7-day-old XVE::aMIRBRM seedlings treated with β-estradiol to induce the amiRNA were gradually decreased by approximately 50% during a 24h time course , indicating that the amiRNA was effective . SVP transcript levels showed a similar reduction kinetics in the time course ( Fig . 5D ) . This result reveals that proper BRM activity is required for SVP expression . To further verify that BRM activates SVP expression at the transcriptional level , we obtained a previously developed SVP promoter-GUS fusion reporter line ( ProSVP:GUS ) [5] , and introduced it into the brm-1 background by genetic crosses ( brm-1 ProSVP:GUS ) . As shown in Fig . 5E , GUS activity in brm-1 ProSVP:GUS was almost invisible compared to that in ProSVP:GUS at all three time points ( Fig . 5E ) , suggesting that the promoter of SVP has no detectable activity when BRM is absent . As negative controls , we also stained Col wild-type and brm-1 mutants but saw no signals ( Fig . 5E ) . Documented Arabidopsis gene expression data indicate a temporal and spatial overlap of the SVP and BRM expression patterns in leaves ( S7 Fig . ) [50] , which is consistent with a role for BRM as a positive regulator of SVP in developing seedlings . These observations demonstrate a positive spatial and temporal correlation between BRM and SVP expression , and when combined with our BRM-GFP ChIP data ( Fig . 4B and 4C ) that showed a direct BRM binding to the SVP locus , indicate that BRM directly promotes SVP expression during vegetative development . Having shown above a positive role for BRM in regulating SVP expression , we next sought to investigate whether the BRM-SVP module can largely explain the early flowering phenotype of the brm mutant . Both brm and svp single mutants show early flowering phenotypes under long-day conditions [6 , 7 , 33 , 39 , 40] , but it is not known whether there is a common molecular mechanism underlying their flowering phenotypes . We first estimated the flowering time in the two mutants by counting the number of leaves at bolting ( Fig . 6A and 6B , top panel ) . brm-1 and svp-31 , a null T-DNA insertion mutant [6] , flowered at roughly the same time . The svp-31 heterozygous plants flowered significantly later than their homozygous siblings but earlier than wild-type Col plants , indicating that SVP controls flowering in a dosage-dependent manner , consistent with previous observations [7] . Next , taking advantage of the dosage-dependent nature of flowering control by SVP , we quantified SVP transcript levels by qRT-PCR in the mutant plants to estimate the contribution of SVP to the flowering control by BRM ( Fig . 6B , middle panel ) . Our qRT-PCR data confirmed that svp-31 is a null allele and the heterozygous plants accumulated approximately half the amount of SVP transcripts found in wild-type plants ( Fig . 6B , middle panel ) . SVP expression in brm-1 was drastically decreased to less than half that of svp-31 heterozygous plants . In brm-1 ProBRM:BRM-GFP plants , both the flowering time and SVP expression were restored to the wild-type level ( Fig . 6A and 6B ) , further confirming that BRM activity is responsible for the normal expression level of SVP . Our quantification of flowering time and SVP transcript levels in brm-1 , when compared quantitatively to those from svp-31 mutants , suggests that 1 ) BRM is a major activator of SVP expression; and 2 ) The early flowering phenotype of the brm-1 mutant can largely be accounted for by the down regulation of SVP transcription in the mutant , although other BRM targets also have minor contributions . To provide additional evidence to strengthen our conclusion , we tested whether restoration of SVP in brm mutant background could overcome its early flowering phenotype by expressing SVP from a promoter that is not controlled by BRM ( Pro35S:SVP ) [51] in brm-5 , a chemically induced mutant that has a single nucleotide change in the region encoding the ATPase domain [33] . Indeed , introduction of Pro35S:SVP into brm-5 could rescue the early flowering of the brm-5 mutant ( Fig . 6C ) . We also generated a brm-1 svp-31 double mutant to test the genetic interaction between BRM and SVP in flowering time control . The brm-1 svp-31 double mutant flowered only slightly earlier than either single mutant ( Fig . 6B ) , suggesting that BRM and SVP act largely in a common pathway in determining flowering time , and only minor contributions from other BRM targets . It is worth mentioning that three other flowering time genes also displayed increased H3K27me3 levels in the brm mutant ( S2 Dataset and S8 Fig . ) . When we checked the expression of these genes , we only saw a clear decrease of AGAMOUS-LIKE24 ( AGL24 ) expression but not the other two in brm-1 ( S8 Fig . ) . AGL24 is a MADS-box protein involved in flowering time control . agl24 mutants show delayed flowering while agl24 svp double mutants are early flowering as svp single mutants [52] . The data thus suggest that the early flowering phenotype of brm mutants is unlikely caused by these flowering time genes . In addition , we also examined the expression of FT , a well-established SVP target , in the various genetic backgrounds ( Fig . 6B , bottom panel ) . As expected , FT transcript levels correlated negatively with those of SVP and positively with flowering time in the corresponding genetic backgrounds . In summary , our observations strongly suggest that BRM represses flowering mainly through activating SVP .
In both animals and plants , a group of proteins that counteract PcG function have been described and referred to as TrxG proteins [9 , 13] . Several putative TrxG proteins have been proposed in Arabidopsis , including the H3K4 methyltransferase ATX1 [53] , the SAND-domain DNA binding protein ULTRAPETALA1 ( ULT1 ) [54] , the chromatin remodeling ATPase PICKLE ( PKL ) [55] , the H3K27me3 demethylase REF6 [19] , and the SWI2/SNF2 ATPases SPLAYED ( SYD ) and BRM [34] . A challenge for the field is to understand the specific roles of the putative TrxG proteins and the functional relationship among them in antagonizing PcG . The nature of the antagonism between SWI/SNF-type chromatin remodeling ATPases and PcG proteins has been investigated in several studies in animals; and several models of counteraction have been proposed [23 , 25 , 56–58] . Interestingly , a very recent report in Arabidopsis showed that BRM overcomes the repression of AG and AP3 by the PcG pathway during the initiation of floral development [34] , however , how it does so and to what extent BRM is required for antagonizing PcG function in plants remains unclear . Our genome-wide study shows that BRM deficiency led to an increase in H3K27me3 levels at several hundred genes across the genome during vegetative development in Arabidopsis . We further observed increased occupancy of CLF/SWN-containing PcG complex ( es ) at these genes when BRM is not located there ( Fig . 3; S1 Table ) . Considering that there are low but significant levels of H3K27me3 ( Fig . 2B–2D ) and CLF/SWN occupancy ( Fig . 3 ) at SVP in wild-type plants , we favour a model of antagonism between BRM and PcG , in which BRM might function to prevent high levels H3K27me3 and CLF/SWN occupancy instead of excluding them ( Fig . 7 ) . It is also possible that BRM could function to keep PcG in an inactive state . In addition to chromatin remodelers , plants might employ transcription factors to counteract PcG . A recent study showed that the binding of transcription factor AG to the promoter of zinc finger repressor KNUCKLES ( KNU ) causes the eviction of the PcG proteins from the locus , leading to the induction of KNU [59] . Thus , both transcription factors and chromatin remodeling proteins could be involved in counteracting PcG . It will be interesting to determine whether and how these two machineries work together in antagonizing PcG function . Our genome-wide analysis of H3K27me3 occupancy in brm mutant indicates that BRM does not only antagonize PcG function during plant development , but also cooperates with PcG at some loci ( Fig . 1D and 1E ) . For example , the H3K27me3 level at WRKY23 is decreased and the expression of the gene is up-regulated in both brm and fie mutant ( FIE is a PcG subunit ) seedlings [17] ( this study ) , suggesting that both BRM and PcG are required for the proper expression of WRKY23 . Further , we show that the decreased H3K27me3 observed at WRKY23 loci in brm mutant could be because of , at least partly , the decreased CLF binding . Therefore , this observation suggests that BRM may work with the PcG proteins at some of the common loci and thus repress the targets expression . WRKY23 was recently found to be needed for proper root development and the over-expression of WRKY23 results in the reduction of root length [60] . It will be interesting to test whether the increased transcription of WRKY23 could explain the short root phenotype of brm [28] . The synergistic relationship between BRM and PcG reported here was also observed by a study in human embryonic stem cell showing that an embryonic stem cell specific SWI/SNF complex acts synergistically with PRC2 at all four Hox loci [25] . The mechanism by which BRM cooperates with PcG is currently unknown . One possibility would be that BRM directly interacts with PcG and facilitates the targeting of PcG to genes . Indeed , we found that BRM co-localizes with H3K27me3 at the WRKY23 locus in wild-type seedlings ( Fig . 4 ) , suggesting that BRM might interact with PcG proteins . However , no study so far has demonstrated a direct physical interaction between BRM and PcG proteins . It is possible that these two complexes might interact transiently or indirectly . Nevertheless , the synergistic relationship between BRM and PcG found in both animals and plants might suggest its biological relevance and warrants further studies . The proper transition from vegetative growth to flowering is critical for the reproductive success of angiosperm plants and must be controlled precisely . BRM has been proposed as a repressor of flowering as suggested by the early flowering phenotype and the elevated FT expression of brm mutants [39 , 40] . However , it was not clear whether BRM acts directly or indirectly to repress FT . SVP has been demonstrated to be a direct repressor of FT [5 , 6] , and thus serves as a key repressor of floral transition . The precise regulation of SVP is obviously of critical importance for our understanding of flowering control . Thus far , however , no direct upstream activator of SVP has been identified . In this work , we provide evidence demonstrating that BRM represses FT by directly maintaining a high level of SVP expression ( Fig . 7 ) . First , loss of BRM activity results in decreased expression of SVP ( Fig . 5A–5E ) , which is associated with increased H3K27me3 levels ( Fig . 1D and 1E ) and increased occupancy of CLF and SWN ( Fig . 3C and 3D ) . Second , BRM directly binds to the SVP locus in vegetative tissues where SVP is highly expressed ( Fig . 4B and 4C ) . Together , these observations suggest that BRM represses the floral transition through directly activating SVP . This is consistent with the genetic evidence showing that the brm-1 svp-31 double mutant displays almost the same early flowering phenotype as brm-1 and svp-31 single mutants ( Fig . 6A and 6B ) . Although our data support a scenario that BRM represses flowering mainly through SVP , some evidence suggests that BRM may also repress flowering through other pathways . For example , the expression of CONSTANS ( CO ) , an activator of FT in the photoperiod pathway , was increased in brm mutants [40] . In addition , elevated expression of both FLC and FT in brm mutants was also reported previously [39 , 40] . Since FLC is a repressor of FT expression [61] , it seems hard to understand why the expression levels of both FLC and FT were increased in brm mutants . Our results presented here provide an explanation for this apparent discrepancy: mutation of BRM results in reduced expression of SVP and consequently lower abundance of the SVP-FLC repressor complex , ultimately leading to activation of FT , regardless of the increased expression of FLC . It is also relevant to note that down-regulation of BAF60/SWP73B was recently reported to cause increased FLC expression and delayed floral transition [37] . The apparently opposing flowering time phenotype of brm mutants and the BAF60 knockdown line is puzzling . It is unknown whether and how BAF60 regulates SVP expression . It might be possible that the presence of BAF60 in a SWI/SNF complex inhibits the activity of BRM , thus reduction of BAF60 could allow BRM to activate SVP expression , which , in turn , leads to delayed floral transition . Alternatively , it might also be possible that BRM and BAF60 are present in distinct complexes that differ in their regulatory activities and target genes , e . g . , BRM activates SVP , while BAF60 represses FLC . Our genome-wide H3K27me3 profiling data also reveal that BRM is involved in the regulation of a number of other important developmental genes including , most noticeably , members of the miR166 and miR156 families ( S2 Dataset ) . It is well established that the miR166 family miRNAs target the transcripts of the HD-ZIPIII genes , controlling the level and domain of their expression to allow their proper functions in plant development [62–64] . More recently , we uncovered a new role for miR166 in repressing the seed maturation program during vegetative development [65] . An earlier study demonstrated the involvement of BRM in repression of the seed maturation genes in leaves [33] – a brm mutation was isolated in a reporter-assisted genetic screen for Arabidopsis mutants exhibiting ectopic expression of seed storage protein genes [33 , 65 , 66] . Our new data presented here thus provide a potential link between the two early studies [33 , 65]: it strongly suggests that BRM promotes the accumulation of miR166 , which in turn represses seed maturation genes in developing seedlings . In conclusion , our work demonstrates that BRM promotes vegetative development by harnessing PcG proteins ( mainly by preventing their activities ) at key developmental genes .
Arabidopsis seeds were stratified at 4°C for 3 days in dark condition . Then the seeds were sown on soil or on agar plates containing 4 . 3 g/L Murashige and Skoog nutrient mix ( Sigma-Aldrich ) , 1 . 5% sucrose , 0 . 5 g/L MES ( pH 5 . 8 ) , and 0 . 8% agar . Plants were grown in growth rooms with 16-h-light/8-h-dark cycles ( Long-day , LD ) at 22°C or 16°C . T-DNA insertion mutants were obtained from the ABRC , unless otherwise indicated . The brm-1 ( SALK_030046 ) , brm-5 , clf-29 ( SALK_021003 ) , tfl2-1 ( CS3796 ) , svp-31 ( SALK_026551 ) and SWN-4 ( SALK_109121 ) mutants are all in the Col background and have been described previously [6 , 28 , 33 , 46 , 67 , 68] . Homozygous T-DNA insertion mutants were identified by PCR-based genotyping . Transgenic plants ProBRM:BRM-GFP , ProSVP:GUS , Pro35S:SVP , Pro35S:GFP-CLF and Pro35S:GFP have been described [5 , 43 , 47 , 69] . ChIP was carried out as described [70 , 71] with minor modifications . Briefly , two grams of 14-day-old seedlings grown on MS agar were cross-linked with 1% formaldehyde and then ground into fine power with liquid nitrogen . Chromatin was isolated and sheared into 200–800 base pair fragments by sonication . The sonicated chromatin was immunoprecipitated with 5 μL of anti-H3K27me3 ( 07–449 , Millipore ) , anti-GFP ( ab290 , Abcam ) or anti-H3 ( Ab1791 , Abcam ) antibodies . The precipitated DNA was then recovered with the MiniElute PCR Purification Kit ( Cat#28004 , Qiagen ) according to the manufacturer’s instructions . ChIP-qPCR was performed with three technical replicates , and results were calculated as percentage of input DNA according to the Champion ChIP-qPCR user manual ( SABioscience ) . If fold enrichment was used , the calculated percentage input of the wild-type control plant at the regions tested was set to 1 . The fold enrichment represents the fold change relative to the wild-type . Independent ChIP experiments were performed at least two more times and similar results were obtained . Primer sequences used for ChIP-qPCR were listed in S2 Table . Ten ng of ChIP DNA immuoprecipitated by the anti-H3K27me3 antibody was used for ChIP-seq library construction . End repair , adapter ligation and amplification were carried out using the Illumina Genomic DNA Sample Prep Kit according to the manufacturer’s protocol . Illumina Genome Analyser IIx or HiSeq 2500 was used for high-throughput sequencing of the ChIP-seq library . The raw sequence data were processed using the Illumina sequence data analysis pipeline GAPipeline1 . 3 . 2 . Then Bowtie [72] was employed to map the reads to the Arabidopsis genome ( TAIR10 ) [73] . Only perfectly and uniquely mapped reads were retained for further analysis . Then the data were analyzed as described [19] . Briefly , the alignments were first converted to WIG files using MACS [74] . Then the data were imported to Integrated Genome Browser ( IGB ) [75] for visualization . Secondly , the program SICER [76] was used to identify ChIP-enriched domains ( peaks ) in histone modification signals . Thirdly , quantitative comparisons between wild-type Col and mutants were performed using the ChIPDiff program [77] . Regions with more than twofold changes were kept for further analysis . Finally , the identified regions were annotated according to the Arabidopsis annotation gff file ( TAIR10 , www . arabidopsis . org ) using a customized Perl script . Two independent biological replicates were used for sequencing , and only the regions of H3K27me3 found in both replicates were included in the analyses . Total RNA was isolated from ∼100 mg of plant tissues using the RNeasy Plant Mini kit ( Qiagen ) . One μg RNA was reverse transcribed into cDNA using the High Capacity cDNA Reverse Transcription kit ( ABI ) . Random primers from the kit were used as primers . Real-time quantitative PCR was conducted using the SsoFast EvaGreen Supermix kit with the Bio-Rad CFX96 real-time PCR detection system ( Bio-Rad Laboratories , Inc . ) . The data shown in the figures are the average of three technical replicates . Results were repeated with two additional independent RNA samples ( biological replicates ) . GAPDH served as the internal reference . PCR primers used in real-time PCR are listed in S2 Table . For genome-wide expression analysis , three biological replicates of Col , brm-1 , clf-29 and brm-1 clf-29 samples were analyzed on Affymetrix ATH1 arrays . Genes showing 1 . 5 fold changes were considered to be differentially expressed . The SWN gene without the stop codon was amplified by PCR and cloned into the pDONR221 vector ( Invitrogen ) by BP reaction according to the manufacturer’s instructions . The resulting transgene in the entry vector was sequenced to make sure that no mutation was introduced during PCR . The transgene was then transferred into the pEarlyGate 104 Gateway-compatible destination vector [78] by LR reaction , according to the manufacturer’s instructions , to make Pro35S:YFP-SWN . The construct was introduced into Agrobacterium tumefaciens GV3101 , which was then used to transform swn-4 mutant plants using the floral dip method [79] . Transgenic plants were selected in MS agar media containing 50 μg/ml of Hygromycin B and allowed to grow in soil to maturity to yield seeds . PCR primers used in making the construct are listed in S2 Table . For generating the XVE::aMIRBRM construct , the pRS300 vector [80] was used as the backbone to first generate aMIRBRM . The primers used were designed according to WMD3 ( http://wmd3 . weigelworld . org/cgi-bin/webapp . cgi ) and are listed in S2 Table . aMIRBRM was subcloned into the pDONR221 vector ( Invitrogen ) , confirmed by sequencing , and then recombined into the pMDC7 Gateway-compatible destination vector [78] where the aMIRBRM transgene is controlled by a Estradiol-inducible promoter . The construct was transformed into Col wild-type plants by the floral dip method [79] . Transgenic plants were selected for Hygromycin B resistance and allowed to grow to maturity to yield seeds . Seven-day-old T2 transgenic plants were treated either by 10μmol β-Estradiol or DMSO ( as the mock control ) and samples were collected at different time points after the treatment . The standard GUS staining solution ( 0 . 5 mg/mL 5-bromo-4-chloro-3-indolyl-glucuronide , 20% methanol , 0 . 01 M Tris-HCl , pH 7 . 0 ) was used . Seedlings immersed in GUS staining solution were placed under vacuum for 15 min , and then incubated at 37°C overnight . The staining solution was removed and samples were cleared by sequential incubation in 75% and 95% ethanol . Wild-type and mutant plants were grown side by side in soil at 22°C or 16°C with 16-h-light/8-h-dark cycles . The number of rosette leaves was counted when the length of the inflorescence stem was 1 cm . For each genotype , at least 20 plants were analyzed , and the analysis was repeated 3 times independently . Sequence data from this article can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: BRM ( AT2G46020 ) , SVP ( AT2G22540 ) , CLF ( AT2G23380 ) , BEL1 ( AT5G41410 ) , TCP2 ( AT4G18390 ) , WRKY23 ( AT2G47260 ) , miR156D ( AT5G10945 ) , LHP1/TFL2 ( AT5G17690 ) , EMF2 ( AT5G51230 ) , FIE ( AT3G20740 ) , SWN ( AT4G02020 ) , VRN2 ( AT4G16845 ) , AP1 ( AT1G69120 ) , TA3 ( AT1G37110 ) , ACTIN2/7 ( AT5G09810 ) , AG ( AT4G18960 ) , AP3 ( AT3G54340 ) , FLC ( AT5G10140 ) , AGL24 ( AT4G24540 ) , SMZ ( AT3G54990 ) and FT ( AT1G65480 ) . All raw ChIP-seq dataset and ATH1 expression array dataset have been deposited in the Gene Expression Omnibus database under accession number GSE47202 and GSE53623 .
|
In flowering plants , the proper transition from vegetative growth to flowering is critical for their reproductive success and must be controlled precisely . Multiple genes have been shown to regulate the floral transition in response to environmental and endogenous cues . Among them is SHORT VEGETATIVE PHASE ( SVP ) , a key flowering repressor gene in Arabidopsis . SVP is highly expressed during the vegetative phase to promote growth , but the mechanism by which the high expression level of SVP is maintained remains unknown . Here , we report a genome-wide study to examine the functional interplay between the BRM chromatin remodeler and the PcG proteins that catalyze trimethylation of lysine 27 on histone H3 ( H3K27me3 ) , a histone mark normally associated with transcriptionally repressed genes . We identify BRM as a direct upstream activator of SVP . BRM acts to keep the levels of H3K27me3 low at the SVP locus by inhibiting the binding and activities of the PcG proteins . Thus , our work identifies a previously unknown mechanism in regulation of flowering time and demonstrates the power of genome-wide approaches in dissecting regulatory networks controlling plant development .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
The Arabidopsis SWI2/SNF2 Chromatin Remodeler BRAHMA Regulates Polycomb Function during Vegetative Development and Directly Activates the Flowering Repressor Gene SVP
|
Post kala-azar dermal leishmaniasis ( PKDL ) , a sequel to visceral leishamaniasis ( VL ) in 5–15% cases , constitutes a parasite reservoir important in disease transmission . The precise immunological cause of PKDL outcome remains obscure . However , overlapping counter regulatory responses with elevated IFN-γ and IL-10 are reported . Present study deals with ex-vivo mRNA and protein analysis of natural regulatory T cells ( nTreg ) markers ( Foxp3 , CD25 and CTLA-4 ) and IL-10 levels in lesion tissues of PKDL patients at pre and post treatment stages . In addition , correlation of nTreg markers and IL-10 with parasite load in tissue lesions was investigated . mRNA levels of nTreg markers and IL-10 were found significantly elevated in pre-treatment PKDL cases compared to controls ( Foxp3 , P = 0 . 0009; CD25 & CTLA-4 , P<0 . 0001; IL-10 , P<0 . 0001 ) , and were restored after treatment . Analysis of nTreg cell markers and IL-10 in different clinical manifestations of disease revealed elevated levels in nodular lesions compared to macules/papules . Further , Foxp3 , CD25 and IL-10 mRNA levels directly correlated with parasite load in lesions tissues . Data demonstrated accumulation of nTreg cells in infected tissue and a correlation of both IL-10 and nTreg levels with parasite burden suggesting their role in disease severity in PKDL .
Leishmaniasis constitutes various forms of globally widespread group of neglected diseases caused by an obligatory intracellular protozoan parasite of genus Leishmania . It is currently endemic in 88 countries and overall prevalence is estimated as 12 million with 350 million at risk . Visceral leishmaniasis ( VL ) is the most severe form , fatal if not treated . 90% of all VL cases world wide occur in India , Bangladesh , Nepal , Brazil and Sudan [1] . In India , Leishmania donovani causes VL or Kala azar ( KA ) and Post kala azar dermal leishmaniasis ( PKDL ) while L . tropica is responsible for cutaneous leishmaniasis ( CL ) in humans . PKDL is an unusual dermatosis that develops in 5–15% of apparently cured VL cases in India and in about 60% of cases in Sudan [2] . This chronic skin condition produces gross cutaneous lesions in the form of hypopigmented macules , erythema and nodular stages . So far , little is known about the parasite/host factors that drive the parasite to shift from site of initial infection viscera ( spleen or bone marrow ) to the dermis or about the clinical manifestation of the disease . Inadequate treatment is considered to be a factor in PKDL development; however , the disease may develop even after complete treatment . Factors such as genetics and nutrition may be important [2] and remain to be explored in Indian PKDL . The precise immunological cause remains obscure . In Sudanese PKDL , immune suppression , reinfection or reactivation is considered to be the major underlying cause of PKDL development [2] . Reactivation of disease in the form of PKDL is suggested , on account of retention and maintenance of residual IL-10 and TGF-β levels in sodium antimony gluconate ( SAG ) treated KA individuals [3] . However , current reports suggest that PKDL may develop even after treatment with anti-leishmanial drugs such as Amphotericin B or Miltefosine . Thus , other mechanisms may be responsible for disease development . Like human VL , elevated levels of IFN-γ and TNF-α are reported systemically or in lesion tissues of PKDL with simultaneous presence of immunosuppressive cytokine , IL-10 , suggesting that there is no defect in mounting antigen specific responses [4] . Direct correlation between circulating IL-10 levels with parasite load suggests role of IL-10 in compromising the effector T cell function in human VL [5] , [6] . In , addition IL-10 knockout mice are highly resistant to L . donovani infection and treatment with anti-IL-10 receptor antibody promotes clinical cure [7] . Several IL-10–producing CD4+ T cell subpopulations have been described , among them naturally occurring CD4+CD25+Foxp3+ regulatory T ( nTreg ) cells are one such sub-population with unique ability to inhibit the response of other T cells . It is characterized by the constitutive expression of IL-2R-α chain ( CD25 ) and by expression of the transcriptional factor , Foxp3 . Evidence from experimental murine models of L . major infection suggests that nTreg cells promote survival of Leishmania parasites and reactivation of disease [8] . In addition , in human CL intralesional nTreg have been associated with SAG unresponsiveness and disease pathology [9] , [10] . In our previous report we proposed the role of nTreg in PKDL in the context of overlapping counter-regulating cytokines and chronic persistent infection [4] . The present study was carried out to evaluate if nTreg and IL-10 have an influence on disease persistence in human . The study supports the role of IL-10 and nTreg in PKDL pathology , as evident from a direct correlation of IL-10 and Foxp3 mRNA levels with parasite load within lesion tissues .
Lesional skin tissues were collected from PKDL patients ( n = 25 ) originating from Bihar and reporting to the Department of Dermatology , Safdarjung Hospital , New Delhi ( Table 1 ) . Biopsies were collected from face ( n = 19 ) or shoulder region ( n = 6 ) . All patients were HIV seronegative and diagnosis was confirmed by microscopy and polymerase chain reaction , described previously [4] . Furthermore , based on lesion types , PKDL patients were categorized in 3 groups , nodular ( N ) ( n = 12 ) , macular or papular ( M/P ) ( n = 10 ) and polymorphic ( n = 3 ) . Patients were treated with oral Miltefosine ( 150 mg/day ) for 2 months . Follow-up samples were collected from the same site as at pre treatment stage one month after completion of treatment in 8 cases , all of which showed apparent clinical cure . Further no parasites were detectable by real time PCR in any of these cases at this stage . 5 normal skin tissues were collected from the shoulder region of healthy individuals . The study was approved by and carried out under the guidelines of the Ethical Committee of the Safdarjung Hospital , India . All patients provided written informed consent for the collection of samples and subsequent analysis . Punch biopsy ( 4–6 mm ) samples were collected from PKDL patients and healthy individuals in RNAlater ( Ambion , Austin , TX , USA ) and stored in liquid nitrogen until use . Total RNA was isolated using Trizol reagent , in accordance with the manufacturer's instructions , and quality of RNA was assessed using Bioanalyzer ( Agilent , Foster City , CA , USA ) . Quantity of RNA was determined by Nanodrop spectrophotometer ( Thermo Scientific , Wilmington , DE , USA ) . cDNA was synthesized from 2 µg of total RNA using High capacity cDNA synthesis kit ( Applied Biosystems , Foster City , CA , USA ) . Incubation conditions for reverse transcription were 10 min at 25°C , followed by 2 hours at 37°C and were performed on a MasterCycler Gradient ( Eppendorf , Hamburg Germany ) . Real-time polymerase chain reaction was performed on an ABI Prism 7000 sequence detection system ( Applied Biosystems ) using cDNA specific FAM-MGB–labeled Taqman primer/probe sets ( Applied Biosystems ) for IL-10 ( Hs00174086_m1 ) , CD25 ( Hs00166229_m1 ) , CTLA-4 ( Hs00175480_m1 ) , Foxp3 ( Hs00203958_m1 ) . VIC-MGB labeled 18S rRNA ( 4319413E ) was used as endogenous control for relative amount of mRNA in each sample . The relative quantification of products was determined by the number of cycles over endogenous control required to detect the gene expression of interest . PKDL skin lesion tissue ( n = 12 ) was collected in NET buffer [150 mM NaCl , 15 mM Tris-HCl ( pH 8 . 3 ) and 1 mM EDTA] . Tissue was homogenized in liquid nitrogen and DNA was extracted using QIAamp DNA Tissue kit ( QIAGEN ) according to manufacturer's instructions . All samples were processed on the same day and isolated DNA was stored at −80°C until use . DNA samples from 3 patients who were part of our previous study [5] were also included . SYBR Green I based Real-time PCR was used for accurate quantification of parasite load as described previously [5] . Briefly , PCR reaction was performed in an ABI Prism 7000 sequence detection system ( Applied Biosystems ) using forward primer ( 5′-CTTTTCTGGTCCTCCGGGTAGG-3′ ) , reverse primer , ( 5′-CCACCCGGCCCTATTTTACACCAA-3′ ) . A standard curve was constructed using 10-fold serially diluted L . donovani parasite DNA corresponding to 104 to 0 . 1 parasite per reaction . Punch biopsy skin tissue was collected in neutralized formaline . The tissue was paraffin embedded and 5 µm sections were prepared on polylysine coated glass slides from all skin specimens . Immunohistochemical staining was performed using a standard polymer-peroxidase technique ( Dako , Carpinteria , CA , USA ) . After deparaffination , rehydration and blockade of endogenous peroxidase activity , heat-induced antigen retrieval was performed in Tris-EDTA buffer ( pH 9 . 0 ) . After antigen retrieval sections were covered with serum-free protein block ( Dako ) for 1 hr , followed by incubation with anti-human Foxp3 ( e-biosciences , San Diego , USA ) for 1 hr and EnVision1 anti-mouse/horseradish peroxidase ( HRP ) polymer ( Dako ) for 30 min at room temperature . Color was developed using diaminobenzidine ( DAB1 ) chromogen system . Statistical analysis was performed using Graph Pad Prism 5 ( GraphPad Software , Inc . , San Diego , CA , USA ) . Correlation was evaluated using Spearman/Pearson correlation test . P values of less than 0 . 05 were considered significant .
Natural Treg markers and IL-10 mRNA level were evaluated in skin lesion tissues of PKDL patients and compared with healthy controls tissues . mRNA analysis showed significantly elevated levels of nTreg markers in pre-treatment cases compared to control ( Foxp3 , P = 0 . 0009; CD25 & CTLA-4 , P<0 . 0001 ) ( Figure 1A ) . After treatment , a significant decrement in mRNA levels ( Foxp3 , P = 0 . 0025; CD25 , P = 0 . 0002 & CTLA-4 , P<0 . 0001 ) was noticed in paired samples ( Figure 1B ) . In addition , we evaluated mRNA levels of IL-10 , a molecule produced by adaptive Treg or Tr1 cells , and frequently associated with experimental or human VL pathology [5]–[7] . IL-10 mRNA level was significantly elevated in PKDL compared to control ( P<0 . 0001 ) . A significant drop in IL-10 mRNA levels was noticed in paired samples ( P = 0 . 0004 ) ( Figure 1 ) . To further evaluate whether there is any direct association between expression of these molecules and parasite burden , we evaluated the parasite load in PKDL lesion tissues ( n = 15 ) . The median parasite load was 776 parasites/µg tissue DNA ( range = 3 to 590 , 000 parasites/µg tissue DNA ) , with a higher parasite load in nodular lesions ( median 2 , 244 parasites/µg tissue DNA , n = 10 ) as compared to that in papular/macular lesions ( median 28 parasites/µg tissue DNA , n = 5 ) . Ex vivo analysis showed a direct correlation between parasite load and mRNA levels in lesion tissues ( CD25 , r = 0 . 691; Foxp3 , r = 0 . 817; IL-10 , r = 0 . 821 ) , such correlation was not noticed for CTLA-4 ( Figure 2 ) . No parasite was detected in any of the post treatment sample . To authenticate mRNA expression at protein level , IHC staining for Foxp3 expression was evaluated between groups . Three PKDL samples showed intense nuclear staining for Foxp3 in tissue infiltrate region . After treatment there was a substantial reduction in Foxp3+ cells and cell infiltrates . Representative examples are illustrated in Figure 3 . We investigated preponderance of localized nTregs and IL-10 in nodular tissues compared to macular or papular clinical manifestations . mRNA analysis showed enhanced CD25 , Foxp3 and IL-10 mRNA levels in nodular tissues compared to other forms ( CD25 & Foxp3 , P<0 . 001; IL-10 , P = 0 . 006 ) ( Figure 4 ) .
During infection , a precise controlled immune response is desired to protect the host from invading parasites , preventing untoward immune responses and maintaining homeostasis . nTreg constitute one such arm of the regulatory network acting as a double edged sword , on one hand limiting inflammation mediated pathology , at the same time promoting parasite replication [8] . Recently we have reported a rich parasite burden in nodular tissue of PKDL patients compared to other forms . Further , we also showed positive correlation between circulating IL-10 and parasite burden in human VL [5] . Here we provide evidence for accumulation of nTreg in PKDL lesion tissues and demonstrate a direct correlation between CD25 , Foxp3 and IL-10 mRNA levels with parasite load . At the pre-treatment stage the expression of nTreg cells surface markers ( CD25 , CTLA-4 and Foxp3 ) was found elevated in lesion tissue compared to control , implicating a role of nTreg in PKDL . Reports on human CL have documented a possible role of intra-lesional Treg cells for local control of effector T cell functions and correlation with drug unresponsiveness [9] , [10] . In contrast to PKDL , human VL was not linked with nTreg accumulation in blood or spleen , nor was antigen-specific IFN-γ response rescued following depletion of CD25+ cells [6] . Possible discrepancy between VL and PKDL could be context dependent due to ( i ) different niche and clinical manifestation; ( ii ) infection induced inflammation ( iii ) presence of Treg inducing and proliferating factors . Of the human CD4+ T cells , approximately 30% cells express CD25 . 1–3% of these express CD25 at high levels ( CD25++ ) that possess suppressor activity [11] . The present study demonstrated elevated mRNA expression of CD25 in lesion tissue of patients and the level subsided after treatment , suggesting accumulation of Treg in lesion tissues supported by immunohistochemical identification of Foxp3+ cells . In Sudanese PKDL , scanty CD25+ cells in tissue biopsy are documented [12] , contrary to this we noticed enhanced CD25 expression in all patients , which could be due to differences in the ethnic composition of the two populations or strains of the parasites . Thus , the finding suggests the variation in regulatory cells population according to the immune environment or the degree of inflammation and also suggests a distinct PKDL pathology in comparison with Indian VL or Sudanese PKDL . The correlation between advent of host immune responses and parasite persistence has been demonstrated in various Leishmania infections [5] , [13] , [14] . Analysis of Foxp3 and CD25 mRNA levels and parasite load showed direct correlation in lesion tissues . Furthermore , nodular lesions have rich parasite burden compared to other forms , indicating that chronic infection in the skin might have generated a population of Treg cells that have influence on parasite propagation and the level of immunity . Numerous recent observations have shown influence of nTreg on functional immunity against several microbes including human malaria parasite [15] . Similar correlation with parasite load was lacking for CTLA-4 , although mRNA was enhanced at pretreatment stage . CTLA-4 ( CD152 ) is expressed on activated CD4+ and CD8+ T cells and binds to the costimulatory ligands , B7-1 ( CD80 ) and B7-2 ( CD86 ) , with a 20-fold higher affinity than CD28 [16] , [17] . CTLA-4 expression is not detected on naive T cells , but transcriptionally induced after T cell activation [18] . CTLA-4 can out-compete CD28 for binding with co-stimulatory ligands , especially when these molecules are limiting , and low level of CD28 expression on circulating T cells is reported in PKDL [19] . Thus , in the context of enhanced CTLA-4 at pretreatment stage , data suggests inhibitory local signals in lesion tissues . In murine VL , blockage of CTLA-4 results in beneficial effect , in the form of low parasite burden and increment in the frequency of IFN-γ and IL-4 producing cells in spleen and liver [20] . Further studies are needed to understand the extent to which CTLA-4 contributes its effect either on regulatory T cells or activated Th1 cells or on both and if blockage of CTLA-4 has any influence on functional immunity in human PKDL . IL-10 is repeatedly implicated as an immunosuppressive factor in both human and experimental leishmaniasis . The role of IL-10 in KA and PKDL is well documented [4] , [5] and a consistent correlation between IL-10 levels and the development of PKDL in Sudanese KA patients has been established [21] . Like nTreg markers , IL-10 mRNA level was enhanced in PKDL lesion tissues at pretreatment stage , the level subsided following treatment , similar to previous observations [4] , [6] . Additionally , like Foxp3 and CD25 , IL-10 mRNA level directly correlated with parasite burden , and was high in nodular lesions compared with other form , probably due to preponderance of nTreg and other cellular infiltrates in nodular lesions . Furthermore , in mice IL-10 is found to be important for the maintenance of Treg activity [22] . Immunopathology of different clinical manifestations of PKDL is not well understood . There is higher cellular infiltration including T cells and plasma cells in nodular PKDL compared to macular or papular forms [23] , [24] . In addition , we have recently demonstrated [5] , and also observed in the present study , a higher parasite burden in nodular lesions compared to macular or papular lesions . Abundance of nTreg with IL-10 in tissue infiltrates of nodular form , observed in the present study , may be the driving factors for high parasite burden resulting in disease aggravation . Collectively the data suggests a possible role of Treg and IL-10 in parasite establishment in PKDL patients . Because nTreg have been shown to produce IL-10 and TGF-β associated with immune suppression in experimental systems , studies are warranted to explore antigen specific IL-10 source in PKDL lesion tissues . Furthermore functional studies are required to support the association of Treg and immunosuppression in PKDL . Such findings will lead to new targets for immunotherapeutic or vaccine strategies against PKDL , important from the perspective of parasite reservoir/transmission and a barrier towards VL eradication .
|
Post kala azar dermal leishamniasis ( PKDL ) , an unusual dermatosis develops in 5–15% of apparently cured visceral leishmaniasis cases in India and in about 60% of cases in Sudan . PKDL cases assume importance since they constitute a major human reservoir for the parasite . Inadequate treatment of VL , genetics , nutrition and immunological mechanisms that allow renewed multiplication of latent parasites or reinfection predispose to PKDL . Immunopathogenesis of PKDL is poorly understood . IL-10 is widely accepted as an immuno-suppressive cytokine and produced by diverse cell populations including , B cells , macrophages and CD4+ T cells . Natural T regulatory ( nTreg ) cells are subpopulation of CD4+ T cells that inhibit the response of other T cells . In this study we reported the accumulation of nTreg cells in lesion tissues of PKDL patients . Further correlation of Treg markers and IL-10 with parasite load in lesion tissues suggested a role of IL-10 and Treg in parasite establishment or persistence . Further studies are warranted to explore antigen specific IL-10 source in lesion tissues and unravel the concerted induction or accumulation of Treg in PKDL .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"immunopathology",
"immunology",
"biology",
"microbiology",
"immune",
"response",
"pathogenesis"
] |
2011
|
Foxp3 and IL-10 Expression Correlates with Parasite Burden in Lesional Tissues of Post Kala Azar Dermal Leishmaniasis (PKDL) Patients
|
Deciphering the specific contribution of individual motifs within cis-regulatory modules ( CRMs ) is crucial to understanding how gene expression is regulated and how this process is affected by sequence variation . But despite vast improvements in the ability to identify where transcription factors ( TFs ) bind throughout the genome , we are limited in our ability to relate information on motif occupancy to function from sequence alone . Here , we engineered 63 synthetic CRMs to systematically assess the relationship between variation in the content and spacing of motifs within CRMs to CRM activity during development using Drosophila transgenic embryos . In over half the cases , very simple elements containing only one or two types of TF binding motifs were capable of driving specific spatio-temporal patterns during development . Different motif organizations provide different degrees of robustness to enhancer activity , ranging from binary on-off responses to more subtle effects including embryo-to-embryo and within-embryo variation . By quantifying the effects of subtle changes in motif organization , we were able to model biophysical rules that explain CRM behavior and may contribute to the spatial positioning of CRM activity in vivo . For the same enhancer , the effects of small differences in motif positions varied in developmentally related tissues , suggesting that gene expression may be more susceptible to sequence variation in one tissue compared to another . This result has important implications for human eQTL studies in which many associated mutations are found in cis-regulatory regions , though the mechanism for how they affect tissue-specific gene expression is often not understood .
Gene expression is initiated by the binding of transcription factors ( TFs ) to cis-regulatory modules ( CRMs ) such as enhancer elements , which give rise to specific patterns of temporal and spatial activity [1] . Recent years have seen a dramatic increase in the ability to identify the location of regulatory elements using genome-wide information on TF occupancy [2] , [3] , [4] , [5] , [6] , [7] , cofactor recruitment [8] , [9] and chromatin modifications [10] , [11] , [12] , [13] , [14] , [15] . These studies have identified thousands to hundreds-of-thousands of regulatory elements that could be potentially active at a given time and in a given cell-type during development . Given this extensive regulatory landscape , it has become an enormous challenge to decode how CRMs function in terms of their spatio-temporal activity . Detailed dissection of individual elements has revealed developmental enhancers whose function is dependent on the presence of individual TF motifs [16] , [17] , combinations of motifs [18] , [19] , [20] , and the specific arrangement [21] , [22] , [23] , [24] , [25] , [26] or not [2] , [27] , [28] , [29] of those motifs . These examples have inspired much debate over the relative importance of each of these variables to enhancer function , but systematic rules to understand their contribution have not emerged . While dissection of endogenous enhancers has proved to be an extremely powerful approach to understand enhancer function [19] , [20] , [29] , [30] , [31] , [32] , individual enhancers instantiate only one of many possible solutions that can give rise to a specific gene expression pattern [7] , [28] , [33] , thereby limiting the range of functional rules that are generally explored . Synthetic elements offer the possibility to examine a wide range of motif compositions and motif positioning rules while ensuring , as much as possible , the neutrality of non-motif ( i . e . spacer ) sequences . Synthetic promoter-YFP libraries in yeast , for example , were used to quantify and model the effect of different promoter motif configurations on the levels of YFP expression [34] , [35] . Similarly , synthetic constructs combined with massively parallel sequencing were used to dissect the relationship between DNA sequence and activity of constructs transiently transfected into cell lines [36] or injected into mouse tail veins [37] . These approaches offer the clear advantage of scale , as thousands of elements with different DNA sequence combinations can be examined simultaneously . However , they are limited by the simplicity of the read-out , which is a relative measure of the levels of the CRM's activity at a single time point or condition . For most developmental enhancers , the impact of sequence changes on the timing or tissue-specificity of gene expression is equally important . It is also not clear to what extent episomal DNA from transient transfection or tail-vein injections recapitulates the impact of chromatin context and nucleosome positioning on gene expression . In these respects , reporter assays in transgenic animals provide invaluable information . Although generally difficult to scale , multi-stage assays using stable transgenic embryos yield precise information about when and where an enhancer is active in an in vivo chromatinized context . In this study , we systematically engineered 63 synthetic elements that differ with respect to the number and kinds of TF motifs they contain , as well as the relative spacing and orientation of these motifs . These elements were specifically designed to assess three properties of enhancer motif organisation relevant to metazoan development: ( 1 ) the ability of homotypic clusters of individual motifs to function as developmental enhancers , ( 2 ) the ability of combinations of different kinds of motifs ( heterotypic motif clustering ) to generate new emergent activity , and ( 3 ) the effect of changes in motif organisation , such as number , spacing and orientation , on the robustness of enhancer activity in both space and time . While many studies have focused on the effect of point mutations on enhancer activity , there is an enormous amount of structural variation within natural populations , including the sequence of Drosophila [38] , [39] and humans [40] . To assess the influence of small insertions or deletions , we have systematically changed the spacing between motifs for four heterotypic pairs of TF motifs . To examine these properties , we focused on ten motifs recognized by TFs that form part of a highly studied regulatory network that governs Drosophila mesoderm development [41] , [42] , including the downstream effector TFs of Wingless ( known as Wnt in vertebrates ) and Dpp ( BMP in vertebrates ) signaling pathways . We generated very simple elements consisting of six motifs in either a homotypic or heterotypic combination , which were stably integrated into the Drosophila genome . The resulting patterns of expression driven by these elements suggest a number of interesting features governing CRM function . First , despite the known importance of combinatorial activity to refine spatial patterns of expression , elements with multiple copies of an individual motif can drive complex patterns of expression . For example , although pMad is known to act cooperatively with lineage TFs in both Drosophila and vertebrates , it is sufficient to transduce Dpp activity when its motifs are in the appropriate configuration . Second , combining motifs for as few as two TFs can lead to novel emergent expression . Although the importance of cooperative DNA binding in the regulation of development has long been supported by other studies , the sufficiency of so few sites speaks to the extent to which even minor changes in cis-regulatory sequence can lead to the evolution of novel expression profiles . Third , the spacing and orientation of motifs is not only essential for enhancer activity in terms of binary on-off effects , but also has more quantitative effects on the robustness of gene expression , including inter- and intra- embryo variability . Fourth , these effects of motif organization , often referred to as motif grammar , are tissue-specific . The same ‘two-TF’ enhancer can function using very flexible motif spacing in one tissue , yet have rigid constraints in another , demonstrating an additional way in which the organization and function of CRMs acts to reduce the constraints of pleiotropy for regulatory mutations .
For each factor , synthetic CRMs were generated by combining six motifs , separated by a spacer sequence of defined length . Therefore , there were two aspects to the design of the synthetic elements; the choice of motif instance used for each TF and the sequence of the spacer . First , for the TF motifs , we selected high affinity motifs for each factor , as much as possible . TFs often recognize the same or highly similar sequences . This includes not only members of the same family of TFs , e . g . GATA factors ( such as Pnr ) , but also TFs with apparently unrelated DNA binding domains , e . g . the bHLH factor Twist and the zinc finger TF Snail . While we cannot change this inherent property of TFs , we did try to increase the potential specificity of the motif ( or word instance ) used here for the particular TFs we are interested in by using motifs derived from in vivo occupancy data for nine of the ten factors [2] , [7] . For clarity , we refer to each construct by the name of the TF from which ChIP data was used to learn the motif . However , the sequence of all motif instances used , as well as their similarity to other TF motifs is provided in Table S1 . Second , we tried to select a neutral spacer sequence that does not include known TFBS , based on our current knowledge . This is not trivial , as in addition to the spacer sequence itself , the border sequence bridging the spacer and the known motif ( red bar , Figure 1A ) can in itself create an additional binding site . To minimize this possibility , rather than using a common spacer , we computed an optimal spacer sequence for each combination of motifs to minimize the chance of inadvertently generating additional binding sites , based on current information ( Supplemental Methods ( Text S1 ) ) . During the course of this study , our results show that the spatio-temporal activity of the designed elements is primarily driven by the intended TF motifs and not from the spacer sequence , as indicated by two lines of evidence: ( 1 ) The concordance between the activities of homotypic and heterotypic elements for the same TFs , which were designed with different spacer sequences ( described below ) and 2 ) the very similar spatio-temporal activity of two heterotypic CRMs , which were specifically designed with identical TF motifs but with two different spacer sequences ( pMad-Tin , see below ) . Each synthetic CRM , which were on average 105 bp in length , was cloned into a common minimal lacZ reporter vector and stably integrated into the same location in the Drosophila genome using the phiC31 system [43] , to allow for a direct comparison of enhancer activities . The ability of each CRM to drive spatio-temporal lacZ expression was assessed during all stages of embryogenesis by fluorescent in situ hybridization to examine the full regulatory potential of the DNA sequence . The combinatorial binding of TFs provides complexity to regulate refined spatial patterns of expression , and forms the basis for logical operations within Gene Regulatory Networks driving development [44] . However , in addition to combinatorial activity , homotypic clusters of an individual TF's motif are also present in regulatory regions in the vicinity of developmental genes in both Drosophila [45] and vertebrates [46] . Although prevalent in vivo , the role of these clusters in regulating gene expression , and the properties governing how they function , is currently not clear . Studies examining the ability of clusters of motifs to regulate gene expression in transgenic reporter assays have had varied success: this includes ( although not exhaustive ) homotypic clusters of motifs that were sufficient to drive activity [47] , [48] , [49] , [50] , [51] , [52] and those that were not [48] , [53] , [54] , [55] , [56] . However , as the elements in each of these individual studies were designed and tested in different ways , it is impossible to deduce any functional inference across studies . We initiated this study by systematically examining the activity of elements composed of a cluster of six identical motifs . In total we tested homotypic CRMs with motifs for ten essential factors , encompassing multiple types of DNA binding domains: bHLH ( Twist ( Twi ) ) , homeobox ( Tinman ( Tin ) , Bagpipe ( Bap ) ) , T-box ( Dorsocross ( Doc ) ) , GATA zinc finger ( Pannier ( Pnr ) ) , MADS box ( Myocyte enhancer factor 2 ( Mef2 ) ) , FoxF ( Biniou ( Bin ) ) , HMG-domain ( T-cell factor ( TCF ) ) , Ets-domain ( Pointed ( Pnt ) ) and the MH1 domain ( pMad ) . Seven out of ten of these very simple elements were able to drive sequence-specific spatio-temporal expression ( Figures 1B–F , 2 , S1 ) . In six cases , the expression profiles driven by clusters of a single motif were sufficient to partially ( and in the case of two , almost completely ) recapitulate the expression of the TF whose in vivo occupied motif was used to construct the CRM . For example , the synthetic CRMs containing Doc enriched and GATA motifs drove expression in the presumptive amnioserosa , cells where the endogenous Doc and pnr genes are expressed ( white boxes , Figure 1B″ , C″ ) . Similarly the synthetic CRM containing Tin motifs drove expression in a subset of the dorsal mesoderm , colocalizing with the expression of the endogenous tin gene ( Figure 1D″ , arrow ) . Although multimerizing the preferred Twist E-box motif was not sufficient to drive mesoderm expression at early stages of development , this synthetic CRM could activate expression in the hindgut visceral mesoderm ( VM ) at later stages , overlapping the endogenous twist gene's expression ( Figure 1E″ , arrow ) . The activity of two homotypic CRMs , containing motifs for either pMad ( discussed below ) or Bin , stood out as they were sufficient to recapitulate almost the entire domain of the TF's activity . For the Bin CRM , this included the foregut , midgut and hindgut VM ( Figures 1F″ , S2 ) through multiple stages of development . This result suggests that motifs for this FoxF factor are sufficient to regulate expression throughout the VM at multiple stages of development , consistent with Bin's essential requirement for VM development [57] and extensive enhancer occupancy [7] , [58] . In contrast to the six CRMs that gave ‘expected’ activity , the activity of the homotypic dTCF CRM only partially overlaps wingless expression , ( the ligand that activates the cascade leading to dTCF activation ) . The CRM's activity is restricted to segmental groups of cells that are in close proximity to , but not always adjacent to , the wg stripes ( Figure S1A″ ) , which may reflect more ‘long’-range signaling from Wg , or alternatively the activity of an additional TF that can occupy this dTCF motif . Taken together , these results indicate that homotypic clusters of an individual motif are often sufficient to regulate specific patterns of spatio-temporal activity , analogous to the more commonly studied combinatorial elements . Multimerizing single TF motifs can provide remarkable specificity , as demonstrated by the non-random overlap of CRM activity with part or all of the TF's expression in six out of ten cases tested . Conversely , it is interesting that not all activator clusters are sufficient to drive expression , in contrast to what might be assumed from the activity of yeast GAL4 sites . For example , three synthetic CRMs ( multimers of Pnt , Mef2 and Bap motifs ) yielded no activity ( Figure S1B–D ) . In the case of Mef2 and Twist , the lack of general mesoderm activity is surprising given that both TFs have very well characterised DNA binding specificities in both Drosophila and vertebrates . This diversity in CRM output may reflect inherent differences in the regulatory potential among different TFs or in their ability to act cooperatively in a homotypic manner to potentiate their activity . The dpp gene , coding for the ligand of the Dpp signaling cascade , is expressed in the foregut , hindgut and midgut VM in parasegment 3 ( PS3 ) and 7 ( PS7 ) at stage 13/14 of embryogenesis [59] . dpp expression is restricted to these VM domains through the integration of activating inputs from Ubx and Bin [57] , [59] , and repression via Wg signaling [60] . The downstream effector TF of Dpp signaling is the phosphorylated form of Mad ( pMad ) . Our synthetic CRM containing six pMad motifs could recapitulate almost the entire expression of dpp in the VM during these stages of embryogenesis ( Figure 2A ) . Therefore , despite the complexity in the network of upstream factors regulating dpp expression , once dpp is expressed , sites for its downstream effector ( pMad ) alone are sufficient to activate enhancer activity in these sub-tissue domains . The spatial boundaries of the enhancer's activity are most likely refined through the action of Brinker ( Brk ) . Brk is a transcriptional repressor whose expression is negatively regulated by Dpp signaling [61] . This results in cells with high levels of pMad and low Brk ( Dpp responding cells , where our synthetic CRM is active ) and those with high Brk and low pMad ( neighbouring cells outside the Dpp signaling domain , where our CRM is inactive ) . As Brk and pMad can recognize the same motif [62] , there is often direct competition between the two TFs to regulate enhancer activity , where the relative levels of both TFs serve to limit the spatial domain of Dpp target gene expression , as nicely demonstrated for the endogenous Ubx enhancer [62] . The sufficiency of pMad motifs alone to activate enhancer activity was unexpected given previous studies in both Drosophila [56] and vertebrates [63] , [64] , indicating that Mad proteins bind to enhancers cooperatively with other tissue-specific ‘lineage’ factors , and have little activity alone . We therefore further investigated the regulatory potential of pMad sites in isolation by directly comparing the activity of elements containing six , four or two motifs , in the same orientation and with the same spacer sequence ( Figures 2 , S3 ) . The expression of lacZ in PS7 of the midgut VM ( Figure 2 ) is particularly informative of pMad activity: the Dpp signaling cascade activates pMad in an autocrine [65] and paracrine [66] fashion , which likely results in the highest levels of pMad activation in VM cells in PS7 and lower levels in adjacent parasegments . The homotypic CRM containing six pMad motifs reflects this paracrine signaling , having a larger domain of expression covering neighboring cells compared to the dpp expressing cells ( Figure 2A ) . In contrast , the synthetic CRM containing four motifs had more restricted activity to a narrow domain anterior and posterior to the dpp expressing cells ( Figure 2B ) , while the CRM containing only two motifs was active only in the Dpp producing cells ( autocrine signaling ) ( Figure 2C ) . Therefore , pMad sites alone , when present in close proximity , are sufficient to activate enhancer activity in the absence of specific lineage TFs . However , the extent of the activity is dependent on the number of available pMad motifs and on the balance between pMad and Brk concentration . In addition to the number of cells in which the enhancer was active , we also observed a strong correlation between the number of sites and the strength of the CRM . In the amnioserosa , for example , the CRM with six pMad sites drove robust expression at stage 11 ( Figure S3A″ ) , while four sites resulted in reduced activity , and the CRM with two sites drove only very weak expression in the amnioserosa ( Figure S3B , C ) . This trend was even more dramatic at stage 14 , when the pMad concentration in the amnioserosa decreases [67] . While the CRM with six pMad sites drove strong activity at stage 14 ( Figure S3D ) , the CRMs with either four or two sites were inactive in the amnioserosa at this stage . The number of pMad sites , thus , appears to be able to compensate for a decrease in the amount of accessible pMad protein at stage 14 , presumably by providing a larger platform for cooperative binding . Cooperative interactions between TFs can sometimes occur through direct protein-protein interactions ( PPI ) , which may introduce constraints in the organization of the TFs' motifs within CRMs [1] . Taking advantage of the design flexibility of synthetic elements , we generated heterotypic CRMs composed of motifs for TFs known to exhibit protein-protein or genetic interactions: namely motifs for Tin with either Pnr , Doc , dTCF , or pMad [68] , [69] , [70] , [71] , to systematically explore two properties: 1 ) The influence of changes in the relative spacing and orientation of motifs on the robustness of CRM activity in different tissues , and 2 ) The ability of different combinations of TF motifs to generate emergent expression patterns in a particular tissue , not observed for CRMs containing only one kind of motif . Each heterotypic CRM contained three pairs of TF motifs for either Doc-Tin , GATA-Tin , dTCF-Tin , or pMad-Tin . For each TF pair , on average nine constructs were tested in transgenic animals , in which the spacing and/or orientation of the Tin motif was systematically altered ( Table S2 ) . Heterotypic CRMs containing GATA-Tin or Doc-Tin combinations resulted in a pattern of expression that was largely the sum of the activity of each of the corresponding homotypic CRMs alone ( data not shown ) . Changing the spacing and orientation of the sites ( Table S2 ) had little or no effect on CRM activity , indicating a minimal role for motif positioning , or grammar , in these instances . The dTCF-Tin heterotypic CRMs did not have any activity , despite the fact that the homotypic CRMs containing six dTCF or Tin sites drove specific activity in segmental groups of cells in the ectoderm and mesoderm , respectively ( Figures 1D , S1A ) . This lack of activity in the heterotypic context most likely reflects the reduction in the number of motifs used , from six in the homotypic constructs to three motifs for each factor in the heterotypic CRM . Correspondingly , homotypic construct containing four dTCF sites [53] , [54] had no embryonic activity . In contrast to the GATA-Tin and Doc-Tin heterotypic pairs , whose activity was robust to changes in motif organisation , the activity of the pMad-Tin heterotypic CRMs changed depending on the motif spacing . As discussed above , pMad motifs alone are sufficient to regulate expression in PS3 and PS7 of the midgut VM . In these cases , the strength of activity was dependent on the number of motifs present , with a dramatic reduction in expression observed going from four to two sites ( Figure 2 ) . A homotypic CRM containing three pMad sites separated by 13 bp also showed a dramatic reduction in VM activity ( Figure 3A ) , a result that may stem from either or both the reduction in sites or the increase in spacing ( from 6 bp to 13 bp ) . Interestingly , when a Tin motif is inserted within each 13 bp ‘spacer’ sequence , this pMad-Tin heterotypic design can restore activity in the midgut VM ( Figure 3B , white square ) , giving a pattern of expression similar to the pMad homotypic CRM with 6 bp spacing . The ameliorative effects of the Tin sites were limited by distance . Increasing the spacing between adjacent pMad and Tin motifs from 2 to 8 bp , which increases the spacing between pMad motifs from 13 to 25 bp , caused a progressive reduction in lacZ expression in the midgut VM ( Figure 3 ) . This VM activity is primarily driven through the pMad and Tin motifs and not the spacer sequence , as changing the spacer sequence had a minimal effect on VM enhancer activity , while a similar distance effect was observed by altering the length of spacer between the two sites ( compare Figures 3 and S5 ) . These results suggest that the occupancy of Tin sites acts either to bridge the distance between pMad sites ( up to ∼20 bp ) facilitating cooperative pMad regulation or alternatively to facilitate pMad-Tin combinatorial regulation of VM expression . To distinguish between these two possibilities , we examined the effect of altering the relative orientation of the sites , reasoning that if Tin occupancy acts indirectly to help pMad recruitment altering the orientation of the Tin site should have little effect on activity . However , we observed that changes in motif orientation did influence CRM activity: one orientation of the Tin motif ( arbitrarily referred to as antisense ( A ) ) typically drove stronger VM expression compared to CRMs with the motif in the opposite orientation ( referred to as sense ( S ) ) ( Figure 3 , compare C to G ) . This result indicates that the Tin motifs are not just neutral sequences bridging the spacing between homotypic pMad motifs , but rather suggest a specific mechanism of cooperative Tin-pMad DNA binding , and therefore transcriptional regulation . Combining pMad and Tin sites resulted in CRM activity in the dorsal mesoderm , VM and amnioserosa –representing essentially the summation of the expression profiles of the respective homotypic CRMs ( Figures 1D , 2A , S4 ) . However , in contrast to GATA-Tin and Doc-Tin CRMs , whose joint expression profiles were largely additive , the heterotypic pMad-Tin CRMs were also sufficient to drive expression in a new domain , the developing heart ( Figure 4 ) . Cardioblast specification requires the action of a large number of TFs and signaling cascades ( for review , see [72] ) . Tin , Pnr ( GATA factor ) , and Doc together with Dpp and Wg signaling are essential for heart development [73] , [74] , [75] , [76] , [77] , [78] , a tissue in which these factors regulate each others' expression [70] , [75] , [76] , [77] , [78] , [79] , [80] , and act as a collective unit to regulate enhancer activity [2] , [27] . Given this complex regulation , we did not anticipate that a simple element containing only pMad and Tin sites would be able to drive expression in the heart . Importantly , the emergence of this pattern was highly dependent on the CRM architecture: Among our constructs , heart activity was only observed in CRMs containing three pMad and three Tin motifs placed in close proximity ( within 2–4 bp ) to one another ( Figure 4B–E ) . In contrast , no heart activity was observed in CRMs where the spacing between motifs was increased to 6 bp or 8 bp . Even in the ‘optimal’ motif configuration , we observed embryo-to-embryo variability , indicating that expression in the heart is significantly less robust ( i . e . more influenced by stochastic events ) than is expression driven by pMad-Tin constructs in the VM and other tissues . To better assess the robustness of these constructs activity , we made use of P-element transgenesis to integrate our CRMs into random locations in the genome . We reasoned that non-robust CRM activity would be more highly influenced by variation in chromatin context at different genomic positions compared to more robust expression profiles . Across these random-insertion sites , the pMad-Tin CRMs drove highly consistent expression in the VM , amnioserosa , and dorsal mesoderm , while heart activity in those same embryos varied dramatically as a function of genomic location ( Table S3 ) . In fact , the pMad-Tin A2 CRM was only active in the heart in the context of a transgenic fly line obtained using a P-element ( Figure 4B ) , and not with the phiC31-mediated integration ( Figure 3B ) . These results indicate that the heart activity of pMad-Tin CRMs is teetering on the edge of activation , being highly sensitive to both the motif context within the enhancer and the enhancer context within its chromatin environment . Thus , while the combined activity of these two TFs can give rise to emergent activity in the developing heart , this is not a robust mechanism to generate heart expression . Examining CRM activity in the heart revealed that some CRMs exhibit varied activity among embryos , and even within an embryo , as demonstrated by the pMad-Tin S4 construct , which drove expression throughout the entire heart in some embryos ( Figure 3G ) , but only in a posterior portion of the heart in others ( Figure 4E ) . To assess this variability in a systematic manner , and to provide quantitative data to which we could fit a model explaining CRM activity ( see below ) , we applied two measures of the robustness of CRM activity: ( 1 ) Penetrance , defined as the fraction of embryos within a population that show CRM activity in the relevant tissue ( in this case the VM or heart ) , and determined by any spatial overlap of lacZ expression with dpp ( VM ) or tin ( heart ) at a defined developmental stage ( Figures 5A , S6A ) . ( 2 ) CRM expressivity , defined as the fraction of tissue-specific regions within an embryo ( i . e . the proportion of the VM or heart ) that display CRM expression ( Figures 5A , S6A ) . In the midgut VM , for example , there are four domains of CRM activity ( Figure 5A ) . If the CRM is active in all four domains , it has an expressivity of 1 , while activity in two out of four domains has an expressivity of 0 . 5 . To minimize systematic error , we implemented automated image analysis ( see Materials and Methods ) of embryos at a consistent stage of development ( Figures 5A , S6A ) . In using penetrance and expressivity as quantitative metrics of CRM activity , we avoided issues arising from directly comparing non-linearly amplified signals from standard in situ hybridization of lacZ levels between CRMs . Penetrance provides a reliable measure of enhancer embryo-to-embryo variability , while expressivity provides a readout of intra-embryo enhancer variability – both of which we exploit to assess the effect of motif organization on the robustness of activity in two tissues . The CRM penetrance of around one hundred embryos was measured for each of the 8 pMad-Tin synthetic CRMs ( Table S4 ) . With this quantification , differences in activity became more striking in several regards . First , within a tissue , subtle differences in activity between CRMs with different motif configurations become clear . For example , in the antisense orientation , changing the motif spacing from 2 bp to either 4 bp or 6 bp spacing resulted in only a slight decrease in penetrance , from 1 to 0 . 91 in the VM . However , the expressivity of these CRMs was quite distinct . The CRM with a 4 bp spacing had an expressivity of 1 , while this number was nearly halved ( 0 . 59 ) for the CRM with a 6 bp spacing ( Figure 5B; Table S5 ) . The extreme effect of a small ( 2 bp ) change in motif spacing suggests that direct , and potentially cooperative , interactions among bound factors have been disrupted , leaving the enhancer's activity more prone to variation . A second striking observation is the extent to which changes in CRM architecture impacts activity in a tissue-specific manner . This is made clear by relative differences in CRM penetrance between the VM and heart . In the antisense orientation , for example , CRM penetrance remains almost unchanged in the VM when the motif spacing between the pMad and Tin sites is changed from 2 to 6 bp ( Figure 5B ) . In contrast , the penetrance in the heart drops from 0 . 89 at a 4 bp spacing to zero when the motifs are spaced by 6 bp . A similar dramatic effect was observed by flipping the orientation of the Tin motif from antisense to sense , at a 4 bp spacing , which caused the penetrance of heart activity to drop from 0 . 89 to 0 . 48 ( Figure 5C ) . These switch-like transitions in penetrance indicate that heart activity can only occur with a very restricted motif organization , which relies on close proximity of all TF motifs . Taken together these results highlight an important property of cis-regulatory activity in multicellular organisms: An enhancer element ( e . g . composed of only two types of motifs as described here ) can require a very restricted motif configuration to regulate expression in one tissue ( heart ) , but yet be much more flexible in its motif organisation to drive robust activity in another ( VM ) . The extent to which cooperative interactions , including higher-order interactions across multiple proteins , contribute to enhancer activity is difficult to assess by simply visualizing expression patterns . Here , computational models can be extremely helpful to explore the effect of potential interactions on CRM function [33] , [81] . To better understand the contribution of higher-order interactions among TFs on our synthetic CRMs , we used fractional site occupancy modeling ( Figures S7; Methods ( Text S1 ) ) , which describe DNA-protein and protein-protein interactions as thermodynamic processes , an approach that has been successfully used to understand other regulatory elements [82] , [83] , [84] , [85] , [86] , including Drosophila enhancers [26] , [87] . As fractional site occupancy models analyze the probability of every possible configuration of binding events , the complexity of the models increases exponentially with the number of TFBSs . To avoid these complications , we aimed to identify the simplest model that recapitulates the observed CRM activity . In line with our observations ( as seen in Figures 3 and 5 ) , and with research in another tissue [71] , the model assumes the presence of direct cooperative interactions between neighbouring pMad and Tin proteins . Protein-protein interactions were modeled as the extent of overlap between spheres of “interaction space” around each bound protein , with sense and antisense orientations having different effective spherical radii ( Figure S7 , and includes parameters for the strength of possible cooperativity between bound TFs ( Supplemental Methods ( Text S1 ) ) . The model specifically explores whether additive interactions between Tin-pMad pairs are sufficient to recapitulate the observed experimental data or if potential ‘higher order’ cooperativity among pMad-Tin pairs with nearby pMad or Tin bound proteins are required to drive robust CRM activity . For VM activity driven by the pMad-Tin heterotypic CRMs , a model that includes an additional degree of cooperativity beyond pMad-Tin pairs fit the data better compared to a model in which pMad-Tin pairs act in an independent additive manner ( Figures 6A , S8A ) . A key difference between the models lies in their predictions of the robustness of shorter CRMs: if higher order interactions are central for robust CRM activity then shorter CRMs will have a sharper decrease in robustness compared to pMad-Tin interactions alone . To experimentally test this , we halved the size of two heterotypic CRMs , generating CRMs with pMad-Tin-pMad motifs in S2 and A4 configuration . Such small CRMs could drive expression in the midgut VM ( Figure 6B , C ) , albeit at a reduced level . In contrast , CRMs with only one pMad and Tin site , representing the smallest possible cooperative binding configuration ( a pMad-Tin pair ) , drastically reduced all VM activity ( Figure 6D , E ) . Incorporating higher level cooperativity into the model , without any further fitting , significantly improved the quality of the prediction of the shortened CRMs ( Figures 6F , G , S8B ) . This suggests that pMad-Tin-pMad is the minimal configuration essential for robust VM expression . Importantly , this model was robust to a ‘leave-two-out’ cross validation iterated over all possible orderings of the 12 CRMs , arguing against over-fitting ( Figure S8C ) . Finally , to test the model's prediction that additive interactions between pMad-Tin pairs with other bound pMad or Tin proteins are insufficient to drive robust VM expression , we tested the activity of a CRM with the motif configuration Tin-pMad-Tin . This CRM had very weak VM activity ( Figure S8D ) , consistent with a requirement for higher order interactions between multiple pMad-Tin pairs for robust CRM activity ( Figure 6G ) . In summary , increasing the complexity of TF cooperativity resulted in significantly improved consistency with experiment compared to considering only independent pMad-Tin cooperative pairs ( Figure 6G ) . Next , we addressed how heterotypic pMad-Tin CRMs lead to activity in the heart . Simple models with cooperativity between bound Tin and pMad can recapitulate the observed penetrance and expressivity in the heart for CRMs ( Figures 6H , S8E , F ) . However , only the model including higher-order cooperative interactions between three neighboring units of Tin-pMad-Tin was consistent with both the shorter constructs and the six TF motif CRMs ( Figures 6H , S8E , F ) . We note that the true minimal motif arrangement to generate robust heart activity is likely to be much more complex . In line with this , the Tin-pMad-Tin CRM has no heart activity . Interestingly , the effective range of cooperative TF interactions learned by the model for the heart was considerably lower ( <5 bp ) than for the VM ( <9 bp ) ( Table S6 ) . In summary , while our experimental data is suggestive of cooperative TF interactions being likely necessary for CRM activity , the modeling has formalized this and systematically identified the range of interactions and the likely minimum level of higher-order TF cooperativity required for activity in both the VM and heart . Taken together , the modeling provides regulatory rules that explain how the same two TF motifs can give rise to activity in two different tissues ( VM and heart ) depending on the motif organization within the CRM .
While quantifying the activity of a simple ‘two-TF motif’ CRM ( pMad-Tin ) , our results show that enhancer activity can exhibit very different sensitivity to motif organization in one tissue compared to another ( Figure 7 ) . Several mechanisms could account for this interesting effect , including different concentrations of the TF ( i . e . pMad or Tin ) in the different tissues , the availability of tissue-specific co-factors , or tissue-specific priming of the enhancer , which may increase the ease by which the enhancer is activated . An elegant dissection of the endogenous spa enhancer demonstrated that completely rearranging the relative order and spacing of TF binding sites could switch its cell type-specific activity from cone cells to photoreceptors in the eye [20] . In comparison , the changes in motif organisation introduced in our study were much more subtle such that the relative order of motifs was completely preserved . Yet only changing the spacing or orientation of motifs altered the robustness of enhancer activity in a tissue-specific manner . This result indicates that small insertions or deletions in CRMs , that do not affect the TF motifs themselves , could still have significant effects on gene expression in one tissue while having no effect in another . A study examining the activity of neuroectoderm enhancers between Drosophila species supports this model , where reduced spacing between Dorsal and Twist sites results in broader neuroectodermal stripes of CRM activity , while increased motif spacing resulted in progressively narrower stripes [88] . Studies of both endogenous enhancers and the synthetic CRMs described here provide compelling evidence that the exact positioning of motifs within CRMs is crucial for the robustness of their activity in one tissue , while it may be largely dispensable in another . Different cell types can therefore interpret the same motif content of a given enhancer in different manners . The Drosophila heart is composed of two cell types , cardioblasts and pericardial cells , each of which requires the integration of many regulatory proteins for proper specification and diversification [72] . A characterized pericardial enhancer , eve MHE , for example , contains pMad and Tin binding sites in addition to sites for dTCF , Twi , Ets proteins , and Zfh1 [17] , [89] . Given this complexity , it was surprising that a simple element built from pMad and Tin sites alone was sufficient to drive expression in the heart , albeit at a later developmental stage . Our analyses indicate that this activity is due to cooperativity binding between Tin and pMad , facilitated by a very specific motif arrangement . Using crystal structure data from close homologues of pMad [90] and Tin [91] , we modelled the two TFs interaction on DNA , using a similar range of motif spacing ( Figure S9 ) . This 3D structural model indicates that it is possible for the DNA binding domains of these two proteins to both bind to DNA at a 2 bp spacing and to physically interact at a 2 bp and 4 bp spacing , but not at 6 bp spacing . Although done by homologue mapping , this structural data is consistent with our functional analyses of CRM activity , and further supports direct DNA binding cooperativity between these two TFs . It is interesting to note , that although pMad and Tin sites are sufficient to drive expression in the heart from stage 13 to 14 ( when placed in a limited motif arrangement ) , nature appears to use other enhancer configurations to regulate this critical function . There are two important aspects to this finding . First , heart activity arising from CRMs containing pMad and Tin sites alone is not robust . The enhancers are on ‘the edge’ of activation , where subtle changes in motif positioning or enhancer location switch activity between embryos and within embryos . Second , endogenous enhancers that are bound only by pMad and Tin – with no known input from other factors – direct expression in the dorsal mesoderm and not in the heart , at stage 10 [2] , [92] . In our synthetic situation , pMad and Tin sites also drive robust expression in the dorsal mesoderm , in addition to variable weak expression in the heart . Therefore , although pMad and Tin sites alone are sufficient to drive heart activity in limited motif contexts , this mechanism is most likely not robust enough to be generally used to drive heart expression in vivo . This is consistent with recent studies showing that heart enhancer activity is elicited by the collective action of many TFs , which can occupy enhancers with considerable flexibility in terms of their motif content and configuration [2] , [27] . Our pMad-Tin synthetic elements uncovered a very simple , although not very robust , alternative mechanism to regulate heart activity , and represent a nice example of how combinatorial regulation can lead to emergent expression profiles more than the simple sum of its parts . The expression of key developmental genes is generally buffered against variation in genetic backgrounds and environmental conditions . This may occur at many levels including RNA polymerase II pausing [93] , [94] and the presence of partially redundant enhancers [95] , [96] , [97] , [98] . However , robust expression may also be buffered by the motif content within an enhancer to ensure a stable regulatory function . CRMs , for example , often include additional binding sites to those that are minimal and necessary [99] . In the context of the pMad-Tin synthetic CRMs , the motif organization can also act to ensure robust activity . Our results demonstrate that even in situations where the composition of motifs and their relative arrangement are maintained , subtle changes in the spacing between the motifs could have dramatic effects on enhancer output . Interestingly , this effect seems to be very tissue-specific , with some tissues maintaining robust activity whilst others lost all enhancer activity . Taken together , the data presented in this study demonstrate that subtle alterations in motif organization can affect the ability of different tissues to ‘read’ an enhancer , which in turn may allow each tissue to fine-tune enhancer activity based on fluctuations in its molecular components .
Binding affinity models ( PWMs ) for Twi , Tin , Mef2 , Bap , Bin , dTCF , pMad , and Doc2 and GATA were derived from ChIP-chip data analyses [2] , [7] . The model for Pnt was generated using published footprints ( Supplemental Methods ( Text S1 ) ) . PWMs were first trimmed on each side to remove positions with an information content ( IC ) of less than 0 . 4 ( trimming stopped at the first IC position > = 0 . 4 ) . The sequence that best fits the PWM model was then determined for each trimmed PWM and is referred to hereafter as “TFBS” . All TFBS sequences used to design the synthetic CRMs are available in Table S1 . For each CRM , a ‘neutral’ spacer sequence ( a linker sequence placed between motifs ) was heuristically determined by minimizing the sequence affinity for known TF PWM models ( Supplemental Methods ( Text S1 ) ) . Synthetic CRMs were generated from long oligonucleotides synthesized by Eurofins MWG Operon with compatible cohesive ends upon annealing for cloning . The forward and reverse strand oligonucleotides were phosphorylated , annealed and subsequently ligated into the pDUO2n [7] , to generate stable , transgenic Drosophila lines using the phiC31 site-specific integrase [43] . The pH-Pelican [100] vector was use to test the robustness of enhancer activity at different genomic locations by random P-element transgenesis . The sequence of each CRM was verified to ensure that there were no synthesis errors and is provided in Table S2 . CRM activity was assessed in embryos from transgenic flies using fluorescent in situ hybridization as described previously [101] . The following ESTs or full length cDNAs from the Drosophila Gene Collection ( DGC ) were used to generate probes: RE13967 ( bap ) , RE40937 ( doc2 ) , RE20611 ( dpp ) , GM04312 ( dTCF ) , SD02611 ( pnr ) and AT15089 ( twi ) . cDNAs used for bin and tin , lacZ , Mef2 and pnt were generous gifts from M . Frasch , U . Elling and M . Taylor respectively . Double or triple in situ hybridizations were performed using anti-fluorescein-POD , anti-DIG-POD and anti-biotin-POD antibodies ( Roche , 1∶2000 dilution ) and were developed sequentially with Cy3 , fluorescein , and Cy5 tyramide signal amplification reagents ( Perkin Elmer TSA kit ) . The lacZ expression patterns were imaged using Zeiss LSM 510 FCS or LSM 510 META confocal microscopes with A-Plan 10×/0 . 25 PH1 objective . Background subtraction of both the CRM and tissue-specific channels was performed using a morphological opening with disk size greater than the largest relevant VM region ( typical disk size of 25 pixel radius ) . The tissue-specific subset of images ( e . g . dpp or tin in situs ) were segmented using the Ilastik software package ( www . ilastik . org ) . The segmented images were analysed using Matlab . The segmented regions in each image were smoothened by performing dilation ( disk size of 5 pixel radius ) followed by equivalent erosion . An area threshold ( >200 pixels ) was used to remove small , segmented regions . Finally , the perimeter of the segmented regions was calculated ( using Matlab function bwperim ) and overlaid onto the CRM expression data . The penetrance was calculated for approximately 100 embryos for each of the twelve lines ( Table S4 ) . Bootstrapping was used to estimate the error in the penetrance measurements . The expressivity was calculated from around 16 carefully staged and positioned embryos ( based on morphology and markers for VM ( dpp ) and heart ( tin ) tissues ) for each line ( Table S5 ) . Embryos aligned dorsally were imaged and four regions of midgut VM were assigned , as shown in Figure 5A . The observed heart expression ( also viewed dorsally ) occurs in two rows of cells along either side of the embryo . We separated each row into an anterior and posterior segment , resulting in four heart regions ( Figure S6A ) . The posterior segments , to the right of the PS7 VM region , correspond roughly to the heart proper , while the defined anterior heart segments correspond roughly to the region often referred to as the ‘aorta’ . The penetrance in both the VM and heart was therefore measured as signal in one to four different regions of the tissue . A fractional occupancy model was used to analyze the experimental data [102] . Our methodology was similar to other thermodynamic models used to understand CRM activity in Drosophila ( e . g . [26] , [87] ) . The model had at most four parameters: two parameters described the relevant protein-protein interactions; and two parameters were used to distinguish sense and antisense binding effects . Mathematical details are provided in the Supplemental Methods ( Text S1 ) . This work is carried out in Drosophila , and was conducted in compliance with EMBL's guidelines .
|
Transcription is initiated through the binding of transcription factors ( TFs ) to specific motifs that are dispersed throughout the genome . Genomics methods have helped to discern which motifs for a TF are occupied and which are not , yet it is poorly understood why certain combinations of bound motifs , and not others , drive specific patterns of expression . Here , we take a bottom-up approach to address this question: We constructed simple , synthetic elements containing motifs for only one or two TFs in different orientations and integrated them into the Drosophila genome . By assessing when and where these elements drive expression , we could model specific rules governing tissue-specific enhancer activity . Despite the general importance of TF combinatorial interactions during development , elements with a single TF's motif were often sufficient to drive complex expression . By combining motifs for two factors , we observed non-additive expression in the heart . While the enhancer's activity could tolerate changes in motif spacing and orientation in many tissues , the robustness of heart expression was very sensitive to subtle sequence changes . These results highlight an important property of enhancers—as their readout is context-specific , so too are the effects of mutations within them , including small insertions that may alter a gene's expression in one tissue , but not in another .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"systems",
"biology",
"developmental",
"biology",
"model",
"organisms",
"genetics",
"synthetic",
"biology",
"biology"
] |
2014
|
Subtle Changes in Motif Positioning Cause Tissue-Specific Effects on Robustness of an Enhancer's Activity
|
Insulin , the primary hormone regulating the level of glucose in the bloodstream , modulates a variety of cellular and enzymatic processes in normal and diseased cells . Insulin signals are processed by a complex network of biochemical interactions which ultimately induce gene expression programs or other processes such as translation initiation . Surprisingly , despite the wealth of literature on insulin signaling , the relative importance of the components linking insulin with translation initiation remains unclear . We addressed this question by developing and interrogating a family of mathematical models of insulin induced translation initiation . The insulin network was modeled using mass-action kinetics within an ordinary differential equation ( ODE ) framework . A family of model parameters was estimated , starting from an initial best fit parameter set , using 24 experimental data sets taken from literature . The residual between model simulations and each of the experimental constraints were simultaneously minimized using multiobjective optimization . Interrogation of the model population , using sensitivity and robustness analysis , identified an insulin-dependent switch that controlled translation initiation . Our analysis suggested that without insulin , a balance between the pro-initiation activity of the GTP-binding protein Rheb and anti-initiation activity of PTEN controlled basal initiation . On the other hand , in the presence of insulin a combination of PI3K and Rheb activity controlled inducible initiation , where PI3K was only critical in the presence of insulin . Other well known regulatory mechanisms governing insulin action , for example IRS-1 negative feedback , modulated the relative importance of PI3K and Rheb but did not fundamentally change the signal flow .
Insulin , the primary hormone regulating the level of glucose in the bloodstream , modulates a variety of cellular and enzymatic processes in normal and diseased cells [1]–[7] . The regulation of cellular function by insulin and insulin-like growth factors I/II ( IGF-I/II ) is a highly complex process [8]–[14] . Insulin and IGF-I/II interact with insulin receptors ( IR ) , and type I/II IGF receptors ( IGF-IR/IIR ) in addition to other transmembrane receptors [10] . These interactions ultimately induce gene expression programs or other processes such as translation initiation . Translation rates of many cell cycle and survival proteins are modulated by growth factor , hormone or other mitogenic signals [15] . Insulin induces the activation of class I Phosphoinositide 3-kinases ( PI3Ks ) , which in turn activate the serine/threonine protein kinase Akt and the mammalian target of rapamycin ( mTOR ) . The PI3K/Akt/mTOR signaling axis is important to a variety of cellular programs , including apoptosis [16] , cell size control [17] and translation initiation . Among other functions , activation of the PI3K/Akt/mTOR axis results in the phosphorylation of eukaryotic translation initiation factor 4E-binding protein ( 4E-BPx ) family members [18] . Phosphorylation of 4E-BPx causes the release of the eukaryotic translation initiation factor 4E ( eIF4E ) , which is critical to directing ribosomes to the 7-methyl-guanosine cap of eukaryotic mRNAs . Previously , the availability of eIF4E has been shown to be rate limiting for translation initiation in many eukaryotic cell-lines [15] , [19] . Given its central role in cell biology , evolutionarily optimized infrastructure like translation might be expected to be robust or highly redundant . Surprisingly , deregulated translation , especially involving growth-factor or insulin induced initiation mechanisms , has been implicated in a spectrum of cancers [20] . Despite the wealth of literature on insulin signaling , the relative importance of the components linking insulin with translation initiation remains unclear . Many investigators have explored this question using both experimental and computational tools . For example , Caron et al . recently published a comprehensive map of the mTOR signaling network , including a detailed portrait of insulin induced mTOR activation and its downstream role in translation initiation [21] . Taniguchi et al . proposed three criteria to identify the critical nodes of insulin signaling: network divergence , degree of regulation and potential crosstalk [10] . Using these criteria , they identified insulin-receptor ( IR ) , PI3K and Akt as the critical nodes of insulin action . Several insightful mathematical models of insulin-signaling have also been published [22]–[25] . While these models vary in their focus and biological scope , none has exclusively focused on how insulin stimulates translation initiation . This particular question was addressed by Nayak et al . , who analyzed a family of detailed mathematical models of growth factor and insulin induced translation initiation [26] . Like the Taniguchi et al . hypothesis , their study suggested that Akt/mTOR were structurally fragile , and likely the key elements integrating growth factor signaling with translation . However , the Nayak et al . model neglected several key features of insulin processing , e . g . , negative feedback of IR resulting from mTOR activity . The objective of this study was to rank-order the importance of components of insulin-induced translation initiation using computational tools . Toward this objective , we analyzed an ensemble of mechanistic mathematical models of insulin induced translation initiation that was a significant extension of our previous work [26] . First , we expanded the original model connectivity to include a detailed description of the regulation and activity of insulin , insulin-like growth factor and platelet-derived growth factor ( PDGF ) receptor family members ( including negative feedback ) . Second , we refined the description of the phosphorylation state of Akt and its downstream role in the activation of the mTORC1 and mTORC2 complexes . Lastly , we used new model estimation and interrogation techniques to generate and analyze an uncorrelated population of initiation models that were simultaneously consistent with 24 qualitative and quantitative data sets . Interrogation of this model population , using sensitivity and robustness analysis , identified an insulin-dependent switch that controlled translation initiation . Without insulin , a balance between the pro-initiation activity of the GTP-binding protein Rheb and anti-initiation activity of PTEN controlled basal initiation . Rheb knockdown simulations confirmed decreased initiation in the majority of the model population , while translation initiation increased for all models in the population following a PTEN deletion . On the other hand , a combination of PI3K and Rheb activity controlled insulin inducible initiation . PI3K deletion in the presence of insulin removed the ability of the network to process insulin signals , but did not remove initiation altogether . PI3K deletion reduced initiation to approximately 60% of its maximum level . Interestingly , the relative contribution of PI3K versus Rheb to the overall initiation level could be tuned by controlling IRS-1 feedback . In the absence of feedback , PI3K was more important than Rheb to signal propagation , while the opposite was true in the presence of feedback . Taken together , our modeling study supported the Taniguchi et al . hypothesis that PI3K was a critical node in the insulin-induced initiation network . However , we also found that the role of PI3K was nuanced; PI3K in combination with Rheb controlled initiation in the presence of insulin , while the combination of PTEN and Rheb controlled basal initiation .
The translation initiation model consisted of 250 protein , lipid or mRNA species interconnected by 573 interactions ( Fig . 1 ) . The model described the integration of insulin and growth-factor signaling with 80S assembly . While other eukaryotic translation initiation mechanisms exist , we focused only on cap-mediated translation as the dominant translation mechanism [27] . The model interactome was taken from literature ( SBML file available in the supplemental materials Protocol S1 ) ; the connectivity of insulin- and growth-factor induced translation initiation has been extensively studied [14] , [28] . The model interactome was not specific to a single cell line . Rather , it was a canonical representation of the pathways involved in insulin and growth-factor induced initiation . Using a canonical network allowed us to explore general features of insulin or growth-factor induced translation initiation without cell line specific artifacts . Binding of insulin or IGF-I/II with IR or IGF-I/IIR promotes the autophosphorylation of the cytosolic domains of these receptors at tyrosine residues . Receptor autophosphorylation promotes the formation of adaptor complexes , which are anchored in place by insulin receptor substrate ( IRSx ) family members; IRSx are required for the assembly of adaptor complexes involving the SHC-transforming protein 1 ( Shc ) , Son of Sevenless ( SoS ) , growth factor receptor-bound protein 2 ( Grb2 ) and Ras proteins [29]–[31] . In the model we considered only the IRS-1 protein and neglected other IRSx family members . Adaptor complex formation ultimately culminates in the activation of the catalytic subunit of PI3K . Among their many roles , PI3Ks catalyze the phosphorylation of the phospholipid PIP2 to PIP3 [6] . PIP3 is critical to the localization of 3-phosphoinositide-dependent kinase 1 ( PDK1 ) to the membrane , where it phosphorylates the master kinase Akt at Thr308 [32] . Akt is further phosphorylated at Ser473 by the rictor-mammalian target of rapamycin ( mTORC2 ) protein [33] . Once phosphorylated , Akt promotes translation initiation by directly or indirectly activating the mTORC1 protein [1] . Akt directly activates mTORC1 through a novel binding partner known as PRAS40 [34] , [35] . However , mTORC1 can also be activated by the GTP bound form of the Ras homologue enriched in brain ( Rheb ) protein . Without insulin , Rheb is regulated by the tuberous sclerosis complex TSC1/2 , which has GTPase activating protein ( GAP ) activity . Akt directly phosphorylates TSC1/2 which inhibits its GAP activity and allows Rheb-mediated activation of mTORC1 [36] , [37] . Activated mTORC1 plays two key roles in translation initiation; first , it activates ribosomal protein S6 kinase beta-1 ( S6K1 ) and second it phosphorylates eukaryotic translation initiation factor 4E-binding protein ( 4E-BPx ) family members [38] . In this study , we included only 4E-BP1 and modeled a single deactivating phosphorylation site . Phosphorylated 4E-BP1 releases eIF4E which , along with other initiation factors , is critical to directing ribosomes to the 7-methyl-guanosine cap structure of eukaryotic mRNAs [28] . Several mechanisms attenuate insulin and growth-factor induced translation initiation . First , insulin signal propagation can be controlled by disrupting adaptor complex formation . For example , we included tyrosine phosphatases and competitive inhibitors such as protein-tyrosine phosphatase 1B ( PTP1B ) , src homology phosphotyrosyl phosphatase 2 ( SHP2 ) , growth factor receptor-bound protein 10 ( Grb10 ) and suppressor of cytokine signaling 1/3 ( SOCS1/3 ) which interfere with adaptor complex formation and activity [10] , [39]–[41] . Second , several mechanisms control PIP3 formation , PDK1 recruitment and Akt phosphorylation [10] . In the model , we included the phosphatase and tensin homolog ( PTEN ) protein , which dephosphorylates PIP3 [42] , as well as the SH2 ( Src homology 2 ) -containing inositol phosphatase-1 ( SHIP1 ) protein which hydrolyses the 5-phosphates from PIP3 [43] . Lastly , S6K1 inhibits IRS-1 activity by phosphorylation at Ser318 [44] . S6K1/IRS-1 feedback has been shown to be important in insulin resistance and cancer [14] , [45]–[47] . Translation initiation was modeled using mass-action kinetics within an ordinary differential equation ( ODE ) framework . ODEs and mass-action kinetics are common methods of modeling biological pathways [48]–[50] . However , ODEs have several important limitations that could be addressed with other model formulations e . g . , Partial Differential Equation ( PDE ) based models . PDEs naturally describe spatially distributed intracellular processes or can be used to model population dynamics using population balance methods [51] . However , the computational burden associated with solving and analyzing systems of PDEs , especially at the scale of the current study , would be substantial . Alternatively , we have addressed both of these ODE shortcomings ( without resorting to a PDE formulation ) by including well-mixed compartments to account for spatially localized species and processes and have considered an ensemble of models in our analysis to coarse-grain population phenomena . Irregardless of whether we have an ODE or PDE model formulation , both classes of model typically require the identification of a large number of unknown model parameters . The initiation model had 823 unknown parameters ( 573 kinetic parameters and 250 initial conditions ) , which were not uniquely identifiable ( data not shown ) . We estimated an experimentally constrained population of parameters using multiobjective optimization . Model parameters were estimated , starting from an initial best fit parameter set , using 24 in vitro and in vivo data sets taken from literature ( Table 1 ) . These training data were taken from multiple independent studies ( in different cell lines ) exploring insulin and IGF-I/II signaling or in-vitro translation initiation . These data were largely western blot measurements of the total or phospho-specific abundance of proteins following the addition of a stimulus or inhibitor . While the use of multiple cell-lines was not ideal , it did allow us to capture a consensus picture of insulin or IGF-I/II initiated signaling ( which was useful in understanding the general operational principles of the network ) . However , one should be careful when applying consensus models to specific cell lines or tissues , as these generally may behave qualitatively differently . The residual between model simulations and each of the experimental constraints was simultaneously minimized using the multiobjective POETs algorithm [52] . We used a leave-three-out cross validation strategy to independently estimate prediction and training error during parameter identification ( Table 1 ) . Additionally , a random control ( 100 random parameter sets ) was run to check the training/prediction fitness above random ( Table 1 ) . The training error for 23 of the 24 objectives was statistically significantly better than the random control at a 95% confidence level . Additionally , for 20 of the 24 objectives , the model prediction error was also significantly better than the random control ( p0 . 05 ) . Of the four remaining objectives ( O4 , O5 , O12 and O13 ) , three involved phosphorylated Akt ( O4 and O12 ) or IRS-1 ( O13 ) , each of which had redundant measurements in the objective set that were significant . While the remaining objective , which involved IRS-1 levels ( O5 ) , was not significantly better than the random control , the absolute error was small . The ensemble of translation models recapitulated diverse training data across multiple cell lines . POETs generated 18 , 886 probable models with Pareto rank 4 . Model parameters had coefficients of variation ( CV ) ranging from 0 . 65 to 1 . 10 . Further , 89% ( 512 of 573 ) of the model parameters were constrained with a CV 1 . The performance of 5 , 818 rank-zero models is shown in Fig . 2 . The majority of objective functions were uncorrelated e . g . , O4O13 or O12O13 or directly proportional e . g . , O3O11 or O9O15 . Uncorrelated or proportional objectives suggested the model population simultaneously described each training constraint . However , several other objectives were inversely proportional e . g . , O12O14 . For these pairs , the model was unable to simultaneously fit both training data sets . Surprisingly , these objectives were the same protein pAkt ( Thr308 ) O9O12 and pS6K1 ( Thr389 ) O3O14 , taken from either different cell lines or different labs . This suggested conflicts in the data e . g . , cell line variation or differences in specific laboratory protocols , rather than structural inaccuracies in the model , were responsible for the inverse relationship . The key indicators of eukaryotic translation initiation are the phosphorylation of S6K1 and 4E-BP1 [38] . Both Tzatos et al . and Villalonga et al . performed insightful studies exploring the dynamics of S6K1 and 4E-BP1 phosphorylation in L6 Myotubes and RhoE 3T3 cells [53] , [54] . The ensemble recapitulated these observations with error distributions that were statistically significantly better than random parameters ( , ; , ) ( Fig . 3A and 3B , Table 1 ) . The model population also recapitulated IGF1 induced Akt and S6K1 phosphorylation ( , ; , ) ( Fig . 3E and 3F , Table 1 ) . Lorsh et al . studied ribosomal assembly dynamics in rabbit reticulocytes , suggesting the formation of the eIF2∶GTP∶Met-tRNA tertiary complex was rate limiting in 80S formation [55] . Our model captured 80S assembly dynamics , including the crucial lag phase in the first two minutes of stimulation ( , ) ( Fig . 3C , Table 1 ) . Inhibitor data was also used for model training . Without insulin , PI3K was not activated and pAkt ( Ser473 ) levels remained low ( Fig . 3D , lane 1 ) . Following insulin stimulation , PI3K activation resulted in increased pAkt ( Ser473 ) levels ( Fig . 3D , lane 2 ) . Wortmannin , a PI3K inhibitor , significantly decreased pAkt ( Ser473 ) ( Fig . 3D , lane 3 ) . While our model population qualitatively captured this decrease , the levels of pAkt ( Ser473 ) were higher than those observed experimentally . The model was not trained using mTORC1/2 measurements , however species immediately upstream and downstream of mTORC1/2 , namely pAkt ( Ser473 ) or S6K1 were used in model training . Without insulin , pAkt ( Ser473 ) and S6K1 ( Thr421/Ser424 ) levels were low ( Fig . 3E/F , lanes 1 ) . Addition of insulin increased pAkt ( Ser473 ) and S6K1 ( Thr421/Ser424 ) . Upon rapamycin addition , mTORC1 was inhibited and the levels of phosphorylated S6K1 decreased ( Fig . 3E , lane 3 ) . However , because of its position upstream of mTORC1 , pAkt ( Set473 ) levels were unchanged ( Fig . 3E , lane 3 ) . The model was validated by comparing simulations with in vivo and in vitro data sets not used for training or cross-validation ( Table 2 ) . For four of the five prediction data sets , the model demonstrated errors statistically significantly better than a random control ( p0 . 05 ) . However , the remaining prediction case ( P3 ) , while not significantly different than random , has a small error relative to the other objectives . Data from Lorsh et al . was used to validate the dynamics of intermediate ribosomal complexes [55] . The level of 43S mRNA was quantified using both GTP and a non-degradable GTP-like homologue GMP-PNP ( Fig . 4A ) . Data involving GMP-PNP was used for training while data involving GTP was used only for validation ( , ) . Garami et al . explored insulin-induced Rheb activation and the role of TSC1/2 in the presence and absence of wortmannin and rapamycin [56] . We first compared measured versus simulated Rheb-GTP levels , with and without insulin , in the absence of inhibitors . While we captured the qualitative trends , we over-predicted the percentage of GTP bound Rheb ( , ) ( Fig . 4B ) . The model also failed to predict sustained Rheb-GTP levels in the presence of rapamycin . This suggested that sustained pAkt ( Ser473 ) levels ( observed in Fig . 3E ) were not correlated with increased Rheb-GTP activity . Garami et al . also measured the levels of GTP bound Rheb in both wild-type and TSC2 knockout cells . Because of TSC2's regulatory role , a TSC2 knockout significantly increased Rheb-GTP levels ( , ) ( Fig . 4C ) . Lastly , the model predicted the levels of 4E-BP1 bound eIF4E in response to heat shock ( , ) ( Fig . 4D ) [57] . Because the model was not trained on stress-induced translation inhibition , this result further demonstrated the predictive power of the model population . Sensitivity analysis generated falsifiable predictions about the fragility or robustness of structural features of the initiation architecture . First order sensitivity coefficients were computed for 40 parameter sets selected from the ensemble ( materials and methods ) , time-averaged and rank-ordered for the 250 species in the model , in the presence and absence of insulin and IRS-1 feedback . The sensitive components of insulin signaling shifted from Rheb in the absence of insulin to a combination of Rheb and PI3K in the presence of insulin . Sensitivity coefficients ( ) were calculated with and without insulin over the complete 100 min response ( Fig . 5A ) . Globally , processes involved with 80S formation were consistently ranked among the most sensitive , irrespective of insulin . However , the sensitivity of other signal processing components changed with insulin status . For example , without insulin , Rheb/Rheb-GDP were highly fragile ( rank0 . 25 ) , while PI3K , PIP2 , PIP3 and PTEN were highly robust ( rank0 . 0 ) . Surprisingly , the relative sensitivity of these network components changed in the presence of insulin . While the fragility of Rheb/Rheb-GDP shifted modestly upward with insulin , the sensitivity of PI3K and its downstream complexes increased dramatically ( rank0 . 45 ) following insulin stimulation . This suggested that the combination of PI3K and Rheb activity was critical to insulin action over the full 100 min time window . However , it was unclear whether PI3K was always important , or if there was a temporal window in which PI3K became important following insulin stimulation . To explore this question , we time-averaged the sensitivity coefficients over early- and late-phase time periods following insulin stimulation ( Fig . 5B ) . The 0–5 minute time period captured the initial network dynamics , while the 30–100 minute time period captured the network at a quasi-steady state . Generally , network components were more sensitive under dynamic operation ( species beneath the 45 line ) , compared with steady state . However , there were exceptions to this trend . For example , PI3K , PTEN and TSC1/2 were equally sensitive in both time frames , suggesting these species played important roles in both dynamic and steady state signaling . On the other hand , the Rheb rank decreased from to as the network moved toward steady state . Taken together , the sensitivity results suggested that Rheb activity controlled the background level of translation initiation while the PI3K axis in combination with Rheb regulated insulin-induced initiation . Moreover , the transition between PTEN and PI3K control occurred directly after the addition of insulin , giving rise to switch like behavior . IRS-1 phosphorylation , a well known negative feedback mechanism [14] , [45]–[47] , attenuated PI3K sensitivity . We explored the role of IRS-1 feedback by comparing sensitivity coefficients under insulin stimulation in the presence and absence of IRS-1 feedback ( Fig . 5C ) . The most significant change without feedback was the sensitivity of the IR∶IRS-1 and adaptor complexes ( Fig . 5C , black fill ) ; IR∶IRS-1 , which anchors the adaptor complex to the activated receptor and is immediately upstream of PI3K activation , changed from NSS rank 0 . 04 to 0 . 32 . The sensitivity of the PI3K/Akt signaling axis also increased in the absence of feedback ( Fig . 5C , grey fill ) . Surprisingly , the sensitivity of Rheb and many ribosomal components decreased in the absence of feedback . Similar results were observed when sensitivity coefficients were time averaged over the 0 to 5 min time window ( Fig . 5D ) . These sensitivity calculations suggest that IRS-1 feedback plays a significant role in insulin signaling by modulating the relative importance of PI3K versus Rheb . Thus , IRS-1 feedback though not directly identified as a fragile regulatory motif , has significant effects on network function . Lastly , the architectural features of the initiation network identified by sensitivity analysis , as either fragile or robust , were likely parameter independent . While first-order sensitivity coefficients are local , we sampled a family of uncorrelated parameter sets ( mean correlation of approximately 0 . 6 ) to generate a set of consensus conclusions . By sampling over many uncorrelated sets , we calculated how our conclusions changed with different unrelated parameter sets . The distribution of ranking ( standard-error shown in Fig . 5 ) suggested that despite parametric uncertainty , sensitivity analysis over an uncorrelated model population produced a consensus estimate of the strongly fragile or robust elements of the insulin signaling network . Previously , we ( and others ) have shown that monte-carlo parameter set sampling produced similar results in several studies across many signaling networks [49] , [58]–[60] . Knockdown simulations were conducted for 92 proteins to estimate the functional connectedness of the initiation network . The effects of the perturbations were quantified by calculating the relative change ( ) in translational activity ( 80S formation ) for each simulated knockout in the presence ( Fig . 6A ) and absence ( Fig . 6B ) of insulin . Knockdown simulations were conducted using 400 models selected from the ensemble based on error and correlation ( materials and methods ) . Proteins were classified based on their impact on translational activity: little or no effect ( , white fill ) , moderate decrease ( , dark grey ) , critical ( , light grey ) and increase ( , black ) . Generally , knockdowns in the presence of insulin were more likely to decrease initiation ( Fig . 6A ) . Knockdown analysis identified 24 proteins ( or 26% of the network ) that were critical to translation initiation irrespective of insulin status; these critical components included mTORC1 , S6K1 , several initiation factors and other ribosomal components . Sensitivity analysis suggested basal translation was governed by Rheb , while insulin-induced initiation was governed by PI3K . Robustness analysis showed that perturbations in PI3K signaling , in the presence of insulin , restored initiation control to Rheb . Initiation was reduced by 40% by disrupting species immediately upstream or downstream of PI3K; a moderate reduction in the presence of insulin demonstrated that initiation was governed by both PI3K and Rheb . Lastly , deletion of TSC1/2 ( negative regulator of Rheb ) or 4E-BP1 ( sequesters the cap-binding protein eIF4E ) , increased initiation in the presence of insulin . Interestingly , for several proteins the direction or magnitude of change in initiation activity depended upon the presence or absence of insulin . For example , PTEN deletion significantly increased initiation ( 1 ) in the absence of insulin , but had no effect when insulin was present . On the other hand , PI3K deletion had a moderate reduction on 80S formation in the presence of insulin , but only a small effect in the absence of insulin ( Fig . 6B ) . These results suggested that PI3K and PTEN were conditionally fragile proteins; in the presence of insulin , PI3K is a critical signal processing node , while PTEN acts to restrain inadvertent basal initiation . Paradoxically , Rheb and mTORC2 subunit ( sin1 , rictor ) knockdowns increased initiation . Our expectation from sensitivity analysis was that a Rheb knockdown would reduce initiation , irrespective of insulin status . However , this was not universally true; some members of the model population showed increased initiation ( Fig . 6C ) . Following the deletion of PTEN , approximately 80% ( or 323 of the 400 models sampled ) had increased initiation in the absence of insulin . Of these models , 16% ( or 51 of 323 ) had at least a two fold increase in translational activity . This result was expected; deletion of a protein species resulted in a qualitatively similar change in initiation across the ensemble of models . However , for Rheb knockdowns , members of the ensemble demonstrated qualitatively different behavior . For 84% ( or 334 of 400 ) of the models sampled , Rheb knockdowns significantly down-regulated initiation . Thus , the vast majority of models behaved as expected . Interestingly , 20 models ( or 5% of the models sampled ) had increased translation initiation in the presence of a Rheb knockdown , with 15 models demonstrating greater than a two-fold change ( Fig . 6C ) . Thus , the model population estimated by POETs contained models with qualitatively different behavior . Histograms of sin1 and rictor knockdowns showed a similar trend ( results not shown ) . We explored the flux vectors of these outlying parameter sets to better understand the mechanistic effect of Rheb and rictor/sin1 knockouts . All of the outlying models were in regions of parameter space where the association between Rheb and GTP was very high . Strong Rheb/GTP binding resulted in abnormally high signal flux to mTORC1 despite the inhibitory effects of TSC1/2 ( Fig . 6D , top-left ) . Consequently , less GTP was available for the energy-dependent steps of translation initiation ( i . e . formation of eIF2-GTP-met-tRNA tertiary complex ) . Additionally , strong association between Rheb and GTP resulted in high levels of activated mTORC1 and S6K1 . However , despite the high levels of mTORC1 , GTP-dependent pre-initiation reactions were rate limiting ( Fig . 6D , labeled* ) . Thus , Rheb knockdown released the network from its GTP limitation and shifted the predominant signaling mode to mTORC2 . This shift in signaling , while lowering the activated mTORC1/S6K1 level , ultimately resulted in higher levels of initiation ( Fig . 6 bottom-left ) . On the other hand , the rictor/sin1 knockdown behaved differently . The rate-limiting step for the rictor/sin1 knockdowns was mTORC1 activation: more Rheb-GTP was present than there was mTORC1 to be activated ( Fig . 6D top-right ) . Thus , knockdown of rictor/sin1 prevented the assembly of mTORC2 and freed the mTOR subunit to be used for mTORC1 assembly . This shift toward mTORC1 assembly and activation relieved the Rheb-GTP/mTORC1 bottleneck , resulting in increased initiation .
In this study , we developed and analyzed a population of insulin and growth factor induced translation initiation models . These models described the integration of insulin and growth-factor signals with 80S assembly . A family of model parameters was estimated from 24 transient and steady state data sets using multiobjective optimization . In addition to the training data , the model family also predicted novel data sets not used during model training . The population of initiation models was analyzed using sensitivity and robustness analysis to identify the key components of insulin-induced translation initiation . Without insulin , a balance between the pro-initiation activity of the GTP-binding protein Rheb and anti-initiation activity of PTEN controlled basal initiation . Rheb knockdown simulations confirmed decreased initiation in the majority of the model population . Surprisingly , we also identified a model subpopulation in which deletion of Rheb or mTORC2 components increased initiation . In these cases , removal of Rheb or mTORC2 components relieved a rate-limiting bottleneck e . g . , constrained levels of GTP , leading to increased initiation . On the other hand , in the absence of insulin , translation initiation increased for all models in the population following a PTEN deletion . In the presence of insulin , Rheb and PTEN were no longer the dominant arbiters of initiation; a combination of PI3K and Rheb activity controlled inducible initiation , where PI3K was only critical in the presence of insulin . PI3K deletion in the presence of insulin removed the ability of the network to process insulin signals , but did not remove initiation altogether . PI3K deletion reduced initiation to approximately 60% of its maximum level . Interestingly , the relative contribution of PI3K versus Rheb to the overall initiation level could be tuned by IRS-1 feedback . In the absence of feedback , PI3K was more important than Rheb to signal propagation , while the opposite was true in the presence of feedback . PI3K and PTEN in combination with Rheb are components of a switch that regulates inducible and basal translation initiation . In the absence of insulin , a balance between the pro-initiation activity of Rheb and the anti-initiation activity of PTEN regulated basal initiation . On the other hand , in the presence of insulin , control shifted to a combination of Rheb and PI3K , where PI3K activity regulated the inducible fraction of initiation . Thus , deletion of PTEN , constitutive activation of PI3K or constitutively active Rheb could all induce aberrant translation initiation without an insulin or growth factor signal . Yuan and Cantley noted that every major species in the PI3K pathway is mutated or over-expressed in a wide variety of solid tumors [6] . For example , activating mutations in PIK3CA , the gene encoding the catalytic subunit of PI3K , induces oncogene signaling in colon , brain and gastric cancers [61] . On the other hand , PTEN mutations have long been implicated in a spectrum of cancer types [62] . Both PIK3CA and PTEN mutations induce a pro-initiation operational mode in the absence of growth factor . Likewise , constitutive Rheb activity induces a variety of pleiotropic traits involving translation . For example , Saucedo et al . showed that Rheb over-expression in Drosophila melanogaster increased cell size , wing area and G1/S cell cycle progression [63] . Rheb and TSC1/2 mutations are also frequently observed in cancer [64] , [65] . Taken together , our study supports the supposition of Taniguchi et al . that PI3K is a critical arbiter of insulin-induced translation initiation [10] . However , we have also shown that initiation control and particularly the role of PI3K was more nuanced; while insulin or growth-factor inducible initiation was controlled by PI3K , basal initiation was controlled by Rheb . Moreover , in the absence of insulin , PTEN was the critical upstream initiation regulator , not PI3K . This suggested that the relative level of the phosphorylated phospholipids PIP2 and PIP3 was actually the key mediator of initiation . Lastly , Taniguchi et al . suggested that Akt was also a key node involved in insulin action . Our previous model directly supports this , however , the current model does not . Rather , our analysis suggested that Rheb was the downstream controller of initiation . These two points of view are not contradictory however , as Rheb activation is driven by phosphorylated Akt . The initiation model connectivity was assembled from an extensive literature review , however , several potentially important signaling mechanisms were not included . First , we should revisit the role of PRAS40 . Currently , PRAS40 acts as a cofactor that aids in pAkt ( Ser473 ) -mediated activation of mTORC1 . Sancak et al suggested that PRAS40 sequesters mTORC1 , and only after phosphorylation by Akt does it releases from mTORC1 [34] . Other groups have also shown that mTORC1 can phosphorylate and inhibit PRAS40 , thus providing a positive feedback mechanism for Akt-mediated mTORC1 activation [66] , [67] . A more complete description of PRAS40 will enhance our ability to interrogate Akt dependent mTORC1 activation . Second , we need to refine the description of IRS-1 feedback . Currently , we assume a single deactivating phosphorylation event at Ser308 . However , several studies have shown that IRS-1 can be phosphorylated at multiple serine sites , which are both activating and deactivating [44] , [68] . Additionally , PTEN is known to dephosphorylate activated PDGF receptors and attenuate their activity , a feature not included currently [69] . A more complete description of IRS-1 phosphorylation could help define how , and under what conditions , IRS-1 regulation attenuates PI3K activation . Third , we modeled the regulation of 4E-BPx as a single phosphorylation event where phosphorylated 4E-BPx was unable to bind to eIF4E . In reality , 4E-BPx family members , such as 4E-BP1 , have several phosphorylation sites [70] and the release of eIF4E is driven only after multiple conserved phosphorylation events [71] . Additionally , eIF4E can itself be phosphorylated at Ser209; while there is agreement that the phosphorylation of eIF4E does have a regulatory significance , the data is contradictory as to whether it is positive or negative [72] . Fourth , signaling downstream of mTORC1 has also been shown to mediate translation modes beyond those included in our model . eIF3 has been identified as a scaffolding protein that recruits mTORC1 to untranslated mRNA and facilitates S6K1 and 4E-BP1 phosphorylation [73] . S6K1 can also activate eIF4B , a protein that helps eIF4A to unwind the secondary structure of untranslated mRNA [74] . Further , a recently discovered scaffold protein , SKAR , has been shown to assist S6K1 recruitment to mRNA [75] . Lastly , because of mTORC1's unique cellular role , it would be interesting to explore how other aspects of metabolism interact with insulin signaling to mediate decisions between translation , lipid synthesis or proliferation . In these studies , one could imagine constructing in-vivo mouse models to explore the physiological role of mTORC1 signaling in important diseases such as diabetes or cancer .
The translation initiation model was formulated as a set of coupled non-linear ordinary differential equations ( ODEs ) : ( 1 ) The symbol denotes the stoichiometric matrix ( ) . The quantity denotes the concentration vector of proteins ( ) . The term denotes the vector of reaction rates ( ) . The element of the matrix , denoted by , described how protein was involved in rate . If , then protein was consumed in . Conversely , if , protein was produced by . Lastly , if , then protein was not involved in rate . We assumed mass-action kinetics for each interaction in the network . The rate expression for interaction was given by: ( 2 ) The set denotes reactants for reaction while denotes the stoichiometric coefficient ( element of the matrix ) governing species in reaction . The quantity denotes the rate constant governing reaction . All reversible interactions were split into two irreversible steps . Model equations were generated using UNIVERSAL from an SBML input file ( available in the supplemental materials Protocol S1 ) . UNIVERSAL is an open source Objective-C/Java code generator , which is freely available as a Google Code project ( http://code . google . com/p/universal-code-generator/ ) . The model equations were solved using the LSODE routine in OCTAVE ( v 3 . 0 . 5; www . octave . org ) on an Apple workstation ( Apple , Cupertino , CA; OS X v10 . 6 . 4 ) . When calculating the response of the model to the addition of insulin or other growth factors , we first ran to steady state and then issued the perturbation . The steady state was estimated numerically by repeatedly solving the model equations and estimating the difference between subsequent time points: ( 3 ) The quantities and denote the simulated concentration vector at time and , respectively . The vector-norm was used as the distance metric , where s and = 0 . 001 for all simulations . We used multiobjective optimization in combination with cross-validation to estimate an ensemble of initiation models . Multiobjective optimization in combination with cross-validation allowed us to address qualitative conflicts in the training data , and to protect against model over-training . While computationally more complex than single-objective formulations , multiobjective optimization is an important tool to address qualitative conflicts in training data that arise from experimental error or cell-line artifacts [76] . Multiobjective optimization balances these conflicts allowing us to identify a consensus model population . In this study we used the Pareto Optimal Ensemble Technique ( POETs ) to perform the optimization . POETs integrates standard search strategies e . g . , Simulated Annealing ( SA ) or Pattern Search ( PS ) with a Pareto-rank fitness assignment [52] . Denote a candidate parameter set at iteration as . The squared error for for training set was defined as: ( 4 ) The symbol denotes scaled experimental observations ( from training set ) while denotes the scaled simulation output ( from training set ) . The quantity denotes the sampled time-index and denotes the number of time points for experiment . In this study , the experimental data used for model training was typically the band intensity from immunoblots , where intensity was estimated using the ImageJ software package [77] . The scaled measurement for species at time in condition is given by: ( 5 ) Under this scaling , the lowest intensity band equaled zero while the highest intensity band equaled one . A similar scaling was defined for the simulation output . By doing this scaling , we trained the model on the relative change in blot intensity , over conditions or time ( depending upon the experiment ) . Thus , when using multiple data sets ( possibly from different sources ) that were qualitatively similar but quantitatively different e . g . , slightly different blot intensities over time or condition , we captured the underlying trends in the scaled data . We computed the Pareto rank of by comparing the simulation error at iteration against the simulation archive . We used the Fonseca and Fleming ranking scheme [78] to estimate the number of parameter sets that dominate . Parameter sets with increasing rank are progressively further away from the optimal trade-off surface . The parameter set was accepted or rejected by POETs with probability : ( 6 ) where is the annealing temperature and denotes the Pareto rank for . The annealing temperature was discretized into 10 quanta between and and adjusted according to the schedule where was defined as . The initial temperature was given by , where was used in this study and the final temperature was . The epoch-counter was incremented after the addition of 100 members to the ensemble . Thus , as the ensemble grew , the likelihood of accepting parameter sets with a large Pareto rank decreased . To generate parameter diversity , we randomly perturbed each parameter by . We performed a local pattern search every steps to minimize the residual for a single randomly selected objective . The local pattern-search algorithm has been described previously [79] . A leave-three-out cross-validation strategy was used to simultaneously calculate the training and prediction error during the parameter estimation procedure [80] . The 24 training data sets were partitioned into eight subsets , each containing 21 data sets for training and three data sets for validation . The leave-three-out scheme generated 18 , 886 probable models . From the approximately 6000 rank zero models , we iteratively selected 50 random models from each cross-validation trial with the lowest correlation and shortest Euclidian distance to the origin ( minimum error ) . This selection technique produced sub-ensembles with low set-to-set correlation ( 0 . 50 ) and minimum training error . Sensitivity coefficients were calculated for 40 models selected from the ensemble ( rank-zero , low-correlation , minimum error selection ) . First-order sensitivity coefficients at time : ( 7 ) were computed by solving the kinetic-sensitivity equations [81]: ( 8 ) subject to the initial condition . The quantity denotes the parameter index , denotes the number of parameters in the model , denotes the Jacobian matrix , and denotes the th column of the matrix of first-derivatives of the mass balances with respect to the parameters . Sensitivity coefficients were calculated by repeatedly solving the extended kinetic-sensitivity system for forty parameters sets selected from the final 400 member ensemble . These sets were chosen to be comparable to the final 400 member ensemble on the basis of parametric coefficient of variation ( CV ) ; the sets selected for sensitivity analysis had a mean CV of 0 . 850 . 5 and a mean correlation of approximately 0 . 6 . Thus , there were diverse and uncorrelated . The Jacobian and the vector were calculated at each time step using their analytical expressions generated by UNIVERSAL . The resulting sensitivity coefficients were scaled and time-averaged ( Trapezoid rule ) : ( 9 ) where denotes the final simulation time . The time-averaged sensitivity coefficients were then organized into an array for each ensemble member: ( 10 ) where denotes the index of the ensemble member , denotes the number of parameters , denotes the number of ensemble samples and denotes the number of model species . To estimate the relative fragility or robustness of species and reactions in the network , we decomposed the matrix using Singular Value Decomposition ( SVD ) : ( 11 ) Coefficients of the left ( right ) singular vectors corresponding to largest singular values of were rank-ordered to estimate important species ( reaction ) combinations . Only coefficients with magnitude greater than a threshold ( = 0 . 001 ) were considered . The fraction of the vectors in which a reaction or species index occurred was used to determine its importance ( sensitivity ranking ) . The sensitivity ranking was compared between different conditions to understand how control in the network shifted as a function of perturbation or time ( Fig . 5 ) . Robustness coefficients were calculated as shown previously [60] . Robustness coefficients ( denoted by ) are the ratio of the integrated concentration of a network marker in the presence ( numerator ) and absence ( denominator ) of a structural or operational perturbation . The quantities and denote the initial and final simulation time , respectively , while and denote the indices for the marker and the perturbation respectively . If , then the perturbation increased the marker concentration . Conversely , if the perturbation decreased the marker concentration . Lastly , if the perturbation did not influence the marker concentration . Robustness coefficients were calculated over 400 models selected from the ensemble ( rank-zero , low-correlation , minimum error selection ) . Convergence analysis suggested that the qualitative conclusions drawn from the robustness analysis would not change if more than N = 400 parameter sets were sampled ( Fig . S1 ) .
|
Insulin is a hormone produced by the body that regulates uptake of glucose from the bloodstream . The cellular response to insulin is governed by a complex network of intracellular interactions that ultimately influence cell growth and metabolism . Because of its central role in physiology , insulin signaling has been extensively studied . Yet despite this wealth of research , the relative importance of components in insulin signaling remains unclear . Mechanistic computer simulations have been shown to provide insight into the function of complex systems , such as insulin signaling . In this work we constructed and interrogated a mathematical computer simulation of insulin signaling to better understand the important components of the insulin signaling network . We determined the most important network components and identified network perturbations that can induce dramatic shifts in cellular phenotype . Our results offer an in-depth analysis of the insulin signaling pathway and provide a unique paradigm towards understanding how malfunctions in insulin signaling can result in numerous disease states .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"diabetic",
"endocrinology",
"cancer",
"risk",
"factors",
"signaling",
"networks",
"computerized",
"simulations",
"oncology",
"mathematical",
"computing",
"mathematics",
"hormonal",
"causes",
"of",
"cancer",
"insulin",
"theoretical",
"biology",
"endocrinology",
"regulatory",
"networks",
"computing",
"methods",
"biology",
"diabetes",
"and",
"endocrinology",
"nonlinear",
"dynamics",
"systems",
"biology",
"biochemical",
"simulations",
"computer",
"science",
"computer",
"modeling",
"computational",
"biology",
"genetics",
"and",
"genomics"
] |
2011
|
Computational Modeling and Analysis of Insulin Induced Eukaryotic Translation Initiation
|
rs143383 is a C to T transition SNP located in the 5′untranslated region ( 5′UTR ) of the growth differentiation factor 5 gene GDF5 . The T allele of the SNP is associated with increased risk of osteoarthritis ( OA ) in Europeans and in Asians . This susceptibility is mediated by the T allele producing less GDF5 transcript relative to the C allele , a phenomenon known as differential allelic expression ( DAE ) . The aim of this study was to identify trans-acting factors that bind to rs143383 and which regulate this GDF5 DAE . Protein binding to the gene was investigated by two experimental approaches: 1 ) competition and supershift electrophoretic mobility shift assays ( EMSAs ) and 2 ) an oligonucleotide pull down assay followed by quantitative mass spectrometry . Binding was then confirmed in vivo by chromatin immunoprecipitation ( ChIP ) , and the functional effects of candidate proteins investigated by RNA interference ( RNAi ) and over expression . Using these approaches the trans-acting factors Sp1 , Sp3 , P15 , and DEAF-1 were identified as interacting with the GDF5 5′UTR . Knockdown and over expression of the factors demonstrated that Sp1 , Sp3 , and DEAF-1 are repressors of GDF5 expression . Depletion of DEAF-1 modulated the DAE of GDF5 and this differential allelic effect was confirmed following over expression , with the rs143383 T allele being repressed to a significantly greater extent than the rs143383 C allele . In combination , Sp1 and DEAF-1 had the greatest repressive activity . In conclusion , we have identified four trans-acting factors that are binding to GDF5 , three of which are modulating GDF5 expression via the OA susceptibility locus rs143383 .
Osteoarthritis ( OA ) is a common disease of the synovial joints , affecting millions of people worldwide . It is a chronic , highly disabling disease , characterised by the progressive loss of articular cartilage , changes in the subchondral bone , and variable levels of synovial inflammation [1] . Many patients suffer from joint pain and tenderness , limiting the functioning of the joint and thus having a significant impact on quality of life . Furthermore , evidence is now emerging of an increased mortality risk in OA patients [2] . Non-steroidal anti-inflammatory drugs ( NSAIDs ) and cyclo-oxygenase 2 ( COX-2 ) inhibitors are recommended for the pharmacological management of OA . Although these have proven to be effective for pain relief and suppression of inflammation , these treatments are failing to target the underlying cause and progression of disease . There has been limited success so far in trials of disease-modifying drugs , with arthroplasty remaining the basis for curative therapy [3] . There are a number of risk factors for OA , including age , gender , mechanical injury and obesity . Genetics contribute a significant risk to developing the disease , with heritability estimates ranging from 39–79% dependent on the joint site affected [4] . A number of genes have been found to harbour OA susceptibility alleles and genome wide association scans have provided additional loci worthy of investigation [5] . When a susceptibility allele has been identified it is necessary to investigate the functional effect of the polymorphism in order to enhance understanding of its role in disease aetiology . This information can then be used to assist in diagnosis , prognosis and to alleviate detrimental genetic effects by modulating or restoring gene function or expression . To date , the most reproducible association with OA has been to rs143383 , a C/T single nucleotide polymorphism ( SNP ) located within the 5′untranslated region ( 5′UTR ) of the growth differentiation factor 5 gene GDF5 ( HUGO Gene Nomenclature Committee ( HGNC ) number 4420 ) . The T allele of the SNP was first associated with increased risk of OA in an Asian population , with this association subsequently replicated in Europeans [6]–[8] . Haplotype analysis combined with an examination of promoter activity following the sequential deletion of the GDF5 promoter/5′UTR demonstrated that rs143383 is the causal SNP , with its T allele mediating reduced expression relative to its C allele [6] . This phenomenon is known as differential allelic expression ( DAE ) . A subsequent analysis of RNA extracted from the joint tissues of OA patients heterozygous for the SNP revealed that the GDF5 DAE is active during the disease process , with DAE observed in cartilage , ligament , synovium , fat pad and meniscus [7] , [9] . Overall , these studies demonstrated that a reduction in GDF5 expression mediated by the T allele of rs143383 is a risk factor for OA . GDF5 protein has a vital role in the formation and repair of joints . It acts as an extracellular signalling molecule , activating the expression of genes involved in the formation of cartilage and bone [10] . During joint specification , GDF5 is present within the joint interzone , and has been found to have a pivotal role during chondrogenesis [11] . It is expressed in both normal and OA cartilage , and has been proposed to also be important in cartilage repair following trauma [12]–[16] . Rare and highly penetrant mutations of GDF5 underlie several severe musculoskeletal conditions , including Hunter-Thompson syndrome , Grebe syndrome and Brachdactyly Type C [17]–[20] . These conditions present with joint dislocations , which are found to mainly occur in the knees and hips , shortened limb bones , abnormalities in the development of the phalangeal joints and brachydactyly . This essential role of GDF5 during joint development and joint maintenance has been further demonstrated in the mouse brachypodism mutation , which is a premature termination codon of Gdf5 that results in an absence of functional protein from the mutant allele . Homozygous mice have a number of developmental abnormalities of both bone and soft tissues whereas heterozygous mice show no overt growth abnormalities but when challenged are more susceptible to develop an OA-like phenotype [21] , [22] . We have previously reported on DEAF-1 ( HGNC:14677 ) as a potential trans-acting factor that binds to rs143383 [9] . The aim of our latest study was to perform a more detailed analysis of DEAF-1 and to identify additional factors that bind differentially to the two alleles of rs143383 and that could account for the GDF5 DAE that is mediated by this SNP . We used the human liposarcoma cell line SW872 for our research since 1 ) the cell line expresses GDF5; 2 ) it is heterozygous for rs143383; 3 ) it also demonstrates GDF5 DAE and 4 ) it is amenable to a variety of in vitro experimental manipulations . Since SW872 cells exhibit GDF5 DAE it was assumed that the trans-acting factors that mediate the DAE were expressed in these cells . We used two different approaches to identify the novel trans-acting factors . The first utilised bioinformatics software to predict protein binding based on the sequence surrounding rs143383 , followed by electrophoretic mobility shift assays ( EMSAs ) to screen these potential candidates . The second approach used an oligonucleotide pull down assay to isolate proteins binding to the promoter region of GDF5 , followed by quantitative mass spectrometry , enabling both the identification and quantification of proteins binding to the C and T alleles of rs143383 . Chromatin Immunoprecipitation ( ChIP ) , luciferase assays and RNA interference ( RNAi ) were then used to confirm binding of the newly identified candidate proteins in vivo and to assess their role in mediating GDF5 DAE . The EMSA and RNAi results were then confirmed using a combination of the chondrosarcoma cell line SW1353 , the osteosarcoma cell line MG63 and human articular chondrocytes . This study has identified four trans-acting factors that are binding to GDF5 , three of which are modulating the expression of this important growth factor .
As we previously described , the human liposarcoma cell line SW872 is heterozygous at rs143383 , expresses GDF5 and demonstrates DAE [23] . In this cell line there is a DAE imbalance of 1 . 5 between the C and T alleles ( Figure 1 ) , which is comparable to the average DAE observed in human joint tissues [9] . In that study the level of DAE at rs143383 was found to be similar between all the joint tissues examined , and was confirmed in several different cell lines using luciferase reporter assays [9] . This indicates that the imbalance is not due to a tissue or cell type specific factor , but instead implies that the same trans-acting factors are regulating the expression of GDF5 via rs143383 in a number of cell types . We therefore used the SW872 heterozygous cell line as a model system for the discovery and investigation of these trans-acting factors . We investigated protein complex binding using SW872 nuclear extract and fluorescently labelled C and T allele probes ( Figure 2A ) . We observed a similar pattern of protein complex binding to the two probes . We confirmed the specificity of the assay by adding unlabelled C and T allele competitors , and the two specific complexes binding revealed a differential affinity for the two alleles . For both complexes , binding to the C allele probe was outcompeted with excess unlabelled C and T allele competitor , and vice versa for the T allele probe . Higher concentrations of C allele unlabelled competitor were required to outcompete binding to the T allele probe and complex binding was competed from the C allele probe at a lower concentration of T allele competitor compared to C allele competitor . These results suggest the two protein complexes bind more avidly to the T allele , compared to the C allele . We used smaller sized unlabelled competitors to refine the region of binding of the two complexes; this assay suggested that the majority of the sequence of the probe including the rs143383 polymorphic site is required for the binding of the two complexes ( Figure S1 ) . There is a small degree of competition using the 50× concentration of competitor 1 ( −15 to +2 relative to rs143383 ) and competitor 2 ( −6 to +6 ) but not with competitor 3 ( −3 to +10 ) suggesting that the region upstream of the polymorphism may be more important for complex 1 and 2 binding . Using the online databases TransFac , Tess and Promo 3 . 0 , we identified a number of transcription factors that were predicted to bind to GDF5 within the region containing rs143383 . We refined the number of potential factors using competitors containing the consensus binding sequence of each factor ( competitor sequences are listed in Table S1 ) . If binding of either complex to the GDF5 probes was competed , the factors were investigated further by the addition of an antibody targeting the protein to the EMSA binding reaction . On the addition of a shared Sp1/Sp3/ETF consensus competitor , binding of both complexes to the GDF5 probes was competed ( Figure 2B ) . Sp1 ( HGNC:11205 ) and Sp3 ( HGNC:11208 ) had been identified by all three databases . The addition of an antibody targeting Sp1 resulted in a supershift of the upper complex and addition of an antibody targeting Sp3 supershifted both the lower , and one of the upper complexes ( Figure 2C ) . The Sp1 and Sp3 antibodies were the only ones tested that resulted in supershifts; Figure S2 shows examples of trans-acting factors that did not supershift , along with a supershifted Sp1 . These results confirm the binding of Sp1 and Sp3 to GDF5 in vitro in SW872 cells . We subsequently confirmed the binding of Sp1 and Sp3 using nuclear extracts from the chondrosarcoma cell line SW1353 , the osteosarcoma cell line MG63 and from primary human articular chondrocytes ( HACs; Figure S3A and S3B ) . We performed an oligonucleotide pull down assay using C and T allele DNA probes and then identified and quantified the binding of proteins to each allele using tandem mass tag ( TMT ) 6-plex isobaric labelling followed by mass spectrometry . The binding of activated RNA polymerase II transcriptional coactivator p15 ( P15; also known as SUB1 and PC4; HGNC:19985 ) was identified in both the C and T allele DNA samples . However , P15 was reproducibly found to be more abundant in the T allele sample , in comparison with the C allele sample with an average C/T ratio of 0 . 67 . This protein was absent in the background control sample . P15 does not have a known binding consensus sequence and we were therefore not able to use an EMSA to investigate competition for binding to the fluorescently labelled C and T allele probes . However , on the addition of an antibody targeting P15 , we observed a decrease in the two specific protein complexes binding to the two probes ( Figure 2D ) . This was also observed in SW1353 and MG63 cells and in HACs ( Figure S3C ) . Following our previous report that the DEAF-1 consensus competitor sequence was able to compete binding of proteins to C and T allele probes [9] , we investigated the effect of adding an antibody targeted against DEAF-1 to our EMSA reaction . We observed a supershifted complex in both C and T allele probe reactions , with the complex appearing to be more intense in the T allele probe sample ( Figure 2E ) . The supershifted complex was also confirmed using nuclear extract from HACs , with the protein complexes binding to the C and T allele probes being less intense than those observed in the SW872 cells ( Figure S3D ) . P15 was discovered by the oligonucleotide pull down experiment but this technique did not detect Sp1 , Sp3 or DEAF-1 , which were instead detected by the EMSA analysis . A possible explanation for this is the different binding conditions used , including different salt concentrations , in the pull down assay versus EMSA . To assess this , we repeated the EMSA using salt concentrations equivalent to those used in the pull down and observed that Sp1 and Sp3 were no longer able to bind to the C and T allele probes ( Figure S4 ) . We suspect therefore that this accounts for the different results obtained between pull down and EMSA . This result justifies our use of two distinct techniques for identifying trans-acting factors . Following the identification and confirmation of the binding of the Sp1 , Sp3 , P15 and DEAF-1 trans-acting factors to a GDF5 probe in vitro , we next sought to confirm the binding of these factors to the GDF5 locus in vivo using ChIP followed by PCR . In the PCR reaction we amplified the GDF5 exon 1 region , encompassing rs143383 , and the intensities of the PCR products were clearly greater following ChIP with anti-Sp1 , anti-Sp3 and anti-P15 antibodies relative to the IgG negative control antibody ( Figure 2F ) . This suggests that this region of GDF5 is enriched for Sp1 , Sp3 and P15 binding . We were unable to examine binding of DEAF-1 in vivo due to the unavailability of a specific ChIP grade antibody for this protein . After confirming the binding of these four factors to GDF5 , we then sought to assess if each factor regulates the expression of GDF5 . We first confirmed the expression of Sp1 , Sp3 , P15 and DEAF-1 in patient tissue samples; all four genes , in addition to GDF5 , were expressed in cartilage ( from OA and non-OA patients ) , synovium and fat pad ( Figure S5 ) . We next analysed the effect of Sp1 , Sp3 , P15 and DEAF-1 depletion on GDF5 expression by RNAi in the SW872 cells . The depletion of the mRNA for each gene was confirmed by real time RT-PCR and of Sp1 , Sp3 and P15 protein by immunoblotting ( Figures 3A and 3B ) . Due to the low expression levels of DEAF-1 within SW872 cells , we had difficulty in confirming the knockdown of the endogenous protein . We therefore confirmed that the siRNA is able to deplete DEAF-1 protein following the over expression of DEAF-1 EGFP fusion protein ( Figure S6 ) . The overall expression of GDF5 was increased following depletion of each factor . For Sp1 , Sp3 and P15 depletion , these increases in GDF5 expression were not significant , whilst a significant fold change ( p<0 . 001 ) was observed upon DEAF-1 knockdown ( Figure 3C ) . We next used allele specific real time PCR to assess if any of the four factors differentially affects expression of the two alleles of rs143383 , and as such could contribute to the DAE mediated by this SNP . Depletion of Sp1 and Sp3 resulted in small and non-significant increases in the C to T ratio ( ratio of 2 . 1 in the control ( NTsiRNA ) to 2 . 7 ( Sp1 siRNA ) or 2 . 4 ( Sp3 siRNA ) ) whilst P15 depletion did not alter the DAE ( Figure 3D ) . DEAF-1 depletion increased the DAE from a C/T ratio of 2 . 1 in the control ( NTsiRNA ) to 4 . 7 ( DEAF-1 siRNA ) and this was highly significant ( p<0 . 001 , Figure 3D ) . We confirmed the effect on overall GDF5 expression in SW1353 cells , with knockdown of the four factors increasing GDF5 expression . In line with that observed in SW872 cells , the increases in GDF5 expression were not significant following Sp1 , Sp3 and P15 depletion but a significant fold change was observed upon DEAF-1 knockdown in this chondrosarcoma cell line ( Figure S7 ) . Additionally Sp1 , Sp3 , P15 and DEAF-1 depletion experiments were performed in HACs . The depletion of the mRNA for each gene was confirmed by real time RT-PCR and of Sp1 , Sp3 and P15 protein by immunoblotting ( Figure S8; as for SW872 , endogenous DEAF-1 was not detectable in HACs ) . Depletion of P15 and DEAF-1 resulted in small and non-significant increases in GDF5 expression , whilst Sp3 depletion increased GDF5 expression significantly ( p<0 . 05 ) . These data suggest that all four factors are involved in the transcriptional activity of GDF5 , each repressing GDF5 expression , with DEAF-1 having significant repressive effects and also clearly contributing to GDF5 DAE in the SW872 cells . We next over expressed each of the four factors in combination with a reporter vector that contained the GDF5 promoter and the 5′UTR sequence encompassing rs143383 and which drove expression of the luciferase gene . We used two constructs , one containing the T allele and the other the C allele of the SNP . These experiments were performed in the chondrosarcoma cell line SW1353 . We first assessed what effect this single nucleotide difference mediated on luciferase activity and observed that the presence of a T allele at rs143383 significantly reduced the luciferase activity , with an average C/T allelic ratio of 1 . 2 ( p<0 . 001 , Figure 4A ) , confirming previous findings [9] . Over expression of Sp1 , Sp3 , P15 and DEAF-1 fusion proteins was then confirmed by immunoblotting and immunofluorescence ( Figure 4B and Figure S9 respectively ) . Over expression of Sp1 decreased the promoter activity of both C and T allele constructs , with a significant repressive effect on the T allele ( p<0 . 05; Figure 4A ) , significantly increasing the C/T ratio to 1 . 38 ( p<0 . 01 ) . Over expression of Sp3 decreased the promoter activity of both the C and T allele constructs , and this effect was significant with the T allele construct ( p<0 . 001; Figure 4A ) significantly increasing the allelic ratio to 1 . 48 ( p<0 . 001 ) . P15 over expression decreased the promoter activity of both alleles , however , this repressive effect was not significant ( Figure 4A ) . Finally , DEAF-1 over expression significantly repressed the promoter activity of both alleles ( C and T alleles p<0 . 001; Figure 4A ) , but most notably repressed the T allele construct , decreasing its activity to near that of the empty control and significantly increasing the allelic ratio to 1 . 37 ( p<0 . 01 ) . These results confirm that Sp1 , Sp3 and DEAF-1 are significantly repressing GDF5 expression , and this repression is greater for the T allele of rs143383 . Conversely , P15 only appears to be mediating a minor , non-significant repressive effect . We next assessed whether the repressive effects seen in the above experiment would be stronger if the factors were co-transfected and over expressed together . When Sp1 and Sp3 were jointly over expressed there was a significantly greater reduction in expression of both the C and the T alleles relative to when they were over expressed alone ( Figure 5A ) . Furthermore , the C/T allelic ratios significantly increased from 1 . 38 for the Sp1 over expression and 1 . 48 for the Sp3 over expression to 1 . 70 for the joint over expression ( p<0 . 001 for the joint over expression versus Sp1 alone and p<0 . 05 for the joint over expression versus Sp3 alone; Table S2 ) . When Sp1 and DEAF-1 were jointly over expressed there was a reduction in expression of both the C and T alleles relative to when they were over expressed alone ( Figure 5B ) . The C/T allelic ratios increased significantly from 1 . 38 for Sp1 and 1 . 37 for DEAF-1 to 1 . 55 for the joint over expression ( p<0 . 001 versus C/T ) . However , these C/T allelic ratio changes were not significant when compared with Sp1 or DEAF-1 over expression alone ( p = 0 . 1 ) . Finally , when Sp3 and DEAF-1 were jointly over expressed , the C/T allelic ratios increased from 1 . 48 for Sp3 and 1 . 37 for DEAF-1 to 1 . 6 for the joint over expression , and this was a significant C/T difference compared to DEAF-1 over expression alone ( p = 0 . 01 ) . Over expression of P15 in combination with Sp1 , Sp3 or DEAF-1 did not contribute any further significant repressive effects compared to over expression of the factors alone ( data not shown ) . Finally , we performed co-immunoprecipitation experiments using nuclear extracts from SW1353 cells to show that Sp1 , Sp3 , P15 and DEAF-1 directly interact . We observed co-immunoprecipitation of Sp1 when Sp3 and DEAF-1-EGFP were immunoprecipitated ( Figure S10A ) . In the reciprocal experiment , Sp3 and DEAF-1 were co-immunoprecipitated upon Sp1 immunoprecipitation ( Figure S10B and S10D ) . P15 was co-immunoprecipitated following Sp1 , Sp3 and DEAF-1 EGFP immunoprecipitation ( Figure S10C ) . Finally , Sp3 was co-immunoprecipitated following DEAF-1 EGFP immunoprecipitation ( Figure S10B ) , and the reciprocal experiment revealed DEAF-1 co-immunoprecipitation with Sp3 immunoprecipitation ( Figure S10D ) .
The rs143383 T allele has been reproducibly associated with increased risk of OA , and produces a lower level of expression of GDF5 relative to the C allele . This DAE is apparent in all tissues of the articulating joint and also within the rs143383 heterozygote cell line SW872 , which therefore provided us with an ideal model system to investigate the trans-acting factors mediating this DAE [9] , [23] . Using a variety of techniques we identified Sp1 , Sp3 , P15 and DEAF-1 as proteins that bind to the two alleles of rs143383 . Depletion of all four increased the expression of GDF5 , whilst DEAF-1 depletion significantly modulated the DAE . Conversely , the over expression of Sp1 , Sp3 and DEAF-1 repressed C and T allele expression , repressing the T allele more strongly . When over expressed together , DEAF-1 and Sp1 mediated the greatest overall repressive effect whereas over expression of Sp1 and Sp3 together mediated the greatest differential allelic effect , repressing the T allele to a greater extent than the C allele . Using co-immunoprecipitation we demonstrated that these four factors directly interact with each other . Overall therefore we have identified trans-acting factors that bind differentially to the alleles of rs143383 and which contribute to the DAE that is mediated by this important OA susceptibility locus . Sp1 and Sp3 are well characterized transcription factors that have a high degree of conservation between their zinc finger DNA binding domains ( 95% homology ) and which bind to related DNA sequences [24] , [25] . Sp1 is usually considered a potent activator of gene expression , although repressive activity has been reported , whereas Sp3 is known to possess both activator and repressor functions [26]–[28] . Both proteins are ubiquitously expressed and bind with high affinity to GC rich motifs , which are promoter elements present in a diverse range of genes . The proteins also form a multi-protein complex to synergistically regulate gene expression [29] . Promoters that do not contain a TATA binding site are commonly known to have an Sp protein-binding site . In these TATA-less promoters Sp1 has been reported to play a critical role in anchoring the basal transcription machinery to promote transcriptional initiation . Sp1 facilitates the binding of TFIID through binding to TBP ( TATA binding protein ) associated factors ( TAFs ) which then recruit RNA polymerase II [30] . GDF5 does not contain a TATA box and thus it appears likely that in binding to the GDF5 5′UTR , Sp1 and Sp3 may be mediating interactions with the basal transcriptional machinery to modulate transcription of this gene . In our EMSA experiments , a comparison of the complex formation of Sp1 and DEAF-1 to the GDF5 probes revealed that there is an abundance of Sp1 protein relative to DEAF-1 protein . DEAF-1 however has the most significant repressive effect on GDF5 expression . Sp1 is known to form homomultimers when it is bound to the promoters of genes , where it can serve as a docking site for the binding of other proteins [31] . This Sp1 multimerisation may account for the relative abundance of this protein . Sp1 and Sp3 have been previously reported to interact with HDAC1 in order to mediate gene repression [32] . Our analysis did not however provide evidence of HDAC1 binding to rs143383 or to its immediate flanking sequence . The importance of Sp1 and Sp3 during joint development is highlighted by the large number of target genes that they regulate , the expression of which are key for the formation of the joint and include SOX9 , COL1A1 and RUNX2 [33]–[36] . P15 is a small , highly abundant nuclear protein with multiple functions in transcription , replication and DNA repair [37] . As a transcriptional co-activator , P15 mediates functional interactions between transcription factors and the general transcription machinery [38] . P15 has also been reported to stabilize multi-protein complexes and has previously been reported to act as a co-activator of Sp1 , where it was reported to function as a linker between Sp1 and the pre-initiation complex ( PIC ) [39] , [40] . Repressor functions of P15 have also been reported [41] . P15 knockout mice are lethal , highlighting the important role of this factor during development; however heterozygous knockout mice display no overt phenotype indicating there may be a threshold level of P15 that is required for normal development . Sp1 , Sp3 and DEAF-1 were not identified by the oligonucleotide pull down experiment . We hypothesised that this may be due to the different salt conditions used between pull down and EMSA and we then demonstrated that this was the case . This highlights the importance of using more than one method for the discovery of trans-acting factors . Another difference between our pull down and EMSA experiments was the length of the genomic DNA sequence used , which was 212 bp in the pull down and 33 bp in the EMSAs . We chose to use a long sequence in the pull down in order not to limit the capture of proteins that may bind over large DNA regions . It is possible however that by using such a long sequence we captured non-specific proteins that may have disrupted the binding of Sp1 , Sp3 and DEAF-1 . The use of a shorter DNA sequence or of repeat concatamers of rs143383 and its immediate flanking sequence , combined with varying salt concentrations , may have led to the identification of Sp1 , Sp3 and DEAF-1 by the oligonucleotide pull down approach . Of all of the four trans-acting factors that we identified , DEAF-1 appears to repress GDF5 expression most significantly . The lack of a ChIP grade antibody precluded us from demonstrating the binding of DEAF-1 in vivo . However , the EMSA supershift that we observed combined with the significant changes in both overall and allelic GDF5 expression following DEAF-1 depletion , and the significant repressive effects observed following DEAF-1 over expression , provided us with compelling evidence that this trans-acting factor is modulating GDF5 expression at rs143383 . DEAF-1 is repressing the T allele more avidly , compared with the C allele , thus following DEAF-1 depletion we expected to observe a greater increase in the expression of the T allele , and a decrease in the C/T allelic ratio . Conversely , we observed an increase in the C/T allelic ratio . We believe this may be either a result of the incomplete depletion of DEAF-1 protein , or because the other factors forming part of the repressive complex are continuing to differentially repress GDF5 expression . DEAF-1 is expressed in many neuroendocrine and reproductive tissues and is expressed at high levels in the foetus , suggesting an important role during development [42] . DEAF-1 regulates the expression of a number of genes and its transcriptional activity can be modulated by a single base-pair change to its binding site , with its repressive regulation of the expression of the serotonin auto-receptor 1A ( 5HT1A ) gene reduced following a C to G transversion [43] , [44] . This study confirms our observation that the activity of DEAF-1 is sensitive to subtle changes in its binding sequence . DEAF-1 knockout mice display skeletal abnormalities including rib cage defects , with a large proportion of the animals suffering from defective neural tube closure that causes death shortly after birth [45] . Using our experimental data and the predicted binding regions for each protein we have prepared a model for how we believe Sp1 , Sp3 , P15 and DEAF-1 are interacting relative to rs143383 ( Figure 6 ) . The core consensus site for DEAF-1 is TCGG , which resides directly over the SNP , whereas the Sp1/Sp3 GC binding motif is immediately upstream . Although we have confirmed the binding of P15 to GDF5 both in vitro and in vivo , P15 is not mediating a significant repressive effect on GDF5 expression . We propose therefore that DEAF-1 , Sp1 and Sp3 are forming a repressive complex that forms directly over rs143383 and are differentially modulating the expression of the C and T alleles . P15 may be interacting with this complex and serving as a linker between Sp1 and the general transcription machinery . We have very recently identified YY1 as a transcriptional activator that binds 80 bp upstream of rs143383 , within the GDF5 promoter [46]; YY1 and Sp1 have previously been shown to jointly modulate the expression of genes and so it is possible that YY1 may indirectly interact with the complex at rs143383 [47] . The relevance of our results extend beyond OA , since the T allele of rs143383 has been associated with a number of other musculoskeletal phenotypes including congenital hip dysplasia [48] , Achilles tendinopathy [49] , lumbar disc degeneration [50] , variation in normal height , hip axis length , and an increased risk of fracture [51] , [52] . Transcription factors are now becoming more widely considered as targets for therapeutics to modulate the expression of genes . One approach that has proven effective in vivo and which is being considered for clinical application is the inhibition of transcription factors with molecules that mimic the transcription factor binding site [53] . This is known as transcription factor decoy and Sp1 has already been targeted using this approach in breast cancer [54] . The factors that we have identified could therefore serve as novel therapeutic targets , with their depletion restoring the expression levels of GDF5 in patients with the OA susceptibility T allele .
SW872 cells were cultured in Dulbecco's modified eagles medium: Hams F12 nutrient mix , GlutaMAX in a 3∶1 ratio ( Invitrogen , Life Technologies , Paisley , UK ) containing 5% ( v/v ) foetal bovine serum ( FBS ) , 100 U/ml penicillin and 100 µg/ml streptomycin ( Sigma-Aldrich , St . Louis , USA ) . SW1353 cells were cultured in Dulbecco's modified eagles medium: F12 ( 1∶1 ) ( Invitrogen ) containing 10% ( FBS ) , 100 U/ml penicillin , 100 µg/ml streptomycin and 2 mM L-glutamine ( Sigma-Aldrich ) . Monolayer cultures were maintained in vented T75 cm2 flasks at 37°C , in a 5% CO2 ( v/v ) atmosphere . MG63 cells were cultured in Dulbecco's modified eagles medium ( Invitrogen , Life Technologies ) containing 10% ( v/v ) foetal bovine serum ( FBS ) , 100 U/ml penicillin , 100 µg/ml streptomycin ( Sigma-Aldrich ) and 2 mM of L-glutamine ( Sigma-Aldrich ) . Human articular chondrocytes ( HACs ) were isolated from articular cartilage obtained from patients with osteoarthritis undergoing total hip or knee replacement surgery . HACs were also obtained from non-OA patients who had undergone joint replacement due to neck-of-femur ( NOF ) fracture . Ethical approval and informed consent were obtained prior to surgery ( research ethics committee reference 09/H0906/72 issued by the UK National Research Ethics Service ) . Enzymatic digestion and HAC culture was performed as previously described [55] . Genomic DNA , total RNA and total protein were simultaneously extracted from SW872 cells using a spin column extraction kit according to the manufacturer's instructions ( Nucleospin Triprep , Macherey-Nagel , supplied by Fisher , UK ) . Nucleic acids were quantified using a NanoDrop ND-1000 Spectrophotometer ( NanoDrop Technologies , Wilmington , USA ) . 1 µg of total RNA was DNase treated with 2 units of Turbo DNase ( Ambion , Life Technologies ) and reverse transcribed using Moloney Murine Leukemia Virus ( M-MLV; Invitrogen ) following manufacturer's instructions . Gene expression of GDF5 , SP1 , SP3 , P15 and DEAF-1 was determined by real time RT-PCR and normalised to the housekeeping gene HPRT1 using the delta ct method ( 2− ( ct test gene ) - ( ct HPRT1 ) ) . Gene expression assays were purchased from either Applied Biosystems ( ABI , Life Technologies ) or Integrated DNA Technologies ( IDT , Belgium ) . Differential Allelic Expression ( DAE ) analysis , to assess the expression of the C and T alleles of rs143383 , was performed using a custom SNP genotyping assay ( ABI , Life Technologies ) containing forward and reverse primers and allele specific probes ( VIC or FAM labelled ) . For analysis , the cDNA C/T allelic ratio was normalised to the genomic DNA ( gDNA ) C/T allelic ratio ( representing a 1∶1 ratio ) for each treatment group . An ABI PRISM 7900HT Sequence detection System was used for all real time PCR quantification . In SW872 cells for both overall gene expression and DAE analysis , three independent experiments were performed , with three biological replicates per experiment ( n = 9 ) . For each DNA and cDNA sample we performed three pipetting replicates , which were averaged prior to analysis . For SW1353 cells , three independent experiments were performed ( n = 3 ) . For HACs twelve biological replicates were performed . Statistical analysis of % knockdown was performed using a Students 2-tailed t-test whilst the one-way analysis of variance ( ANOVA ) test was used for GDF5 fold change and DAE analysis . The primer and probe sequences are listed in Table S3A . For the extraction of nuclear protein , cells were seeded at a density of 15×106 on 500 cm2 plates ( Corning , USA ) . Two buffers were used sequentially to isolate nuclear proteins; following centrifugation ( 10 , 000 g 30 seconds ) cell pellets were re-suspended in 1 ml hypotonic buffer ( 10 mM HEPES pH 7 . 6 , 1 . 5 mM MgCl2 , 10 mM KCl , 1 mM DTT , 10 mM NaF , 1 mM Na3VO4 , 0 . 1% Tergitol ( v/v ) , 1× complete protease inhibitor cocktail tablet per 50 ml solution ( Roche , UK ) ) and incubated on ice for 15 minutes . After a second centrifugation , the cell pellet was re-suspended in 500 µl high salt buffer ( 20 mM HEPES , pH 7 . 9 , 420 mM NaCl , 20% glycerol ( v/v ) , 1 mM DTT , 10 mM NaF , 1 mM Na3VO4 , 1× complete protease inhibitor cocktail tablet per 50 mls of buffer ) and incubated on ice for 30 minutes . Following a final centrifugation ( 10 , 000 g , 2 minutes ) , the supernatant containing nuclear protein was stored at −80°C . PROMO 3 . 0 , TESS , and TransFac online databases were used to predict protein binding to the C and T-alleles of GDF5 . Fluorescently labelled oligonucleotides for both alleles ( Eurofins MWG Operon , Ebersberg , Germany ) were re-suspended to a final concentration of 100 pmol/µl in water ( Sigma-Aldrich ) . Single-stranded oligonucleotides were incubated at 95°C for 5 minutes in a solution containing EMSA annealing buffer ( 100 mM Tris-HCl pH 7 . 5 , 500 mM NaCl , 10 mM EDTA ) to a final concentration of 20 pmol/µl and cooled slowly to room temperature for 2 hours to generate double stranded annealed probes . The annealed probes were diluted to 100 fmol/µl in water ( Sigma-Aldrich ) prior to the EMSA reaction . A native 5% ( weight/volume ) polyacrylamide gel was prepared the day before the EMSA and allowed to set at 4°C overnight . The EMSA was then carried out as per manufacturer's instructions using the Odyssey Infrared EMSA kit ( LiCor Biosciences , Cambridge , UK ) . The optimal binding reaction contained 1× Binding Buffer , 2 . 5 mM DTT , 1 µg Poly ( dI∶dC ) , 5 mM MgCl2 , 200 fmol annealed oligonucleotide and 5 µg nuclear extract . The gel was visualised using an Odyssey Infrared Imager ( LiCor Biosciences ) . For competition assays to test binding of predicted proteins , single stranded unlabelled oligonucleotides ( Sigma-Aldrich ) containing the consensus binding sequence of the protein were annealed as previously described for the labelled probes . For supershift EMSAs , 2 µg of antibody was added to the binding reaction . Table S1 lists the nucleotide sequences of the labelled probes and unlabelled competitor sequences . Table S4 provides details of the antibodies . A 212 bp DNA region encompassing rs143383 was amplified by PCR using a biotinylated 5′ primer and unlabelled 3′ primer ( Sigma-Aldrich ) ( Table S3B ) . Two PCRs were performed , using homozygous C or T template DNA at the polymorphic site . 40 pmol of PCR product was coupled to 2 mg of Streptavidin Dynabeads ( Invitrogen ) following the manufacturers instructions . A sample containing no DNA was used as a control . DNA-beads complexes were blocked as described previously [56] . SW872 cell nuclear lysates were extracted as described above , transferred to a tube for dialysis ( Tube-O-dialyzer , VWR , UK ) , and dialyzed in a low salt buffer ( 20 mM HEPES pH 7 . 9 , 20% ( v/v ) glycerol , 0 . 1 M KCl , 0 . 2 mM EDTA , 0 . 5 mM PMSF , 0 . 5 mM DTT ) for 4 hours at 4°C . The buffer was replaced and the lysates were dialysed for a further 16 hours at 4°C . Following this , the nuclear lysate was pre-cleared for 1 hour with 50 µl Streptavidin Dynabeads ( Invitrogen ) . DNA-bead complexes were then re-suspended in 1 mg of the prepared SW872 protein extract and incubated for 2 hours at 4°C with shaking . Beads were washed six times with BC-100 buffer and re-suspended in 1× SDS sample buffer ( 4% SDS , 0 . 2 M Tris-HCL , 4% glycerol , 0 . 01% bromophenol blue , 2% β-mercaptoethanol ) . Complexes were eluted from the beads following incubation at 95°C for 5 minutes and isolated following magnetic separation . The samples ( CC , TT and no DNA ) were loaded on to a 12% gel , and subject to separation by electrophoresis , followed by coomassie blue staining . Quantitative mass spectrometry was performed as previously described [57] . Briefly , following peptide digestion overnight using trypsin , labelling of the three conditions was carried out with a TMT isobaric mass tagging kit ( Thermo Scientific , Surrey , UK ) . Labelled samples were mixed prior to off-gel fractionation of the peptides . Following liquid chromatography tandem mass spectrometry ( LC-MS/MS ) , quantitative analysis was carried out using ProteinExplorer , version 1 . 0 ( Thermo Scientific ) and the search engine MASCOT ( Matrix Science Company ) used for identification of proteins . These results were then sorted according to detection in the background sample and ranked with the most robust hits being proteins with high confidence values , based on the identification of more than 2 unique peptide sequences , the coverage of peptides in the protein and those with low variability between peptide quantification values . Proteins known to have a role in transcriptional activation or repression were prioritised for further analysis . ChIP experiments were performed as recommended by the manufacturer using the Magna ChIP A kit ( Merck , Millipore , Consett , UK ) . Briefly , SW872 cells were cultured until 70% confluent on 500 cm2 culture plates ( Corning ) . Cells were cross linked for 10 minutes with 1% ( w/v ) formaldehyde , 1 . 25 M glycine was then added for 5 minutes to quench unreacted formaldehyde . The cells were then washed twice and harvested in cold PBS containing protease inhibitors . Cells were then centrifuged for 8 minutes at 720 g , re-suspended in lysis buffer and incubated on ice for 20 minutes . The cell suspension was sonicated using a Soniprep150 probe sonicator ( MSE UK , London , UK ) to shear the chromatin , and then pre-cleared with magnetic protein A beads for 30 minutes at 4°C . 100 µg of chromatin was incubated with rotation overnight at 4°C in addition to 10 µg of either rabbit IgG antibody ( negative control ) , anti-acetyl histone H3 ( positive control ) or 10 µg of the antibody of interest and 40 µl magnetic protein A beads ( the antibodies used are listed in Table S4 ) . Using a magnetic separator immunoprecipitated DNA/protein complexes were isolated and washed as recommended . Cross-linking was reversed by incubating the DNA/protein complexes and the input control ( 10% of sonicated chromatin ) in elution buffer with proteinase K at 65°C for 2 hours . DNA was purified and analysed by PCR ( Table S3B ) . 2 µl of immunoprecipitated DNA was added to a 15 µl PCR reaction , the thermocycling conditions as follows; 94°C 14 minutes , followed by 32 cycles of 94°C 30 seconds , 57°C for 30 seconds ( annealing temperature for GDF5 ChIP primers ) , 72°C for 30 seconds and a final step of 72°C for 5 minutes . PCR products were electrophoresed through a 2% ( w/v ) agarose gel containing ethidium bromide . Three ChIP experiments in total were performed for each antibody , each showing consistent results . SW872 cells were seeded at 350 , 000 cells per well in a 6 well culture plate ( Costar , UK ) . After 24 hours , cells were transfected using 100 nM Dharmacon ON-TARGETplus Smartpool siRNAs targeted against SP1 , SP3 , P15 , DEAF-1 and a Non-Targeting Pool control in addition to Dharmafect 4 lipid reagent ( Thermo Fisher , UK ) . After 48 hours the cells were harvested , nucleic acid and protein isolated and RNA reverse transcribed as described above . Depletion of mRNA expression was calculated compared to cells transfected with the ON-Targetplus Non-Targeting Pool control siRNA ( Thermo Fisher ) . SW1353 cells were seeded at 250 , 000 cells per well in a 6 well culture plate and transfected as described for SW872 cells using Dharmafect 1 lipid reagent ( Thermo Fisher ) . Human articular chondrocytes were seeded at 300 , 000 cells per well in a 6 well culture plate and transfected as described for SW872 cells using Dharmafect 1 lipid reagent ( Thermo Fisher ) . The GDF5 promoter and part of the 5′UTR region spanning −97 to +305 ( relative to the transcriptional start site ) was subcloned from the GDF5 pGL3-Basic vector [46] into the Mlu/BglII sites of the purified pGL3-Enhancer Vector ( Promega , UK ) . The Sp1 , Sp3 and P15 open reading frames ( ORF ) were amplified from cDNA using the primers listed in Table S3B , ligated into the EcoR1 and SacII sites of the pEGFP-N1 vector ( Clontech ) and transformed into MACH1 competent bacterial cells ( Invitrogen ) . The DEAF-1-EGFP-N1 expression plasmid was kindly donated by C . Garrison Fathman [58] . Plasmid DNA was extracted using a Qiagen Maxiprep Kit ( Qiagen , Crawley , UK ) . SW1353 cells were seeded at a density of 17 , 500 cells per well in a 48-well cell culture plate ( Costar , UK ) and cultured for 48 hours prior to transfection . Cells were transfected with 2 µg of plasmid DNA ( containing 1 µg of GDF5 pGL3 enhancer vector and several combinations of either 1 µg empty pEGFP-N1 vector , 500 ng empty pEGFP-N1 and one of the transcription factor expression plasmids , or 500 ng each of two transcription factor expression plasmids ) in addition to 15 ng of pTK-RL Renilla using ExGen 500 in vitro transfection reagent ( Fermentas , York , UK ) . Four wells were transfected per condition and a total of three individual experiments were performed . After 24 hours , transfected cells were lysed and luciferase and renilla activity measured using the Dual Luciferase Assay system ( Promega , UK ) with the MicroLumat Plus LB96V luminometer ( Berthold Technologies UK , Harpenden , UK ) . Statistical analysis was performed using the Students 2-tailed t-test . To assess siRNA knockdown of our candidate proteins and successful over expression , total protein was isolated as described , quantified ( Bradford reagent , Expedeon ) and 10 µg was resolved on SDS-10% ( w/v ) polyacrylamide gels . Protein was then transferred to Immobilon-P PVDF membranes ( Merck Millipore ) . The antibodies detailed in Table S4 were used to assess protein levels following siRNA knockdown and over expression in SW872 and SW1353 cells respectively . A monoclonal β-Actin antibody was used as a loading control . For the examination of protein over expression , 250 , 000 cells per well were seeded in 6-well culture dishes and transfected with plasmid vectors and ExGen500 as described for the 48-well plate , but the relative amounts of each were increased according to the culture volume . For the over expression of DEAF-1 EGFP , followed by DEAF-1 siRNA treatment , SW872 cells were transfected with DEAF-1 EGFP as described above , and after 6 hours the cells were treated with NTsiRNA or with DEAF-1 siRNA and then harvested after 48 hours . To examine the overexpression of EGFP-N1 vectors , SW1353 cells were seeded at a density of 10 , 000 cells/well in a chamber slide ( Nagel Nunc International , USA ) and after 48 hours , transfected with 1 µg plasmid vector using ExGen 500 . After 24 hours , cells were washed in PBS and fixed with 4% ( w/v ) Paraformaldehyde in PBS for 10 minutes , washed again in PBS and mounted using vectashield with DAPI ( 4′6-diamidino-2-phenylindole ) ( Vector Laboratories , Burlingame , CA ) . Fluorescence was detected using a LEICA DMLB fluorescent microscope and a SPOT-RT camera . SW1353 cells were cultured until 70% confluent on 500 cm2 culture plates ( Corning ) and nuclear protein was extracted as described above . 10 µg of antibody and 200 µg of nuclear extract ( diluted 1 in 5 in high salt lysis buffer ) was incubated over night at 4°C with shaking . 70 µl magnetic protein A beads were added to each immunoprecipitation , and this mixture was incubated at 4°C with shaking for 4 hours . Using a magnetic separator , immunoprecipitated protein complexes were isolated and washed with lysis buffer twice and with PBS once . The magnetic beads were then re-suspended in Laemmli buffer and the samples were heated to 95°C for 5 minutes . The supernatant was then taken forward for analysis by SDS-PAGE .
|
GDF5 is an important growth factor that plays a vital role in the development and repair of articulating joints . rs143383 is a polymorphism within the regulatory region of the GDF5 gene and has two allelic forms , C and T . Genetic studies have demonstrated that the T allele is associated with an increased risk of osteoarthritis in a range of ethnic populations whilst previous functional studies revealed that this allele mediates its effect by producing less GDF5 transcript than the C allele . In this study , we sought to identify transcription factors that are binding to rs143383 and that are responsible for mediating this differential level of expression . Using two different approaches we have identified four factors and our functional studies have revealed that three of these factors repress GDF5 expression and that DEAF-1 modulates the differential expression of the two rs143383 alleles . The factors that we have identified could serve as novel therapeutic targets , with their depletion restoring the expression levels of GDF5 in patients with the osteoarthritis susceptibility T allele . The relevance of our results extends beyond osteoarthritis , since the T allele of rs143383 is also a risk factor for a number of other musculoskeletal diseases .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"rheumatology",
"gene",
"expression",
"genetics",
"gene",
"regulation",
"molecular",
"genetics",
"biology",
"human",
"genetics",
"genetics",
"of",
"disease",
"dna",
"transcription",
"gene",
"function"
] |
2013
|
The Identification of Trans-acting Factors That Regulate the Expression of GDF5 via the Osteoarthritis Susceptibility SNP rs143383
|
Hemorrhagic fever with renal syndrome ( HFRS ) is a rodent-borne disease caused by Hantaviruses . It is endemic in all 31 provinces , autonomous regions , and metropolitan areas in mainland China where human cases account for 90% of the total global cases . Shandong Province is among the most serious endemic areas . HFRS cases in Shandong Province were first reported in Yutai County in 1968 . Since then , the disease has spread across the province , and as of 2005 , all 111 counties were reported to have local human infections . However , causes underlying such rapid spread and wide distribution remain less well understood . Here we report a spatiotemporal analysis of human HFRS cases in Shandong using data spanning 1973 to 2005 . Seasonal incidence maps and velocity vector maps were produced to analyze the spread of HFRS over time in Shandong Province , and a panel data analysis was conducted to explore the association between HFRS incidence and climatic factors . Results show a rapid spread of HFRS from its epicenter in Rizhao , Linyi , Weifang Regions in southern Shandong to north , east , and west parts of the province . Based on seasonal shifts of epidemics , three epidemic phases were identified over the 33-year period . The first phase occurred between 1973 and 1982 during which the foci of HFRS was located in the south Shandong and the epidemic peak occurred in the fall and winter , presenting a seasonal characteristic of Hantaan virus ( HTNV ) transmission . The second phase between 1983 and 1985 was characterized by northward and westward spread of HFRS foci , and increases in incidence of HFRS in both fall-winter and spring seasons . The human infections in the spring reflected a characteristic pattern of Seoul virus ( SEOV ) transmission . The third phase between 1986 and 2005 was characterized by the northeast spread of the HFRS foci until it covered all counties , and the HFRS incidence in the fall-winter season decreased while it remained high in the spring . In addition , our findings suggest that precipitation , humidity , and temperature are major environmental variables that are associated with the seasonal variation of HFRS incidence in Shandong Province . The spread of HFRS in Shandong Province may have been accompanied by seasonal shifts of HTNV-dominated transmission to SEOV-dominated transmission over the past three decades . The variations in HFRS incidence were significantly associated with local precipitation , humidity , and temperature .
Hemorrhagic fever with renal syndrome ( HFRS ) , a rodent-borne disease caused by Hantaviruses ( HV ) , is characterized by fever , acute renal dysfunction , and hemorrhagic manifestations . HFRS , initially described clinically at the turn of the 20th century , is primarily distributed in the Asian and European continents , and worldwide approximately 150 , 000 to 200 , 000 hospitalized HFRS cases are reported each year , with the majority occurring in developing countries [1] . HFRS is widely distributed and a major public health concern in China . At present , it is endemic in all 31 provinces , autonomous regions , and metropolitan areas in mainland China where human cases account for 90% of the total global cases [2] . In China , HFRS is mainly caused by two types of Hantaviruses , i . e . , Hantaan virus ( HTNV ) and Seoul virus ( SEOV ) , each of which has co-evolved with a distinct rodent host [3] . HTNV , which causes a more severe form of HFRS than SEOV does , is associated with Apodemus agrarius , while SEOV is typically carried by Rattus norvegicus . Occurrence of HFRS cases is seasonal with a bimodal pattern and studies suggest that the pattern is linked to varying transmission dynamics of the two serotypes of HVs among their animal hosts - HTNV-caused HFRS cases occur year-round but tend to peak in the winter while SEOV-caused infections typically peak in the spring [4]–[8] . Historically , Shandong Province bears the largest HFRS burden in China - the cumulative human cases accounted for 1/3 of the national total [9] . This study aims to , through the use of a 33-year's ( 1973 to 2005 ) record of HFRS cases in Shandong Province , characterize the spatial and seasonal patterns of HFRS distribution and spread and explore associations between meteorological factors and such patterns of distribution and spread of the disease .
The study area covers Shandong Province , a coastal province in Eastern China , located between latitude 34°25′ and 38°23′ north , and longitude 114°35′ and 112°43′ east ( Figure 1 ) . In Shandong Province , the central area is mountainous , and the eastern and the southern areas are hilly . The north and northwest parts of Shandong are composed of the alluvial plain of the Yellow River , which is part of the North China Plain . Plains and basins , mountains and hills , and rivers and lakes make up 63% , 34% , and 3% of total area of the province , respectively . The province includes 111 counties belonging to 17 regions with a total land area of 156 , 700 square kilometers and a population of about 90 million . In 1950 , HFRS was included on the list of Class B Notifiable Diseases in China . Since then , data reporting has followed a standard protocol determined by Chinese Center for Disease Control and Prevention and consistent throughout the country [2] , [5] . Prior to 1983 , a HFRS case was determined by a set of clinical criteria , as defined by a national standard . Since then , antibody-based serological tests ( e . g . MacELISA , IFA ) were also used and coupled with the clinical criteria [4] , [7] . Clinical diagnosis criteria include: exposure history ( i . e . exposure to rodents and their excreta , saliva , and urine within two months prior to the onset of illness ) ; acute illness with at least two of the following clinical symptoms ( i . e . fever , chill , hemorrhage , headache , back pain , abdominal pain , acute renal dysfunction , and hypotension ) ; experience or partial experience of the 5 phases of disease course ( i . e . fever , hypopiesis , oliguresis , hyperdiuresis , and recovery ) ; and abnormity of blood and urine routine parameters [10] . In this study , records for HFRS cases during 1973–2005 were obtained from the Shandong Notifiable Disease Surveillance System ( SNDSS ) and were processed by county and by month . Disease variables ( e . g . cases , deaths , and incidence of HFRS ) were collected and geo-referenced on a digital map of Shandong Province using ArcGIS 9 . 2 ( ESRI Inc . , Redlands , CA , USA ) . Demographic data were also obtained from SNDSS . In addition , monthly meteorological data covering 700 surveillance stations in mainland China from 1973 to 2005 were collected from the China Meteorological Data Sharing Service System ( http://cdc . cma . gov . cn/index . jsp ) . Raster format of meteorological data including monthly average temperature , relative humidity , and monthly cumulative precipitation was created using a spatial interpolation method ( Kriging model ) . The monthly meteorological factors for each county of Shandong Province from 1973 to 2005 were then extracted in ArcGIS 9 . 2 ( ESRI Inc . , Redlands , CA , USA ) . To characterize the spatial and seasonal patterns of HFRS distribution , monthly incidences from 1973 to 2005 and for different epidemic phases in Shandong Province were plotted . Furthermore , annualized incidences and the proportion of monthly average incidence over different epidemic phases for each county were mapped in gradient colors and pie charts , respectively . To explore the diffusion trend of HFRS endemic areas , a vector velocity map [11] of HFRS spread was developed using trend surface analysis ( TSA ) [12]–[14] . TSA is a global smoothing method using polynomials with geographic coordinates , as defined by the central point of each county's polygon . In this study , when the first HFRS case was reported for each county , a trend surface on the month-year was created to explore the diffusion patterns and corridors of spread over time . The month-year of the first recorded case was identified for counties in the database . The x- and y-coordinates of county centroids were then derived according to an Albers conical equal area projection using the Shandong Provincial map in ArcGIS 9 . 2 ( ESRI Inc . , Redlands , CA , USA ) . Least square regression using quadratic polynomials of the x- and y-coordinates to predict year of first reported case was conducted using STATA 9 . 1 software ( StataCorp LP , Texas , USA ) [15] . Partial differential equations ( Δyear/ΔX and Δyear/ΔY ) were derived from the fitted model , generating a vector of the magnitude ( i . e . slope ) and direction of the diffusion trend of HFRS endemic areas for each location . The square root of the slope equates to the velocity of diffusion , as reflected by the size of the arrow on the maps . To explore associations between the HFRS incidence and meteorological factors from 1973 to 2005 , Granger causality ( G-causality ) tests were performed using the monthly HFRS incidence and meteorological factors ( including monthly average temperature , monthly cumulative precipitation , and monthly average relative humidity ) in EViews 3 . 1 software ( Inst . of System Science , Irvine , CA , USA ) . To assess possible time-lag effects resulting from meteorological factors , time lags from 1 to 3 months were included in the analysis [16] . Based on the G-causality analysis at the provincial level ( without reference to count ) , we further conducted panel data analysis at the county level . The panel data analysis was utilized to analyze the multiple cross-sectional data combining time-series and has the advantage of a large number of data points , which increased the degrees of freedom and reduced the collinearity among explanatory variables [17] , [18] . In this study the panel data involved two dimensions of observations that included time series observations ( monthly HFRS incidence and monthly averages of meteorological factors from 1973 to 2005 ) from all 111 counties in Shandong Province . The panel Poisson model with fixed effects was used to assess the impact of the three meteorological factors on HFRS incidence at varying time lags [15] . The percentage change ( PC ) in incidence in response to the change of a variable by a given amount ( which is equal to 100* ( exp ( coefficient ) -1 ) ) , 95% confidence intervals ( CIs ) , and P-values were estimated using the maximum likelihood method . For PC estimation , a 10 mm difference was used for monthly cumulative precipitation , while 10°C and 10% differences were used for monthly average temperature and monthly average relative humidity , respectively . Univariate analysis was conducted to examine the effect of individual variables . Additionally , the time lag variables from 1 to 3 months for the three meteorological factors were tested . Multivariate analysis was then performed using variables with a P-value <0 . 1 from the univariate analysis as covariates . Correlations between covariates were quantitatively assessed and models were optimized by comparing -2 log likelihood when correlated variables were added or removed one by one . It was discovered that adding these variables with the smallest -2 log likelihood value for precipitation , humidity , and temperature , respectively , could derive a more accurate model . STATA ( Version 10 . 0 ) was used in the panel data analysis ( StataCorp LP , Texas , USA ) .
The first HFRS cases in Shandong were reported in 1968 in Yutai County in the southwest region of the province , and no cases were reported again until 1973 . Since then , new cases emerged in the central and southern parts of the province and endemic areas began expanding throughout Shandong Province . A total of 282 , 442 HFRS cases were reported in Shandong Province from 1968–2005 . Because only one case was reported between 1968 and 1972 , the epidemic curve was created to show the temporal distribution of HFRS from 1973–2005 in Shandong Province . The incidence curve over the 33-year period is reasonably characterized by three phases: Phase I spanning from 1973 to 1982 with a typical fall-winter peak of incidence , Phase II covering the period between 1983 and 1985 with an emerging spring peak of incidence , and Phase III spanning from 1986 to 2005 with a dominance of spring peak of incidence ( Figure 2 ) . Figure 2a shows the monthly epidemic curve from 1973 to 2005 in Shandong while Figure 2b shows seasonal shifts of incidence distributions during the three phases . Clearly a bimodal pattern of incidences is seen during phase II and III where a rapid increase of HFRS incidence was shown in both the spring and the fall-winter season during phase II . During phase III the incidence in the spring season continued at approximately the same level , but then declined quickly in the fall-winter season . Figure 3 shows annualized incidences and the proportion of monthly average incidence over three epidemic phases for each county . During phase I , the main endemic areas of HFRS are located in south-central Shandong Province with a single epidemic peak in the fall-winter season mapped by the red color of the pies . In phase II , newly established endemic areas emerged in northwest and southwest Shandong Province , an epidemic peak in the spring season is mapped by the green color in the pies ( Figure 3 ) . In phase III , we found that the epidemic peak in the spring season remained predominant in almost all counties in Shandong Province ( Figure 3 ) . Figure 4 summarizes the spatial trend of expansion of HFRS endemic areas in Shandong Province . Figure 4a shows that expansion of HFRS endemic areas over the past decades since the 1960s as shown in different colors and the number in each county indicates the order ranked by the month-year of the first recorded case in the county . The velocity and direction of diffusion for each coordinate location were mapped to show the movement and instantaneous rate of HFRS diffusion in Shandong Province over the study period . Trend surface analysis with high-order polynomials is sensitive to data anomalies at the edge of the study area [11] . Less data are available at the study area boundaries; velocity vector size and direction are therefore less reliable and may not be accurate at the edge of the study area . For these reasons , 18 velocity vectors were removed from the vector diffusion map . Velocity of movement was lowest early in the epidemic ( Figure 4b ) , when HFRS spread from its primary focus in southern Shandong Province . The epidemic diffused outward with a higher velocity of movement from the south-center Shandong Province in 1980s . The disease moved distinctly north , east , and west into all counties . These results are consistent with detection of cases in 2005 in the northernmost region of Shandong Province . Pairwise G-causality analysis showed the association between HFRS incidence and meteorological factors with different time lags at the provincial level ( Table 1 ) . The results indicated that precipitation and humidity are G-causalities of HFRS incidence , as significant one-way association was seen between meteorological factors and the HFRS incidence but not vice versa . The association between the average temperature and HFRS incidence is not clear from this analysis , as a mixed picture was seen ( i . e . two-way association ) . In the panel data analysis , univariate analysis showed the effects of monthly cumulative precipitation , monthly average relative humidity , monthly average temperature and their time-lag variables from 1 to 3 months separately , and indicated that all variables , except for monthly average relative humidity with the 3-month lag , appeared to be significant factors ( Table 2 ) . Multivariate panel data analysis demonstrated that the three variables , monthly cumulative precipitation with 1-month lag , monthly average relative humidity with 1-month lag and monthly average temperature with 2-month lag , were significantly associated with HFRS incidence ( Table 2 ) .
This study , utilizing a longitudinal dataset spanning 33 years , characterized the spatiotemporal distribution and three phases of HFRS epidemics in Shandong Province . Over the past three decades , HFRS endemic areas have spread from their initial centers - Rizhao , Linyi , and Qingdao regions - towards the northern , western , and eastern parts of the province . Notably , shifts of seasonal peaks of HFRS , as characterized by the three epidemic phases , are suggested to be associated with shifts of causal agents of HFRS - HTNV and SEOV , each with different epidemiological characteristics . For example , the HFRS epidemic shift from phase I ( 1973 – 1982 ) to phase II ( 1983–1985 ) suggest that SEOV may have emerged as dominant causal agent for the spring peak of HFRS incidence . Early studies reported that the transmission of Hantaviruses through A . agrarius mice peaks in the winter , while R . norvegicus rat-associated infections mainly occurred in the spring [4]–[7] . Given the well-documented evidence and the three-phase pattern observed in the present study , we infer that HFRS in phase I ( 1973–1982 ) and phase II ( 1983–1985 ) was primarily caused by HTNV in southeastern Shandong , and SEOV was responsible for the HFRS epidemics after phase II in most areas of Shandong Province . The seasonal and spatial distributions of HFRS cases by these two agents are consistent with the distributions of relevant rodent species - A . agrarius was found predominantly on the eastern coast , and R . norvegicus was distributed in almost all areas of Shandong Province according to surveillance studies of HFRS in China since the 1980s [6] , [7] , [19] Although cross-protective immunity to hantaviruses exists in recovered humans , the type of HFRS endemic areas largely depends on predominant reservoirs and their populations . In mainland China , antigen-positive A . agrarius , Apodemus peninsulae , and R . norvegicus rats remain predominant in rural areas , forest areas , and urban areas , respectively [8] . Noticeably , the newly-established HFRS endemic areas in mainland China since the 1990s , also including those in Shandong Province , are mostly associated with SEOV ( e . g . peridomestic rodents-associated ) with a characteristic spring peak of human infections [4] , [20] , [21] . The results suggest that prioritizing control efforts on peridomestic rodents in residential areas in the spring and on sylvatic rodents in the late autumn and early winter might provide an effective method to target the specific Hantaviruses that causes HFRS . Our meteorological factor analysis shows monthly cumulative precipitations with 1-month lag , monthly average relative humidity with 1-month lag , and monthly average temperature with 2-month lag are significantly associated with the seasonal variation of HFRS incidence in Shandong Province . As we know , rodent populations , the reservoirs of HFRS , respond rapidly to conducive weather conditions [22] . The relationship between rodent population dynamics and meteorological factors is complex , varying by different rodent species and climate regions [23] . These complicated relationships may have different influences on disease transmission . For example , heavy precipitation followed by increased grass seed production was associated with higher deer mouse densities that caused an outbreak of hantavirus pulmonary syndrome in the Four Corners region of the USA ( the New Mexico area ) [24]–[27] . However , excessive rainfall could have a negative impact on rodents by destroying their habitats in Eastern China [28] , [29] . In addition , frequent rain may decrease the likelihood of rodent-to-rodent contact , rodent-to-human contact , and virus transmission due to decreased rodent activity and reduced human exposure [28] . However , underlying mechanisms for the negative correlations between HFRS incidence , temperature , and relative humidity are not yet clear . As a whole , HFRS incidence in the most recent past two decades was highest in the frigid-temperate zone , mostly in northeastern China , followed by the warm-temperate zone , with lower incidence seen in southeastern China , where there are higher temperatures , higher humidity , and greater precipitation [2] , [8] . This is consistent with the results of our analyses of meteorological factors . Thus , we assume meteorological factors may affect rodent dynamics and activity as well as infectivity of Hantavirus . The results support further research related to rodent host ecology . The current study suggests that climate may be used as a predictor of the intensity of HFRS transmission in a larger geographical area , however , future research is needed to better understand the underlying mechanistic effects , in particular those related to rodent ecology . The transmission of Hantaviruses to humans is also related to other factors such as human activities , farming patterns , and rodent abatement strategies [30] . HVs are primarily transmitted from rodent hosts to humans by aerosols generated from contaminated wastes ( e . g . urine and feces ) of rodents , and to a lesser extent , possibly by contaminated food or rodent bites [31] , [32] . We speculate that socio-environmental changes in the past three decades may have impacted human-rodent interactions . From the 1970s to the 2000s , rural areas in Shandong , like many areas in the rest of country , experienced economic transformation and farm mechanization ( Figure 2a ) . Phase I described in this study occurred during the beginning of this transformation , possibly reflecting greater exposure to rodents in the field during the extensive farming activities . During this time , crop storage in households was relatively rare due to the commune system . HFRS cases during this time were probably largely due to sylvatically acquired HTNV . During phase II , from 1983 to 1985 , higher farm output and crop yield resulted in contacts between human and rodents both in fields and in residences . In addition , relatively less precipitation during this phase might have increased the likelihood of rodent-to-human contact and virus transmission [27] . Therefore , HFRS incidence in this phase reflected more or less equally high both in the fall-winter and in the spring . For phase III of the HFRS incidence after 1986 , the shift of manual to mechanized farming and increased storage of crop/food in households resulted in a greater frequency of human-rodent interactions in residences , thus causing a high spring and low winter pattern of incidence . In addition , the annual incidence of HFRS declined from 26 . 0 to 3 . 6 per 100 , 000 people during the 1995–2005 period , which may be related to improved housing conditions , better environmental sanitation , a transformation of farm mechanization in rural area , and control measures , in addition to the influence from climate variation in which the negative effects on incidence from monthly cumulative precipitations with 1-month lag , monthly average relative humidity with 1-month lag , and monthly average temperature with 2-month lag , which were assessed during this phase within this area ( results not shown ) . Furthermore , our results indicating an association between HFRS incidence and meteorological factors with time lags indicate that these meteorological factors may also influence the populations and infection rates of Hantan viruses in rodents via the gestation periods and sexual maturation of rodents ( which is typically 2–3 months ) [33] . Despite insights gained from the present study , the limitations of our study should also be acknowledged . Firstly , due to a lack of time series data on the Hantaviruses of HFRS cases , population densities of rodents , and other influencing factors , it is difficult to further uncover the probable causes of the shifts of seasonal patterns of HFRS incidence from 1973 to 2005 . Secondly , although the accuracy of clinically diagnosed HFRS cases is high ( greater than 80% ) , atypical cases of HFRS and mistaken diagnoses of HFRS patients infected by Seoul virus existed [5] , the increase of incidence beginning from Phase II also could be affected by the change from purely clinical to laboratory based surveillance in that years , when the availability of laboratory testing could encouraged more clinicians to look for the disease , and possible biases in disease reporting could influence the analysis of our results . In addition , the data are from a passive surveillance system . As a result , cases might be underreported , which then might influence the precision of our analyses . Also , several issues remain unaddressed: how changes in land use , housing conditions , and lifestyles might influence the spread of HFRS in the province and how these factors impact transmission dynamics associated with each serotype ? Such knowledge is particularly relevant in the context of the continued spread of HFRS endemic areas throughout Shandong Province , as well as the variation in the seasonal distribution of HFRS incidence . Understanding these issues may further assist in informing prevention and control strategies .
|
Hemorrhagic fever with renal syndrome ( HFRS ) , a rodent-borne disease caused by Hantaviruses , is characterized by fever , acute renal dysfunction and hemorrhagic manifestations . At present , it is endemic in all 31 provinces , autonomous regions , and metropolitan areas in mainland China where human cases account for 90% of the total global cases . Historically Shandong Province bears the largest HFRS burden in China—the cumulative number of human cases accounted for 1/3 of the national total . Here we report a spatiotemporal analysis of human HFRS cases in Shandong using reported case data spanning 1973 to 2005 . Through the analysis of seasonal incidences and use of velocity maps , three phases of seasonal shifts of HFRS epidemics and the expansion pattern of HFRS endemic areas were identified over the 33-year period . In addition , precipitation , humidity , and temperature were found to be significantly associated with the seasonal variation of HFRS incidence in Shandong Province . These findings offer insights in understanding possible causes of HFRS spread and distribution and may assist in informing prevention and control strategies .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"public",
"health",
"and",
"epidemiology",
"public",
"health",
"and",
"epidemiology/epidemiology",
"public",
"health",
"and",
"epidemiology/preventive",
"medicine",
"public",
"health",
"and",
"epidemiology/infectious",
"diseases"
] |
2010
|
Spatiotemporal Trends and Climatic Factors of Hemorrhagic Fever with Renal Syndrome Epidemic in Shandong Province, China
|
The two DNA strands of the nuclear genome are replicated asymmetrically using three DNA polymerases , α , δ , and ε . Current evidence suggests that DNA polymerase ε ( Pol ε ) is the primary leading strand replicase , whereas Pols α and δ primarily perform lagging strand replication . The fact that these polymerases differ in fidelity and error specificity is interesting in light of the fact that the stability of the nuclear genome depends in part on the ability of mismatch repair ( MMR ) to correct different mismatches generated in different contexts during replication . Here we provide the first comparison , to our knowledge , of the efficiency of MMR of leading and lagging strand replication errors . We first use the strand-biased ribonucleotide incorporation propensity of a Pol ε mutator variant to confirm that Pol ε is the primary leading strand replicase in Saccharomyces cerevisiae . We then use polymerase-specific error signatures to show that MMR efficiency in vivo strongly depends on the polymerase , the mismatch composition , and the location of the mismatch . An extreme case of variation by location is a T-T mismatch that is refractory to MMR . This mismatch is flanked by an AT-rich triplet repeat sequence that , when interrupted , restores MMR to >95% efficiency . Thus this natural DNA sequence suppresses MMR , placing a nearby base pair at high risk of mutation due to leading strand replication infidelity . We find that , overall , MMR most efficiently corrects the most potentially deleterious errors ( indels ) and then the most common substitution mismatches . In combination with earlier studies , the results suggest that significant differences exist in the generation and repair of Pol α , δ , and ε replication errors , but in a generally complementary manner that results in high-fidelity replication of both DNA strands of the yeast nuclear genome .
Three processes operate to ensure faithful replication of the eukaryotic nuclear genome [1] , [2] . The first is the ability of DNA polymerases α , δ and ε to selectively insert correct rather than incorrect nucleotides onto correctly aligned rather than misaligned primer-templates . The second is proofreading , the 3′ exonucleolytic excision of errors from the primer terminus during replication . The third is mismatch repair ( MMR ) of errors that escape proofreading ( reviewed in [3]–[7] ) . MMR begins when a mismatch is recognized by homologues of the bacterial MutS homodimer , either Msh2-Msh6 ( MutSα ) or Msh2-Msh3 ( MutSβ ) . This recognition initiates a series of steps that ultimately remove the replication error from the nascent strand and allow new DNA to be synthesized accurately . The origin and nature of the strand discrimination signal used for MMR in vivo remains uncertain . MMR requires the presence of a discontinuity in the newly synthesized strand . At least in vitro , this discontinuity can be a nick or gap located either 3′ or 5′ to the mismatch , with the protein requirements for MMR differing somewhat depending on the location of the DNA ends relative to the mismatch . This provides an attractive possibility ( reviewed in [3] ) , namely that MMR may be directed to the nascent strand by the 3′ ends of growing chains at the replication fork and/or by the 5′ ends of Okazaki fragments that are transiently present during lagging strand replication . That the latter could provide a higher signal density for MMR of lagging strand replication errors was suggested in an earlier study of MMR of a damaged ( 8-oxo-G-A ) mismatch [8] . This leads to a previously unexplored question addressed by the present study , i . e . , is the efficiency of MMR similar or different for mismatches generated during leading and lagging strand replication ? Investigation of this question is complicated by the fact that DNA polymerases α , δ and ε ( Pols α , δ and ε , respectively ) are all required to efficiently replicate the nuclear genome [9] , and these polymerases have different error rates and error specificities [2] , [10] . Over the years , multiple models have been considered for the division of labor among these three polymerases during replication ( reviewed in [9]–[12] ) . Among these models , recent evidence [2] , [13] , [14] suggests that under normal circumstances , the leading strand template is primarily replicated by Pol ε , while the lagging strand template is replicated by Pol α-primase and Pol δ . Although MMR corrects errors made by all three polymerases [2] , [13] , [15]–[21] , it has only recently become possible to determine the extent to which MMR efficiency , and possibly MMR enzymology , varies depending on the replicase that made the error , the nascent strand containing the error and/or the location of the error within a DNA strand . We are investigating these variables using Saccharomyces cerevisiae strains containing mutant alleles of the POL1 ( Pol α ) , POL2 ( Pol ε ) and POL3 ( Pol δ ) genes . These mutant alleles , pol1-L868M [18] , [19] , pol2-M644G [13] and pol3-L612M ( [2] and references therein ) , encode enzymes with single animo acid replacements at the polymerase active site that reduce the fidelity of DNA synthesis . As a consequence , strains harboring these alleles have elevated spontaneous mutation rates , thereby allowing assignment of responsibility for most in vivo errors to a chosen mutator polymerase , rather than its wild type counterparts [2] , [13] . In strains containing these mutator polymerases , URA3 mutation rates and mutational spectra can be determined and used to calculate the rates for specific mutations , e . g . , single base substitutions and insertions/deletions ( indels ) in various sequence contexts . Comparison of these rates in MMR-proficient yeast strains to strains that lack MSH2-dependent MMR yields a calculation of the apparent MSH2-dependent MMR efficiency for a variety of replication errors generated during replication in vivo . Using this approach , we recently described the efficiency of repairing lagging strand replication errors generated by L868M Pol α and L612M Pol δ [21] . Here we extend the effort using yeast strains encoding M644G Pol ε , allowing the comparison of MMR correction efficiencies for replication errors made by each of the three eukaryotic replicative polymerases . The results indicate that on average , MMR balances the fidelity of leading and lagging strand DNA replication , but with exceptions that place some base pairs at high risk of mutation from replication infidelity even in cells with normal MMR .
Our previous inference that Pol ε is a leading strand replicase was based on patterns of rare mutations in one gene ( URA3 ) at one locus ( AGP1 ) [13] . Two recent studies have made it feasible to test Pol ε strand assignment using a different biomarker , ribonucleotide incorporation into nuclear DNA . The first study demonstrated that , in addition to reduced fidelity for single base mismatches , M644G Pol ε also has reduced sugar discrimination , i . e . , it incorporates rNTPs into DNA much more readily than does wild-type Pol ε [23] . In that study , rNMPs incorporated into nascent DNA during replication by M644G Pol ε were detected as alkali-sensitive sites in the nuclear genome of a pol2-M644G rnh201Δ strain , which lacks the ability to repair rNMPs in DNA due to deletion of the RNH201 gene encoding the catalytic subunit of RNase H2 . A more recent study exploited this fact to probe the genomic DNA of a homologous S . pombe polε-M630F rnh201Δ mutant strain by strand-specific Southern blotting [14] . When strand-specific probes flanking ARS3003/3004 were used , the results revealed that more rNMPs were incorporated into the nascent leading strand than into the nascent lagging strand . This led to the interpretation that , as in budding yeast , fission yeast Pol ε is also the primary leading strand replicase [14] . Using this same strategy , we examined the strand specificity of rNMP incorporation in S . cerevisiae pol2-M644G rnh201Δ strains with the URA3 reporter in one of two possible orientations , using alkali treatment and subsequent probing for either the nascent leading or lagging strand with strand-specific URA3 probes ( Figure 1A ) . One of the two strands from each pol2-M644G rnh201Δ strain was preferentially sensitive to alkaline hydrolysis ( Figure 1B ) . In each case , this corresponded to the nascent leading strand products of replication ( probe A in orientation 2 and probe B in orientation 1 ) . These results strongly support the idea that Pol ε preferentially replicates the leading strand template . Note that the distribution of ribonucleotides within the two strands across the whole genome remains to be determined and could differ . The strategy used here to study strand-specific MMR involves measuring spontaneous mutation rates in yeast strains with the URA3 reporter gene present in either of two orientations , both proximal to ARS306 , a well-characterized , early-firing replication origin [24] . In our initial study of the role of Pol ε in replication [13] , we compared mutation rates in MMR proficient ( MSH2+ ) strains with wild type Pol ε ( encoded by the POL2 gene ) to rates in strains with the pol2-M644G mutation . The pol2-M644G strains had elevated mutation rates [13] , an observation that is reproduced here ( Table 1 ) . The majority of 5-FOA resistant mutants had single-base mutations in the URA3 gene . In orientation 1 , these were predominantly A-T to T-A mutations at base pairs 279 and 686 . These mutations were rare in orientation 2 ( partial spectra in [13] , complete spectra in Figure S1A ) . This strong orientation bias , and the fact that the in vitro error rate for template T-dTMP mismatches by M644G Pol ε is much higher than the error rate for template A-dAMP mismatches , implies that Pol ε participates in leading strand DNA replication [13] . Two later studies [2] , [25] indicated that Pol δ primarily acts as a lagging strand polymerase and has a less substantial role in leading strand replication . This further implied that Pol ε not only participates in leading strand DNA replication , but that it is the major leading strand replicase . The pol2-M644G msh2Δ mutants have strongly elevated mutation rates relative to the MSH2+ strains ( Table 1 ) , indicating that the vast majority of the mutations are made by M644G Pol ε . In the absence of mismatch repair , most 5-FOA resistant mutants contained single base changes that were widely scattered throughout the URA3 coding sequence ( Figure S1B ) . As compared to MMR proficient pol2-M644G strains , base pairs 279 and 686 in pol2-M644G msh2Δ strains did not stand out as hotspots for A-T to T-A transversions in orientation 1 , even though base substitution and single base deletion hotspots were observed at several other locations ( Figure S1B ) . The data in Table 1 and Figure S1 were used to calculate rates for single base mutations in the MMR-proficient and msh2Δ strains ( Table S2 ) . The ratio of these rates reflects the apparent MMR correction efficiency for each type of error , and the results can be compared ( see discussion ) to those reported earlier [21] for replication errors made by L868M Pol α and L612M Pol δ . As noted previously [21] , [26]–[29] , certain correction factors could be higher if some mismatches in the MMR proficient strains are not subject to MMR , either because they are damaged or because they are generated during DNA transactions that occur outside of replication . Conclusions about the overall balance of repair between strands and polymerases derive from collective consideration of all single base mismatches . In the pol2-M644G strain background , the MMR correction factor for all single base mismatches is 250-fold ( Table 2; Figure 2A , blue bar; Table S2 ) , i . e . , on average , 249 of 250 single base replication errors generated by M644G Pol ε are corrected by MMR . This correction factor is higher than for L612M Pol δ ( Table 2; Figure 2A , green diamond ) , but lower than for L868M Pol α ( Table 2; Figure 2A , red diamond ) . As a consequence , the mutation rates for all three variant polymerase strains are similar when MMR is operative ( top line in Table 2 ) . Average correction factors are high for each of the four classes of single base changes generated by M644G Pol ε ( Figure 2A ) , in the following order: deletions ( 1 , 500-fold ) , insertions ( 1 , 100-fold ) , transitions ( 440-fold ) and transversions ( 72-fold ) . Correction factors vary widely between specific positions in the URA3 open reading frame . Figure 2B–2D show eight locations where it is possible to compare MMR of the same mismatch generated by M644G Pol ε during leading strand replication ( blue bars ) or by Pol α ( red diamonds ) and δ ( green diamonds ) during lagging strand replication ( expanded from [21] ) . In order to maintain equivalent template context , leading strand errors found in one URA3 orientation in the pol2-M644G strains are always compared to lagging strand errors found in the other URA3 orientation in the pol1-L868M and pol3-L612M strains . For example , the correction factors in Figure 2B for deleting an A-T pair from the three longest runs of A-T pairs in the URA3 coding sequence ( base pairs 174–178 , 201–205 and 255–260 , Figure S1 ) are each inferred to involve a single unpaired T . The comparative MMR correction factors and their implications are considered in the Discussion . In contrast to the efficient repair of most single mismatches , the rate of A-T to T-A transversions at base pair 686 in orientation 1 ( Figure S1 and Table S3 ) is no higher in the pol2-M644G msh2Δ strain than in the MMR-proficient pol2-M644G strain ( Table S2 ) . This indicates that T-T mismatches generated at base pair 686 during leading strand replication by M644G Pol ε are not efficiently corrected by MMR ( Figure 3 , “T-dT , 686 , ” dark blue bar ) . This contrasts with an average of 41-fold correction ( dark blue bar on left ) of the same mismatch inferred at all other A-T base pairs in URA3 , i . e . , A to T substitutions in orientation 1 and T to A transversions in orientation 2 ( Figure S1 ) . Adjacent to base pair 686 is a triplet repeat sequence , 5′-ATT ATT ATT gTT ( designated here as ATT3 ) . For several reasons ( see Discussion ) , we speculated that this sequence might suppress MMR at base pair 686 . To test this , we constructed strains in which ATT3 was modified to 5′-ATA ATC ATA gTT ( designated ATT0 , see Figure 3 ) , with the three ( underlined ) changes interrupting the repeat units without changing the amino acid sequence . We then measured spontaneous mutation rates and generated mutational spectra ( Figure S2 ) to determine if the flanking sequence changes allowed MMR of T-T mismatches at base pair 686 . The results ( Table S3 ) indicate that this is indeed the case . The MMR correction factor at base pair 686 increased to 35-fold ( Figure 3 , p≤0 . 001 ) , indicating that 97% of T-T mismatches are repaired when base pair 686 is flanked by ATT0 . Single base-base mismatches are repaired by MutSα ( Msh2-Msh6 ) but not by MutSβ ( Msh2-Msh3 ) [3]–[7] , implying that the ATT3 sequence is suppressing repair of the T-T mismatch that would normally occur via MutSα . However , given evidence that MutSβ can bind to a non-B-DNA structure that can form in a triplet repeat sequence and promote triplet repeat expansion ( reviewed [30] ) , we examined whether suppression of MMR at base pair 686 might depend on MutSβ . This was done by calculating the A-T to T-A mutation rate at base pair 686 in URA3 orientation 1 in a pol2-M644G msh3Δ rnh201Δ strain [31] . The calculated A-T to T-A rate is 17×10−8 , which is no lower than observed here in the Msh3+ strain ( 6 . 8×10−8 , Table S3 ) . Thus suppression of MMR by ATT3 is independent of MutSβ .
We previously inferred that Pol ε participates in leading strand replication using base substitutions as biomarkers for leading strand replication . These events are rare , occurring approximately once per 10 million incorporations . The present study uses ribonucleotides as an independent and much more abundant biomarker . The preferential presence of ribonucleotides in the nascent leading strand observed here in pol2-M644G rnh201Δ strains ( URA3 orientation 1 and orientation 2; Figure 1 ) strongly supports the inference that Pol ε primarily participates in leading strand replication . This does not preclude occasional Pol ε participation in lagging strand replication . The interpretation that Pol ε primarily participates in leading strand replication lends credibility to the interpretations presented below regarding the efficiency of repairing mismatches made by Pol ε during leading strand replication as compared to mismatches of similar composition made by Pols α and δ during lagging strand replication . An additional notable point here is that the sizes of the nascent leading strand fragments resulting from alkaline hydrolysis of DNA from the pol2-M644G rnh201Δ strains ( Figure 1 ) indicate that approximately one ribonucleotide may be incorporated for every 1 , 000 deoxyribonucleotides . This density of ribonucleotide incorporation into DNA is about four orders of magnitude higher than for A-T- to T-A transversions . Thus ribonucleotides mapped by deep sequencing techniques could serve as high density , genome-wide biomarkers of Pol ε action in vivo during replication and possibly during repair and recombination . The average MMR correction factors for errors made by M644G Pol ε are highest for indels , intermediate for transitions and lowest for transversions ( Figure 2A ) . This rank order is common to E . coli [28] , [29] , [32] and to errors made by yeast Pols α and δ [21] , suggesting that MMR has conserved the ability to most efficiently correct the most potentially deleterious errors ( indels ) , and also the base-base mismatches made at the highest rates by both bacterial and eukaryotic replicases . This general principal is qualified by the observation that MMR efficiency varies , even for the same inferred mismatch ( e . g . , either an extra T , a C-dT or a G-dT mismatch , Figure 2B , 2C and 2D , respectively ) made by the same polymerase ( M644G Pol ε ) during replication of the same ( leading ) strand . Most sequence-dependent variations in MMR efficiency seen here are in the 2- to 10-fold range ( Figure 2 ) depending on the comparison . That such variations are typically small is perhaps expected , since MMR is needed to preserve the stability of nuclear genomes despite their enormous sequence complexity . Variations due to mismatch composition and location are consistent with biochemical studies showing differences in MMR in vitro [33] and with mutational studies in vivo in which the identity of the replicase that made the mismatch was unknown . Several explanations for variations in eukaryotic MMR efficiency can be explored in the future . For example , the efficiency with which E . coli repairs transversion mismatches in phage λ increases with increasing G-C content in neighboring nucleotides [32] , and recognition of certain mismatches by MutSα is influenced by a 6-nucleotide region surrounding the mismatch [34] . Thus it may be that flanking sequences , such as those shown in Figure 2 , influence eukaryotic MMR efficiency in vivo by modulating ( i ) mismatch binding by MutSα , which contacts several base pairs on either side of the mismatch [35] , ( ii ) base pair stacking , since a MutSα-bound mismatched base stacks with a conserved phenylalanine in Msh6 , and/or ( iii ) DNA flexibility , since MutSα-bound mismatched DNA is kinked , and a transition between bent and unbent DNA may be critical for limiting MMR to processing of mismatched as compared to matched base pairs [36] . Variations in MMR efficiency might also depend on proteins that operate downstream of mismatch binding , such as MutLα or exonucleases , or they may reflect other variables , such as the timing of nucleosome reloading behind the replication fork , nucleosome dynamics and/or chromatin remodeling . A striking observation here is the apparent absence of MMR of the A-T to T-A transversion at base pair 686 ( Figure 3 ) , which is inferred to result from a T-T mismatch made by M644G Pol ε during leading strand replication . This lack of repair contrasts sharply with efficient repair at many other locations . For example , the deletion mismatch at base pairs 255–260 , which is predicted to involve a mismatch containing a single unpaired T in the template ( Figure 2B ) , has an approximately 6000-fold higher correction factor than for the T-T mismatch at base pair 686 . Lack of repair at base pair 686 is not due to a general inability to correct A-T to T-A transversion mismatches , because the average correction factor for these events elsewhere in URA3 is 41-fold ( Figure 3 ) . The absence of correction at position 686 led us to test whether MMR was inhibited by the adjacent 5′-ATTATTATTgTT sequence . There were several reasons to suspect that this could be the case . The sequence is A-T rich and may have unusual helical parameters that could diminish MMR . For example , sequences containing larger numbers of ATT repeats can form a non-hydrogen bonded structure [37] , and can be induced into hairpins by the DNA minor groove binding ligand DAPI ( 4′ , 6-diamidino-2-phenylindole ) [38] , [39] . Triplet repeat sequences can form non-B-DNA structures that bind MMR proteins ( reviewed in [30] ) , and they are often associated with genome instability ( reviewed in [40] ) , albeit characterized by indels rather than base substitutions . In addition , recent studies have demonstrated that nucleosomes influence the behavior of MMR proteins and visa versa ( e . g . , see [41]–[44] ) , and nucleosome binding to DNA is influenced by DNA sequence , with A-T-rich dinucleotides such as those present in ATT3 having an important role in nucleosome positioning ( e . g . , see [45] , [46] and references therein ) . For these reasons , we examined MMR at base pair 686 after changing the flanking sequence to eliminate the triplet repeats and decrease A-T content by one base pair . The results indicate that these changes allowed correction of 97% of the mismatches generated by M644G Pol ε at base pair 686 ( 88% correction at the lower 95% confidence limit , Figure 3 ) . This suggests that the ATT3 flanking sequence is a natural cis-acting suppressor of the normal MSH2-dependent MMR machinery . Suppression does not decrease upon deletion of MSH3 , and thus is MutSβ independent , unlike triplet repeat expansion [30] . Collectively , position 686 and ATT3 are an example of what has been called an “At Risk sequence Motif” [47] , i . e . , a naturally occurring DNA sequence that results in inefficient operation of a DNA transaction required for genome stability . The fact that one such sequence exists in the 804 base pair open reading frame of URA3 leads one to wonder how many natural suppressors of MMR might be present in nuclear genomes . This issue is currently being investigated using the deep sequencing approach previously used to infer that Pol δ is a lagging strand replicase across the yeast genome [25] . Experiments are also planned to examine which ( if any ) of the possibilities mentioned in the preceding section may be relevant to inefficient MMR at base pair 686 . We previously suggested that MMR may be directed to the nascent strand by the 3′ ends of growing chains at the replication fork [22] , and/or by the 5′ ends of Okazaki fragments that are transiently present during lagging strand replication [8] . The 5′ ends of Okazaki fragments , and perhaps the PCNA required to process these ends , could potentially provide a higher signal density for MMR of lagging strand replication errors as compared to errors generated during leading strand replication , which is thought to be more continuous . If so , then MMR might be more efficient in correcting lagging strand errors . In an initial test of this hypothesis , we found that mutagenesis due to a mismatch formed at one particular G-C base pair during replication of unrepaired 8-oxo-G in ogg1-deficient yeast was lower for lagging as compared to leading strand replication , and importantly , that this bias was largely eliminated in MMR defective strains [8] . Among several possible explanations that we considered for loss of the strand bias , one was that 8-oxo-G-dA mismatches made during lagging strand replication may be more efficiently corrected than are 8-oxo-G-dA mismatches made during leading strand replication . A major goal of the present study was to test this hypothesis for multiple , natural ( i . e . , undamaged ) mismatches generated at different locations during replication of a larger target sequence . The present study accomplishes this , and allows the first direct comparison of MMR efficiency for errors made by all three replicases , to our knowledge , thereby providing insights into the contribution of MMR to leading and lagging strand replication fidelity . From the results in Figure 2 , we conclude that in general , mismatches made by all three replicases are repaired very efficiently . This is logical given the need to preserve genetic information in both DNA strands . This conclusion is independent of various models regarding which DNA polymerase replicates which strand ( reviewed in [11] , [12] ) . Other implications derive from the model wherein Pols α and δ are the primary lagging strand replicases and Pol ε is the primary leading strand replicase . In our earlier report [21] , we pointed out that correction factors were higher for mismatches made by Pol α than for the same mismatches made by Pol δ , suggesting that the 5′ ends of Okazaki fragments may be strand discrimination signals and that MMR efficiency may be related to the proximity of a mismatch to that signal . This is interesting given that DNA polymerase ε is highly processive , at least as processive as DNA polymerase δ , and that leading strand replication is thought to be largely continuous [48] , [49] , [50] . It is of course conceivable that leading strand replication may not be as continuous as current models imply . If leading strand replication is indeed largely continuous , then the fact that MMR corrects most Pol ε errors about as efficiently as it corrects errors made by Pols α and δ ( Figure 2 ) implies the existence of MMR signals other than the 5′ ends of Okazaki fragments , and these can very efficiently direct MMR to the nascent leading strand . Possible signals for leading strand replication include the above-mentioned 3′ ends of growing chains at the replication fork [22] , [51] , [52] , nicks introduced into the nascent leading strand by nucleases , and/or asymmetrically bound PCNA [8] , [53] . PCNA is a particularly attractive possibility for differentially modulating the efficiency of MMR of errors made by the three replicases , because it is involved in early steps in MMR ( see [3]–[7] for review] ) , it does not influence DNA synthesis by Pol α , and it does stimulate DNA synthesis by both Pol δ and Pol ε , albeit through different PCNA-polymerase interactions ( see [9] and references therein ) . The results in Figure 2 further suggest that , even for the same mismatch ( extra T , G-dT or C-dT ) in a common sequence context , MMR efficiency varies depending on which polymerase made the error . In two of three instances involving deletion of a single template T ( Figure 2B ) , the repair of mismatches made by Pol δ is higher than for mismatches made by Pol ε . This correlates with the observation that Pol δ generates this mismatch in vitro at a higher rate than does Pol ε [54] . Similarly , transitions and transversions ( Figure 2A ) and several site-specific base substitutions ( Figure 2C and 2D ) generated by Pol α are corrected more efficiently than are mismatches generated by Pol δ and Pol ε . Pol α lacks an intrinsic proofreading exonuclease activity and is less accurate than proofreading-proficient Pols δ and ε ( reviewed in [10] , [55] ) . Thus the present study of mismatches generated by Pol ε extends the idea that MMR has evolved to most efficiently correct the most deleterious mismatches ( i . e . , indel mismatches ) . Within classes of similar deleterious potential ( base-base mismatches ) , evolution has produced the highest efficiency versus the most frequently generated mismatches . In a model wherein Pol ε is the major leading strand replicase and Pols α and δ conduct about 10% and 90% of lagging strand replication [2] , respectively , the results ( Table 2; Figure 2A , average repair for single base errors ) further suggest that MMR balances the fidelity of replication of the two strands despite the use of replicases with substantially different fidelity and error specificity .
The strains used in this study , the measurements of spontaneous mutation rates and the sequencing of URA3 mutants were as previously described [2] , [13] , [21] , save that MSH2 was deleted from haploid pol2-M644G strains rather than diploid . The ATT3 to ATT0 conversion was made via site-directed mutagenesis and integration pop-out [56] in a strain with wild type polymerases . PCR product containing the ATT0 URA3 allele was then transformed into msh2Δ backgrounds and proper insertion verified via sequencing . Genomic DNA was isolated from exponentially growing cultures ( grown in YPDA at 30°C ) using the Epicentre Yeast DNA purification kit . Five µg of DNA was treated with 0 . 3 M KOH for 2 h at 55°C and subjected to alkaline-agarose electrophoresis as described [23] . Following neutralization , DNA was transferred to a charged nylon membrane ( Hybond N+ ) by capillary action and probed by Southern analysis . Strand-specific radiolabeled probes were prepared from a PCR-amplified fragment of URA3 template , using a previously described procedure and probe design [14] . See Text S1 .
|
The stability of complex and highly organized nuclear genomes partly depends on the ability of mismatch repair ( MMR ) to correct a variety of different mismatches generated as the leading and lagging strand templates are copied by three polymerases , each with different fidelity . Here we provide the first comparison , to our knowledge , of the efficiency of MMR of leading and lagging strand replication errors . We first confirm that Pol ε is the primary leading strand replicase , complementing earlier assignment of Pols α and δ as the primary lagging strand replicases . We then show that MMR efficiency in vivo strongly depends on the polymerase that generates the mismatch and on the composition and location of mismatches . In one extreme case , a flanking triplet repeat sequence eliminates MMR altogether . Overall , MMR is most efficient for mismatches generated at the highest rates and having the most deleterious potential , thereby ultimately achieving high-fidelity replication of both DNA strands .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetic",
"mutation",
"microbiology",
"nucleotides",
"model",
"organisms",
"mutation",
"types",
"dna",
"replication",
"dna",
"dna",
"synthesis",
"biology",
"molecular",
"biology",
"mutagenesis",
"biochemistry",
"nucleic",
"acids",
"genetics",
"yeast",
"and",
"fungal",
"models",
"dna",
"repair",
"saccharomyces",
"cerevisiae",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] |
2012
|
Mismatch Repair Balances Leading and Lagging Strand DNA Replication Fidelity
|
The precision of multisensory perception improves when cues arising from the same cause are integrated , such as visual and vestibular heading cues for an observer moving through a stationary environment . In order to determine how the cues should be processed , the brain must infer the causal relationship underlying the multisensory cues . In heading perception , however , it is unclear whether observers follow the Bayesian strategy , a simpler non-Bayesian heuristic , or even perform causal inference at all . We developed an efficient and robust computational framework to perform Bayesian model comparison of causal inference strategies , which incorporates a number of alternative assumptions about the observers . With this framework , we investigated whether human observers’ performance in an explicit cause attribution and an implicit heading discrimination task can be modeled as a causal inference process . In the explicit causal inference task , all subjects accounted for cue disparity when reporting judgments of common cause , although not necessarily all in a Bayesian fashion . By contrast , but in agreement with previous findings , data from the heading discrimination task only could not rule out that several of the same observers were adopting a forced-fusion strategy , whereby cues are integrated regardless of disparity . Only when we combined evidence from both tasks we were able to rule out forced-fusion in the heading discrimination task . Crucially , findings were robust across a number of variants of models and analyses . Our results demonstrate that our proposed computational framework allows researchers to ask complex questions within a rigorous Bayesian framework that accounts for parameter and model uncertainty .
Visuo-vestibular integration in heading perception presents an ideal case to characterize the details of the causal inference strategy in multisensory perception . While a wealth of published studies have shown that integration of visual and vestibular self-motion cues increases perceptual precision [9–12 , 14 , 41–43] , and accuracy [14] , such an integration only makes sense if the two cues arise from the same cause—that is optic flow and inertial motion signal heading in the same direction . Despite the putative relevance of causal inference in heading perception , the inference strategies that characterize visuo-vestibular integration in the presence of sensory conflict remain poorly understood . For example , a recent study has found that observers predominantly integrated visual and vestibular cues even in the presence of large spatial discrepancies [33]—whereas a subsequent work has presented evidence in favor of causal inference [34] . Furthermore , these studies did not vary cue reliability—a manipulation that is critical to test whether a Bayes-optimal inference strategy or a suboptimal approximation was used [35] . Another aspect that can influence the choice of inference strategy is the type of inference performed by the observer . In particular , de Winkel and colleagues [33 , 34] asked subjects to indicate the perceived direction of inertial heading—an ‘implicit’ causal inference task as subjects implicitly assessed the causal relationship between visual and vestibular cues on their way to indicate the final ( integrated or segregated ) heading percept . Even in the presence of spatial disparities as high as 90° , one study found that several subjects were best described by a model which fully integrated visual and vestibular cues [33] ( possibly influenced by the experimental design; see also [34] ) . It is plausible that performing an explicit causal inference task , which forces subjects to indicate whether visual and vestibular cues arose from the same or different events , may elicit different inference strategies , as previously reported in category-based induction [44] , multi-cue judgment [45] , and sensorimotor decision-making [46] . While some studies have tested both explicit and implicit causal inference [18 , 21 , 47] , to our knowledge only one previous study contemplated the possibility of different strategies between implicit and explicit causal inference tasks [21] , and a systematic comparison of inference strategies in the two tasks has never been carried out within a larger computational framework . Thus , the goal of this work is two-fold . First , we introduce a set of techniques to perform robust , efficient Bayesian factorial model comparison of a variety of Bayesian and non-Bayesian models of causal inference in multisensory perception . Factorial comparison is a way to simultaneously test different orthogonal hypotheses about the observers [21 , 48–50] . Our approach is fully Bayesian in that we consider both parameter and model uncertainty , improving over previous analyses which used point estimates for the parameters and compared individual models . A full account of uncertainty in both parameter and model space , by marginalizing over parameters and model components , is particularly prudent when dealing with internal processes , such as decision strategies , which may have different latent explanations . An analysis that disregards such uncertainty might produce unwarranted conclusions about the internal processes that generated the observed behavior [37] . Second , we demonstrate our methods by quantitatively comparing the decision strategies underlying explicit and implicit causal inference in visuo-vestibular heading perception within the framework of Bayesian model comparison . We found that even though the study of explicit and implicit causal inference in isolation might suggest different inference rules , a joint analysis that combines all available evidence points to no difference between tasks , with subjects performing some form of causal inference in both the explicit and implicit tasks that used identical experimental setups . In sum , we demonstrate how state-of-the-art techniques for model building , fitting , and comparison , combined with advanced analysis tools , allow us to ask nuanced questions about the observer’s decision strategies in causal inference . Importantly , these methods come with a number of diagnostics , sanity checks and a rigorous quantification of uncertainty that allow the experimenter to be explicit about the weight of evidence .
We compiled a diverse set of computational techniques to perform robust Bayesian comparison of models of causal inference in multisensory perception , which we dub the ‘Bayesian cookbook for causal inference in multisensory perception’ , or herein simply ‘the cookbook’ . The main goal of the cookbook is to characterize observers’ decision strategies underlying causal inference , and possibly other details thereof , within a rigorous Bayesian framework that accounts for both parameter uncertainty and model uncertainty . The cookbook is ‘doubly-Bayesian’ in that it affords a fully Bayesian analysis of observers who may or may not be performing Bayesian inference themselves [51] . Fully Bayesian model comparison is computationally intensive , hence the cookbook is concerned with efficient algorithmic solutions . The cookbook comprises of: ( a ) a fairly general recipe for building observer models for causal inference in multisensory perception ( see Methods and Section 1 of S1 Appendix ) , which lends itself to a factorial model comparison; ( b ) techniques for fast evaluation of a large number of causal inference observer models; ( c ) procedures for model fitting via maximum likelihood , and approximating the Bayesian posterior of the parameters via Markov Chain Monte Carlo ( MCMC ) ; ( d ) state-of-the-art methods to compute model comparison metrics and perform factorial model selection . It is noteworthy that , while the current work focuses on the example of visuo-vestibular heading perception , this cookbook is general and can be applied with minor modifications to multisensory perception across sensory domains . Computational details are described in the Methods section and S1 Appendix . Here we present an application of our framework to causal inference in multisensory heading perception . For ease of reference , we summarize relevant abbreviations used in the paper and their meaning in Table 1 . We demonstrate our framework taking as a case study the comparison of explicit vs . implicit causal inference strategies in heading perception . In this section we briefly summarize our methods . Extended details and description of the cookbook can be found in the Methods and S1 Appendix . We examined how subjects perceived the causal relationship of synchronous visual and vestibular headings as a function of disparity ( svest − svis , nine levels ) and visual reliability level ( high , medium , low; Fig 3A ) . Common cause reports were more frequent near zero disparities than for well-separated stimuli ( Repeated-measures ANOVA with Greenhouse-Geisser correction; F ( 1 . 82 , 18 . 17 ) = 76 . 0 , ϵ = 0 . 23 , p < 10−4 , η p 2 = 0 . 88 ) . This means that observers neither performed complete integration ( always reporting a common cause ) nor complete segregation ( never reporting a common cause ) . Common-cause reports were not affected by visual cue reliability alone ( F ( 1 . 23 , 12 . 33 ) = 1 . 84 , ϵ = 0 . 62 , p = . 2 , η p 2 = 0 . 16 ) , but were modulated by an interaction of visual reliability and disparity ( F ( 7 . 44 , 74 . 44 ) = 7 . 38 , ϵ = 0 . 47 , p < 10−4 , η p 2 = 0 . 42 ) . Thus , observers’ performance was affected by both cue disparity as well as visual cue reliability when explicitly reporting about the causal relationship between visual and vestibular cues . However , this does not necessarily mean that the subjects’ causal inference strategy took visual cue reliability into account . Changes in sensory noise may affect measured behavior even if the observer’s decision rule ignores such changes [35]; a quantitative model comparison is needed to probe this question . We compared a subset of models from the full factorial comparison ( Fig 2A ) , since some models are equivalent when restricted to the explicit causal inference task . In particular , here fixed-criterion models are not influenced by the ‘prior’ factor , and the ( stochastic ) fusion model is not affected by sensory noise or prior , thus reducing the list of models to seven: Bay-C-E , Bay-C-I , Bay-X-E , Bay-X-I , Fix-C , Fix-X , SFu . To assess the evidence for distinct determinants of subjects’ behavior , we combined LOO scores from individual subjects and models with a hierarchical Bayesian approach [54] ( Fig 3B ) . Since we are investigating model factors that comprise of an unequal number of models , we reweighted the prior over models such that distinct components within each model factor had equal prior probability ( Fix models had 2× weight , and SFu 4× ) . In Fig 3B we report the protected exceedance probabilities φ ˜ and , for reference , the posterior model frequencies they are based on , and the Bayesian omnibus risk ( BOR ) , which is the estimated probability that the observed differences in factor frequencies may be due to chance [55] . We found that the most likely factor of causal inference was the Bayesian model ( φ ˜ = 0 . 78 ) , followed by fixed-criterion ( φ ˜ = 0 . 18 ) and probabilistic fusion ( φ ˜ = 0 . 04 ) . That is , fusion was ∼ 24 times less likely to be the most representative model than any form of causal inference combined , which is strong evidence against fusion , and in agreement with our model-free analysis . The Bayesian strategy was ∼ 3 . 5 times more likely than the others , which is positive but not strong evidence [57] . Conversely , the explicit causal inference data do not allow us to draw conclusions about noise models ( constant vs . eccentric ) or priors ( empirical vs . independent ) , as we found that all factor components are about equally likely ( φ ˜ ∼ 0 . 5 ) . At the level of specific models—as opposed to aggregate model factors – , we found that the probability of being the most likely model was almost equally divided between fixed-criterion ( C-I ) and Bayesian ( either X-E or C-I ) . All these models yielded reasonable fits ( Fig 3C ) , which captured a large fraction of the noise in the data ( absolute goodness of fit ≈ 76% ± 3%; see Methods ) ; a large improvement over a constant-probability model , which had a goodness of fit of 14 ± 5% . For comparison , we also show in Fig 3C the stochastic fusion model , which had a goodness of fit of 17 %± 5% . Visually , the Fix model in Fig 3C seems to fit better the group data , but we found that this is an artifact of projecting the data on the disparity axis . Disparity is the only relevant dimension for the Fix model; whereas Bay models fits the data along all dimensions . The visual superiority of the Fix model wanes when the data are visualized in their entirety ( see S1 Fig ) . We verified robustness of our findings by performing the same hierarchical analysis with different model comparison metrics . All metrics were in agreement with respect to the Bayesian causal inference strategy as the most likely , and the same three models being most probable ( although possibly with different ranking ) . BIC and marginal likelihood differed from LOO and AICc mainly in that they reported a larger probability for the constant vs . eccentricity-dependent noise ( probability ratio ∼4 . 6 , which is positive but not strong evidence ) . These results combined provide strong evidence that subjects in the explicit causal inference task took into account some elements of the statistical structure of the trial ( disparity , and possibly cue reliability ) to report unity judgments , consistent with causal inference , potentially in a Bayesian manner . From these data , it is unclear whether observers took into account the empirical distribution of stimuli , and whether their behavior was affected by eccentricity-dependence in the sensory noise . We examined the bias in the reported direction of inertial heading computed as ( minus ) the point of subjective equality for left/rightward heading choices ( L/R PSE ) , for each visual heading and visual cue reliability ( Fig 4A ) . Specifically , for a given value of visual heading svis ( or small range thereof ) , we constructed a psychometric function as a function of svest ( see Methods for details ) . If subjects were influenced by svis and took visual heading into account while computing inertial heading , this would manifest as bias in the psychometric function ( that is , a shifted point of subjective equality ) . If subjects were able instead to discount the distracting influence of svis , there should be negligible bias . As per causal inference , we qualitatively expected that there would be bias for smaller |svis| , but the bias would either decrease or saturate as |svis| increases . However , note that a nonlinear pattern of bias may also emerge due to eccentricity-dependence of the noise , even in the absence of causal inference . The bias was significantly affected by visual heading ( Repeated-measures ANOVA; F ( 0 . 71 , 7 . 08 ) = 19 . 67 , ϵ = 0 . 07 , p = . 004 , η p 2 = 0 . 66 ) . We found no main effect of visual cue reliability alone ( F ( 0 . 85 , 8 . 54 ) = 0 . 51 , ϵ = 0 . 43 , p = . 47 , η p 2 = 0 . 05 ) , but there was a significant interaction of visual cue reliability and heading ( F ( 2 . 93 , 29 . 26 ) = 7 . 36 , ϵ = 0 . 15 , p < 10−3 , η p 2 = 0 . 42 ) . These data suggest that subjects’ perception of vestibular headings was modulated by visual cue reliability and visual stimulus , in agreement with previous work in visual-auditory localization [21] . However , quantitative model comparison is required to understand the mechanism in detail since distinct processes , such as different causal inference strategies and noise models , could lead to similar patterns of observed behavior . We performed a factorial comparison with all models in Fig 2A . In this case , factorial model comparison via LOO was unable to uniquely identify the causal inference strategy adopted by observers ( Fig 4B ) . Forced fusion was slightly favored ( φ ˜ ∼ 0 . 48 ) , followed by Bayes ( φ ˜ ∼ 0 . 27 ) and fixed-criterion ( φ ˜ ∼ 0 . 25 ) , suggesting that all strategies were similar to forced fusion . Conversely , eccentricity-dependent noise was found to be more likely than constant noise ( ratio ∼ 5 . 7 ) , which is positive but not strong evidence , and empirical priors were marginally more likely than independent priors ( ∼ 2 . 1 ) . The estimated Bayesian omnibus risk was high ( BOR ≥ 0 . 29 ) , hinting at a large degree of similarity within all model factors such that observed differences could have arisen by chance . All metrics generally agreed on the lack of evidence in favor of any specific inference strategy ( with AICc and BIC tending to marginally favor fixed-criterion instead of fusion ) , and on empirical priors being more likely . As a notable difference , marginal likelihood and BIC reversed the result about noise models , favoring constant noise models over eccentricity-dependent ones . In terms of individual models , the most likely models according to LOO were , in order , forced fusion ( X-E ) , Bayesian ( X-E ) , and fixed-criterion ( C-E ) . However , other metrics also favored other models; for example , Bayesian ( C-E ) was most likely according to the marginal likelihood . All these models obtained similarly good fits to individual data ( Fig 4C; absolute goodness of fit ≈ 97% ) . For reference , a model that responds ‘rightward motion’ with constant probability performed about at chance ( goodness of fit ≈ 0 . 3 ± 0 . 1% ) . In sum , our analysis shows that the implicit causal inference data alone are largely inconclusive , possibly because almost all models behave similarly to forced fusion . To further explore our results , we examined the posterior distribution of the prior probability of common cause parameter pc across Bayesian models , and of the criterion κc for fixed-criterion models ( Fig 5 , bottom left panels ) . In both cases we found a broad distribution of parameters , with only a mild accumulation towards ‘forced fusion’ values ( pc = 1 or κ c ≳ 90 ° ) , suggesting that subjects were not completely performing forced fusion . Thus , it is possible that by constraining the inference with additional data we would be able to draw more defined conclusions . Data from the explicit and implicit causal inference tasks , when analyzed separately , afforded only weak conclusions about subjects’ behavior . The natural next step is to combine datasets from the two tasks along with the data from the unisensory heading discrimination task in order to better constrain the model fits . Before performing such joint fit , we verified whether there was evidence that model parameters changed substantially across tasks , in which case we might have had to change the structure of the models ( e . g . , by introducing a subset of distinct parameters for different tasks [49] ) . For each model parameter , we computed the across-tasks compatibility probability Cp ( Fig 5 ) , which is the ( posterior ) probability that subjects were most likely to have the same parameter values across tasks , as opposed to different parameters , above and beyond chance ( see Methods for details ) . We found at most mild evidence towards difference of parameters across the three tasks , but no strong evidence ( all Cp > . 05 ) . Therefore , we proceeded in jointly fitting the data with the default assumption that parameters were shared across tasks . For the joint fits there are nine possible models for the causal inference strategy ( three explicit causal inference × three implicit causal inference strategies ) . However , we considered only a subset of plausible combinations , to avoid ‘model overfitting’ ( see Discussion ) . First , we disregarded the stochastic fusion strategy for the explicit task , since this strategy was strongly rejected by the explicit task data alone . Second , if subjects performed some form of causal inference ( Bayesian or fixed-criterion ) in both tasks , we forced it to be the same . This reduces the model space for the causal inference strategy to four components: Bay/Bay , Fix/Fix , Bay/FFu , Fix/FFu ( explicit/implicit task ) . Combined with the prior and sensory noise factors as per Fig 2A , this leads to sixteen models . Factorial model comparison via LOO found that the most likely causal inference strategy was fixed-criterion ( φ ˜ = 0 . 79 ) , followed by Bayesian ( φ ˜ = 0 . 13 ) , and then by forced fusion in the implicit task ( φ ˜ = 0 . 05 paired with Bayesian explicit causal inference , φ ˜ = 0 . 03 paired with fixed-criterion explicit causal inference; Fig 6A ) . This is positive evidence that subjects were performing some form of causal inference also in the implicit task , as opposed to mere forced fusion ( ratio ∼ 11 . 4 ) . Moreover , we found strong evidence for eccentricity-dependent over constant noise ( φ ˜ > 0 . 99 , ratio ∼ 132 . 7 ) . Instead , the joint data were still inconclusive about the prior adopted by the subjects , with only marginal evidence for the empirical prior over the independent prior ( ∼ 2 . 9 ) . In terms of specific models , the most likely model was fixed-criterion ( X-E ) , followed by Bayesian ( X-E ) , and explicit Bayesian / implicit forced fusion ( both X-I and X-E ) . The best models gave a good description of the individual joint data , with an absolute goodness of fit of ≈ 91% ± 1% ( Fig 6B ) . Examination of the subjects’ posteriors over parameters for the joint fits ( Table 2 and Fig 5 , black lines ) showed reasonable results . The base visual noise parameters were generally monotonically increasing with decreasing visual cue reliability; the vestibular base noise was roughly of the same magnitude as the medium visual cue noise ( as per experiment design ) ; both visual and vestibular noise increased mildly with the distance from straight ahead; subjects had a small lapse probability . For Bayesian models , pc was substantially larger than the true value , 0 . 20 ( t-test t ( 10 ) = 10 . 8 , p < 10−4 , d = 3 . 25 ) , suggesting that observers generally thought that heading directions had a higher a priori chance to be the same . Nonetheless , for all but one subject pc was far from 1 , suggesting that subjects were not performing forced fusion either . An analogous result holds for the fixed criterion κc , which was smaller than the largest disparity between heading directions . We found that prior parameters σprior and Δprior had a lesser impact on the models , and their exact values were less crucial , with generally wide posteriors . Finally , we verified that our results did not depend on the chosen comparison metric . Remarkably , the findings regarding causal inference factors were quantitatively the same for all metrics , demonstrating robustness of our main result . Marginal likelihood and BIC differed from LOO and AICc in that they only marginally favored eccentricity-dependent noise models , showing that conclusions over the noise model may depend on the specific choice of metric . All metrics agreed in marginally preferring the empirical prior over the independent prior . In conclusion , when combining evidence from all available data , our model comparison shows that subjects were most likely performing some form of causal inference instead of forced fusion , for both the explicit and the implicit causal inference tasks . In particular , we find that a fixed-criterion , non-probabilistic decision rule ( i . e . , one that does not take uncertainty into account ) describes the joint data better than the Bayesian strategy , although with some caveats ( see Discussion ) . Performing a factorial comparison , like any other statistical analysis , requires a number of somewhat arbitrary choices , loosely motivated by previous studies , theoretical considerations , or a preliminary investigation of the data ( being aware of the ‘garden of forking paths’ [58] ) . As good practice , we want to check that our main findings are robust to changes in the setup of the analysis , or be able to report discrepancies . We take as our main result the protected exceedance probabilties φ ˜ of the model factors in the joint analysis ( Fig 6A , reproduced in Fig 7 , top row ) . In the following , we examine whether this finding holds up to several manipulations of the analysis framework . A first check consists of testing different model comparison metrics . In the previous sections , we have reported results for different metrics , finding in general only minor differences from our results obtained with LOO . As an example , we show here the model comparison using as metric an estimate of the marginal likelihood—the probability of the data under the model ( Fig 7 , 2nd row ) . We see that the marginal likelihood results agree with our results with LOO except for the sensory noise factor ( see Discussion ) . Therefore , our conclusions about the causal inference strategy are not affected . Second , the hierarchical Bayesian Model Selection method requires to specify a prior over frequencies of models in the population [54] . This ( hyper ) prior is specified via the concentration parameter vector α0 of a Dirichlet distribution over model frequencies . For our analysis , since we focused on the factorial aspect , we chose an approximately ‘flat’ prior across model factors ( see Methods for details ) , instead of the default flat prior over individual models ( α0 = 1 ) . We found that performing the group analysis with α0 = 1 did not change our results ( Fig 7 , 3rd row ) . Another potential source of variation is specific model choices , or inclusion of model factors . For example , a common successful variant of the Bayesian causal inference strategy is ‘probability matching’ , according to which the observer chooses the causal scenario ( C = 1 or C = 2 ) randomly , proportionally to its posterior probability [24] . As a first check , we performed the model comparison again using a ‘probability matching’ Bayesian observer instead of our main ‘model averaging’ observer ( Fig 7 , 4th row ) . Results are similar to the main analysis . If anything , the fixed-criterion causal inference strategy gains additional evidence here , suggesting that probability matching is a worse description of the data than our original Bayesian causal inference model ( as confirmed by looking at differences in LOO scores of individual subjects , e . g . for the Bay-X-E model; mean ± SEM: ΔLOO = −17 . 3 ± 5 . 7 ) . A recent study in audio-visual causal inference perception has similarly found that probability matching provided a poor explanation of the data [21] . In the factorial framework we could also have performed the previous analysis in a different way , by considering ‘probability matching’ as a sub-factor of the Bayesian strategy , together with ‘model averaging’ . As we have done before for the explicit causal inference task , we reassign prior probabilities to the models so that they are constant for each factor ( in this case , the two Bayesian strategies get a × 1 2 multiplier ) . Results of this alternative approach show an increase of evidence for the Bayesian causal inference family ( Fig 7 , bottom row ) . The values of φ ˜ for the fusion models are also slightly higher , which is due to an increase of the Bayesian omnibus risk ( the probability that the observed differences in factor frequencies are due to chance , a warning sign that there are too many models for the available data ) . This result and other lines of reasoning suggest caution when model factors contain an uneven number of models ( see Discussion ) . Nonetheless , the main conclusion does not qualitatively change , in that observers performed some form of causal inference as opposed to forced fusion . Finally , we performed several sanity checks , including a model recovery analysis to ensure the integrity of our analysis pipeline and that models of interest were meaningfully distinguishable ( see Methods and S1 Appendix for details ) . In conclusion , we have shown how the computational framework of Bayesian factorial model comparison , which is made possible by a combination of methods described in the cookbook , allows to explore multiple questions about aspects of subjects’ behavior in multisensory perception , and to account for uncertainty at different levels of the analysis in a principled , robust manner .
Our findings in the explicit causal inference task demonstrate that subjects used information about the discrepancy between the visual and vestibular cues to infer the causal relationship between them . Results in the implicit causal inference task alone were mixed , in that we could not clearly distinguish between alternative strategies , including forced fusion—in agreement with a previous finding [33] . However , when we combined evidence from all tasks , we found that some form of causal inference was more likely than mere forced fusion , in agreement with a more recent study [34] . Our findings suggest that multiple sources of evidence ( e . g . , different tasks ) can help disambiguate causal inference strategies which might otherwise produce similar patterns of behavioral responses . Our Bayesian analysis allowed us to examine the distribution of model parameters , in particular the causal inference parameters pc and κc , which govern the tendency to bind or separate cues for , respectively , a Bayesian and a heuristic fixed-criterion strategy . Evidence from all tasks strongly constrained these parameters for each subject . Interestingly , for the Bayesian models we found an average pc much higher than the true experimental value ( inferred pc ∼ 0 . 5 vs . experimental pc = 0 . 2 ) . This suggests that subjects had a tendency to integrate sensory cues substantially more than what the statistics of the task would require . Note that , instead , a Bayesian observer would be able to learn the correct value of pc from noisy observations , provided some knowledge of the structure of the task . Our finding is in agreement with previous studies which demonstrated an increased tendency to combine discrepant visual and vestibular cues [10 , 33 , 43 , 59 , 60] and also a large inter-subject variability in pc , and not obviously related to the statistics of the task [23] . We note that , in all studies so far , the ‘binding tendency’ ( pc or κc ) is a descriptive parameter of causal inference models that lacks an independent empirical correlate ( as opposed to , for example , noise parameters , which can be independently measured ) . Understanding the origin of the binding tendency , and which experimental manipulations it is sensitive to , is venue for future work [23 , 61] . For example , de Winkel and colleagues found that the binding tendency depends on the duration of the motion stimuli; decreasing for motions of longer duration [34] . Previous work has performed a factorial comparison of only causal inference strategies [21] . Our analysis extends that work by including as latent factors the shape of sensory noise ( and , thus , likelihoods ) and type of priors [48 , 49] . Models in our set include a full computation of the observers’ posterior beliefs based on eccentricity-dependent likelihoods , which was only approximated in previous studies that considered eccentricity-dependence [22 , 33 , 34] . Indeed , in agreement with a recent finding , we found an important role of eccentricity-dependent noise [22] . Conversely , our analysis of priors was inconclusive , as our datasets were unable to tell whether people learnt the empirical ( correlated ) prior , or made an assumption of independence . Our main finding , relative to the causal inference strategy , is that subjects performed causal inference both in the explicit and implicit tasks . Interestingly , from our analyses the most likely causal inference strategy is a fixed-criterion strategy , which crucially differs from the Bayesian strategy in that it does not take cue reliability into account—let alone optimally . This finding is seemingly at odds with a long list of results in multisensory perception , in which people are shown to take cue uncertainty into account [9 , 10 , 42 , 62] . We note that this is not necessarily in contrast with existing literature , for several reasons . First , this result pertains specifically to the causal inference part of the observer model , and not how cues are combined once a common cause has been inferred [21] . To our knowledge , no study of multisensory perception has tested Bayesian models of causal inference against heuristic models that take into account disparity but not reliability , as it has been done for example in visual search [56 , 63] and visual categorization [36 , 64] . A quantitative modeling approach is needed—qualitatively analyzing the differences in behavior at different levels of reliability is not sufficient to establish that observers take uncertainty into account; patterns of observed differences may be due to a change in sensory noise even if the observer’s decision rule disregards cue reliability . Second , our results are not definitive—the evidence for fixed-criterion vs . Bayesian is positive but not decisive . Our interpretation of this result is that subjects are following some suboptimal decision rule which happens to be closer to fixed-criterion than to the Bayesian strategy for the presented stimuli and range of tested reliability levels . It is possible that with a wider range of stimuli and reliabilities , and possibly with different ways of reporting ( e . g . , estimation instead of discrimination ) , we would be able to distinguish the Bayesian strategy from a fixed-criterion heuristic . Finally , we note that model predictions of our Bayesian models are good but still show systematic discrepancies from the data for the explicit causal inference task ( Figs 3C and 6B ) . Previous work has found similar discrepancies in model fits of unity judgments data across multiple sensory reliabilities ( e . g . , see Fig 2A in [21] ) . This suggests that there is some element of model mismatch in current Bayesian causal inference models , possibly due to difference in noise models or to other processes that affect causal inference across cue reliabilities , which deserves further investigation . We performed our analysis within a factorial model comparison framework [50] . Even though we were mainly interested in a single factor ( causal inference strategy ) , previous work has shown that the inferred observer’s decision strategy might depend on other aspects of the observer model , such as sensory noise or prior , due to nontrivial interactions of all these model components [37] . Our method , therefore , consisted of performing inference across a family of observer models that explicitly instantiated plausible model variants . We then marginalized over details of specific observer models , looking at posterior probabilities of model factors , according to a hierarchical Bayesian Model Selection approach [54 , 55] . We applied a few tweaks to the Bayesian Model Selection method to account for our focus on factors as opposed to individual models ( see Methods ) . Our approach was fully Bayesian in that we took into account parameter uncertainty ( by computing a metric , LOO , based on the full posterior distribution ) and model uncertainty ( by marginalizing over model components ) . A fully Bayesian approach has the advantages of explicitly representing uncertainty in the results ( e . g . , credible intervals over parameters ) , and of reducing the risk of overfitting , although it is not immune to it [65] . In our case , we marginalized over models to reduce the risk of model overfitting , which is a complementary problem to parameter overfitting . Model overfitting is likely to happen when model selection is performed within a large number of discrete models . In fact , some authors recommend to skip discrete model selection altogether , preferring instead inference and Bayesian parameter estimation in a single overarching or ‘complete’ model [66] . We additionally tried to reduce the risk of model overfitting by balancing prior probabilities across factors , although we noted that this may not be enough to counterbalance the additional flexibility that a model factor gains by having more sub-models than a competitor . Our practical recommendation , until more sophisticated comparison methods are available , is to ensure that all model components within a factor have the same number of models , and to limit the overall number of models . Our approach was also factorial in the treatment of different tasks , in that first we analyzed each bisensory task in isolation , and then combined trials from all data in a joint fit . The fully Bayesian approach allowed us to compute posterior distributions for the parameters , marginalized over models ( see Fig 5 ) , which in turn made it possible to test whether model parameters were compatibile across tasks , via the ‘compatibility probability’ metric . The compatibility probability is an approximation of a full model comparison to test whether a given parameter is the same or should differ across different datasets ( in this case , tasks ) , where we consider ‘sameness’ to be the default ( simplyfing ) hypothesis . We note that if the identity or not of a parameter across datasets is a main question of the study , its resolution should be addressed via a proper model comparison . With the joint fits , we found that almost all parameters were well constrained by the data ( except possibly for the parameters governing the observers’ priors , σprior and Δprior ) . An alternative option to better constrain the inference for scarce data or poorly identified parameters is to use informative priors ( as opposed to non-informative priors ) , or a hierarchical approach that assumes a common ( hyper ) prior to model parameters across subjects [67] . The general goal of a model comparison metric is to score a model for goodness of fit and somehow penalize for model flexibility . In our analysis we have used Pareto-smoothed importance sampling leave-one-out cross-validation ( PSIS-LOO [53] ) as the main metric to compare models ( simply called LOO in the other sections for simplicity ) . In fact , there is a large number of commonly used metrics , such as ( corrected ) Akaike’s information criterion ( AIC ( c ) ) [68] , Bayesian information criterion ( BIC ) [68] , deviance information criterion ( DIC ) [69] , widely applicable information criterion ( WAIC ) [70] , and marginal likelihood [71] . The literature on model comparison is vast and with different schools of thought—by necessity here we only summarize some remarks . The first broad distinction between these metrics is between predictive metrics ( AIC ( c ) , DIC , WAIC , and PSIS-LOO ) [72] , that try to approximate out-of-sample predictive error ( that is , model performance on unseen data ) , and BIC and marginal likelihood , which try to establish the true model generating the data [71] . Another orthogonal distinction is between metrics based on point estimates ( AIC ( c ) and BIC ) vs . metrics that use partial to full information about the model’s uncertainty landscape ( DIC , WAIC , PSIS-LOO , based on the posterior , and the marginal likelihood , based on the likelihood integrated over the prior ) . First , when computationally feasible we prefer uncertainty-based metrics to point estimates , since the latter are only crude asymptotic approximations that do not take the model and the data into account , besides simple summary statistics ( number of free parameters and possibly number of data points ) . Due to their lack of knowledge of the actual structure of the model , AIC ( c ) and BIC can grossly misestimate model complexity [72] . Second , we have an ordered preference among predictive metrics , that is PSIS-LOO ≻ WAIC ≻ DIC ≻ AIC ( c ) [72] . The reason is that all of these metrics more or less asymptotically approximate full leave-one-out cross validation , with increasing degree of accuracy from right to left [53 , 72] . As mentioned before , AIC ( c ) works only in the regime of a large amount of data . DIC , albeit commonly used , has several issues and requires the posterior to be multivariate normal , or at least symmetric and unimodal—gross failures can happen when this is not the case , since DIC bases its estimate of model complexity on the mean ( or some other measure of central tendency ) of the posterior [72] . WAIC is a great improvement over DIC and does not require normality of the posterior , but its approximation is generally superseded by PSIS-LOO [53] . Moreover , PSIS-LOO has a natural diagnostic , the exponents of the tails of the fitted Pareto distribution , which allows the user to know when the method may be in trouble [53] . Full leave-one-out cross validation is extremely expensive , but PSIS-LOO only requires the user to compute the posterior via MCMC sampling , with no additional cost with respect to DIC or WAIC . Similarly to WAIC , PSIS-LOO requires the user to store for each posterior sample the log likelihood per trial , which with modern computers represent a negligible storage cost . The marginal likelihood , or Bayes factor ( of which BIC is a poor approximation ) , is an alternative approach to quantify model evidence , related to computing the posterior probability of the models [71] . While this is a principled approach , it entails several practical and theoretical issues . First , the marginal likelihood is generally hard to compute , since it usually involves a complicated , high-dimensional integral of the likelihood over the prior ( although this computation can be simplified for nested models [73] ) . Here , we have applied a novel approximation method for the marginal likelihood following ideas delineated in [74 , 75] , obtaining generally sensible values . However , more work is needed to establish the precision and applicability of such technique . Besides practical computational issues , the marginal likelihood , unlike other metrics , is sensitive to the choice of prior over parameters , in particular its range [66] . Crucially , and against common intuition , this sensitivity does not reduce with increasing amounts of data . A badly chosen ( e . g . , excessively wide ) prior for a non-shared parameter might change the marginal likelihood of a model by several points , thus affecting model ranking . The open issue of prior sensitivity has led some authors to largely discard model selection based on the marginal likelihood [66] . For these reasons , we chose ( PSIS- ) LOO as the main model comparison metric . As a test of robustness , we also computed other metrics and verified that our results were largely independent of the chosen metric , or investigated the reasons when it was not the case . As a specific example , in our analysis we found that LOO and marginal likelihood ( or BIC ) generally agreed on all comparisons , except for the sensory noise factor . Unlike LOO , the marginal likelihood tended to prefer constant noise models as opposed to eccentricity-dependent models . Our explanation of this discrepancy is that for our tasks eccentricity-dependence provides a consistent but small improvement to the goodness of fit of the models , which can be overrided by a large penalty due to model complexity ( BIC ) , or to the chosen prior over the eccentricity-dependent parameters ( wvis , wvest ) , whose range was possibly wider than needed ( see Fig 5 ) . The issue of prior sensitivity ( specifically , dependence of results on an arbitrarily chosen range ) can be attenuated by adopting a Bayesian hierarchical approach over parameters ( or a more computationally feasibile approximation , known as empirical Bayes ) , which is venue for future work . Model evaluation , especially from a Bayesian perspective , is a time-consuming business . For this reason , we have compiled several state-of-the-art methods for model building , fitting and comparison , and made our code available . The main issue of many common observer models in perception is that the expression for the ( log ) likelihood is not analytical , requiring numerical integration or simulation . To date , this limits the applicability of modern model specification and analysis tools , such as probabilistic programming languages , that exploit auto-differentiation and gradient-based sampling methods ( e . g . , Stan [76] or PyMC3 [77] ) . The goal of such computational frameworks is to remove the burden and technical details of evaluating the models from the shoulders of the modeler , who only needs to provide a model specification . In our case , we strive towards a more modest goal of providing black-box algorithms for optimization and MCMC sampling that exhibit a larger degree of robustness than standard methods . In particular , for optimization ( maximum likelihood estimation ) we recommend Bayesian adaptive direct search ( BADS [78] ) , a technique based on Bayesian optimization [79 , 80] , which exhibits robustness to noise and jagged likelihood landscapes , unlike common optimization methods such as fminsearch ( Nelder-Mead ) and fmincon in MATLAB . Similarly , for MCMC sampling we propose a sampling method that combines the robustness and self-adaptation of slice sampling [81] and ensemble-based methods [82] . Crucially , our proposed method almost completely removes the need of expensive trial-and-error tuning on the part of the modeler , possibly one of the main reasons why MCMC methods and full evaluation of the posterior are relatively uncommon in the field ( to our knowledge , this is the first study of causal inference in multisensory perception to adopt a fully Bayesian approach ) . Our framework is similar to the concept behind the VBA toolbox , a MATLAB toolbox for probabilistic treatment of nonlinear models for neurobiological and behavioral data [83] . The VBA toolbox tackles the problem of model fitting via a variational approximation that assumes factorized , Gaussian posteriors over the parameters ( mean field/Laplace approximation ) , and provides the variational free energy as an approximation ( lower bound ) of the marginal likelihood . Our approach , instead , does not make any strong assumption , using MCMC to recover the full shape of the posterior , and state-of-the-art techniques to assess model performance . Detailed , rigorous modeling of behavior is a necessary step to constrain the search for neural mechanisms implementing decision strategies [84] We have provided a set of computational tools and demonstrated how they can be applied to answer specific questions about internal representation and decision strategies of the observer in multisensory perception , with the goal of increasing the set of models that can be investigated , and the robustness of such analyses . Thus , our tools can be of profound use not only to the field of multisensory perception , but to biological modeling in general .
The Institutional Review Board at the Baylor College of Medicine approved the experimental procedures ( protocol number H-29411 , “Psychophysics of spatial orientation and vestibular influences on spatial constancy and movement planning” ) and all subjects gave written informed consent . We build upon standard causal inference models of multisensory perception [18] . For concreteness , in the following description of causal inference models we refer to the visuo-vestibular example with binary responses ( ‘left/right’ for discrimination , and ‘yes/no’ for unity judgements ) . The basic component of any observer model is the trial response probability , that is the probability of observing a given response for a given trial condition ( e . g . , stimulus pair , uncertainty level , task ) . In the following we briefly review how these probabilities are computed . All analysis code was written in MATLAB ( Mathworks , Inc . ) , with core computations in C for increased performance ( via mex files in MATLAB ) . Code is available at https://github . com/lacerbi/visvest-causinf . For a given model , we denote its set of parameters by a vector θ . For a given model and dataset , we define the parameter log likelihood function as LL ( θ , model ) = log p ( data | θ , model ) = log ∏ i = 1 N trials p ( r ( i ) | s vis ( i ) , s vest ( i ) , c vis ( i ) , θ , model ) = ∑ i = 1 N trials log p ( r ( i ) | s vis ( i ) , s vest ( i ) , c vis ( i ) , θ , model ) ( 10 ) where we assumed conditional independence between trials; r ( i ) denotes the subject’s response ( ‘right’ or ‘left’ for the discrimination trials; ‘common’ or ‘separate’ causes in unity judgment trials ) ; s vis ( i ) and s vest ( i ) are , respectively , the direction of motion of the visual ( resp . vestibular ) stimulus ( if present ) , and c vis ( i ) is the visual coherence level ( that is , reliability: low , medium , or high ) , in the i-th trial . We built different observer models by factorially combining three factors: causal inference strategy ( Bayesian , fixed-criterion , or fusion ) ; shape of sensory noise ( constant or eccentricity-dependent ) ; and type of prior over heading directions ( empirical or independent ) ; see Fig 2A and ‘Causal inference models’ section of the Methods for a description of the different factors . For each subject , we fitted the different observer models , first separately to different tasks ( unity judgment and bisensory inertial discrimination ) , and then performed a joint fit by combining datasets from all tasks ( including the unisensory discrimination task ) . We evaluated the fits with a number of model comparison metrics and via an objective goodness of fit metric . Finally , we combined evidence for different model factors across subjects with a hierarchical Bayesian approach . We verified our ability to distinguish different models with a model recovery analysis , described in S1 Appendix . The Bayesian cookbook for causal inference in multisensory perception , or simply ‘the cookbook’ , consists of a recipe to build causal inference observer models for multisensory perception , and a number of algorithms and computational techniques to perform efficient and robust Bayesian comparison of such models . We applied and demonstrated these methods at different points in the main text; further details can be found here in the Methods and S1 Appendix . For reference , we summarize the main techniques of interest in Table 3 .
|
As we interact with objects and people in the environment , we are constantly exposed to numerous sensory stimuli . For safe navigation and meaningful interaction with entities in the environment , our brain must determine if the sensory inputs arose from a common or different causes in order to determine whether they should be integrated into a unified percept . However , how our brain performs such a causal inference process is not well understood , partly due to the lack of computational tools that can address the complex repertoire of assumptions required for modeling human perception . We have developed a set of computational algorithms that characterize the causal inference process within a quantitative model based framework . We have tested the efficacy of our methods in predicting how human observers judge visual-vestibular heading . Specifically , our algorithms perform rigorous comparison of alternative models of causal inference that encompass a wide repertoire of assumptions observers may have about their internal noise or stimulus statistics . Importantly , our tools are widely applicable to modeling other processes that characterize perception .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"classical",
"mechanics",
"applied",
"mathematics",
"social",
"sciences",
"neuroscience",
"simulation",
"and",
"modeling",
"perception",
"algorithms",
"optimization",
"cognitive",
"psychology",
"mathematics",
"vision",
"research",
"and",
"analysis",
"methods",
"research",
"assessment",
"physics",
"psychology",
"inertia",
"sensory",
"cues",
"research",
"validity",
"biology",
"and",
"life",
"sciences",
"sensory",
"perception",
"physical",
"sciences",
"cognitive",
"science",
"motion"
] |
2018
|
Bayesian comparison of explicit and implicit causal inference strategies in multisensory heading perception
|
Zika virus ( ZIKV ) is a mosquito-borne pathogen which has recently spread beyond Africa and into Pacific and South American regions . Despite first being detected in 1947 , very little information is known about the virus , and its spread has been associated with increases in Guillain-Barre syndrome and microcephaly . There are currently no known vaccines or antivirals against ZIKV infection . Progress in assessing interventions will require the development of animal models to test efficacies; however , there are only limited reports on in vivo studies . The only susceptible murine models have involved intracerebral inoculations or juvenile animals , which do not replicate natural infection . Our report has studied the effect of ZIKV infection in type-I interferon receptor deficient ( A129 ) mice and the parent strain ( 129Sv/Ev ) after subcutaneous challenge in the lower leg to mimic a mosquito bite . A129 mice developed severe symptoms with widespread viral RNA detection in the blood , brain , spleen , liver and ovaries . Histological changes were also striking in these animals . 129Sv/Ev mice developed no clinical symptoms or histological changes , despite viral RNA being detectable in the blood , spleen and ovaries , albeit at lower levels than those seen in A129 mice . Our results identify A129 mice as being highly susceptible to ZIKV and thus A129 mice represent a suitable , and urgently required , small animal model for the testing of vaccines and antivirals .
Zika virus ( ZIKV ) is a mosquito-borne flavivirus first isolated from a sentinel rhesus macaque placed in the Zika forest in Uganda in 1947; thereafter , ZIKV was isolated from a pool of Aedes africanus mosquitoes from the same forested area [1] . During the 1960s-1980s evidence for ZIKV infection was detected in several African and Asian countries including Nigeria [2] , Pakistan [3] , Malaysia [4] and Indonesia [5] . Until an outbreak in 2007 on Yap Island , Micronesia , no transmission of ZIKV had been reported outside of Africa and Asia and only 14 cases of human ZIKV disease had been previously documented [6] . No further transmission was identified until 2013 when French Polynesia reported autochthonous cases [7] which led to an epidemic in which an estimated 28 , 000 people sought medical advice ( 11% of the population ) [8] . The virus then spread rapidly throughout the Pacific region [9] and there was an imported case into the UK , a male tourist who visited the Cook Islands in 2014 . Interestingly , this case showed evidence of ZIKV RNA in semen two months post onset of symptoms and underlined the prolonged potential for sexual transmission of ZIKV [10] . In 2015 , ZIKV was reported in Brazil [11] . Whilst it is not known how it became established in this previously non-endemic region , one theory presented by Brazilian researchers is that it entered the country during the 2014 football World Cup [12]; another is that it was introduced during the Va’a World Sprint Championship canoe race held in Rio de Janeiro , in 2014 [13] . The latter is supported by phylogenetic studies [14] and also by the fact that no ZIKV-endemic Pacific countries competed in the football World Cup . The expansion of ZIKV endemicity follows closely that of chikungunya virus , which now circulates in most inhabited continents where competent mosquito vectors are extant , and has become a global public health problem in the past decade [15] . At the time of writing , the number of countries reporting autochthonous transmission of ZIKV is increasing with the possibility of further spread into non-endemic areas including North America [16] . The first importation of a ZIKV infected traveller into the US during the ongoing outbreak was reported in January 2016 [17] . On 1st February , 2016 , the World Health Organisation declared ZIKV to be a “Public Health Emergency of International Concern” [18] . Recently , ZIKV infection has been associated with GBS , an autoimmune disease that causes acute or subacute flaccid paralysis [19 , 20] . In December 2013 , a patient from French Polynesia presented with GBS a week after confirmation of acute ZIKV infection [21] . Subsequent GBS cases were confirmed and correlated temporally with the ZIKV outbreak [9] . The incidence rates of GBS during the ZIKV outbreak were approximately 20-fold higher than expected given both the size of the French Polynesian population and the established incidence rates of GBS ( 1–2 per 100 , 000 population per year ) [22] . Whilst the recent temporal and spatial association between the ZIKV outbreak and the increase in cases of GBS implies a potential link , at present there are no data to confirm ZIKV as the antigenic stimulus predisposing individuals to this autoimmune disease [9] . Recent widespread attention has been given to increases in the cases of microcephaly and its coincidence with the ZIKV epidemic in South America . Over the past five years , the incidence of microcephaly has been between 130 and 170 cases annually , but in the first nine months of 2015 this figure roughly doubled [23] . In the last three months of 2015 , over 2400 further cases were reported [23] . This coincides with the rapid spread of ZIKV which was first confirmed in Brazil in May 2015 [11]; indeed , ZIKV has been shown to be present in abnormal foetal brain tissue linked to microcephaly [48 , 49] . However , clear scientific evidence of this aetiology has yet to emerge . Several other infections and toxins are known to cause foetal brain insults and teratogenic effects [24] , including: West Nile virus [25 , 26]; rubella virus [27]; cytomegalovirus [28]; herpes simplex virus type 1 [29]; varicella zoster virus [30]; human immunodeficiency virus [31]; lymphocytic choriomeningitis virus [32]; and mosquito larvicide [33] . Furthermore the diagnosis of ZIKV infection is not straightforward [34]; the opportunity to detect viral RNA in serum is often missed; and serological techniques show extensive cross reactivity between antibodies triggered by different flavivirus infections or past vaccination . Thus there is an immediate and urgent need for research work into ZIKV with a priority on experimental studies in vivo . A small animal disease model would advance knowledge about this pathogen , and open avenues for further research including studies of ZIKV vertical transmission , and any direct or indirect effects of infecton on neural development [35] . It would also provide a means for assessing novel interventions , the pathogenicity of different strains , and the effects of mutations . Mouse models reported previously have relied on the use of juvenile animals and/or intracerebral inoculations [1 , 2 , 36–44] . In this work we have developed an adult model and simulated a natural route of infection . We believe this model will enable a wider range of studies to be conducted in order to advance our knowledge of ZIKV .
ZIKV ( strain MP1751 , isolated in 1962 from pools of Aedes africanus [40] ) was obtained from the National Collection of Pathogenic Viruses , UK . Virus was cultivated in Vero cells ( European Collection of Cell Cultures , UK ) grown in with Dulbecco’s Modified Eagle Medium containing GlutaMAX ( Invitrogen , UK ) and supplemented with 2% heat-inactivated foetal bovine serum ( Sigma , UK ) . Infectious virus titres were determined by plaque assay on Vero cells incubated for six days and using a 1% carboxymethyl cellulose overlay . All procedures with animals were undertaken according to the United Kingdom Animals ( Scientific Procedures ) Act 1986 . These studies were approved by the ethical review process of Public Health England , Porton Down , UK and the Home Office , UK via an Establishment Licence ( PEL PCD 70/1707 ) and project licence ( 30/3147 ) . Mice deficient in the IFN-α/β receptor ( A129 ) and congenic control mice ( 129Sv/Ev ) were obtained from an established breeding colony approved by the UK Home Office ( B&K Universal , UK ) . Female mice aged 5–6 weeks were used for all studies . Virus inocula containing 106 plaque-forming units ( pfu ) of ZIKV diluted in phosphate buffered saline ( PBS ) were administered subcutaneously in volumes of 40μL into each of the right and left hind legs just above the ankle of 12 A129 and 12 129Sv/Ev mice . Control groups of a further three animals of each mouse strain were inoculated with 40 μL PBS in each of the hind legs . Mice were monitored three times each day for clinical signs of disease and a numerical score was assigned at each observation ( 0 normal; 2 ruffled fur; 3 lethargy , pinched , hunched , wasp waisted; 5 laboured breathing , rapid breathing , inactive , neurological; and 10 immobile ) . Temperatures were recorded by an indwelling temperature chips and weights were also recorded daily . A set of humane clinical end points were defined by veterinary staff as a 20% weight loss , or 10% weight loss and a clinical symptom , which mandated euthanasia . At 3 and 7 days post-challenge , a group of four animals from each of the A129 and 129Sv/Ev ZIKV-challenged groups were scheduled to be culled to assess local responses . At necropsy , samples of spleen , liver , brain and ovary were collected and immediately frozen at -80°C for virological analysis or inserted into pots containing 10% neutral buffered saline for microscopic analysis . Blood was collected into RNAprotect tubes ( Qiagen , UK ) for viral load testing . Tissue samples were weighed and homogenised into PBS using ceramic beads and an automated homogeniser ( Precellys , UK ) using six 5 second cycles of 6500 rpm with a 30 second gap . 200 μL of tissue homogenate or blood solution was transferred to 600 μL RLT buffer ( Qiagen , UK ) for RNA extraction using the RNeasy Mini extraction kit ( Qiagen , UK ) ; samples were passed through a QIAshredder ( Qiagen , UK ) as an initial step . A ZIKV specific real-time RT-PCR assay was utilised for the detection of viral RNA from subject animals . The primer and probe sequences were adopted from a published method [45] with in-house optimisation and validation performed to provide optimal mastermix and cycling conditions . Real-time RT-PCR was performed using the SuperScript III Platinum One-step qRT-PCR kit ( Life Technologies , UK ) . The final mastermix ( 15μL ) comprised 10μL of 2x Reaction Mix , 1 . 2μL of PCR-grade water , 0 . 2μL of 50mM MgSO4 , 1μl of each primer ( ZIKV 1086 and ZIKV 1162c both at 18μM working concentration ) , 0 . 8μL of probe ( ZIKV 1107-FAM at 25μM working concentration ) and 0 . 8μL of SSIII enzyme mix . 5μL of template RNA was added to the mastermix in order to give a final reaction volume of 20μL . The cycling conditions used were 50°C for 10 minutes , 95°C for 2 minutes , followed by 45 cycles of 95°C for 10 seconds and 60°C for 40 seconds and a final cooling step of 40°C for 30 seconds . Quantification analysis using fluorescence was performed at the end of each 60°C step . Reactions were run and analysed on the 7500 Fast platform ( Life Technologies , UK ) using 7500 software version 2 . 0 . 6 . Quantification of viral load in samples was performed using a dilution series of quantified RNA oligonucleotide ( Integrated DNA Technologies ) . The oligonucleotide comprised the 77 bases of ZIKV RNA targeted by the assay , based on GenBank accession AY632535 . 2 and was synthesised to a scale of 250 nmole with HPLC purification . This RNA oligonucleotide was quantified using Nanodrop technology ( Thermo Scientific , UK ) with a 10-fold dilution series of quantified material used to assess viral load in subject samples . Eight dilutions of quantified ZIKV RNA oligonucleotide control material were used to assess the limit of detection ranging from 1 , 000 , 000 copies per reaction to 0 . 1 copies per reaction; all dilutions were run in triplicate . The limit of detection for this assay was assessed to be 10 genome copies per PCR reaction with 100% detection for this and all higher concentrations . All samples for 1 copy per reaction and 0 . 1 copies per reaction were negative . In-run analysis indicated the reaction efficiency to be 101% with a slope of -3 . 29 , a Y intercept of 40 . 17 cycles and an R2 value of 0 . 998 . Tissue samples were fixed in 10% neutral buffered formalin for 48 hours and processed routinely to paraffin wax . Sections were cut at 3–5 μm , stained with haematoxylin and eosin ( H&E ) and examined microscopically . Lesions referable to infection were scored subjectively using the following scale: within normal limits ( WNL ) , minimal , mild , moderate and marked . The pathologist was blinded to the groups in order to prevent bias .
Groups of 5–6 week old mice either lacking receptors for IFN-α/β ( A129 ) or from the wild-type strain ( 129Sv/Ev ) were subcutaneously inoculated with 106 pfu ZIKV in the lower leg . Survival analysis was compared between the two ZIKV-infected mouse strains and mock-infected animals which received PBS only , and demonstrated that all A129 animals met humane clinical endpoints 6 days after challenge with ZIKV ( Fig 1A ) . Wild-type 129Sv/Ev mice all survived the 14 day length of the study , as did the control animals . When weights were compared between groups , it could be seen that the ZIKV-challenged A129 mice started to lose weight rapidly after day 3 post-challenge , whereas the other groups all demonstrated a gradual increase over the course of the study indicating that animals were healthy ( Fig 1B ) . Similarly , the temperatures of the A129 mice showed differences compared to the other groups , with a gradual increase until day 4 post-challenge and then a rapid decrease ( Fig 1C ) . The A129 ZIKV-challenged mice were the only ones which exhibited signs of disease post-challenge , with signs first being noted on day 5 and increasing until humane endpoints were met on day 6 post-challenge ( Fig 1D ) . None of the other groups had any clinical signs recorded for the duration of the study . Viral RNA levels were quantified in the blood and tissues ( brain , ovary , spleen and liver ) of ZIKV-challenged A129 and 129Sv/Ev mice culled at day 3 post-challenge and days 6 ( A129 ) or 7 ( 129Sv/Ev ) post-challenge ( Fig 2 ) . Results demonstrated that in A129 mice , the viral RNA was detected in all of the samples at day 3 post-challenge , with levels highest in the spleen . By day 6 post-challenge , the levels remained high but were reduced in all of the samples compared to day 3 apart from the brain . For 129Sv/Ev mice , viral RNA was only detected in the blood , ovary and spleen at day 3 post-challenge , and at day 7 post-challenge was still within the organs but was undetected in the circulation . The levels of viral RNA in 129Sv/Ev mice were orders of magnitude lower than those found in A129 mice . Samples from PBS mock-challenged animals were consistently negative for viral RNA . Inflammatory and degenerative changes were observed in the brains of A129 mice challenged with ZIKV . These comprised widespread nuclear fragments , scattered diffusely throughout the grey and white matter ( Fig 3A ) . Perivascular cuffing of vessels was observed in the parenchyma and meninges by mononuclear cells , many having the morphology of lymphocytes ( Fig 3B ) . Varying numbers of polymorphonuclear cells ( PMNs ) were noted in the grey and white matter , often near blood vessels ( Fig 3C ) . Partially degenerated cells having hyper-eosinophilic cytoplasm and irregularly , shaped , partially condensed nuclei were noted amongst the neurons of the hippocampus ( Fig 3D ) . Numerous small well-defined germinal centres with apoptotic bodies and mitotic figures were noted in the splenic white pulp of control animals . In the spleen of challenged A129 mice , large , poorly defined germinal centres were observed in the white pulp , together with numerous apoptotic bodies and a prominent depletion in the number of mature lymphocytes ( Fig 3E ) . Prominent extra-medullary haematopoiesis as well as numerous mature PMNs were present in the red pulp sinuses ( Fig 3F ) . Small foci of extra-medullary haematopoiesis were also observed in the liver . The ovaries were normal . Lesions referable to Zika virus infection were not observed in the brain , spleen , liver or ovaries of the wild-type 129Sv/Ev mice or the control animals ( Table 1 ) .
At the time of writing , our studies provide the first evidence for a susceptible adult mouse model for ZIKV infection after shallow inoculation of virus below the skin surface at one of the body’s extremities , mimicking natural infection via mosquito bite [46] . Additionally , the virus used in our studies did not need adapting to cause lethality , unlike that used in earlier studies where in vivo passaging of virus was needed to strengthen effects [37] . The concentration of 106 pfu used for inoculation was within the limits of studies with another mosquito-borne flavivirus , West Nile virus , which demonstrated a viral dose 104−106 pfu per mosquito bite [47] . Throughout the course of infection , animals were monitored regularly for clinical signs of disease , including changes in weight and temperature fluctuations . Only the A129 mice with a knockout in their type-I interferon receptor showed evidence of disease . Although fever and general malaise are typical of human ZIKV infection [48] , similar observations were not seen in infected mice . The severe weight loss , hypothermia and clinical signs that were recorded however , meant that ZIKV-challenged A129 mice all met humane endpoints after 6 days . While mortality is not a common feature of human illness , the susceptibility allows endpoints to be easily defined for testing the effectiveness of interventions . The 6 day timepoint to death in A129 mice is similar to that seen when the same strain of mice has been used in models of other arboviruses , including Crimean-Congo haemorrhagic fever virus ( CCHFv ) [49] , Hazara virus [50] , o’nyong-nyong virus [51 , 52] and Japanese encephalitis virus [53] . Despite the severe disease and lack of a type-I interferon response , A129 mice have demonstrated protective effects with vaccines for CCHFv [54] , Ross river virus [55] and chikungunya virus [56] . Therefore , by the retention of the type-II interferon response and otherwise normal immune responses [57] , A129 mice provide an appropriate model for investigating the adaptive immune response and performing active protective studies under stringent , frequently lethal , conditions . Histological analysis showed that microscopic lesions attributable to infection with ZIKV were observed in A129 mice and these focused primarily in the brain . The lesions observed share similarities with those reported in previous studies [36]; necrosis with nuclear debris in the Ammon’s horn ( part of the hippocampus ) , with perivascular cuffing and astrocyte hypertrophy , were observed in one day old Webster Swiss white mice inoculated intra-cerebrally and examined at day 7 post-challenge . In the present study , lesions generally appeared to be distributed more diffusely throughout the brain although were also seen in the hippocampus . In addition , scattered PMNs were observed . In human cases of microcephaly associated with ZIKV during pregnancy , pathological analysis has shown calcifications and other significant brain injuries [58 , 59] . While similar pathological findings were not observed during this study , this is likely due to our studying the acute phase of the disease in mice whereas the investigations into foetal brain abnormalities were undertaken months after the initial infection . Whilst the main focus of attention for ZIKV is its possible links with microcephaly , as with other intrauterine infections , it is possible that the reported cases of microcephaly represent only the more severely affected children . It may be that newborns with less severe disease , affecting not only the brain but also other organs , have not yet been identified [58]; a range of aberrations in foetal development including or apart from microcephaly might occur as a result of ZIKV infection , so other possible sequelae may be possible [35] . In three children born with microcephaly that meet criteria for vertical transmission of ZIKV , funduscopic alterations in the macular region of the eyes were observed [60] . Whilst we did not study all organs in this investigation , future work may include detailed examination at ocular and other local sites . The most significant lesions referable to infection with ZIKV in the mice were those observed in the brain . Nuclear fragmentation , perivascular cuffing by inflammatory cells , and degenerative cells in the hippocampus have been reported previously [36 , 37] . It is of interest that these earlier studies involved intra-cerebral inoculation of neonatal mice; in the present study , similar lesions were observed in genetically altered , immunocompromised , adult mice challenged by subcutaneous inoculation . The microscopic lesions in the white pulp of the of A129 mice , namely large , poorly defined germinal centres , numerous apoptotic bodies and a prominent depletion in the number of mature lymphocytes could be a direct effect of virus infection or a reactive process . The marked extra-medullary haematopoiesis in the red pulp and in the liver are likely to be reactive changes to the infection process . However , immunohistochemical ( IHC ) staining may help to clarify these possibilities by identifying whether viral antigen is present in these tissues . Currently antibody preparations against ZIKV which are suitable for IHC analysis are not available; hence these studies have not yet been conducted . When local viral loads were analysed by real-time RT-PCR , ZIKV RNA was detected in the blood , brain , ovary , spleen and liver of all A129 mice . Interestingly , viral loads in the blood , ovary , spleen and liver decreased by approximately 1–2 logs between days three and six post infection despite adverse changes in clinical symptoms , temperature and weight; however , the mean viral load in brain samples increased by over 3 logs during this period . Whilst we cannot discount very low levels of vRNA contamination in the organs from the viremic blood , the assays are quantitative and so the levels we report demonstrate viral RNA levels which are specific to the time points and tissues . Whilst no histological changes were observed in 129Sv/Ev mice challenged with ZIKV , viral RNA was detected in the blood , ovary and spleen at day 3 post-challenge and just in the ovary and spleen on day 7 post-challenge indicating that the virus entered the circulation and seeded into some of the organs . These findings suggest that despite showing no overt clinical disease , the wild-type 129Sv/Ev mice subclinically harbour ZIKV . The short timeframe of detectable viral RNA is in line with observations in humans with transient viraemia even during the acute phase of clinical disease [45 , 61] . Due to the lack of overt clinical symptoms , it could be speculated that 129Sv/Ev mice might offer an option for studying the teratogenic nature of ZIKV as these mice appeared clinically healthy and challenge with ZIKV did not result in a fatal outcome . The model system reported was assessed with respect to challenge with ZIKV strain MP1751 , an African isolate from 1962 [40] . This strain was fully authenticated by titre , next generation sequence and sterility . Since the ZIKV field is still developing , there is currently no “prototype” strain that researchers are using to allow comparison of results between laboratories . ZIKV is subdivided into two phylogenetically distinct genotypes , African and Asian [62] , and further work will look at using A129 mice to determine any differences in pathogenicity of the different strains . In summary , the results we report demonstrate that A129 mice are a susceptible mouse strain to ZIKV infection which we propose as a suitable and informative in vivo model for the testing of vaccines and therapeutics . The challenge route and dose are similar to those found in natural infection from mosquito bite . This is the first report of a suitable murine model and can accelerate the testing of novel interventions against this pathogen declared as public health emergency of international concern .
|
Since first being recognised in 1947 , Zika virus ( ZIKV ) has mainly been associated with a mild illness with symptoms including a limited fever and rash . In 2007 the virus spread from Africa into French Polynesia and then onwards across Pacific regions and into South America . In these new regions , ZIKV has been associated with more severe clinical conditions including Gullain-Barre syndrome ( GBS ) and microcephaly . There are no currently approved antivirals or vaccines available with proven activity against ZIKV , and the World Health Organisation declared the spread of ZIKV as a Public Health Emergency of International Concern . Here , we have used a mouse strain with a deficiency in the type-I interferon receptor ( A129 ) and shown that these mice are susceptible to ZIKV infection after an inoculation that closely resembles the natural route of infection via mosquito bite . Although A129 mice are deficient in the innate interferon response , they retain their adaptive immunity and thus have successfully been used as suitable models for the testing of vaccinations and antivirals . Our study provides details on a suitable small animal model for the testing of future interventions against ZIKV .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"reproductive",
"system",
"immune",
"physiology",
"microcephaly",
"pathology",
"and",
"laboratory",
"medicine",
"body",
"fluids",
"viral",
"transmission",
"and",
"infection",
"spleen",
"pathogens",
"animal",
"models",
"of",
"disease",
"microbiology",
"animal",
"models",
"viruses",
"developmental",
"biology",
"model",
"organisms",
"rna",
"viruses",
"morphogenesis",
"viral",
"load",
"research",
"and",
"analysis",
"methods",
"animal",
"models",
"of",
"infection",
"animal",
"studies",
"medical",
"microbiology",
"birth",
"defects",
"microbial",
"pathogens",
"mouse",
"models",
"congenital",
"disorders",
"hematology",
"ovaries",
"blood",
"anatomy",
"flaviviruses",
"virology",
"viral",
"pathogens",
"physiology",
"biology",
"and",
"life",
"sciences",
"organisms",
"zika",
"virus"
] |
2016
|
A Susceptible Mouse Model for Zika Virus Infection
|
The therapeutic strategy for advanced nasopharyngeal carcinoma ( NPC ) is still challenging . It is an urgent need to uncover novel treatment targets for NPC . Therefore , understanding the mechanisms underlying NPC tumorigenesis and progression is essential for the development of new therapeutic strategies . Here , we showed that TP53-regulated inhibitor of apoptosis ( TRIAP1 ) was aberrantly overexpressed and associated with poor survival in NPC patients . TRIAP1 overexpression promoted NPC cell proliferation and suppressed cell death in vitro and in vivo , whereas TRIAP1 knockdown inhibited cell tumorigenesis and enhanced apoptosis through the induction of mitochondrial fragmentation , membrane potential alteration and release of cytochrome c from mitochondria into the cytosol . Intersecting with our previous miRNA data and available bioinformatic algorithms , miR-320b was identified and validated as a negative regulator of TRIAP1 . Further studies showed that overexpression of miR-320b suppressed NPC cell proliferation and enhanced mitochondrial fragmentation and apoptosis both in vitro and in vivo , while silencing of miR-320b promoted tumor growth and suppressed apoptosis . Additionally , TRIAP1 restoration abrogated the proliferation inhibition and apoptosis induced by miR-320b . Moreover , the loss of miR-320b expression was inversely correlated with TRIAP1 overexpression in NPC patients . This newly identified miR-320b/TRIAP1 pathway provides insights into the mechanisms leading to NPC tumorigenesis and unfavorable clinical outcomes , which may represent prognostic markers and potential therapeutic targets for NPC treatment .
Nasopharyngeal carcinoma ( NPC ) is the most prevalent head and neck malignancy in Southeast Asia , especially in Southern China [1] . A majority of NPC patients are diagnosed at advanced stages , leading to approximately 30% of NPC patients developing treatment failure [2] . Although NPC is a heterogeneous disease , a combination of radiotherapy and platinum-based chemotherapy remains the standard treatment method [3 , 4] . Therefore , identification of effective molecules regulating NPC development and progression is essential for developing novel therapeutic strategies . Sustaining proliferative signaling and resisting apoptosis are typical hallmarks of cancer [5] . Mitochondria are at the core of programmed cell death or apoptosis [6 , 7] . Proteins involved in mitochondrial network could regulate the apoptotic pathway [8–10] . Thus , it is crucial to elucidate the molecular mechanisms of proliferation and mitochondrial apoptosis to excavate potential therapeutic targets for NPC therapy . TP53-regulated inhibitor of apoptosis ( TRIAP1 ) is a small ~9-kDa protein transcriptionally activated by TP53 [11 , 12] . It has been reported that TRIAP1 protects cancer cells from apoptosis through interaction with hear shock protein 70–4 ( HSP70 ) or the repression of cyclin-dependent kinase inhibitor 1 ( p21 ) [12 , 13] . Recent evidence has also revealed that TRIAP1 contributes to the resistance of apoptosis in a mitochondria-dependent manner [14 , 15] . However , the function and clinical value of TRIAP1 remain unknown in NPC . In addition , TP53 is commonly inactivated in tumor cells to escape apoptosis , indicating there may be other mechanisms regulating TRIAP1 expression and extensive investigation is warrant . MicroRNAs ( miRNAs ) are a class of small non-coding RNAs that negatively regulate gene expression by provoking mRNA degradation or suppressing mRNA translation [16–18] . Importantly , miRNAs have important roles in a wide range of biological processes , including cell proliferation , cell death and motility [19–21] . Accumulating evidences have shown that miRNAs are dysregulated and function as either oncogenes or tumor suppressors in different cancer types [22–24] . In our previous microarray study , a profile of deregulated miRNAs is identified in NPC [25 , 26] , and some miRNAs affect cell growth , proliferation and metastasis in NPC [27–29] . However , it is yet unclear whether these miRNAs maintain their apoptotic effects in NPC . Therefore , understanding the role of miRNAs in apoptosis may provide insight into the mechanisms underlying carcinogenesis and aggressiveness in NPC . In the present study , we demonstrated that TRIAP1 functioned as an oncogene in proliferation and apoptosis through preventing mitochondrial fragmentation and cytochrome c release and its overexpression was correlated with poor survival in NPC patients . miR-320b was revealed to negatively regulate TRIAP1 and exhibited proliferative inhibition and apoptotic promotion , which could be rescued by TRIAP1 overexpression . Thus , the altered miR-320b/TRIAP1 pathway contributes to the proliferation and apoptosis of NPC and may provide novel therapeutic targets for NPC treatment .
To investigate the clinical significance of TRIAP1 in NPC , we first examined TRIAP1 mRNA expression in 16 fresh-frozen NPC and 8 normal nasopharyngeal epithelial tissues . The mRNA expression level of TRIAP1 was significantly upregulated in NPC tissues ( Fig 1A , P < 0 . 01 ) and in 6 NPC cell lines compared with the normal nasopharyngeal epithelial cell line NP69 ( Fig 1B ) . In addition , protein immunoblotting analysis confirmed high TRIAP1 expression in various NPC cell lines ( Fig 1C ) . To further evaluate the expression status of TRIAP1 in NPC , we performed immunohistochemistry ( IHC ) for TRIAP1 in 204 NPC specimens . The results showed that TRIAP1 was overexpressed in 47 . 1% ( 96/204 ) of the NPC specimens ( Fig 1D ) . Importantly , the level of TRIAP1 expression was strongly correlated with distant metastasis ( P < 0 . 001 ) and death ( P = 0 . 003; S1 and S2 Tables ) . Patients with high TRIAP1 expression showed significantly shorter 5-year overall survival ( OS; 92 . 5% vs . 71 . 5% , P = 0 . 002 ) and disease-free survival ( DFS; 83 . 1% vs . 71 . 5% , P = 0 . 001; Fig 1E ) rates than those with low TRIAP1 expression . Moreover , multivariate analysis revealed TRIAP1 overexpression was an independent prognostic factor for OS ( HR , 2 . 75; 95% CI , 1 . 50–5 . 03; P = 0 . 001 ) and DFS ( HR , 2 . 54; 95% CI , 1 . 47–4 . 38; P < 0 . 001; S3 Table ) . Taken together , these data demonstrate that TRIAP1 overexpression is a risk factor for a poor prognosis in NPC patients . To explore the biological role of TRIAP1 in NPC , we transiently overexpressed or knocked down TRIAP1 in CNE-2 and SUNE-1 cells ( S1A–S1D Fig ) . While cell proliferation was significantly promoted following ectopic TRIAP1 overexpression , TRIAP1 knockdown remarkably inhibited cell proliferation ( Fig 2A , P < 0 . 001 ) . Furthermore , TRIAP1 overexpression significantly increased the colony-formation rate and anchorage-independent growth ability , which were impaired by TRIAP1 silencing ( Fig 2B–2D , P < 0 . 01 ) . These data suggest that TRIAP1 promotes NPC cell proliferation . Furthermore , we investigated the effect of TRIAP1 on apoptosis through flow cytometric analysis and found that knockdown of TRIAP1 expression induced a significantly higher rate of apoptotic cells compared with the control group ( Fig 2E , P < 0 . 01 ) . These findings demonstrate that TRIAP1 participates in regulating NPC cell apoptosis . Next , we investigated the effect of TRIAP1 on tumorigenesis in vivo through xenograft tumor models . As shown in Fig 3A–3C , TRIAP1 overexpression significantly enhanced tumor growth , with regard to both tumor volume and tumor weight , compared with the control LV-Vector group . TRIAP1 expression in dissected specimens was confirmed by IHC ( Fig 3D ) . In addition , TRIAP1 overexpression displayed a higher proportion of Ki67-positive cells and a lower percentage of TdT-mediated dUTP Nick-End Labeling ( TUNEL ) -positive cells ( Fig 3D , P < 0 . 001 ) , suggesting that cells with ectopic TRIAP1 expression were actively proliferating . Conversely , tumor growth , tumor size and tumor weight were significantly inhibited by TRIAP1 knockdown ( Fig 3E–3G , P < 0 . 05 ) . Meanwhile , the TRIAP1-knockdown group showed a decreased proliferation index and increased apoptotic index ( Fig 3H , P < 0 . 01 ) . All together , these results support that TRIAP1 promotes NPC tumor growth and inhibits cell apoptosis in vivo . To explore the underlying mechanism of TRIAP1 on NPC cell proliferation and apoptosis , we investigated the subcellular location of TRIAP1 . The observation showed that ectopically expressed TRIAP1 accumulated in the mitochondria , indicating that TRIAP1 co-localized with mitochondria ( Fig 4A ) . Furthermore , TRIAP1 knockdown led to marked mitochondrial fragmentation in both live mitochondrial images ( S2 Fig ) and fixed mitochondrial views ( Fig 4B ) . We next investigated the status of mitochondrial membrane potential ( △Ψm ) . TRIAP1 knockdown induced a significantly increased number of depolarized mitochondria ( Fig 4C ) . Taken together , these findings indicate that TRIAP1 participates in the regulation of mitochondrial fragmentation and is required for normal polarized mitochondrial membrane potential . Furthermore , we examined the potential role of TRIAP1 in mitochondria-dependent apoptosis . Immunofluorescent staining displayed that cytochrome c co-localized with mitochondria and TRIAP1 ( Fig 5A and 5B ) . Interestingly , the loss of TRIAP1 induced the release of cytochrome c from mitochondria into the cytosol accompanied by mitochondrial fragmentation ( Fig 5A and 5B ) . Subsequently , the activity of caspase-3 and -7 was significantly increased by TRIAP1 knockdown , revealing a significant induction of apoptosis ( Fig 5C ) . Together , these results suggest that knockdown of TRIAP1 led to apoptosis through mitochondrial fragmentation and the subsequent release of cytochrome c from mitochondria . To investigate the mechanisms of TRIAP1 expression aberration , we used available bioinformatic algorithms as filters to screen miRNAs targeting TRIAP1 . A total of 98 miRNAs were identified as candidates and subsequently intersected with the 33 downregulated miRNAs identified in our published data set ( NCBI/GEO/GSE32960 , n = 330 , including 312 NPC tissues and 18 normal nasopharyngeal tissues; Fig 6A and 6B and S3 Fig ) [25] . Finally , miR-320b was identified as the sole candidate ( Fig 6B , P < 0 . 01 ) . In determining whether miR-320b negatively regulates TRIAP1 expression , we found that miR-320b mimics significantly inhibited TRIAP1 expression at both the mRNA and protein levels , whereas miR-320b inhibitor increased its expression in NPC cells ( Fig 6C–6E , P < 0 . 05 ) . To further confirm the site-specific repression of miR-320b on TRIAP1 , we constructed wild-type and mutant TRIAP1 3′ UTR luciferase reporter vectors ( Fig 6F ) . miR-320b overexpression or inhibition suppressed or increased the luciferase activity of the wild-type TRIAP1 3′ UTR reporter gene but had no inhibitory effect on the mutant reporter ( Fig 6G , P < 0 . 05 ) . Taken together , these data demonstrate that TRIAP1 is a novel direct target of miR-320b in NPC cells . Subsequently , we investigated whether miR-320b has biological roles in NPC progression . Similar to the effect induced by loss of TRIAP1 , overexpression of miR-320b significantly suppressed cell proliferation ( Fig 7A , P < 0 . 01 ) , but led to mitochondrial membrane depolarization , mitochondrial fragmentation and apoptosis ( Fig 7B–7E and S4A and S4B Fig , P < 0 . 05 ) . Inversely , miR-320b inhibition increased cell proliferation , but decreased mitochondrial membrane depolarization and apoptosis ( S5A–S5F Fig , P < 0 . 05 ) . Furthermore , we explored how miR-320b exerts its functional effects . Restoration of TRIAP1 remarkably abrogated the proliferation inhibition , mitochondrial membrane depolarization , fragmentation and apoptosis induced by miR-320b ( Fig 7A–7E and S4A and S4B Fig , P < 0 . 05 ) , while inhibition of TRIAP1 expression significantly abrogated the induction of proliferation , and the suppression of mitochondrial membrane depolarization and apoptosis induced by miR-320b knockdown ( S5A–S5F Fig , P < 0 . 05 ) . In addition , enforced TRIAP1 overexpression prevented cytochrome c release from mitochondria to the cytoplasm ( Fig 7F and S6A and S6B Fig ) . These results suggest that TRIAP1 is a functional mediator of miR-320b on cell proliferation and mitochondria-dependent apoptosis in NPC . Consistent with our previous miRNA microarray data ( NCBI/GEO/GSE32960; S7 Fig ) , miR-320b was significantly downregulated in 16 fresh-frozen NPC compared with 8 normal nasopharyngeal epithelial tissues , as well as in 6 NPC cell lines compared with the normal cell line NP69 ( Fig 8A and 8B ) . Quantitative RT-PCR revealed that miR-320b expression was inversely correlated with TRIAP1 levels in NPC tissues ( n = 204; Fig 8C; P < 0 . 01 ) . When combining the expression of miR-320b and TRIAP1 , patients in group III with low miR-320b expression and high TRIAP1 expression displayed worse OS and DFS than those in groups I and II with low TRIAP1 expression ( n = 204; S8A and S8B Fig; P < 0 . 001 ) . Furthermore , miR-320b overexpression significantly suppressed tumor growth and displayed a lower proliferation index and a higher apoptotic index , while miR-320b inhibition promoted tumorigenesis and inhibited apoptosis in vivo ( Fig 8D–8G , S9 Fig; P < 0 . 05 ) . Taken together , these results support that the miR-320b/TRIAP1 pathway regulates the proliferation and apoptosis by repressing the release of cytochrome c from mitochondria , leading to NPC tumorigenesis and poor clinical outcomes ( Fig 8H ) .
In our current study , we found that TRIAP1 was upregulated and associated with poor clinical outcomes in NPC . Moreover , we firstly reported that TRIAP1 could be post-transcriptionally regulated by miR-320b , and TRIAP1 expression was inversely correlated with miR-320b expression in clinical NPC samples . Furthermore , miR-320b inhibited cell proliferation and increased apoptosis through the release of cytochrome c from mitochondria in a TRIAP1-dependent manner . Therefore , our findings uncovered a novel mechanism post-transcriptionally regulating TRIAP1 expression by miR-320b and its role in tumorigenesis and unfavorable survival in NPC . Sustaining proliferation and resisting apoptosis are hallmarks of cancer [5] . Apoptosis is programmed cell death regulated by intrinsic and extrinsic pathways centralized in the mitochondria [6 , 7] . However , cell death is commonly resisted in cancer cells . Mitochondria constantly undergo fusion and fission , which are required for cells to maintain mitochondrial integrity and respond to intrinsic apoptotic stimuli [30–33] . A low fusion-to-fission ratio has been reported to result in the loss of mitochondrial fusion , the generation of mitochondrial fragmentation and the release of cytochrome c to trigger cell death ( apoptosis ) [8–10] . Proteins involved in mitochondrial fusion and fission may participate in cancer cell resistance to apoptotic stimuli and serve as new therapeutic targets . A number of studies have observed mitochondrial-mediated apoptosis in treating NPC cells [34 , 35] . However , the regulation of mitochondrial network dynamic and apoptosis in NPC remains undefined . Emerging evidence indicates that TRIAP1 promotes cell survival and prevents apoptosis [11–14 , 36] . In our present study , we found that TRIAP1 promoted NPC cell proliferation and suppressed apoptosis in vitro and in vivo , supporting the contribution of TRIAP1 in NPC development and progression . Moreover , we demonstrated that TRIAP1 overexpression was associated with poor survival and was an independent risk factor in NPC , indicating a significant therapeutic implication of TRIAP1 in NPC . As we known , mitochondrial fragmentation is required for apoptosis induction . In this study , we found that knockdown of TRIAP1 induced mitochondrial fragmentation , membrane potential depolarization and the subsequent release of cytochrome c , and enhanced apoptosis in NPC cells , which is consistent with a previous study in colon cancer [14] . Although another study reports that TRIAP1 exerts its function through repressing p21 [13] , we found that knockdown of TRIAP1 did not increased , but decreases p21 expression , suggesting that TRIAP1did not function through repressing p21 in NPC ( S10A Fig ) . Our study elucidates the mechanisms of TRIAP1 regulating mitochondrial fragmentation and apoptosis in NPC , suggesting that TRIAP1 modulation can be a promising therapy for NPC apoptotic resistance . As we known , TRIAP1 is transcriptionally upregulated by TP53 [7 , 8] and it has been reported that TP53 can be activated by EBV encoded protein LMP1 in NPC [37–39] . However , no obvious upregulation of TP53 and TRIAP1 was observed after LMP1 overexpression in NPC cells ( S10B Fig ) , suggesting that TRIAP1 is not regulated though LMP1/TP53 pathway in NPC and there may be other regulatory mechanisms involved in TRIAP1 overexpression . In this study , we provided evidence that miR-320b negatively regulated TRIAP1 expression and exerted its function on mitochondrial fragmentation and apoptosis by targeting TRIAP1 . The inhibitory effect of miR-320b is consistent with the previous study in other cancer types and cardiomyopathy [40–42] . Here , the miR-320b level was inversely correlated with TRIAP1 expression in NPC patients , revealing that loss of miR-320b determines TRIAP1 overexpression and function in NPC . We also acknowledged that the correlation between miR-320b and TRIAP1 expression was modest in NPC patients , which indicating that there may be some other mechanisms involved in regulating TRIAP1 expression . In conclusion , our study revealed TRIAP1 as an oncogene in tumor progression and unfavorable survival , and miR-320b as a novel post-transcriptional regulator of TRIAP1 expression in NPC . The newly identified miR-320b/TRIAP1 pathway uncovers the molecular mechanisms underlying tumorigenesis and poor clinical outcomes of NPC and may facilitate the development of novel therapeutic strategies against NPC .
For Human Subject Research , this study was approved by the Institutional Ethical Review Boards of Sun Yat-sen University Cancer Center ( approval number: L20150201 ) . Written informed consent was obtained from each patient before the study . For Animal Research , all experiments were performed according to the guidelines approved by the Institutional Animal Care and Use Ethics Committee of Sun Yat-sen University Cancer Center ( approval number: 00111032 ) . A total of 204 consecutive patients diagnosed with non-distant metastatic NPC were recruited from Sun Yat-sen University Cancer Center between January 2004 and January 2007 . Paraffin-embedded biopsy specimens of individual patients were histologically-confirmed and collected for immunohistochemistry and quantitative RT-PCR . Written informed consent was obtained from each patient before the study . This study was approved by the Institutional Ethical Review Boards of Sun Yat-sen University Cancer Center . No patient had received radiotherapy or chemotherapy before biopsy . The TNM stage was reclassified according to the 7th edition of the AJCC Cancer Staging Manual . All patients were treated with radiotherapy , as previously described [43] . Patients with stage III-IV NPC received concurrent platinum-based chemotherapy [3 , 44] . The median follow-up time was 81 . 7 months ( range , 8 . 2 to 113 . 9 months ) . The detailed clinicopathological characteristics are listed in S1 Table . Sixteen fresh-frozen NPC samples with histological diagnosis and eight normal nasopharyngeal epithelium samples were collected and stored in liquid nitrogen until required . IHC analysis was performed on individual sections of 204 specimens , using the polyclonal anti-TRIAP1 antibody ( 1:200 , Sigma-Aldrich , Ronkonkoma , NY , USA ) . The degree of immunostaining was independently evaluated by two pathologists blinded to the clinicopathological characteristics of the patients . The scores were determined on the basis of the staining intensity and the percentage of positively stained cells . The staining intensity was graded as follows: 0 , no staining; 1 , weak staining , light yellow; 2 , moderate staining , yellow brown; and 3 , strong staining , brown . The percentages were scored according to the following standard: 1 , < 10% positive cells; 2 , 10–35% positive cells; 3 , 35–70% positive cells; and 4 , > 70% positive cells , as previously described [45] . The staining index was used to evaluate TRIAP1 expression in NPC sections , with possible scores of 0 , 1 , 2 , 3 , 4 , 6 , 8 , 9 and 12 . Cutoff values were determined on the basis of Receiver operating characteristic ( ROC ) curve analysis , and classified as follow: low TRIAP1 expression , staining index < 6; and high TRIAP1 expression , staining index ≥ 6 [46 , 47] . The human immortalized nasopharyngeal epithelial cell line NP69 and human NPC cell lines CNE-2 , SUNE-1 , CNE-1 , HNE-1 , HONE-1 and C666-1 were obtained from Professor Musheng Zeng within 6 months and authenticated by short tandem repeat profiling ( Sun Yat-sen University , Guangzhou , China ) . NP69 was cultured in keratinocyte/serum-free medium ( Invitrogen , Grand Island , NY , USA ) supplemented with bovine pituitary extract ( BD Biosciences , San Diego , CA , USA ) . CNE-2 , SUNE-1 , CNE-1 , HNE-1 , HONE-1 and C666-1 were grown in RPMI-1640 ( Invitrogen ) supplemented with 10% FBS ( Gibco , Grand Island , NY , USA ) ; 293FT cells were maintained in DMEM ( Invitrogen ) supplemented with 10% FBS . Total RNA from cultured cells and NPC specimens was extracted using TRIzol reagent ( Invitrogen ) as previously described [48] . RNA was reverse transcribed using reverse transcriptase ( Promega , Madison , WI , USA ) with random primers ( Promega ) for TRIAP1 or Bulge-Loop miRNA specific RT-primers ( RiboBio , Guangzhou , China ) for miR-320b . Quantitative RT-PCR reactions were performed on a CFX96 Touch sequence detection system ( Bio-Rad , Hercules , CA , USA ) . Using GAPDH or U6 as internal controls for TRIAP1 and miR-320b , respectively , the relative expression levels were calculated by the 2-ΔΔCT method [49] . Cell lysis was performed at 4°C using RIPA buffer containing a protease inhibitor cocktail ( Fdbio Science , Hangzhou , China ) . Equal amounts of protein were separated on 12% SDS-PAGE gels and transferred to polyvinylidene fluoride membranes ( Merck Millipore , Billerica , MA , USA ) . The membranes were incubated with rabbit polyclonal anti-TRIAP1 antibody ( 1:200; Santa Cruz Biotechnology , Beverly , MA , USA ) , followed by incubation with anti-rabbit IgG secondary antibody ( 1:5000; Epitomics , Burlingame , CA , USA ) . Anti-α-tubulin antibody ( 1:1000; Sigma-Aldrich ) was used as a protein loading control . Detection was visualized by enhanced chemiluminescence . Small interfering RNAs targeting TRIAP1 ( siTRIAP1-1 5’-AGGCAUGCACGGACAUGAATT-3’; siTRIAP1-2 5’-GAAAGAGAUUCCUAUUGAATT-3’ ) , miR-320b mimic ( 5’-AAAAGCUGGGUUGAGAGGGCAA-3’ ) and miR-320b inhibitor ( 5’-UUGCCCUCUCAACCCAGCUUUU-3’ ) were purchased from GenePharma company ( Suzhou , China ) . The human TRIAP1 gene , EBV encoded LMP1 gene and synthesized short hairpin RNA targeting TRIAP1 ( shTRIAP1 ) were cloned into the pSin-EF2- puromycin and pSuper-retro-puromycin vectors , respectively ( Addgene , Cambridge , MA , USA ) . CNE-2 and SUNE-1 cells were transfected with oligonucleotides ( 100 nM ) or plasmids ( 2 μg ) using Lipofectamine 2000 reagent ( Invitrogen ) , and then harvested for assays 48 h after transfection . Stable SUNE-1 cell lines expressing TRIAP1 and shTRIAP1 were generated by lentiviral infection using 293FT cells and selected using 0 . 5 μg ml-1 puromycin . For the MTT assay , 1 , 000 transfected cells were seeded in 96-well plates and exposed to MTT ( BD Biosciences ) for 4 h at 1 , 2 , 3 , 4 , and 5 days . The absorbance values were measured at 490 nm . For the colony-formation assay , 500 transfected cells were plated in six-well plates and cultured for 7 or 12 days . The colonies were stained with 0 . 5% crystal violet for quantification after fixation with 4% paraformaldehyde . The anchorage-independent growth assays were performed by soft agar culture of 2 . 5 × 104 transfected cells in six-well plates for 7 or 12 days . The colony numbers were counted using an inverted microscope . The apoptosis assay was performed using the Annexin V-FITC/PI Apoptosis Detection Kit ( KeyGEN BioTECH , Nanjing , China ) . Briefly , 2 to 5 × 105 transfected cells were rinsed twice with PBS and then resuspended in 500 μl of binding buffer , followed by staining with 5 μl of Annexin V-FITC and propidium iodide ( PI ) for 15 minutes at room temperature in the dark . The detection was performed using a flow cytometer on a Beckman Gallios detection system ( Beckman Coulter Inc . , CA , USA ) . For mitochondrial staining , transfected cells were grown on coverslips inside a Petri dish ( Nest Biotechnology , Wuxi , China ) for 24 h and stained with MitoTracker Red CMXRos ( 0 . 04 μM for live mitochondrial imaging; 0 . 4 μM for fixation and permeabilization after mitochondrial staining; ThermoFisher Scientific , Waltham , MA ) for 30 minutes at 37°C . After staining , the staining solution was replaced with pre-warmed PBS and live mitochondrial were either observed using a confocal laser-scanning microscope ( Olympus FV1000 , Tokyo , Japan ) or allowed continue to fix and permeabilize . After permeabilization , coverslips were incubated with either rabbit polyclonal anti-TRIAP1 antibody ( 1:100; Santa Cruz Biotechnology ) or mouse monoclonal anti-cytochrome c antibody ( 1:300; Cell Signaling Technology , Danvers , MA ) and then incubated with species-matched Alexa Fluor 488 or 594 goat IgG secondary antibody ( Life Technologies , Carlsbad , CA , USA ) . After being counterstained with 4′ , 6-diamidino-2-phenylindole ( DAPI ) , cells were imaged using the confocal laser-scanning microscope ( Olympus FV1000 ) . Mitochondrial membrane potentials ( △Ψm ) were detected using a JC-1 Apoptosis Detection Kit ( KeyGEN BioTECH ) . A total of 2 to 5 × 105 transfected cells were collected , washed twice with PBS , and incubated with 500 μl of prewarmed JC-1 incubation buffer at 37°C for 20 minutes . After incubation , cells were centrifuged , rinsed twice with incubation buffer , and resuspended in 500 μl of incubation buffer . The △Ψm analysis was performed using a flow cytometer on a Beckman Gallios detection system ( Beckman Coulter ) . The activity of caspase-3 and caspase-7 was determined using a Caspase-Glo 3/7 Assay Kit ( Promega ) according to the manufacturer’s instructions . Transfected cells were grown in 100 μl of cultured medium in 96-well plates and equilibrated to room temperature before the assay . A total of 100 μl of Caspase-Glo 3/7 reagent was added to each well and incubated for 1 h in the dark . Luminescence was measured using the luminometer ( Promega ) . Six-week-old male BALB/c nude mice were purchased from the Medical Experimental Animal Center of Guangdong Province ( Guangzhou , China ) . The nude mice were implanted with 1 × 106 SUNE-1 cells stably overexpressing TRIAP1 , shTRIAP1 or the corresponding negative control in the dorsal flank . For miR-320b overexpression and inhibition experiments , 1 × 106 SUNE-1 cells were subcutaneously injected into the dorsal flank of nude mice after pre-transfected with agomir-320b , antagomir-320b or NC control ( 200nM , RiboBio ) . After 7 days , when the tumor volume reached 100mm3 , intratumoral injection of agomir-320b ( 5nM ) , antagomir-320b ( 5nM ) or NC control ( 5nM ) was performed twice a week for 3 weeks . The weights and tumor volumes were measured twice weekly . The mice were sacrificed 28–35 days after implantation , and the tumors were dissected , weighted and paraffin embedded . Serial sections were subjected to IHC analysis using anti-TRIAP1 antibody or anti-Ki67 antibody ( ZSGB-Bio , Beijing , China ) . A proliferation index was measured using the percentage of positive Ki67 cells . A TUNEL assay was performed on sections from paraffin-embedded mouse specimens using the TUNEL In situ Cell Death Detection Kit , Biotin POD ( KeyGEN BioTECH ) according to manufacturer’s instructions . The apoptotic index was quantified by the proportion of positive TUNEL cells . All animal experiments were approved by the Institutional Animal Care and Use Ethics Committee . Both the conserved ( position 265–272 ) and poor conserved ( position 550–556 ) binding sites were mutated . The mutation ( Mt ) and wild-type ( Wt ) versions of the TRIAP1 3′ UTR were generated and cloned into the psiCHECK-2 luciferase reporter plasmid ( Promega ) . Cells were seeded into 6-well plates and co-transfected with the TRIAP1 Wt or Mt 3′ UTR reporter plasmids ( 2 μg ) , along with the miR-320b mimics ( 100 nM ) or miR-320b inhibitor ( 100nM ) or miRNA negative control ( miR-Ctrl , 100 nM ) using Lipofectamine 2000 reagent ( Invitrogen ) . Renilla and firefly luciferase activities were measured 24 h after transfection using the Dual-Luciferase Reporter Assay System ( Promega ) . All statistical analyses were performed using SPSS 16 . 0 software ( IBM , Armonk , NY , USA ) . The data are represented as the mean ± SD resulting from at least three independent experiments . The χ2 and Fisher’s exact tests were used to compare Categorical variables . Survival curves were calculated using the Kaplan-Meier method , and the differences were compared using a log-rank test . A multivariate analysis using a Cox proportional hazards model was performed to assess independent prognostic factors . The correlation of TRIAP1 and miR-320b mRNA expression was compared using a Pearson’s χ2 test . Comparisons between groups were evaluated by two-tailed Student’s t-tests . P < 0 . 05 was considered significant .
|
The therapeutic strategy for advanced nasopharyngeal carcinoma ( NPC ) is still challenging . The most urgent need for NPC is novel treatment targets . Therefore , understanding the mechanisms underlying NPC tumorigenesis and progression is essential for the development of new therapeutic strategies . Here , we identified TRIAP1 could serve as a prognostic biomarker in NPC , and function as an oncogene in NPC tumorigenesis and mitochondrial apoptosis through inhibiting the release of cytochrome c . Moreover , miR-320b post-transcriptionally regulated TRIAP1 expression , and exhibited inhibitory effects on proliferation and promoted apoptosis through targeting TRIAP1 . Thus , our study provides new insights into the mechanisms of NPC tumorigenesis and progression and identifies novel therapeutic targets for NPC treatment .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"gene",
"regulation",
"carcinomas",
"cell",
"processes",
"cancers",
"and",
"neoplasms",
"oncology",
"micrornas",
"mitochondria",
"bioenergetics",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"cell",
"proliferation",
"gene",
"expression",
"biochemistry",
"rna",
"carcinogenesis",
"cell",
"staining",
"cell",
"biology",
"nucleic",
"acids",
"apoptosis",
"genetics",
"biology",
"and",
"life",
"sciences",
"energy-producing",
"organelles",
"non-coding",
"rna",
"nasopharyngeal",
"carcinoma"
] |
2016
|
Overexpression of Mitochondria Mediator Gene TRIAP1 by miR-320b Loss Is Associated with Progression in Nasopharyngeal Carcinoma
|
Risk assessment of tick-borne and zoonotic disease emergence necessitates sound knowledge of the particular microorganisms circulating within the communities of these major vectors . Assessment of pathogens carried by wild ticks must be performed without a priori , to allow for the detection of new or unexpected agents . We evaluated the potential of Next-Generation Sequencing techniques ( NGS ) to produce an inventory of parasites carried by questing ticks . Sequences corresponding to parasites from two distinct genera were recovered in Ixodes ricinus ticks collected in Eastern France: Babesia spp . and Theileria spp . Four Babesia species were identified , three of which were zoonotic: B . divergens , Babesia sp . EU1 and B . microti; and one which infects cattle , B . major . This is the first time that these last two species have been identified in France . This approach also identified new sequences corresponding to as-yet unknown organisms similar to tropical Theileria species . Our findings demonstrate the capability of NGS to produce an inventory of live tick-borne parasites , which could potentially be transmitted by the ticks , and uncovers unexpected parasites in Western Europe .
Due to the combination of increased human and animal movement , socio-economic and environmental changes , as well as the complex interactions between reservoirs , pathogens , and human populations , more emerging diseases are being identified and the epidemiology of ancient diseases is changing , particularly that of vector-borne diseases [1] . After mosquitoes , ticks are the most common worldwide vector that can affect both humans and animals , and can transmit the highest variety of pathogens , including viruses , bacteria and parasites . Of these parasites , Babesia sp . or Theileria sp . are two well-known parasites responsible for several diseases that impact both human and animal health worldwide [2] , [3] . Ixodes ricinus is the most prevalent tick in Europe and the vector for several bacterial and viral pathogens [4] , as well as three parasites: B . divergens , B . microti [4] and Babesia sp . EU1 [5] , [6] . To date , no other parasites have been reported to be transmitted by this tick species , even though these ticks feed on a very large spectrum of hosts potentially infected by several parasite species . However , the list of potential or known tick-borne pathogens is constantly evolving , and emergence or re-emergence of tick-borne diseases leads to the development of unknown health risks [4] . Therefore there is a real concern that tick-borne diseases will appear in areas previously free of such diseases , consequently new studies are required to catalog those parasitic communities hosted by , and potentially transmitted by ticks . Traditionally , identification of microorganisms has relied on their cultivation in artificial environments , but it has become evident that ticks harbor a variety of microbes that may have obligate intracellular life histories and/or require highly specific medium for their cultivation , resulting in the impossibility of successfully culturing some microorganisms , especially parasites . Thus , the identification of tick-borne parasites increasingly relies on molecular detection approaches . Classically , pathogen detection in ticks is performed by PCR with specific primers . These are designed to amplify conserved microbial sequences in a predefined list of pathogens known to be transmitted by the specific collected tick species , in the specific geographical area of collection . However , this method is not at all suited to detect new or unexpected pathogens [7] , [8] . In addition , because of the relative paucity of available sequence data for tick-borne parasites , most of these techniques rely on the amplification of the 18S genes which are well conserved among parasites , implying an additional sequencing step in order to identify them at the species level . Finally , the amount of available DNA in a tick sample limits such detection to a limited number of PCR tests . Consequently , a detailed inventory of pathogenic agents carried by ticks must be carried out without a priori , necessitating novel approaches . Recently , the metagenomic profiles of the bacterial communities associated with the Ixodes ricinus tick have been assessed using Next Generation Sequencing ( NGS ) methods , which permits the characterization of the entire tick microbiome based on 16S rRNA sequencing [9] , [10] . However , such an approach does not allow identification of the bacteria at the species level , which is absolutely essential when distinguishing symbionts and commensals from the pathogenic bacteria carried by the ticks . To avoid this problem , we recently and successfully used a similar approach , but which sequenced the entire transcriptome of ticks , generating an in-depth picture of bacteria carried by Ixodes ricinus from Eastern France , and that led to the identification of both known and unexpected tick-borne bacteria [11] . In this study NGS with a similar protocol was used to produce an inventory of known and unexpected parasites carried by I . ricinus in the same area of Eastern France .
A total of 1478 I . ricinus questing nymphs were collected by flagging in three forested areas of Eastern France ( Alsace Department ) : Murbach ( 47°55′05″N , 7°8′46″E ) , Hohbuhl ( 48°27′33″N , 7°17′22″E ) and Wasselonne ( 48°38′09″N , 7°21′45″E ) , a region with abundant ticks and a concomitant high risk of disease transmission . Ticks were pooled into groups of 15 individuals and crushed in 300 µl of Dulbecco's MEM ( DMEM ) medium supplemented with 10% fetal bovine serum . A pool of 15 I . ricinus nymphs from our pathogen-free colony was treated equivalently and used as a reference as previously described [11] . This control colony originated from female ticks collected in Murbach and was reared as previously described [12] . High throughput sequencing of tick pool samples was performed as previously described [11] . Briefly , total RNA , which indicates the occurrence of viable and replicating microorganisms , and total DNA , for specific real-time PCR , were separately extracted . Wild and pathogen-free RNA samples were sequenced to a depth of 100 million and 62 million for 101 bp paired-end reads respectively . As there is no publicly available I . ricinus reference genome , we removed those sequences corresponding to the ticks themselves , or to symbiotic or commensal bacteria naturally found in ticks , by subtracting sequences homologous to sequences from the pathogen-free reference sample using the SOAP2 aligner tool . Finally , 7 787 463 remaining reads out of 70 396 392 reads initially obtained from wild ticks , were used for de novo assembly , producing 174 841 contigs . Contigs were then assigned the closest known taxonomy according to their identity percentage ( Blast search option of the National Center for Biotechnology Information , www . ncbi . nlm . nih . gov/BLAST ) , and distant alignments were not considered . Of the assigned reads , 6 . 65% of the cDNA derived sequences were of a parasitic origin , corresponding to 0 . 73% of the reads obtained from whole wild ticks . Among these sequences , contigs of significant interest were selected based on at least one of the following criteria 1 ) an identity percentage >95% with a particular parasite species , 2 ) known to be responsible for human or/and animal disease and 3 ) a high read number . Real-time PCR was performed on DNA extracted from each pool of ticks to confirm taxonomic species assignment of NGS-derived contigs . Amplification was performed as previously described [11] and the primers newly designed for this study , based on the 18S rDNA , hsp70 and CCTeta sequences present in GenBank , are presented in Table 1 . Babesia and Theileria DNA used for positive controls were kindly provided by Huseyin Bilgic , Faculty of Vet . Med , Turkey; Laurence Malandrin , ONIRIS , France; Emmanuel Cornillot , Montpellier University , France . For phylogenetic analysis , the 28S sequence data obtained via NGS ( Table 2 ) were aligned and subsequently compared with parasitic species data from GenBank using the phyml v2 . 4 . 4 software [13] , [14] . Distance matrices were calculated using the General time reversible ( GTR ) model and bootstrap analysis was performed with 1000 replications [15] . Plasmodium falciparum , a close apicomplexa was used as an out-group .
Three zoonotic Babesia species , B . divergens , B . microti and Babesia sp . EU1 were identified in I . ricinus , in addition to B . major , a parasite that only infects cattle . Transovarial transmission within ticks is characteristic of Babesia spp . , implying that ticks constitute a real parasite reservoir in the field . Following our selection criteria , four sequences were identified as belonging to the Theileria genera ( Table 2 ) . Three were most closely related to T . parva with 94–97% 18S rRNA identity , but with relatively low e-values and numbers of associated reads ( 535 in total ) . The presence of T . parva DNA was however confirmed by qPCR also based on the 18S rRNA sequence . The fourth sequence appeared to be related to T . taurotragi ( 97% 18S rRNA identity ) with higher e-values and read numbers ( 1216 ) , but no amplification could be obtained after qPCR with specific primers for the 18S rRNA encoding gene . These results indicate that some related Theileria species , but different from T . parva or T . taurotragi , are detected in I . ricinus . Rhipicephalus appendiculatus is the most common vector for T . parva and T . taurotragi , but other Rhipicephalus species can also transmit these organisms , implying flexible vector specificity . Both species occur in Africa , where T . parva mainly infects cattle , whereas T . taurotragi was found to have a wider host range [41] . Phylogenetic analysis based on 28S NGS sequence data indicated that all four ambiguous sequences ( 127324 , 131568 , 164638 and 110157 ) seemed to belong to distinct and novel apicomplexa species ( Figure 1 ) . Only one sequence ( 110157 ) , with the highest probability and read number , was confirmed to be related to a Theileria species . The other three seem to belong to Babesia species . However , considering that very few complete parasite genome sequences are available , parasite identification is mainly performed on the basis of 18S or 28S rDNA sequence analysis . These are the most highly represented parasitic sequences in GenBank , but are not the most informative in terms of species assignation . Moreover , this preliminary analysis was performed with short sequences ( 139–197 bp ) , which are not located at the same region within the 28S rDNA , therefore no definite species can be identified . Thus , further investigations are now required to clarify whether the identification of new Theileria or Babesia species in France , similar to tropical species , actually corresponds to an expanded geographical distribution of these species , and whether they have a potential pathogenic effect in mammals . Unfortunately , the absence of tick-borne parasite genome data causes difficulties in realizing such phylogenetic studies . However , in spite of the low level of robustness of theses phylogenetic analyses , our results are confirmed by other studies , in particular those demonstrating that the B . microti group is entirely divergent from either Babesia sensu stricto or Theileria species [42] . The inventory of parasitic RNA content in I . ricinus performed by NGS revealed the presence of expected viable parasites belonging to the Babesia genus , some of them being identified in France for the first time . However , the epidemiological relevance of these results must of course be interpreted with caution . Unfortunately , complete genomic data on tick-borne parasites is scarce , likely due to large genome complexity compared to the relatively small number of research teams in this field . In addition , their small genome size and the strong inter-species conservation of available sequences ( essentially 18S rRNA ) , does not permit clear species identification . Moreover , unknown species with too distant alignment and the fewest database sequences could not be identified in this context . The increased number of sequences relative to tick-borne parasites in data banks should facilitate an increase in the power of NGS techniques to detect tick-borne parasites in the future . In addition , detecting pathogenic RNA within ticks does not imply that these pathogens are actually transmitted by this arthropod . Therefore competence and epidemiological studies are also required in order to verify whether I . ricinus is implicated in the transmission of those tick-borne diseases which are present or emerging in France . And finally , further studies are also required to confirm whether the unexpected Theileria species detected here is actually novel , and whether the detection of parasitic species similar to other tropical species in France , corresponds to increasing geographical species distribution .
|
Diseases transmitted by ticks have diverse etiology ( viral , bacterial , parasitic ) and are responsible for high morbidity and mortality rates around the world , both in humans and animals . The emergence or re-emergence of tick-borne diseases is increasingly becoming a problem as the geographical distribution of several tick species is expanding , as well as the numbers of potential or known tick-borne pathogens are constantly evolving . It is thus necessary to know which microorganisms circulate within communities of this major vector to ensure adequate epidemiological surveillance . In this study , we evaluated the potential of Next-Generation Sequencing techniques ( NGS ) to produce , without a priori , an inventory of both predicted and non-expected parasites carried by Ixodes ricinus , the most prevalent human biting tick in France . Our findings suggest that NGS strategies could be used to produce an inventory of live parasites residing in ticks from a selected area , thereby expanding our knowledge base of tick-associated parasites .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"and",
"Discussion"
] |
[
"biology",
"and",
"life",
"sciences",
"veterinary",
"science",
"medicine",
"and",
"health",
"sciences"
] |
2014
|
Identification of Parasitic Communities within European Ticks Using Next-Generation Sequencing
|
The reliable response to weak biological signals requires that they be amplified with fidelity . In E . coli , the flagellar motors that control swimming can switch direction in response to very small changes in the concentration of the signaling protein CheY-P , but how this works is not well understood . A recently proposed allosteric model based on cooperative conformational spread in a ring of identical protomers seems promising as it is able to qualitatively reproduce switching , locked state behavior and Hill coefficient values measured for the rotary motor . In this paper we undertook a comprehensive simulation study to analyze the behavior of this model in detail and made predictions on three experimentally observable quantities: switch time distribution , locked state interval distribution , Hill coefficient of the switch response . We parameterized the model using experimental measurements , finding excellent agreement with published data on motor behavior . Analysis of the simulated switching dynamics revealed a mechanism for chemotactic ultrasensitivity , in which cooperativity is indispensable for realizing both coherent switching and effective amplification . These results showed how cells can combine elements of analog and digital control to produce switches that are simultaneously sensitive and reliable .
Bacterial chemotaxis enables the cell to move towards favorable environments . This sensing ability relies closely on collective coordination of several operation modules in the signal transduction pathway ( reviewed in [1][2] ) . The first component of this system is responsible for detecting environmental signals and converting them into intracellular signals . At the surface of the cell , detection of attractants and repellents is mediated by a series of chemoreceptors in the cytoplasmic membrane , the methyl-accepting chemotaxis proteins ( MCPs ) . The second component is the intracellular chemotactic pathway , which processes extracellular signal and converts it into one that is used to determine the behavior of the bacterial flagellar motors: the concentration of the soluble cytoplasmic protein CheY . Binding of repellents induces phosphorylation of CheY , whereas binding of attractants results in CheY dephosphorylation . At the end of the chemotactic pathway lies the final component of the system – the motor block – which changes its switching bias in response to changes in CheY-P ( phosphorylated CheY ) concentration . On a typical E . coli cell surface , there are 4–5 functioning bacterial flagellar motors . When most of the motors on the membrane spin counterclockwise ( CCW ) , flagellar filaments form a bundle and propel the cell steadily forward; if a few motors ( can be as few as one ) spin clockwise ( CW ) , flagellar filaments fly apart and the cell tumbles . Therefore the cell repeats a ‘run’-‘tumble’-‘run’ pattern to perform a biased random walk for chemotaxis in a low Reynolds number world [3] . The essential feature of the motor that allows effective chemotaxis is its ability to switch direction quickly and reliably in response to small changes in environmental conditions . Previous studies have revealed that the motor switching responds ultrasensitively to changes in intracellular CheY-P concentration: a high concentration of CheY-P in the cytoplasm of the cell stimulates more CW rotation , while a low concentration of CheY-P results in more CCW rotation . In WT E . coli , the cytoplasmic concentration of CheY-P is around 3 mM and the flagellar motors show stochastic reversals of rotation every second or so [4] . A small change in CheY-P concentration up or down disrupts this equilibrium and produces a large shift toward either CW or CCW rotation . The sensitivity coefficient for the change in rotational bias ( time spent in CCW vs . CW ) as a function of CheY-P concentration ( the Hill coefficient ) is ∼10 at the most sensitive part of the region of operation [5] . How the flagellar motor accomplishes this switching behavior is not fully understood , partly because structural data are difficult to obtain . It is known that CheY-P molecules interact with a ring-shaped assembly of about 34 identical FliM protein subunits and this unit is believed to be responsible for determining the direction of rotation [6][7] . For several decades , a series of models have attempted to explain the dynamic behavior of the motor switch and identify the underlying kinetic mechanisms that control the steady state behavior of motor switches [8][9] . Tu and Grinstein [10] used a theoretical argument to suggest that in a dynamical two-state ( CW and CCW ) model , temporal changes in CheY-P concentration drive the switching behavior of the motor at long time scales and produces a power-law distribution for the durations of the CCW states . Bialek et al . [11] used the bacterial motor as a model system to evaluate the noise limitation of intracellular signaling , concluding that the motor switch operates close to the theoretical limit imposed by diffusive counting noise . A key test for any model of motor switching is the ability to explain how small changes in extracellular concentration are converted into large changes in motor output . To explain this ultrasensitivity , the possibility of cooperative binding of CheY-P to the FliM subunits of the motor switch complex has been suggested [12] . However , studies focused on this binding step [13][14] have determined a Hill coefficient of ∼1 for it , which eliminates the possibility that the amplification is driven by cooperative CheY-P binding to the motor and suggests that a separate , post-binding step within the switch complex is responsible . Duke et al . [15] described a stochastic allosteric model that qualitatively reproduces the ultrasensitive switching and locked state behavior of the motors assuming energetic coupling between neighbor units on the FliM ring inspired by the classic Ising phase transition theory . In particular , this model can reproduce the Hill coefficient of the switch , the nonlinear dependence of rotational bias on CheY-P concentration and the equilibrium between the CW and CCW locked states . The model was based on two assumptions: ( a ) each subunit of the ring can exist in one of two conformations: CCW and CW state and undergoes a conformational change catalyzed by the binding of CheY-P and ( b ) a coupling between neighboring subunits favors a coherent configuration and this leads to the propagation of conformational changes along the ring . Although this model is able to qualitatively reproduce the equilibrium behavior of the motor switch , further work is needed to test its ability to reproduce the dynamics of the switching behavior . Here we investigated in detail the behavior of the conformational spread switching model and its ability to reproduce measurements of locked state intervals , the Hill coefficient and other measures of motor dynamics . We then performed a parameter space search to identify the parameters required to best match experimental findings and make further predictions .
Our Monte Carlo model is based on the approach of Duke et al . [15] and our previous work [16] , which we briefly describe here . In addition , we make it more general by extending the assumption of symmetry in their original model to include asymmetric cases . The centerpiece of the model is a multi-protein complex or oligomer ( to simulate the FliM ring ) , the individual protomers of which are identical to one another and arranged in a closed ring of size 34 . The interface between adjoining ring units represents domains at the boundary between proteins in a biological multi-protein complex . Each protomer can at any time be in either an active ( here denoted A and shaded dark in Figure 1a ) or inactive ( here denoted I and left unshaded in Figure 1a ) state , leading to CW and CCW rotation state , respectively . Each protomer can also be bound ( here denoted B ) or not bound ( here denoted N ) to a single CheY-P molecule . Then each protomer can make transitions between four possible states , AB↔AN↔IN↔IB↔AB . The model assumes that each protomer can flip reversibly between the two mechanical conformations ( CW and CCW ) and ligand binding/unbinding changes its chemical conformations , all of which together contribute to a free energy diagram shown in Figure 1a . To reproduce high sensitivity , the model further assumed a coupling energy between the mechanical conformations of adjacent protomers , which favors alike conformations between neighbors , but that the rate constant for CheY-P binding to a protomer is affected only by the conformation of the protomer itself . In the original model of Duke et al . [15] , it is assumed that the free energy of the active state , relative to that of the inactive state , changes from +EA to −EA when a protomer binds ligand , for simplicity ( Figure 1a ) . In our model , to make it more general , we introduce two separate energy differences between the active and inactive states: EA0 when a protomer is unliganded and EA1 when it is liganded ( Figure 1b ) . Under the assumption of energy symmetry , the free energy change associated with CheY-P binding can be modeled as , where is the CheY-P concentration required for neutral bias . In the asymmetric case , we use the same definition of , however , does not lead to neutral bias since . In later calculations , we use a numerical method to search for c0 . 5 ( asymmetric ) as a function of . Finally , the model includes a cooperative energy term ( EJ , here called cooperativity ) so that the free energy of a protomer is lowered by EJ for each neighbor that is in the same state ( Figure 1c ) . This interaction is crucial because it leads to the stochastic creation of semi-stable ‘domains’: regions of the ring whose constituents are either all in the active or all in the inactive state . These domains can then either ( a ) shrink and disappear , returning the ring to its previous coherent state or ( b ) grow to encompass the entire ring , a state in which it will remain until another stochastically growing domain of the opposite type will lead to another ring switch . Given all these energy combinations , a protomer can make transitions between all possible states at rate constants describing mechanical conformational changes by: is the sum of the free energy changes associated with changes in activity and interaction . The fundamental flipping frequency , ωa , was set as 104 s−1 , a typical rate of protein conformational change and consistent with previous modeling of the switch complex [15] . Lacking information about λa , the parameter that specifies the degree to which changes in the free energy affect forwards as opposed to backward rate constants , an intermediate value of λa = 0 . 5 was selected . The free energy associated with CheY-P binding depends only on the conformation of the protomer bound , not on adjacent protomers . So the rate constants describing chemical conformational changes are:where c is the concentration of CheY-P , c0 . 5 is the concentration of ligand at which protomers of the ring are 50% occupied on average under the symmetry assumption . ωb is the characteristic binding rate and ΔG ( N→B ) is the free energy associated with CheY-P binding . A value of ωb = 10 s−1 was selected based on the experimentally determined CheY-P binding rate [17] , and consistent with previous modeling of the switch complex [15][16] , and λb = 0 such that the binding rate is independent of protomer conformation . In the case of asymmetric EA , the CheY-P concentration corresponding to neutral bias can only be solved numerically . We use custom written C++ code to generate a Monte-Carlo simulation of the conformational spread model [16] ( including cases of both symmetric and asymmetric energy ) . At the beginning of each simulation , each protomer on the ring is set to active and with CheY-P bound . Later on , each protomer n of the ring is assigned two transition times , An and Bn , at which it will undergo a conformational change associated with ( A ) change between CCW and CW state and ( B ) CheY-P molecule binds on or off . The program progresses iteratively by locating the event in the [A1 , A2……A34 , B1 , B2……B34] array with the earliest execution time , and after change its state accordingly ( either mechanical state or chemical state ) , new transition time An and Bn of that protomer is updated by t−t0 = −ln ( rand ) /k , where k is the rate constant for the next transition , t0 is the simulation time when the calculation is made and rand is a random number generated in the interval 0 to 1 [18] . If the transition was associated with a change in mechanical conformation of that protomer , then transition times An+1 and An−1 for the two adjacent protomers are also recalculated ( for a closed ring of protomers , we defined An+1 for n = 34 to be A1 and An−1 for n = 1 to be A34 ) . The activity of all protomers on the ring is recorded at integer number MΔt , where Δt = 0 . 1 ms as the output sampling time interval and M goes up to 50 , 000 , 000 . The algorithm continues to update protomer activities on the ring until the simulation time exceeds a specified maximum . Following Duke et al . [15] , we assume that switching in the bacterial motor is controlled by the C-ring in the motor complex , which contains 34 copies of the protein FliM and therefore set n = 34 in our model unless otherwise stated . In our model we have in total 4 free parameters: EA0 , EA1 , EJ , c . The CheY-P concentration c is expressed in the unit of c0 . 5 and when we change it we see the ring operate at different bias and therefore the response curve can be plotted . In the following sections , when we make predictions about ring switching time , switching interval etc . , we searched the parameter space EA0 , EA1 and EJ across the ranges 0 . 5 kBT≤EA0≤1 . 5 kBT , 0 . 5 kBT≤EA1≤1 . 5 kBT , and 3 . 5 kBT≤EJ≤4 . 5 kBT ( shown in Table 1 ) , but for each parameter set , we only present results at neutral bias for simplicity .
A typical screenshot of the ring with multiple domains , labeled with a symbol legend , is shown in Figure 2a . We simulated the qualitative behavior of the ring for a few carefully chosen special cases under the symmetry assumption . Typical screenshots of the ring representing different regimes in the parameter space are shown in Figure 2b . If the activation energy is zero ( EA = 0 , Figure 2b , top row ) , growing domains can only form at random through cooperativity between neighbors , but are unstable and unable to grow sufficiently quickly to encompass the ring because any one of their constituent protomers has a high probability of flipping . When the cooperativity energy becomes high , a coherent ring conformation starts to emerge as the coupling between neighboring protomers is sufficiently strong to lock the whole ring in one conformation . However , as the activation energy is zero , the switch complex loses its ability to respond to ligand concentration changes and switching between coherent inactive and coherent active states can be very slow . In the absence of cooperativity ( EJ = 0 , Figure 2b , middle row ) , the ring displays random salt-and-pepper patterns reflecting the underlying stochasticity of the ligand binding and unbinding process . When the activation energy becomes high , absolute coupling between chemical conformation and mechanical conformation starts to emerge , and the protomers exist in inactive form only when there is no ligand bound and change to active form once ligand binds . In this case , a coherent active conformation of the ring only exists when there are 34 ligands bound to the ring and for a coherent inactive conformation of the ring we find 0 ligand bound . When cooperativity and activation energy are both present at appropriate magnitudes ( EA = 1 kBT , EJ = 4 kBT as discussed in reference [15] , Figure 2b , bottom row ) , the ring spends most of its time locked in either a coherent inactive or active conformation , with transitions between the two ( switches ) accomplished rapidly by means of a spreading domain . In order to achieve both ring stability and coherent switching , cooperativity is needed to ensure the presence and growth of domains , and activation energy is needed to stabilize these domains by ligand binding . In this energy regime of the model parameter space , the conformational spread model best simulate the performance of the flagellar motor switching responding to external signals . Ring activity , locked states and switching . We simulated the behavior of the 34-protomer ring with the method introduced earlier and EA0 = EA1 = 1 kBT , EJ = 4 kBT . A typical result is shown in Figure 3 . The top panel is a time series of the number of active ring protomers ( 34 active protomers correspond to the CW state and 0 to CCW in our convention ) . The middle panel shows the number of protomers with bound ligand . This graph matches that of the locked states ( i . e . the two fit on top of each other if superimposed ) along both axes: ligand binding makes the active state more favorable and the active state binds ligand more strongly and the two effects cooperate to produce locked state and switching behavior . The bottom panel shows the number of independent domains ( see Figure 2a for an illustration of typical domain formation present on the ring at any time ) . Domains appear within a locked CCW or CW state because of stochastic flipping events in protomers and their growth is driven by ligand binding and unbinding and protomer-protomer cooperativity . The domains are transient features of the ring's behavior . They can either ( a ) disappear or ( b ) grow ( alone or by fusing with nearby domains ) to encompass the whole ring , with these latter events corresponding to switches and occurring very rarely: <1% of domains lead to a switch . We find that at the parameter values identified by Duke et al . [15] , the number of independent domains almost never exceed 6 ( and that such a ring state only exists ∼0 . 0279% of the time ) . The vast majority of the time , the ring either contains one domain ( coherent state , 83 . 77% of the time ) or contains two domains ( 15 . 34% of the time ) . Four domains are present 0 . 8685% of the time . In addition to being able to reproduce the locked coherent state on the ring and fast switching behavior , a separate key test of the conformational spread model is its ability to reproduce the relationship between changes in CheY-P concentration and motor bias . The Hill coefficient ( the maximum sensitivity of the switch ) is defined in this case using the relation:where Y is the CW bias , h is the Hill coefficient , c is the concentration of the CheY-P and c0 . 5 is the concentration required for neutral bias ( in the case of asymmetric model , replace c0 . 5 to c0 . 5 ( asymmetric ) ) . Here we estimated the Hill coefficient by fitting a linear equation to a plot of log[Y/ ( 1−Y ) ] against log ( c/c0 . 5 ) , with the slope of this line corresponding to h . For each parameter set , its characteristic Hill curve can be generated by long time simulation with varying c , and plot CW bias of the simulated trace as a function of c . The sensitivity of ring activity to changes in ligand concentration depends more strongly on the activation energy and considerably less on the cooperativity . A lower sensitivity can be brought about by a lower activation energy or by a lower cooperativity , with the activation energy having the dominating influence . However , with a 34 protomer ring , the cooperativity can be no less than the critical cooperativity required for coherent switching to occur , i . e . EJ>3 . 5 kBT [15] . In table 1 , we present the Hill coefficient calculated for parameters EA0 , EA1 and EJ across the ranges 0 . 5 kBT≤EA0≤1 . 5 kBT , 0 . 5 kBT≤EA1≤1 . 5 kBT , and 3 . 5 kBT≤EJ≤4 . 5 kBT . We see that the experimentally determined Hill coefficient ∼10 can be reproduced by a large parameter sets . The behavioral features of the ring can be further characterized by the distributions of ( a ) the times spent in the two locked states and ( b ) the times required for both CCW→CW and CW→CCW switches . Here we used simulated ring state data to obtain the theoretical length of the locked state intervals predicted by the model . Because we have direct access to the fundamental protomer states , filtering and threshold algorithms are not needed to identify the intervals ( and switches , respectively ) . Because transitions between the two locked states are not instantaneous , we needed an unambiguous way to define CCW and CW intervals , respectively . We defined such an interval as the time ( in simulation steps ) between when the ring enters a fully locked state ( 0 or 34 active protomers , respectively ) and when it next enters the other fully locked state ( i . e . 34 or 0 active protomers , respectively ) . Distributions of locked state intervals obtained from simulation traces ( EA0 = EA1 = 1 kBT and EJ = 4 kBT at neutral bias ) equivalent to 30000 seconds of real time are shown in Figure 4 ( a ) . To make a comparison , the log-linear plot of the distributions at low ( 0 . 2 ) and high ( 0 . 8 ) CW biases are also shown in Figure 4 ( b ) . Least-squares fitting of exponential curves to the simulation data are shown overlaid . We see that the locked state distribution follows an exponential distribution . In table 1 , we presented the mean locked state interval values calculated for parameter EA0 , EA1 and EJ across the ranges 0 . 5 kBT≤EA0≤1 . 5 kBT , 0 . 5 kBT≤EA1≤1 . 5 kBT , and 3 . 5 kBT≤EJ≤4 . 5 kBT . Within the parameter range of our simulation , the minimum value of mean locked state time is 0 . 13 s and the maximum value is 22 . 17 s ( shown in Table 1 ) . The mean locked state time increases when the energy of activation or cooperativity is increased , with the activation energy EA having the dominant influence . The essential feature of interest of the model proposed by Duke et al . [15] is that the ring can simultaneously achieve very rapid switches and very stable locked states . This qualitatively matches what is observed in the rotary motors of flagellar bacteria such as E . coli , which can rotate at hundreds of RPM stably for a long period but switch direction quickly ( on the order of ms ) and stochastically . Distinct from the classic MWC model , which requires coherent switches to happen instantaneously , in our model switches occur by a mechanism of conformational spread . We defined a switch time as the time ( in simulation steps ) between when the ring leaves a fully locked state ( 0 or 34 active protomers , respectively ) and when it next enters the other fully locked state ( i . e . 34 or 0 active protomers , respectively ) . We simulated the behavior of the ring at the optimal activation energy and cooperativity values identified earlier ( EA0 = EA1 = 1 kBT , EJ = 4 kBT ) for different values of bias . The empirical distributions thus determined are shown in Figure 5 . In contrast to the distributions of locked state intervals , the switch times follow a peaked gamma distribution . At the parameter value chosen , the mean lies between 58–61 ms for low ( 0 . 2 ) , middle ( 0 . 5 ) and high ( 0 . 8 ) biases and these are statistically independent of bias and of direction of switch . In table 1 , we presented the mean switch time values calculated for parameter EA0 , EA1 and EJ across the ranges 0 . 5 kBT≤EA0≤1 . 5 kBT , 0 . 5 kBT≤EA1≤1 . 5 kBT , and 3 . 5 kBT≤EJ≤4 . 5 kBT . The result of our simulation shows that the mean switch time values changes across the ranges from 12 . 03 ms to 322 . 65 ms . The mean switch time increases when activation or cooperativity energy is increased , with the activation energy having the dominant influence . To confirm that typically one domain of opposite conformation ( rather than several ) grows to encompass the entire ring , we also computed power spectra for the ring activity traces in order to characterize the spectral properties of the ring switch complex . If switching events are associated with a single nucleation event ( a Possion step ) , we expect the power spectra of the trace to be monotonically decreasing with a ‘knee’ , i . e . display a Lorenzian profile . In contrast , if switching events are associated with multiple hidden steps , as for example in a closed biochemical system with hidden reactions , then we expect a non-Lorentzian profile with a peak ( a local maximum ) at a characteristic frequency related to the number of steps involved [19] . Our simulation results ( Figure 6 ) show the spectra thus obtained are Lorentzian without a local maximum at long times . This behavior is observed at different values of bias . These results offer an internal confirmation of the model results shown in Figure 4 , which indicate that the distributions of times spent in the locked states are exponential . Such a system would be expected to display power spectra with Lorentzian profiles . However , because the power spectra and locked state time distributions are computed independently and by different methods , the result that they predict the same behavior is an important internal test of the model . In particular , the power spectra results confirm that the locked state time distributions are not an artifact of our algorithm for detecting the start and end of a locked state . In our recent experimental paper [16] , we used a high-resolution optical system to measure the switching time and locked state interval of bacterial flagellar motors . The experimental observations confirmed that the switching time distribution follows a broad gamma distribution with mean switch time 18 . 72 ms and the locked state interval follows an exponential distribution with mean interval value 0 . 75 s at neutral bias . Hence we use mean switch time ( ∼18 ms ) , mean locked state time ( ∼0 . 75 s ) and Hill coefficient ( ∼10 ) to parameterize our model . Table 1 shows a coarse parameter search of our model with predictions of the switching time , locked state interval , and Hill coefficient . We identify the 0 . 55 kBT≤EA0≤0 . 95 kBT , 0 . 55 kBT≤EA1≤0 . 95 kBT , and 4 . 05 kBT≤EJ≤4 . 25 kBT region for a fine parameter search ( Table 2 ) . We see with only a few parameter sets , the conformational spread model is able to reproduce the three experimentally determined quantities . With the results shown in Table 2 , we find 3 groups of values that fit the experimental values determined by Bai et al . [16] . They are EA0 = 0 . 55kBT EA1 = 0 . 75kBT EJ = 4 . 15 kBT , EA0 = 0 . 75kBT EA1 = 0 . 55kBT EJ = 4 . 15 kBT , EA0 = EA1 = 0 . 65kBT EJ = 4 . 15 kBT . We therefore expect a conformational spread model with activation energy ∼0 . 65 kBT and coupling energy ∼4 . 15 kBT can well reproduce experimental observations . Please see Figure 7 for a visual summary of our computational results . For simplicity , we only showed those values with EA0 = EA1 = EA and the best-fit parameter set has been labeled by a square . Although we have identified a best-fit parameter set that can well reproduce experimental findings , we have to note that these fit values are sensitive to the parameters fixed earlier , especially to the fundamental flipping frequency ωa . Here we investigate how mean locked state time and mean switch time respond to changes of ωa while other parameters remain fixed . We see in Figure 8 that the fundamental flipping frequency is a scaling factor of the system . Both mean locked state time and mean switch time are inversely scaled by the flipping frequency . When flipping frequency is higher , each protomer on the ring makes more attempts to flip to the opposite conformation and therefore the locked state becomes less stable ( hence mean locked state time decreases ) and transition becomes much faster ( hence mean switch time decreases ) ; when flipping frequency is lower , each protomer on the ring makes fewer attempts to flip to the opposite conformation and therefore locked state becomes more stable ( hence mean locked state time increases ) and transition becomes much shorter ( hence mean switch time increases ) . In the above sections , we have determined the best-fit model parameters using experimental results . It will be interesting to test the ring behavior at different sizes using those values . When Duke et al . [15] first proposed the conformational spread model , they identified that EJ>kBT ln N ( N is the size of the ring ) is the condition under which a large ring has the characteristic of a coherent switch . In the case of 34 protomers , this condition requires that EJ>3 . 5kBT . When this condition is met , in time series of ring activity , we see for the majority of time that , the ring stays in complete active ( active protomer = 34 ) or complete inactive ( active protomer = 0 ) state . This invokes an empirical mathematical definition of ‘coherent switch’: the active number of protomers on the ring has to be in 0 or N for greater than 65% of the total simulation time . We then simulated the ring activity at sizes of 10 , 60 , 100 protomers with activation energy 0 . 65 kBT and coupling energy 4 . 15 kBT at neutral bias and the result is shown in Figure 9 . From the requirement of EJ>kBT ln N we expect to see coherent switch behavior for ring sizes at 10 , 34 , and 60 , but not at 100 . Indeed , in Figure 9 , we see with the same parameter set , the smaller the ring is , the easier a switch happens . At ring size of 10 , 34 , 60 protomers , we see clear locked states in the trace and the switching events are fast . However , at the ring size of 100 , a switch across the ring becomes very difficult and the time spent during a switch is comparable to the time the ring stays in a locked state . In Table 3 , we made predictions about the mean locked state intervals , mean switch times and Hill coefficient of the switch response with activation energy 0 . 65 kBT and coupling energy 4 . 15 kBT at different ring sizes .
In this study , we undertook a comprehensive numerical simulation analysis of a general model of stochastic allostery in a protein ring and evaluated the ability of such a model to explain the switching , sensitivity and locked state behavior of the rotary bacterial motor . We modeled the gearbox of the motor as a ring of 34 identical protomers , a geometry inspired by the FliM structure in the motor complex , believed to be responsible for motor switching . The model is able to qualitatively reproduce the motor behavior , such as locked rotation in CCW or CW state and fast switching between the two . Furthermore , based on a comprehensive parameter space search , the model can also quantitatively account for the experimentally determined switch time , locked state interval and Hill coefficient of the motor . Specifically , we found a unique set of values that fit the experimental value best , activation energy must be around 0 . 65kBT and the cooperativity around 4 . 15kBT . The bounds around these values are tight . Smaller or larger energies result in rings that either ( a ) spend too long or too little time in the locked states , ( b ) do not have the required sensitivity or ( c ) far away from this parameter regime , fail to switch coherently . With the ring operating in this parameter set , time traces of ring state ( measured as the number of active protomers ) indicate that the ring spends most of its time in one of the two locked states , with rare ( every 0 . 5–2 seconds ) switches between the two being accomplished very rapidly ( on the order of milliseconds ) . The trace of ligand activity ( measured as the number of ring protomers having bound ligand molecules ) mirrors the ring state , with the two driving each other: binding of more ligand drives active domain formation , which in turn leads to a preference for having ligand bound and conversely . Rather than being completely locked in one stable state with all protomers being either active or inactive , the ring displays constant activity in the form of nascent domains of the opposite state to the locked state , seen as ‘noise’ . For the vast majority of the time , only one such domain exists , and the presence of two ( but no more ) growing domains is frequently , but not always , associated with a switching event . The model predicts that the time spent by the ring in the locked states corresponding to CW ( all protomers active ) and CCW ( all protomers inactive ) is exponentially distributed . The model can also predict the Hill coefficient ( ∼10 ) measured for the sigmoidal curve that relates CheY-P concentration to motor bias . Near the optimal parameter point identified , the distributions of the switching times are gamma-like with a peak around 5–8 ms . To be effective , a switch must achieve two globally conflicting properties . It must accomplish sensitivity by amplifying small changes in the effector , but only over a narrow critical range ( the switching point ) . Outside this range , it must accomplish reliability by being unresponsive to changes in the effector . The allosteric switching model explains how the motor simultaneously meets these competing design requirements . Near the critical CheY-P concentration , a highly cooperative mechanism ( EJ≫1kBT ) is used to amplify small stochastically occurring nascent domains that can rapidly grow to encompass the entire switching complex . The resulting digital switch displays the desired selective ultrasensitivity but switches chaotically . In order to ensure switch reliability , CheY-P binding must moderately stabilize the protomer active state , providing a mechanism for biasing the whole switch complex by continuously varying the CheY concentration over a large range: a strategy typical of analog control . The values and ratio of the strengths of the two mechanisms must be tightly controlled in order for the switch complex to be functional . We hypothesize that this control is accomplished through the biochemical structure of the protomers and ring , which are genetically determined and so robust to intracellular noise during the cell's life . In light of recent studies of digital cellular signaling [20] , we wish to further suggest that the combination of analog and digital control here proposed to explain the behavior of the bacterial switch complex may be a motif typical of biological switch design . In this study we have focused on a ring consisting of 34 protomers because it is believed that the C-ring in the E . coli flagellar motor , which consists of 34 copies of the protein FliM , acts as the motor direction switch . However , numerous examples of protein rings and other interconnected protein complex geometries are known , including DNA polymerase sliding clamps , voltage-gated ion channels , ATP synthase etc . Each of these rings may hypothetically accomplish its function using a conformational spread mechanism , but would consist of different numbers of protomers . In our model this can be simulated by changing N , the number of elements in the ring . In this paper , we have narrowed our study to a closed protein ring . However , we have to point out that the conformational spread model as well as the numerical method we presented here can be easily modified to describe one dimensional allostery regulation in a protein chain or a strand of DNA molecules . The model can also be modified to describe signal transduction and amplification on a two dimensional plane , which will be of great use in studying functions of cellular receptors . Allostery is a widespread mechanism in biology and conformational change is the basis for a large subset of all protein function . Since protein complexes are the workhorses of the cell , we expect models similar to this and the idea of conformational spread in general to be increasingly important in systems biology and biophysics . Investigating the applicability of conformational spread models to other biological systems will be the subject of future work . Bacterial chemotactic exploration depends on the ability of the flagellar motors at the base of the flagella to perform two tasks: ( 1 ) remain stable in their current direction of rotation for long periods ( seconds ) as required and ( 2 ) switch quickly between the two directions in response to the environmental changes detected by the chemotaxis pathway . These properties make the bacterial switch an exquisite computational element that combines ultrasensitivity and reliability . In this paper we presented an analysis of a model featuring conformational spread that aims to explain the mechanism of the motor switch . Simulations confirm that this model is able to reproduce the characteristics of the motor observed in experiments . We speculated that stochastic models of conformational spread will be a common theme in protein allostery and signal transduction .
|
Bacteria swim to find nutrients or to avoid toxins . Their swimming is powered by the rotation of flagella ( hair-like structures ) that act as propellers . Each flagellum is driven by a rotary molecular engine ( the bacterial flagellar motor ) that can rotate in either a counterclockwise or clockwise direction and switches between the two directions are frequent and rapid . Although the motor has been studied in detail , we do not understand how it is able to reliably switch direction – a critical function that gives bacteria the ability to steer . In this paper we examined a mathematical model describing how a potential gearbox in the motor might work inside a ring of identical proteins . We compared the output of this model with experimental data on switching speed and other measures of motor function , finding excellent agreement . This is an exciting finding not only because the operation of the motor itself is important , but also because protein complexes play an important and ubiquitous role in cellular signal transduction and therefore , “conformational spread” may be a widespread mechanism for signal propagation in biology .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"theoretical",
"biology",
"biology",
"computational",
"biology",
"biophysics"
] |
2012
|
Conformational Spread in the Flagellar Motor Switch: A Model Study
|
An analysis of gene expression variability can provide an insightful window into how regulatory control is distributed across the transcriptome . In a single cell analysis , the inter-cellular variability of gene expression measures the consistency of transcript copy numbers observed between cells in the same population . Application of these ideas to the study of early human embryonic development may reveal important insights into the transcriptional programs controlling this process , based on which components are most tightly regulated . Using a published single cell RNA-seq data set of human embryos collected at four-cell , eight-cell , morula and blastocyst stages , we identified genes with the most stable , invariant expression across all four developmental stages . Stably-expressed genes were found to be enriched for those sharing indispensable features , including essentiality , haploinsufficiency , and ubiquitous expression . The stable genes were less likely to be associated with loss-of-function variant genes or human recessive disease genes affected by a DNA copy number variant deletion , suggesting that stable genes have a functional impact on the regulation of some of the basic cellular processes . Genes with low expression variability at early stages of development are involved in regulation of DNA methylation , responses to hypoxia and telomerase activity , whereas by the blastocyst stage , low-variability genes are enriched for metabolic processes as well as telomerase signaling . Based on changes in expression variability , we identified a putative set of gene expression markers of morulae and blastocyst stages . Experimental validation of a blastocyst-expressed variability marker demonstrated that HDDC2 plays a role in the maintenance of pluripotency in human ES and iPS cells . Collectively our analyses identified new regulators involved in human embryonic development that would have otherwise been missed using methods that focus on assessment of the average expression levels; in doing so , we highlight the value of studying expression variability for single cell RNA-seq data .
The regulatory program that ensures that a human embryo can develop successfully starting from a single cell zygote is one of the most fascinating examples of systems-level genetic control . During development , individual cells must quickly respond to internal and external signals while the number of cells that make up an embryo increases at a rapid rate . How an embryo is able to coordinate signals cooperatively across all cells , while subsets of cells undergo diverse fate transitions to specific lineages remains an open question . Inherent in an embryo’s regulatory program is the need to balance flexibility with robustness to ensure that development can continue in spite of perturbations that may occur . Studying how individual cells alter their transcriptomes as an embryo transitions through each developmental stage presents an opportunity to understand the core of the regulatory program , and specifically how robustness is maintained throughout development . Single cell technology has revolutionized almost every arena of the biological sciences but perhaps none more significantly than developmental biology . Profiling the transcriptomes of individual cells provides a means to disentangle heterogeneous properties that can identify small numbers of distinct or rare cells amongst a population of cells that otherwise appear identical based on a handful of markers [1] . As early as the 2000s , studies have demonstrated the limitations of inferences derived from bulk cell approaches , where transcriptomes from multiple cells are combined to create an ensemble representation of a generalized single cell [2–4] . This ensemble-based model , referred to as an “averaged cell” by Levsky and Singer [5] is unable to capture the variability inherent in gene expression in cell populations and therefore provides only a marginal insight into transcriptional regulation . Applying single cell profiling to understand developmental processes has been invaluable for pinpointing specific genes that direct cell fate transitions towards distinct lineages [6] . The ability to resolve heterogeneity in gene expression amongst cell populations has been useful in revising and identifying more specific cell types; and in the process , we are only beginning to appreciate just how diverse the transcriptional states underlying each developmental stage can be [7] . Studying transcription at the single cell level has forced us to confront the fact that gene expression consists of both signal and noise [8–10] . While the recognition that transcription is a stochastic process predates single cell technology [11 , 12] , it is only recently that we have come to appreciate the insight that variability in gene expression can shed on understanding regulatory control [13–15] . In the context of single cells , inter-cellular variability represents a measure of consistency or dispersion of a specific signal amongst a cell population . As a byproduct , expression variability also inversely reflects our ability to predict or have confidence in the transcriptional state for a new cell . For instance , a gene with low expression variability has high generalizability for any future cells sampled at random and therefore may be valuable as a marker of that cell’s state . Conversely , a gene with high expression variability is one whose expression levels fluctuate widely across cells in the population , and this heterogeneity could be due to several factors that can provide information about its regulation , e . g . a difference in cell cycling , cell fate , or stochastic component of the regulatory program used by the cell [16–18] ( see Fig 1A ) . When characterizing a phenotype based on differential gene expression , the statistical methods that are typically used , such as a t-test or ANOVA , are based on deviations detected in average expression and assume constant levels of variability . While this is a sensible approach for identifying overall shifts , we acknowledge that studying dispersion or variability in expression is also important to understanding regulation in single cells [19 , 20] . Single molecule studies performed at single cell resolution have demonstrated that gene expression is inherently stochastic , where even amongst isogenic cells , a gene’s expression level is not identically distributed [21] . Most of this stochasticity is because the production of a transcript requires the coordination of multiple components in a cell , some of which are present at very low concentrations . Gene expression can be modeled as a probabilistic event , and thus in a cell population , each cell contains a gene that is expressed in an all-or-nothing , binary manner according to a certain probability . This generates cell-to-cell heterogeneity in the population where some cells have genes whose expression is switched on while in other cells , expression is off [22] . Consequently , variability represents a very real component to understand gene expression , and by assuming constant variability and not studying this property directly , we may miss out on identifying key regulators in single cells [23 , 24] . Recently , Yan et al . [25] used single cell RNA-sequencing ( RNA-seq ) to elucidate the transcriptional landscape of preimplantation human embryos from distinct stages of development ( Fig 1B ) . This study used RNA-seq to comprehensively profile individual cells that were derived from the same human embryo at early stages of development . Major classes of genes with key temporal changes over the course of development were identified , including novel lncRNAs regulators , potentially new protein-coding transcripts , as well as dynamic patterns of alternative splicing . The application of single cell profiling provided new insights into transcriptional regulation of human embryos . It is worth highlighting , however , that the analytical methods used by Yan et al . [25] are based on averaging expression from the cell populations . While this is adequate for identifying average trends , their study fails to use the valuable opportunity presented in this data set to understand how variability in gene expression is distributed at the cell-to-cell level . For single cell RNA-seq data , an analysis of expression variability is useful in identifying genes that are invariantly or heterogeneously expressed in cell populations , which in turn may provide insight into regulation . Here , we use the data set generated by Yan et al . [25] to demonstrate the additional utility that analyzing variability of gene expression can bring to our understanding of transcriptional control in human embryonic development ( see Fig 2 for an overview of the analysis performed ) . Our analysis approach is composed of the following key steps: ( 1 ) identification of stable genes across development , ( 2 ) detection of stage-specific variability markers ( 3 ) identification of control states based on different levels of expression variability . Our results show that certain genes do have extremely stable inter-cellular expression during development , and that this stability appears to play an important role in regulating the cell . Stable genes are expressed across a range of levels , and those with low expression are enriched for cell type-specific processes whereas those with high expression are involved in fundamentally important cellular processes like maintenance , and metabolism . We also identified genes that could function as markers of stage , as these genes are noticeably expressed with increased homogeneity for cells of a particular developmental stage . Overall , our findings point to variability in gene expression as a regulatory feature of developmental processes in human embryos .
The embryos that were profiled in Yan et al . [25] were donated by women who had already undergone successful in vitro fertilization ( IVF ) , delivered a healthy baby , and elected to donate the remaining high-quality embryos that had been cryopreserved . On average , the women were 30 years old , all had tubal-factor infertility , and all had partners with normal semen parameters . To be included in the profiling study , the embryos underwent extensive screening for good morphology and high quality . Quality assessments were performed using externally-derived , established scoring criteria for viability of the embryos [26] , in conjunction with explicit definitions of embryonic stage ( as outlined in the Materials and Methods section of Yan et al . [25] ) . To the best of our knowledge , we are confident that these embryos were viable . We tested the data to ensure that an appropriate standard of data quality was met for studying gene expression variability . Stringent thresholds were used to discard genes that were expressed below a specific quality threshold in the majority of cells that were profiled ( Materials and Methods , S1 Text ) . To measure inter-cellular expression variability , we adopted a statistic termed the SDC that represents the standard deviation ( SD ) of gene expression between cells of the same embryo that is averaged across multiple embryos ( Materials and Methods , Fig 1A ) . The data were inspected to ensure that patterns of inter-cellular variability were not influenced by the number of cells contributing to each stage ( see S1 Text ) . We also investigated the consistency of the inter-cellular variability measures between the embryos and observed that the transcriptomes of the embryos were relatively stable compared to each other ( see Fig 3B ) . This observation lends support to treating the embryos as replicates to estimate the inter-cellular expression variability in our study ( see S1 Text ) . We also evaluated the possibility that chromosomal copy number mosaicism could affect the variability analysis for those genes expressed on an aneuploid chromosome in a subset of cells . To check for chromosomal copy number effects on expression , we compared the normalized gene expression distribution for every chromosome ( see S2 Text ) and found no obvious aberrations in chromosome-wide expression levels [27] . The analysis of gene expression variability remains an area where guidelines for the best statistical practices are still maturing . For this reason , we investigated the utility of using two statistics , the SDC , based on the SD and the coefficient of variation ( CV ) for studying gene expression variability . The CV is a standardized measure of variability or dispersion that is calculated by taking the ratio of the SD and the average . It is a measure that has previously been used in studies of expression variability for microarray-based data [13 , 14] . One advantage of the CV is that it addresses any potential correlation between the variability and average . Nevertheless , it can be more difficult to interpret what a CV value represents for an individual gene with respect to its average expression and expression variability . Another potential issue is that as a ratio of two statistics , the CV is also in theory subject to zero-inflation for genes that have very low levels of average expression . Such genes will be falsely assigned a higher level of variability as measured by the CV , independent of how dispersed the data for this gene really is . A factor that often arises when studying expression variability is the assumption that the mean and variability of a gene are correlated . Based on our comparison of the CV and SD , the two statistics appear to handle the nature of this correlation quite differently ( see S3 Text ) . We found that the CV had a higher negative correlation with the average expression than the SDC for all four stages . We also conducted simulation studies to test the ability of the two statistics to identify genes with three different levels of variability . We tested the SDC and CV on three different test cases and overall , the SDC appeared to be the better performing statistic with the ability to identify the correct number of variability states more decisively than the CV , and also in general , increased precision in classifying genes to their correct level of variability . We also repeated our analyses of expression variability to compare the results of our main findings when the CV statistic was used . We found that although there was an overlap in the stable genes identified , the two statistics still measure different properties of expression variability . We also noticed that within a specific developmental stage , genes with a low SDC and high average expression , were more likely to have a low CV value , however the converse was not true . Genes with a low CV had both average values and SDC values that spanned a wide spectrum ( in some cases greater than the third quartile for SDC , and less than the first quartile for average expression ) . Therefore , for applications such as marker detection , using CV may be limited as it does not provide the kind of control that comes from setting specific filters on the SDC and average expression directly . Overall , our analyses , which are outlined in more detail in S3 Text , point to the SDC as being the more informative statistic to study gene expression variability for the Yan data set . Focusing on genes whose expression remains invariant amongst all cells across the developmental stages may reveal the subset of core regulators that are integrally important to the developing embryo . Genes expressed at very precise levels in a cell population may correspond to widespread regulators that are ubiquitously switched on to maintain homeostasis , or reflect specialized processes that must be systematically turned off until the embryo is ready . We used Levene’s test to determine those genes with a non-significant change in expression across development ( adjusted P-value > 0 . 05 ) , and clustered the SDC values of these genes using a Normal mixture model to identify the gene clusters with low , medium and high variability . We identified 955 genes whose variability in gene expression varied the least across the four stages ( see Materials and Methods , S1 Table , S1 Fig ) and this group was designated as the group of stable genes . Inspection of the stable genes revealed that they could be classified into three distinct modes of absolute expression; corresponding to low , medium , and high levels of expression ( Fig 4 ) . To understand how these different modes may be contributing to developmental regulation , we identified pathways and functions through Ingenuity Pathway Analysis ( IPA ) that were significantly enriched in each of these three groups . Overall , we observed that stable genes with low levels of expression were involved in pathways regulating specialized cell types ( see S2 Table ) . For all four stages , we saw terms that related to disease processes affecting a specific cell type , e . g . melanoma and chronic leukemia . For specific stages , we also observed enrichment of tissue-specific diseases that affect different organs , such as the kidney ( renal cancer , 4-cell and 8-cell and blastocyst ) , endometrium ( endometrium carcinoma , 4-cell , and blastocyst ) , head and neck ( 8-cell ) , and colon ( blastocyst ) ( see S2 Table ) . For the stable genes with medium levels of expression , 59% were found to be common to all four stages , suggesting that they form a core set of housekeeping-type genes that are expressed in a non-stage specific nature ( S3 Table , see S2 Fig for single cell profiles of some of the stable genes with medium expression ) . This gene set was enriched for fundamentally important pathways , including those controlling transcription ( cleavage and polyadenylation of pre-mRNA ) , translation ( EIF2 signaling , regulation of eIF4 and p70S6K signaling ) , protein ubiquitination , mTOR signaling and DNA damage ( cell cycle: G2/M damage checkpoint regulation ) ( see S3A Table ) . The medium-expressed group of stable genes were also enriched for signaling pathways involving key growth factors and receptors ( e . g . ERK5 signaling , estrogen signaling , and ephrin signaling ) . We observed enrichment for processes related to structural remodeling ( e . g . remodeling of epithelial adherens junctions , epithelial adherens junction signaling , and regulation of actin-based motility by Rho ) . These are important processes for the growing embryo , as critical inter-cellular communication is known to be transmitted via adherens junctions . Enrichment in functional terms obtained from IPA support these themes as well , where enriched key terms include “initiation of translation of mRNA” , “protein and expression of RNA” ( see S3B Table ) . Significant terms related to infection were also observed ( “infection of embryonic cell lines” , “epithelial cell lines” , “viral infection” , “HIV infection” , “infection by HIV-1” , “infection by RNA virus” ) , and these terms are likely to reflect the massive degree of cellular proliferation occurring in the embryo . The stable genes with high levels of average expression were among the smallest subset identified . These genes overlapped significantly with pathways already detected in the stable gene set with medium levels of average expression . Specifically , these genes were enriched for EIF2 signaling , regulation of eIF4 and p70S6K signaling , and mTOR signaling at all stages ( S4 Table ) . For the blastocyst stage , significant enrichment of processes involving cell cycle ( G1/S checkpoint regulation ) and NADH repair was observed . We used the list of stable genes that were expressed at the 4-cell stage to make inferences on which genes may be contributed by the maternal transcriptome . We identified genes that were highly expressed in the oocytes but lowly expressed in human embryonic stem cells ( hESCs ) and intersected this list of genes with the stable genes that had either medium or high average expression at the 4-cell stage embryos . We also identified genes that were likely to be part of the early zygote transcriptome by looking for stable genes that were either expressed at high or medium average levels at the 4-cell stage , but not highly expressed in the oocytes . We found 90 genes that were likely to be contributed by the maternal transcriptome ( S5 Table ) . For the early zygote genes , we identified 626 genes that were active at the 4-cell stage , and 239 genes that were repressed or lowly-expressed ( S6 and S7 Tables ) . One of the genes likely to be contributed by the maternal transcriptome that we identified was DPPA3 , a known maternal factor in the mouse that has a key role in embryonic development . DPPA3 was actually the only gene that was highly expressed at the 4-cell stage ( the remaining 89 genes had medium expression at the 4-cell stage ) . We also showed that these stable , early zygotic genes were significantly enriched in genes that were known to be targets for the human pluripotency transcription factors SOX2 and NANOG ( P-value < 0 . 05 ) but not OCT4 [28] . The analyses and results are further described in S4 Text . While we hypothesize that many of the stably expressed genes serve an important regulatory function for human embryonic development , experimentally validating this hypothesis is a significant challenge . Instead , we can use the genome-wide catalogues that have been collected from studies of human disease genes , or genes that represent essential or robust elements of the genome to computationally infer the functional impact of the stable genes that we have identified , on the cell and embryonic development . We conducted a meta-analysis using seven gene catalogues that represent different aspects of essentiality or perturbation of the human genome . The catalogues that were used were the Genome-wide Association Study Catalog ( GWAS Catalog ) [29] , the Online Mendelian Inheritance in Man ( OMIM ) Gene Map [30] , a set of human orthologs of mouse essential genes [31] , a set of the top 10% most ubiquitously expressed human genes [32] , a set of human haploinsufficient genes [33] , a set of human recessive disease genes that had a DNA copy number variant ( CNV ) deletion [34] , and a set of loss-of-function genetic variants in human protein-coding genes [35] . We found that the stable genes were significantly associated with an enrichment of the essential genes and the top 10% of ubiquitously expressed genes ( two-sided Fisher’s exact test P-value < 0 . 05 , see S5 Text ) . The stable genes were marginally significant in overlap with the haploinsufficient genes and the OMIM Gene Map ( P-value < 0 . 10 ) . We also found a significant depletion of stable genes from the loss-of-function variants and the list of recessive disease genes affected by a CNV deletion ( two-sided Fisher’s exact test P-value < 0 . 05 ) . These results provided some evidence to support the claim that the stable genes had important functional consequences for regulation of the cell and by extension , we assume , the embryo . Stable genes were significantly enriched for genes that were essential or ubiquitously expressed , and less likely to be associated with loss-of-function variants or a known recessive disease gene affected by a CNV deletion . Variability is a statistical property that reflects how the distribution of gene expression in all cells is shrinking or expanding . Genes that have changes in variability at different stages of development may therefore shed light on important stage-specific regulators of the embryo . To identify such genes , we applied the following criteria ( 1 ) a gene must have a statistically significant change in expression variability across the four stages based on Levene’s test [36] , ( 2 ) the minimum level of variable expression for a specific stage ( all other stages had a higher SDC ) , ( 3 ) an average expression level at a specific stage that was higher than all other stages ( see Fig 5B–5D ) . To avoid confusion with markers that are found using average-based approaches that are typically employed , we refer to genes satisfying the three criteria listed , as variability markers . For morulae and blastocyst stages , we identified eight and eleven stage-specific variability markers , respectively ( Table 1 , Fig 5C and 5D ) . These genes share a common theme in functioning to ensure the proper development of the embryo , and they mainly play a role in embryonic development , cell protection , or metabolism . For example , mesoderm development candidate 2 ( MESDC2 ) , a morulae variability marker , functions as a key mesenchymal chaperone protein for Wnt co-receptors [37] ( see S3 Fig ) . For blastocysts , an example is epithelial cell adhesion molecule ( EPCAM ) , which has a critical role in the formation of trophectoderm , the outer layer of the blastocyst that will eventually become the placental interface [38 , 39] . Some of these genes act specifically to protect the cell from free radicals and other toxins , such as blastocyst variability markers peroxiredoxin 6 ( PRDX6 ) [40–42] and glutathione S-transferase pi 1 ( GSTP1 ) [43] . For single cell expression profiles of the variability marker genes , and functional information from the literature , see S3 and S4 Figs for the morula set , and S5 and S6 Figs for the blastocyst set . When considering the functions that are served by the blastocyst marker genes , one theme was metabolism . In order to ascertain that this was not an artefact of the local environment that the human embryos were stored at , we looked at the expression patterns of the blastocyst marker genes in human induced pluripotent stem cell lines ( iPSCs ) and hESCs from three other studies [44–46] as well the hESCs that were profiled by Yan et al . [25] . For these variability marker genes , we saw high levels of expression relative to the global distribution of expression in all cell lines tested , providing some evidence to suggest that the patterns we observed for the blastocyst marker genes were not likely to be solely a product of the embryo’s environment ( see S6 Text ) . We also applied an ANOVA model to determine which variability markers could be identified using a standard approach ( see S7 Text ) and which were specific to our analysis based on gene expression variability . Some of the variability markers had statistically significant P-values from the ANOVA model , and these genes have been highlighted in bold and underlined in Table 1 . The largest degree of overlap was observed for the blastocyst variability markers , where only two genes , LAMTOR1 and RPL17 , were uniquely detected based on our variability-based approach . To demonstrate the functional impact of the stage-specific variability markers , we validated one of the blastocyst variability markers , HDDC2 , by shRNA-mediated knockdown in a human iPSC line . qPCR experiments confirmed that the knockdown of HDDC2 mRNA caused a significant decrease in the expression of key pluripotency markers DNMT3B and NANOG , two genes that are critical for embryonic development and maintenance of pluripotency in iPSCs ( see Fig 6 ) . We also tested the impact of the up-regulation of the HDDC2 locus on hESC differentiation using a hESC line with a stably-integrated inducible CRISPRa/Cas9-VP64 artificial transcriptional activator system . Over-expression of HDDC2 attenuates the drop in expression of the pluripotency marker NANOG and induction of the neuroepithelial marker PAX6 during the early stages of neural differentiation ( see Fig 7 ) . While these experiments cannot provide evidence for the effects of HDDC2 on cell viability or embryonic development , the results from the HDDC2 knockdown suggest that this gene plays a role in the maintenance of pluripotency . From the transcriptional effects caused by HDDC2 over-expression on early neural differentiation , we can infer that HDDC2 is able to either reinforce the persistence of the pluripotent phenotype , or is involved with specific interference of the neural differentiation process , or both . In early developmental processes , evidence supports the role of gene expression variability as a necessary initiating step that precedes lineage commitment in progenitor cells [8 , 47 , 48] . As embryos transition from the 4-cell to the blastocyst stage , the global distribution of inter-cellular variability shifts from lower to higher values ( Fig 3A ) . Based on the shape of the density distribution , we see an increase in the overall number of genes that are expressed more heterogeneously between cells as the embryo differentiates . This result likely reflects the diversification of transcriptomes observed in cells that are undergoing the necessary cell fate changes to become distinct lineages and specialized cell types [49] but which also contain inherent stochasticity [50] . The shape of the inter-cellular variability distribution demonstrates how at each developmental stage , the transcriptome is expressed with levels of precision that span a wide spectrum ( Fig 3A ) . This is unsurprising given that during the 4-cell to the 8-cell stage , the embryo undergoes activation of the zygotic genome ( ZGA ) and a massive reprogramming of the transcriptome occurs [51 , 52] . The stochastic nature of reprogramming has been observed in stem cells as well as the existence of alternative stem-cell states , and hence we expected to see higher levels of SDC [53] . Analyzing expression variability gives us a window into which components of the transcriptome are being used in cells of the developing embryo at different levels of precision . We are interested therefore in characterizing how precision changes as the embryo develops , and which genes are involved at either end of the regulatory control spectrum . We refer to these different levels of expression variability as control states . Mixture models were applied to identify within the data , the number of control states present at each stage and classify the subsets of genes observed under different levels of regulatory control ( see Fig 8 , Materials and Methods ) . The 4-cell stage had the largest number of control states , where each state could be interpreted as one of four distinct levels of variable expression ( low , medium , high , very high ) ( Fig 8A ) . During development , we observed that these levels collapse down into simpler states , e . g . for the morula stage , there are three levels ( see Fig 8C ) and for the blastocyst stage , there are only two ( Fig 8D ) . We investigated whether genes with the same level of regulatory control were enriched for certain pathways or processes using IPA software . Much like the variability markers , we observed consistent changes in pathways that appear to protect and ensure adequate gene regulation in the developing embryo ( S8A–S8D Table ) . For example , four pathways were enriched for low-variability ( most stable , mixture 1 ) genes , EIF2 signaling , regulation of eIF4 and p70S6K , mTOR signaling , and protein ubiquitination . This overlap was significant for all four developmental stages , and represents pathways that control important functions that are required by every living cell ( for an overview of the important enriched terms occurring per stage , see S7 Fig ) . For the 4-cell stage only , the low-variability genes were enriched for the “DNA Methylation and Transcriptional Repression Signaling pathway” ( S8A Table ) . The enriched genes ( SIN3A , SAP18 , SAP130 , RBBP7 , RBBP4 , CHD4 ) were involved in regulation of histone deacetylation , and DNA methylation ( DNMT1 ) . Proper regulation of histone deacetylation is known to be a requirement for embryonic development [54] . DNMT1 is involved in maintaining single-stranded methylation of newly replicated DNA , and its expression is activated by cell cycle-dependent transcription factors in the S phase [55] . We also observed enrichment of a key developmental pathway “Ephrin B signaling” for the 4-cell low-variability group but at no other stage . Genes enriching this pathway have been directly implicated in fish and mammalian embryo development , including Rho-associated , coiled-coil containing protein kinase 2 ( ROCK2 ) , beta catenin ( CTNNB1 ) and ras homolog family member A ( RHOA ) [56–58] . Pathways associated with telomere extension by telomerase , and telomere signaling were statistically significant amongst variably-expressed genes at the 4-cell and the blastocyst stages ( mixtures 3 and 2 respectively , S8A and S8D Table ) . The importance of telomerase to preimplantation embryos has previously been established via its link to reproductive potential [59] . At the 4-cell stage , telomeres are thought to reset during genome activation , which may explain the higher number of telomerase-associated genes that have variable expression at this stage . The six genes that were annotated to the telomerase extension pathway by IPA ( TERF2 , TNKS , HNRNPA2B1 , TERF2IP , TNKS2 , POT1 , XRCC5 , RAD50 , and NBN ) have roles in double-strand DNA break repair and damage response , where they are responsible for protecting the telomere extension process . Telomere DNA elongation has been observed to occur between the 8-cell and the blastocyst stages in both animal [60 , 61] and human embryo studies [62] . It is likely that this corresponds to the second enrichment of variable genes observed at the blastocyst stage . Although successful telomerase elongation is critical to the embryo , variability in expression of genes associated with this process may be due to the slightly different rates at which cells are using and completing this process . The transition from the morula to the blastocyst features the appearance of the first cell type specification with the organization of the trophectoderm . Integrin binding and activation is an essential element of blastocyst implantation and trophoblast differentiation [63–65] . Integrins are a class of heterodimeric transmembrane cell surface receptors that participate in cell-cell interactions and regulate signals for cell adhesion , growth and survival . We found that genes in the IPA pathway “Integrin Signaling” first appeared in the gene clusters for the morula stage , and continued to be significantly enriched in the blastocyst stage , possibly indicating the variable process of cell differentiation within the morula before the visible segregation of the trophectoderm ( S8C and S8D Table ) . For the blastocyst stage , we find that low-variability genes were enriched for metabolic pathways , such as oxidative phosphorylation , oxidative stress , and glycolysis-related pathways ( S8D Table ) . In fact , glycolysis and gluconeogenesis pathways are uniquely observed to occur for low-variability genes in the blastocyst stage . This may indicate the appearance of a more diverse regulatory program in blastocyst metabolism . Blastocysts are known to require a higher metabolic load due to the burst in developmental change that occurs at this stage . For example , the cytosolic form of malate dehydrogenase ( MDH1 ) , which is featured in both oxidative phosphorylation and gluconeogenesis pathways has been shown to be important for blastocyst development in mice [66] . On the other hand , by the blastocyst stage , because of the emergence of distinct lineages and cell types , cells across the embryo may begin to show signs of asynchronicity in cell cycling . We see this effect reflected in the statistically significant enrichment of cell-cycle related processes in variably-expressed genes at the blastocyst stage . Dependence on specific developmental and signal transduction pathways is also expected to be more heterogeneous at this stage owing to the distinct cell types that have different expression programs . Variably-expressed genes at the blastocyst stage were enriched for several signaling pathways , including ATM signaling , PI3K/AKT signaling , JAK/Stat signaling , ERK/MAPK signaling , and others ( S8D Table ) . While our analysis so far has concentrated on identifying the gene-specific measures that vary in a cell population ( see Fig 9A ) , it is worthwhile to highlight those cells that are the most aberrant or consistent in the population , with respect to global gene expression . Sampling a fixed number of cells and analyzing patterns of variability allowed us to look at the overall heterogeneity of the cell population based on the global gene expression variability distributions . Inspection of the density plots obtained for each 4-cell sample revealed interesting properties that provide insight into the transcriptional noise occurring at a specific development stage ( Fig 9B ) . Multiple draws of cells from the population can highlight how good of a representation any single cell sampled at random is likely to be . For situations where it is not possible to sample every cell in an organism or tissue , this kind of analysis gives us the ability to assess which parts of the transcriptome are more generalizable and stable ( Fig 9B ) , and conversely , which cells may be unusual or aberrant . For the 8-cell stage , we observed that all 4-cell combinations produce overlapping densities and suggest that most cells are relatively similar to each other in the embryo at this stage of development with respect to global variability ( Fig 9C ) . The shape of the densities also indicate that most of the transcriptome is expressed with lower variability . For the morulae stage , it is clear that there are three dominant sub-clusters of 4-cell combinations and that this stage has the greatest degree of cellular diversity . For the blastocyst stage , this diversity appears to subside , and cells are more homogeneous with respect to their overall variability profile . The density for the blastocyst combinations suggest that although there is reduced cellular heterogeneity compared to the morulae stage , there are more genes expressed at higher levels of inter-cellular variability . These densities may reflect a transition signature of the morula cells . In blastocysts , there are at least two major tissue types ( trophectoderm , inner cell mass ) whereas the morulae are still in an undifferentiated state where cells are only starting to transition into two different types . The blastocyst , on the other hand , is where we see the first differentiation events occurring . One limitation of the human embryo data set was that it featured relatively small numbers of embryos , and hence to test the robustness of our results , we looked to two separate studies that were performed in mouse as a source of data to validate our key findings ( see S8 Text ) . The first study from Guo et al . [7] generated gene expression data for 48 genes in five to seven embryos per stage using the Fluidigm Biomark System 48 . 48 Dynamic Arrays . We selected the data from the 4-cell , 8-cell , 16-cell , 32-cell and 64-cell stages as these paralleled the most closely with the human developmental stages that were used in our analysis of the human embryos . Using the Guo data set we were able to verify the existence of a subset of stably expressed genes that had invariant expression across the development of the embryo , and these genes showed low , medium and high levels of average expression . We also detected sets of variability markers that displayed changes in average expression and expression variability that appeared to delineate a specific stage of development . Although only 48 genes were profiled in this data set , we saw that the distribution of expression variability for the total set of genes adopts wider values as the mouse embryos developed . Even with a higher number of mouse embryos , the distribution of inter-embryo variability remained relatively constant as we had observed with the human data . In the second mouse study , Deng et al . [67] applied Smart-seq and Smart-seq2 single cell RNA-sequencing to profile the transcriptomes of cells taken from early stage embryos . Their experimental design was similar to that of the Yan human embryo data , but with a higher number of embryos and cells . We used the 4-cell , 8-cell , 16-cell and late blastocyst stages from their study to compare our results . Our analysis of the Deng data set also yielded results that aligned with those observed from our study of the human embryos . Moreover , because the number of genes in both data sets was more similar , we were able to compare the overlap between the lists of stable and variable genes that were identified in the mouse and human embryos . The overlap was statistically significant ( two-sided Fisher’s exact test P-value < 10−23 , odds ratio estimate 3 . 744 ) , even after permutation tests ( see S8 Text ) . We tested the gene lists from the human-mouse comparisons for enrichment of biological pathways and processes using the MSigDB database [68] ( based on Hallmark gene set terms and C4 computational gene set terms ) to further investigate the nature of both the overlap between the human and mouse stable genes , but also the species-specific differences . Genes that were stable in both human and mouse embryos , were enriched most strongly for MYC targets and housekeeping genes . We found that the discordant genes ( either stable in mouse but variable in human , or vice versa ) were enriched for more disease-specific terms and processes . Two terms that were unique to the list of genes variable in human but stable in mouse , reflected gene sets that regulate pluripotency ( the Wong Embryonic Stem Cell Core [69] and the Mueller PluriNet protein-protein interaction network [70] ) . Based on these results , genes that are discordant with respect to expression variability may play a role in human-specific ( or mouse-specific ) diseases , or be required for species-specific embryonic development and maintenance of pluripotency .
We have taken a novel approach to uncovering how the transcriptome is regulated in human embryos . The use of technologies enabling analysis with a single cell resolution enables an embryo to be modeled as a defined population of cells , and by studying variability between these cells directly , we can identify genes whose expression is either homogeneously or heterogeneously distributed in the cell population . Our analyses have highlighted key developmental pathways that show different degrees of regulatory control , and we have demonstrated that analyses of expression variability can provide functional insights into stage-specific markers . It is important to emphasize that all our inferences on regulatory control of the transcriptome are based on data collected from human embryos . This has an impact on improving our understanding of transcriptional regulation in human developmental processes . A recent study has revealed that even for mice , a species that is often believed to be a good model of human biology , the differences in transcriptomes from humans are larger than expected [71] . In embryology , there are many technical , ethical and scientific limitations that understandably make the use of non-human embryos more feasible for—omic studies; however , it is useful to keep in mind that the results of our study provide valuable insights into how the transcriptome is controlled and used during the development of an actual human embryo . The greater diversity of control states observed for the 4-cell stage is likely to be a result of the transcriptome being in a continued state of flux and reorganization following zygotic genome activation ( ZGA ) . As the embryo transitions away from the 4-cell stage , plasticity in global gene expression gives way to the commencement of the more specific differentiation programs . This phenomenon may overlap with what other studies have described as “waves of transcriptional activation” [52] . We see genes in the blastocyst stage segregate to one of two different control states ( see Fig 8D ) , and from this , we speculate that control of the transcriptome has polarized into low and high levels of precision , with fewer intermediately-variable genes observed than in earlier developmental stages ( see Fig 8A and 8B ) . The decreased complexity of regulatory control observed across development is also consistent with our model of how ZGA is influenced by the number of completed cell cycles . Our general understanding is that repressors and activators have opposing activity during ZGA that changes as cells of the embryo transition through successive cell cycles [72] . During early embryogenesis , the expression activity of maternal repressors is high at first but as these cells divide further their concentration is diluted in the embryo and therefore this activity eventually decreases . Activators , on the other hand , are initially present at low levels and may require successive cell cycles before attaining a threshold level for successful activation . The interplay between activators and repressors gives rise to the resulting levels of mRNA copies , and associated degree of heterogeneity observed in the embryo as it develops . We identified genes with stable expression for the 4-cell to blastocyst stages , and these may form the core set of genes required for embryonic development to occur successfully . The stable genes were expressed at one of three distinct modes that spanned low , medium and high levels of expression , and this is interesting to note because it suggests that invariance of expression may be part of regulating genes required at all expression levels , and is not only confined to genes that are highly activated or need to be silenced . The assumption that variability diminishes inversely with the absolute level of transcription is not consistent with our observations . The first two modes of expression ( low and medium ) could be interpreted as genes that are turned off and on , respectively . Stable genes with low expression may be the result of a gene that has been switched off and their non-zero expression is the result of background signal observed in the RNA-seq experiment . Alternatively , these genes may be weakly expressed due to other mechanisms , e . g . from leaky transcription , where initiation of transcription occurs but not enough of the necessary components from the remaining machinery are available to follow through . The third expression mode ( stable genes with high expression ) may operate as accelerators or as higher level activators that support or function in roles related to the stable genes with medium expression . The fact that the pathways with significant enrichment for these genes overlap with those enriched for the medium-expressed ( Fig 4 ) , stable genes supports this interpretation . Identifying genes based on their lack of variability has some parallel to developing catalogues of housekeeping genes , a concept that has been a fixture in molecular biology for nearly fifty years now [73] . While housekeeping genes represent a practical use in standardization and quality control , they have also garnered interest from a systems biology perspective because they represent a core , minimal set of genes necessary to support the cell . With growing technology , we continue to revise our definition of housekeeping genes; however , as far as we are aware , none of these approaches evaluate changes in variability directly . Our approach therefore has cross-over utility for the purpose of identifying potentially new housekeeping-type genes in single cells . Similarly , our approach may also have utility for defining stage-specific control genes that can be used to normalize or calibrate expression data of other genes with more heterogeneous expression . We compared the stable genes that showed low , medium and high expression across all stages ( see Fig 4 and S1 Table ) with two sources of housekeeping genes , the top 10% of ubiquitously expressed genes collected by de Jonge et al . [32] , and a list of housekeeping genes for RNA-seq data collected by Eisenberg et al . [73] . For both sets of housekeeping genes , we saw considerable overlap between the stable genes that had medium or high levels of expression compared to very few lowly-expressed stable genes ( S9 and S10 Tables ) . Across species , the property that divergence occurs at early and late , but not the middle part of development trajectories is a phenomenon referred to as the hourglass effect [74] . Other studies have shown that developmentally important genes demonstrate remarkable sequence similarity across a wide range of organisms , pointing to shared mechanisms that have been evolutionarily conserved . It has also been shown that gene expression control follows a similar pattern of conservation or stability , and experimental evidence to that effect has been most recently reported by Gerstein et al . [75] . Using transcriptomes of distant species generated by the ENCODE and modENCODE consortia , genes exhibiting developmental stage-specific expression showed diminished inter-species expression variability in the middle stages of development but were heterogeneous before and after . A recent paper by Liu et al . [76] also analyzed the same human embryo data set generated by Yan et al . [25] where they built co-expression gene modules and performed evolutionary conservation analysis on gene sequences as well as transcriptional regions upstream of each gene . They concluded that genes from stage-specific co-expressed modules have increased selective pressure at the 8-cell stage versus earlier stages ( zygote to the 4-cell stage ) and later ( 8-cell to the late blastocyst stage ) . In our study , the variability marker genes that were identified for the 8-cell and the morula stages also follow a pattern of diminished variability at these intermediate development stages ( see Fig 5B–5D ) . This observation corresponds to Liu et al . ’s discovery of increased selective pressure in co-expressed module genes at the 8-cell stage . Higher evolutionary pressure represents an increase of evolutionary constraint . Both independently derived observations point to an ‘hourglass’ in the preimplantation embryo , with lower evolutionary constraint and higher variability flanking a period of higher evolutionary constraint and lower variability . The stage-specific variability markers we have identified show promise as potential determinants of embryo stage , however , they also shed light on understanding embryogenesis . It is likely that these markers are consistently expressed for all cells of the embryo because they provide or support critical functions for the developing embryo . This hypothesis could be readily tested in the context of model organisms like zebrafish and Drosophila . For human cells , these markers may be useful in future studies for identifying healthy embryos based on transcriptional states . Some of these applications may be in laboratory embryology where analysis of the least variable components of the transcriptome may serve as a scaffold to create predictive signatures of the developmental stage and potential of a biopsied embryo’s cells , especially for expression effects in genetically modified embryos . In conjunction with other existing criteria , information based on expression variability may eventually be used as a screen during routine day 3–5 biopsy for preimplantation embryos in IVF and other applications to improve pregnancy rates of women using these procedures . We also see potential clinical applications in genetic diagnosis for reproductive potential for human embryo pre-implantation . Embryos that are otherwise chromosomally healthy are known to have a significant risk of implantation failure [77] . While there are many reasons underlying implantation failure , one significant cause may be due to innate errors in development due to disordered regulatory programs in the developing embryos that may result from genetic , epigenetic and environmental factors . As trophectoderm biopsies taken on day 5 ( blastocysts ) are regularly used as material for preimplantation genetic diagnosis ( PGD ) , part of the biopsy material could be used to isolate cells and test the transcriptional regulatory state of the embryo in the trophectoderm for both expression level and SDC . This effect could be studied in a clinical trial through genetic fingerprinting analysis , determining the origin of a single birth following a double embryo transfer into the same patient . The expression state of cells in the implanting embryos versus the non-implanting embryos can be quantified and compared among a cohort , including the SDC of genes that represent important biomarkers for embryo reproductive potential . There is value to understanding the variability of genes as a marker of cell state and regulatory potential [19] . Other variability markers may also unlock intriguing clues into how blastocysts are regulated ( see S5 and S6 Figs ) . NDUFA12 helps to generate ATP for the cell via oxidative phosphorylation in mitochondria [78] . Mutations in this gene are often the cause of mitochondrial oxidative phosphorylation diseases , suggesting that NDUFA12 is critical for maintaining healthy human cells , and for blastocysts , NDUFA12 may play an important role in supplying these cells with sufficient energy [79] . LAMTOR1 is a membrane protein whose function is currently only known in the context of late endosomes and lysosomes , which are organelles that the cell relies upon to perform waste disposal . LAMTOR1 is anchored to the surface of these organelles , and through other intermediates can activate the mTORC1/MAPK signaling pathways that lead to cell growth and control of energy homeostasis [80] . At this point , it is difficult to say what impact LAMTOR1 has on blastocysts; however , a mouse study has shown that LAMTOR1 is responsible for proper epidermal development by regulation of lysosome-mediated catabolism [81] . It is possible that these processes may affect other tissues in the blastocysts more generally , and hence we see this gene being flagged as a variability marker . Another marker of interest , ACTN4 is part of the α-actinin family , a set of related proteins that bind a wide range of molecules , including actin , to regulate a host of important processes in the cell [82] . α-actinin appears to operate in cell type-specific roles in the human body . Specifically , ACTN4 has been found to be ubiquitously expressed and distinct from the other α-actinins in that it interacts with transcription factors [83] . While its connection to blastocysts is also unclear at this point in time , ACTN4 has been detected in higher levels in cells with higher motility [84] and may be linked to metastatic processes in cancer [85 , 86] . Motility in cells of the blastocyst may be important for blastocyst activation and trophectoderm motility [87] as previously observed in mouse embryos . We observed that some of the variability markers that were identified using our approach overlapped with markers that can be found with standard approaches such as an ANOVA model . One example was EPCAM , a blastocyst variability marker identified also by ANOVA ( see S5 and S6 Figs ) . EPCAM acts as a cell surface protein that is often used as a stem cell marker [88] because of its association with elevated levels in undifferentiated human embryonic stem cells [89] . Depending on the cell type that the expression profile was captured in , and other biological factors , there may be genes showing distinct changes in variability and/or average expression and both statistics are informative . Decreases in variability may be indicative of increased criticality , whereas increases in average expression may suggest a general trend of the cell population . While transcription is fundamentally a stochastic phenomenon that affects all genes generically , the observation that some genes are more variably expressed than others , points to the existence of physical mechanisms underlying gene expression variability . These mechanisms can be grouped broadly into two major sources of variation . The first source refers to mechanisms embedded in the genome sequence , where the content of the promoter region and other regulatory sequences affects the degree of expression variability of a downstream gene . Nucleosome positioning , for example , is one such mechanism that can generate variable gene expression where a nucleosome typically occupies a region where transcription start sites and TATA elements exist [90] . Another example involves the interaction of long-range gene regulation where the expression of some genes require the interaction of other chromosomal segments . This mechanism can generate increased expression variability for genes that are regulated by this kind of nuclear architecture [91] . Other epigenetic mechanisms such as the presence of chromatin modifications have also been attributed to expression variability [92] . The second source of variation refers to the networks or interactions that a gene participates in that can influence its expression variability . We have seen through both experimental and simulation studies , that certain circuit formations , such as auto-regulatory motifs and feedback loops , can propagate signals to create different degrees of noise or variability [15 , 16] . Pathways can be viewed as a collection of integrated circuits of genes , and the makeup of these circuits have been refined through evolutionary processes to become highly gene-specific and well-tuned . Therefore , depending on the placement of genes in their specific circuits , they may be more highly sensitive to exhibit fluctuations in their expression [93] . Related to this is the interaction with other molecules that have repressive or activator effects on a mRNA signal . For example , the interaction with microRNAs and shRNAs in the cell may deplete transcript levels of a gene however due to concentration , spatial regulation and other intracellular effects , this repression is likely to be observed non-uniformly in a cell population .
We used the RPKM ( reads per kilobases per million reads ) normalized RNA-seq data set generated by Yan et al . [25] . The data was downloaded from GEO using the accession number GSE36552 . Whole-genome RNA-seq data from three embryos were available for the four embryo stages used in our study ( 4-cell , 8-cell , morula , blastocyst ) except for morula where there were only two embryos . For morula and blastocyst , eight and ten cells were profiled respectively . In order to eliminate genes that may be affected by poor quality , genes with an RPKM-expression value ≥ 0 . 1 in at least 75% of cells from the same stage were retained . This resulted in 8105 genes that were used for further analysis . To measure the inter-cellular expression variability of each gene g , we adopted the following statistic SDCg: SDCg=1E∑j=1E1Nj∑i=1Nj ( xij−x¯j ) 2 where xij is the expression level of gene g in the i-th cell of the j-th embryo for a total number of E embryos . For each embryo j , there are a total of Nj cells that have been profiled . x-j represents the average expression occurring in the j-th embryo . The SDC captures the SD of expression levels observed between cells belonging to the same embryo , that is then averaged across all embryos to give an overall measure of inter-cellular variability . We make the assumption that the embryos represent biological replicates , and therefore genes showing consistent inter-cellular variability between embryos are given higher weight by the SDC . This assumption is reasonable to make given that the embryos were screened extensively and only those passing stringent morphological criteria were included for RNA-sequencing ( Fig 1 ) . As a contrast to the SDC , and as a means to evaluate inter-embryo variability of expression , we also computed the SDEg: SDEg=1E∑j=1E[1Nj∑i=1Nj ( xij−x¯j ) 2−1E∑j=1E1Nj∑i=1Nj ( xij−x¯j ) 2]2 . All analysis and code is available upon email request . For each stage , the SDC values were clustered using Normal mixture models to identify how many variability states were present . We used the Mclust function in the mclust R package ( version 5 . 0 . 1 ) where the number of states was inferred by the function , and selected from possible values ranging from 1 to 9 . We used the Ingenuity Pathway Analysis database ( Summer Release 2014 ) to identify significantly enriched pathways and processes for genes in different variability states at each stage . Significance was determined using a Fisher’s exact P-value that was adjusted for multiple testing correction using the Benjamini-Hochberg method [94] . Cut-offs were applied and are specified in the corresponding table legends . We applied three simultaneous filters to identify stage-specific variability markers . To qualify as a marker of stage X , a gene had to have ( 1 ) the smallest SDC in stage X compared to all other stages; ( 2 ) a higher average expression in stage X compared to all other stages; ( 3 ) a statistically significant change in variability as detected by Levene’s test ( adjusted P-values < 0 . 05 ) . For the knockdown experiments , a transgene-free human iPS cell line ( clone C11 [95] ) was transduced with lentiviral particles delivering pLKO . 1-Puro family vectors constitutively expressing shRNAs against either control ( EGFP ) or HDDC2 ( x2 ) mRNAs ( see S11–S13 Tables ) . Puromycin selection ( 2μg/mL in the culture medium ) was applied at days 4–6 after transduction , after which RNA was harvested for qPCR analysis . For inducible up-regulation experiments , a human ES cell line inducibly expressing Cas9-VP64 , an artificial transcriptional activator , was transduced with lentiviruses allowing for a puromycin-selectable expression of chimeric gRNAs targeting either unrelated gene or HDDC2 . 2 days after the start of puromycin selection commencing at day 4 after transduction , induction of gene activation was started by addition of doxycycline at 1μg/mL . 2 days after that neuronal differentiation was started using a standard dual Smad inhibition protocol ( on days 0 and 2 of differentiation , KOSR medium without FGF but supplemented with 5μM SB431542 , 10 5μM Dorsomorphin and doxycycline at 2μg/ml ) . RNA was harvested for qPCR analysis on day 3 . We first identified genes that failed to meet statistical significance for Levene’s test ( adjusted P-value > 0 . 05 ) . These genes were clustered according to their SDC values into three groups that corresponded to low , medium and high levels of variability . We used the hclust function to perform agglomerative hierarchical clustering . Functional enrichment analysis was performed on the low variability group using IPA software . The set of stable genes were taken to be the cluster with the lowest level of variability ( Cluster 1 , S1 Fig ) . All possible 4-cell combinations were elucidated for each stage , and the variability profile constructed for each combination .
|
In order to function properly , cells express specific sets of genes that are regulated via complex transcriptional programs . During early stages of development , when an embryo consists of only a handful of cells , it is vital that these cells work together so that the embryo can develop into a healthy baby . Single cell studies allow us to understand how each cell contributes to ensuring proper regulation of the embryo , as well as identify the critical genes whose expression is important for development . While we understand that regulation of a gene occurs through the timing of when it is expressed and also the quantity of its expression , more recently we have come to recognize that the variability of a gene’s expression across single cells may also contribute to the viability of the organism . In this study , we analyzed the gene expression variability of human embryos at different developmental stages . We discovered distinctive patterns of variability across cells in the embryo; some genes had extremely stable expression , and others were variable but with increased homogeneity in expression at a particular developmental stage . We validated one of these stage-specific markers and found that it played a role in the maintenance of pluripotency of human pluripotent stem cells . Overall , these results can help unlock additional clues into understanding how embryonic development is regulated in humans .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Variability of Gene Expression Identifies Transcriptional Regulators of Early Human Embryonic Development
|
A key priority in infectious disease research is to understand the ecological and evolutionary drivers of viral diseases from data on disease incidence as well as viral genetic and antigenic variation . We propose using a simulation-based , Bayesian method known as Approximate Bayesian Computation ( ABC ) to fit and assess phylodynamic models that simulate pathogen evolution and ecology against summaries of these data . We illustrate the versatility of the method by analyzing two spatial models describing the phylodynamics of interpandemic human influenza virus subtype A ( H3N2 ) . The first model captures antigenic drift phenomenologically with continuously waning immunity , and the second epochal evolution model describes the replacement of major , relatively long-lived antigenic clusters . Combining features of long-term surveillance data from the Netherlands with features of influenza A ( H3N2 ) hemagglutinin gene sequences sampled in northern Europe , key phylodynamic parameters can be estimated with ABC . Goodness-of-fit analyses reveal that the irregularity in interannual incidence and H3N2's ladder-like hemagglutinin phylogeny are quantitatively only reproduced under the epochal evolution model within a spatial context . However , the concomitant incidence dynamics result in a very large reproductive number and are not consistent with empirical estimates of H3N2's population level attack rate . These results demonstrate that the interactions between the evolutionary and ecological processes impose multiple quantitative constraints on the phylodynamic trajectories of influenza A ( H3N2 ) , so that sequence and surveillance data can be used synergistically . ABC , one of several data synthesis approaches , can easily interface a broad class of phylodynamic models with various types of data but requires careful calibration of the summaries and tolerance parameters .
Many infectious pathogens , most notably RNA viruses , evolve on the same time scale as their ecological dynamics [1] . One of the perhaps best documented examples are human influenza A viruses , which cause substantial morbidity and mortality as they escape host immunity predominantly through the evolution of their surface antigens [2] . The resulting , dynamical interaction between the ecological and evolutionary processess can be better understood through the formulation and simulation of so-called “phylodynamic” mathematical models , e . g . [3]–[8] . However , while data on disease incidence as well as viral genetic and antigenic variation are increasing for many viruses , e . g . [9]–[13] , fitting and assessing phylodynamic models to these data is still not commonly done . Historically , epidemiological time series data have been pervasively used to analyze hypotheses of host-pathogen interactions at the population level [14]–[17] . However , time series data capture the underlying evolutionary processes of pathogens only very indirectly . For flu , this has limited the type of infectious disease models that can be statistically interfaced with time series data , and the number of epidemiological parameters that can be simultaneously estimated [18] , [19] . Consequently , the disease behavior of rapidly evolving pathogens is increasingly studied under additional , complementary data sets [1] , most typically in ways that attempt to qualitatively reproduce prominent disease attributes [3]–[8] . More recently , coalescent-based statistical methods have been used to elucidate the disease dynamics of RNA viruses from molecular genetic data alone [20] . These methods have been particularly useful to reconstruct epidemiological transmission histories , identifying when and where transmission occurred and how viral populations change over time . For example , coalescent-based analyses have highlighted the importance of the tropics in the complex circulation dynamics of human influenza A ( H3N2 ) virus ( in short: H3N2 ) [9] , [21] , [22] . However , most coalescent methods estimate past population dynamics within a class of flexible demographic functions including exponential and logistic growth as well as the nonparametric Bayesian skyride [23] , [24]; but see also [25] . These demographic functions do not explicitly describe the non-linear population dynamics of RNA viruses . Thus , assessing which ecological interactions underlie observed patterns of sequence diversity , and estimating the respective strength of these interactions , is difficult within this framework . Because of these limitations , we adopt a different statistical approach known as Approximate Bayesian Computation ( ABC ) to infer the phylodynamics of RNA viruses . ABC allows mechanistic phylodynamic models to be simultaneously fitted against both sequence and surveillance data . This method circumvents explicit likelihood calculations by simulating instead from the stochastic model that defines the likelihood [26] . Recent extensions of ABC allow for model assessment to be carried out at no further computational cost [27] . We further suggest incorporating variable selection procedures to quantify if and to what extent the data provide support for the inclusion of specific model components [28] . To demonstrate the utility of our approach , we consider the phylodynamics of interpandemic H3N2 . We obtained weekly reports of H3N2 incidence in the Netherlands from 1994–2009 by combining influenza-like-illness ( ILI ) surveillance data with detailed records of associated , laboratory-confirmed cases of flu by type and subtype [29] , [30] , and similarly for France and the USA; see Figure 1 and the supplementary online material ( Text S1 ) . In addition , we reconstructed the ladder-like phylogeny of H3N2's haemagglutinin gene ( HA ) from dated European sequences collected in 1968–2009 ( see Figure 1 and Text S1 ) . To represent H3N2's global phylodynamics , we focus on a class of spatially structured phylodynamic compartmental models that formalize probabilistically how evolving , antigenic variants interact epidemiologically . These antigenic variants might correspond to the major antigenic clusters that are distinguishable in H3N2 antigenic maps [31] , but can in principle also represent a different phenotypic resolution . The evolutionary dynamics of viral genotypes are separately formulated for each antigenic phenotype because genetic distances do not necessarily easily translate into phenotypic relationships [5] . Spatial substructure has been incorporated in several models of H3N2 phylodynamics to reflect the global circulation of the virus [4] , [8] , [32] . We adopt here a simple source-sink framework , where the sink is thought of as the Netherlands into which viral genetic diversity and antigenic strains are imported on a seasonal scale from a source population where the virus persists [9] , [33] . We fit and assess two distinct models to the combined features of sequence and incidence data described in Figure 1 and Table 1 . The first model captures H3N2's antigenic drift phenomenologically through gradual loss of immunity , and the second model describes the antigenic evolution of the virus explicitly with particular assumptions on the tempo of antigenic change .
To perform phylodynamic inference and goodness-of-fit analyses for complex phylodynamic models , we adopt a simulation-based approach that has become known as Approximate Bayesian Computation ( ABC ) [26] . Our first goal is to estimate the posterior density ( 1 ) of epidemiological and evolutionary model parameters under approximations to the likelihood of observed population incidence and phylogenetic data . The prior density can be used to incorporate existing information or limit the range of plausible values of model parameters . Our second goal is to assess fitted phylodynamic models based on a recent extension of ABC [27] . ABC methods circumvent computations of the likelihood by comparing the observed data to simulated data in terms of many , lower-dimensional summary statistics , , such as those in Figure 1 . Using a distance function that compares summaries , each simulation is weighted according to the magnitude of the summary error under a weighting scheme , and this value is used in place of the likelihood term in Monte Carlo algorithms . In essence , ABC is a particular auxiliary variable Monte Carlo method , where the summary errors take on the role of auxiliary variables . Integrating these errors out , the ABC likelihood approximation adopted here is ( 2 ) where the weighting scheme is typically the Indicator ( 3 ) with tolerance parameter or the Exponential ( 4 ) with . Intuitively , the summary errors indicate how well a parameterized model reproduces the observed data . Once Monte Carlo algorithms such as the Markov Chain Monte Carlo ( MCMC ) sampler proposed by Marjoram et al . [34] have converged , the magnitude of the summary errors can be used to diagnose goodness-of-fit with respect to each of the summaries . To use this detailed information on each summary , we prefer using ( 2 ) to the Mahanalobis approximation ( see [26] ) . Although uncommon , we typically use the log ratio so that the errors can be uniformly interpreted as fold-deviations . Parameter inference using ABC is approximate in that the ABC target density approaches the posterior density ( 1 ) as tends to zero if the summaries are sufficient for [26] . We use a Monte Carlo algorithm that is very similar to the MCMC sampler in Figure 2 . A full specification of the algorithm is given in Text S1 . It is typically difficult to establish the sufficiency of phylodynamic summaries analytically , and instead a small set of summaries is chosen such that model parameters of interest can be estimated [26] . Table 1 lists basic features of H3N2 epidemiological and phylogenetic data that were primarily considered in this study . Phylodynamic models were fitted and assessed against the features of the Dutch incidence data and the viral phylogeny derived under the Exponential clock model . The differences between these summaries and those derived from the remaining data in Figure 1 were used to set the ABC tolerances large enough so that inference is robust to the choice of phylogenetic reconstruction method and reporting country . Although smaller tolerances can be computationally feasible , these were not supported by the additional data considered . We typically use the Indicator weighting scheme ( 3 ) with tolerances that encompass differences in summary values across reporting countries and/or reconstruction methods , see Table 1 . When a model never fits a particular summary well , we use ( 4 ) to give a mild prior preference to small errors [27] . See Text S1 for further details . A frequent problem in phylodynamic modeling is to determine if a specific model parameter should be included . For example , it can be unclear which types of ecological interactions between antigenic variants underlie pathogen phylodynamics , or if the residual selection parameter in ( 5c–5d ) is required in addition to a given ecological mechanism that induces immune selection . Following existing variable selection procedures [28] , we use an additional indicator variable to denote whether a single model parameter is present ( ) or absent ( ) and estimate its posterior probability under equation ( 2 ) . Here , we use a standard spike-and-slab variable selection procedure [28] to estimate inclusion probabilities of the residual selection parameter .
To illustrate ABC methodology with the summaries in Table 1 , we begin with a classical phenomenological model that implicitly accounts for antigenic drift through gradual loss of immunity [37] . H3N2 phylodynamics are represented with a spatial two-tier system of equations that is a special case of ( 5 ) when the antigenic emergence rate is set to ( 6 ) For simplicity , we will refer to ( 5 ) without antigenic variants as the SEIRS model . While several models have been able to simulate phylodynamics that are consistent with some aspects of the observed data , most notably the ladder-like phylogeny of H3N2's haemagglutinin gene [4] , [5] , [41] , none have been quantitatively fitted and tested against a set of epidemiological and molecular genetic features such as those in Figure 1 . Here , we focus on the epochal evolution model as formulated in [7] within the above spatial framework , which is identical to ( 5 ) when antigenic variants are interpreted as major antigenic clusters . To fit ( 5 ) to the serial replacement of 11 major antigenic clusters within 1968–2002 [31] , we define an antigenic cluster as any antigenic unit that survives for at least years and use the summaries in Table 1 as well as the number of antigenic clusters generated in 1968–2002 ( nclust ) . Following [7] , the emergence rate is set to increase with age , ( 7 ) , and the scaling parameter is estimated . For simplicity , we refer to ( 5 ) with this antigenic emergence rate and an antigenic resolution that is determined by nclust as the epochal evolution model .
Fitting mechanistic models to infectious disease dynamics of RNA viruses that may escape immunity is notoriously difficult , and key epidemiological parameters such as can be estimated only under tacit assumptions from incidence time series [18] , [19] , [43] . Currently , alternative statistical synthesis approaches are explored to harness the information in complementary data sources [25] , [44] , [45] . Considering summaries of interpandemic H3N2 sequence and surveillance data , we show here that ABC can be used to fit and assess complex phylodynamic models which describe how evolutionary and ecological processes of the influenza virus may interact . Key phylodynamic parameters could be estimated under relatively weak assumptions ( Table 2 ) , and ABC diagnosed readily if and in which direction the two considered models deviate from all the available data taken together . Phylodynamic parameter inference and goodness-of-fit analyses rely critically on the possibility to combine epidemiological and molecular genetic data . In particular , H3N2 case report data were not sufficient to disentangle the reporting rate from epidemiological parameters , and measures of sequence divergence and diversity were not sufficient to separate the population size from evolutionary parameters . To the extent that other RNA viruses are characterized by different phylodynamic behavior , different sets of summaries must be identified in each case to replace likelihood calculations . ABC relates evolutionary and epidemiological data mechanistically through an evolving dynamic system and thereby allows us to investigate empirical phylodynamic hypotheses more directly than is possible with other statistical data synthesis approaches [44] , [45] . Whenever the evolution and ecology of the virus are inseparably linked [1] , case report and phylogenetic summaries are co-dependent . In general , this reduces the degrees of freedom of a phylodynamic model in reproducing features of both types of data simultaneously , and may reveal model inconsistencies . For example , the fitted epochal evolution model could not reproduce the TMRCA's and the population attack rates at the same time ( Figure 5L ) . The reported parameter estimates and summary errors are derived by conditioning only on the phylodynamic summaries and weighting schemes described in Table 1 . ABC is sensitive to the chosen summary statistics and the tolerances since they determine how the prior is re-weighted in light of the presented evidence ( see for example Table S3 and Figure S4 in Text S1 ) [26] . Here , we chose broad enough tolerances such that the weighting schemes are robust to differences in surveillance time series from the Netherlands , France and the US . This approach seems appropriate to avoid overfitting in the context of the limitations of syndromic influenza surveillance , but may be less suited in the analysis of other viral infectious diseases . Figures 3 and 5 illustrate that the resulting dimension reduction regularizes the underlying , intractable likelihood into a smooth , yet well-defined surface such that key phylodynamic model parameters are identifieable and goodness-of-fit can be characterized . It remains unclear to what degree the use of sufficient statistics or the full historical data would be desirable . Indeed , when infectious pathogens escape immunity , the likelihood surface can be especially complex [18] , [43] . Likelihood-based inference is then sensitive to small changes in the complete historical data [46] , [47] , which can be problematic when the reported incidence time series or viral phylogeny is itself subject to considerable uncertainty and/or bias [2] , [48] . We used ABC to fit mechanistic phylodynamic models of interpandemic influenza A ( H3N2 ) to summaries of surveillance data from the Netherlands and sequence data from Northern Europe . Influenza is a globally circulating virus , and the mechanistic models considered must account for the replenishment of genetic variants from outside Northern Europe in order to reproduce features of influenza's phylogeny . In contrast , semi- or non-parametric models of population dynamics that are used in coalescent methods do not necessarily require this layer of spatial complexity [9] , [20] . Here , the mechanistic structure of Eqns . ( 5 ) constrains the set of possible phylodynamic trajectories in such a way that influenza's global disease dynamics must be explicitly accounted for . Put more generally , the quantitative features of H3N2 sequence and incidence data contain sufficient information to determine at least some basic aspects of phylodynamic process models statistically . The two models we analyzed show clear limitations in their ability to replicate features of H3N2 sequence and surveillance data simultaneously , and the ABC error diagnostics give some indication how these models could be refined ( Figures 3 and 5 ) . For example , the phylogenies generated under the SEIRS and the epochal evolution models have , across time , more lineages than the observed HA phylogeny ( Table 3 ) . One possible explanation is that localized extinctions may not occur sufficiently often under the re-seeding source-sink framework , suggesting that models with more detailed population structure , either in space or by age , may result in thinner phylogenies . Accounting for these types of population structure can be critical for understanding viral phylodynamics; here we showed that the fit of the epochal evolution model to both sequence and incidence data depends critically on the assumed spatial model structure and the associated φ ( see Text S1 ) . The SEIRS model could not generate the irregularity in observed incidence data . In comparison , our analysis of the epochal evolution model demonstrates that epochal evolutionary processes can easily excite irregular between-season dynamics that match observed data ( see Figures 5I and S18 in Text S1 ) . Since the virus is known to be under intense immune selection [2] , it seems plausible that antigenic evolution is an important co-factor in explaining influenza's irregular seasonality in temperate regions [49] . Several alternative models have been proposed to reproduce H3N2's narrow HA phylogeny . Here , we identified an additional , testable constraint for these models on surveillance data , that arises through the phylodynamic interactions in Eqns . ( 5 ) . The cluster-specific duration of immunity must be sufficiently long to avoid deep phylogenetic branching . If the fitted values of and are correlated , this in turn implies a characteristic range of population level attack rates that can be tested against available data as in Figure 5L . In particular , the duration of immunity can be lower if the time between replacement events is shorter . Thus , while H3N2's limited standing genetic diversity provides information on the strength of immune interactions between H3N2 antigenic variants , this second constraint may help identify the tempo of antigenic evolution . For the epochal evolution model with source-sink migration dynamics , the average simulated waiting time is years from the emergence of the current antigenic cluster to the next successfully invading offspring antigenic cluster , and this implies phylodynamics that are inconsistent with the molecular genetic and epidemiological summaries in Table 1 taken together . More frequent and more gradual transitions between antigenic variants that are smaller than H3N2's antigenic clusters would allow for lower estimates of that are more in line with observed population level attack rates , break weaker refractory oscillations in their onset , and might also provide sufficient , continual selection pressures to explain the fast divergence in H3N2's HA phylogeny [8] . In this case , sequence and surveillance data would point to a finer antigenic resolution than the one suggested through antigenic map analyses [31] . Alternatively , it is also possible that finer population structure , either in space or by age , could increase extinction rates and thereby allow for a narrow HA phylogeny under a broader , more realistic set of epidemiological parameters without accelerating the tempo of antigenic evolution per se . More broadly , both types of data are now increasingly becoming available for RNA viruses [9]–[13] . This study indicates that these data , when considered simultaneously , may drastically constrain parameter space and readily expose model deficiencies , so that ABC appears as a well-suited tool to explore the phylodynamics of RNA viruses .
|
The infectious disease dynamics of many viral pathogens like influenza , norovirus and coronavirus are inextricably tied to their evolution . This interaction between evolutionary and ecological processes complicates our ability to understand the infectious disease behavior of rapidly evolving pathogens . Most statistical methods for the analysis of these “phylodynamics” require that the likelihood of the data can be explicitly calculated . Currently , this is not possible for many phylodynamic models , so that questions on the interaction between viral variants cannot be well-addressed within this framework . Simulation-based statistical methods circumvent likelihood calculations . Considering interpandemic human influenza A virus subtype H3N2 , we here illustrate the effectiveness of these methods to fit and assess complex phylodynamic models against both sequence and surveillance data . We find that combining molecular genetic and epidemiological data is key to estimate phylodynamic parameters reliably . Moreover , the information in the available data taken together is enough to expose quantitative model inconsistencies . Methods such as ABC which can combine sequence and surveillance data appear to be well-suited to fit and assess mechanistic hypotheses on the phylodynamics of RNA viruses .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"public",
"health",
"and",
"epidemiology",
"evolutionary",
"ecology",
"mathematics",
"ecology",
"evolutionary",
"biology",
"epidemiology",
"statistics",
"theoretical",
"ecology",
"biology",
"population",
"biology",
"statistical",
"methods"
] |
2012
|
Phylodynamic Inference and Model Assessment with Approximate Bayesian Computation: Influenza as a Case Study
|
Insecticidal crystal toxins derived from the soil bacterium Bacillus thuringiensis ( Bt ) are widely used as biopesticide sprays or expressed in transgenic crops to control insect pests . However , large-scale use of Bt has led to field-evolved resistance in several lepidopteran pests . Resistance to Bt Cry1Ac toxin in the diamondback moth , Plutella xylostella ( L . ) , was previously mapped to a multigenic resistance locus ( BtR-1 ) . Here , we assembled the 3 . 15 Mb BtR-1 locus and found high-level resistance to Cry1Ac and Bt biopesticide in four independent P . xylostella strains were all associated with differential expression of a midgut membrane-bound alkaline phosphatase ( ALP ) outside this locus and a suite of ATP-binding cassette transporter subfamily C ( ABCC ) genes inside this locus . The interplay between these resistance genes is controlled by a previously uncharacterized trans-regulatory mechanism via the mitogen-activated protein kinase ( MAPK ) signaling pathway . Molecular , biochemical , and functional analyses have established ALP as a functional Cry1Ac receptor . Phenotypic association experiments revealed that the recessive Cry1Ac resistance was tightly linked to down-regulation of ALP , ABCC2 and ABCC3 , whereas it was not linked to up-regulation of ABCC1 . Silencing of ABCC2 and ABCC3 in susceptible larvae reduced their susceptibility to Cry1Ac but did not affect the expression of ALP , whereas suppression of MAP4K4 , a constitutively transcriptionally-activated MAPK upstream gene within the BtR-1 locus , led to a transient recovery of gene expression thereby restoring the susceptibility in resistant larvae . These results highlight a crucial role for ALP and ABCC genes in field-evolved resistance to Cry1Ac and reveal a novel trans-regulatory signaling mechanism responsible for modulating the expression of these pivotal genes in P . xylostella .
The Gram-positive entomopathogen Bacillus thuringiensis ( Bt ) is the most widely used biopesticide due to its highly specific activity and environmental safety [1] . The insecticidal activity of Bt is largely attributed to diverse δ-endotoxins ( Cry toxins ) produced during sporulation [2] . Transgenic crops harboring Bt toxin genes ( Bt crops ) are the most successful insecticidal biotechnology , with >75 million hectares planted worldwide [3] . However , high adoption of Bt crops and concurrent use of Bt pesticides represent high selection pressure for insect resistance evolution . To date , cases of field-evolved resistance to Bt sprays or Bt crops have been reported in at least seven insect species [4 , 5] . The economic and environmental importance of Bt insecticides highlight the significance of clarifying the molecular mechanisms of insect field-evolved resistance to Bt . The mode of action of Bt Cry toxins includes a critical binding step to receptors in the insect midgut , which is conducive to formation of a toxin pore on the enterocyte membrane that leads to osmotic cell death [6] . The importance of this binding step is further highlighted by high levels of resistance to Bt Cry toxins being almost exclusively associated with alterations in receptor genes [7] . While a number of insect midgut proteins have been proposed as functional receptors for diverse Cry toxins [8] , high levels of resistance to Cry1 toxins due to reduced toxin binding have been genetically linked to mutations or expression alterations of receptor genes such as cadherin , aminopeptidase-N ( APN ) and alkaline phosphatase ( ALP ) [6] . Recently , mutations in ATP-binding cassette transporter subfamily C member 2 ( ABCC2 ) gene [9–13] have been reported to be linked to high levels of resistance to Bt Cry toxins in diverse lepidopteran insects , and it has been proved to be a functional receptor for Bt Cry toxins in Bombyx mori [14] . Although expression alterations of ABCC genes have been reported to result in chemical insecticide resistance in many insects [15 , 16] , whether the expression alterations of ABCC genes can be involved in insect Bt resistance is unclear . In addition , although altered ALP gene expression seems to be commonly associated with lepidopteran resistance to Cry1 toxins [17 , 18] , there is currently no available functional or genetic evidence for these ALP proteins in Bt resistance . The diamondback moth , Plutella xylostella ( Lepidoptera: Plutellidae ) , is a global notorious pest that can rapidly evolve resistance to insecticides and cause US $4–5 billion in management costs annually [19] . Thus far , field-evolved resistance to Bt sprays has only been described in P . xylostella [20] and greenhouse populations of Trichoplusia ni [21] . In both cases , resistance was monogenic and transmitted as an autosomal recessive trait associated with reduced toxin binding to the midgut [22 , 23] . In T . ni , this reduced Cry1Ac toxin binding is associated with reduced expression of a midgut aminopeptidase gene ( APN1 ) [24] possibly trans-regulated by an unidentified gene located in a resistance locus containing the ABCC2 gene [10] . In P . xylostella , cis-acting mutations in putative toxin receptor genes are not linked to field-evolved resistance to Cry1Ac [25 , 26] . As in T . ni , resistance to Cry1Ac in P . xylostella also mapped to a multigenic resistance locus ( BtR-1 ) containing the ABCC2 gene [10] . However , the detailed genetic makeup of this resistance locus and the potential trans-acting regulatory effect on Cry toxin receptors by resident genes remain unknown . High-level resistance phenotype to Bt Cry toxins in insects is often autosomal recessive and controlled by a single gene , however , the fact that many players responsible for resistance suggests there may be a common pathway that links all these receptor genes . Mitogen-activated protein kinase ( MAPK ) signaling pathway has been described to control immune defensive responses to Bt Cry toxins in nematodes [27 , 28] and insects [29 , 30] . The MAPK signaling pathway consists of a four-kinase cascade module ( MAP4K , MAP3K , MAP2K and MAPK ) that can positively or negatively regulate expression of diverse functional genes via different transcription factors [31 , 32] . Therefore , it is plausible that the MAPK signaling pathway may be the common pathway that can regulate the expression of diverse receptor genes to result in insect resistance to Bt Cry toxins . In this study , we identify a novel major mechanistic pathway that the MAPK signaling cascade trans-regulating differential expression of ALP and ABCC genes confers high-level resistance to Cry1Ac in both field-evolved and laboratory-selected strains of P . xylostella . This discovery greatly advances our comprehensive understanding of insect resistance mechanisms to Bt Cry1Ac toxin and provides new insights into how insects evolve resistance to Bt entomopathogen .
Bioassays confirmed high-level Bt resistance in diamondback moth strains originally collected from Florida ( DBM1Ac-R , >3500-fold ) , Shenzhen ( SZ-R , 458-fold Cry1Ac ) and Shanghai ( SH-R , 1890-fold , Bt var . kurstaki ) ( S1 Table ) . A fourth near-isogenic strain ( NIL-R ) was generated to control for variation in genetic backgrounds that may be observed between strains , and was highly resistant to both Cry1Ac ( >3900-fold ) and Btk ( >2800-fold ) . Despite their diverse origins , reduced Cry1Ac toxin binding to midgut BBMV proteins was a common phenotype observed in all resistant samples compared to a Bt susceptible reference strain DBM1Ac-S ( Fig 1 ) , suggesting midgut receptor alterations as a likely resistance mechanism . Multiple previously reported midgut receptors for Bt Cry toxins including cadherin , APN and ALP were first investigated . Recently , we have determined that the midgut cadherin is not involved in Cry1Ac resistance in all of our Cry1Ac/Btk resistant P . xylsotella strains [33] . In this study , we further detected that reduced Cry1Ac toxin binding was significantly associated with reduced ALP enzymatic activity in BBMV from larvae of all resistant strains ( P < 0 . 05; Holm-Sidak’s test; n = 3 ) , while APN activity did not differ ( Fig 2A ) . In contrast , we did not detect significant differences in ALP or APN enzymatic activity when comparing gut luminal contents ( Fig 2B ) , supporting a reduction of membrane-bound ALP ( mALP ) might be responsible for reduced toxin binding , but not soluble ALP ( sALP ) . As ALP activity was significantly reduced in resistant strains , we cloned the full length cDNA of a novel mALP gene ( PxmALP , GenBank accession no . KC841472 ) from the DBM1Ac-S larval midgut tissue ( S2 Table ) . The deduced PxmALP protein sequence displays typical structure features of a mALP protein ( S1 Fig ) , and GenBank database search with the PxmALP protein sequence detected high identity to mALPs in diverse insect species . Phylogenetic analysis ( S2 Fig ) showed that PxmALP is clearly grouped into the same cluster of a clade containing many other lepidopteran mALPs reportedly involved in Bt resistance , suggesting the PxmALP may be a functional Cry1Ac toxin receptor as other lepidopteran mALPs [34 , 35] . Full length PxmALP sequencing using larval midgut cDNA from four resistant P . xylostella strains didn’t identify any constant non-synonymous substitutions or indels , suggesting PxmALP mutations are not linked with Cry1Ac resistance . However , qPCR analysis confirmed a significant reduction ( >50% ) in PxmALP expression in larvae from all the resistant strain ( Fig 2C ) , which is also reflected by our RNA-Seq transcriptome profiling data [36] ( S3 Table ) . Moreover , the reduced expression level of PxmALP gene is congruent with reduced ALP activity in the resistant strains . To determine whether PxmALP can serve as a functional Cry1Ac receptor in P . xylostella , heterologous expression of recombinant PxmALP was conducted in Spodoptera frugiperda Sf9 cell culture , and we detected it by Western blotting and ALP activity assays using cells transfected with an empty bacmid or a bacmid containing the Arabidopsis thaliana β-glucuronidase ( GUS ) gene as controls ( S3A Fig ) . Localization of PxmALP to the surface of transfected cells was demonstrated by releasing GPI-anchored proteins through cleavage with phosphatidylinositol-specific phospholipase C ( PI-PLC ) and detecting most of the recombinant ALP activity in the supernatant ( S3B Fig ) . When expressed on the surface of the Sf9 cells , PxmALP bound Cry1Ac toxin as detected by confocal fluorescence microscopy ( Fig 3A ) and ELISA assays ( S3C Fig ) , while no Cry1Ac toxin binding was detected in untransfected or cells expressing the GUS gene . As expected from the GPI-attachment to the membrane of the recombinant PxmALP , release of PxmALP from the surface of Sf9 cells by PI-PLC treatment resulted in Cry1Ac toxin binding being localized to aggregates in the media , probably containing the released PxmALP , rather than to the surface of the Sf9 cells ( Fig 3A , compare panels 2F and 3F ) . Moreover , binding of Cry1Ac toxin to the Sf9 cells expressing PxmALP was conducive to cytotoxicity ( Fig 3B ) , while cell viability was unaffected in untransfected and Sf9 cells expressing the GUS gene . To further test PxmALP as functional Cry1Ac receptor , we silenced PxmALP gene expression and detected larval susceptibility to Cry1Ac protoxin . Both dsRNA concentration and timing of silencing were optimized in preliminary experiments ( S4 Fig ) . As negative controls , we used non-injected or buffer-injected larvae , while to control for unintended off-target effects , we used larvae injected with dsRNA targeting the PxmALP ortholog in Helicoverpa armigera ( HamALP1 , GenBank accession no . EU729322 . 1 ) . Sequence similarity in the dsRNA fragments targeting PxmALP or HamALP1 reached to about 56% , but no consensus motifs were longer than 19 bp to avoid possible off-target effects [37] . Microinjection of dsRNA targeting an internal region of PxmALP ( nucleotides 510 to 883 ) resulted in about 80% reduction in expression levels compared to controls 48 h post-injection , whereas no significant changes in expression were detected when injecting dsRNA targeting HamALP1 ( Fig 3C ) . Subsequent bioassays at 48 h post-injection for 72 h demonstrated that PxmALP silencing resulted in significantly decreased larval susceptibility to Cry1Ac protoxin ( P < 0 . 05; Holm-Sidak’s test; n = 3 ) compared to controls ( Fig 3D ) . Specifically , about 55% mortality was observed in control larvae treated with 1 . 0 μg/ml of Cry1Ac , while only 19% mortality was observed in larvae injected with dsPxmALP ( mortality in non-injected larvae fed control diet was < 5% ) . When the toxin concentration was increased to 2 . 0 μg/ml , 95% mortality was observed in controls while only 48% mortality was detected in larvae injected with dsPxmALP . While resistance to Cry1Ac in the DBM1Ac-R ( previously called Cry1Ac-R ) and NO-QA P . xylostella strains maps to the same BtR-1 resistance locus [38] , the PxmALP gene is located on a separate chromosome [26] . We assembled the approximately 3 . 15 Mb chromosome region representing BtR-1 locus ( Fig 4 ) assisted by linkage mapping data [10] , genomic data of B . mori [39] and P . xylostella [40] , and the genetic synteny between P . xylostella and B . mori . The BtR-1 locus contains four P . xylostella genome scaffolds and more than 130 annotated genes ( S4 Table ) . Presence of seven known genetic mapping marker genes [10] were confirmed and ten candidate resistance genes , including two P450 genes ( CYP18A1 and CYP18B1 ) , five ABCC genes ( ABCC1-5 ) and three genes involved in the MAPK signaling pathway ( two MAPK genes and one MAP4K gene ) , were identified ( Fig 4 ) . Considering previous reports suggesting their involvement in Bt Cry toxins intoxication [9 , 29] , we focused our subsequent work on potential alterations in the ABCC and MAPK genes in BtR-1 . Previous study showed that a 30-bp deletion in exon 20 of ABCC2 gene is linked to Cry1Ac resistance in the NO-QAGE strain of P . xylostella [10] , however , we did not detect any indels or constant non-synonymous substitutions in this region in our resistant strains , which led us to further study all the five ABCC genes in the BtR-1 resistance locus . We assembled the full-length coding sequences of PxABCC1-5 by in silico analysis , PCR cloning and sequencing , and the bona fide full-length cDNA sequences of these five ABCC genes ( GenBank accession nos . KM245560–KM245564 ) as they were incorrectly annotated in the draft P . xylostella genome . Cloning and comparison of the full-length PxABCC1-5 cDNA sequences from midgut pools of susceptible or resistant larvae detected sequence variations in the region encompassing exons 4 to 11 of PxABCC1-3 , which did not result in changes in the number or size of amplified bands in PCR assays targeting PxABCC1 ( S5A Fig , left figure ) , but led to additional amplicons observed for PxABCC2 and PxABCC3 ( S5B and S5C Fig , top figures ) . Sequencing of these amplicons from each strain allowed us to detect one PxABCC1 isoform , eleven PxABCC2 isoforms and four PxABCC3 isoforms , probably resulting from alternative splicing of the PxABCC1-3 mRNA precursor ( S6 Fig ) . Some of the identified alternative splicing isoforms contained premature stop codons in the transmembrane domain ( TMD ) or the subsequent nucleotide binding domain 1 ( NBD1 ) , which would result in truncated and possibly non-functional proteins . However , their relative distribution was similar ( S5A Fig , right figure; S5B and S5C Fig , bottom figures ) among untreated individual susceptible ( DBM1Ac-S ) larvae and larvae of the NIL-R resistant strain surviving exposure to an extremely high dose of Cry1Ac protoxin ( 10000 μg/ml , causes 100% mortality in DBM1Ac-S larvae ) , supporting no associations between ABCC isoforms in that region and resistance to Cry1Ac . The lack of large inversions or deletions was also tested and confirmed using one-step amplification of the full-length PxABCC1-5 cDNA followed by nested PCR with overlapping primer sets ( S5–S9 Tables ) . Since no association between mutations in PxABCC genes and resistance was observed , we subsequent compared levels of expression for all five ABCC genes in the BtR-1 locus in susceptible and resistant strains using qPCR ( Fig 5A ) . These analyses revealed that three of the five ABCC genes ( PxABCC1-3 ) showed significant differences in gene expression between susceptible and resistant strains , whereas no obvious expression alteration for PxABCC4 or PxABCC5 was found . Interestingly , while PxABCC2 and PxABCC3 were significantly down-regulated , PxABCC1 was dramatically up-regulated in all the resistant compared to susceptible P . xylostella strains ( P < 0 . 05; Holm-Sidak’s test; n = 3 ) . These observations ( except PxABCC2 differences ) were also detected after re-analyzing of our RNA-Seq transcriptome profiling data [36] ( S3 Table ) . The observed association between differential PxABCC gene expression and resistance led us to investigate the relationship among the three PxABCC genes . Protein sequence analysis showed that these genes share typical structural features of ABCC family members , including two transmembrane domains ( TMDs ) and two nucleotide binding domains ( NBDs ) . The genomic structure showed that these three PxABCC genes display high protein sequence similarity ( about 59% ) , extremely similar exon size and number , and the same intron phase ( S10 Table; see also S7A Fig ) , which indicating they are paralogs derived from an ancient gene duplication event . Phylogenetic analysis showed that the P . xylostella ABCC1-3 genes share high sequence identity with homologs from Spodoptera exigua ( S7B Fig ) , which have been recently proved to be involved in resistance to Cry1 toxins [12] . To determine the effect of PxABCC2 and PxABCC3 down-regulation on susceptibility to Cry1Ac toxin , we silenced their expression by RNAi and tested susceptibility ( LC50 and LC90 ) in silenced larvae . Expression levels for both genes were significantly reduced at 24 h after dsRNA injection , with lowest expression levels detected after 48 h and lasting at least 72 h in both cases ( Fig 5B and 5C ) . Silencing was specific to each ABCC gene and did not affect PxmALP expression levels ( Fig 5B and 5C ) . Bioassays performed at 48 h post-injection for 72 h showed a marked decrease in susceptibility to Cry1Ac toxin in both dsPxABCC2- and dsPxABCC3-treated larvae compared to the buffer- or dsEGFP-injected larvae . Moreover , silencing of multiple genes simultaneously ( combinational RNAi ) by injection of a combination of dsRNAs targeting PxmALP , PxABCC2 and PxABCC3 ( dsMultigenes ) resulted in a comparatively higher reduction in susceptibility to Cry1Ac ( P < 0 . 05; Holm-Sidak’s test; n = 3 ) ( Fig 5D ) . Specifically , about 50% mortality was observed in control larvae treated with 1 . 0 μg/ml of Cry1Ac ( LC50 value ) , while only 12% , 15% and 6% mortality was observed in larvae injected with dsPxABCC2 , dsPxABCC3 and dsMultigenes , respectively ( mortality in non-injected larvae fed control diet was <5% ) . When using 2 . 0 μg/ml ( LC90 value ) , 92% mortality was observed in controls while only 32% , 37% and 15% mortality was detected in larvae injected with dsPxABCC2 , dsPxABCC3 and dsMultigenes . All three silencing treatments had significant effects ( LSD test; P < 0 . 05; n = 3 ) on fitness components , including decreased pupation rate , reduced pupal weight , shortened pupal time and lower eclosion rate when compared to control treatments , and no differences ( LSD test; P > 0 . 05; n = 3 ) were detected among control treatments ( S11 Table ) . We further performed genetic linkage analysis to test non-cosegregation of alternative ABCC1-3 gene splicing isoforms or cosegregation of differentially altered expression of the PxmALP and PxABCC1-3 genes with resistance to Cry1Ac toxin in the NIL-R strain . Reciprocal F2 backcross families from crossing the near-isogenic NIL-R ( resistant ) and DBM1Ac-S ( susceptible ) strains were generated and selected on cabbage with or without a lethal dose of Cry1Ac protoxin ( S8 Fig ) . Comparison of distribution and sequencing of PxABCC1-3 cDNA isoforms among individual larvae from backcross families exposed or not to Cry1Ac toxin demonstrated no association between PxABCC1-3 isoforms and resistance to Cry1Ac ( S9 Fig ) . Quantification of PxmALP and PxABCC1-3 expression levels in individual larval midguts from backcross families not exposed to Cry1Ac selection showed two distinct groups ( Fig 6 ) . One group demonstrated significantly reduced expression levels of PxmALP ( < 0 . 4-fold ) , PxABCC2 ( < 0 . 15-fold ) and PxABCC3 ( < 0 . 25-fold ) , while the other group displayed expression levels similar to larvae from the susceptible parental strain ( DBM1Ac-S ) or the F1 generation from NIL-R × DBM1Ac-S crosses ( Fig 6A , 6C and 6D ) . The ratio between the numbers of individuals in each group , 8:10 , 9:9 and 9:9 in backcross family a and 9:9 , 9:9 and 10:9 in backcross family b , for the PxmALP , PxABCC2 and PxABCC3 genes , respectively , were statistically validated to follow the 1:1 random assortment ratio ( P > 0 . 10 or P = 1 . 0; χ2 test ) . Rearing of neonates from both backcross families on Cry1Ac resulted in about 50% mortality ( 55% backcross family a , 47 . 5% in family b ) , consistent with the expected Mendelian inheritance of the recessive resistance trait . All the surviving larvae in both backcross families had reduced PxmALP ( < 0 . 4-fold ) , PxABCC2 ( < 0 . 15-fold ) and PxABCC3 ( < 0 . 25-fold ) expression levels compared to larvae from the DBM1Ac-S strain or the F1 generation , demonstrating tight linkage ( cosegregation ) with resistance to Cry1Ac in NIL-R ( P < 0 . 001 , χ2 test ) . However , the expression levels for the PxABCC1 gene in both Cry1Ac-selected and non-selected larvae were similarly up-regulated ( 4- to 14-fold ) compared to susceptible larvae ( Fig 6B ) . Although the unselected backcross individuals exhibited two distinct groups with differing PxmALP , PxABCC2 and PxABCC3 expression levels , both groups of larvae had similar PxABCC1 expression levels ( t test , P > 0 . 10 ) , supporting no correlation between down-regulation of PxmALP , PxABCC2 , or PxABCC3 and up-regulation of PxABCC1 expression in the larvae . Analysis of the BtR-1 locus found three genes ( two MAPK genes and one MAP4K gene ) involved in MAPK signaling pathways ( S4 Table; see also Fig 4 ) . Unlike the two MAPK genes , the PxMAP4K4 gene locates extremely close to the three PxABCC genes within the core BtR-1 locus and shows perfect genetic synteny between P . xylostella and B . mori ( S4 Table; see also S10 Fig ) . Using specific primers ( S12 Table ) , we cloned and corrected the full-length cDNA sequence of the incorrectly annotated PxMAP4K4 gene in the P . xylostella genome ( DBM-DB , Gene ID Px002422 ) , the bona fide full-length cDNA sequence has been deposited in the GenBank database ( accession no . KM507871 ) . Sequence alignment of the deduced amino acid sequences showed conserved N-terminal kinase ( Serine/threonine kinase catalytic domain , STKc ) and the C-terminal regulatory ( citron/NIK homology domain , CNH ) domains of PxMAP4K4 gene in homologs from different species ( S11 Fig ) . Comparisons of PxMAP4K4 expression levels between susceptible and resistant strains by qPCR showed that this gene was constitutively up-regulated in larvae from all resistant strains compared to the susceptible strain ( Fig 7A ) . Moreover , toxin induction assays showed that the expression level of PxMAP4K4 was significantly increased ( P < 0 . 05; Holm-Sidak’s test; n = 3 ) in the susceptible strain DBM1Ac-S but didn’t alter ( P > 0 . 05; Holm-Sidak’s test; n = 3 ) in the resistant strain NIL-R when treated with respective median lethal concentration of Cry1Ac in both strains , suggesting that the high expression levels of PxMAP4K4 in resistant strain was constitutive rather than induced ( S12 Fig ) . To confirm the significance of this observation , we silenced PxMAP4K4 expression by RNAi in resistant NIL-R larvae and tested larval susceptibility to Cry1Ac post-RNAi . Microinjection of dsRNA targeting the CNH domain region of the PxMAP4K4 mRNA resulted in about 55% reduction ( 0 . 45-fold ) in expression levels at 48 h post-injection , with expression returning to control levels at 120 h post-injection ( Fig 7B ) . Correspondingly , the expression levels of PxmALP , PxABCC2 and PxABCC3 were significantly increased by 2 . 1- , 4 . 8- and 2 . 5-fold at 48 h post-injection , whereas the expression level of PxABCC1 was dramatically reduced 0 . 28-fold ( 77% reduction ) at this time point ( Fig 7B ) . Subsequent bioassays performed at 48 h post-injection for 72 h demonstrated that silencing of PxMAP4K4 gene expression resulted in a significant increase in larval susceptibility to Cry1Ac protoxin ( P < 0 . 05; Holm-Sidak’s test; n = 3 ) when compared to the buffer- or dsEGFP-injected larvae ( Fig 7C ) . Specifically , about 3% and 12% mortality was observed in respective control larvae untreated or treated with 1000 μg/ml of Cry1Ac ( LC10 value ) , while approximately 72% mortality was observed in larvae injected with dsPxMAP4K4 , and mortality in dsPxMAP4K4-treated larvae not exposed to toxin was < 4% .
Field insect populations can develop resistance to entomopathogens used as biopesticides , such as B . thuringiensis ( Bt ) , limiting their potential efficacy for pest management . Multiple Cry1A midgut receptors have been reported in Lepidoptera , which should theoretically make resistance evolution difficult , however , genetic analysis has commonly shown resistance to be a single autosomal locus [41] . Data in this study provides a comprehensive mechanistic description of resistance to Cry1Ac and a Btk biopesticide in larvae from diverse P . xylostella strains . Although previous reports supported that mutations in the PxABCC2 gene localized to the BtR-1 locus are responsible for resistance to Cry1Ac in P . xylostella [10] , we did not detect any mutations in the PxABCC2 or other PxABCC genes in BtR-1 associated with resistance in any of our tested strains . In contrast , our findings clearly support that differential expression of a midgut membrane-bound alkaline phosphatase ( PxmALP ) gene and a suite of PxABCC genes ( including PxABCC2 ) is associated with high levels of resistance to Cry1Ac and Btk in P . xylostella . This is the first report showing that expression alterations , not gene mutations , of ABCC2 and other ABCC genes can be involved in insect Bt resistance . More importantly , for the first time , we identify a transcriptionally-activated upstream gene in the MAPK signaling pathway ( PxMAP4K4 ) within the BtR-1 locus can trans-regulate differential altered expression of the PxmALP and PxABCC genes in BtR-1 to result in Cry1Ac resistance . This novel molecular mechanism of Cry1Ac resistance in P . xylostella is summarized in Fig 8 . Resistance to Cry1Ac in our field-evolved strain DBM1Ac-R [42] and the near isogenic strain NIL-R [43] fits the “Mode 1” type characterized by high levels of resistance to at least one Cry1A toxin , recessive inheritance , reduced binding of at least one Cry1A toxin , and lack of cross-resistance to Cry1C toxin [44] . Reduction in Cry1Ac binding is associated with “Mode 1” type resistance in nearly all cases of field-evolved P . xylostella resistance [7] . This observation , coupled with the Cry toxin binding site model developed for P . xylostella [45] and our Cry1Ac toxin binding data , clearly suggested that “Mode 1” resistance in our strains was due to alterations in at least one Cry1A toxin receptor . While a number of putative Cry receptors have been proposed [8] , only alterations in cadherin , APN , ALP and ABCC2 genes have been found to associate with high resistance to Cry1Ac in Lepidoptera [6] . Dramatically reduced ALP enzymatic activity in BBMV samples of resistant P . xylostella larvae prompted us to clone the full-length cDNA of this gene for further functional studies . Sequence analysis showed that the cloned PxmALP is identical to the partial ALP1 gene sequence ( GenBank accession no . EF579960 , a partial genomic sequence including an intron ) . In agreement with Cry1Ac resistance not genetically mapping to ALP1 [26] , we did not detect any mutations in PxmALP associated with resistance among our P . xylostella strains , however , the PxmALP expression levels were significantly reduced . This observation is also supported by our prior reports demonstrating down-regulation of ALP1 in the DBM1Ac-R ( Cry1Ac-R in that report ) strain [46] and a more detailed analysis of our recent RNA-Seq survey ( S3 Table ) [36] . In agreement with these observations , current data support ALPs as relevant Cry1A toxin binding proteins [47 , 48] . Importantly , altered ALP levels were described as associated with “Mode 1” type resistance to Cry1Ac in lepidopteran hosts [17 , 18] , yet no mechanistic or linkage evidence were provided . Based on the functional and linkage data here , we propose a model for “Mode 1” resistance in P . xylostella in which PxmALP serves as a “lethal receptor” for Cry1Ac toxin . Down-regulation of PxmALP expression in resistant larvae results in reduced toxin binding to the midgut cells and survival . Our definition of PxmALP as a “lethal receptor” is based on its effectiveness as Cry1Ac receptor in cell assays , and the similar susceptibility in heterozygous and homozygous susceptible larvae , as previously suggested in Cry1Ac-resistant H . virescens [49] . Previous complementation tests have suggested that the same resistance locus ( BtR-1 ) containing the mutant PxABCC2 gene [10] is responsible for resistance to Cry1Ac in P . xylostella strains from the continental US ( PEN from Pennsylvania , SC1 from South Carolina and DBM1Ac-R originally collected from Florida ) , Hawaii ( NO-QA and NO-QAGE ) , and China ( SZBT ) [22 , 25 , 38] . Since that the PxmALP gene is not located in the BtR-1 locus [26] , our first hypothesis to reconcile the available data on P . xylostella resistance was that mutations or altered expression of PxABCC2 or other PxABCC genes in BtR-1 could result directly or indirectly in reduced expression of PxmALP . In support of this observation , resistance to a Bt pesticide in S . exigua [12 , 50] and to Cry1Ac in T . ni [10 , 24] was linked to mutations in ABCC2 and a concomitant down-regulation of an APN gene . However , unlike previous reports [10] , we did not detect any mutations in PxABCC2 associated with resistance . Instead , we detected alternative splicing of ABCC subfamily genes , as previously reported in other insects [51–53] . Some of the splicing isoforms can lead to truncated ABCC proteins , which in heterozygotes would be masked by the susceptible allele , as previously proposed for cadherin gene [49] . Thus , these alternative splicing isoforms may represent a natural system to generate recessive gene mutation pools for Bt resistance selection . Intriguingly , alternative ABCC splicing was limited to the exon 4–11 region , suggesting that this cDNA region is more prone to allow mutations . This mechanism would explain rapid appearance of field-evolved resistance to Bt sprays in P . xylostella . Although gene mutations of PxABCC genes were excluded , our further study corroborated differential expression alterations of these PxABCC genes were associated with Cry1Ac intoxication in P . xylsotella . However , our RNAi results showed that silencing of PxABCC2 or PxABCC3 genes did not affect PxmALP expression , which suggests PxABCC genes can’t regulate the expression level of PxmALP gene . Then , an independent role of PxABCC and PxmALP genes in Cry1Ac susceptibility is suggested as our second hypothesis . In support of this observation , similar effects on Cry1Ac susceptibility were detected when silencing PxABCC or PxmALP genes separately , and the comparatively higher reduction in Cry1Ac susceptibility was observed after performing combinatorial RNAi to simultaneous silencing of PxmALP , PxABCC2 and PxABCC3 genes . Therefore , it is plausible to postulate that an uncharacterized trans-regulatory gene in the BtR-1 locus could potentially control differential altered expression of both PxABCC and PxmALP resistance genes . Not incidentally , a similar hypothesis was proposed to explain allelic expression alteration of ABCC2 and APN3 genes in B . mori [54] , and APN down-regulation in Cry1Ac-resistant T . ni [24] and Cry1Ab-resistant Ostrinia nubilalis [55] . While genes modulating ALP and ABCC expression in insects have not been reported , genes in the MAPK signaling pathway have been shown to modulate mammalian ALP [56–58] and human ABCC gene expression [59–62] . The MAPK signaling pathway can be activated as a defensive response to Bt Cry toxins [27–30] . Consequently , it is plausible that altered PxmALP and PxABCC gene expression in resistant P . xylostella may result from a primary enhanced defensive response involving activation of a trans-acting gene in the MAPK signaling pathway located in the BtR-1 locus . As expected , we found three MAPK genes in BtR-1 and identified the PxMAP4K4 gene in proximity to the three ABCC genes ( S10 Fig ) . Of particular note , homologs of this PxMAP4K4 gene in mammals , Caenorhabditis elegans , and Drosophila are all upstream components of the MAPK signaling pathway and play important physiological roles in these species [63–65] . Accordingly , we found that the expression level of this gene can be induced in the susceptible strain when challenged by low concentration of Cry1Ac toxin , and functional data in this study demonstrated that this gene is constitutively transcriptionally-activated in resistant larvae to trans-regulate PxmALP and PxABCC expression levels thereby dramatically affecting Cry1Ac susceptibility , which finally attests to the involvement of the MAPK signaling pathway in P . xylostella Cry1Ac resistance . Since that cadherin can mediate the intracellular MAPK signaling pathway in mammal or insect cells [32 , 66 , 67] , and considering that alteration of cadherin gene expression is associated with resistance to Cry toxins in several other lepidopteran insects [68 , 69] , it will be very interesting to examine the possible feedback regulation of MAPK signaling in cadherin gene expression regulation . Moreover , since diverse upstream cytokines in the MAPK signaling pathway can regulate expression of APN genes in mammals [70 , 71] , and considering that alteration of APN gene expression is associated with resistance to Cry toxins in several other lepidopteran insects [24 , 50 , 54 , 55 , 72] , it will be very interesting to examine the possible involvement of MAPK signaling in APN gene expression regulation . Recently , we have found that down-regulation of a novel ABC transporter gene ( PxABCG1 or Pxwhite ) possibly trans-regulated by the MAPK signaling pathway can also be involved in P . xylostella Cry1Ac resistance , suggesting the MAPK signaling pathway may trans-regulate numerous ABC transporters from different subfamilies [73] . Therefore , it is plausible that this novel trans-regulatory mechanism might be a common regulation event of diverse Bt receptor genes in all of these cases . Duplication or amplification of functional genes is thought to be a major driving force for adaptive evolution of insect response to environmental stress [74] and development of insecticide resistance [75 , 76] . Whole genomic analyses support that the ABC transporter superfamily has undergone apparent gene duplication in the P . xylostella genome [40] , and this duplication may allow for functional redundancy . Based on their extremely similar genomic structure and high sequence similarity , it is highly possible that the PxABCC1-3 genes in BtR-1 locus may have been generated through gene duplication and share similar functions . Considering that PxABCC1 gene up-regulation was not linked to Cry1Ac resistance , we speculate that PxABCC1 may have lost the ability to bind Cry1Ac toxin but retained substrate transport function to functionally rescue the reduced PxABCC2 and PxABCC3 phenotype in resistant larvae . Likewise , transcriptionally-activated MAPK signaling induced by Cry5B intoxication in C . elegans up-regulated a target cation efflux transporter gene ( ttm-1 ) possibly involved in removing cytotoxic cations generated by toxin-induced pore formation [27] . This phenomenon would also resemble up-regulation of APN6 in Cry1Ac-resistant T . ni with down-regulated APN1 expression [24] . Since silencing of PxABCC2 and PxABCC3 genes result in obvious fitness costs in P . xylostella larvae , rescue of ABCC gene function in resistant P . xylostella by PxABCC1 would also help explain lack of fitness costs in the DBM1Ac-R strain [77] . These data suggest that functional redundancy in ABCC genes can reduce fitness costs and thus increase the probability of resistance evolution in the field , which may threaten the continued effectiveness of Bt sprays/Bt crops and the currently adopted refuge strategy . In this case , we should attach great importance to this observation and perform continuous field monitor of insect resistance in such form . In summary , the present study shows that alteration in expression of multiple putative Cry1Ac receptors is linked to “Mode 1” type resistance in P . xylostella , and that this altered gene expression is trans-regulated by the MAPK signaling pathway . Although the data presented do not directly address the participation of additional elements and the full repertoire of the MAPK signaling pathway , they do provide strong evidence for an important role of this signaling pathway in insect susceptibility to Cry1Ac toxin . Further work is needed to identify additional toxin receptor genes ( e . g . cadherin and APN ) that may also be controlled by this pathway and other MAPK genes or downstream transcription factors involved in Bt resistance in insects . The present data deepens our understanding of how insect target cells counter Cry intoxication through functionally sophisticated intracellular responses to result in Bt resistance . Moreover , the identified pivotal genes and their expression regulation mechanism responsible for resistance to Cry toxins in this study are critical for sensitive and efficient monitoring and management practices to delay field-evolved insect resistance to Bt pesticides and Bt crops .
The susceptible DBM1Ac-S and resistant DBM1Ac-R ( previously referred to as Cry1Ac-R ) strains of P . xylostella were originally provided by Drs . J . Z . Zhao and A . Shelton ( Cornell University , USA ) in 2003 . The DBM1Ac-R strain originated from insects with field-evolved resistance to Javelin ( Bt var . kurstaki ) from Loxahatchee ( Florida , USA ) [78] that were crossed with the DBM1Ac-S ( Geneva 88 ) strain ( originated from Geneva , NY , USA ) and further selected with Cry1Ac-expressing broccoli [79] . Resistance to Cry1Ac in DBM1Ac-R is autosomal , incompletely recessive and mostly monogenic [77] . The SZ-R ( previously referred to as T2-R ) and SH-R strains were originated from moths collected in China at Shenzhen ( 2003 ) and Shanghai ( 2005 ) , respectively . The SZ-R strain was generated by selection in the laboratory with Cry1Ac while the SH-R strain was selected with a Bt var . kurstaki ( Btk ) formulation ( WP with potency of 16000 IU/mg , provided by Bt Research and Development Centre , Agriculture Science Academy of Hubei Province , China ) . The DBM1Ac-S strain was kept unselected while the DBM1Ac-R and SZ-R strains have been kept under constant selection with a Cry1Ac protoxin solution killing 50–70% of the larvae sprayed on cabbage leaves . The near-isogenic NIL-R strain has been generated at the time this study was carried out and has been described elsewhere [43] . For this work , all strains were reared on JingFeng No . 1 cabbage ( Brassica oleracea var . capitata ) without exposure to any Bt toxins or chemical pesticides at 25°C , 65% RH and 16D:8L photoperiod . Adults were fed with a 10% sucrose solution . The Cry1Ac protoxin was extracted and purified from Bt var . kurstaki strain HD-73 as previously described [80] . Both purified Cry1Ac protoxin and trypsin-activated toxin were quantified by densitometry as described elsewhere [81] . Toxicity of Cry1Ac toxin or Btk formulation in 72 h bioassays with larvae from five different strains of P . xylostella using a leaf-dip method as described elsewhere [33] . Ten third instar P . xylostella larvae were tested for each of seven toxin concentrations and bioassays replicated four times . Mortality data were corrected using Abbott’s formula [82] and experiments with control mortality exceeding 10% were discarded and repeated . The LC50 values were calculated by Probit analysis [83] . Fourth-instar larval midguts ( about 2000 ) from each P . xylostella strain were dissected in cold MET buffer [17 mM Tris–HCl ( pH 7 . 5 ) , 5 mM EGTA , 300 mM mannitol] plus protease inhibitors ( 1mM PMSF ) . Midgut brush border membrane vesicles ( BBMV ) were prepared as described elsewhere [84] . Purified BBMV proteins were quantified using the method of Bradford [85] with bovine serum albumin ( BSA ) as standard , and then flash frozen and kept in aliquots at -80°C until used . Between 5–8 fold enrichment in specific APN activity using L-leucine-p-nitroanilide ( Sigma ) was detected when comparing to initial midgut homogenates . Midgut luminal contents were obtained as described elsewhere [18] by homogenization of pools of dissected midguts of actively feeding fourth-instar P . xylostella larvae with an electric pestle in a 1 . 5 ml centrifuge tube containing 100 μl of phosphate-buffered saline ( PBS ) buffer ( 137 mM NaCl , 2 . 7 mM KCl , 10 mM Na2HPO4 , 2 . 0 mM KH2PO4 , pH 7 . 4 ) . Homogenates were vigorously vortexed and centrifuged at 4°C ( 10 min at 16000×g ) . Supernatants were used for subsequent enzymatic assay measurements or flash frozen and kept in aliquots at -80°C until used . Specific APN and ALP activity assays were performed using L-leucine-p-nitroanilide and p-nitrophenyl phosphate disodium ( pNPP ) as substrates , respectively , as described elsewhere [86] . Enzymatic activity was detected as the changes in optical density ( OD ) at 405 nm for 5 min at room temperature in a SpectraMax M2e ( Molecular Devices ) microplate reader . One enzymatic unit was defined as the amount of enzyme that would catalyze production of the chromogenic product from specific substrate per min and mg of BBMV protein at 37°C . Data shown are the means from triplicate measurements from three independent BBMV preparations and were analyzed for significance with a one-way ANOVA using Holm-Sidak’s tests ( overall significance level = 0 . 05 ) with the SPSS Statistics ( ver . 17 . 0 ) software ( SPSS Inc . ) . The seven mapped genes in Baxter et al . [10] were used as markers to find P . xylostella genome scaffolds within the chromosome region of the BtR-1 resistance locus in the Diamondback moth Genome Database ( DBM-DB , http://iae . fafu . edu . cn/DBM ) , and their locations determined the final jointed pattern of all P . xylostella genome scaffolds . Homologs of the genes within this locus between P . xylostella and B . mori were identified through Blastp searches in each genome database . The detailed genetic makeup of the BtR-1 resistance locus is listed in S4 Table . Fourth-instar larvae from different P . xylostella strains were anesthetized on ice and the midgut tissues were immediately dissected in RNase-free water containing 0 . 7% NaCl . Total RNA was extracted from single or pool of these dissected midguts using TRIzol reagent ( Invitrogen ) according to different experiments . Integrity of the RNA was determined using 1% TBE agarose gel electrophoresis , and then quantified by a NanoDrop 2000c spectrophotometer ( Thermo Fisher Scientific Inc . ) . For gene cloning , the first-strand cDNA was prepared using 5 μg of total RNA with the PrimeScript Ⅱ 1st strand cDNA Synthesis Kit ( TaKaRa ) following manufacturer’s recommendations . For qPCR analysis , the first-strand cDNA was prepared using 1 μg of total RNA with the PrimeScript RT kit ( containing gDNA Eraser , Perfect Real Time ) ( TaKaRa ) following manufacturer’s recommendations . The synthesized first-strand cDNA was immediately stored at -20°C until used . For PxmALP cloning , degenerate primers ( S2 Table ) were designed to target two highly conserved regions among selected lepidopteran mALP sequences ( Bombyx mori , BAB62745; Helicoverpa armigera , ACF40806; Heliothis virescens , ACP39712; and Ostrinia furnacalis , AEM43806 ) . The PCR parameters were as follows: one cycle of 94°C for 3 min; 35 cycles of 94°C for 30 s , 59°C for 45 s and 72°C for 1 min; a final cycle of 72°C for 10 min . The generated 536 bp midgut cDNA fragment was sequenced and used in Rapid Amplification of cDNA Ends ( RACE ) to obtain the full length ALP cDNA with 3′-Full RACE Core Set Ver . 2 . 0 ( TaKaRa ) and SMARTer RACE cDNA Amplification ( Clontech ) kits following manufacturer’s protocols . Once a full-length cDNA for PxmALP was obtained the full coding sequence was validated by PCR amplification . Large-scale sequencing and comparing of the full-length midgut PxmALP was performed among all the susceptible and resistant P . xylostella strains . The full length midgut PxmALP cDNA sequence has been deposited in the GenBank ( accession no . KC841472 ) . To first detect the reported PxABCC2 mutation ( 30 bp deletion in Exon 20 ) correlated with Bt resistance in P . xylostella strain NO-QAGE [10] , one specific primer pair covering this region was designed based on the published partial PxABCC2 cDNA sequence ( GenBank accession no . JN030490 ) ( S6 Table ) . The PCR amplification of all the cDNA sequences of five ABCC genes in the BtR-1 resistance locus was performed using the strategy described elsewhere [87] . We first in silico assembled and corrected the full-length sequences of ABCC1-5 based on their putative coding sequences in the DBM-DB database ( http://iae . fafu . edu . cn/DBM , Gene ID: Px002418+19 , Px002416 , Px002414+15 , Px009835 and Px009834 ) and their unigenes in our P . xylostella transcriptome database [88] ( S3 Table ) , then their cDNA was amplified by designing full-length primer pairs with four or five overlapping fragments ( S5–S9 Table ) . The full length ABCC1-5 cDNA sequences have been deposited in GenBank ( accession nos . KM245560–KM245564 ) . In addition , the occurrence of large inversions or deletions was tested in a second amplification strategy by amplifying the whole cDNA sequence with specific full-length primers ( S5–S9 Table ) in a first step , followed by nested PCR amplification using the same full-length primer pairs with four or five overlapping fragments as described above . Amplicons were sequenced and the distribution of alternatively spliced transcripts of ABCC1-3 was compared among samples from untreated and larvae surviving Cry1Ac exposure . Using the same strategy as for ABCC genes , we cloned and obtained the full-length cDNA sequence of the PxMAP4K4 gene ( GenBank accession no . KM507871 ) . The PCR reactions ( 25 μl total volume ) contained 18 . 5 μl of double-distilled H2O ( ddH2O ) , 2 . 5 μl of 10×LA Taq or Ex Taq Buffer , 2 μl of dNTP Mix , 5 μM of each specific primer , 1 μl of first-strand cDNA template , and 0 . 25 μl LA Taq HS or Ex Taq HS polymerase ( TaKaRa ) . Reactions ( 35 cycles ) were then performed in an S1000 or C1000 Thermal Cycler PCR system ( BioRad ) with the following parameters: one cycle of 94°C for 6 min; 35 cycles of 94°C for 30 s , 50–59°C for 45 s and 72°C for 5 min; a final cycle of 72°C for 15 min . The nested PCR parameters were as follows: one cycle of 94°C for 6 min; 35 cycles of 94°C for 30 s , 59°C ( PxABCC2 and PxMAP4K4 ) /54°C ( other four PxABCC genes ) for 45 s and 72°C for 2 min; a final cycle of 72°C for 10 min . All the cloning primers for each gene were designed in the Primer Premier 5 . 0 software ( Premier Biosoft ) . Amplicons of the expected size were excised from 1 . 5–2 . 5% agarose gels , purified using the Gel Mini Purification Kit ( Generay ) , and subcloned into the pEASY-T1 ( Transgen ) or pMD18-T vectors ( TaKaRa ) before transformation into Escherichia coli TOP10 competent cells ( Transgen ) for sequencing . Gene sequence assembling , multiple sequence alignment and exon-intron analysis were carried out with DNAMAN 7 . 0 ( Lynnon BioSoft ) . The open reading frame of the target nucleotide sequence is found by the ORF Finder tool at NCBI website ( http://www . ncbi . nlm . nih . gov/gorf/gorf . html ) . The nucleotide sequence-similarity analyses were performed through BLAST tool at NCBI website ( http://blast . ncbi . nlm . nih . gov/ ) . The deduced protein sequence was obtained by an ExPASy translate tool Translate ( http://web . expasy . org/translate/ ) from the Swiss Institute of Bioinformatics . The N-terminal signal peptide was determined using the SignalP 4 . 0 server ( http://www . cbs . dtu . dk/services/SignalP/ ) . The transmembrane region and membrane topology was analyzed by the TOPCONS online software ( http://topcons . cbr . su . se/ ) . Protein specific motif was searched and analyzed using the Myhits software ( http://myhits . isbsib . ch/cgi-bin/motif_scan ) , the Prosite software ( http://www . expasy . ch/prosite/ ) and CDD ( conserved domain database ) at NCBI . Two GPI modification site prediction servers ( big-PI Predictor: http://mendel . imp . ac . at/sat/gpi/gpi_server . html and GPI-SOM: http://gpi . unibe . ch/ ) were used to predict the GPI-anchor signal sequence and GPI anchoring site . Presence of N- and O-glycosylation sites on the predicted protein sequence were tested using the NetNGlyc 1 . 0 ( http://www . cbs . dtu . dk/services/NetNGlyc/ ) and NetOGlyc 4 . 0 server ( http://www . cbs . dtu . dk/services/NetOGlyc/ ) , respectively . Protein sequences of the ALP and ABCC genes used for phylogenetic analyses were extracted from different databases: GenBank ( http://www . ncbi . nlm . nih . gov/ ) , SilkDB ( http://silkworm . genomics . org . cn/ ) , DBM-DB ( http://iae . fafu . edu . cn/DBM ) and Manduca Base ( http://agripestbase . org/manduca/ ) , and sequence alignment was carried out after eliminating vast redundant ALP or ABCC sequences . All the selected insect ALP and ABCC amino acid sequences were subjected to analysis through Clustal W alignment using Molecular Evolutionary Genetic Analysis software version 5 . 0 ( MEGA 5 ) [89] , then the phylogenetic tree was constructed using the neighbor-joining ( NJ ) method with “p-distance” as amino acid substitution model , “pairwise deletion” as gaps/missing data treatment and 1000 bootstrap replications . Gene-specific primers to the PxmALP gene were selected and used in PCR reactions ( 25 μl ) containing 11 . 95 μl of ddH2O , 11 . 25 μl of 2 . 5×SYBR Green MasterMix ( TIANGEN ) , 4 μM of each specific primer , and 1 μl of first-strand cDNA template . The qPCR program included an initial denaturation for 6 min at 94°C followed by 40 cycles of denaturation at 94°C for 30 s , annealing for 30 s at 61°C , and extension for 35 s at 72°C . Gene-specific primers for PxABCC and PxMAP4K4 genes were designed in the cDNA regions without alternative splicing and used in PCR reactions ( 25 μl ) containing 9 . 5 μl of ddH2O , 12 . 5 μl of 2×SuperReal PreMix Plus ( TIANGEN ) , 7 . 5 μM of each specific primer , 1 μl of first-strand cDNA template and 0 . 5 μl 50×ROX Reference Dye ( TIANGEN ) . The qPCR program included an initial denaturation for 15 min at 95°C followed by 40 cycles of denaturation at 95°C for 15 s , annealing for 30 s at 53°C ( PxABCC2 ) /55°C ( other four PxABCC genes ) /63°C ( PxMAP4K4 ) , and extension for 32 s at 72°C . For melting curve analysis , an automatic dissociation step cycle was added . Reactions were performed in an ABI 7500 Real-Time PCR system ( Applied Biosystems ) with data collection at stage 2 , step 3 in each cycle of the PCR reaction . Amplification efficiencies were calculated from the dissociation curve of quadruplicate replicates using five 2-fold serial dilutions ( 1:1 , 1:2 , 1:4 , 1:8 , and 1:16 ) . Only results with single peaks in melting curve analyses , 95–100% primer amplification efficiencies , and >0 . 95 correlation coefficients were used for subsequent data analysis . Negative control reactions included ddH2O instead of cDNA template , which resulted in no amplified products . The amplified fragments were sequenced to confirm that potential expression differences were not due to sequence mutations in the targeted genes . Relative quantification was performed using the 2-ΔΔCt method [90] and normalized to the ribosomal protein L32 gene ( GenBank accession no . AB180441 ) as validated elsewhere [40 , 91] . Four technical replicates and three biological replicates were used for each treatment . One-way ANOVA with Holm-Sidak’s tests ( overall significance level = 0 . 05 ) were used to determine the significant statistical difference between treatments . The Bac-to-Bac Baculovirus Expression System ( Invitrogen ) was used to express the recombinant PxmALP protein in Spodoptera frugiperda Sf9 cell cultures . The full-length PxmALP cDNA was cloned and amplified by high fidelity PCR with specific primers ( S2 Table ) . Amplicons were purified , subcloned and sequenced as described above . Recombinant plasmids with correct insertion were verified by endonuclease digestion , PCR and sequencing . The verified positive clone was digested with EcoRI and XbaI for 3 h and then ligated into the pFastBac TH B donor plasmid vector to generate the recombinant pFastBac HT B-PxmALP bacmid . The recombinant plasmids ( pFastBac HT B-PxmALP ) were then transformed into DH10Bac competent cells ( Invitrogen ) and positive recombinant bacmid DNAs were detected by antibiotic selection and confirmed by PCR amplification . For heterologous expression , transfections were performed in sterile six-well plates ( Costar ) . Briefly , Sf9 cultures ( 8×105 cells/well ) with >97% viability were cultured in Grace’s insect medium supplemented with 10% fetal bovine serum and transfected with 1 μg of pFastBac HT B-PxmALP in Cellfectin II Reagent ( Invitrogen ) following manufacturer’s instructions . Cells were incubated at 27°C until the viral infection was clear ( 3 days post-infection ) and then the P1 viral stock was harvested by centrifugation at 480×g for 5 min at room temperature . The viral titer was determined using absolute quantification with standard curve by qPCR . The optimized viral stock with multiplicity of infection ( MOI ) of 0 . 1 was used to infect 2 . 0×106 Sf9 cells/well , and the supernatant at 72 h post-infection representing the P2 viral stock was collected and used to infect Sf9 cells ( 2 . 0×106 cells/well ) at an optimized high MOI value ( 3–5 ) . Non-infected cells and Sf9 cells infected with either an empty bacmid or a bacmid containing the Arabidopsis thaliana β-glucuronidase ( GUS ) gene ( pFastBac-GUS ) were used as controls . Transfected Sf9 cell pellets were harvested 3 days post-infection , washed three times with PBS buffer ( pH 7 . 4 ) and lysed using the I-PER Insect Cell Protein Extraction Reagent ( Thermo Fisher Scientific Inc . ) plus 1 μg/ml aprotinin ( Sigma ) with gentle agitation at 4°C for 10 min . After centrifugation at 15000×g at 4°C for 15 min , the supernatants containing recombinant proteins were quantified as described above , and stored at -80°C until used . Binding of Cry1Ac to BBMV proteins was tested as described elsewhere [92] in 100 μl ( final volume ) reactions containing 10 nM Cry1Ac toxin and 10 μg P . xylostella BBMV in PBS binding buffer ( PBS , pH 7 . 4 containing 0 . 1% BSA and 0 . 1% Tween-20 ) . After electrophoresis with a constant current of 300 mA at 4°C for 1 h and then incubation with blocking buffer ( PBS , 0 . 1% Tween-20 , 3% BSA ) for 1 h with constant shaking , bound toxin was detected with rabbit anti-Cry1Ac polyclonal antisera ( 1:100000 dilution ) followed by goat anti-rabbit secondary antibody conjugated to horseradish peroxidase ( HRP ) ( 1:5000 dilution , CWBIO ) . The bound Cry1Ac was visualized using the SuperSignal West Pico ( Pierce ) reagent . Relative Cry1Ac binding was quantified using densitometry with the ImageJ v . 1 . 47 software ( http://rsbweb . nih . gov/ij/ ) with intensity in the DBM1Ac-S BBMV sample considered 100% binding . Data presented are the means and standard errors from assays using three independent BBMV experiments per strain . Immunolocalization of Cry1Ac toxin binding to Sf9 cells expressing PxmALP was tested as previously [93] with slight modifications . Transfected cell cultures were incubated in 300 μl of PBS ( pH 7 . 4 ) alone or with 1 U of phosphatidylinositol-specific phospholipase C ( PI-PLC ) ( Invitrogen ) at 4°C for 2 h with gentle agitation , and then washed thrice with PBS and incubated with Cry1Ac toxin ( 100 μg/ml ) at 27°C for 2 h . Cultures were washed and then fixed in ice-cold 4% paraformaldehyde for 15 min . After washing and blocking with 1% BSA for 1 h at room temperature , the cells were probed sequentially with primary rabbit polyclonal anti-Cry1Ac antibody and FITC-conjugated goat anti-rabbit secondary antibody , each with 1:100 dilution and incubation for 1 h at 27°C . Finally , the cells were pipetted onto glass slides , mounted with coverslips and examined immediately under a LSM 700 confocal laser scanning microscope ( Carl Zeiss ) using excitation at 488 nm and 20× objective with additional zooming . Image acquisition of the controls ( Non-infected Sf9 cells and GUS-infected Sf9 cells ) and data processing were performed under the same conditions . Enzyme Linked Immunosorbent Assays ( ELISA ) were performed as described elsewhere [47 , 94] . To test for Cry1Ac binding to PxmALP , 10 nM trypsin-activated Cry1Ac toxin was fixed into ELISA plates ( Costar ) overnight at 4°C , followed by five washes with 200 μl PBST buffer ( PBS , pH 7 . 4; 0 . 05% Tween-20 ) . The plates were then blocked by incubating with 100 μl 1% BSA at 37°C for 1 . 5 h , and washed five times with 200 μl PBST . After incubating with 0 . 5 μg of solubilized Sf9 cell culture proteins transfected with empty bacmid , expressing the GUS protein or PxmALP , bound PxmALP to Cry1Ac was detected using a 1:5000 dilution of anti-His antibody coupled to horseradish peroxidase ( HRP ) and subsequent 1:5000 dilution of anti-mouse antibody ( CWBIO ) . Finally , the plates were incubated with 150 μl TMB ( 3 , 3′ , 5 , 5′-tetramethylbenzidine ) Horseradish Peroxidase Color Development Solution ( Beyotime ) , and the enzymatic reaction was stopped with 50 μl 2M H2SO4 and absorbance values ( OD values ) were read at 450 nm in microplate reader . As controls , wells coated with Cry1Ac but incubated without any of the three expressed proteins and revealed with the same antibodies above . The OD values of controls were all below 0 . 2 and subtracted from the experimental OD values . The experiments were repeated for three times using protein samples from independent batches and each with three replications . Susceptibility to Cry1Ac in transfected Sf9 cell cultures was assessed by counting cells stained by trypan blue using an IX-71 Inverted Microscope ( Olympus ) . Cells at 3-day post-infection were washed twice with PBS and then incubated with 100 μg/ml of Cry1Ac toxin in Sf9 cell medium . After 3 h at 27°C with gentle agitation , cells were washed once with 1 ml of PBS and then resuspended in 1 ml of a 0 . 4% trypan blue solution in PBS . The numbers of live ( unstained ) and dead ( stained blue ) cells in three replicates for each cell type and from three independent transfections were counted in a hemocytometer . Relative percentage mortalities were calculated using the total cell number in each replicate . Mortality data from diverse treatments were tested for significant differences using two-way analysis of variance ( ANOVA ) and Holm-Sidak’s multiple pairwise comparison tests ( overall significance level = 0 . 05 ) . Toxin induction assays of the PxMAP4K4 gene were performed using third instar larvae from the susceptible strain DBM1Ac-S and the near-isogenic resistant strain NIL-R . We selected the DBM1Ac-S or NIL-R larvae with 1 or 3500 μg/ml Cry1Ac protoxin ( respective LC50 value in each strain ) for 72h as the leaf-dip method used in bioassay , the unselected larvae from both strains were used as control groups . After 72h , larval midguts were dissected from survivors , and subsequent total RNA extraction , cDNA synthesis , qPCR analysis of the PxMAP4K4 gene expression were as described above . Three independent experiments were conducted , and one-way ANOVA with Holm-Sidak’s tests ( overall significance level = 0 . 05 ) were used to determine the significant statistical difference between control and treatment groups . The expression of PxmALP , PxABCC2 , PxABCC3 and PxMAP4K4 genes was silenced using injection of dsRNA in early 3rd instar P . xylostella larvae . Specific primers containing a T7 promoter sequence at the 5′ end to generate dsRNA targeting PxmALP ( GenBank accession no . KC841472 ) and mALP1 from H . armigera ( GenBank accession no . EU729322 . 1 ) , or EGFP ( GenBank accession no . KC896843 ) were designed using the SnapDragon tool ( http://www . flyrnai . org/cgi-bin/RNAi_find_primers . pl ) . Primers to generate dsRNA to PxABCC2 ( GenBank accession no . KM245561 ) and PxABCC3 ( GenBank accession no . KM245562 ) were designed to the specific transmembrane region lacking alternative splicing and not in the intergenic conserved nucleotide binding domain ( NBD ) to avoid potential off-target effects . Primers for dsRNA of PxMAP4K4 ( GenBank accession no . KM507871 ) were designed to the constant C-terminal CNH domain region lacking alternative splicing ( S11 Fig ) . After amplification from P . xylostella or H . armigera total larval midgut RNA and confirmation by sequencing , the amplicons ( 438 bp for dsPxmALP , 538 bp for dsHamALP1 , 469 bp for dsEGFP , 603 bp for dsPxABCC2 , 531 bp for dsPxABCC3 , and 582 bp for dsPxMAP4K4 ) were used as template for in vitro transcription reactions to generate dsRNAs using the T7 Ribomax Express RNAi System ( Promega ) . The synthesized dsRNAs were suspended in injection buffer ( 10 mM Tris–HCl , pH 7 . 0; 1 mM EDTA ) , and then they were subjected to 1% agarose gel electrophoresis and quantified spectrophotometrically prior to microinjection . To increase dsRNA stability and facilitate dsRNA delivery , injection was carried out with a 1:1 volume ratio of Metafectene PRO transfection reagent ( Biontex ) after incubation for 20 min at 25°C . A combinatorial RNAi approach involving simultaneous knockdown of PxmALP , PxABCC2 and PxABCC3 genes was performed by mixing equal amounts ( 300 ng each ) of the corresponding dsRNAs for microinjection of larvae from the susceptible DBM1Ac-S strain . In contrast , silencing of PxMAP4K4 was performed in the near-isogenic Cry1Ac resistant strain NIL-R displaying increased PxMAP4K4 expression . None of the larvae were exposed to Cry1Ac toxin before dsRNA microinjection to avoid detection of transcriptome changes due to exposure to the toxin . Optimal time to detect silencing and dsRNA amounts were optimized for PxmALP in preliminary experiments ( S4 Fig ) . Microinjection was carried out under a SZX10 microscope ( Olympus ) . The volume of sample microinjected into each larvae was determined to result in <20% larval mortality 5 days post-injection . The Nanoliter 2000 microinjection system ( World Precision Instruments Inc . ) with sterilized fine glass capillary microinjection needles pulled by P-97 micropipette puller ( Sutter Instrument ) were used to deliver 70 nanoliters of injection buffer ( containing Metafectene PRO solution ) or dsRNAs ( 300 ng ) into the hemocoel of early 3rd instar DBM1Ac-S or NIL-R P . xylostella larvae . Larvae were starved for 6 h and anesthetized for 30 min on ice before microinjection . More than twenty or fifty larvae were injected for each treatment and three independent experiments performed . Injected larvae were allowed to recover for about 3 h at room temperature and then returned to normal rearing conditions for the subsequent qPCR assays to determine gene silencing and bioassays . Effectiveness of RNAi was tested by qPCR 0–120 h post-injection using cDNA prepared from isolated total midgut RNA as described above . Leaf-dip bioassays were performed for 72 h using larvae at 48 h after dsRNA injection and Cry1Ac protoxin concentrations representing approximately the LC50 ( 1 μg/ml ) and LC90 ( 2 μg/ml ) values for non-injected DBM1Ac-S larvae and LC10 ( 1000 μg/ml ) values for non-injected NIL-R larvae . Bioassays were performed with forty larvae per RNAi treatment and toxin concentration , and each bioassay replicated three times . Mortality in control treatments was below 5% and bioassay data processing was as described above . One-way or two-way ANOVA with Holm-Sidak’s tests ( overall significance level = 0 . 05 ) were used to determine the significant statistical difference between qPCR and bioassay treatments , respectively . Effects of RNAi on fitness costs were analyzed by comparing biological parameters , including pupation percentage , pupal weight , pupation duration and eclosion percentage . Larvae injected with buffer containing Metafectene PRO transfection reagent were used as a negative control . All the larvae used in the test were fed on fresh cabbage leaves without exposure to Cry1Ac toxin . Each treatment was replicated three times with ten larvae per replicate . Least squared difference ( LSD ) tests ( overall significance level = 0 . 05 ) were used to determine statistical significance of differences in biological parameters between control and treated groups . The near-isogenic NIL-R ( resistant ) and DBM1Ac-S ( susceptible ) strains were used for genetic linkage analysis as described elsewhere [24] . A single-pair cross was prepared between a male from the NIL-R and a female from the DBM1Ac-S strain to generate an F1 progeny . A diagnostic Cry1Ac toxin dose killing 100% of the F1 ( heterozygous ) larvae was determined in bioassays as described above . Reciprocal crosses between an F1 and NIL-R moths were made to generate two backcross families ( S8 Fig ) . The progenies from each backcross family ( total of 40 larvae per family ) were reared on control ( cabbage ) or experimental ( cabbage with 20 μg/ml of Cry1Ac toxin ) diets . Purified RNA from single backcross family individuals surviving no treatment or exposure to Cry1Ac was used for cDNA synthesis . Linkage between the existence of multiple PxABCC gene isoforms or differential alteration of PxmALP and PxABCC gene expression and resistance to Cry1Ac was tested using PCR amplification and qPCR conditions as described above .
|
Biopesticide and transgenic crops based on Bacillus thuringiensis ( Bt ) Cry toxins are widely used worldwide , yet the development of field resistance seriously threatens their sustainability . Unraveling these resistance mechanisms are of great importance for delaying insect field resistance evolution . The diamondback moth was the first insect to evolve field resistance to Bt biopesticides and it is an excellent model for the study of Bt resistance mechanisms . In this work , we present strong empirical evidence supporting that ( 1 ) field-evolved resistance to Bt in P . xylostella is tightly associated with differential expression of a membrane-bound alkaline phosphatase ( ALP ) and a suite of ATP-binding cassette transporter subfamily C ( ABCC ) genes , and ( 2 ) a constitutively transcriptionally-activated upstream gene ( MAP4K4 ) in the MAPK signaling pathway is responsible for this trans-regulatory signaling mechanism . These findings identify key resistance genes and provide the first comprehensive mechanistic description responsible for the field-evolved Bt resistance in P . xylostella . Given that expression alterations of multiple receptor genes result in Bt resistance in many other insects , it can now be tested to determine whether the previously unidentified trans-regulatory mechanism characterized in this study is also involved in these cases .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
MAPK Signaling Pathway Alters Expression of Midgut ALP and ABCC Genes and Causes Resistance to Bacillus thuringiensis Cry1Ac Toxin in Diamondback Moth
|
Soil-transmitted helminths colonize more than 1 . 5 billion people worldwide , yet little is known about how they interact with bacterial communities in the gut microbiota . Differences in the gut microbiota between individuals living in developed and developing countries may be partly due to the presence of helminths , since they predominantly infect individuals from developing countries , such as the indigenous communities in Malaysia we examine in this work . We compared the composition and diversity of bacterial communities from the fecal microbiota of 51 people from two villages in Malaysia , of which 36 ( 70 . 6% ) were infected by helminths . The 16S rRNA V4 region was sequenced at an average of nineteen thousand sequences per samples . Helminth-colonized individuals had greater species richness and number of observed OTUs with enrichment of Paraprevotellaceae , especially with Trichuris infection . We developed a new approach of combining centered log-ratio ( clr ) transformation for OTU relative abundances with sparse Partial Least Squares Discriminant Analysis ( sPLS-DA ) to enable more robust predictions of OTU interrelationships . These results suggest that helminths may have an impact on the diversity , bacterial community structure and function of the gut microbiota .
Today , approximately a quarter of the world's population carries soil-transmitted helminths [1] . In an endemic tropical environment , more than 70% of the population may be infected with helminths [2] , which is indicative of our helminth exposure during the course of human evolution . Just as the commensal microbiota has coevolved with mammalian hosts [3] , they must also have coevolved with helminths within their mutual hosts [4] . However , there is little understanding of how the presence of helminths may affect the microbial ecology of the human gut . While the healthy human microbiota from individuals of the developed world is well-characterized [5] , residents of developing countries harbor very different bacterial communities on the skin [6] , and in the gut [7] , with an increased abundance of the genus Prevotella [8] . While diet has been assumed to contribute most significantly to the differences of the intestinal tract , it is possible that helminth infections also may have a substantial impact on the human microbiome [9] . Alterations in the human microbiome have been associated with a range of conditions in the developed world , including inflammatory bowel disease ( IBD ) [10]–[12] , obesity [13]–[15] autism [9] , type 2 diabetes [16] and allergies [17]–[19] . These diseased states have often been linked with decreased diversity of the microbiota , perhaps because healthy species-rich communities are more stable and resistant to pathogenic invasions [20] . Acute gastrointestinal infections also are associated with reductions in microbial diversity and increased volatility of microbial communities [21] . In contrast , helminth exposure has been shown to restore microbial diversity to macaques suffering from idiopathic chronic diarrhea [22] . While recent studies have begun to establish changes to microbial communities in response to intestinal helminths [4] , [23] , most studies to date have been on animal models and not on human subjects . As Malaysia experiences rapid development , helminth infections have generally decreased , however they are still highly prevalent among the Malaysian indigenous populations [24] , who live in the rural and semi-urban areas of Peninsular Malaysia . The aim of our study was to compare the gut microbiota of helminth-infected individuals within this community , with individuals who were not colonized by helminths . Based on our previous studies with macaques experimentally infected with helminths [22] , we hypothesized that helminth infected individuals may have increased microbial diversity relative to uninfected individuals . Genomic surveys of the microbiota require the development of new approaches to examine the interactions of microbes with each other and their hosts . The abundance of each microbial operational taxonomic unit ( OTU – i . e . the best known taxonomic specification of a population as measured by high throughput sequencing of the bacterial 16S gene ) reflects the product of the total microbial community size times the relative fitness of that OTU's population in the measured community/environment . Ecological relationships should be detectable via similarity measures ( e . g . - correlations ) between microbial abundances and host variables ( e . g . helminth infection status ) . However , compositional artifacts are a major confounder hindering functional microbial inference with standard statistical approaches and requiring the development of new methods . These compositional artifacts are a result of the fact that community total size varies and that ‘components’/OTU abundances are measured as relative proportions of this unknown total community size ( for example in this study OTU abundances for each sample must sum to 1 in a typical microbiome datasets in order to normalize for sampling or sequencing depth ) . Another common problem with the analysis of these datasets is that the number of variables of interest is usually much larger than the number of available samples , and thus we are typically undersampled with respect to analysis that aim to detect relationships between microbes in complex communities . In order to address these two problems for our dataset , we devised a strategy to combine centered log-ratio ( clr ) transformation [25] for OTU relative abundances with sparse Partial Least Squares Discriminant Analysis ( sPLS-DA ) [26] in order to identify OTU features that are more predictive of helminth infection status . This approach is general and could also be easily applied to other microbiome studies in the future . R source code for our analysis can be found at ( bonneaulab . bio . nyu/ ) . Sequencing 16S ribosomal RNA may reveal alterations to taxonomic groups and species composition but it does not provide data on metabolic activity and function of the microbial communities that may be altered by helminth colonization . Microbial function in communities can be assessed using shotgun metagenomics [7] , [27] , but this approach is expensive and challenging to analyze . PICRUSt is an approach for inferring the metagenome of the closest available whole genome sequences using 16S gene sequence profiles [28] , [29] , hence it provides a way to predict changes to microbial function likely to be associated with changes in OTU-abundance detected via 16S-sequencing . In this study , we used PICRUSt [28] to investigate functional differences in the microbiota of helminth colonized individuals .
The Medical Ethics Committee of University Malaya Medical Center ( UMMC ) , Kuala Lumpur , Malaysia under the MEC reference number 943 . 14 , approved the study on indigenous Malaysians . All the adult participants provided written informed consent . Written informed consent was obtained from legal representatives of the children that participated in the study . New York Harbor Veterans Affairs Hospital Institutional Review Board approval was obtained before enrolling subjects in the study in New York . Written informed consent was received from all the participants . The 51 volunteers ( ages 0 . 4–48 years ) from the indigenous community in Malaysia , also known as Orang Asli in local terms , lived in two villages of Kampung Ulu Kelaka ( N 3° 6′ 0″ , E 102° 4′ 1″ ) ( n = 22 ) and Kampung Dusun Kubur ( N 2° 56′ 22 . 6″ , E 102° 4′ 32 . 3″ ) ( n = 29 ) that were located in the Jelebu district , Negeri Sembilan state , approximately 80 km east of Kuala Lumpur , capital of Malaysia . Both communities were of the Temuan ethnic subgroup , with the majority of the villagers living below the poverty line ( i . e . , RM 800 or USD 256 per month ) . The main economic activities are rubber tapping , farming and collecting forest products . Most of the houses in Kampung Ulu Kelaka were built under the government-coordinated Housing Aid Program ( PBR ) and were equipped with basic amenities such as piped water , toilets and had electric supply . However , in Kampung Dusun Kubur , the villagers were still living in wooden or partial brick houses , without piped water and toilets . While increasingly some of the younger villagers have been able to complete high school education , most of the adult villagers from these two communities have only attended elementary school . Sample collection was carried out in September 2012 . The research subjects from New York were recruited at the Manhattan Veteran Affairs Hospital and were all over the age of 50 . They were at an average risk for colon cancer and presented for screening colonoscopy . The median age was 60 . 5 . Ethnicity was 30% ( n = 6 ) Caucasian , 25% ( n = 5 ) African American , 30% ( n = 6 ) Hispanic and 15% ( n = 3 ) Other . Stool samples with 2 . 5% potassium dichromate as a preservative were collected in screw-capped containers and stored in 4°C for several days prior to DNA extraction . DNA was extracted from the 51 fecal samples using MACHEREY-NAGEL NucleoSpin Soil kit ( MACHEREY-NAGEL GmbH & Co . KG , Düren , Germany ) . The SL 2 buffer was used with 150 µl Enhancer SX in the sample lysis step . DNA was eluted in 50 µl of Buffer SE and stored at −20°C . The formalin-ether concentration technique was used for microscopic examination . One to two gram ( s ) of stool sample was mixed with 7 ml of 10% formalin and 3 ml of ethyl acetate in 15 ml falcon tubes . The samples were then centrifuged at 2500 rpm for 5 min . Fecal smears were made , stained with 0 . 85% iodine and observed by light microscope under the magnification of 100× for helminths ( Trichuris spp . , Ascaris spp . , and hookworm ) . Each subject was scored for presence or absence of each of these three helminth species . V4 rRNA paired-end sequencing was performed on the DNA from the fecal samples collected using the protocol modified from Caporaso et al [30] . The forward primer construct ( 5′- AAT GAT ACG GCG ACC ACC GAG ATC TAC ACT ATG GTA ATT GTG TGC CAG CMG CCG CGG TAA -3′ ) contained a 5′ Illumina adapter , a forward primer pad , the 515F primer and a two-base linker sequence ( ‘GT’ ) . The reverse primer ( 5′- CAA GCA GAA GAC GGC ATA CGA GAT NNN NNN NNN NNN AGT CAG TCA GCC GGA CTA CHV GGG TWT CTA AT -3′ ) contained the 3′ Illumina adapter , a unique 12-base error-correcting Golay barcode , the reverse primer pad , a two-base linker sequence ( ‘CC’ ) and the 806R primer . Polymerase chain reaction ( PCR ) was carried out in triplicate using the Bio-Rad CFX 96 system ( Bio-Rad , Hercules CA , USA ) . The PCR mix contained 0 . 2 µM forward and reverse primers , 1 µl template DNA , 10 µl 5 Prime Hot Master Mix ( 5 PRIME , Gaithersburg MD , USA ) and 12 µl of MoBio PCR certified water ( MO BIO Laboratories , Calsbad CA , USA ) . Thermal cycling consisted of 94°C for 3 min , followed by 35 cycles of 94°C for 45 s , 50°C for 60 s and 72°C for 90 s , with a final extension of 10 min at 72°C to confirm full amplification . Replicate amplicons were pooled and the DNA concentrations were determined using the Quant-iT PicoGreen dsDNA reagent and kit ( Invitrogen , Grand Island NY , USA ) based on the manufacturer's instructions . Fluorescence was measured on the Perkin-Elmer Victor Plate reader using the 490/535 nm excitation/emission filter pair with measurement time 0 . 1 s . The amplicons were then pooled in equimolar ratios and purified using QIAquick PCR purification kit ( Qiagen Inc , Chatsworth , CA , USA ) . The final concentration of cleaned DNA amplicon was determined using the Qubit PicoGreen dsDNA BR assay kit ( Invitrogen , Grand Island , NY , USA ) . Amplicon sequencing was performed on the Illumina MiSeq system ( Illumina , San Diego CA , USA ) . We first improved the joining potential of the paired-end raw sequences by trimming low quality bases at the overlapping ends of reads . We used EA-utils [31] , [32] to trim and join reads by iterating over a range of Phred quality scores ( 1–20 ) as the threshold for terminal base removal . We then joined paired-end reads at each quality score , only accepting the sequence set that resulted in the maximum number of joined sequences and then removed unjoined barcodes . Finally , we proceeded with processing of sequence reads using the Quantitative Insights Into Microbial Ecology ( QIIME ) software package [33] to filter sequence read quality ( minimum quality score of 25 , minimum/maximum length of 200/1000 , no ambiguous bases allowed and no mismatches allowed in the primer sequence ) and to split multiplexed libraries . To identify and quantitate abundances of Operational Taxonomic Units ( OTUs ) from the sequence data , we used a combination of reference-based and de novo sequence clustering ( pick_subsampled_reference_otus_through_otu_table . py ) . For closed reference-based picking , we aligned sequences to the Greengenes 12_10 reference collection ( available at http://greengenes . secondgenome . com/downloads ) . 0 . 1% of the sequences that failed to align to the reference were randomly subsampled and clustered de novo using UCLUST [34] , with an OTU cluster defined at a sequence similarity of 97% . The centroid sequence of each cluster was chosen as the new reference set for another round of closed-reference OTU picking . OTU assignments for reads that failed to align to this reference collection were picked by another round of de novo clustering . We performed all closed-reference picking by dividing the task into 20 jobs and ran the alignment in parallel on a high performance computing cluster environment . For de novo OTU clusters , representative sequences were picked for taxonomic identity assignment using the Ribosomal Database Project ( RDP ) classifier [35] , [36] . The taxonomy assignment for each sequence was truncated at the most specific taxonomic level with a confidence score of at least 0 . 8 . The PyNAST alignment algorithm [37] was used to align the OTU representative sequences against the Greengenes core database set with a minimum alignment length of 189 and a minimum identity of 75% and FastTree [38] was used to construct a phylogenetic tree . We then generated a final OTU table for downstream analysis by excluding the sequences that had failed to align by PyNAST . To compare the sequencing data from the Malaysian and New York samples , we preprocessed , quality-filtered and split the libraries of the sequence reads from the two separate MiSeq runs , independently . The files containing the resulting sequence data were concatenated into a single file and a new mapping file with combined sample data were taken together for OTU picking and downstream analysis . Samples were evaluated for beta diversity ( community diversity divergence between samples ) and alpha diversity ( microbial diversity within samples ) calculations in QIIME . Beta diversity was calculated using the QIIME default beta diversity metrics for weighted and unweighted UniFrac distances on both uneven and evenly subsampled OTU tables . UniFrac is a measure of the amount of evolutionary history that is unique between samples of at least two different environments [39] , [40] . Unweighted UniFrac makes comparisons based on the presence and absence of members , while the weighted version also incorporates abundance information . To identify environments that could drive groupings of similar communities , principle coordinate analysis ( PCoA ) was performed on the UniFrac distance matrices generated from beta diversity calculation and the resulting PCoA plots were visualized using the KiNG graphics program ( http://kinemage . biochem . duke . edu/index . php ) . Environments producing distinct clustering of samples were subjected to significance testing using the non-parametric statistical analysis ANOSIM via QIIME . Alpha rarefaction was performed using the phylogenetic distance [41] and Shannon index [42] metrics . We rarefied OTU tables so that all sample sizes matched the minimum sampling depth , then randomly subsampled sequences over a range of specified depths and calculated the alpha diversity for each sample at each point , with 10 independent iterations at each depth . We assessed statistical significance of between-group alpha diversity metrics by two-sample t-test as implemented in QIIME . To identify taxa with differentiating abundance in the different environments , the LDA Effect Size ( LEfSe ) algorithm was used with the online interface Galaxy ( http://huttenhower . sph . harvard . edu/galaxy/root ) . Helminth infection status and country were assigned as the respective comparison classes in two separate analyses ( one comparing countries and one comparing helminth status within Malaysia . LEfSe first identified features that were statistically different among the different biological classes . It then performed non-parametric factorial Kruskal-Wallis ( KW ) sum-rank test and Linear Discriminant Analysis ( LDA ) to determine whether these features are consistent with respect to the expected behavior of the different biological classes [43] . To generate better predictive models of Helminth-microbiome component interactions , we used centered log-ratio ( clr ) transformations of the relative abundance data ( to circumvent the compositional bias problem ) , in combination with sparse Partial Least Squares Discriminant Analysis ( sPLS-DA ) ( to address the “noise” problem ) , in order to generate more compositionally robust , predictive models of OTU features . Partial Least Squares discriminant analysis ( PLSDA ) is a technique to predict a discrete ‘response’ ( e . g . - parasite infection status ) by regressing on many sample features ( e . g . OTU abundances ) , and finding the set of latent , orthogonal factors that maximizes the covariance between predictors and response variables [26] . It has been used previously on microbiome relative abundances , for example , to identify OTUs associated with antibiotic treatment [44] . However , PLSDA , along with other statistical methods , suffers from problems of compositional artifacts due to unit-sum constraint of relative abundance data . To account for these compositional artifacts , we first transformed relative abundances using the centered log-ratio ( clr ) transformation . This maps compositional data to a corresponding Euclidean space by dividing each component by the geometric mean of all the components in a sample and then taking the log of that ratio [25] . Typically , this transformation results in data singularity , which is a major disadvantage for covariance-based techniques . However , this is not a problem for PLSDA , which can efficiently deal with multi-colinearity that can result from the clr transformation [45] , [46] . Using a combination of custom R scripts and the caret and spls packages [47] , [48] , we clr transformed the data and filtered the OTU table to the set of highly variable OTUs . We performed feature selection using sparse PLS-DA and 5-fold cross validation to tune algorithm parameters ( sparsity and number of latent components ) and to check model validity . To assess the statistical significance of a feature's contribution to the model , we obtained bootstrapped and null ( by randomly permuting data ) estimates of PLS-DA coefficients and report the p-value of an OTU as the fraction of bootstrapped coefficients lying at or above the tails of the null distribution . We visualized the resulting models by projecting the clr-transformed data points onto the PLS loadings of sparse PLS-DA and permutation-selected features and visualize these projections as biplots . Code used to perform this analysis is available at bonneaulab . bio . nyu . edu . The demultiplexed data files from the processed sequencing reads were subjected to another round of closed-reference OTU picking for use in PICRUSt . This was done by aligning the sequences against the newest Greengenes reference OTUs ( downloaded from http://greengenes . secondgenome . com/downloads/database/13_5 ) and OTUs were assigned at 97% identity . The resulting OTU table was then used for microbial community metagenome prediction with PICRUSt on the online Galaxy interface . PICRUSt was used to derive relative Kyoto Encyclopedia of Genes and Genomes ( KEGG ) Pathway abundance [28] . Supervised analysis was done using LEfSe to elicit the microbial functional pathways that were differentially expressed in individuals with different helminth infection status .
We compared the stool microbiota of 51 residents from two villages in Malaysia collected on a single field trip ( Table 1 ) . The majority ( n = 36; 70 . 6% ) of the 51 fecal samples collected from the Malaysian indigenous communities contained helminths , with Trichuris spp . ( n = 28; 54 . 9% ) being the most common , followed by Ascaris spp . ( n = 21; 41 . 2% ) and hookworm ( n = 5; 9 . 8% ) . Hence , individuals were often colonized by more than one helminth ( n = 17; 33 . 3% ) . The V4 region of bacterial 16S rRNA was PCR-amplified and sequenced on the MiSeq ( Illumina ) platform from the 51 Malaysian fecal samples . A total of 931 , 078 quality-filtered sequences were obtained from these samples with an average of 19 , 002±6 , 451 ( SD ) sequences per sample . These reads were clustered into 83 , 457 unique OTUs with an average of 1 , 703 OTUs per subject . The number of observed OTUs was similar across the different age groups ranging from 1000–4500 ( Figure 1A ) . Averaging across samples , the most abundant phyla were Firmicutes ( 55 . 9% ) , Bacteroidetes ( 23 . 5% ) , and Proteobacteria ( 10 . 1% ) . This pattern was largely similar across the individual microbiota , with the exception that in two younger subjects , age 5 months and 3 years old , in whom we found unusually high abundance of Actinobacteria ( 35 . 23 and 55 . 49% ) , which was resolved to the genus Bifidobacterium ( Figure 1B ) . When subjects were classified based on helminth infection status , the PCoA plot generated from the unweighted UniFrac distance matrices on an uneven OTU table suggested clustering of helminth-positive subjects along the first principle coordinate ( PC1 ) , representing 8 . 17% of intersample variance ( Figure 2A ) . This difference in bacterial communities was significant , as determined using the non-parametric statistical test analysis of similarity ( ANOSIM ) , where R = 0 . 18 ( p = 0 . 04 ) . Alpha diversity analysis was performed on samples after rarefaction to 1139 sequences/sample ( minimum sampling depth ) . Rarefaction curves generated for the Phylogenetic Distance and Shannon index showed that the helminth-positive subjects demonstrated greater diversity than the helminth-negative subjects ( Figures 2B and 2C ) . Two-sample t-test performed on both metrics showed that the helminth-positive subjects had significantly greater species richness ( total number of species present ) than the helminth-negative subjects ( p = 0 . 04 ) , but the difference in evenness ( number of organisms per species ) between both groups were insignificant ( p = 0 . 3 ) . The number of observed OTUs was also significantly different ( p = 0 . 0326 ) between the helminth positive and negative subjects ( Figure 2D ) . These results suggest that helminth colonization significantly affects the gut microbiota and may increase the species diversity of the bacterial communities . We next performed a supervised comparison of the microbiota between helminth-positive and helminth-negative subjects by utilizing the LEfSe algorithm to identify taxonomic differences associated with helminth infection status ( Figure 3 ) . We used a logarithmic LDA score cutoff of 3 . 0 to identify important taxonomic differences between infected and uninfected individuals . This analysis revealed that helminth-infected individuals have increased abundance sequences representing Paraprevotellaceae , Mollicutes , Bacteroidales , and Alphaproteobacteria . Helminth-negative subjects had an increased abundance of Bifidobacterium . While LEfSe provides an LDA score for class comparisons in order to identify bacterial components that are different between helminth positive and negative individuals through the Kruskal-Wallis ( KW ) sum-rank test , it may be subject to compositional artifacts and noise common to microbiome datasets . To address these problems , we developed a method for combining centered log-ratio ( clr ) transformations of the relative abundance data ( to circumvent the compositional bias problem ) , in combination with sparse ( to ensure selection of relatively few OTUs ) Partial Least Squares Discriminant Analysis ( sPLS-DA ) to generate predictive models of OTU features ( see Methods ) . We first used this approach to identify OTUs associated with overall helminth infection status ( Figure 4A ) , and also with the subset of subjects infected with either Trichuris alone ( Figure 4B ) or Ascaris alone ( Figure 4C ) . In our implementation , we perform cross validation analysis to select algorithm parameters , and generate bootstrapped ( n = 1000 ) estimates of model coefficients , selecting only OTUs with statistically significant pseudo p-values ( alpha = 0 . 05 ) . These models were then visualized using biplots of the first two PLS discriminant components ( Figure 4 ) using the ggplot2 package [49] . With helminth colonization as a whole , there was an unclassified Bacteroidales associated with infection . However , other families ( Lachnospiraceae and Prevotellaceae ) had OTUs that are both positively and negatively associated with colonization , which was difficult to interpret . What was more interesting was that Trichuris alone was strongly associated with Paraprevotellaceae , consistent with the LEfSe analysis , indicating that Trichuris colonization maybe the driving force for this association . We next utilized inferred metagenomics by PICRUSt [28] to investigate functional differences in the microbiota of the 51 individuals . As noted above , the relative abundance of different bacterial taxa among the villagers varied considerably between individuals ( Figure 1B ) . However , when we assessed the microbial metabolic and functional KEGG pathways of these communities by inferred metagenomics using PICRUSt [28] , [29] , the pathways were more evenly distributed and consistent between individuals ( Figure 5A ) . This stability of metabolic pathways despite variability of microbial taxa was also observed in the Human Microbiome Project [27] . To identify microbial functional pathways that may be altered by helminth infection , we performed supervised comparisons with LEfSe . The gut microbiota of helminth positive individuals have functional pathways more abundant for genetic information processing , particularly pathways for translation , replication and repair ( Figure 5B ) . The pathways for nucleotide metabolism , as well as pathways for cell growth and death were also of significantly higher abundance in the microbiota of helminth positive individuals ( Figure 5B ) . This effect is largely due to the presence of Trichuris worms , since many of the same functional pathways are identified when comparing individuals colonized with Trichuris alone with individuals that are helminth negative ( Figure 5C ) . At a higher KEGG pathway hierarchy level , there were differences in microbial metabolic pathways between the helminth positive and negative individuals ( Figure S1 ) . Carbohydrate and xenobiotic metabolism pathways were enriched in the microbiota of helminth negative individuals , while microbiota of helminth positive individuals encoded increased abundance of metabolic pathways involving nucleotides , amino acids , terpenoids and polyketides , novobiocin biosynthesis and for vitamins and co-factors involving one carbon pool by folate . The reduced usage of carbohydrate metabolic pathways is likely to be driven by the presence of Ascaris worms , since this pathway was also identified as differentially enriched in helminth negative individuals when performing a comparison with individuals colonized only by Ascaris ( Figure 5D ) . To validate our sampling , sequencing and analysis approaches , we wanted to replicate findings from previous studies describing microbiota differences between developing and developed countries . We compared the microbiota of the 51 individuals from these Malaysian indigenous communities with individuals living in a westernized environment such as the USA . Nineteen fecal samples from healthy adult men in New York City previously analyzed with the same MiSeq sequencing platform was used as this reference group . The combined data set of 51 fecal samples from the Malaysian indigenous communities and 19 fecal samples from New York gave a total of 1 , 029 , 891 sequences and 92 , 701 OTUs . By examining unweighted UniFrac distance matrices on an uneven OTU table , these differences translated to distinct clusters visualized on a PCoA plot ( Figure 6A ) . The separation between Malaysian and US samples was best seen along PC1 , which captured 13 . 94% of intersample variance . ANOSIM showed that the microbial community differences between the Malaysian and New York subjects were highly significant with R = 0 . 69 ( p = 0 . 001 ) . This was a much larger difference than the difference between helminth infected and non-infected people in Malaysia . When we examined alpha diversity ( Figures 6B and 6C ) , rarefaction performed on OTU table rarefied to a minimum sampling depth of 1450sequences/sample showed that the stool microbiota of the Malaysian subjects were more diverse than the stool microbiota of the New York subjects . Two-sample t-tests performed for both metrics showed that the differences in microbiota diversity between the Malaysian and New York subjects were highly significant ( p = 0 . 001 , p = 0 . 004 ) . A supervised comparison using LEfSe was then performed to statistically define ( at log LDA threshold of 3 . 50 ) the particular differences in microbial composition between Malaysian and US samples . This confirmed that bacteria from the phylum Firmicutes was more abundant in the New York individuals and the phyla Cyanobacteria , Actinobacteria , Tenericutes and Proteobacteria were more abundant in Malaysian individuals ( Figures S2 and S3 ) . More specifically , Erysipelotrichi , Ruminococcus , Bacteroides and Blautia were more abundant for the individuals from New York , while Mollicutes , Gammaproteobacteria , Faecalibacterium , Prevotella and Ruminococcaceae were more abundant among the Malaysian indigenous people ( Figure 6D ) . We then utilized inferred metagenomics by PICRUSt to compare microbial functions between the indigenous Malaysians and New Yorkers ( Figure 7 ) . Biological pathways encoded were relatively stable ( Figure 7A ) , but when we utilized LEfSe to identify specific differences ( Figure 7B ) , it was interesting to note that many of the pathways associated with helminth colonization ( e . g . Genetic information processing , replication and repair , and translation ) were also enriched in the indigenous Malaysians vs . New Yorkers comparison ( Figure 7B ) . These results suggest that helminth colonization may contribute towards the large differences observed between individuals that live in developing countries and industrialized countries .
In this study , we found that helminth colonization was associated with a significant effect on the gut microbiota of Malaysian indigenous people , with increased diversity and increased abundance of a particular Paraprevotellacae , which may be driven by Trichuris infection . However , these differences were much smaller than the differences observed between the Malaysian indigenous people and residents of New York City . This finding is not surprising , since the US residents had much lower bacterial diversity and evenness when compared to the rural Malaysian population , consistent with previous studies comparing rural residents of developing countries with urban residents of developed countries ( e . g . Bangladesh vs . US [50] , Burkina Faso vs . Italy [8] , Amazonas and Malawians vs . US [7] ) . Although the samples from New York that we used to compare with the Malaysians were processed differently did not come from age-matched individuals , they enabled a basis for validating our sampling and sequencing approach . All of these data are consistent with the hypothesis that socioeconomic development is associated with the disappearance of the ancestral microbiota [19] , [51] . To our knowledge , this is the first time that the gut microbiota of rural indigenous Malaysians has been investigated . As elsewhere , the phyla Firmicutes and Bacteroidetes dominated the microbiota of most of the fecal samples collected from these Malaysian indigenous communities . Taxa abundances across the different age groups was largely similar , consistent with studies showing that the phylogenetic composition of the human intestinal microbiota evolved to resemble an adult-like configuration within the first three years of life [7] . Since among the 51 subjects from whom we collected the fecal samples , only five subjects were three years old or younger , data was limited for this early period . Nevertheless , the unusually high abundance of Bifidobacterium in two of the young subjects is consistent with prior observations of its enrichment in the infant gut [7] , [8] . In animal models , the nematode Heligmosomoides polygyrus was found to alter the gut microbiota of healthy mice , with increased Lactobacillaceae after infection [52] . Trichuris suis infection also has an effect on the intestinal microbiota of pigs , with an increase of the mucus colonizing Mucispirillum bacteria , and shotgun sequencing showed reduce carbohydrate metabolism , with significant reductions in cellulolytic Ruminococcus [53] . There were also significant differences in microbial composition between pigs able to clear adult worms compared to those remaining colonized [36] . Pigs with heavy parasite burden were associated with greater Campylobacter abundance [36] . In cattle , infection with the nematode parasite Ostertagia ostertagi , which infects the abomasum and may be less sensitive to the mucosal immune responses elicited by these helminths , led to minimal microbiota changes [54] . Helminth infection of macaques with colitis can reverse dysbiosis , restoring communities to resemble those in healthy macaques [22] . Since helminths co-exist in the intestinal tract with the gut microbiota , significant interactions between these organisms are not surprising . Trichuris muris eggs utilize the cecal microbiota as environmental cues to enable hatching and the exit of the larvae [55] . Fewer Heligmosomoides polygyrus adult worms were recovered from germ-free than conventional mice during infection , associated with increased eosinophilia , granulomata , and thickening of the small intestinal wall [56] , indicating that the commensal bacteria may reduce the inflammatory response against the worms , and that H . polygyrus requires them to develop appropriately . In summary , there appears to be significant cross-talk between helminths and the gut microbiota . The increase in bacterial alpha-diversity among helminth-infected individuals could be important , because higher microbiota diversity has generally been associated with better health [57] . This concept of ecosystem stability is supported by both animal models and clinical studies [22] , [50] , [58] . However , we did not observe differences in community evenness ( Figure 2C ) . The helminth effects on bacterial diversity but not evenness may reflect our small study size . From an ecological perspective , microbial evenness may correlate with community function after stressor induced perturbations [59] . Macaques with colitis had reduced diversity of the mucosal microbiota that was increased following therapeutic helminth infection [22] . However , it is important to note that a recent study on school children in Ecuador did not observe statistically significant effects of Trichuris trichiura infection on bacterial composition or microbial diversity of the microbiota [60] . More studies in the future will be needed to determine if these discrepancies are a result of geographical and population differences or a result in the approach used in each study . Indeed the differences we observed between helminth-infected and negative individuals were not dramatic , hence a larger study size in our population is needed to confirm these observations . More importantly , longitudinal studies after anti-helminth therapy of infected individuals would provide a more direct test of the impact of helminths on the gut microbiota . We collected samples from an area where the majority of individuals were infected , indicating heavy helminth exposure . The helminth-negative individuals are likely to be continually exposed to helminths , and some may be infected but not detected through our microscopy-based examinations . One possibility is that past helminth exposure may have already altered the gut microbiota , although the subjects may have been negative at the time of study . Although the samples from New York that we used to compare with the Malaysians did not come from age-matched individuals , they enabled a basis for validating our sampling and sequencing approach , replicating prior studies comparing developing country rural communities with developed country urban communities [7] , [8] , [50] . The greater diversity and evenness of the gut microbiota among the Malaysian indigenous community was consistent with prior findings showing that rural Amerindian and Malawian adult populations had significantly more diverse gut microbiota than American adults [7] , [8] , [50] . Bacteria from the phylum Firmicutes , which was more abundant among the New York subjects , has been linked to obesity [15] , [61] and with antibiotic treatment [62] . Findings of increased Prevotella among the Malaysian samples and increased Bacteroides among the New York subjects were also consistent with previous reports on the Prevotella/Bacteroides difference between developed and developing countries . This difference has been attributed to dietary differences , with Bacteroides associated with diets rich in animal protein , several amino acids , and saturated fats ( common in developed countries ) and Prevotella being associated with carbohydrates , simple sugars , and high fiber diet ( common in developing countries ) [8] , [63] . However , it is important to note that storage conditions for stool samples prior to DNA extraction were different for the Malaysian subjects and VA subjects , which could have uncontrolled effects on bacterial diversity . While stool samples from New York were frozen immediately , samples from Malaysia were stored at 4 degrees for up to a week prior to extraction . We have not experimentally determined that storage conditions do not affect bacterial diversity , however all of the samples from the Malaysian subjects were collected on the same day and DNA was also extracted from all of the samples on the same day . Hence , all of the samples were stored for the same period of time and storage conditions should have had the same effect on all the samples with regards to bacterial diversity . Nonetheless , future studies should evaluate freshly frozen stool samples collected in the field . The inference of microbial function by the prediction of bacterial metagenome using PICRUSt added another dimension in characterizing the differences of the microbiota between helminth positive and helminth negative individuals . The differences in gene contents of the intestinal bacteria between helminth positive and helminth negative individuals were largely driven by Trichuris infection . We have previously suggested that Trichuris infection can lead to an increase in mucus production and epithelial cell turnover , consequently reducing the number of bacteria attached to the intestinal wall and restoring the diversity of mucosal bacteria [22] . As such , the higher rate of bacterial cell turnover in the gut of Trichuris positive individuals , as reflected by the increased abundance of genetic information processing and cell cycle pathways , could be an effect of the microbial diversity restoration process associated with ongoing Trichuris infection . However , since these differences were not observed in other studies [60] , it will require further confirmation to better characterize the functional effects of helminth induced microbiota compositional changes . In conclusion , this study provides a preliminary view of the effects of helminth infection on the human gut microbiota in the indigenous communities of Malaysia . Although differences between helminth-positive and negative subjects are not as substantial as those with urban US residents , greater bacterial diversity appears associated with helminth colonization . While we cannot determine if these are causal relationships , these results will help direct future investigations of the relationship between helminths and the gut microbiota in developing countries . Perhaps in the future , these relationships can be exploited therapeutically for the treatment of autoimmune diseases , since helminths already are being studied in clinical trials .
|
Soil-transmitted helminths are carried by large numbers of people in developing countries . These parasites live in the gut and may interact with bacterial communities in the gut , also called the gut microbiota . To determine whether there are alterations to the gut microbiota that are associated with helminth infections , we examined the types of bacteria present in fecal samples from rural Malaysians , many of whom are helminth-positive and find it likely that helminth colonization alters the gut microbiota for rural Malaysians .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"helminth",
"infections",
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"neglected",
"tropical",
"diseases",
"tropical",
"diseases",
"soil-transmitted",
"helminthiases",
"parasitic",
"diseases"
] |
2014
|
Helminth Colonization Is Associated with Increased Diversity of the Gut Microbiota
|
In October 2013 , a locally-acquired case of dengue virus ( DENV ) infection was reported in Western Australia ( WA ) where local dengue transmission has not occurred for over 70 years . Laboratory testing confirmed recent DENV infection and the case demonstrated a clinically compatible illness . The infection was most likely acquired in the Pilbara region in the northwest of WA . Follow up investigations did not detect any other locally-acquired dengue cases or any known dengue vector species in the local region , despite intensive adult and larval mosquito surveillance , both immediately after the case was notified in October 2013 and after the start of the wet season in January 2014 . The mechanism of infection with DENV in this case cannot be confirmed . However , it most likely followed a bite from a single infected mosquito vector that was transiently introduced into the Pilbara region but failed to establish a local breeding population . This case highlights the public health importance of maintaining surveillance efforts to ensure that any incursions of dengue vectors into WA are promptly identified and do not become established , particularly given the large numbers of viraemic dengue fever cases imported into WA by travellers returning from dengue-endemic regions .
Dengue fever is caused by a mosquito-borne flavivirus , dengue virus ( DENV ) , and is responsible for a significant disease burden globally with an estimated 390 million cases of dengue fever per annum , of which 96 million result in symptomatic disease [1] . The incubation period is 3 to 14 days [2] , and clinical manifestations range from mild symptoms , typically including fever , headache , rash , myalgia and arthralgia , to life-threatening haemorrhagic fever and severe shock [3] . In recent decades there has been a resurgence of dengue fever in many tropical and sub-tropical regions following the re-introduction of the primary mosquito vector species Aedes ( Stegomyia ) aegypti ( L . ) [4] . Transmission of DENV in Australia is currently restricted to urban areas of north Queensland and the Torres Strait where Ae . aegypti remains well established and outbreaks regularly occur following importations of dengue viruses by returning travellers infected overseas [5–7] . Recent incursions of Ae . aegypti into towns in the Northern Territory have occurred , but the vectors were eliminated without any dengue activity being detected [8–10] . In WA , outbreaks of dengue were reported in northern parts of the state up until the mid-1940s [6] and Ae . aegypti was recorded as far south as Harvey , approximately 150km south of the capital city , Perth , in the southwest of the State up until the 1950s [6] . However , the distribution of this species receded dramatically following changes in water storage practices and the introduction of scheme water supply , and disappeared completely by the end of the 1960s [6] . Other potential vectors of dengue , such as Aedes ( Stegomyia ) albopictus ( Skuse ) , are not present in WA and no native mosquito species are known to have the potential to transmit DENV . Dengue fever is a notifiable disease in WA , meaning that doctors and laboratories are legally compelled to report cases . In WA , public health units follow up all dengue fever cases to determine where the infections were acquired , if the notifying doctors have not provided the information . The vast majority of cases have been acquired overseas and , in the past five years , these have increased to several hundred cases annually , reflecting both the increase in Western Australians travelling to , and increasing dengue activity in nearby dengue-endemic countries , particularly Bali in Indonesia [11] . The Commonwealth Department of Agriculture also conducts routine surveillance to monitor incursions of exotic mosquitoes in areas surrounding international air and sea ports in WA . In early October 2013 , the Department of Health WA was notified of a laboratory-confirmed case of dengue in a male who had not travelled outside of WA in over a decade . This was the first report of a locally acquired case of dengue fever in the state for more than 70 years .
The case was interviewed to establish his symptomatology , date of onset and travel history in the two weeks prior to symptom onset to determine the most likely location of exposure . The Western Australian Notifiable Infectious Diseases Database ( WANIDD ) was examined to determine if any other locally-acquired cases of dengue had occurred in the area . A serum sample was collected from the patient on 1 October 2013 and referred for confirmation to PathWest Laboratory Medicine WA ( PathWest ) , the state reference laboratory , for testing of dengue antibody titre by flavivirus haemagglutination inhibition ( HI ) , dengue IgM by SD Dengue IgM Capture ELISA and dengue NS1 antigen by Bio-Rad Platelia Dengue NS1 Ag . The dengue serotype was also determined using in-house , type-specific , reverse transcriptase real-time double PCR . In addition , virus isolation was attempted in African green monkey kidney epithelial cells ( Vero cells ) . A second sample collected 6 days later was retested using the same methods . In WA , dengue virus testing is not routinely requested for patients presenting with a dengue-like illness unless they have travelled overseas or to northern Queensland . Therefore , in order to exclude other possible locally-acquired cases , all blood samples sent to PathWest laboratories in the Pilbara requesting routine arbovirus testing between 1 July 2014 and 13 October 2013 were re-tested for dengue IgM and NS1 antigen . Intensive mosquito surveillance was carried out at the likely locations of exposure ( Point Samson and Wickham townships in the Pilbara region in north-west WA ( Fig 1 ) ) between 17 and 25 October 2013 . Adult mosquitoes were collected using carbon dioxide-baited encephalitis virus surveillance traps ( EVS/CO2 traps ) [12] modified to suit local environmental conditions , Biogents ( BG ) sentinel traps [13] , and sticky ovitraps [14] . Extensive larval surveys of potential container breeding habitats were also undertaken . A second mosquito survey of Point Samson , Wickham and Cape Lambert wharf ( Fig 1 ) was undertaken after the start of the wet season in January 2014 when desiccation-resistant eggs of container-breeding exotic mosquito species would be expected to hatch .
The case resided in a caravan park in the small coastal town of Point Samson in the Pilbara region of WA , located approximately 1500km north-east of Perth ( Fig 1 ) , and frequently travelled to Wickham , 10 km south-west of Port Samson . Both Point Samson and Wickham towns service mining operations in the region and Port Walcott ( Cape Lambert wharf ) , a major iron ore exporting port located 3km north-west of Point Samson ( Fig 1 ) where international cargo vessels dock . The case reported onset of symptoms typical of dengue fever , including fever , headache , lethargy and rash , on 24 September 2013 while in Bunbury in the southwest of WA , two days after departing Point Samson ( Fig 1 ) . He denied having travelled outside WA for many years , which was confirmed by the Commonwealth Department of Customs and Immigration , and had not been in the vicinity of any international airports or seaports during the two weeks prior to onset of symptoms with the exception of Port Walcott , which he was unable to recall specifically if he had visited or been bitten by mosquitoes there during the potential exposure period . He also reported wearing heavy duty clothing with long sleeves for work duties . However , the case did recall being bitten by mosquitoes at his residence in Point Samson during this period . Therefore , the most likely place of exposure was considered to be Point Samson , followed by Wickham where he travelled frequently for work , or Port Walcott where he may have also visited . In 2013 , prior to this locally acquired case , eight other cases of dengue fever were notified among Pilbara residents; all reported recent travel to dengue-endemic countries . Of these , the most recent case had been reported over six weeks earlier and resided in Karratha ( 50km west of Point Samson ) . None of these eight cases were residents of Point Samson or Wickham and the patient had not visited Karratha . Testing by the reference laboratory confirmed dengue virus infection in two serum samples taken six days apart . The results from the first serum sample tested was dengue IgM positive , dengue NS1 antigen positive , PCR positive for DENV-1 and had a high flavivirus HI antibody titre [1:640] . The second sample yielded the same results , with the exception of a negative NS1 antigen test . Cultures for dengue virus were unsuccessful and further subtyping of the virus was also not possible as a DNA sequence from the PCR product could not be obtained . One hundred and sixteen serum samples from 115 people submitted for routine arbovirus testing by PathWest’s Pilbara laboratories between 1 July 2014 and 13 October 2013 were re-tested for dengue IgM and NS1 antigen . One sample tested positive for dengue NS1 antigen and IgM and two further samples tested positive for dengue IgM alone . Further investigations revealed all cases had acquired a dengue infection overseas in a dengue-endemic country . The climate in the region of Point Samson is generally hot and dry with a mean annual rainfall of approximately 300mm , almost two-thirds of which falls during the wet season between January and March . No exotic mosquito species were detected during the intensive mosquito surveillance undertaken at Point Samson and Wickham . The dominant native species collected as adults were Culex quinquefasciatus and various species in the Aedes ( Macleaya ) subgenus which were also the only species collected as larvae . Small numbers of adult mosquitoes of the following species were also collected: Cx . annulirostris , Cx . sitiens , Ae . vigilax , Ae . bancroftianus and Anopheles novaguinensis . None of these are known or suspected dengue vector species . While properties within the area of the survey contained a large numbers of potential breeding receptacles , only six actual breeding sites were found . Overall , mosquito abundance at the time of the survey was extremely low . There had been no rainfall since 24 June 2013 so it is likely that mosquito abundance would also have been low during mid-September when this man was most likely to have acquired his dengue infection . Further intensive adult and larval mosquito surveys conducted in January 2014 following rainfall associated with the wet season also did not detect any known dengue vector species . Routine surveillance for exotic mosquitoes by the Commonwealth Department of Agriculture at international air and sea ports in WA did not detect the dengue vectors Ae . aegypti and Ae . albopictus in the months prior to this case ( personal communication , Aaron Maxwell , Assistant Director–Operational Science Services , Department of Agriculture Western Australia ) . The Department of Health issued a media statement on 15 October 2013 to warn the public of the possibility of local dengue virus transmission while the Pilbara Population Health Unit requested local doctors to test patients presenting with dengue-like symptoms . Other mosquito borne flaviviruses are relatively common in the Pilbara , so doctors were asked to include requests for dengue virus when testing for arboviral infections , whether or not the person had travelled overseas .
This paper reports the first case of dengue fever acquired in WA for the past 70 years . The case had not travelled outside WA for over a decade and laboratory test results confirmed a recently acquired DENV infection . Point Samson was the likely location of exposure as he recalled being bitten by mosquitoes there , but Wickham could not be ruled out as the case had also travelled there frequently for work during the potential exposure period , and neither could Port Walcott where he may have also visited . As northern WA has had epidemic dengue in the past , this case has renewed concern about the potential for re-establishment of dengue in the area . The initial concern was that the DENV infection in this case may have been transmitted by an exotic dengue vector mosquito that had become established in the Pilbara region and been infected with DENV by feeding on a viraemic returned traveller . This is analogous to the ongoing situation in northern Queensland [5–7] . While the existence of an established local breeding population of dengue vector mosquitoes could not be completely excluded by this investigation , neither the initial nor follow-up mosquito surveys found any evidence to support this scenario . In addition , no other cases were identified in the retrospective laboratory testing of patient sera with suspected arbovirus infections , although the possibility that further cases occurred but were not tested cannot be completely excluded . Similarly , in the absence of any DENV activity in mosquitoes or humans in the Pilbara prior to this case , it is very unlikely that it was introduced into a local native species with previously unrecognised vector competence . The most likely explanation for this locally acquired case is that a DENV-infected mosquito harbouring in luggage or in other items was transported into the region either by ship recently arrived from overseas , by road from northern Queensland , or via a direct international flight into Perth or Port Hedland International Airports ( Fig 1 ) from dengue-endemic countries . The case may have been bitten by the infected mosquito while visiting Port Walcott or the mosquito could have been transported to Point Samson , where it escaped and fed on the patient , but did not survive to lay eggs and reproduce in the dry and inhospitable environment at that time of year . While the probability of such a scenario occurring is low , it certainly cannot be discounted entirely and is the best explanation based on the available evidence . Transient introduction by air of an infected dengue vector was regarded as the most plausible cause of a locally-acquired dengue fever case in Darwin in the Northern Territory in 2010 [15] . Passive transportation of mosquitoes by both air and sea has also been demonstrated in previous studies in the south-east Asian region [16 , 17] and this has been suggested as a mechanism for spread of dengue fever internationally [4] . Further evidence to support this scenario is provided by the detection of exotic mosquitoes , including Ae . aegypti , at Perth , Adelaide and Melbourne International Airports ( Fig 1 ) in 2014 . The most regular detections occurred at Perth International Airport , but there is no evidence to suggest that Ae . aegypti became established locally [18] . We cannot be certain about the mechanism for local infection with DENV in this case . However , intensive surveys of mosquito fauna in the region of likely exposure immediately after serological confirmation of the case and again following wet season rains , together with enhanced screening and surveillance for other dengue cases found no evidence of an ongoing risk of dengue infection in the region . This case , and recent detections of Ae . aegypti at several international airports in Australia , highlights the potential risk of the incursion and establishment of DENV vectors in areas of Australia outside northern Queensland . If such vectors become re-established , the risk of local dengue outbreaks occurring would be very high due to the large number of infected returning travellers [11] . The ongoing public health surveillance and rapid response measures described in relation to this case remain an important public health priority in view of the well documented history of the widespread occurrence of Ae . aegypti and outbreaks of dengue fever in WA .
|
Dengue fever transmission in Western Australia ceased in the 1940s and there are currently no known dengue vector species present . Despite this , a locally acquired of dengue fever was reported in 2013 with the most likely location of exposure the township of Point Samson in Pilbara region in the north-west of the State . A comprehensive follow-up investigation was undertaken and while the dengue case was confirmed , no other cases were identified and not exotic dengue vector mosquitoes were detected around the location of exposure . The exact mechanism for the locally acquired infection could not be determined but the most likely explanation is that a dengue infected mosquito was transported into the region , fed on the patient , but did not survive to lay eggs and establish a local breeding population . This case highlights the potential risk of the re-emergence of dengue fever in regions of Australia outside where ongoing transmission is currently limited to in northern Queensland , particularly given the large number of imported dengue cases that occur among travellers returning from dengue endemic regions .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Investigation of the First Case of Dengue Virus Infection Acquired in Western Australia in Seven Decades: Evidence of Importation of Infected Mosquitoes?
|
The membrane-associated and membrane-spanning constituents of the Schistosoma mansoni tegument surface , the parasite's principal interface with the host bloodstream , have recently been characterized using proteomic techniques . Biotinylation of live worms using membrane-impermeant probes revealed that only a small subset of the proteins was accessible to the reagents . Their position within the multilayered architecture of the surface has not been ascertained . An enzymatic shaving approach on live worms has now been used to release the most accessible components , for analysis by MS/MS . Treatment with trypsin , or phosphatidylinositol-specific phospholipase C ( PiPLC ) , only minimally impaired membrane integrity . PiPLC-enriched proteins were distinguished from those released in parasite vomitus or by handling damage , using isobaric tagging . Trypsin released five membrane proteins , Sm200 , Sm25 and three annexins , plus host CD44 and the complement factors C3 and C4 . Nutrient transporters and ion channels were absent from the trypsin fraction , suggesting a deeper location in the surface complex; surprisingly , two BAR-domain containing proteins were released . Seven parasite and two host proteins were enriched by PiPLC treatment , the vaccine candidate Sm29 being the most prominent along with two orthologues of human CD59 , potentially inhibitors of complement fixation . The enzymes carbonic anhydrase and APD-ribosyl cyclase were also enriched , plus Sm200 and alkaline phosphatase . Host GPI-anchored proteins CD48 and CD90 , suggest ‘surface painting’ during worm peregrination in the portal system . Our findings suggest that the membranocalyx secreted over the tegument surface is not the inert barrier previously proposed , some tegument proteins being externally accessible to enzymes and thus potentially located within it . Furthermore , the detection of C3 and C4 indicates that the complement cascade is initiated , while two CD59 orthologues suggest a potential mechanism for its inhibition . The detection of several host proteins is a testimonial to the acquisitive properties of the tegument surface . The exposed parasite proteins could represent novel vaccine candidates for combating this neglected disease .
The persistence of adult schistosomes in the bloodstream for decades means they must deploy unique and effective immune evasion strategies at their interface with the host . The 1 cm-long worms are covered by a naked syncytial layer of cytoplasm , the tegument , connected by cytoplasmic tubules to underlying cell bodies that contain the machinery for protein synthesis , packaging and export . The tegument surface has a multilaminate appearance , interpreted as a plasma membrane overlain by a lamellate secretion , the membranocalyx [1] . This complex molecular architecture is maintained by export of the contents of multilaminate vesicles , which originate in the cell bodies of the syncytium . There is experimental evidence for slow turnover of the membranocalyx to the external environment [2] whilst recycling of the plasma membrane by internalisation has been anticipated but not conclusively demonstrated [3] . Initial observations suggested the membranocalyx was an amphipathic bilayer , probably composed of phospholipids , which served as a physical barrier to prevent antibody binding or host leukocyte attachment to the underlying plasma membrane . In addition , its supposed hydrophobic properties coincided with a demonstrable ability of worms in the bloodstream to acquire host molecules , particularly erythrocyte glycolipids ( the so-called host antigens ) [4] . Whether this acquisition is a deliberate process that benefits the parasite or an accidental consequence of the membranocalyx properties , and any relevance it has to immune evasion , remain unclear . Building on techniques developed in 1980s to detach the tegument by freeze-thaw and enrich the surface membrane complex by differential centrifugation [5] , we have characterized its composition using proteomic techniques [6] , [7] . We developed a differential extraction scheme for the membrane preparation , with chaotropic agents of increasing strength , which enabled us to identify both membrane-associated and membrane-spanning constituents . These compositional findings demonstrated the importance of the tegument for nutrient uptake and maintenance of solute balance , as well as the presence of several hydrolases in the surface layers [6] . In a second study we incubated live worms with membrane-impermeant probes to biotinylate the most externally-accessible proteins and then recovered the tagged molecules by affinity chromatography for MS/MS identification [7] . This approach revealed that only a small subset of transporters , membrane structural proteins , enzymes , and schistosome-unique proteins were labelled , together with host immunoglobulins and complement C3 . We concluded that these represented the most “exposed” surface constituents but we could not place them within the multilayered architecture of the surface with any certainty . To add another dimension to our understanding of tegument surface organization we have now used an enzymatic shaving approach on live worms to release the components most accessible to the selected enzymes , for MS/MS analysis . By analogy with techniques for stripping adherent cells from culture flasks , we used trypsin to cleave exposed protein loops or domains without impairing membrane integrity . We also incubated worms with phosphatidylinositol-specific phospholipase C ( PiPLC ) to release any externally accessible GPI-anchored proteins . As a control for proteins released by the vomiting of gut contents during the incubation period , or damage due to handling , we compared protein release +/− PiPLC , using the iTRAQ technique . Finally we used a phospholipase A2 ( PLA2 ) preparation purified from snake venom to erode the lipid bilayer complex to determine if proteins could be selectively detached . We report that both trypsin and PiPLC removed a small subset of proteins whilst inflicting minimal damage on the worms whereas PLA2 was more destructive . We show that the iTRAQ technique identified both parasite and host proteins enriched by PiPLC treatment whereas trypsin released a different subset , with only Sm200 common to both . The proteins we have identified in the adult are also present as transcripts in the lung schistosomulum . We suggest that collectively they may be candidates for a schistosome vaccine , especially if responses can be targeted to the lungs to interfere with intravascular migration of incoming larvae .
The procedures involving animals were carried out in accordance with the UK Animals ( Scientific Procedures ) Act 1986 , and authorised on personal and project licences issued by the UK Home Office . The study protocol was approved by the Biology Department Ethical Review Committee at the University of York . A Puerto Rican isolate of S . mansoni was maintained using albino Biomphalaria glabrata snails and NMRI strain mice as laboratory hosts . All animal experiments were approved by the Ethical Review Process Committee of the Department of Biology , University of York . Adult parasites were obtained by portal perfusion of mice seven weeks after exposure to 200 cercariae , using RPMI1640 medium ( minus phenol red ) buffered with 10 mM HEPES ( both from Invitrogen , Paisley , UK ) . Parasites were extensively washed in the same medium and tissue debris and any damaged individuals removed under a dissecting microscope . No attempt was made to separate males from females . Approximately 800 freshly perfused parasites from 20 mice were used in each of four replicate experiments with trypsin as the shaving enzyme . They were incubated in a 30 mL Corning flask ( Corning , NY , USA ) containing 5 mL of buffered RPMI , with trypsin MS ( Promega , Southampton , UK ) added at 10 µg/mL , for 30 min at room temperature ( RT ) . The supernatant was recovered , transferred to a 15 mL Falcon tube and centrifuged at 500×g for 30 min to remove any insoluble material such as the haematin particles in gut vomitus . Streptomycin and penicillin were then added to a final concentration of 100 µg/mL to prevent microbial growth and tryptic digestion was continued overnight at 37°C , after which peptides were reduced and alkylated . Reduction was performed in the presence of 20 mM DTT for 30 min at 65°C in a water bath and , after cooling , alkylation was performed in the presence of 80 mM iodoacetamide for 1 h at RT in the dark . Trifluoroacetic acid ( TFA ) was then added to a final concentration of 0 . 1% before recovery of peptides by passage through a solid phase Strata C18-E extraction cartridge ( 55 µm , Phenomenex , Macclesfield , UK ) , followed by several column washes in 0 . 1% TFA and final elution in 750 µL of 50% acetonitrile/0 . 1% TFA . The eluted fraction was concentrated under vacuum to dryness and peptides resuspended in 20 µL 0 . 1% TFA . A 3 µL aliquot of the tryptic peptide preparation was injected onto a reversed-phase PS-DVB monolith column ( 200 µm i . d . ×5 cm , LC Packings , Amsterdam , Netherlands ) . Peptides were separated using a two-step linear gradient of 2–31 . 4% ( v/v ) acetonitrile in 0 . 1% aqueous heptafluorobutyric acid over 60 min , followed by 31 . 4–51% ( v/v ) in the same solvent over 5 min , at a flow rate of 3 µL/min; UV absorbance at 214 nm was monitored . Fractions were collected onto a MALDI target plate using a Probot ( Dionex , Bannockburn , USA ) with simultaneous addition of matrix solution ( 6 mg/mL α-cyano-4-hydroxycinamic acid ( CHCA , Sigma , Poole , UK ) in 60% ( v/v ) acetonitrile ) . Positive-ion MALDI mass spectra ( MS ) were obtained using a 4700 Proteomics Analyzer with TOF-TOF Optics ( Applied Biosystems , Framingham , USA ) in reflector mode , over the m/z range 800–4000 and monoisotopic masses obtained from centroids of raw , unsmoothed data . The precursor mass window was set to a relative resolution of 50 , and the metastable suppressor was enabled . The default calibration was used for MS/MS spectra , which were baseline-subtracted ( peak width 50 ) and smoothed ( Savitsky-Golay with three points across a peak and polynomial order 4 ) ; peak detection used a minimum S/N of 5 , local noise window of 50 m/z , and minimum peak width of 2 . 9 bins . The twenty strongest peaks from each fraction , having a signal to noise ( S/N ) greater than 50 and a fraction-to-fraction precursor exclusion of ±0 . 2 Da , were selected for CID-MS/MS analysis . Singly-charged peptides were fragmented with Source 1 collision energy of 1 keV , and air as the collision gas . Peak lists from the MS/MS data , containing all m/z values from m/z 20 to the precursor m/z - 60 , with a minimum S/N of 10 , were provided by TS2 software ( version 1 . 0 . 0 , Matrix Science Ltd . , London , UK ) . Each list , corresponding to one MALDI plate , was then submitted to a local copy of the Mascot program ( version 2 . 1 , Matrix Science ) and searched against the SmGenesPlusESTs ( 260448 sequences; 71029272 residues ) , an in-house database derived from the publically available data in http://www . genedb . org/genedb/smansoni/ ) , and the NCBInr Mus musculus database ( 139457 sequences ) for host proteins . Search parameters specified only tryptic cleavages and allowed for up to one missed site , the variable carbamidomethylation of cysteines , and oxidation of methionines; precursor and product ion mass error tolerance was set to ±0 . 3 Da . A decoy database , generated by Mascot , was used with the significance threshold for protein identification set to achieve a false positive rate of 1 to 2% and peptide threshold set to ‘least identity’ . A protein was considered positively identified if the ion score for a particular peptide had an expect value less than 0 . 05 . GPI-anchored proteins were recovered from live worms by in vitro incubation with PiPLC as the shaving enzyme , the experiment being performed twice to provide biological replicates . Downstream analysis required the worms from 40 mice , which were perfused and treated in two separate batches to minimise the time ex vivo; supernatants were then combined . Each batch was incubated at 37°C for 1 h in the presence of PiPLC ( from Bacillus cereus , Sigma ) at 1 . 25 Units/mL , with conditions as for trypsin . The supernatant was removed and concentrated at 4°C using a 5000 Da cut-off centrifugation device ( Vivaspin 6 , West Sussex , UK ) . The control for secretion , vomitus production and parasite damage due to handling comprised an identical experiment , minus PiPLC . This also served as a ‘background’ control for the other enzyme treatments . For one experiment , PiPLC-released proteins were obtained using PiPLC from a different source ( from Bacillus thuringiensis , Europa Bioproducts , Wicken , Cambridigeshire , UK ) employing the same conditions as above . On that occasion the GPI-released fraction was used for a 2-DE separation . The composition of released material was evaluated by 1-DE using a pre-cast NUPAGE 4–12% Bis-Tris gel ( Invitrogen ) after a 45 min run at 200 V . The gel was then fixed in 40% methanol , 10% acetic acid for 30 min , stained with SYPRO Ruby ( Invitrogen ) for 2 h in the dark and imaged using a Molecular Imager FX ( Bio-Rad , Bath , UK ) . Protein content in each lane of the gel was then estimated by densitometric analysis using Quantity One software ( Bio-Rad ) . Fifty µg of the PiPLC-treated sample was also evaluated by mini 2-DE essentially as previously described [8] , [9] . After electrophoresis the gel was first stained with SYPRO Ruby , imaged as above , and restained with Bio-Safe Coomassie ( BioRad ) ; all visible spots were selected for “in gel” digestion [8] . An aliquot of 1–2 µL of the digestion supernatant containing the peptides was spotted on a MALDI plate and dried before the addition of 0 . 6 µL of a saturated solution of CHCA matrix ( in 50% acetonitrile/0 . 1% TFA ) . Peptide fragmentation data from each gel spot was processed by GPS Explorer Software ( Applied Biosystems ) underpinned by Mascot ( settings as above ) , to provide a putative identity for the protein . The relative composition of PiPLC-treated and control samples was characterized using isobaric tagging ( the iTRAQ labelling technique ) following the protocol provided by the manufacturer ( Applied Biosystems ) . Prior to labelling , two aliquots of treated and control samples containing 10 µg protein were taken to provide technical replicates . Briefly , 10 µg of both control and PiPLC-treated samples were individually denatured , reduced , and alkylated with reagents supplied in the iTRAQ kit . Peptides were generated by trypsin digestion using a 1∶20 enzyme/protein ratio , at 37°C for 24 h , and labelled with iTRAQ reagents at lysine , terminal amine groups and partially at tyrosine residues . Test samples were labelled with tags 116 or 117 and control samples with tags 114 or 115 , respectively . A peptide mixture was made by combining the four tagged samples and cleaned up using a strong cation-exchange cartridge to remove the detergents and excess iTRAQ reagents . The peptides were then affinity-purified in a Strata C18-E cartridge , eluted as for tryptic peptides , dried using a vacuum concentrator and resuspended in 10–20 µL of 0 . 1% TFA . LC-MS/MS was performed as described above . Protein identification and peptide quantification was achieved by submitting the TS2-generated MS/MS raw data files to Mascot , searching against SmGenesPlusESTs and NCBInr databases . Search parameters were tryptic peptides , with 0–1 missed cleavage; fixed modifications , β-methylthiolation of cysteines , iTRAQ tagging of lysines and N-terminal amine groups; variable modifications were oxidation of methionines and iTRAQ tagging of tyrosines . Precursor and product ion mass error tolerance was set to ±0 . 3 Da . The 114-tagged sample ( C1 ) was taken as the reference for calculating ratios . The Mascot software displays the median normalised geometric mean ratios and a factor from which the geometric standard deviation can be derived . Automatic outlier removal was performed by the Mascot software . A protein was considered enriched if the mean T1/C1 and/or T2/C1 ratios minus the 95% confidence limit exceeded the corresponding C2/C1 ratio plus its 95% confidence limit . For single peptide identifications , a protein was considered enriched if it appeared in both biological replicate experiments , and had a mean treated/control ratio >2 . The final approach for recovery of parasite surface molecules using enzymatic shaving , involved the treatment of live parasites with PLA2 . This enzyme isolated from the venom of Crotalus durissus terrificus ( deposited under accession number P24027 at NCBInr ) was kindly provided by Prof . Andreimar Soares ( Faculty of Pharmaceutical Sciences , University of Sao Paulo , Brazil ) . During the shaving experiment , approximately 800 freshly perfused parasites were incubated in a 30 mL Corning flask , containing 6 mL of buffered RPMI1640 , with PLA2 added at 16 µg/mL for 1 h , at RT . Another batch of parasites was used in a control and parallel experiment in which no enzyme was added to the culture medium . The supernatants from control and treated samples were concentrated using a Vivaspin 6 filtration device ( 5000 Da cut off ) at 500×g , at 4°C until the volume reached 500 µL . The samples were then transferred to 1 . 5 ml eppendorf tubes and centrifuged at 25 , 000×g for 20 min . After this step a membranous pellet , recovered from PLA2-treated parasites only , was extracted in 50 µL of 0 . 5% Triton-X100 , yielding approximately 20 µg of protein . Reduction , alkylation , trypsin digestion and peptide clean-up were performed as described for the iTRAQ protocol ( omitting the labelling steps ) . LC-MS/MS of the peptide mixture was performed essentially as described above . Host plasma membrane proteins on the surface of adult parasites were investigated either on live worms , perfused from mice using RPMI medium , or by preparing OTC-embedded cryostat sections . Both were incubated with primary antibodies at 1∶100 dilution in PBS for parasite sections and in RPMI for live worms , containing 5% normal goat serum , for 1 h at RT . Monoclonal antibodies used were rat anti-mouse CD44 ( 558739 ) , rat anti-mouse CD90 ( Thy-1; 553016 ) and hamster anti-mouse CD48 ( 553682 ) , all from BD Biosciences Pharmingen , New Jersey , USA . Labelling was detected by the use of goat anti-rat IgG conjugated to Alexa fluor 488 at 1∶500 dilution for 30 min in the same buffer , or goat anti-hamster IgG conjugated to Alexa fluor 568 under the same conditions . A searchable database was created in 2005 , comprising all S . mansoni transcripts then available from dbEST ( http://www . ncbi . nlm . nih . gov/ ) , the Sao Paulo Schistosoma mansoni EST genome project ( http://bioinfo . iq . usp . br/schisto/ ) , and the Wellcome Trust Sanger Institute ftp site ( ftp://ftp . sanger . ac . uk/pub/pathogens/Schistosoma/mansoni/ESTs ) . All proteins of interest identified by the enzymatic shaving approach were searched against the compiled EST data to determine the number of transcripts detected in adults and lung stage schistosomula; this provided a very approximate guide to the relative levels of expression in the two life cycle stages .
Freshly perfused live parasites were subjected to trypsin digestion under controlled conditions , in order to release from the surface membrane complex of the tegument any proteins , or segments thereof , accessible to the enzyme . The vast majority of worms retained a normal appearance and activity over the 30 min incubation . The culture supernatant was recovered and released proteins/peptides allowed to digest further overnight . Tryptic peptides were then reduced and alkylated before their recovery and separation using reversed-phase chromatography ( Figure S1A ) for subsequent mass-spectrometric identification . The identities obtained by Mascot searching of the MS/MS data were then categorized according to their molecular function and , by inference , their potential cellular origin ( Table 1 ) . The host complement proteins C3 and C4 and the leukocyte surface marker CD44 were released in two or more of the four replicate experiments . Host haemoglobin alpha and beta chains were also identified in one experiment , presumably an indication that regurgitation from the worm gut was occurring . A number of proteins , known to be associated with the tegument surface , were also released by the treatment . They included three phospholipid-binding annexins and the membrane protease calpain . Of the annexins , Smp_077720 was found in all four experiments and the others on two and one occasions , respectively; calpain was detected in only one experiment . The schistosome-unique proteins of unknown function , Sm200 and Sm25 , were found on three and four occasions , respectively . The final membrane protein , not previously associated with the tegument , shows homology with cell surface proteoglycans . A pair of BAR-domain-containing endophilins was also released by the treatment in all four experiments while the putative potassium-channel inhibitor , SmKK7 , was found on one occasion . Although worm incubations were short-term , two known gut-derived proteases , asparaginyl endopeptidase and cathepsin B1 were detected . In addition , a total of 12 proteins with unknown function and localization were found ( Tables S1 and S2 ) , some containing domains , e . g . Ig-like and EGF-like , which may indicate a surface position . In spite of strenuous efforts to maintain worm viability , the trypsin treatment affected the integrity of the surface membranes to some extent , as evidenced by the appearance of seven cytoskeletal and 17 cytosolic proteins in the medium ( Tables S1 and S2 ) . Among the former , actin , fimbrin and severin have been proposed as constituents of the tegumental spines that reside immediately beneath the surface membrane complex . The more numerous cytosolic proteins , comprising glycolytic enzymes , chaperones , and antioxidants indicate the leakage of internal components in at least some of the parasites . Known tegumental surface transporters , ion channels , and enzymes ( other than calpain ) were conspicuous by their absence in the trypsin preparation . The purpose of treatment with PiPLC was to release GPI-anchored proteins accessible to the externally applied enzyme in live worms; incubation with the enzyme for 1 h appeared to have no morphologically obvious deleterious effect . A 1-DE gel separation of material released by treated and control worms revealed a complex pattern of protein bands with Mr ranging from 10 to >250 kDa ( Figure 1A ) . The two preparations displayed a strong similarity , with only three bands ( arrowed ) visibly enriched by the PiPLC treatment versus the control; two were of high molecular mass ( approx . 200 kDa ) while the other was located at the bottom of the gel ( approx . 12 kDa ) . The PiPLC treatment released sufficient protein to permit a mini 2-DE separation ( Figure 1B ) for subsequent MS/MS analysis of gel spots . The analysis revealed the presence of proteins known to be GPI-anchored , such as ‘Surface protein’ ( Sm200 ) and alkaline phosphatase . Protein orthologues of CD59 and carbonic anhydrase IV were novel features of this 2D map . The absence of gut proteases Sm31 and Sm32 was notable . The identities of other spots revealed the presence of cytosolic and cytoskeletal contaminants , including thioredoxin , fatty acid binding protein ( Sm14 ) , Sm22 . 6 , enolase and triose phosphate isomerase ( Table S3 ) . The remaining spots on the SYPRO-stained gel ( Figure 1B ) were not detected by Coomassie staining so no identification could be assigned using single spot tryptic digestion . As proteins were released into the culture medium during the 1 h incubation irrespective of whether PiPLC was present or not , we used the iTRAQ technique to determine the degree of enrichment due solely to the enzymatic shaving . Within an individual labelling protocol , splitting both the control and treatment samples provided technical replicates . After LC separation of the tagged peptide mixture ( Figure S1B ) , fragmentation spectra of the most abundant peptides were generated , each containing signature peaks for the four reporter tags ( Figure 2 and Figure S2 ) . In most instances the fragmentation spectra yielded four peaks of approximately equal area ( Figure 2A ) . The 115/114 ( C2/C1 ) ratios approximated to unity , indicating approximately equal protein contributions of the two control samples to the iTRAQ mixture . However , the majority of 116/114 ( T1/C1 ) and 117/114 ( T2/C1 ) ratios were almost invariably less than one , sometimes significantly so ( Figure 2A and Figure S2 ) . We have no reason to believe that the control and treatment incubations differed except in the addition of enzyme , and so we must attribute the ratios of less than unity to a dilution effect on the background proteins in the treated samples . This dilution resulted from the addition of PiPLC plus the extra proteins released from the surface by its action . In a minority of fragmentation spectra the four peaks representing the reporter tags were not of approximately uniform area ( Figure 2B and Figure S2 ) . The reporter ion peaks from the treated samples were from 2 to 10 fold greater intensity indicating enrichment of the parent peptide by the PiPLC treatment . A total of 52 identities was obtained from fragmentation of the tagged peptides ( Figure S2 and Tables S4 and S5 ) ranging from 1 to 15 per identity , primarily of cytosolic or cytoskeletal origin . The PiPLC-enriched proteins of parasite origin were , in descending order of abundance , Sm29 , CD59a , Sm200 , carbonic anhydrase , CD59b , alkaline phosphatase and ADP-ribosyl cyclase ( Figure 3 ) . In addition two host proteins , CD48 and CD90 ( Thy1 . 2 ) , were also enriched by PiPLC treatment . Proteins of known gut origin , α2-macroglobulin and saposin B , were present in control and treated samples in equivalent amounts indicating similar regurgitation of gut contents over the one hour incubation . Equivalent amounts of cytosolic ( e . g , enolase , 14-3-3 ) and cytoskeletal proteins ( e . g , actin , Sm20 . 8 ) in control and treated samples ( Figure 3 and Table S4 ) indicated a certain degree of surface damage , but not inflicted by the PiPLC treatment . Although not enriched by PiPLC hydrolysis , because they lack the GPI-anchor , three other proteins are worth noting . Two of these , CD63/tetraspanin ( TSP-2 ) and annexin IV ( Smp_074140 ) were already known to be associated with the tegument surface and the third is an 8 kDa low molecular weight protein ( LMWP ) . This last protein , present in a range of trematodes [10] , has a signal peptide and so may be a true secreted protein released along with the membranocalyx; its status as a tegument surface protein needs to be confirmed . Unlike the other two enzymatic treatments PLA2 had a dramatic effect on worm appearance and viability . Moderate concentrations of PLA2 produced visible worm damage and death , whereas greater dilutions had little selective effect in removing known tegumental surface components whilst still causing leakage of cytoskeletal and cytosolic components ( Table S6 ) . The known tegument surface proteins released were Sm29 , LMWP and dysferlin . The PLA2 approach was therefore discontinued . Confocal microscopy of adult worm sections revealed that host CD44 was confined entirely to the tegument with no staining of internal structures ( Figure 4A ) . Examination of a Z-stack through the tegument of an intact male worm revealed that the pattern of staining was not uniform , being concentrated on numerous , parallel , transverse ridges and especially the spines on the dorsal tubercles ( Figure 4B ) . On close inspection , the individual spines had a definite inverted V appearance the most intense staining being at the tip ( Figure 4C ) . Antibodies to CD48 and CD90 failed to detect their respective targets on the worm surface . Our in-house EST database was interrogated for the occurrence of transcripts in lung schistosomula encoding the proteins of interest released from adult worms by the enzyme treatments . Although it is difficult to make inferences about abundance as some data were obtained from normalised libraries , nearly all transcripts were represented in both larvae and adults in roughly similar numbers ( Table 2 ) . The exceptions were the proteoglycan ( 5L and 0A ) and ADP-ribosyl cyclase ( 0L and 1A ) . Transcripts for calpain and Sm200 were particularly abundant , whilst the largest number recorded ( 63 ) was for one of the CD59 orthologues in the lung schistosomulum . Based on the transcript evidence we tentatively conclude that the most exposed tegument surface proteins of adults are also likely to be present on the surface of the migrating lung schistosomulum . Only one tegument surface protein , Sm200 , was released by both the trypsin and PiPLC treatments of live worms . The difference between the two was that PiPLC released the entire GPI-anchored molecule for subsequent trypsinisation as a separate step . Conversely the trypsin treatment released peptide fragments from Sm200 molecules only during the incubation period , which were then recovered using reversed-phase chromatography . Mapping of the peptide hits onto the primary amino acid sequence revealed a very different pattern of distribution ( Figure S3 ) . A total of 15 peptides , with a significant Mascot score , was identified for the PiPLC-released protein , distributed throughout the entire molecule . In contrast only six peptides were identified for the trypsin-released protein and four of these were clustered towards the C-terminus of the protein , i . e . the region located nearest to the GPI anchor . This may indicate that only part of the Sm200 protein is accessible to trypsin in the live worm . In this context , the programs NetNGlyc and NetOGlyc both at http://www . cbs . dtu . dk/services/ predicted five N ( scores >0 . 6 ) but no O-glycosylation sites on the amino acid sequence coding for Sm200 . It is notable that four of the five potential N-linked sites are located furthest from the GPI anchor site in the N-terminal region where few peptides were identified following trypsin treatment of the live worms . Thus it is plausible that attached N-glycans protect the native Sm200 polypeptide chain in situ on the worm surface from trypsin attack .
Our previous biotinylation studies have provided insights into the disposition of proteins in the complex molecular structure of the schistosome tegument surface . In the present study we have taken an alternative approach to obtain information about the location of tegument components . This involved shaving the surface of live adult worms with selected enzymes to release accessible proteins , and their identification by proteomics . It is axiomatic in this approach that the integrity of the worm surface is not compromised by the treatment . Our results indicate that trypsin and PiPLC largely fulfilled this criterion but PLA2 did not . We must assume that this last enzyme , via its attack on the lipid bilayers , rapidly caused generalized erosion of the surface membranes and loss of integrity . We shall therefore consider only the results of the trypsin and PiPLC treatments to make inferences about tegument surface organization . In addition , the male parasite has at least two times the surface area of the female , assuming all surfaces are accessible to the enzymes . If the gynaecophoric canal is effectively sealed , then the male dorsal surface would have contributed the bulk of released protein . It was inevitable that the in vitro culture of live worms resulted in the contamination of the enzyme-released proteins by gut content . Thus the detection of proteins such as α2-macroglobulin , saposin B , and hemoglobinase ( asparaginyl endopeptidase ) in the two enzymatic shaving experiments was to be expected . Conversely , even with the most careful handling of parasites including their recovery from mice by perfusion with culture medium , the presence of cytosolic and cytoskeletal proteins must be attributed to some degree of worm damage . In the present study , we used the iTRAQ labelling technique on test and control samples to discriminate enzymatic enrichment from normal secretion or protein leakage due to invisible damage . This approach successfully overcame the dominance of abundant peptides from contaminating cytosolic and cytoskeletal proteins to highlight the few proteins released by PiPLC treatment . Nevertheless , one must bear in mind that as enzyme accessibility has not been addressed in this investigation , caution should be exercised when considering the fold enrichment found for GPI-anchored molecules . In this regard , it is possible that a higher fold enrichment for a given protein may only indicate a more exposed/accessible location at the parasite surface rather than imply protein abundance . A major finding of our study is that seven parasite proteins can be detached from the surface by treatment with PiPLC and enriched compared to control incubations . From this we infer that each is inserted into the parasite surface by a GPI-anchor . As these anchors are invariably located at the C-terminus of such proteins we conclude that the extraneous 30 kDa PiPLC enzyme is able to access the anchor in proximity to a lipid bilayer . Whether this bilayer is the secreted membranocalyx or the underlying plasma membrane is problematic . Access to the latter would mean that the digesting enzyme had to pass through the protective membranocalyx . One possible inference is that the seven GPI-anchored proteins are inserted into the membranocalyx . For carbonic anhydrase this seems unlikely because of its assumed function in regulating acid-base balance at the parasite surface . It would be best placed to do this if located between the two lipid bilayers so that , by analogy with the erythrocyte , CO2 generated inside the worm could be converted to HCO3− for diffusion into the bloodstream . Indeed , the reverse diffusion of a chloride ion through the membranocalyx would be required to balance the charge , bringing it into close proximity with the anion transporters , which our previous studies have shown are present in the tegument plasma membrane [6] . The gene model for carbonic anhydrase is incomplete , lacking both the N-terminal exon encoding the signal peptide and the C-terminal exon ( s ) encoding the site for the attachment of a GPI-anchor . However , its enrichment by PiPLC treatment of live worms attests to the presence of a GPI-anchor . Two isoforms of CD59 orthologues , identified by possession of the CCxxDxCN motif , were also highly enriched by the PiPLC treatment , implying their outer location at the surface . These proteins are potentially significant because the human CD59 protein protects self-cells against complement fixation by blocking formation of the C5 to C9 membrane attack complex [11] . A similar role for the two schistosome molecules is very attractive , as a component of the parasite's mechanisms of immune evasion . It is of note that in the S . mansoni genome there are four additional gene models encoding CD59 orthologues , which could also be tegument-associated [12] . These CD59 orthologues in the tegument surface are better candidates for inhibition of complement fixation than the SCIP-1 protein [13] , subsequently identified as paramyosin [14] , [15] . The protein most highly enriched by PiPLC treatment was Sm29 that has no known function . It was originally selected by a bioinformatic search as a protein with a single membrane spanning domain [16] . It was identified in the final pellet after differential extraction in a compositional analysis of the tegument membranes [6] and was accessible to biotinylation in live worms [7] . Subsequent confocal microscopy revealed its association with the tegument although technical issues do not allow a firm conclusion about its precise location [17] . Our PiPLC shaving results demonstrate both its peripheral location and the fact that it is GPI-anchored , the latter prediction also made by big-PI Prediction server ( http://mendel . imp . ac . at/gpi/gpi_server . html ) . The peripheral location of Sm200 , again with no known function , was also revealed by PiPLC shaving . This protein first cloned by Hall et al . [18] was already known to be GPI-anchored and located in the tegument surface [19] , [20] . Surprisingly therefore it was not found in compositional analysis of the tegument membrane [6] , but was accessible to biotinylation in live worms [7] . These observations suggest that whilst definitely surface-attached it is readily lost during the processing of tegument membranes by differential extraction for MS analysis , a feature that may be linked both to its size and GPI-anchor . Moreover , Sm200 has recently been identified in circulating lipoprotein particles from the blood of schistosome-infected humans , confirming its turnover into the vascular environment [21] . The two remaining GPI-anchored proteins shown to be enriched by iTRAQ labelling after PiPLC treatment were alkaline phosphatase and ADP-ribosyl cyclase . The former is well characterized and was previously shown to be GPI-anchored in schistosomula [22] . It has been used as a membrane marker in the development of methods for tegument surface isolation [5] . It was identified by compositional analysis of the tegument surface membranes in the urea/thiourea/CHAPS/sulfobetaine ( UTCS ) fraction and insoluble pellet [6] , and is accessible to biotinylation in live worms [7] . Its presence in the UTCS fraction indicates that it is , at least partially , loosely associated with the surface membranes i . e not membrane spanning . The recently characterised ADP-ribosyl cyclase was also shown to be GPI-anchored and localized to the outer tegument of the adult schistosome [23] . Both schistosome enzymes must be able to access their substrates in live worms but their precise function at the tegument surface remains to be established . In the case of ADP-ribosyl cyclase , a role in calcium mobilization has been proposed [23] but alternatively it could function in immune evasion by regulating ecto-NAD+ levels , thereby reducing substrate availability for CD38- and CD157-mediated effector functions of lymphocytes . Corroborating the identification of GPI-anchored molecules using the iTRAQ technique , four out of the seven proteins assigned as GPI-anchored were also detected by our 2-DE analysis . These were Sm200 , alkaline phosphatase , two isoforms of CD59 orthologues and carbonic anhydrase . A definite proof of their enrichment due to PiPLC's activity is the fact that these molecules are underrepresented and are not easily identifiable in 2-DE maps produced using soluble worm preparations [8] or membrane-extracted proteins from crude or differentially-extracted S . mansoni tegument [6] . Use of the iTRAQ technique to identify proteins enriched by trypsin treatment was not possible because the released material was already partially digested . This prevents accurate quantification of protein for labelling and therefore the trypsin shaving fraction was subsequently processed as a peptide digest . We have taken as a reference of cytosolic/cytoskeletal contaminants in the trypsin shaving , the number and diversity of those proteins released during 1 h incubation at 37°C in the absence of added enzyme . When we compared the number of proteins that could indicate leakage or vomitus with the number of proteins in the trypsin-treated parasites , we observed we had less contamination , most likely due to a shorter incubation time , 30 min . In addition , we are assuming that as trypsin should not be able to permeate the parasite tissues , the peptides originating from membrane proteins are likely to represent the most exposed/accessible domains at the parasite surface . We therefore focus only on ( 1 ) membrane and membrane-associated , ( 2 ) vesicular pathway and secreted , and ( 3 ) host proteins . The peripheral location of Sm200 was again confirmed but it was the only one of the seven PiPLC-released proteins that was also released by trypsin . This suggests that the rest may be protected from proteolysis , potentially by their N and/or O-linked glycans or by a sequestered location . The release of three annexins by the trypsin treatment indicates their superficial location . One ( Smp_077720 ) was previously detected by both biotinylation and compositional analysis and two are new to this study ( Smp_074140 and Smp_074150 ) . The known phospholipid-binding properties of this group of proteins means that they could have a role in promoting adhesion of the membranocalyx to the plasma membrane , acting like a molecular ‘velcro’ via the four binding domains that each possesses . More recently , certain annexin isoforms have been implicated in immunomodulatory functions such as the resolution of inflammation [24] . They are able to interact with receptors on the surface of leukocytes to control apoptosis and their clearance by macrophages . The existence of such a process at the surface of the tegument , mediated by schistosome annexins , accords with the absence of leukocyte binding to adult worms in vivo [25] . However , a comparison of the orthologies of the schistosome and human annexins is difficult because of the evolutionary distance . Thus , BLAST searching of the schistosome sequences against the NCBInr database reveals the closest homologues of Smps 077720 , 075150 and 075140 are human annexins A13 , A7 and A8 respectively , not the A1 isoform that has been most implicated as an anti-inflammatory agent . The conjecture will only be resolved by expression of the schistosome annexins for assays of function . The presence of Sm25 as a tegument surface protein has a chequered history . It was proposed as a vaccine candidate because anti-Sm25 antibody levels correlated with protection in mice vaccinated with a crude tegument membrane preparation [26] . It was then cloned and designated as an N-glycosylated integral membrane protein [27] but later characterized as a palmitoylated protein with the implication that it was on the cytosolic leaflet of the plasma membrane [28] . Further immunocytochemical studies suggested it was distributed throughout the tegument syncytium but not associated with the surface membranes [29] and it could only be biotinylated when parasites were permeabilized by Triton X-100 [30] . It was not found in either our compositional or biotinylation studies on the tegument surface . The removal of Sm25 from live worms by trypsin , when so few other membrane proteins are released , suggests a unique accessibility . The proteolytic enzyme calpain was previously identified at the tegument surface by both compositional analysis [6] and biotinylation [7]; its potential as a vaccine candidate has already been exploited [31] , [32] . The final member of this group is a heparin sulphate-like proteoglycan which has not been identified in any previous studies and merits further investigation . In the secreted protein category , this is the first report of SmKK7 release from the tegument surface; it has previously only been reported from cercarial secretions [33] . The detection of a pair of BAR domain proteins in all four experiments after trypsin treatment is intriguing since these banana-shaped molecules located on the cytosolic surface of plasma membranes function to promote membrane curvature in exo- and endocytosis [34] . This suggests that the trypsin may be entering the fine tubules located at the base of tegument pits where multilaminate vesicle fusion with the tegument plasma membrane occurs [35] , [36] . That conclusion would place Sm25 and the proteoglycan in a similar location where constituent proteins are poorly protected by overlying membranocalyx , perhaps loosely analogous to the relationship of the flagellar pocket of trypanosomes [37] to variant surface glycoprotein export . The release of complement C3 from the parasite surface by trypsin treatment is consistent with our previous biotinylation study [7] . We can now add C4 , its precursor in the complement cascade , but not the C5 to C9 components of the membrane attack complex , suggesting complement fixation but then inhibition of the cascade . The failure to detect any immunoglobulins , when murine IgM , IgG1 and IgG3 heavy chains were biotinylated [7] , is puzzling as they would be expected to initiate complement fixation . However , there is evidence that these proteins are resistant to trypsin degradation , particularly under the low concentration conditions employed in our experiments [38] . In fact , aiming to preserve worm viability , trypsin was used at a concentration approximately 100 times lower than that typically employed for stripping adherent cells from culture flasks . A positive consequence of such a gentle shaving treatment on live worms is justified by the reasonable number of genuine membrane proteins that were identified . In contrast , a previous study on S . bovis with trypsin treatment , used methanol-fixed parasites [39] and found an overall higher number of protein identifications . However , membrane-associated and/or integral proteins were poorly represented , and the results are not comparable with our approach using live worms . Our enzyme shaving experiments extend the list of host proteins known to be firmly associated to the parasite surface . The release of CD48 and CD90 ( Thy1 . 2 ) by PiPLC confirms their possession of GPI-anchors and most likely accounts for their transference from the host ( leukocytes ) in the same manner as host glycolipids transfer from erythrocytes through a process currently termed cell-painting [40] . Supporting this finding is the demonstrable ability of purified Thy-1to reincorporate into the plasma membrane of murine Thy-1- cells directly from aqueous suspension and without the use of detergents [41] . However , failure to detect the presence of both CD48 and CD90 by immunocytochemistry indicates their relative paucity . The detection of host CD44 is more surprising since it is an extracellular protein anchored in the membrane of a range of cell types by a single transmembrane domain [42] . However , the intense staining of the tips of spines on the dorsal tubercles , the point of contact between the worm and the vascular endothelium during peregrination around the portal vasculature , suggests that the transfer occurs when the apposing membranes of tegument and endothelium are pressed together . The observations on CD44 staining are a testament to the acquisitive properties of the schistosome surface . Our studies on the surface accessibility of tegument proteins are relevant to the development of schistosome vaccines . In mice immunized with attenuated cercariae challenge schistosomula are eliminated in the lungs by inflammatory foci that block their onward migration [43] . Both priming of the immune response [44] in skin-draining lymph nodes and the pulmonary effector responses appear to be mediated by proteins on the schistosomular tegument surface [45] . Although we have characterised surface proteins on the adult worm , the presence of the encoding mRNAs in the lung schistosomulum provides support for the suggestion that those same proteins are exposed on the larval tegument surface . In this context Sm29 has already been used successfully as a vaccine candidate [17] . We therefore suggest that the other six PiPLC-anchored proteins would also repay investigation as putative vaccine candidates , especially Sm200 because it was also removed by trypsin . Indeed , the presence of a GPI-anchor may make proteins in this group more prone to detachment from the surface for transcytosis across the pulmonary capillary endothelium , processing by accessory cells and presentation to reactive CD4+ Th1 cells in the lung parenchyma . To this list we can add the three annexins released by trypsin and LMWP that may be a tegument secretion , noting also that calpain is a proposed vaccine candidate [31] . It is conceivable that a cocktail of these proteins would elicit strong protection against intravascular migrating schistosomula , especially if targeted to the lungs , whereas immunisation with a single protein has only modest success [46] .
|
Adult schistosome parasites can reside in the host bloodstream for decades surrounded by components of the immune system . It was originally proposed that their survival depended on the secretion of an inert bilayer , the membranocalyx , to protect the underlying plasma membrane from attack . We have investigated whether any proteins were exposed on the surface of live worms using incubation with selected hydrolases , in combination with mass spectrometry to identify released proteins . We show that a small number of parasite proteins are accessible to the enzymes and so could represent constituents of the membranocalyx . We also identified several proteins acquired by the parasite on contact with host cells . In addition , components of the cytolytic complement pathway were detected , but these appeared not to harm the worm , indicating that some of its own surface proteins could inhibit the lytic pathway . We suggest that , collectively , the ‘superficial’ parasite proteins may provide good candidates for a schistosome vaccine .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"infectious",
"diseases/helminth",
"infections",
"chemical",
"biology/protein",
"chemistry",
"and",
"proteomics"
] |
2011
|
Enzymatic Shaving of the Tegument Surface of Live Schistosomes for Proteomic Analysis: A Rational Approach to Select Vaccine Candidates
|
The Type III Secretion System ( T3SS ) is a macromolecular complex used by Gram-negative bacteria to secrete effector proteins from the cytoplasm across the bacterial envelope in a single step . For many pathogens , the T3SS is an essential virulence factor that enables the bacteria to interact with and manipulate their respective host . A characteristic structural feature of the T3SS is the needle complex ( NC ) . The NC resembles a syringe with a basal body spanning both bacterial membranes and a long needle-like structure that protrudes from the bacterium . Based on the paradigm of a syringe-like mechanism , it is generally assumed that effectors and translocators are unfolded and secreted from the bacterial cytoplasm through the basal body and needle channel . Despite extensive research on T3SS , this hypothesis lacks experimental evidence and the mechanism of secretion is not fully understood . In order to elucidate details of the T3SS secretion mechanism , we generated fusion proteins consisting of a T3SS substrate and a bulky protein containing a knotted motif . Because the knot cannot be unfolded , these fusions are accepted as T3SS substrates but remain inside the NC channel and obstruct the T3SS . To our knowledge , this is the first time substrate fusions have been visualized together with isolated NCs and we demonstrate that substrate proteins are secreted directly through the channel with their N-terminus first . The channel physically encloses the fusion protein and shields it from a protease and chemical modifications . Our results corroborate an elementary understanding of how the T3SS works and provide a powerful tool for in situ-structural investigations in the future . This approach might also be applicable to other protein secretion systems that require unfolding of their substrates prior to secretion .
T3SS are found in numerous Gram-negative bacteria and share strong homologies among different invasive pathogens . Using the T3SS , bacteria are able to secrete effector proteins that translocate into the host-cell where they target metabolic or signal transduction pathways for example [1][2] . The T3SS is a key instrument in interactions between bacteria and eukaryotes , as it is used also by symbiotic bacteria in plants [3][4] . One pathogen depending on T3SS-mediated virulence is Shigella flexneri , a human pathogen of the intestine that depends on effector protein delivery to establish an infection . In S . flexneri serovar 5a M90T , the T3SS is encoded on a 210 kb-extrachromosomal plasmid [5] , where genes encoding the NC are clustered in distinct operons . About 25 genes that lie in the membrane-expression of invasion plasmid antigen ( mxi ) -locus and surface presentation of antigen ( spa ) -locus constitute the NC . Together with the invasion plasmid antigen ( ipa ) -operon , this region is referred to as the entry region which is necessary and sufficient for invasion of host cells [6][7] . The translocators IpaB and IpaD regulate secretion by forming a complex at the tip of the needle [8] and deletion of either ipaB or ipaD causes hypersecretion of effectors [9] . As for other bacterial T3SS [10][11] , the NC from S . flexneri shows striking structural similarity to a syringe with a basal body and a needle-like hollow tube that can be isolated from the bacterial envelope [12] . The basal body is made of stacked protein rings that are inserted into the inner and outer membrane . Together , these rings form a conduit which narrows into the needle that protrudes from the bacterium . The needle is made of many copies of one small subunit protein that assembles into a helical tube . Both the basal body and the needle form a continuous channel that ranges from the bacterial cytoplasm to the extracellular environment . The inner diameter of the needle channel was estimated to be 2–3 nm [12][13] , and recent structural analysis defined a 2 . 5 nm channel in Salmonella enterica serovar Typhimurium SPI-1 with a conserved architecture in S . flexneri [14][15] . Type III secretion mechanism has been studied using fusions of effectors and stably folded protein domains [16][17][18][19][20] . For example , fusion proteins consisting of Yersinia enterocolitica effectors fused to dihydrofolate reductase ( DHFR ) or ubiquitin either obstructed the T3SS [17][18][19] or were rejected [18] . If the fold of DHFR or ubiquitin was destabilized by mutations or by the action of a chaperone , fusions were readily secreted by the T3SS [17][18] . This implies that fused effectors can be secreted by the T3SS if the substrate is unfolded prior to secretion . Taken together , whether or not a T3SS substrate is compliant for secretion through the NC channel seems to depend on its fold and structural stability . These data support the hypothesis that effector proteins need to be unfolded in order to be efficiently secreted through the needle channel . While the model of secretion through the NC channel has been widely accepted , no experimental evidence for this model exists [21][22] . Furthermore , the whole idea of the T3SS working as a microsyringe injecting effectors into the host cell has been questioned [22] . A substrate has neither been pictured in contact with the NC nor has an actively secreting NC been used for structural investigations . Therefore , our strategy was to trap a substrate inside the S . flexneri NC , using a fusion protein which cannot be unfolded by the T3SS . We designed fusion proteins that consist of the translocator IpaB and the RNA 2′-O-ribose methyltransferase RrmA ( PDB ID 1IPA ) . RrmA has a trefoil-knot in its C-terminal region [23] and we will refer to RrmA as “Knot” . Here , we present a direct visualization of the NC together with IpaB-Knot . We show that the NC channel physically encloses its substrate and experimentally confirm a substantial hypothesis of the T3SS secretion mechanism . To our knowledge , this is the first demonstration of a substrate being transported through the NC channel .
IpaB is a multifunctional protein that induces pyroptosis in macrophages by lysosomal leakage and activation of Caspase-1 [24][25] . We constructed translational fusions consisting of IpaB followed by the Knot ( IpaB-Knot ) with either a leucine-glutamine linker , a TEV protease site between both proteins or the Knot expressed without IpaB-fusion ( Fig . 1A ) . To investigate whether IpaB was still functional when fused to the Knot , induction of pyroptosis by purified IpaB-Knot was quantified in a cytotoxicity assay . Induction of pyroptosis was measured by LDH released from murine bone marrow-derived macrophages ( mBMM ) that were treated with increasing concentrations of IpaB-Knot ( Fig . 1B ) . The fusion protein used in this assay was free of contaminants and wildtype IpaB ( e . g . resulting from fusion protein instability ) which could give false positive results ( Fig . S1 ) . We observed a similar dose-dependent release of LDH with IpaB-Knot as has been demonstrated for IpaB alone [25] . 50 µg/ml of recombinant fusion protein caused LDH release as efficiently as 1% detergent ( Triton X-100 ) . These results demonstrate that IpaB is still functional when fused to the Knot . We also addressed the folding status of the Knot-domain in IpaB-Knot by measuring the change in tryptophan fluorescence upon unfolding ( Fig . 1C ) . IpaB contains a single tryptophan whereas the Knot domain has four tryptophans , three of which are part of the trefoil-knot in the C-terminal region . This indicates that the fluorescence emission spectrum mainly represents the folding status of the Knot . We compared spectra of purified IpaB-Knot and the Knot alone , either in buffer or in the presence of the chaotropic salt Guanidine HCl ( GuHCl ) . The native Knot protein has a fluorescence emission peak at 335 nm which represents tryptophans in the folded protein ( Fig . 1C , solid grey line ) . The fluorescence spectrum of IpaB-Knot is similar to the native Knot spectrum with an emission peak at 335 nm ( Fig . 1C , solid black line ) , suggesting that the Knot is folded when fused to IpaB . Protein unfolding in the presence of 6 M GuHCl leads to a red shift of the emission peak from 335 nm to 355 nm which indicates a change of the molecular environment of the respective tryptophans in both IpaB-Knot and Knot alone . This change in fluorescence is caused by structural changes of the protein and demonstrates that the Knot domain was folded in both cases and is sensitive to unfolding by GuHCl . The analysis of folded parts of the Knot domain in IpaB-Knot was in line with these findings . Protein fragments were obtained by treatment with the non-specific protease Proteinase K . Protein fragments protected from cleavage by their native fold were separated by 2D-gel electrophoresis and further analyzed by mass spectrometry ( MS ) according to the procedure described by Jungblut et al . [26][27] . We detected polypeptides that cover almost the entire sequence of the Knot domain and the complete sequence of the trefoil-knot motif . Subfragments were confirmed by MS/MS analysis ( Fig . S2 ) . Only the C-terminal helix and N-terminal parts of the Knot domain , but not the trefoil-knot , were cleaved , suggesting that this motif is folded and consequently inaccessible to Proteinase K . Together with our findings from fluorescence analysis , these results confirm a tightly folded core domain covering the trefoil knot-motif in IpaB-Knot which is inaccessible to Proteinase K and sensitive to unfolding by the chaotropic agent GuHCl . IpaB is essential during S . flexneri invasion of epithelial cells [7] . Since IpaB-Knot is functional , we analyzed whether bacteria with a genomic ipaBknot fusion allele were still invasive compared to the wildtype . The fusion ipaBknot was introduced by insertion of the knot-gene ( together with a 3′ strep-tag ) downstream of the ipaB-open reading frame in S . flexneri M90T . Hence , ipaBknot was expressed under the native promoter of the ipgCipaBCDA-operon ( M90T::ipaBknot ) . M90T::ipaBknot was compared to wildtype M90T and the invasion-deficient ipaB strain . Invasiveness was quantified in a gentamicin protection assay [28] where bacteria are allowed to replicate inside host cells after invasion . Clearly , M90T::ipaBknot is attenuated in invasion and is not significantly different from the negative control ipaB ( Fig . 2A ) . This attenuation does not result from altered expression of IpaB-Knot compared to IpaB in M90T ( Fig . 2B ) or a deficit in expressing structural proteins of the NC . Levels of IpaB are equal in all strains tested as were levels of MxiG which constitutes the inner membrane ring , an integral part of the needle complex [12] . As we have already demonstrated that IpaB is functional in IpaB-Knot , the non-invasive phenotype might result from effects on the T3SS pathway . In order to elucidate effects on secretion , we inserted the functional fusion encoding ipaBknot in the virulence plasmid of an ipaD-deficient strain ( ipaD::ipaBknot ) . The ipaD strain secretes IpaB as well as other effectors without the need for T3SS induction [9] . Therefore , we compared protein secretion in ipaD , ipaD::ipaBknot , and a secretion-deficient negative control ( mxiHipaD ) . Secreted proteins were analyzed by Coomassie stain of SDS PAGE ( Fig . 3A ) and Western blot ( Fig . 3B ) . Supernatants from the ipaD strain show a pattern of various proteins secreted . In contrast , secretion of proteins in the ipaD::ipaBknot strain was dramatically reduced . Attenuation of secretion was limited to T3SS-dependent proteins , since secretion of SepA , a substrate to a type V secretion system [29] , was not influenced by IpaB-Knot ( Fig . 3A , 110 kDa and 3B , upper panel ) . Western blot analysis supported our findings of limited secretion in the ipaD::ipaBknot strain ( Fig . 3B ) . We probed for the translocator IpaC in supernatants and bacterial lysates from ipaD and ipaD::ipaBknot . Reduction of secreted IpaC in supernatants was not due to a diminished synthesis , as intracellular levels of the protein were equal in all three strains . Importantly , IpaB was secreted by the ipaD mutant whereas IpaB-Knot was only detected in bacterial lysates of ipaD::ipaBknot , indicating its cytoplasmic localization ( Fig . 3B ) . We conclude that fusing the knot to IpaB prevents secretion of IpaB itself and impedes secretion of other T3SS effectors . We wondered whether additional effector-knot fusions could block the T3SS , or if the phenotype we observed was limited to IpaB-Knot . Therefore , we fused the knot to the early secreted effector IpaA and the late secreted effector IpaH7 . 8 [30] . As seen with IpaB-Knot , IpaA-Knot and IpaH7 . 8-Knot fusions dramatically reduced secretion of effectors ( Fig . 3C . We further delineated secretion by Western blot analysis and found that IpaA-Knot and IpaH7 . 8-Knot remained cytosolic while blocking secretion of other effector proteins ( Fig . 3D ) . The overall reduction in secretion was not due to a lack of T3SS needle complexes as the expression of T3SS structural components , indicated by MxiG , was not altered by expression of substrate-knot fusions . Interestingly , secretion of IpaC was attenuated in ΔipaD::ipaBknot and ipaD::ipaAknot strains but not in the late effector-fusion allele ipaD::ipaH7 . 8knot . As secretion of early effectors is a prerequisite for transcription of late effectors , this likely reflects the hierarchy of secretion that is characteristic of Shigella . Taken together these results suggest that effector-Knot fusions can be used as a general tool to obstruct T3SS . As a control , the Knot alone was expressed by insertion of the knot-encoding gene at the same locus without fusion of the ipaB-gene ( ipaD::knot , Fig . 1A ) . knot was inserted with a stop-codon for ipaB together with the ipaC ribosomal binding site mediating translation . In these experiments , secretion of T3SS effectors from ipaD::knot were the same as for ipaD ( Fig . S3A ) . We found that expression of the knot alone did not affect the T3SS pathway , including effector synthesis , nor did it exert toxicity , as IpaB and MxiG were equal to ipaD ( Fig . S3B ) . In summary , only the fusion of IpaB to the Knot reduced effector secretion by the T3SS in the otherwise hypersecreting ipaD strain . Next , we studied the interaction of IpaB-Knot and NCs from ipaD::ipaBknot by cesium chloride density fractionation . NCs were isolated from the bacterial membrane and separated from soluble proteins by centrifugation as it migrates to high-density fractions . Isolated NCs were detected with an anti-MxiG antibody . In line with the hypothesis of IpaB being unable to bind to the NC in the ipaD strain [8] , residual IpaB was separated from the NCs . IpaB remained in the low-density fraction whereas NCs migrated to high-density fractions ( Fig . 4 , left panel , ipaD ) . In contrast , IpaB-Knot migrated with isolated NCs from ipaD::ipaBknot to high-density fractions , which indicates an interaction with the NC ( Fig . 4 , middle panel , ipaD::ipaBknot ) . As a control , IpaB-Knot from recombinant expression ( rIpaB-Knot ) was added to purified NCs from ipaD cells ( Fig . 4 , right panel , ipaD+rIpaBKnot ) . Purified IpaB-Knot remained in low-density fractions and did not migrate with NCs to high-density fractions . A co-migration was only observed for endogenous IpaB-Knot and not purified IpaB-Knot that was added to the NCs . Therefore , interaction of IpaB-Knot and the NCs from ipaD::ipaBknot most likely results from its attempted secretion and subsequent obstruction of the T3SS by IpaB-Knot . Based on our previous results , we tested if IpaB-Knot is arrested inside the NC channel as the Knot of the fusion remains folded and is impassable for the T3SS . Therefore , we analyzed the presence of IpaB-Knot with isolated NC channels after density fractionation using immuno-electron microscopy ( iEM ) . Fractions that contained both IpaB-Knot and MxiG ( Fig . 3 , middle panel , fractions 4 and 5 ) were analyzed with an anti-IpaB monoclonal ( H16 ) and an anti-Strep tag monoclonal antibody . The anti-IpaB monoclonal antibody recognizes an N-terminal stretch between the residues 118 and 179 [31] . The IpaB domain of IpaB-Knot was detected either at the needle tip or at the base of isolated NCs ( Fig . 5A and Figure S4 ) , whereas the C-terminal Strep-tag was only detected at NC bases ( Fig . 5B ) . IpaB was not detected in samples of isolated NCs from the ipaD strain , which is consistent with our finding that IpaB does not co-purify in this background ( Fig . 4 , left panel ) . To assess whether the observed co-localization between IpaB-Knot and NC is specific , 50 random iEM images were quantified and coordinates for gold labels against IpaB and NC centers recorded . We found an average of 80 NC and 20 labels per image ( Fig . S5 ) . In order to obtain a random distribution of gold and NCs , respective coordinate X and Y values were randomized according to counts per image within the range of the image dimensions . Both counted or simulated coordinates were subjected to a nearest-neighbor analysis , quantifying distances between NC centers and the closest adjacent gold particles . The distribution of counted NC/gold distances peaks around 30–50 nm . In comparison , distances obtained from the random distribution peaked around 100 nm–120 nm . We also labeled NCs with anti-IpaB and anti-Strep tag antibodies simultaneously . The anti-IpaB monoclonal antibody was expressed in a human cell line which resulted in a human Fc region . After labeling the NC with these primary antibodies , anti-human secondary antibody coupled to 6 nm gold and anti-mouse secondary antibody coupled to 12 nm gold were used . In a few cases , simultaneous labeling of both epitopes at the same NC was observed ( Fig . 5D ) , with IpaB-labeling at the tip and Strep-labeling at the base . These results show co-localization of IpaB-Knot with the NC , resulting from a stable interaction between IpaB-Knot and NC . We examined whether IpaB-Knot was inserted into the NC channel . According to the hypothesis that secretion of effectors and translocators occurs through the channel , part of IpaB-Knot should be enclosed and therefore be inaccessible to modifications . Our construct features a TEV protease-specific cleavage site between IpaB and the Knot domain together with a C-terminal Strep-tag ( Fig . 1A ) . We would predict that , once IpaB-TEV-Knot is cleaved by TEV protease , IpaB' will be released from the beads into the sample supernatant . IpaB-TEV-Knot from ipaD::ipaBTEVknot NC isolates and IpaB-TEV-Knot purified from E . coli BL21 were coupled to Strep-Tactin sepharose beads via the Strep-tag and subsequently treated with TEV protease . Both supernatants and bead fractions were analyzed by Western blot after protease treatment using IpaB and MxiG antibodies . IpaB-TEV-Knot co-purified with NCs showed almost no cleavage products ( IpaB' ) after treatment with increasing concentrations of TEV protease ( Fig . 6A , upper panel ) . The recombinant IpaB-TEV-Knot , however , was efficiently cleaved at low concentrations of TEV protease ( Fig . 6A , lower panel ) . As IpaB-TEV-Knot decreased proportionally with increasing TEV protease concentrations , released IpaB' accumulated in supernatants and increased in direct correlation with the amount of protease . Some IpaB' attached to the beads after cleavage which could be due to a non-specific interaction between IpaB' and the beads . We conclude that IpaB-TEV-Knot co-purified with NC is largely protected from enzymatic cleavage by TEV protease . Cleavage of purified IpaB-TEV-Knot indicates an accessible TEV recognition site which was only observed for purified IpaB-TEV-Knot alone . We also investigated whether IpaB-Knot can be modified by the addition of crosslinking PEG molecules ( PEGylation ) . Each PEG adds 1 . 1 kDa to the proteins molecular weight and has an arm length of 8 . 8 nm and therefore cannot penetrate or diffuse into the NC channel for sterical reasons . PEGylation occurs via crosslinking of primary amines ( lysine residues and N-terminus ) to the NHS-group of PEG . IpaB-Knot has 58 PEGylation sites in total which are evenly distributed across the protein . Similar to the TEV protease assay , Strep-Tactin beads were saturated with either purified IpaB-Knot or isolated NCs containing IpaB-Knot and PEGylated . Samples were analyzed using a Strep-tag antibody since the tag has no PEGylation sites and the antibody epitope is not affected . We observed a size shift of PEGylated IpaB-Knot to about 110 kDa , which corresponds to PEGylation at 20–30 amines ( Fig . 6B , left panel ) . On the other hand , PEGylation of IpaB-Knot isolated together with NCs occured to a lesser extent as the protein shifted not more than 10 kDa ( Fig . 6B , right panel ) . We hypothesize that the difference is due to inaccessibility of amines resulting from the presence of the NC channel . Together with the results from the TEV protease-assay , we conclude that protection is conferred by the NC channel which surrounds IpaB-TEV-Knot and partially covers the protein .
In this study we demonstrate that T3SS substrate proteins travel through the NC channel during type III secretion . While largely assumed , our experiments provide convincing evidence for this model . Based on the strong conservation of the NC among different bacterial species [10][11] , we would predict that our findings are of general relevance for T3 secretion . We generated fusion proteins consisting of the translocator IpaB as a T3SS substrate and a protein with a trefoil-knot in its C-terminal region ( Knot ) . We show that IpaB-Knot has a functional IpaB domain and a folded Knot domain which indicates that the fusion protein is folded prior to secretion . However , the fusion protein attenuates invasion as observed for an ipaB-deficient mutant . In order to pass through the narrow NC channel , the T3SS needs to unfold the Knot . We observed that IpaB-Knot inhibits secretion of T3SS translocators and effectors in the hypersecretor S . flexneri ipaD . Obstruction may occur because the T3SS cannot unfold the Knot domain within IpaB-Knot and consequently , IpaB-Knot blocks the channel . Our results are supported by previous reports which show a direct correlation between T3SS secretion and the ability of a protein to be thoroughly inserted into the channel and its folding status [17][18] . As IpaB-Knot obstructed the effector secretion , we next analyzed the interaction between IpaB-Knot and NCs . Interaction was indicated as IpaB-Knot migrated with NCs in gradient fractionations . This co-migration was limited to endogenous IpaB-Knot and therefore takes place only inside the bacterium as it could not be restored by mixing purified IpaB-Knot with isolated NCs . This suggests an active mechanism mediating this interaction , probably also with the contribution of other factors . At this point , more experimental work is required to elucidate the underlying mechanism . IpaB-Knot was detected with isolated NCs and this co-localization was shown to be specific . To our knowledge , this is the first visualization of a T3SS substrate together with the NC . Based on the N-terminal signal sequence and chaperone binding domains , the N-terminus of effectors provides a T3SS-specific secretion signal [21][32] . We found that IpaB-Knot is at least partially secreted as its N-terminus could be detected at the tips of isolated NCs . This implies that , besides carrying the signal for secretion , the N-terminus of IpaB is also secreted first since our fusion does not allow secretion of IpaB by the C-terminus . We show that the channel physically encloses the fusion protein . An embedded TEV protease-cleavage site in IpaB-TEV-Knot isolated with NCs was inaccessible to TEV protease . Furthermore , the protein is partially inaccessible to PEGylation . From these experiments we conclude that IpaB-TEV-Knot is secreted through the NC channel . This is supported by our observation that the interaction between the fusion and the NC is stable . We isolated the NCs with IpaB-TEV-Knot based on the interaction of a C-terminal Strep-tag in IpaB-TEV-Knot . These results confirm the linear secretion model as well as secretion through the needle of the NC . We provide a tool for structural investigations of the NC in a stabilized active state and in association with a substrate protein . This allows the possibility to identify novel interaction partners for the T3SS substrates , which may still be associated with the complex at the stage where secretion is arrested . Also , the interaction of the substrate with structural components inside the channel can be studied in detail . This approach might also be used for other secretion systems , where proteins are required to unfold in order to be secreted . Ultimately , this contributes to the understanding of an essential mechanism of bacterial virulence which in turn might lead to novel antibacterial strategies .
The bacterial strains and plasmids used in this study are listed in table 1 . Bacterial cultures were grown in Tryptic Soy Broth ( TSB ) medium supplemented with antibiotics as selection markers . Strains were kept on TSB agar plates supplemented with antibiotics and 0 . 01% ( w/v ) Congo red ( Sigma ) . Mutant strains were generated following the protocol from Datsenko and Wanner [33] using the pKD3 for chloramphenicol cassette amplification . The cassettes were flanked by homologous regions of the insertion sites which were included in the oligonucleotides for cassette amplification . For gene inactivation , the open reading frame was replaced with the antibiotic cassette in the bacterial genome . For fusion constructs , the fusion partner-gene was ligated to the antibiotic cassette by PCR and inserted into the genome . Oligonucleotides used for genetic modifications of bacteria are listed in table 2 . Bacteria were grown over night and cell density ( OD ) was measured by absorption at 600 nm . Culture aliquots corresponding to an optical density of 2/ml ( about bacteria ) were harvested and the supernatant was saved . Bacteria were washed once with phosphate buffered saline ( PBS ) and resuspended in SDS sample buffer . The supernatant was filtered ( ) and proteins were precipitated with 10% ( v/v ) Trichloroacetic acid for 5 min at and harvested subsequently by centrifugation at 16 . 100 rcf for 30 min at . Pellets were washed once with 1 ml ice-cold acetone and centrifuged again for 30 min . Acetone was discarded directly after centrifugation and the pellets dried at room temperature and resuspended in 3 M TrisHCl , pH 8 . 5 , plus SDS sample buffer . Supernatants and 10% of the equivalent bacterial lysates were analyzed by SDS PAGE and Western blotting . Invasion of Shigella strains was quantified as described previously [34] . HeLa cells per well were seeded in 24-well plates and grown overnight in Dulbeccos Modified Eagle Medium ( DMEM , Life Technologies ) +10% fetal calf serum ( FCS ) . Cell culture medium was exchanged the next day to DMEM without FCS ( serum-free medium , SFM ) . HeLa cells were infected with bacteria ( log phase ) in PBS for 30 min at and centrifuged at 1 . 340 rcf for 10 min . Medium was exchanged to SFM+gentamicin to avoid continuous reinfection and bacteria were allowed to replicate inside cells for 2 h at 37 . HeLa cells were then washed with PBS and lysed in the presence of 1% ( v/v ) Triton X-100 in PBS and bacteria were plated on LB agar . Bacterial invasion was quantified by number of colony-forming units ( CFU ) per ml culture after overnight incubation at . All SDS PAGE ( except 2D SDS PAGE ) were performed using TrisHCl gradient gels ( Criterion AnykD or 4–20% , Bio-Rad ) . Proteins were transferred on nitrocellulose membrane ( GE Healthcare ) with for 1 h and blocked with 5% ( w/v ) milk powder in PBS with 0 . 05% ( v/v ) Tween-20 . Proteins were detected using monoclonal IpaB antibody ( H16 ) [31] , monoclonal MxiG antibody ( 7G1 , this study ) , monoclonal DnaK antibody ( Stressgen ) , monoclonal Strep-tag antibody ( QIAGEN ) or polyclonal IpaC antibody ( Institut Pasteur ) . Visualization was performed using secondary antibodies coupled to horse radish peroxidase and enhanced chemiluminescence using a DuraStable West-Kit ( Thermo Scientific ) . IpaB-Knot or IpaB-Knot in complex with IpgC were purified as previously described [32][35] . E . coli BL21 co-transformed with pET28a::ipgC ( with a 3′ 6his-tag ) and pASK-IBA3+::ipaBknot ( with a 3′ strep-tag ) were grown to early log-phase and expression of ipgC was induced with 0 . 5 mM ( IPTG , Thermo Fisher ) for 30 min . Then , expression of ipaBknot or ipaB was induced with ( w/v ) anhydrotetracycline ( Thermo Fisher ) . IpgC and IpaB-Knot was co-purified using 5 ml HisTrap HP cartridges ( GE Healthcare ) according to manufacturers instructions . The complex binds to the resin via IpgC and IpaB-Knot is eluted with 0 . 1% ( w/v ) Lauryldimethylamine N-oxide ( LDAO ) . For fluorescence spectrometry and Proteinase K digestion , full complex of IpgC/IpaB-Knot was eluted with imidazole . Additional affinity chromatography in the presence of 0 . 05% LDAO using 5 ml Strep-Tactin columns ( IBA ) was performed to separate full length IpaB-Knot from IpaB contaminants which might occur because of instability of the fusion protein . Native protein was separated from aggregates by size exclusion chromatography using a Superdex 200 16/60 column ( GE Healthcare , Freiburg ) with 20 mM 4- ( 2-hydroxyethyl ) -1-piperazineethanesulfonic acid ( HEPES ) , pH 7 . 4 , and 150 mM NaCl . The knot was expressed with a C-terminal His-tag ( pASK-IBA33+::knot ) in E . coli BL21 and purified via a 5 ml HisTrap HP column in 20 mM TrisHCl , pH 7 . 4 , 300 mM NaCl , 1 mM Dithiothreitol ( DTT ) , 40 mM Imidazole ( ultra-pure ) and eluted with 400 mM Imidazole under same buffer conditions . Proteins were concentrated to 1–2 mg/ml . IpaB cytotoxicity was quantified as described before [25] . Murine bone marrow macrophages ( mBMM ) from C57BL/6J mice ( Jackson Lab ) were seeded in 96-well plates ( mBMM/well ) in DMEM with 2% FCS and incubated overnight . Medium was replaced with DMEM with reduced FCS ( 0 . 5% ) . Different concentrations of recombinant IpaB-Knot or bovine serum albumin ( BSA ) were added to cells . The cells were then incubated for 2 h at . Supernatants were tested for LDH using a colorimetric assay ( Promega ) according to the manufacturers instructions . Purified protein ( either IpaB-Knot in complex with IpgC or Knot ) was used at concentrations at . Samples were excited with 295 nm monochromatic light using 2 . 5 nm or 5 nm slit-width . Emission spectra were recorded at 300–400 nm at a slit-width of 5 nm or 10 nm . Control spectra ( buffer with IpgC as a control for IpaB-Knot in complex with IpgC ) were subtracted from sample protein spectra . Emissions were normalized to their respective maximum . Purified IpaB-Knot in complex with IpgC was treated with Proteinase K at a molar ratio of 1∶100 ( enzyme to substrate ) for 1 h at room temperature . The reaction was stopped by the addition of complete EDTA-free protease inhibitor cocktail ( Roche ) and the cleavage products were separated by 2D SDS PAGE . Respective protein cleavage products were trypsinated and analyzed by MS , whereas fragments of specific interest were confirmed by MS/MS . Needle complexes were isolated from bacteria as described previously [12] . 2 L of bacterial culture were inoculated 1∶50 and grown to an optical density of 1 . 2–1 . 6/ml and washed once with PBS . Cells were osmotically shocked by resuspension in 0 . 5 M sucrose . 0 . 1 M TrisHCl , pH 8 . 0 and 5 mM EDTA were added . Bacteria were incubated in the presence of 1–2 mg lysozyme ( Novagen ) for 30–60 min and lysed with 2% ( v/v ) Triton X-100 . Debris was removed by centrifugation for 20 min at 45 . 000 rcf and needle complexes were harvested by pelleting the supernatant at 110 . 000 rcf for 1 h . The pellet was washed in 10 mM TrisHCl pH 8 . 0 , 150 mM KCl , 5 mM EDTA , 1% ( v/v ) Triton X-100 and 0 . 3% ( w/v ) Sarcosyl . Centrifugation steps were repeated and pelleted needle complexes were resuspended in 1–2 ml 50 mM TrisHCl , pH 8 . 0 , 50 mM EDTA , 0 . 1% Triton X-100 ( TET buffer ) over night . Needle complexes were further separated either by CsCl gradient centrifugation ( 27 . 5% ( w/v ) CsCl in TET ) and size exclusion chromatography ( Superose 6 16/60 , GE Healthcare , in TE buffer ) or affinity purifaction via Strep-Tactin sepharose ( IBA ) in TET via the C-terminal Strep-tag of IpaB-Knot . Needle complexes were eluted with TET +5 mM desthiobiotin . aliquots of the needle complex preparations were applied to freshly glow-discharged carbon- and pioloform-film-coated copper grids and allowed to adsorb for 10 min . The grids were then washed with PBS , blocked for 15 min and incubated with mouse or human IpaB monoclonal antibody , anti-Strep antibody or both for 2–3 h at room temperature . After washing with PBS the samples were incubated with goat-anti-mouse antibody adsorbed to 12 nm gold particles for 30 min and washed with PBS . After three washes with distilled water , the grids were contrasted with 4% phospho-tungstic-acid ( PTA ) , 1% Trehalose , pH 7 . 0 , touched onto filter paper and air-dried . The grids were examined in a LEO 906 ( Zeiss AG ) electron microscope operated at 100 kV and images were recorded with a Morada ( SIS-Olympus ) digital camera . Coordinates for gold labels and needle complex centers from 50 iEM images were manually obtained using the ImageJ analysis package [36] . The values were processed by R [37] and subjected to a nearest-neighbor analysis with the Biobase package [38] . Isolated needle complexes or recombinant protein was added to 100 Strep-Tactin beads ( IBA ) and incubated for 1–2 h at with gentle rotation ( 12–15 rpm ) . Beads were washed twice with TET buffer and subsequently washed twice with 1 ml 50 mM Tris , pH 8 . 0 , 0 . 5 mM EDTA , 1 mM DTT ( TEV assay buffer ) and finally resuspended with TEV assay buffer . 4 aliquots of 40 µl bead slurry were adjusted to room temperature and subsequently incubated with either TEV assay buffer ( negative control ) , , or TEV protease at 400 rpm at . The beads were collected by centrifuging at 2000 rcf for 2 min . aliquots of supernatant were collected and supplemented with SDS sample buffer . The bead fraction was washed 2 times with 1 ml TEV assay buffer and the beads resuspended in TEV assay buffer and supplemented with SDS sample buffer . Samples were analyzed by SDS PAGE and Western blotting . Isolated needle complexes or recombinant protein was added to Strep-Tactin beads ( IBA ) and incubated for 1–2 h at 4°C with gentle rotation ( 12–15 rpm ) . Beads were washed twice with TET buffer and subsequently washed twice with PBS and finally resuspended with 50 µl PBS . 2 aliquots of bead slurry were adjusted to room temperature and subsequently incubated with 0 . 5 mM PEGylation reagent MS ( PEG ) 24 ( Thermo Scientific ) for 1 h at room temperature . The reaction was quenched with the addition of 100 mM TrisHCl for 5 min at room temperature and the reaction was directly incubated with SDS sample buffer and analyzed by SDS PAGE and Western blotting .
|
Type III Secretion Systems ( T3SS ) secrete bacterial effector proteins from the cytoplasm across the cell wall , but mechanistic details of this process remain mostly elusive . We locked the T3SS of Shigella flexneri in an actively secreting state by expression of substrate fusions that consist of a functional translocator and a stably-folded knotted protein . Although recognized as T3SS substrates , the fusions are not released from secreting Shigella and impede transport of other effector proteins by obstructing the T3SS channel . We localized the fusion at isolated channels and observed that the translocator is secreted with the N-terminus first . We further demonstrate that the channel physically encloses the partially transported substrate . Our analysis elucidates important steps of the T3SS mechanism . Furthermore , we developed fusion proteins useful for advanced structural investigations of one of the most complex bacterial virulence devices known and our approach may help to also understand other protein transport mechanisms .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"bacteriology",
"medicine",
"shigellosis",
"infections",
"protein",
"interactions",
"macromolecular",
"assemblies",
"host-pathogen",
"interaction",
"microbiology",
"globular",
"proteins",
"bacterial",
"biochemistry",
"protein",
"classes",
"bacterial",
"diseases",
"protein",
"folding",
"protein",
"structure",
"bacterial",
"pathogens",
"infectious",
"diseases",
"inflammation",
"proteins",
"microbial",
"pathogens",
"structural",
"proteins",
"biology",
"recombinant",
"proteins",
"biophysics",
"biochemistry",
"protein",
"chemistry",
"gram",
"negative",
"transmembrane",
"proteins",
"immunity",
"transmembrane",
"transport",
"proteins"
] |
2014
|
A Substrate-Fusion Protein Is Trapped inside the Type III Secretion System Channel in Shigella flexneri
|
Helminth infections have proven recalcitrant to control by chemotherapy in many parts of Southeast Asia and indeed farther afield . This study isolates and examines the influence of different aspects of the physical and social environment , and uneven intervention effort contributing to the pathogenic landscape of human Opisthorchis viverrini infections . A cross-sectional survey , involving 632 participants , was conducted in four villages in northeast Thailand to examine the impact on prevalence and parasite burden of the reservoir dam environment , socio-economic , demographic , and behavioral factors , and health center intervention efforts . Formalin-ether concentration technique was used for diagnoses , and multivariate models were used for analyses . The importance attributed to O . viverrini infections varied among health centers in the four study villages . Villages where O . viverrini infections were not prioritized by the health centers as the healthcare focus were at a higher risk of infection ( prevalence ) with odds ratio ( risk factor ) of 5 . 73 ( 3 . 32–10 . 27 ) and p-value < 0 . 01 . Priority of healthcare focus , however , did not appear to influence behavior , as the consumption of raw fish , the main source of O . viverrini infections in the study area , was 11 . 4% higher in villages that prioritized O . viverrini infections than those that did not ( p-value = 0 . 01 ) . Landscape variation , notably proximity to reservoir , affects vulnerability of local population to infection . Infection intensity was higher in population located closer to the reservoir with risk ratio of 2 . 09 ( 1 . 12–4 . 02 ) and p-value < 0 . 01 . Patterns of infection intensities among humans were found to match fish infection intensity , where higher infection intensities were associated with fish obtained from the reservoir waterbody type ( p-value = 0 . 023 ) . This study demonstrated the importance of environmental influence and healthcare focus as risk factors of infections in addition to the socio-economic , demographic , and behavioral factors commonly explored in existing studies . The reservoir was identified as a crucial source to target for opisthorchiasis intervention efforts and the need to consider infection intensity in disease control efforts was highlighted . The holistic approach in this study , which underscores the close relationship between the environment , animals , and humans in development of human infections or diseases , is an important contribution to the framework of One Health approach , where consideration of helminth diseases has largely been overlooked .
Helminthiases , which include foodborne trematodiases , lymphatic filariasis , schistosomiasis , and soil-transmitted helminthiases , are the most common neglected tropical diseases ( NTDs ) in southeast Asia [1] . They disproportionally affect the poor or marginalized population in developing countries , trapping the afflicted in a vicious cycle of poor health outcomes and poverty , and costing billions of dollars in treatment each year [2] . The increasing recognition of the burden caused by helminth infections has brought about large-scale control programs by the World Health Organization and other nationwide control programs in countries in Asia [3 , 4] , Latin America [5] , and sub-Saharan Africa [6] , where helminthiases are prevalent . These programs have primarily relied upon chemotherapy for helminthiases control [7] . Many chemotherapy programs have relatively limited objectives , resulting in reduced infection levels in the short-term [8] . Re-emergence of the disease , and possibly even development of resistant strains of parasites , is common once a program has been terminated , however [7] . Evidence already exists of the reduced efficacy of drugs used to combat lymphatic filariasis [9] and schistosomiasis [10] , with frequent treatment involving anthelmintic drugs appearing to hasten the development of drug resistance in some animals [11] . While chemotherapy has reduced levels of infection in the short-term , ensuring that positive health benefits extend beyond the cessation of chemotherapy programs has been challenging without improvements in the other factors that predispose populations to helminthiases [12 , 13] . Helminth infections , and indeed many infectious diseases , are strongly influenced by environmental and socio-economic conditions , and by human behavior and the effectiveness of health service provision [13] , or what Lambin et al [14] term the pathogenic landscape for disease . A major increase in schistosomiasis following the construction of dams and irrigation infrastructure has been well-documented [15] , as has the eradication of schistosomiasis in Japan through modernization of agricultural practices [16] and reduced hookworm infections as a result of improvements in sanitation and housing [17] . A One Health approach permits consideration of vulnerabilities at the environment-animals-humans interface [18 , 19] , accounting for the complex and highly dynamic process of infection , where a change in one underlying factor can drastically alter the situation for the other conditions , leading to the uneven distribution of diseases even in places with seemingly similar conditions . Such unevenness is observed in opisthorchiasis , an infection caused by the foodborne trematode Opisthorchis viverrini , where large variations in the disease burden may be observed in a relatively small geographic area [20 , 21] . Despite the close association of helminth parasite life cycle and life strategies with the physical environment and animal hosts , the One Health approach has rarely been applied to study of helminthiasis [22] . Yet , understanding such factors that underpin infections with a high focality can provide important contributions to the framework of One Health approach to a broader range of diseases , enabling intervention efforts that are tailored to local pathogenic landscapes , and in particular finely resolved vulnerabilities to the disease , to better accommodate future variations [23] . Moreover , the influence of , for example , environmental conditions can be easily masked by other factors that contribute to the extent and severity of a disease outbreak , such as health intervention efforts [24] . For example , intensive chemotherapy efforts have mitigated schistosomiasis burdens associated with recent hydro-infrastructure developments [25 , 26] , but such effects are palliative and temporary if the underlying factors causing infections , including infection of animal hosts and environmental conditions that promote and maintain pathogenesis , remain [27] . Opisthorchiasis is a major NTD in southeast Asia , and in the Mekong River basin in particular . The parasite involved , O . viverrini , is one of only three metazoan pathogens classified as a group 1 carcinogen , with sufficient evidence to establish a link between O . viverrini and cancer in humans [28] . Carcinogenicity of opisthorchiasis stems not only from prolonged infection and re-infection but also from the repeated treatment involving praziquantel anthelminthic , which can induce DNA damage leading to the development of hepatobiliary abnormalities , including cholangiocarcinoma ( CCA ) [29 , 30] . CCA is among the leading causes of cancer-associated mortality in the Mekong River basin [31] . O . viverrini is closely associated with wetland ( rice ) -based agriculture where drainage canals can facilitate infection of fish hosts by snail-shed cercariae [32] . The trematode has a three-host life cycle with freshwater Bithynia spp . snails and cyprinid fish as , respectively , the first and second intermediate hosts , and humans as the definitive host [33] . Human infection occurs through the consumption of raw or undercooked cyprinid fish , which is a common practice in the Mekong River basin . Small-scale freshwater fishing activities provide a major source of protein and additional income for local communities [21] , while raw fish consumption has led to the persistence of opisthorchiasis in many parts of the region despite decades of control efforts [34 , 35] . The control efforts have , to date , largely been restricted to chemotherapy and education campaigns , where the measure of success of control programs is limited to prevalence reduction instead of reinfection rate and long-term sustainability [36] . Despite the close relationship of opisthorchiasis with the physical and social environment , research on the range of factors that underpin the cycle of infection and reinfection has largely been neglected [18 , 37] . Particularly , there is little information on the association between human infection intensity and fish infection variation in different waterbody types . In fact , infection intensity is much less frequently reported than infection prevalence in O . viverrini studies [eg . 38–40] . The same is the case for other helminthiases , including soil-transmitted helminthiases [41] . This is problematic because infection intensity enables a very different understanding of the disease transmission and life strategies as compared with infection prevalence , in addition to being a factor in the most severe forms of infectious disease , including the risk of developing CCA in the case of O . viverrini infections . The focus of this study is the pathogenic landscape for opisthorchiasis , in particular the epidemiological role of dam construction and subsequent reservoir creation , socio-economic conditions , demographic factors and behavior , and variations in the efficacy of the provision of health services . This study illustrates the causes of an uneven distribution of disease burden , identifying contributing factors of infection while controlling for existing chemotherapy control efforts . Infection intensity is determined in addition to infection prevalence , and the variations in factors shaping intensity and prevalence examined . This study has the potential to facilitate improved health intervention efforts that take into account the high focality of opisthorchiasis . The approach and results have wider applicability , to the study of other NTDs , especially those with complex , environmentally sensitive life cycles .
Ethical approval for this study was obtained from the institutional review board of National University of Singapore , Singapore ( Reference code: A-14-122 , approved on 20 August 2014 ) and Khon Kaen University , Thailand ( Reference code: HE571229 , approved on 22 July 2014 ) . Permission for fieldwork was obtained from the subdistrict health centers . Meetings were held with heads of the health centers and health center workers to explain the purpose , procedures , risks , and benefits of the study . Health center workers were briefed , using Thai language , on the participant information sheet and the need to obtain written consent from the participants , and on how to administer the questionnaire , and to obtain fecal samples . All adult subjects were informed about the study design and objectives , and all study subjects gave written consent . No children were involved in this study . Identifiable information collected including names were anonymized using code numbers . After fecal examination , for participants tested positive with parasitic infection , personal information and corresponding infection results were made available only to the health center in the village so that treatment could be administered . Deworming medication was provided to the health centers for treatment of participants who were tested positive with infection . Those infected with O . viverrini were treated with praziquantel at an oral dose of 40 mg/kg . All medications were administered by certified nurses from the health centers . After the survey , only code numbers were retained by the principle investigator with the infection results and survey responses . No identifiable information was kept nor published . This study was conducted in four villages in the catchment for the Ubolratana reservoir ( 16°43’40°N , 102°34’45°E ) , northeast Thailand ( Fig 1 ) . Two of the villages , Sai Mun and Huay Bong , are located in the province of Nong Bua Lamphu , to the north of the reservoir . The other two , Fa Luem and Pho Tak , are in the province of Khon Kaen , to the south of the reservoir ( Fig 1 ) . According to Ong et al [42] , levels of O . viverrini infection of intermediate fish hosts were greater in fish caught in the main body of the reservoir when compared with those captured in rivers draining into the reservoir . In order to examine the influence of the physical environment on human O . viverrini infection , villages of varied levels of exposure to fish infected with O . viverrini were sampled . Two of the villages sampled , one in the north and one in the south of the reservoir , are located along the river inlets in the study area , and two , one in the north and one in the south of the reservoir , are located along the shore of the main body of the reservoir . Hereinafter , the villages are referred to as north ( N ) -river , N-reservoir and south ( S ) -river , and S-reservoir . O . viverrini infection prevalence and intensity were compared between villages located along the river inlets and reservoir to highlight and examine possible environmental influences . Samples in the north and south of the reservoir were compared to determine the association of infection with inter-provincial health jurisdiction . For reference , infection prevalence and intensity for each village were also presented , but no analyses were performed on them . A cross-sectional study was conducted between August and December , 2014 . Fecal samples and questionnaire-based surveys on socio-economic , demographic , and behavioral factors of participants ( S1 File and S2 File ) were collected from August 2014 , and any infected individuals identified from the results were treated during the months of November and December 2014 . Participating households were selected from information provided by the local health center using a random number generator . All members from the selected households who were 21 years or older at the time of the survey were invited to participate . Using StatCalc in Epi Info 7 . 1 . 5 software at confidence interval level of 95% and margin of error at 5% , a sample size of 125 was needed for each village . A total of 756 participants from the four study villages were eventually invited . Current infection status involving O . viverrini , other foodborne parasites , and soil-transmitted helminths were determined from the analysis of fecal samples . A single fecal sample was provided by each participant . The samples were returned to the health center on the same day and kept on ice . Samples were transported to the laboratory the following morning where they were stored at -20°C until analyzed for their parasite content . To increase the number of fecal samples returned , each village was visited on two consecutive mornings for the transportation of samples . Samples were processed using the formalin-ether concentration technique [43] and examined under the microscope by experienced laboratory technologists . The formalin-ether concentration technique is the current gold standard diagnostic for O . viverrini infection [44] , although immunological and molecular techniques to increase the sensitivity and specificity of diagnoses are being developed [34 , 44] . O . viverrini eggs were counted and recorded , and evidence of other intestinal parasites noted . Infection prevalence was tabulated by dividing the number of infected people with the total number of people sampled , while infection intensity was determined as the number of O . viverrini eggs per gram ( epg ) of fecal sample . Infection statuses of participants were provided to the head of the health centers along with medications for the treatment of O . viverrini and other intestinal parasites . Information on past treatment of O . viverrini was obtained from both health center records and completed questionnaires ( the latter were used to identify participants who received O . viverrini treatment from institutions other than health centers , including hospitals ) . A questionnaire-based survey was conducted to determine the association of socio-economic , demographic , behavioral factors with O . viverrini infection prevalence and intensity . Variables used in this study were selected based upon existing studies on O . viverrini risk factors [45 , 46] , while the set of possible responses in the multiple-choice questionnaire were formulated based on preliminary semi-structured interviews conducted with 251 respondents in the catchment of the Ubolratana reservoir . Demographic information , such as age and gender , of participants were provided by the health centers . Age was tabulated based on the year of birth of the participant and was expressed as a continuous variable . Other data , including level of education and occupation , were obtained through the questionnaires . As each participant may have more than one occupation , the various occupation types were each presented as an explanatory variable . Per capita income was calculated by dividing household income by the number of household members . Participants were considered as living “Below poverty line” or “Above poverty line” by comparing their household’s per capita income to average 2014 poverty line values from the National Economic and Social Development Board of Thailand for the provinces of Nong Bua Lamphu ( 2357 baht ) and Khon Kaen ( 2514 baht ) [47] . Participants were given the option of whether they wished to disclose information on their income . Levels of awareness of the hazard of O . viverrini infection and patterns of consumption of the raw fish dishes Koi pla ( freshly prepared raw fish salad ) , Mum pla , and Pla som ( both of which are lightly fermented raw fish dishes ) , which are commonly eaten in the study area , were determined through the questionnaire survey . Participants were also asked for the reasons behind their consumption/non-consumption of raw fish . Variables examined in this study were summarized in S1 Table . The Isarn Agenda , a program aimed at CCA prevention and control in northeast Thailand , was introduced in 2012 . The program involves fecal examination , ultra-sound scan for CCA above 40 years of age , exhibiting risky behavior , notably the consumption of raw fish . People found with opisthorchiasis are treated . Education programs are also created for primary school children . The Isarn agenda is not equally applied throughout northeast Thailand , however , as each province has the autonomy to decide on health priorities locally . In Khon Kaen province , in the southern part of the study area , only two districts , which are not included in this study , adopted the Isarn agenda , while other districts opted to focus on non-communicable diseases , such as cardiovascular diseases and diabetes . All districts in Nong Bua Lamphu , in the northern part of the study area , adopted the Isarn agenda . Thus , of the four villages examined in this study , the N-river and N-reservoir villages adopted the Isarn agenda , and the S-river and S-reservoir village did not . The prevalence and intensity of infections in the four study villages were compared with past O . viverrini infection diagnostic tests conducted by the health centers . Unlike mass drug administration efforts for other helminth parasites , praziquantel anthelmintic were given as a treatment for O . viverrini only for patients who were tested positive for infection by the parasite . As such , past attempts in O . viverrini diagnostic tests can also be used to determine past chemotherapy efforts by the health centers . Furthermore , information on local health priorities and perceptions of opisthorchiasis was also obtained from the heads of the health centers . Criteria for inclusion in the analyses included providing consent , not having withdrawn from the study , provision of suitable stool sample , and having a completed questionnaire . The prevalence and intensity of O . viverrini infections , and the reasons for/for not consuming raw fish , were analyzed for their association with environmental factors . The most notable environmental factor included in analysis was the type of waterbody from which the fish used in raw fish dishes originated from ( river or reservoir ) . Examined social factors included O . viverrini awareness , age , gender , and occupation . The possible influence of interventions by health centers , and variations in their level of implementation , was also investigated ( S1 Table ) . Bivariate analyses were first performed on each explanatory variable; variables with p-values below 0 . 2 were next entered into multivariate models . To examine prevalence , data from all participants were used in analyses . To examine intensity , only participants who tested positive for infection were included . To examine reasons for consumption , only participants who consumed raw fish were used for the analyses; conversely , in examining reasons for non-consumption , only participants who do not consume raw fish were used . Logistic regression was employed for analyzing infection prevalence , reasons for consumption , and reasons for non-consumption . The models were simplified with backward elimination and variable deletion determined using a chi-squared test for non-significant difference in deviance . Quasi-Poisson regression was used for analyzing infection intensity in the case of overdispersion . The model was simplified with backward elimination and variable deletion determined using F-test for non-significant difference in deviance . In addition , chi-squared test was used to test for variation in proportion of raw fish consumption by location , gender , and O . viverrini awareness . Variations in level of infections of fish by waterbody type ( reservoir or river inlet ) were also analyzed , using data from Ong et al [42] . A t-test with unequal variances was used on log ( x+1 ) transformed data , as data were not normally distributed . Differences in levels of infections in fish according to waterbody type [42] are compared with results from this study . Results from these analyses were used as a basis for examining the role of factors that have contributed to the pathogenic landscape for opisthorchiasis in the study area .
Of the 756 participants invited , 632 suitable samples were obtained ( 83 . 60% ) . The mean age of participants is 52 . 6 years . Among the participants , 54 . 2% were females and 45 . 8% were males . Comparison of the O . viverrini infection prevalence and intensity of the four villages showed that the S-river village had the highest prevalence at 40 . 21% , while the N-reservoir village had the highest infection intensity at 99 . 41 epg ( Fig 2A ) . When the villages were grouped according to their provincial health jurisdiction , infection prevalence in the north villages in Nong Bua Lamphu province ( 5 . 45% ) was statistically significantly lower than that of the south villages in Khon Kaen province ( 26 . 42% ) ( Fig 2B ) . When villages were grouped according to their proximity to waterbody types , infection prevalence did not vary much between villages located close to the reservoir and to the river inlets , but infection intensity was significantly higher for the reservoir villages at 93 . 72 epg than for the river villages at 38 . 54 epg . Bivariate analyses of the infection status of other foodborne parasites and soil-transmitted helminths revealed that O . viverrini infection prevalence was higher in participants who were also infected with other parasites , particularly those infected with soil-transmitted helminths ( Fig 3A ) . Participants who had raw fish consumption behavior were found to be significantly associated with higher O . viverrini prevalence ( Fig 3A ) , while the intensity of infection was significantly higher for participants who had not been dewormed ( 121 . 19 epg ) than for those who had been ( 56 . 7 epg ) ( Fig 3B ) . However , such associations were not observed in the multivariate models . Past O . viverrini deworming history did not greatly influence O . viverrini infection prevalence ( Fig 3A ) , and no statistically significant association was observed between O . viverrini awareness and both the O . viverrini infection prevalence and intensity . The mean age of participants found infected with O . viverrini was 56 . 1 years while the mean age that of the uninfected was 52 . 0 . Age was positively associated with infection prevalence in the multivariate model . Among social factors ( Fig 4 ) , both the bivariate and multivariate analyses suggested that gender was significantly associated with infection prevalence , while farming as an occupation and poverty line were significantly associated with infection intensity . Bivariate analyses indicated that the explanatory variables of the presence of soil-transmitted helminths , location , age , gender , education , farming as an occupation , above or below the poverty line , and raw fish consumption were significantly associated with O . viverrini infection prevalence ( p < 0 . 05 ) . Other foodborne parasitic infection was associated with O . viverrini infection prevalence at p < 0 . 20 . These variables were hence entered into a multivariate regression model . Results of the multivariate logistic regression model showed that the likelihood of infection was higher among villagers in the south , increased with age , and was the greatest in males and in those who consumed raw fish ( Table 1 ) . However , the consumption of raw fish is higher ( 63 . 6% ) in the two villages located in the province of Nong Bua Lamphu ( N-river and N-reservoir ) when compared with the two villages studied in the province of Khon Kaen ( S-river and S-reservoir ) ( 52 . 2% ) ( χ2 = 7 . 31 , df = 1 , p-value = 0 . 01 ) . Males were also more likely to consume raw fish ( 63 . 5% ) than females ( 54 . 2% ) ( χ2 = 4 . 83 , df = 1 , p-value = 0 . 03 ) . O . viverrini awareness was found to be negatively associated with raw fish consumption , with 28 . 6% of participants who were unaware of O . viverrini reported not eating raw fish as compared to 48 . 7% of participants who were aware of O . viverrini ( χ2 = 7 . 86 , df = 1 , p-value = 0 . 01 ) . The variables of past O . viverrini deworming , waterbody type , farming as an occupation , and income relative to the poverty line were significantly associated with infection intensity ( p < 0 . 05 ) . Other foodborne parasitic infection and soil-transmitted helminth infection were associated with O . viverrini infection intensity at p < 0 . 20 . Consequently , these variables were examined together in a multivariate Quasi-Poisson regression model . Results of the multivariate Quasi-Poisson regression model revealed that increased infection intensity was found in participants from villages located closest to the reservoir , participants who were not farmers , and participants who chose not to disclose their income information ( Table 2 ) . Comparison of spatial variation in human and fish infection indicated that , similar to human infection intensity , fish infection density was significantly higher in the reservoir waterbody type than the river waterbody type ( t = 2 . 66 , df = 10 . 22 , p-value = 0 . 023 ) , but not significantly different between the north and south ( t = -0 . 04 , df = 9 . 39 , p-value = 0 . 97 ) . Participants who were unaware of O . viverrini were more than twice as likely to state that they ate raw fish because it tasted delicious , while those living in the S-river and S-reservoir villages were more likely to state that they ate it out of habit . The odds of eating raw fish because of friends decreased with every one-year increase in age . Males were also more than twice as likely to eat raw fish because of friends as females ( Table 3 ) . The likelihood of not eating raw fish in order to avoid being infected by O . viverrini was at least eight times higher among participants who were aware of the risks of infection . When asked about the reason for selecting the option of avoiding O . viverrini despite having responded “No” in the question regarding O . viverrini awareness , some of the participants explained that they had been encouraged by health volunteer workers or nurses to avoid eating raw fish because of the parasite , even though they were unsure about what the parasite was . Participants who knew about O . viverrini were also about seven times more likely to avoid eating raw fish due to other health reasons . Participants who did not know about O . viverrini were more likely to avoid eating raw fish because they dislike it . In addition , participants who said that they disliked raw fish were more likely not to have received treatment in the past , are younger , or live in either N-river or N-reservoir village ( Table 4 ) . In the S-river village , there have been no attempts to determine O . viverrini infections , including fecal examination , for at least 10 years . The priorities of the health center of the S-river village focused on the health effects of pesticide use and respiratory tract infections ( Table 5 ) . By comparison , in the S-reservoir village , fecal tests of O . viverrini infection were performed in 2007 and 2008 , with infection prevalence estimated at 0% and 2% , respectively ( Table 5 ) . Because of funding constraints , the direct smear technique was employed and only relatively few people were tested in the village . A recent fecal examination done in 2014 , in villages belonging to the same health jurisdiction as the S-reservoir village , yielded infection prevalence similar to that of the S-reservoir village in 2007 and 2008 . Similar to the S-river village , the S-reservoir health center staff did not view opisthorchiasis as a top priority; instead , diabetes and hypertension were the main concerns . In the N-river village , fecal examination was performed in 2011 and 2012 with infection prevalence estimated at 5 . 25% and 2 . 26% , respectively . The local health center prioritized teenage pregnancy and parasitic diseases as the top health concerns , with campaigns aimed at reducing rates of teenage pregnancy organized by health center staff . In the N-reservoir village , fecal examination was performed in 2012 and 2013 , and O . viverrini prevalence was estimated at 8 . 24% and 0 . 27% , respectively . Different from the south villages , the local health centers of both N-reservoir and N-river villages used the single Kato-Katz thick smear technique for O . viverrini infection test . Hypertension and diabetes , the top health concerns of the S-reservoir village , were also prioritized by the local health center staff of the N-reservoir village as the major health concerns , among work related injuries and gastrointestinal diseases ( Table 5 ) .
The results show that examining prevalence alone risks ignoring important parasitic infection trends . Although there was not a significant difference in O . viverrini infection prevalence between villages located near river inlets as compared with villages near the Ubolratana reservoir , infected villagers from near the reservoir had more than double the parasite intensity as compared with villagers from near the river . The pattern of infection intensities among humans thus matched the infection density of fish collected from these locations , with higher overall fish infection associated with the reservoir when compared with river inlets [42] , while there was no difference in infection density in the fish from the south or north reservoir . Distance to waterbody had an impact on where villagers tended to source the fish used in raw fish dishes; villagers who lived close to the river tended to procure fish from the river , while those living close to the reservoir tended to procure fish from the reservoir . As fish in the reservoir is more plentiful , people who lived farther away from both river inlets and reservoir also tended to rely upon fish caught from the reservoir [42] . Differences in fish infection levels depending on waterbody can affect the level of exposure of humans to the risk of infection , as is evident in the results; the average intensity of infection in the S-river village was low despite the lack of chemotherapy effort , and lower than both N-reservoir and S-reservoir villages , despite the recent chemotherapy treatment efforts in the N-reservoir village in particular . Using only infection prevalence as the measure of success for intervention effort can problematically lead to individuals with high infection intensities in low prevalence areas being overlooked . Even in individuals with low infection intensity , it is possible to develop CCA , as observed in this study and other biomedical studies . During the course of the survey in this study , a participant who was tested negative for infection was diagnosed with CCA and passed away shortly after diagnosis . The participant had a history of raw fish consumption and no record of past O . viverrini treatment . The apparent absence of O . viverrini eggs in the fecal sample could have been due to a low intensity of infection or bile duct obstruction [48] . Biomedical studies show that opisthorchiasis-induced inflammation can lead to the development of O . viverrini-induced advanced periductal fibrosis ( APF ) and CCA , which are driven by common cellular mechanisms , marked by elevated level of plasma interleukin-6 [49] . Participants with the most elevated level of plasma interleukin-6 were found to have an increased risk of 19 and 150 times of developing APF and CCA , respectively , as compared with other O . viverrini infected individuals with no detectable plasma interleukin-6 [49] . The risk of developing APF was found to increase with increased infection intensity [50 , 51] and duration of infection [50] . The findings in this study are of relevance to the concept of One Health , as they highlight the close relationship between the health of humans and that of the health/infection status of the animal hosts and physical environment . The findings identify the reservoir as an important target for opisthorchiasis intervention efforts and also underscore the importance of considering infection intensity in the understanding of the pathways through which the parasite is transmitted . Comparative multilocality studies are necessary to gain useful insights into the similarity or difference in relationships between opisthorchiasis and the environment in such reservoir systems . Far higher infection prevalence in males than in females accords with findings from some previous studies [52 , 53] . Little difference in prevalence between genders has also been reported [38 , 54] , although Hasewell-Elkin et al [54] notes that the frequency of high infection intensities may be higher among males . Males are also more likely to die from opisthorchiasis . As males are often the main income earners in families in Thailand , opisthorchiasis can exert a disproportionate economic toll on those affected [49] . One reason for a higher infection prevalence and intensity among males is likely to be their socializing behavior: raw fish dishes are often available for consumption at social gatherings of males . Infection prevalence also tended to increase with age from 21 years in this study . This finding is at odds with existing results , which indicate a plateauing of infection prevalence in the late teens followed by a decline in later life [34 , 35] . In some studies , fishermen and/or farmers were found to have higher infection prevalence [38 , 45] . This is because local fishermen often make a dish of Koi pla from their catch to celebrate that day’s fishing [23] . Farmers may also harvest fish from their rice paddies and prepare and consume the catch on the spot . Conversely , in this study , infection in fishermen and farmers was not significantly higher than for other occupations . Higher infection intensities were found only in participants who were not farmers . The participants who were not farmers have other occupations including contract worker , craftsman , fisherman , foodseller , office worker , stay at home , and others . There was however no significant difference in intensity among people who belonged to those occupation types and those who do not ( Fig 4 ) , suggesting that the observed higher infection intensity in people who are not farmers is not determined by a single occupation type . Higher infection intensity was also found only in participants who chose not to disclose their income information . No clear pattern was observed between occupation types and the disclosure of income ( S2 Table ) . Consequently , the socio-economic and demographic factors selected in this study could not identify the specific groups of people at risk of higher infection intensity . Recent chemotherapy efforts in three of the four villages may have weakened links with the range of factors that result in infections . While there was no significant difference in infection prevalence and intensity with O . viverrini awareness , O . viverrini awareness appeared to reduce the proportion of people who reported consuming raw fish . Participants who were aware of O . viverrini were also more likely to avoid raw fish consumption in order to avoid opisthorchiasis and other health issues , while participants who were unaware of O . viverrini were more likely to avoid consumption due to personal dietary preferences . Awareness campaigns may be able to affect personal health decisions to a certain extent , although more holistic effort is needed to tackle this long-standing issue . The pattern of villagers residing in the south of the reservoir being much more likely to be infected than villagers in the north may reflect inter-provincial differences in health priorities and treatment efforts . Use of praziquantel to treat infections can result in a sharp decline in prevalence [55] . For example , praziquantel administration brought about an immediate decline of O . viverrini prevalence from approximately 60% to 14% , while infections among the control , untreated group increased from 65% to 71% within the same time frame [56] . Likewise , a similar decline in prevalence ( 67% to 16% ) during three years of praziquantel administration is reported in Sripa et al [37] . Favorable results following chemotherapy-based treatment efforts do not necessarily imply long-term success of a campaign , however . Resurgence of infection has been observed soon after the cessation of a campaign [57] . This study revealed disparate healthcare concerns and opisthorchiasis control efforts . While the particular focus of the health center can be tailored to the needs of the villages within the sub-district [58] , funding allocation for healthcare is decided at provincial level . As the S-river and S-reservoir villages are part of districts in Khon Kaen province , where the Isarn Agenda was not implemented , limited funding was made available for opisthorchiasis control efforts . The lack of fecal examination for O . viverrini infection for the past decade may account for the high infection prevalence recorded in the S-river village . In the S-reservoir village , where a relatively limited treatment program was in place , infection prevalence was second only to the S-river village . Direct smear was used in both villages to test for infections as it is the most affordable , despite it being the least sensitive method [59] . The low sensitivity of the test may have led to erroneous results in the form of low prevalence data . Due to an apparent low prevalence of O . viverrini and increasing prevalence in chronic diseases , particularly diabetes and hypertension , it is not unexpected that the local health center staff increasingly prioritize such chronic diseases as their top health concerns . Coupled with the affordability and simplicity of testing for diabetes and hypertension , regular blood sugar tests and blood pressure tests are offered by the health centers , which may in turn shift the health focus of the villagers to such chronic diseases . Indeed , during the course of this study , the villagers and health center staff of the south villages have expressed that O . viverrini infection is not an issue of concern in the village . Coincidentally , fecal examinations were carried out by the health center staff in 2014 , the same year of this present study , to survey O . viverrini infections in villages within the sub-district of the S-reservoir village . As the health center knew about our intent of sampling in the S-reservoir village , the health center sampling was conducted in all villages of the sub-district except the S-reservoir village . Their survey reported an overall infection prevalence of 0 . 25% for those villages , which was close to the prevalence observed for the S-reservoir village in 2007 and 2008 ( Table 5 ) . Nevertheless , the prevalence was in stark contrast to the much higher levels obtained in this study ( i . e . , 18 . 45% for the S-reservoir and 40 . 21% for the S-river , Fig 3A ) , with the disparity likely due to the difference in sensitivity of fecal examination methods employed . Unfortunately , disparate healthcare focus , coupled with limited funding and a less sensitive opisthorchiasis screening method may have given villagers–and health center staff–a false impression of the importance of opisthorchiasis . Villagers who consume raw fish may be lulled into a false sense of security when any tests for O . viverrini infection generate negative results , despite the frequent consumption of raw fish , as mentioned by several villagers interviewed . Fecal examinations were carried out by the health centers concerned on a greater number of individuals in the N-river and N-reservoir villages . The diagnostic tests for O . viverrini infection also relied upon the more sensitive single Kato-Katz thick smear . Infection prevalence in the N-river village , at 5 . 3% and 2 . 3% in , respectively , 2011 and 2012 , was close to the 4 . 6% prevalence obtained in this study . Similarly , close results were found for the N-reservoir village ( 8 . 2% and 0 . 3% in , respectively , 2012 and 2013 , compared with 6 . 4% in this study ) . Despite the increased focus on opisthorchiasis and CCA after the implementation of Isarn Agenda , the lower infection prevalence , and no significant differences in awareness of the risks of O . viverrini infection , the proportion of participants who reported eating raw fish remained high in these two north villages . This study emphasizes the influence of health center focus on O . viverrini infection prevalence . While most of the prior work has emphasized on human behavior and social risk factors for helminth diseases including opisthorchiasis , healthcare focus and provision can greatly affect the risk of infection and the vulnerability of local populations [60] . Healthcare focus and provision can also sheds light on the varying stakeholders’ values determining the pathogenic landscape of diseases . Stakeholders’ values can influence the outcome and direction of healthcare provision as illustrated in the variation in provincial funding and health center focus in this study . In the cases of other disease intervention efforts that substantially rely on external donor funding , there can be potential conflicting interests between local population , funding donors , or even pharmaceutical companies [61] . The influences of healthcare focus and interests of other stakeholders thus need to be considered when deciphering the factors contributing to disease risks . The holistic approach in this study has identified important features of helminth parasitism , specifically , opisthorchiasis , which include the connectivity of animal hosts and humans facilitated by waterbodies and human behavior; human behavioral and physical environmental conditions that facilitated reinfection; and the influence of healthcare interventions on infection prevalence . Identification of such features of parasitism is an important contribution to the framework of One Health approach [23] , where consideration of helminth diseases has largely been overlooked [22] .
While the role of socio-economic , demographic , and behavioral risk factors on O . viverrini infection have been investigated in previous studies , this study identified other additional influential environmental and healthcare implementation risk factors in O . viverrini infection . Humans interact with the environment reciprocally , thereby influencing their risks of disease infection . Human modifications of the environment , particularly in the form of dam construction and reservoir creation , have changed the aquatic habitats for the O . viverrini intermediate fish hosts . As O . viverrini infection intensities in the fish vary across different waterbody types , humans affect their risks of consuming O . viverrini infected raw fish through fish procurement location preferences . In opisthorchiasis studies and that of other helminthiasis , infection intensity is still much less frequently reported . The importance of considering infection intensity in a cross-sectional infection study is exemplified in this study , owing to the critical role of intensity in the most serious forms of many infectious diseases , including opisthorchiasis , and in providing insights into parasite transmission risks . Healthcare focus can directly affect human infection prevalence through chemotherapy and indirectly guide villagers’ risk perceptions through the choices of health campaigns or monitoring programs . Chemotherapy in the case of helminthiases such as opisthorchiasis is only palliative , with re-infections quickly occurring if the underlying factors that expose humans to infection are not dealt with . There is thus a need for a holistic approach to integrate the factors accounting for the broader pathogenic landscape within which diseases such as opisthorchiasis persist .
|
Many of the large-scale helminth control programs around the world have primarily relied upon drug treatment . Reliance on drug treatment alone does not deal with the ultimate causes of infection , resulting in reduced infection levels only in the short-term . Re-emergence of infections and possibly even development of drug resistance in parasites are common once the programs have been terminated . There is thus a need for consideration of a broader context , including environmental influence and healthcare focus , within which infections thrive . This study examines the roles of a reservoir dam environment , inter-provincial healthcare focus variation , and socio-economic , demographic , and behavioral factors to highlight the varying roles of such factors contributing to this disease landscape . The findings underscore the importance of a holistic approach in infection studies in order to provide more sustainable disease treatment and elimination outcomes .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion",
"Conclusion"
] |
[
"medicine",
"and",
"health",
"sciences",
"ecology",
"and",
"environmental",
"sciences",
"rivers",
"demography",
"cancer",
"treatment",
"clinical",
"oncology",
"tropical",
"diseases",
"vertebrates",
"parasitic",
"diseases",
"animals",
"health",
"care",
"freshwater",
"fish",
"oncology",
"foodborne",
"trematodiases",
"clinical",
"medicine",
"pharmaceutics",
"aquatic",
"environments",
"bodies",
"of",
"water",
"neglected",
"tropical",
"diseases",
"fishes",
"marine",
"and",
"aquatic",
"sciences",
"people",
"and",
"places",
"helminth",
"infections",
"opisthorchiasis",
"chemotherapy",
"freshwater",
"environments",
"earth",
"sciences",
"biology",
"and",
"life",
"sciences",
"drug",
"therapy",
"health",
"education",
"and",
"awareness",
"organisms"
] |
2016
|
Uncovering the Pathogenic Landscape of Helminth (Opisthorchis viverrini) Infections: A Cross-Sectional Study on Contributions of Physical and Social Environment and Healthcare Interventions
|
Leptospirosis is the most common bacterial zoonoses and has been identified as an important emerging global public health problem in Southeast Asia . Rodents are important reservoirs for human leptospirosis , but epidemiological data is lacking . We sampled rodents living in different habitats from seven localities distributed across Southeast Asia ( Thailand , Lao PDR and Cambodia ) , between 2009 to 2010 . Human isolates were also obtained from localities close to where rodents were sampled . The prevalence of Leptospira infection was assessed by real-time PCR using DNA extracted from rodent kidneys , targeting the lipL32 gene . Sequencing rrs and secY genes , and Multi Locus Variable-number Tandem Repeat ( VNTR ) analyses were performed on DNA extracted from rat kidneys for Leptospira isolates molecular typing . Four species were detected in rodents , L . borgpetersenii ( 56% of positive samples ) , L . interrogans ( 36% ) , L . kirschneri ( 3% ) and L . weilli ( 2% ) , which were identical to human isolates . Mean prevalence in rodents was approximately 7% , and largely varied across localities and habitats , but not between rodent species . The two most abundant Leptospira species displayed different habitat requirements: L . interrogans was linked to humid habitats ( rice fields and forests ) while L . borgpetersenii was abundant in both humid and dry habitats ( non-floodable lands ) . L . interrogans and L . borgpetersenii species are widely distributed amongst rodent populations , and strain typing confirmed rodents as reservoirs for human leptospirosis . Differences in habitat requirements for L . interrogans and L . borgpetersenii supported differential transmission modes . In Southeast Asia , human infection risk is not only restricted to activities taking place in wetlands and rice fields as is commonly accepted , but should also include tasks such as forestry work , as well as the hunting and preparation of rodents for consumption , which deserve more attention in future epidemiological studies .
The World Health Organization ( WHO ) estimates the global burden of leptospirosis at over one million severe human cases per year , with a growing number of countries reporting leptospirosis outbreaks [1] , [2] . Leptospirosis can represent up to 20–40% of idiopathic febrile illness [3] , [4] . Symptoms vary widely and mimic those of other diseases , including malaria , viral hepatitis , yellow fever , dengue , bacterial and viral meningitis , as well as many others [5] , [6] , [7] . Thus , leptospirosis patients may be misdiagnosed with these regionally more common or well-known diseases . In addition , many cases occur in tropical locations without adequate health care , surveillance and reporting , these factors are therefore likely to influence an underestimation of case numbers . For example , Thailand , which has a relatively good health system , reports several thousand cases of leptospirosis each year , while its neighbors , Cambodia and Lao PDR , report very few . This discrepancy is almost certainly due to under-reporting [8] , [9] . An additional problem is the limited understanding surrounding basic aspects of the disease epidemiology . Leptospirosis is caused by infection with members of the genus Leptospira that includes nine pathogenic species and at least five intermediate species [4] . Most cases of human leptospirosis , however , are not identified at the species , serogroup , or serovar level , hindering environmental risk awareness . In addition to studies aimed at understanding leptospirosis from a genetic standpoint , there have been numerous attempts to understand its transmission . Exposure to virulent leptospires may be direct , via contact with urine or tissues from infected animals , or indirect , where freshwater or humid environments are contaminated with an infected animal's urine . Socio-economic variables and occupations such as mining , cleaning sewers , working in a slaughterhouse , farming , and cattle breeding are known to increase the risk of contracting leptospirosis in Southeast Asia [9] , [10] , [11] . Limited research has been conducted on the distribution of leptospires in both the environment and in reservoir species . Many species can act as reservoirs , but wild rodents are usually considered to be the main reservoirs for human leptospirosis . Rodents generally acquire leptospirosis as pups , and maintain it as a chronic infection in the renal tubules , excreting bacteria in their urine throughout their life span , often in increasing amounts [6] . Once leptospires are shed into the environment , they can survive in water or soil , depending on physiochemical conditions [10] , [12] . In a previous study , Ganoza et al . [13] showed that the concentration and species of leptospires found in environmental surface water correlated with the risk of severe leptospirosis in humans . However , whether leptospiral species have different natural habitat and landscape distribution requirements remains largely unexplored . In the present study , we aimed to ( 1 ) describe Leptospira prevalence , species and strains in rodents from seven localities in Southeast Asia ( Thailand , Laos PDR and Cambodia ) ; ( 2 ) compare isolates from humans living in regions where rodents were sampled; ( 3 ) determine whether certain habitats or rodent species increase the prevalence of infection with specific Leptospira species . Finally , we discuss the outcomes from this combination of approaches , and their implication for infection routes and environmental risks for humans .
Leptospira cultures from human patients analyzed in this study were previously isolated by the Mahidol University in Thailand as part of the national surveillance for leptospirosis . The strains and DNA samples derived from these cultures were analyzed anonymously for this research study . Systematic field sampling was carried out by joint Asian and French research institute teams . Traps were set within houses with the approval of the owner or tenant . Outdoors , traps were set with the agreement of the village chief . None of the rodent species investigated are on the CITES list , nor the Red List ( IUCN ) . Animals were treated in accordance with the guidelines of the American Society of Mammalogists , and with the European Union legislation ( Directive 86/609/EEC ) . Each trapping campaign was validated by the national , regional and local health authorities . Approval notices for trapping and investigation of rodents were given by the Ministry of Health Council of Medical Sciences , National Ethics Committee for Health Research ( NHCHR ) Lao PDR , number 51/NECHR , and by the Ethical Committee of Mahidol University , Bangkok , Thailand , number 0517 . 1116/661 . Cambodia has no ethics committee overseeing animal experimentation . The ANR-SEST ( Agence Nationale pour la Recherche , Santé-Environnement et Santé-Travail ) program on rodent-born diseases in Southeast Asia , which provided part of the funding for this project , has been approved by the Managing Directors from both the Asian and French research institutes . In addition , regional approval was obtained from the regional Head of Veterinary Service ( Hérault , France ) , for the sampling and killing of rodents and the harvesting of their tissues ( approval no . B 34-169-1 ) carried out during this study . Seven localities were sampled for rodents during 2009 and 2010: Nan ( 19 . 15 N; 100 . 83 E ) , Loei ( 17 . 39 N; 101 . 77 E ) and Buriram ( 14 . 89 N; 103 . 01 E ) in Thailand , Luang Prabang ( 19 . 62 N; 102 . 05 E ) and Champasak ( 15 . 12 N; 105 . 80 E ) in Laos PDR , and Preah Sihanouk ( 10 . 71 N; 103 . 86 E ) and Mondolkiri ( 12 . 04 N; 106 . 68 E ) in Cambodia . Within localities , samplings were conducted over an area of about 10 kilometers squared . Four main habitats were distinguished , namely 1 ) forested ( rubber and teak plantations , secondary and primary forest ) ; 2 ) non-floodable lands ( shrubby wasteland , young plantations , orchards ) , ( 3 ) floodable lands ( cultivated floodplains , rice fields ) , and ( 4 ) human dwellings ( in villages or cities ) . For each habitat , 10 trapping lines , which each consisted of 10 wire live-traps ( hand-made locally , about 40×12×12 cm ) every five meters , were installed over a period of four days . Additional captures were also conducted by locals . Captured rodents were collected each day and taken back to the laboratory for dissection according to the protocol of Herbreteau et al . [14] . Where possible , rodent species were determined in the field using morphological criteria from Pages et al . [15] , but as morphological criteria were not fully discriminant between some genera , molecular approaches were also carried out . The mt gene was used for barcoding Mus species and some Rattus species ( R . tanezumi , R argentiventer , R . sakeratensis , R . adamanensis ) [16] . In accordance with Pages et al [17] , the mt lineages “Rattus lineage II” and “Rattus lineage IV” of Aplin et al [18] , were considered as conspecifics and named R . tanezumi . Barcoded samples were identified using the webservice RodentSEA [19] .
We detected 64 Leptospira-positive rodents from the 901 tested , giving a mean prevalence of 7 . 1% . Nineteen shrews ( Suncus murinus ) were also tested and all were found negative . Leptospires were detected in six localities ( Figure 1 ) with highly variable prevalence across localities , from 0% to 18% . Twelve rodent species ( over 18 tested ) were found positive and prevalence varied from 0 to 19% across species ( Table 1 ) . The rrs PCR assay was performed on the 64 samples which were positive for Leptospira . After 25 of these samples returned negative results following direct PCR , they were then analyzed by nested PCR . All PCR products were then sequenced and the Leptospira species were categorized based on phylogenetic analysis of the rrs fragment . Four species were determined: L . borgpetersenii ( n = 36 ) , L . interrogans ( n = 25 ) , L . kirschneri ( n = 2 ) and L . weilli ( n = 1 ) . The amplification of secY was successful in 31 of the 64 samples positive for Leptospira , including 20 L . borgpetersenii and 11 L . interrogans . No amplification could be detected for L . kirschneri and L . weilli . Lack of amplification was probably due to low levels of Leptospira DNA in the samples . The alignment of the 549-bp secY fragments distinguished a total of eight distinct alleles ( GenBank accession numbers: KF770694-KF770731 ) , including two alleles for L . borgpetersenii ( A and B ) and six for L . interrogans ( C to H ) . There was no clear association between secY-identified strains and either locality or rodent species ( Figure 2 and Text S1 ) . MLVA positively identified 15 of the 25 tested samples , which were positive for L . interrogans ( Text S1 ) . The MLVA patterns are in close agreement with the alleles determined by secY sequencing . Comparison with our reference strains indicates that our samples share an identical secY sequence and MLVA profile to strains of the Canicola , Pyrogenes and Autumnalis serogroups ( Text S1 ) . Two secY alleles from L . interrogans were recovered from both rodents and humans ( Figure 2 , Text S1 ) . The secY C allele was recovered from the wild mouse , M . cookie , from Loei , northern Thailand . This allele had previously been characterized from the ST34 clone , which corresponds to a Autumnalis serogroup human isolate associated with the northern Thailand outbreak between 1999 and 2003 ( Thaipadungpanit et al . 2007 ) . Secondly , the secY D allele ( MLVA pattern 640/750/650 ) of the Pyrogenes serogroup was found in both rodent and human samples from Loei , northern Thailand . Statistical analysis revealed that rodent locality , habitat and sex , significantly affected individual infection ( Table 2 ) . Rodents living in households showed significantly lower infection rates ( Figure 3 ) . Males were significantly more likely to be infected than females ( Figure 3 ) . By contrast , rodent species showed no correlation with infection . As there was potential non-independence between the distributions of rodent species among habitats; as three species are strictly restricted to households; we re-analyzed the data after removing all household rodent data . We found similar results ( data not shown ) , again with significant effects due to locality and sex , but not species , indicating that those species living outside human dwellings have an overall similar level of infection . L . borgpetersenii and L . interrogans infection were then investigated separately , which once again showed that locality , habitat and sex , but not rodent species , were the major determinants of infection ( Figure 3 and Text S1 ) . This last analysis suggested a difference in ecological niche for both Leptospira species . In particular , L . borgpetersenii was much more abundant in dry habitats ( non-floodable lands ) than L . interrogans .
Most research on the presence of leptospires in rodents has been conducted in urban areas , or rural areas in the vicinity of households [36] , [37] , [38] . In contrast , only a few studies have investigated the prevalence of Leptospira in rodents within their various habitats [39] . Our study suggested that the mean prevalence in rodents across localities was approximately 10% , when we excluded rodents trapped in human dwellings where prevalence was very low ( 2% ) . Leptospira prevalence was similar between floodable areas , forests and non-floodable agricultural fields . Our results then challenge the widely accepted belief that leptospires mainly circulate in wetlands . Two potential leptospirosis transmission routes are generally assumed; direct transmission between individuals , or via the external environment . The relative importance of these routes in rodents is unknown; however , our results suggest that direct transmission could explain the circulation of leptospires in dry habitats . Individual variation in susceptibility to infection is a common outcome of epidemiological studies . In the context of pathogenesis , infection may vary with numerous individual features such as sex , age , physiological condition , behavior and immunogenetics [40] , [41] . Recently , Perez et al showed that meteorological conditions might also influence Leptospira carriage in rodents , with hot and rainy seasons associated with both high abundance and increased prevalence in rodents [38] . Taking into account this inter-individual variability greatly enhances both our understanding of disease epidemiology , and our ability to predict the outcomes of epidemics by using adapted epidemiological models [42] . Statistical analyses of our dataset revealed that males were clearly more susceptible to Leptospira infections than females , consistent with many reports on vertebrate infections [43] . Differing infection rates observed due to sex might result from endocrine-immune interactions . Androgens have immunosuppressive effects , explaining the reduced efficiency of the male immune system and its association with higher infection rates . Moreover , steroid hormones alter rodent behavior which then influences susceptibility to infection . Males of most mammals are more aggressive than females , more likely to disperse , and have larger home ranges with more intense foraging activities; all these behaviors cause increased pathogen exposure . Additionally , rodent susceptibility to Leptospira infection did not significantly vary across rodent species . Most rodent species were found to be infected by Leptospira and our statistical modelling did not highlight “species” as a significant factor explaining Leptospira infection . The observed variation in prevalence across rodent species is most probably an indirect consequence of their specific habitat requirements . For instance the Pacific rat , Rattus exulans , was rarely found infected , but this probably results from its close association with human dwellings where Leptospira prevalence is consistently lower than in other habitats . Although rarely documented , the different rodent species investigated here display clear habitat preferences [39] , [44] . Some species are more abundant in rain-fed paddy fields ( Bandicota indica , R . argentiventer ) or forests ( Leopoldamys edwardsi , Maxomys surifer ) or non-flooded fields ( Mus cervicolor , Mus cooki ) . In line with other studies [39] , [45] , our results confirmed the importance of Bandicota and Rattus species as hosts for Leptospira strains of human health importance . However , high prevalence of pathogenic species and strains were also observed in rarely investigated rodents such as forest species ( Berylmys sp . , Maxomys sp . ) and wild mice ( Mus sp . ) . This observation suggests that rodent reservoirs for human leptospirosis are probably more diverse than previously thought ( see [35] ) . Finally we discuss any of the present study's limitations , which may mitigate some of the above interpretations . As reported in other agricultural systems ( see [46] , [47] for instance ) rodents may move among habitats , either as part of the dispersal process ( i . e . the movement of an organism from its birth place to its first breeding site , or from one breeding site to another ) , or in response to the seasonal variation in habitat quality ( i . e . amount of food , shelter availability , competition with other rodents , predation etc . ) . In the case of Southeast Asian rodents , one can imagine seasonal movements between flooded and non-flooded habitats , or between other habitats , but we lack data on these movements , which have not been the subject of publication to our knowledge . Because these movements may involve rodents infected with Leptospira , this process could have important consequences on Leptospira distributions within Southeast Asian agricultural landscapes ( see [48] for an example of the importance of rodent movements for the epidemiology of a rodent-borne hantavirus in Europe ) . As has already been pointed out by Singleton and collaborators [49] , more data is needed on the ecology of rodents in Southeast Asia , and such data would probably significantly increase our understanding of Leptospira epidemiology . Another insight of our study is that the two most abundant Leptospira species , L . interrogans and L . borgpetersenii , both of which are of great significance to human disease in Asia [35] , [50] , [51] , [52] , may have different epidemiological cycles . L . interrogans infection in rodents was restricted to humid habitats while L . borgpetersenii infection was equally frequent in both humid and dry habitats . This new ecological data on rodents is consistent with previous data gained from experimental and genomic studies . Experimental data suggest that survival in water is highly reduced for some strains of L . borgpetersenii when compared to L . interrogans . L . borgpetersenii serovar Hardjo lost >90% viability after 48 h in water , whereas L . interrogans retained 100% viability over the same period [53] . L . interrogans would thus be able to survive in such an environment , especially in surface water , allowing transmission from contaminated water . Whereas L . borgpetersenii would not survive outside its host , forcing direct host-to-host transmission . This difference in ecological niche is reflected in the genomic composition of the two species . L . borgpetersenii serovar Hardjo strains have a smaller genome than L . interrogans . Genome rearrangement in these strains of L . borgpetersenii mainly affect the ability to sense the external environment , which may indicate that these strains are in the process of becoming specialized for direct transmission . In contrast , L . interrogans has many environmental sensing genes and exhibits large shifts in protein expression when moved from a natural environment-like medium to a host-like medium [54] . While we cannot presume that the change reported by Bulach et al . [53] is representative of all L . borgpetersenii strains or is only restricted to certain strains , their study demonstrates that genome composition and habitat preference may largely differ across strains and species . These in vitro results are consistent with our ecological observations . Whether environmental conditions ( outside the host ) determine Leptospira species distribution in nature remains largely unexplored . Ganoza et al . [13] showed differential distribution of isolates in urban or rural water sources in Peru , reflecting rates found among human isolates from both urban and rural settings . Perez et al [38] demonstrated that seasonal variations influence Leptospira prevalence in rats and mice from New Caledonia . Very little data has been published concerning the epidemiology of both Leptospira in humans and wildlife , however some human epidemiology reports suggest that L . interrogans is commonly acquired from contaminated surface water , whereas a host-to-host transmission cycle is more likely to occur for L . borgpetersenii [53] . However our results suggest a lower transmission risk from rodents to humans for local L . borgpetersenii strains , in comparison with L . interrogans strains . Most studies in Southeast Asia currently focus on human infection linked to humid habitats and rice cultivation [9] , [49] . Without calling into question the importance of this route of transmission , our results suggest alternate routes of infection , which deserves further study . Human infection could also occur in other humid habitats , such as standing water and forest streams . Moreover , rodents are the subject of traditional hunting and trade in many parts of Southeast Asia . Close contact between rodents and humans during these activities , as well as rodent preparation before consumption , could present a significant route of infection , which should be evaluated . In line with our results , frequent human activity in forests was identified as a significant risk factor in Laos [9] . On the other hand , and fortunately , human contamination by commensal small mammals is probably low in Southeast Asia , despite the abundance of rodents and shrews in human dwellings . To our knowledge , our work is the first ecological evidence supporting different transmission routes for L . interrogans and L . borgpetersenii species in nature . Clearly this last point deserves more study , notably in order to strictly demonstrate the predominance of L . borgpetersenii direct transmission in ecological systems , as well as to determine if this transmission mode holds true for all borgpetersenii serovars , or for only some specific serovars . Together , this work brings to light novel perspectives on leptospiral epidemiology , reinforces the existence of species-specific transmission routes in nature , and stresses the need for the precise diagnosis of Leptospira involved in human and animal infections in order to better understand and foresee epidemics .
|
Leptospirosis is the most prevalent bacterial zoonosis worldwide . Rodents are believed to be the main reservoirs of Leptospira , yet little epidemiological research has been conducted on rodents from Southeast Asia . Previous studies suggest that activities which place humans in microenvironments shared by rodents increase the probability of contracting leptospirosis . We therefore investigated the circulation of leptospiral species and strains in rodent communities and human populations in seven localities scattered throughout Southeast Asia; in Thailand , Lao PDR and Cambodia . Molecular typing assays were used to characterize leptospiral species and strains in both rodents and humans , which demonstrated common strains between humans and rodents . Additionally , we observed that the two most abundant leptospiral species; L . borgpetersenii and L . interrogans , have different habitat requirements , which supposes different modes of transmission . Lastly , in Southeast Asia , the risk of leptospiral transmission to humans is not solely limited to wetlands and rice paddy fields , but is also linked to forested areas , and activities such as the hunting and/or preparation of rodents for consumption .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"leptospirosis",
"bacterial",
"diseases",
"infectious",
"diseases",
"veterinary",
"diseases",
"zoonoses",
"medicine",
"and",
"health",
"sciences",
"environmental",
"epidemiology",
"epidemiology",
"neglected",
"tropical",
"diseases",
"biology",
"and",
"life",
"sciences",
"population",
"biology",
"tropical",
"diseases",
"veterinary",
"science"
] |
2014
|
Epidemiology of Leptospira Transmitted by Rodents in Southeast Asia
|
The presence of amyloid deposits consisting primarily of Amyloid-β ( Aβ ) fibril in the brain is a hallmark of Alzheimer's disease ( AD ) . The morphologies of these fibrils are exquisitely sensitive to environmental conditions . Using molecular dynamics simulations combined with data from previously published solid-state NMR experiments , we propose the first atomically detailed structures of two asymmetric polymorphs of the Aβ9-40 peptide fibril . The first corresponds to synthetic fibrils grown under quiescent conditions and the second to fibrils derived from AD patients' brain-extracts . Our core structure in both fibril structures consists of a layered structure in which three cross-β subunits are arranged in six tightly stacked β-sheet layers with an antiparallel hydrophobic-hydrophobic and an antiparallel polar-polar interface . The synthetic and brain-derived structures differ primarily in the side-chain orientation of one β-strand . The presence of a large and continually exposed hydrophobic surface ( buried in the symmetric agitated Aβ fibrils ) may account for the higher toxicity of the asymmetric fibrils . Our model explains the effects of external perturbations on the fibril lateral architecture as well as the fibrillogenesis inhibiting action of amphiphilic molecules .
A number of human diseases known as amyloidoses [1] , [2] are associated with the presence of amyloid plaques in organs and tissues . The main constituents of these plaques are fibrillar aggregates arising from the pathological self-assembly of normally soluble proteins . The etiology of amyloidoses is poorly understood , and the causative agents in cellular toxicity have been associated with soluble oligomers [3]–[6] as small as dimers[6] , protofibrils [7]–[10] and mature fibrils[11] . The fibrillar products of aggregation ( these include protofibrils as well as mature fibrils ) share common structural features: they are enriched in β-sheet structure and possess a common cross-β sheet motif , in which the β-strands lay perpendicular to the main axis of the fibril [12]–[16] . In most cases , the atomic structure of the fibrils is not known , although recent computational and solid-state NMR studies have begun to provide detailed models of amyloid fibrils . [11] , [17]–[26] Perhaps the most clinically relevant amyloidosis is Alzheimer's disease ( AD ) , the leading cause of late-life dementia . The protein implicated in AD is the 40–42 amino-acid long amyloid-β ( Aβ ) peptide , derived from proteolytic cleavage of the transmembrane amyloid precursor protein . [27]–[29] Experimental studies have shown that the morphology of Aβ fibrils is exquisitely sensitive to environmental conditions . Gentle mechanic shaking [11] , small chemical modifications ( e . g the oxidation of Met 35/M35ox[19] ) or ligand binding ( e . g small peptidic [30] or non-peptidic inhibitors[31] ) can affect the interactions ( salt bridges , hydrophobic side-chain packing etc . ) between the cross-β subunits ( protofilaments ) constituting the fibril . This can lead to large scale changes in fibril morphology , and even to altered toxicity[11] . For instance , at pH 7 . 4 and 24°C , and under conditions of gentle mechanic sonication , Aβ40 peptides are seen to form amyloid fibrils ( “agitated fibrils” ) that predominantly contain 2 cross-β subunits with untwisted , “striated ribbon” morphologies . [32] Based on a combination of data from solid state NMR and scanning transmission electron microscopy ( STEM ) , Tycko and co-workers showed that the agitated amyloid fibrils are 2-fold-symmetric ( i . e have 2 equivalent cross-β-subunits ) . In sharp contrast , under the same solution conditions , but in the absence of sonication , the resulting “quiescently” grown Aβ40 fibrils predominantly contain 3 cross-β subunits with a “twisted pair” morphology . [33] , [34] These quiescent fibrils appear to be more toxic than the agitated fibrils , based on studies on rat embryonic hippocampal neurons . [11] Even more striking is the fact that a slight alteration in the quiescent growth conditions leads to a different symmetry for the fibril: in one case , the 3 cross-β subunits are arranged in an asymmetric manner ( 2 equivalent cross-β-subunits and one 1 nonequivalent cross-β-subunit ) [11] , and in the other , in a symmetric manner ( 3 equivalent cross-β-subunits ) . [18] Recently , Tycko and co-workers [35] have performed solid state NMR and mass-per-length ( MPL ) studies on fibrils obtained from AD patients' brain extracts . These brain-seeded fibrils , which presumably reflect the relevant fibrils structures found in diseased brains , show yet another morphology , albeit one bearing strong similarities to an asymmetric quiescently grown synthetic Aβ fibril . Both predominantly contain 3 cross-β subunits that show two sets of chemical shifts for many 13C-labeled sites , and the primary difference between the synthetic quiescent and brain-derived fibrils appears to lie in the orientation of the side chains in the C-terminal β-strand of the fibril . While experimental and computational models of the agitated fibril with 2-fold symmetry [17] , [20] , [22] , [32] and of the quiescent fibril with 3-fold symmetry [18] have been proposed based on experimental and computational studies , there is presently no atomically detailed model of the asymmetric quiescent synthetic fibril or of the brained-seeded fibril . Based on the structure of the solved fibrils of Aβ and analysis of the experimental data for the unsolved fibrils , it emerges that all Aβ fibrils ( agitated or quiescent ) studied by Tycko and co-workers share the same fundamental building block: a common cross-β subunit . This subunit ( shown in Fig . 1 A–B ) consists of stacked β-sheets formed from the parallel in-registry assembly of a U-shaped β-strand-loop-β-strand motif . In this cross-β subunit , the β-strands are oriented perpendicular to main chain hydrogen bonding direction , with the hydrogen bonding direction laying parallel to the fibril axis . Two such cross-β subunits stack laterally ( the normal direction to the β-sheet surface ) to form the 2-fold symmetric agitated fibril[32] , while three such units arrange in a triangle to form the 3-fold symmetric quiescent structure . [18] The atomic details of the cross-β subunits differ slightly in the agitated and quiescent models . In the quiescent asymmetric case , a slight conformational difference has been reported in the side-chains of the solvent exposed loop region ( residues 23–29 ) , but the β-sheet-to-β-sheet stacking that determines the overall morphology of the fibril is the same . In the case of the brain-derived subunits , the side-chain orientations of some of the residues are inverted with respect to those in the agitated subunit . In this study , we use the cross-β subunit of reference [32] ( the structure based on the most recent refinement work by Tycko and co-workers , and one that is consistent with the original predictions of Nussinov and coworkers , ref [36] ) as a starting point for our simulations . Using stacking simulations between cross-β subunits , we propose a structural model for the asymmetric quiescent Aβ fibrils and for the brain-seeded fibrils . In the case of the brain-seeded fibril , we introduce appropriate modifications ( as detailed in the methods section ) to capture the correct orientation of the side chains . Our simulations are akin to quiescent assembly conditions as we are not including the effects of mechanical agitation in our stacking simulations . We validate our resulting models using the experimental data provided in the work of Tycko and co-workers . [11] , [35] , [37] , [38] We also propose a unifying lateral stacking mechanism that explains the variations in fibril's lateral architecture and toxicity under different external perturbations ( mechanical shaking , M35 oxidation , and ligand binding ) .
A starting cross-β subunit is extracted from the 2-fold symmetric model of the agitated Aβ9-40 fibrils . This structure corresponds to the most recent refined structure obtained by Tycko and workers [17] , [32] It consists of two β-sheet layers , with each layer containing 6 Aβ9-40 peptides , in which each Aβ9-40 peptide ( Fig . 1A ) is arranged in a β-strand-loop-β-strand/U fold: a N-terminal β-strand ( residues 10–22 ) , a loop ( residues 23–29 ) and a C-terminal β-strand ( residues 30–40 ) . We use the following nomenclature: since the exposed side of the C-terminal β-strand contains only hydrophobic residues ( G29_I31_G33_M35_G37_V39 ) , we refer to it as the hydrophobic “H” β-strand . In contrast , since the exposed side of the N-terminal β-strand contains hydrophobic residues separated by charged or polar residues ( Y10_V12_H14_K16_V18_F20_E22_V24 ) , we refer to it as the polar “P” β-strand . This nomenclature enables us to distinguish the N- and C-terminal β-strand . The implication of a uniform hydrophobic ( H ) surface as opposed to one interdispersed with polar residues ( P ) will be discussed later in the text . We used the same cross-β subunit in modeling fibrils containing multiple cross-β subunits . It should be noted that other researchers have reported differences in the exact position of the residues in the β-sheets with the length of the β-strands sometimes changing [39] . We also note that we have treated residues 1–8 of the Aβ peptide in the fibril as disordered , based on experimental data from the studies of Tycko and coworkers . For this reason , we are only modeling residues 9–40 of Aβ in our stacking simulations . It is entirely possible that in some polymorphs these residues become structured . Throughout this paper , we denote the cross-β subunit ( Fig . 1B ) as HUP where U represents the parallel in-registry assembly of a β-strand-loop-β-strand motif , H the hydrophobic ( residues 29–39 ) and P the polar ( residues 10–22 ) β-sheet surfaces ( Depending on the arrangement of the cross-β-subunit as part of a larger assembly , the cross-β-subunit will appear as HUP , PUH , H∩P or P∩H ) . We first considered the stability of this cross-β-subunit via four 20 . 0 ns long simulations at 310K . The subunit was found to be stable , as judged from the small ( less than 2 Å ) root mean square deviation ( RMSD ) from the starting structure . The fact that this subunit is stable is consistent with recent mass-per-length ( MPL ) data from the Tycko group in which a peak ( ∼9 kD/nm ) corresponding to a single layer of Aβ1-40 is seen for the agitated and the symmetric quiescent fibril of Aβ[38] ( 1 subunit is ∼9 kD/nm , hence the number of subunit is equal to MPL/9 ) . Similar studies using the new apparatus reported in reference 35 have not yet been performed on the asymmetric Aβ fibrils . The core of the 6-member cross-β-subunit ( consisting of the 4 inner peptides ) was very stable , while the 2 outer peptides showed more fluctuations . This is to be expected as the outer peptides have only one neighboring peptide that can provide stabilizing interactions . Since the aim of this study is to investigate lateral assembly and not on β-sheet extension of the cross-β-subunit[40] , we only consider the outer peptides in the energetic , but not the structural analysis . Having established that the cross-β-subunit is a stable entity , we used it as a building block to construct a profibril containing two such cross-β-subunits . Several possible arrangements are possible , and we considered all 6 possibilities based on a combination of 3 interfaces and 2 orientations between two cross-β-subunits ( PUH and H∩P ) , as listed in Text S1 . The 3 possible interfaces are HH ( hydrophobic-hydrophobic ) , PP ( polar-polar ) and mixed PH ( hydrophobic-polar ) and the 2 possible stacking orientations are parallel ( p ) and antiparallel ( a ) . Rather than starting with a pre-assembled fibril and testing its stability [21]–[24] , we initiated our simulations with two separated cross-β-subunits and monitored their assembly ( e . g . Fig . 1B ) . This enables us to study both assembly and stability . The number of side-chain atom contacts and 6 additional structural order parameters were used to characterize the β-sheet-to-β-sheet stacking process ( see Text S1 ) . Of the 6 possible constructs , an ordered and stable fibril interface was observed only in constructs aPP , aHH , and pHH . Snapshots of the final structure from a representative trajectory for each of the 6 constructs studied shown in Figure 2 . We summarize the structural features of the three ordered interfaces below , with the other three disordered interfaces described in the Supplemental Material . For aPP ( Fig . 2-A1 ) , the interface is stabilized by two hydrophobic pairs ( F20-V18 and V18-F20 , viewed from left to right ) , and two salt bridges ( E22-K16 and K16-E22 , viewed from left to right ) in the cross section ( along the β-sheet stacking direction ) of the ordered four β-sheet layers . The side-chains at the sheet-to-sheet interface are packed head-to-head without interdigitating ( “zipping” ) leading to a large layer-to-layer distance ( 13 . 6±0 . 4 Å ) . In the case of aHH ( Fig . 2-C3 ) , a tight hydrophobic interface is formed by five hydrophobic pairs ( G29-V39 , I31-G37 , G33-M35 , M35-G32 and G37-I31 , viewed from left to right ) between two N-terminal β-strands ( i . e . G29_I31_G33_M35_G37_V39 ) . The lack of side-chains of the glycine residues provides a groove on one face into which the large hydrophobic side chains of the opposite cross-β unit can fit . As a result of the insertion of the large hydrophobic side chains ( V39 , M35 and I31 ) into the grooves formed by G29 and G33 on the opposite face , the resulting layer-to-layer distance ( 7 . 1±0 . 8 Å ) in aHH is shorter than seen in aPP ( 13 . 6±0 . 4 Å ) . ” In additions , the β-sheets at the interface of system aHH are slightly less twisted than those of system aPP ( twist angle of ∼3° in aHH versus aPP ∼6° for aPP ) . For pHH , a tight hydrophobic interface is formed by four hydrophobic pairs ( G29-I31 , I31-G33 , G33-M35 and M35-G37 , viewed from left to right ) as a result of a one-residue shift of the β-strand along the β-strand direction . In order to gain further insight into the relative stability of the fibrils with different interfaces , we calculated the binding energy between two cross-β subunits over time for each system using the MM-GBSA module in AMBER . The convergence was observed in the last 5 ns ( see aHH system as an example in Text S1 ) . The results over the last 5 ns are shown in Figure 3 . A clear relative trend emerges: the ordered complexes aPP and aHH have the lowest binding energies ( −159 . 9±7 . 4 and −156 . 2±9 . 5 kcal/mol respectively ) . The pHH construct has a less favorable binding energy ( −108 . 4±6 . 8 kcal/mol ) than aHH . The significant difference ( ∼48 kcal/mol or ∼4 kcal/mol per peptide ) in binding energies between aHH and pHH illustrates that stability is not only determined by the hydrophobicity of the interface alone; the interdigitation of the side-chains at the interface also plays a key role[41] . It is interesting to note that 2D 13C-13C NMR experiments [32] have identified the presence of contact pairs I31-G37 and M35-G33 in Aβ40 fibrils , which further support a construct with an aHH interface over a pHH one [21] , [32] in 2-cross-β-subunit fibrils . Indeed , these contact pairs are among the contact pairs ( G29-V39 , I31-G37 , G33-M35 , M35-G32 and G37-I31 ) present in our aHH model fibril , but not in the pHH fibril . From an energetic perspective , aPP and aHH are the most favorable interfaces ( indistinguishable within error from each other based on our binding energy calculations ) . However , from an entropic perspective , one could argue that the aHH interface might be slightly more favorable than the aPP interface ( larger ΔS ) . The energetic basin associated with hydrophobic interactions ( ie , the HH interface ) is much broader than the narrow basin associated with distance dependent electrostatic interactions ( the salt bridges at the PP interface ) . As a result , the HH interface can accommodate much more structural fluctuations and disorder than the PP interface . Fluctuations leading to a shifting of the two cross-β-subunits along the β-strand direction or disorder related to mis-registry can be tolerated at the HH interface , but not at the PP interface where such effects would lead to breaking of the salt-bridges and hence an overall destabilization of the fibril . As a result , the aHH interface would be the most favorable in terms of free energy , with aPP the close second . Our results are in a qualitative agreement with a recent stability studies [22]–[24] of pre-constructed 2-cross-β-subunit species of Aβ40 modeled by another popular CHARMM force field[42] . Having established that the 1-cross-β-subunit and the 2-cross-β-subunit constructs with aHH and aPP interfaces are stable , we now turn to the assembly of a larger profibril based on the 1- and 2- cross-β-subunits . In particular , we wish to construct a model for the asymmetric 3-subunit quiescent fibril seen in the experiments of Tycko and co-workers[11] . The asymmetry is suggested by the fact that two sets of chemical shifts were observed in experiment for several 13C-labeled sites , indicating that the sidechains of these residues are in different environments , In order to satisfy this asymmetry , two different types of interfaces between three cross-β-subunits are required . Based on our previous calculations , we expect one of the interfaces to be aHH ( the most stable interface ) , and the other one to be aPP ( the second most stable interface ) . The experiments of Tycko also indicate the presence of a smaller amount of a 4 cross-β-subunit fibril . Similarly to the 3 cross-β-subunit fibril , this structure will also involve the two types of interfaces . A 3-cross-β-subunit protofibril can arise either from a 3 body assembly ( 1+1+1 ) , or from a 2 body assembly ( 2+1 ) . Here we only model the 2+1 assembly pathway , as a two-body assembly is more probable than a three-body assembly for entropic reasons . The ( 2+1 ) stacking would involve in a first step the formation of a 2-subunit fibril ( PUHH∩P ) with an aHH interface ( such as the model proposed for the agitated 2-cross-β-subunit fibril ) . It would be followed by the lateral stacking of another cross-β-subunit such that the final fibril has 3 stacked cross-β-subunits ( PUHH∩PPUH ) with two interfaces aHH and aPP ( Fig . 4 left ) . Similarly , to obtain a 4-subunit profibril , “1+1+1+1” , “2+1+1” , “2+2” ( PUHH∩P + PUHH∩P ) and “3+1” ( PUHH∩PPUH + H∩P ) stackings are possible . We focus our study on the aPP interface formation in the ( 2+2 ) pathway . The resulting fibril would be arranged as PUHH∩PPUHH∩P with three interfaces aHH , aPP and aHH ( Fig . 4 right ) . We investigated the assembly of the 2+1 ( PUHH∩P + PUH ) construct for the 3-subunit profibril and the 2+2 ( PUHH∩P + PUHH∩P ) construct for the 4-subunit profibril . The simulations were initiated with the components ( PUHH∩P and PUH for the 3-subunit fibril and PUHH∩P + PUHH∩P for the 4-subunit fibril ) separated by 10 Å ( ∼3 water layers ) along the β-sheet stacking direction . Four 20 ns simulations were performed for the 2+1 and 2+2 systems at 310 K and the formation of the aPP interface was monitored . An ordered and stable aPP interface was formed in all eight simulations ( Text S1 ) . A representative structure of the resulting 3 and 4 cross-β-subunit quiescent fibrils is shown in Figure 4 . The binding energy for forming the aPP interface in the “2+1” or the “2+2” constructs was −163 . 4±9 . 9 kcal/mol , comparable to the number ( −159 . 9±7 . 4 kcal/mol ) seen for forming the in “1+1” aPP interface . The structural parameters are also comparable ( data not shown ) . Our proposed 3-cross-β-subunit asymmetric fibril ( PUHH∩PPUH ) structure has the following features: 1 ) the surface side chains of each of the 3 cross-β-subunits are not structural equivalent due to different environment ( e . g either exposed to solvent or buried at the aHH or aPP interface ) ; 2 ) the two exposed β-sheet surfaces of the fibril differ in hydrophobicity: one is polar/charged ( exposed residues H14 , K16 and E22 ) ; the other is quite hydrophobic ( exposed residues I31 , M35 and V39 ) ; 3 ) whereas the K16 and E22 residues at the aPP interface forms salt-bridges , the K16 and E22 residues at the surface are exposed to solvent and do not form salt bridges . 4 ) the 3 cross-β-subunits are tightly stacked and the thickness of fibril is ∼60 Å . Our proposed 4-cross-β-subunit fibril ( PUHH∩PPUHH∩P ) has two-fold symmetry and the two exposed surfaces are polar/charged . In addition , only half of the K16 and E22 residues from all 4 cross-β-subunits formed salt-bridges ( those at the aPP interface ) . We note that Tycko and co-workers report a slight conformation difference in the side-chains of the loop region ( residues 23–29 ) in the core cross-β-subunit between the agitated and quiescent structures . The use of the agitated cross-β-subunit as our initial building block should not affect our resulting structural model . Indeed , the loop is exposed to the solvent and plays little role in the β-sheet-to-β-sheet stacking that determines the overall morphology of the fibril . It is important to note that the loop region is highly flexible ( dynamic ) compared to the β-sheet regions . It is quite possible that if we ran the simulation longer , we would see some changes in the loop structure of the non-equivalent cross-β-subunit that experiences a different environment from the one seen in the symmetric agitated fibril . Recent experiments by Tycko and co-workers [35]on brain-seeded Aβ fibrils indicate that these fibrils bear strong morphological resemblance to the quiescently grown asymmetric synthetic fibrils . Both chemical shifts and dipole-dipole couplings [35]show the peptide in brain-seeded fibrils adopts the same β-strand-loop-β-strand conformation as in the asymmetric quiescent fibrils ( e . g . F19 , A30 , I31 , L34 and M35 in β-strands; D23 , V24 and G25 in non-β-strand conformation; presence of a D23-K28 salt bridge ) . MPL data indicate that the brain-seeded structures ( again like the quiescent structures ) consist primarily of fibrils with 3 cross-β subunits and NMR experiments show two sets of chemical shifts for many 13C-labeled sites . The primary difference between the brain-seeded and asymmetric quiescent fibrils lies in the orientation of the side-chains . 2D radiofrequency-assisted diffusion ( RAD ) spectra[35] indicate an additional F19-I31 side chain – side chain contact , suggesting the side chains in the C-terminal β-strand are “up-down” flipped as compared with the asymmetric quiescent fibrils . This flipping could be enabled by the flexible backbone of G29 residue , which could accommodate either orientation of side-chains in the C-terminal β-strand . Using the 3-fold asymmetric quiescent fibril model as a template , we construct a model for the brain-seeded fibril by flipping the side chains at the C-terminal β-strand ( Fig . 5 ) . The 3 cross-β-subunit model has both aHH and aPP interfaces . While the interactions at the aPP interface are the same as in the asymmetric quiescent fibrils , the detailed interactions at the aHH interface is changed as the sidechains are flipped ( i . e the side chains of I32 , I34 and V36 now interdigitate ) . The stability of our brain-seeded fibril model was confirmed by four 20 . 0 ns MD simulations at 310K in which the brain-seeded fibril was found to be stable , as judged from the small ( less than 2 Å ) root mean square distance ( RMSD ) from the starting structure . The binding energies for forming the aHH and aPP interfaces are respectively −155 . 4±5 . 9 and −160 . 4±7 . 9 kcal/mol , which are comparable to those in the synthetic fibrils with 3-cross-β-subunits . Much like mechanical agitation , the chemical oxidation of M35 can dramatically alter fibril lateral formation . In the case of the Aβ42 peptide , the β-sheet-to-β-sheet stacking process is completely blocked such that the resulting Aβ42 ( M35ox ) fibril contains only a single cross-β-subunit [19] . Since both the Aβ40 and the Aβ42 cross-β-subunits contain similar β-strand-loop-β-strand motifs ( they differ in the precise location of the loop ) , one would expect the M35 oxidation to affect Aβ40 fibrils in a similar manner as Aβ42 . The structure of the M35ox variant of Aβ9-40 has not been solved experimentally . Here , we consider a M35ox variant of Aβ9-40 and investigate the effects of the oxidation , first on the single cross-β-subunit , then on the assembly ( monitored by stacking simulations ) of the aHH and pHH constructs . We find that the stability of the 1-cross-β-subunit in our simulations is not affected by the single oxidation of M35 , likely a result of the fact that the side chains of the Met residues are exposed to the solvent and hence do not contribute to the stability of the cross-β-subunit . In contrast , the “1+1” assembly simulations ( with the M35 oxidation ) show a reduction in the number of trajectories that lead to an ordered assembled complex ( from 4 to 2 for aHH and from 3 to 1 for pHH out of a total four trajectories for each construct ) ( see Text S1 ) . This confirms that hydrophobic interactions play an important role in stabilizing the pHH and aHH interfaces . Introduction of a polar side-chain at the interface level ( here via single oxidation of the hydrophobic M35 residue ) significantly affects the formation of the hydrophobic interfaces . The stronger hydration tendency of the M35ox residues in the aHH ( M35ox ) and pHH ( M35ox ) constructs in comparison to the M35 residues in constucts aHH and pHH is directly supported by the presence of more water molecules in the first solvation shell ( <2 . 8 Å ) of these side chains , averaged over the last ns of the simulations ( Table 2 ) . ∼22 and ∼28 waters are present in systems aHH ( M35ox ) and pHH ( M35ox ) , respectively , while only ∼10 and ∼12 waters are present for systems aHH and pHH , respectively . Binding energy calculations also reveal a weaker binding energy ( less favorable binding ) between the two cross-β-subunits in constructs aHH ( M35ox ) and pHH ( M35ox ) than that in constructs aHH and pHH by ∼41 . 4 and ∼22 . 6 kcal/mol , respectively ( See Fig . 3 ) . Again , our finding is in a qualitative agreement with a recent stability study [22] of pre-constructed 2-cross-β-subunit species of Aβ40 M35ox mutants modeled using the CHARMM force field[42] . We predict that Aβ40 M35ox mutants would predominantly exist in a single layer structure . It would be interested to see MPL data on this system to confirm this prediction .
Amyloid fibrils are often generated via mechanical agitation in the laboratory , as this process speeds up fibril formation . Fibril formation in the brain , however , more likely resembles quiescent conditions . Indeed , MPL measurements performed by Tycko and co-workers [35] have recently shown that fibrils seeded from Alzheimer's brain-derived fibrils ( likely reflecting the relevant structures present in AD brains ) adopt a structure that has higher similarity to quiescent synthetic fibril structures[11] ( a 3 cross-β subunit structure ) than to agitated fibrils ( a 2 cross-β subunit structure ) . Furthermore , the brain-seeded fibrils show much greater morphological similarities to the asymmetric quiescent fibril structure than to the symmetric quiescent polymorph , presumably because more perturbations were involved in the seeding and growth procedure that generated the fibrils with symmetric structure . [18] Structures have been proposed for both the 2-fold agitated fibrils [20] , [23] , [32] and for the 3-fold symmetric , quiescently grown fibrils [18] . In both cases , the fundamental building block is the same cross-β subunit consisting of stacked β-strand-loop-β-strand motifs ( see Figure 1 ) . In the agitated fibril , two such cross-β-subunits are stacked laterally . In the symmetric quiescent structure , 3 cross-β-subunits are arranged in a triangular configuration . The atomistic structure of the asymmetric 3-unit quiescent fibril and the brain-seeded fibril , on the other hand are not known . In the present work , we propose the first atomistic structure for the asymmetric 3-subunit quiescent synthetic fibril using molecular dynamics that probe the assembly of the core cross-β subunits . This structure is then used as a template for a brain-seeded model that differs primarily from the synthetic quiescently grown fibrils in the orientation of the side chains at the C-terminal β-strand . Our simulations suggest that the asymmetric quiescent fibrils contain 3 cross-β subunits arranged in 6 tightly stacked β-sheet layers ( PUHH∩PPUH ) with two interfaces aHH and aPP ( Fig . 4 left ) . Our proposed structural model is consistent with the known constraints experimentally identified by Tycko and coworkers [11] , [32] , [43] . The experimental observations are the following: ( A ) the quiescent fibrils share similar cross-β subunit with the agitated fibril; ( B ) the quiescent fibrils predominantly contains 3 cross-β-subunits rather than the 2 cross-β-subunits seen in the agitated fibrils ( this information is obtained from analysis of the mass per length ( MPL ) values from STEM experiments ) ; ( C ) the quiescent fibril contains two structurally equivalent and one structurally non-equivalent parts . This conclusion is drawn from the fact that many residues exhibit two sets of 13C chemical shifts , with an approximate 2∶1 ratio of NMR signal intensities . In particular , splitting of I31 was observed even after three generations of the quiescent fibrils ( Fig . 2 of Ref . 10 ) . ( D ) Partial occupation of an intermolecular K16-E22 salt bridge . Tycko and co-workers report the presence of dipole-dipole couplings between side-chain Cσ carbons of E22 residues and side-chain Nζ nitrogens of K16 residues in quiescent fibrils , but not in agitated fibrils [11] . Our proposed structure clearly satisfies constraints A and B . Constraint C is satisfied as well: Our model ( See Fig . 4 left ) contains two structurally equivalent and one structurally non-equivalent parts ( PUHH∩PPUH ) : 2 equivalent layers at the interfaces ( aHH and aPP ) and one non-equivalent outer sheet-layer exposed to solvent ( P and H ) . Hence , the side chains on the peptide surface ( Residues H14 , K16 , V18 , F20 , E22 , and V24 of the polar β-strand/P and I31 , M35 and V39 of the hydrophobic β-strand/H ) would experience two chemical environments with a ratio of 2∶1 , consistent with the experimentally observed chemical shift splitting with a ratio of 2∶1 . As a specific example , we turn to residue I31 for which two sets of 13C chemical shifts , with an approximate 2∶1 ratio of NMR signal intensities , are observed experimentally . The implication is that this residue is found in two difference chemical environments . This is consistent with our three-layer asymmetric structure . One environment corresponds to the I31 residues being buried at the interface; the second corresponds to the I31 residues being exposed to the solvent . There are two instance where the I31 is buried , and one where it is exposed , corresponding to the experimentally observed 2∶1 splitting ratio . In terms of constraint D , our construct indeed shows partial occupancy of the K16-E22 salt bridge . K16-E22 salt bridges are formed at the aPP interface between the upper two sheet-layers ( PUHH∩PPUH ) . The K16 and E22 salt bridges on the outer polar surface are still exposed to water , leading to a 2/3 occupancy of the K16-E22 salt bridges ( See Fig . 4 left ) . It is important to note that the observation of multiple sets of NMR signals for a single labeled site in the fibril ( as seen in the experiments of Tycko cite ) does not rule out the presence of a co-existing population of symmetric structures along with asymmetric structures . Indeed , an alternate explanation for multiple sets of NMR signals is that the sample in reality contains a mixture of fibrils ( e . g different symmetric and asymmetric morphologies ) . However , one can argue in the case of the quiescently grown fibrils of reference[11] , [37] that the presence of both a K16-E22 salt bridge coupled to the presence of a 3 cross-β unit structure sufficiently implies that even in a polymorphic sample , the asymmetric structure would be the major species . The brain-seeded fibrils have not been characterized to the extent of the quiescent fibrils and many more NMR contacts remain to be established . Further experimental data for the brain-seeded fibrils ( for instance , a clear signature of a K16-E22 salt bridge at the aPP interface and further contacts at the aHH interface ) are required to fully validate our brain-seeded model . At present , the experimental data does not seem consistent with a symmetric 3 cross-β unit fibril as a major species , although such polymorphs may be present in the brain . It is important to note that the final morphology of a fibril is dictated by both thermodynamic and kinetic factors . The data of the 2005 Tycko paper [11] ( reporting the asymmetric structure ) and 2008 paper ( reporting the symmetric structure ) [18] pertain to fibrils grown under different conditions . It is apparent that the symmetric “triangle” structure cannot be energetically more stable that the asymmetric 3-layer structure , given the fact that there are far fewer hydrophobic contacts between the subunits . Entropically , the formation of the symmetric structure ( if one considers that it forms from pre-formed subunits , which may not be the case ) , would have to occur in a concerted 3-body 1+1+1 manner . An “open” 1+1 complex on its own would likely not be stable ( or at least not as stable as a closed stacked form ) . Thermodynamically , the stacked asymmetric structure is certainly going to be favored , with the symmetric structure likely a result of kinetic trapping during the experimental procedure . It is compelling to note that the brain derived structure , one that has formed slowly in the brain , perhaps even over decades ( ie that had more opportunity to find a thermodynamically stable structure ) , does not appear to be consistent with the triangle structure , but rather with a layered structure . Our stacking simulations enable us to propose a lateral growth mechanism for the formation of a multiple layer protofibril ( <24 peptides ) . This protofibril acts as a seed for the growth of mature fibrils by the addition of peptides to the two edges ( via the nucleation-growth mechanism[1] ) . This mechanism is shown in Figure 6 . In the first step , a 2-cross-β-subunit protofibril is assembled from two 1-cross-β-subunit protofibrils[38] by forming an aHH interface , which is stabilized by hydrophobic and van der Waals ( VDW ) interactions via interdigitation of the facing side chains . In the second step , the 2-cross-β-subunit protofibril with an aHH interface further assembles with another 1-cross-β-subunit into a 3-cross-β-subunit protofibril . The new interface aPP is stabilized by salt bridges , hydrophobic and VDW interactions . Growth to a 4-cross-β-subunit protofibril is possible , following one of two 2-body assembling pathways: formation of an aPP interface between two 2-cross-β-subunit protofibril ( 2+2 ) or formation of an aHH interface by adding a cross-β-subunit on top of 3-cross-β-subunit protofibril ( 3+1 ) . Further lateral growth into larger ( 5 or greater ) cross-β-subunit complexes is likely limited by the twisting of the β-sheet-layer and other structural defects in the cross-β-subunit which prohibits subunit-to-subunit stacking . In other words , the lateral growth is limited by a faster increase of the entropic cost ( i . e . fast decrease of translation , rotation and conformation entropy upon stacking ) than the increase of the favorable interactions . In fact , a maximum of 4 peptide layers/cross-β-subunits in the fibril is seen experimentally , as opposed to the ∼103 peptide repetition along the fibril axis for a ∼µm length fibril . This lateral growth mechanism explains the effects of external perturbation on synthetic fibril formation . For example , mechanical shaking of the solution kinetically blocks the formation of the aPP interface ( which is less stable than aHH interface ) probably induced by the air water surface . It would hinder the formation of a 3 or 4 cross-β-subunit leaving the 2-cross-β-subunit protofibril ( with aHH interface ) for growing mature 2-cross-β-subunit fibril as the major product . This is consistent with the experimental observation that under conditions of mechanical agitation , the predominant product is a 2-cross-β-subunit fibril . Another example is a chemical perturbation via oxidation that affects the structure of the Aβ fibrils . Our simulations suggest that , much as is the case for the Aβ42 ( M35ox ) fibrils [19] , the Aβ40 ( M35ox ) fibrils would exist predominantly in a single cross-β-subunit form . The oxidation significantly destabilizes the aHH interface ( we see disordered stacking trajectories and a weaker binding energy in system aHH ( M35ox ) and pHH ( M35ox ) . This prevents the formation of multiple ( >2 ) cross-β subunit fibrils , hence leading to predominance of a 1-cross-β-subunit fibril . It is tempting to speculate about why asymmetric quiescent fibrils are more toxic than agitated fibrils . Although the precise mechanism of toxicity of fibrils and early aggregates is still a matter of debate , it is likely that the exposure of hydrophobic side chains , normally buried in a folded protein or dispersed in an unfolded ensemble , is a key component in toxicity [1] . In the most stable 2-cross-β-subunit fibrils aHH ( PUHH∩P ) ( the most likely candidate for the structure of the agitated fibril ) , the continuous hydrophobic surfaces are buried , with the exterior sheet-layers hydrophilic . The solvent exposed surface of the 2-subunit ( PUHH∩P ) fibril ( e . g Y10_V12_H14_K16_V18_F20_E22_V24 of the N-terminal β-strand ) , with small hydrophobic patches interdispersed with non-polar residues , resembles the surface of a folded protein . In contrast , our proposed asymmetric quiescent Aβ40 fibrils with 3 cross-β-subunits ( PUHH∩PPUH ) has a ( large ) exposed continuous hydrophobic face ( -UH ) to the solvent ( i . e . G29_I31_G33_M35_G37_V39 of the C-terminal β-strand ) . This surface may interfere with the normal function of other proteins possibly by binding to and disabling them . In the same spirit , the 1-cross-β-subunit fibril with the M35ox substitution also has a large exposed hydrophobic surface , which may also be one of the factors responsible for the higher toxicity of the Aβ42 ( M35ox ) fibrils [44] over the wild type fibrils . If indeed having a the large exposed hydrophobic surface of fibrils leads to higher toxicity , then “detergent-like” ligands may provide an effective therapeutic for amyloidoses: they could be used to cover the hydrophobic surface by binding their hydrophobic part to the hydrophobic surface , thus exposing their hydrophilic part to the solvent . The exposed hydrophilic part would help improve the solubility of the protofibrils . In addition , these ambiphilic ligands might also cap the lateral growth of protofibrils by blocking the formation of the aHH interface . This may explain the mode of action of both a novel class of peptidic inhibitors designed by Soto et al . [30] and a weaker non-peptidic inhibitor ( Congo red ) [45] , both of which exhibit this ambiphilic feature ( hydrophobic/aromatic side chains on one face; hydrophilic on the other ) .
As the 1-cross-β-subunit is stable and rigid , we can define a local coordinate system as follows: The origin is set to the center-of-mass ( COM ) of the interfacing sheet-layer of the two sheet-layers for 1-cross-β-subunit; the three coordinates are along the β-sheet extension direction , β-strand direction and β-sheet stacking direction ( perpendicular to the β-sheet surface ) . Hence six parameters ( α , β , γ , a , b and c ) are used to characterize the structural relationship ( rotation and translation ) between two interfacing β-sheet-layers of the two 1-cross-β-subunits under a rigid body assumption: α , β and γ are the rotation angles of the β-sheet extension , β-strand and β-sheet stacking directions , respectively and ( a , b and c ) are translation distances along the three directions , respectively . The β-strand direction is defined by the direction of the third or fourth β-strand in the interfacing β-sheet-layer of the 1-cross-β-subunit . The β-sheet direction is defined by the same residues ( Cα atoms ) of the second and fifth β-strands; and the β-sheet stacking direction is obtained by the cross-product of the first two directions .
|
Amyloid diseases are characterized by the presence of amyloid fibrils on organs and tissue in the body . Alzheimer's disease , Parkinson's diseases and Type II Diabetes are all examples of amyloid diseases . Determining the structure of amyloid fibrils is critical for understanding the mechanism of fibril formation as well as for the design of inhibitor molecules that can prevent aggregation . In the case of the Alzheimer Amyloid-β ( Aβ ) peptide , the structure of fibrils grown under conditions of mechanical agitation has been elucidated from a combination of simulation and experiments . However , the structures of the asymmetric quiescent Aβ fibrils ( grown under conditions akin to physiological conditions ) and of Alzheimer's brain–derived fibrils are not known . In this paper , we propose the first atomically detailed structures of these two fibrils , using molecular dynamics simulations combined with data from previously published experiments . In additions , we suggest a unifying lateral growth mechanism that explains the increased toxicity of quiescent Aβ fibrils , the effects of external perturbations on fibril lateral architecture and the inhibition mechanism of the small molecule inhibitors on fibril formation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"computational",
"biology/molecular",
"dynamics",
"public",
"health",
"and",
"epidemiology/global",
"health",
"biophysics/protein",
"folding"
] |
2010
|
Molecular Structures of Quiescently Grown and Brain-Derived Polymorphic Fibrils of the Alzheimer Amyloid Aβ9-40 Peptide: A Comparison to Agitated Fibrils
|
The resident skin microbiota plays an important role in restricting pathogenic bacteria , thereby protecting the host . Scabies mites ( Sarcoptes scabiei ) are thought to promote bacterial infections by breaching the skin barrier and excreting molecules that inhibit host innate immune responses . Epidemiological studies in humans confirm increased incidence of impetigo , generally caused by Staphylococcus aureus and Streptococcus pyogenes , secondary to the epidermal infestation with the parasitic mite . It is therefore possible that mite infestation could alter the healthy skin microbiota making way for the opportunistic pathogens . A longitudinal study to test this hypothesis in humans is near impossible due to ethical reasons . In a porcine model we generated scabies infestations closely resembling the disease manifestation in humans and investigated the scabies associated changes in the skin microbiota over the course of a mite infestation . In a 21 week trial , skin scrapings were collected from pigs infected with S . scabies var . suis and scabies-free control animals . A total of 96 skin scrapings were collected before , during infection and after acaricide treatment , and analyzed by bacterial 16S rDNA tag-encoded FLX-titanium amplicon pyrosequencing . We found significant changes in the epidermal microbiota , in particular a dramatic increase in Staphylococcus correlating with the onset of mite infestation in animals challenged with scabies mites . This increase persisted beyond treatment from mite infection and healing of skin . Furthermore , the staphylococci population shifted from the commensal S . hominis on the healthy skin prior to scabies mite challenge to S . chromogenes , which is increasingly recognized as being pathogenic , coinciding with scabies infection in pigs . In contrast , all animals in the scabies-free cohort remained relatively free of Staphylococcus throughout the trial . This is the first experimental in vivo evidence supporting previous assumptions that establishment of pathogens follow scabies infection . Our findings provide an explanation for a biologically important aspect of the disease pathogenesis . The methods developed from this pig trial will serve as a guide to analyze human clinical samples . Studies building on this will offer implications for development of novel intervention strategies against the mites and the secondary infections .
Scabies is a skin disease caused by the parasitic mite Sarcoptes scabiei variety hominis in humans . It is common worldwide , predominantly affecting overcrowded , socio-economically disadvantaged populations [1] . Scabies is ubiquitous and a significant public health burden in the developing world with prevalence of up to 70% in rural India and between 18 and 42% in the South Pacific , and erratic reports from Africa and South America [2] , [3] . Recently scabies outbreaks are also reported regularly in economically rich regions , particularly from institutional settings such as health care facilities , elderly homes , prisons and child care centers [4] , [5] , where it is a well recognized and serious problem and control is notoriously difficult . Some studies suggest that the general population in economically stable societies has recently become more affected [6] . Moreover , scabies is a significant public health burden in the Indigenous population of tropical northern Australia [7]–[9] . Crusted scabies , a highly contagious manifestation of the disease presenting with extreme parasite numbers , can be seen predominantly in elderly individuals or in immunocompromised patients , especially those with infections due to HIV and human T lymphotropic virus 1 , or drug-induced immunosuppression [10] . Epidemiological studies indicated a link between the epidermal infestation with S . scabiei and cutaneous bacterial infections ( pyoderma , impetigo ) , particularly in tropical settings [9] , [11]–[14] . Secondary bacterial skin infections commonly associated with scabies infestations are primarily caused by two clinically important pathogens , i . e . Streptococcus pyogenes and Staphylococcus aureus , including methicillin-resistant and methicillin-sensitive strains [8] , [15] . These bacterial pathogens potentially cause life-threatening invasive infections , if left untreated . One obvious factor for the close association between mite infestation and bacterial disease is the breach of the physical barrier , i . e . the epidermal layers of the intact skin , through mechanical infringement by the burrowing scabies mites . This creates a suitable niche for pathogens to establish in the mite burrows . Our recent studies showed that scabies mites interfere locally with human complement mediated protection [16] , [17] , thereby promoting growth of S . pyogenes [18] and S . aureus ( Swe et al . , manuscript in preparation ) . However , the tripartite interactions between host , mites and bacteria are largely unexplored . In particular it is unknown whether scabies mites facilitate the transmission of pathogens or rely on obligatory endosymbionts for survival . Also , scabies-induced changes in the skin microbiome could provide a surrogate diagnostic biomarker for ordinary scabies infections , which are notoriously difficult to diagnose by conventional methods . Investigations into the identification of all bacteria associated with scabies mite infestation are important in developing novel control strategies as well as in improving current management and prevention policies . Obtaining clinical samples from human scabies patients is generally a logistical challenge , as scabies is not a notifiable disease , it is difficult to diagnose and outbreaks are sporadic . Ordinary scabies patients harbor only few mites , which complicates targeted sampling . In humans , scabies manifests at multiple skin sites providing multiple , highly diverse habitats for bacteria [19] . Additional elements such as climate , hygienic procedures , age and genetics potentially increase the variation of mite associated microbiota further . Finally , a study in human patients over the course of an infection , in line with ethical limitations , would be very difficult . Therefore a longitudinal pilot study of a defined infection site in a controlled porcine animal model was undertaken . Pigs are natural hosts for S . scabiei var . suis , developing clinical manifestations closely resembling human scabies [20] . The integumentary system , the innate immunity and many biochemical parameters are incredibly similar between pigs and humans [21]–[26] making the pig a well-recognized model to study human infectious diseases [27] . Focusing on the natural site of primary mite infestations in young pigs , i . e . the inner surface of the ear pinnae , we investigated the impact of a S . scabiei var . suis infection on the skin microbiota . In a longitudinal study conducted over 21 weeks we compared the 16S rRNA gene sequences reflecting the development of the normal skin microbial community structure in healthy control animals with the microbiota present prior to , during and after S . scabiei var . suis infection .
Animal care and handling procedures used in this study followed the Animal Care and Protection Act , in compliance with the Australian code of practice for the care and use of animals for scientific purposes , outlined by the Australian National Health and Medical Research Council . The study was approved by the Centre for Advanced Animal Science ( CAAS ) and the QIMR Berghofer Medical Research Institute Animal Ethics Committees ( DEEDI-AEC SA2012/02/381 , QIMR A0306-621M ) . Pigs were housed at the CAAS , Gatton , QLD . A total of 15 female , 3 weeks old Sus scrofa domesticus “Large White” breed siblings from the same pig breeding facility were adjusted to standard stable and feed conditions for 3 weeks prior to the start of the trial . After one week pigs were allocated randomly to 3 experimental groups ( n = 5 ) and housed in identical but separate standardized , climate-controlled rooms , set at an average temperature of 27°C–30°C . All rooms were run on a continuous flow basis and concrete floors were cleaned twice daily . To assist monitoring and sampling , the formation of skin lesions and crusts associated with mite infestation was recorded on a weekly basis using an established scoring system [28] . The tractable experimental porcine scabies model was developed previously [28] where high mite numbers are observed in the inner surface of the ear pinnae of young piglets . The development of crusted scabies is generally seen within 10 to 13 weeks after mite infestation , but only in a subset of infected animals . The majority of individuals , analogously to humans , develops ordinary scabies and then , depending on the mite challenge , overcomes the infection without developing severe infestation and crust formation . Therefore , a previously developed strategy [28] was adopted for one cohort of the trial , where extreme parasite infestation was achieved by treatment with the synthetic gluco-corticoids immune-suppressant Dexamethasone ( Provet , Brisbane ) . Synthetic gluco-corticoids are commonly used to promote infection in animal models [29] , and the crusted scabies following corticosteroid therapy has been observed in humans [30] . The schedule of 1 week as adjustment period to stable conditions and of subsequent 2 weeks as the period to achieve in all individuals the Dexamethasone levels required for fast mite infestation was previously established [28] . The three experimental cohorts of pigs served as a Mite infected group ( M ) , a Mite infected and Dexamethasone treated group ( MD ) , and an uninfected , untreated Control group ( C ) ( Figure S1 , supplemental information ) . The MD group was administered a daily oral dose of 0 . 2 mg Dexamethasone per kg of body weight , starting 2 weeks prior to scabies mite challenge and sustained continuously throughout the trial ( week -2 ) . At the start of the trial ( week 0 ) , prior to infection of the cohorts M and MD , a scraping of approximately 1×1 cm2 of the epidermal layer of the skin was sampled with a sterile curette ( Ø = 7 mm , Stiefel Laboratories Pty Ltd ) . This baseline sample was taken from the same site on the inside of the right ear pinna of each pig across all cohorts . Both MD and M cohorts were then challenged with comparable dosages of the scabies mite S . scabiei var . suis , as described previously [28] . In brief , crusts , sourced from a single site of a heavily infested individual , were dissected into approximately 0 . 5 cm2 pieces containing a few hundred mites and inserted into the vertical ear canal of both ears of the piglets . The animals were temporarily restrained , which prevented dislodgement of the crusts by agitation and ensured successful infestation . Skin scrapings were taken every two weeks unless otherwise stated , from all cohorts to monitor the development of the infection from healthy to moderate and subsequently to severe status of disease . Scrapings were performed in the same manner for every pig , alternating ears and following a predetermined map to ensure maximal conformity and to avoid repeated sampling of the same site . In the case of severe infestation , crusts were first lifted and collected for mite isolation . Subsequently the exposed skin area was sampled . After sample collection at week 16 , pigs in all cohorts including the mite-naive cohort C were treated with the acaricide Doramectin ( 1% ivermectin solution , Pfizer Animal Health ) by intramuscular injection at the recommended dosage of 300 mg per kg of body weight . The skin was allowed to heal completely for 5 weeks after Doramectin treatment . Final skin scrapes were taken at week 21 . All skin samples were collected in 2 ml reinforced centrifuge tubes ( Precellys , Bertin Technologies ) containing 200 µl of enzymatic lysis buffer ( 20 mM Tris , pH 8 . 0 , 2 mM EDTA , 1 . 2% ( v/v ) Triton-X100 ) and stored at −80°C until further processing . Samples representing base line ( week 0 ) , mild infestation ( week 7 ) , severe infestation ( week 10 ) and after healing ( week 21 ) were taken at pigs' ages of 6 , 13 , 16 , 27 weeks , respectively ( Figure S1 , supplemental information ) . The skin crusts collected from severe scabies infections were placed in lidded glass petri dishes and incubated over a moderately warm light source , which allowed mites to leave the substrate . Between 50 and 200 mites were collected into 2 ml reinforced centrifuge tubes ( Precellys , Bertin Technologies ) . Mites were washed twice for 7 minutes with shaking in 4% paraformaldehyde ( PFD ) followed by one wash in PBS to remove external bacteria . Mites were centrifuged for 2 min at 10 , 000 rpm , wash solutions were removed and the mite pellet was stored at −80°C . The skin samples were first incubated in lysozyme ( 20 mg/ml ) for 30 min in a 37°C water bath . Isolated mite samples were not treated with lysozyme . To facilitate homogenization , six 2 . 8 mm stainless steel beads ( Precellys , Bertin Technologies ) were added to both skin and mite samples . The samples were processed in a tissue homogenizer ( Precellys24 , Precellys , Bertin Technologies ) at 6 , 800 rpm for 30 s . Beads were removed and DNA was extracted from the samples using the DNeasy Blood and Tissue Kit ( Qiagen ) according to the manufacturer's instructions with the minor modification as follows . Samples were first incubated with 200 µl of Buffer AL , 40 µl Proteinase K in a 56°C water bath overnight and the standard protocol was followed for all subsequent steps . Purified genomic DNA was eluted in 100 µl buffer AE and the purity and concentration of the samples were analyzed by a spectrophotometer , NanoDrop 2000 ( Thermo Scientific ) . DNA samples were stored at −20°C until required . The bacterial 16S rDNA tag-encoded FLX-titanium amplicon pyrosequencing ( bTEFAP ) was contracted to Mr . DNA Molecular Research LP , Texas [31] , [32] . Briefly , 16S rDNA was amplified from purified genomic DNA using the primer pair 27F ( 5' AGRGTTTGATCMTGGCTCAG 3' ) -519R ( 5' GTNTTACNGCGGCKGCTG 3' ) spanning V1-V3 region ( ∼500 bp product ) . The concentrations of all DNA samples were adjusted to a nominal 20 ng/µl and a 1 µl aliquot of each sample was used per 50 µl PCR reaction . A single-step fusion 30 cycle PCR using HotStarTaq Plus Master Mix Kit ( Qiagen , Calencia , CA ) was performed under the following conditions: 94°C for 3 min , followed by 28 cycles of 94°C for 30 s , 53°C for 40 s , 72°C for 1 min , and a final elongation step at 72°C for 5 min . Amplicon products from different samples were mixed in equal concentrations and purified using Agencourt Ampure beads ( Agencourt Bioscience Corporation , MA , USA ) . Samples were then sequenced utilising Roche 454 FLX-titanium instruments and reagents following manufacturer's guidelines . The 16S rDNA sequences were processed using the software packages QIIME 1 . 5[33] . Barcode sequences were removed by searching for exact matches and primer sequences were trimmed allowing 1 mismatch . Chimeras were removed using ChimeraSlayer [34] . Taxonomic assignments were retrieved by the RDP Classifer v2 . 2 [35] with a confidence threshold of 0 . 6 . Operational taxonomic units ( OTUs ) were generated using the USEARCH package v5 . 2 . 32 [36] with an identity threshold of 97% . A representative sequence was selected for each OTU and taxonomically assigned with the RDP Classifier with a confidence cut-off of 0 . 6 . Subsequently , statistical analysis was performed using R and the Calypso software ( bioinfo . qimr . edu . au/calypso ) . The statistical analysis was done using one relative genus and OTU abundances , i . e . the number of reads assigned to each OTU or genus divided by the total number of reads obtained for each sample . Significant changes in the abundance of genera at different time points were detected by paired t-tests . Shannon index was used to estimate microbial community diversity ( OTU level ) . Representative sequences of OTUs assigned to Staphylococcus or Streptococcus by the RDP Classifier were used for phylogenetic analysis . Only OTUs with a relative abundance of at least 0 . 2% and 0 . 1% were included for the Staphylococcus and Streptococcus trees , respectively . Multiple alignments of reference 16S sequences of the corresponding genera were retrieved from the RDP database [37] and used as a reference to align the representative OTU sequences using HMMER3 [38] , [39] . Phylogenetic trees were reconstructed using FastTree [40] with a generalized time-reversible ( GTR ) model .
All control animals ( C group ) remained clear of mites and healthy throughout the entire trial . At week 7 all scabies treated animals ( M and MD groups ) showed symptoms of successful mite infestation , such as a typical rash over large parts of the body leading to scratching behavior . No other symptoms of compromised health were detected in the M and MD animals . Severe infestations with crust formation in the inner part of the ear pinnae were seen between weeks 10 and 16 , allowing isolation of mites . Crusts were formed in one member of the M group and in all animals treated with Dexamethasone ( MD group ) . The antiparasitic drug Doramectin® was administered by intramuscular injection at the start of week 16 , at a recommended dosage to kill off the mites within a week [41] . Infected skin healed within 5 weeks before the final scrapings were taken at week 21 . A total of 140 skin samples were collected over the 21 week trial period . The sampling site was restricted to the inner sebaceous surface of the ear pinnae because this is the primary site of early scabies infestations in pigs . Ninety six samples were subjected for pyrosequencing bacterial 16S rDNA , yielding 1 , 366 , 477 high quality 16S sequences with an average length of 407 base pairs . Sequence data has been deposited at http://www . ncbi . nlm . nih . gov/sra with the accession number SRX392076 . Samples taken from week 2 were excluded due to redundancy . A remaining subset of 57 samples collected in weeks 0 , 7 , 10 , 13 , 16 and 21 yielding 744 , 225 16S sequences were subjected to further analysis . Twenty samples from the inner sebaceous surface of the ear pinnae were obtained from the Control cohort C containing on average 8 , 053 sequences per sample ( range: 5 , 036–30 , 698 ) . In these twenty skin samples we identified 204 different bacterial genera with at least 5 assigned sequences . The major genera are listed in Table 1 . We observed significant changes in the skin microbiota of healthy animals during the 21 weeks period of the trial ( Figure 1 ) . At week 0 , Streptococcus was the most abundant genus of the skin microbiota of healthy pigs ( 23% of 16S sequences ) followed by Lactobacillus ( 13% ) ( Table 1 , Figure 1 ) . While Streptococcus remained relatively constant throughout the trial , Lactobacillus transiently dropped in abundance to 2% in weeks 7 and 10 ( p = 0 . 01 ) , but then became the most abundant genus with 45% of 16S sequences at week 21 ( p = 0 . 001 ) . We also observed a considerable reduction in microbial diversity: Lactobacillus and Streptococcus together represented about 30% of sequences at week 0 and had risen to about 70% at week 21 , thereby having largely replaced the next 10 abundant genera . This change in community composition was reflected in a reduction of the community diversity measured by Shannon index . The Shannon index at week 21 was significantly lower than in weeks 0 ( p = 0 . 009 ) and 10 ( p = 0 . 008 ) ( Figure 2bi ) and the evenness had dropped from 0 . 75 to 0 . 66 ( Figure S2 , supplemental information ) . During the first 10 weeks the microbial diversity fluctuated only moderately , as indicated by similar Shannon indices ( Figure 2bi ) . In summary , the skin microbiota became gradually less diverse as the healthy piglets matured , reducing from a complex and diverse assemblage during the juvenile stage to two dominating genera , i . e . Lactobacillus and Streptococcus at week 21 . While there is no comparable data from human infants , Grice et al . have previously reported that bacterial communities in adult human sebaceous microenvironments were less diverse than in other skin sites [42] . In total , 57 samples with a median of 8 , 798 sequences ( range: 2 , 467–55990 ) were included to study changes of the skin microbiota associated with scabies . Prior to S . scabies infection in week 0 the control ( C ) and scabies treated ( M ) cohorts showed similar skin microbial community profiles , whereas the skin microbiota of the scabies+Dexamethasone treated cohort ( MD ) was significantly different . Streptococcus and Lactobacillus were the most dominant genera in cohorts C and M ( Figure 2a ) but Lactobacillus ( 24 . 9% ) and Aerococcus were the most abundant genera within the cohort MD . Also , the microbial diversity was significantly lower in the MD cohort compared to the M and C cohorts ( Figure 2b ) . Dexamethasone treatment has previously been reported to improve neutrophil-mediated killing of streptococci in a rat animal model [43] , which could explain the observed differences in community composition . Notably , Dexamethasone treatment did not impact on the main changes in the skin microbiota seen in mite infested animals , as outlined below . At week 0 , before scabies mites were introduced , only a small percentage ( ≤0 . 5% ) of Staphylococcus was present in all three cohorts ( Figure 2a ) . During moderate scabies infection at week 7 , a major increase in Staphylococcus was observed . At week 7 , Staphylococcus abundance ranged from 6% to 76% in cohort M , and from 1% to 20% in cohort MD ( Figure 2a ) . In contrast , Staphylococcus was nearly absent in the mite-free control cohort C , ranging from 0% to 0 . 2% . During severe scabies infestation ( week 10 ) , similar abundance of Staphylococcus remained in M and MD cohorts as in week 7 , with one animal of the MD cohort reaching to 49% , compared with low numbers ( 0 . 2% to 1% ) in the mite-free control cohort C ( Figure 2a ) . In week 10 , when crust formation had occurred in one individual of the M cohort ( M2 ) and all animals of the MD cohort , a dramatic reduction in community diversity was observed in samples taken from crusted sites in these animals ( Figures 2bii and 2biii ) . The microbial community profiles from these skin samples , which were highly infested with mites , comprised up to 80% of Corynebacterium at the expense of other genera ( Figure 2a ) . In week 16 , pigs were treated with the antiparasitic drug Doramectin . At week 21 , after the mite infection was cleared in cohorts M and MD , the skin microbiota showed a similar diversity in all three cohorts ( Figure 2b ) . However , the relative abundance of individual genera was markedly different between the control cohort C and the previously mite infected groups M and MD ( Figure 2a ) . Staphylococcus remained abundant in most samples of the previously mite infected animals ( reaching 27 . 4% in cohort M and 35 . 5% in cohort MD ) compared to a minuscule presence of 0 . 2 to 1% in the control animals ( Figure 2a ) . At the same time Lactobacillus was present at low levels in cohorts M ( 14% , Table 1 ) and MD ( 12% ) while it was the dominant genus in the control cohort C ( 45% ) . Streptococcus was more dominant in the previously mite infected groups M ( 21% ) and MD ( 29% ) compared to the control cohort C ( 18 . 6% ) . At week 21 , when all pigs were free of mites , Corynebacterium was equal to or below 6% in all cohorts . The significant apparent increase in the Staphylococcus abundance in the scabies infected pigs indicated a correlation between S . scabiei infestations and Staphylococcus growth possibly combined with selective removal of other genera . To further identify the tentative species of Staphylococcus present , we constructed a phylogenetic tree of pig skin OTUs and known Staphylococcus species ( Figure 3a ) . Prior to trial commencement at week 0 , an OTU closely related to S . hominis ( OTU3 ) was present at a very low abundance across all cohorts ( Figure 3b ) . Staphylococci remained low throughout the trial in the scabies free control C . In contrast , from week 7 onwards Staphylococcus abundance significantly increased in cohorts M and MD and the community shifted from OTU3 to OTUs 1 , 6 , 8 , 9 and 14 , which are closely related to S . chromogenes ( Figures 3a and 3b ) . After the scabies mites had been killed by drug treatment at week 16 , staphylococci persisted in week 21 in cohorts M and MD . Two major taxonomic units , OTU1 and OTU2 ( closely related to S . chromogenes and S . auricularis respectively ) were in high abundance , followed by a lower abundance in OTU3 , OTU4 ( closely related to S . hominis and S . pasteuri ) and OTU7 ( closely related to S . felis ) ( Figure 3b ) . Staphylococci are part of the healthy skin microflora but are also common causative agents of pyoderma in pigs and other animals [44]–[46] , including humans [47] , with different species predominating in pig and human skin [42] , [48] , [49] . S . hominis is a normal skin commensal of human and animal skin , whereas S . chromogenes is recognized as the causative agent of exudative epidermitis in pigs [45] . Further , S . chromogenes was described to play a role in skin lesions , dermatitis and otitis media in sheep , associated with infestation by the sheep scab mite Psoroptes ovis , and has been identified in the skin microbiota associated with P . ovis [50] . S . auricularis , S . pasteruri and S . felis are coagulase negative , skin residents of human and animals , and were reported to occasionally cause diseases [39] , [51]–[53] . S . epidermidis prevails on healthy human skin while S . aureus is considered an important primary pathogen . In contrast , S . aureus was identified as the predominant species on healthy adult pig skin while S . epidermidis was less frequent [48] . A global increase in transmission of pathogenic methicillin resistant S . aureus strains has been reported between humans and animals [54] . In humans the link between scabies and S . aureus infections is well documented , however exclusively in epidemiological studies [55]–[60] . On human skin S . epidermidis usually has a benign relationship with its host [61] and was proposed to have a protective role in preventing colonisation with pathogenic bacteria , such as S . aureus [62] . Since S . aureus and S . epidermidis had been isolated from pigs before [48] we expected to detect these species on healthy skin and/or an increase in S . aureus during scabies infection; however neither was the case . Their absence in this experimental setting may be due to the sampling site being the sebaceous pinnae of the ears and not the back of the pigs , where they were detected previously [48] . Moreover , S . aureus likely becomes more abundant on the porcine skin after environmental exposure and human contact when housed under normal farming conditions [48] . The piglets in this trial were housed in a controlled facility without contact to other animals or humans , except for handlers who wore gloves and freshly washed overalls for every procedure . Notably , in the trial presented here , staphylococci were abundant only in the pigs that had encountered scabies mites ( cohorts M and MD ) , but barely measurable in the microbiome of mite free pigs in the control cohort C ( Figure 3b ) . The presence of Staphylococcus in large abundance unveils an obvious risk factor for future recurrent bacterial infections in scabies infected animals . While the overall microbial diversity of the mite free skin in control cohort C was similar to that of scabies infected skin in cohorts M and MD , the significant increase of the genus Staphylococcus in the skin of scabies infected pigs strongly suggests that the mite infection selectively favored the establishment of staphylococci . Complement appears to be a major primary defense mechanism of the vertebrate host targeted at mites as well as bacteria . Scabies mites release proteins that inhibit complement [16] , [17] and by reducing complement defense in their vicinity the mites may provide a microenvironment that fosters the survival of pathogenic bacteria [18] . S . aureus also displays an impressive arsenal of complement interference mechanisms [63] . Intriguingly , during human scabies infestations a substantial increase of S . aureus pyoderma is observed [64] , implying that the combined presence of mites and bacteria may further amplify the inflammation response . Similarly , the dominance of the pathogenic S . chromogenes over the commensal S . hominis in the mite infested pig skin may be driven by the production of a range of virulence factors produced by S . chromogenes and the mites . S . pyogenes is another important human pathogen commonly isolated from scabies associated pyoderma in human patients , a species that has not been reported in pigs . However , the genus Streptococcus was the most stable genus and relatively prominent in almost all skin samples throughout the duration of this trial ( Figure 2a ) . Importantly , this Streptococcus population was not severely affected by S . scabiei infection , presenting at a constant abundance prior to , during and after the scabies mite infection . A phylogenetic tree of Streptococcus was constructed to study changes on OTU level ( Figure 4a ) . OTUs closely related to S . alactolyticus dominated the streptococcal population ( OTUs 1 , 2 , 4 , 8 and 9 ) ( Figure 4b ) . S . alactolyticus is considered to be part of the normal gut microbiota of pigs and has been isolated from gastrointestinal tracts of newly weaned piglets and feces [65] , [66] . Our data showed that S . alactolyticus is also likely part of normal skin microbiota in juvenile pig ears . Interestingly OTU5 and 6 most similar to an opportunistic pathogen Streptococcus suis was present in a small proportion in the majority of the samples analyzed , but did not increase at any time point . S . suis is primarily an opportunistic pathogen of pigs but also an emerging human pathogen in the tropics [67] . Although S . suis is known to cause diseases in humans , particularly among pig handlers [68] , our results suggest that handling scabatic pigs may not pose additional risk in this regard . In week 10 , when crust formation had occurred in individuals M2 and MD1–5 , a dramatic reduction in community diversity was observed ( Figures 2bii and 2biii ) . Skin samples taken from crusted sites comprised up to >70% of Corynebacterium at the expense of other genera ( Figure 5 ) . While in an ordinary scabies skin scraping sample even a single mite is very rarely seen , the immediate skin layer below a crust generally contains high numbers of mites , which together with their internal bacteria are part of the skin microbiota . The samples taken from crusted areas in week 10 were heavily infested with mites . Coincidently , a high proportion of the same Corynebacterium sequences were also seen in samples generated from isolated washed mites , indicating that Corynebacteria are part of the mite internal microbiota . Consequently , the high abundance of Corynebacterium sequences detected in the skin scrapings taken from the severely infected sites is likely due to the mite internal microbiota from mites present in the scraping . At week 21 the previously mite infected but now mite free cohorts M and MD showed low levels of Corynebacterium ( Figure 2a , Table 1 ) , reinforcing the hypothesis that this genus could be enriched predominantly within the gut of the mites themselves . Members of this genus are facultative anaerobes and hence well suited to the locally lowered O2 within the mite gut beneath dense skin crusts . Symbiotic Corynebacteria have been isolated from the alimentary systems of a variety of arthropods feeding on skin , such as Triatoma infestans [69] and the tick species Ixodes ricinus , Dermacentor reticulatus and Haemaphysalis concinna [52] , providing potential novel strategies to control the transmission of diseases [70] . Thus , a subsequent study will focus on a thorough characterization of the scabies mite-internal microbiome and potential symbionts which could be targeted for scabies control . We demonstrated that scabies infestation has an impact on the host's skin microbiota . In an experimental porcine skin model , abundance of potentially pathogenic Staphylococcus species increased with the onset of infection , and remained beyond treatment and healing . At the end of the trial previously scabies infested animals showed a much reduced presence of Lactobacillus compared to the control animals . The local Streptococcus population remained stable in all cohorts throughout the trial and seemed almost unaffected by scabies . Corynebacterium abundance in heavily mite infested skin samples was possibly related to the mite-internal microflora . This study provides the first in vivo demonstration of a mite induced shift in the healthy skin microbiota , supporting direct evidence of the previously alleged link between scabies and pyoderma due to Staphylococcus infections , as seen in humans [9] , [11]–[13] . The study focused on the primary site of mite infestations in young piglets and the experimental setup allowed monitoring of the site over time , i . e . prior to , during and after scabies mite infestation . It provides a basis for future investigation in human patients . The study highlights that scabies mite infestation is not a simple ‘itch’ but should be viewed as a complex disease involving a change in the status of the skin microbiota , which gives rise to serious secondary infections .
|
Scabies is a neglected , contagious skin disease caused by a parasitic mite Sarcoptes scabiei . It is highly prevalent world-wide , and now recognized as a possible underlying factor for secondary bacterial infections with potential serious downstream complications . There is currently few experimental data demonstrating directly that mite infestation promotes bacterial infections . Due to remarkable similarities in terms of immunology , physiology and skin anatomy between pigs and humans , we developed a sustainable porcine model enabling in vivo studies of scabies mite infestations . Here , we investigated the impact of the scabies mite infection on the normal pig skin microbiota in the inner ear pinnae in young piglets . Samples obtained prior to , during infection and after acaricide treatment were analyzed by sequencing of bacterial 16S rDNA . We report that scabies infestation has an impact on the host's skin microbiota . Staphylococcus abundance increased with the onset of infection and remained beyond treatment and healing . A shift from commensal to pathogenic Staphylococci was observed . This study supports the link between scabies and Staphylococcus infections , as seen in humans . It is the first in vivo demonstration of a mite induced shift in the skin microbiota , providing a basis for a similar study in humans .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"ecology",
"staphylococcus",
"medical",
"microbiology",
"host-pathogen",
"interactions",
"microbial",
"pathogens",
"parasitology",
"biology",
"and",
"life",
"sciences",
"microbiology",
"pathogenesis",
"bacterial",
"pathogens",
"microbial",
"ecology"
] |
2014
|
Scabies Mites Alter the Skin Microbiome and Promote Growth of Opportunistic Pathogens in a Porcine Model
|
The pathogenesis of persistent viral infections depends critically on long-term viral loads . Yet what determines these loads is largely unknown . Here , we show that a single CD8+ T cell epitope sets the long-term latent load of a lymphotropic gamma-herpesvirus , Murid herpesvirus-4 ( MuHV-4 ) . The MuHV-4 M2 latency gene contains an H2-Kd -restricted T cell epitope , and wild-type but not M2− MuHV-4 was limited to very low level persistence in H2d mice . Mutating the epitope anchor residues increased viral loads and re-introducing the epitope reduced them again . Like the Kaposi's sarcoma–associated herpesvirus K1 , M2 shows a high frequency of non-synonymous mutations , suggesting that it has been selected for epitope loss . In vivo competition experiments demonstrated directly that epitope presentation has a major impact on viral fitness . Thus , host MHC class I and viral epitope expression interact to set the long-term virus load .
Gamma-herpesviruses characteristically persist in lymphocytes . Since the pool of latent genomes is constantly drained by viral reactivation , it must be replenished by virus-driven lymphoproliferation; this in turn is limited by host T cells; the steady-state viral load reflects an equilibrium of these fluxes . Viral loads are remarkably constant in one individual , yet vary hugely between them [1] . What determines the set point ? Such questions are difficult to address without animal models . Murid herpesvirus-4 ( MuHV-4 ) is one of the best established . It is genetically closer to Kaposi's sarcoma associated herpesvirus ( KSHV ) than to Epstein-Barr virus ( EBV ) , but shares with EBV a lymphoproliferative infectious mononucleosis syndrome [2] and persistence in memory B cells [3]–[5] . The pathogenesis of KSHV infection is presumably similar . The steady state MuHV-4 latent load does not appear to reflect the inoculating virus dose [6] , suggesting that it is set instead by host and viral genetic polymorphisms . Both host immunity and viral evasion contribute to MuHV-4 pathogenesis . Evasion dominates during acute latency amplification . The MuHV-4 K3 protein promotes this [7] by degrading MHC class I heavy chains and TAP [8] , [9] . K3 is transcribed in latently infected germinal centre B cells [7] , but also functions in lytically infected myeloid cells [10] , [11] , which are present in lymphoid tissue [4] and are probably an important source of the viral M3 chemokine binding protein [12] , [13] . The quantitative contribution of M3 to CD8+ T cell evasion remains controversial [14] , [15] . However , it is clearly capable of such a role [16] . M4 is another secreted lytic gene product that promotes latency amplification [17] , [18] . Thus , K3 may act both directly and by allowing lytically infected cells to protect latently infected cells in trans [19] . The MuHV-4 ORF73 episome maintenance protein has a further cis-acting CD8+ T cell evasion mechanism , equivalent to that of EBV EBNA-1 [20] , [21] , that is again vital for host colonization [22] . Despite immune evasion , virus-driven lymphoproliferation is brought under control by 3–4 weeks post-infection , at least in part by CD8+ T cells [23]–[25] . H2d mice mount a CD8+ T cell response against the M2 latency gene product at this time [26] . EBV latent loads and associated pathologies are similarly controlled by CD8+ T cells that recognize viral latent antigens [27] , [28] . EBNA-1-specific CD4+ T cells can also suppress EBV lymphoproliferation in vitro [29] , [30] , but whether equivalent recognition occurs in vivo is unclear [31]: even optimized latent CD4+ T cell epitope expression has little effect on MuHV-4 host colonization [32] . Most evidence would therefore suggest that gamma-herpesvirus latency is controlled principally by latent antigen-specific CD8+ T cells [27] , [28] . Vaccination with the M2 latency epitope has little effect on MuHV-4 latency establishment [33] because viral evasion dominates this setting . However , the impact of M2 recognition on the steady state viral load has not been defined . The balance of immunity and evasion could be subtly different here , for example if M3 function is now blocked by antibody . M2 itself promotes acute latency amplification [34]–[36] by modulating Vav-dependent B cell signaling [37] , [38] . The EBV LMP-2A [39] , [40] and KSHV K1 [41] , [42] have equivalent roles . M2 also has anti-interferon and anti-apoptotic functions [43] , [44] , although what these contribute to latency is unclear . An unusual feature of the M2 knockout phenotype in BALB/c ( H2d ) mice is that despite an acute latency deficit , long-term latency is increased [36] . C57BL/6 mice ( H2b ) , which are not known to recognize an M2 epitope , show the same acute latency deficit , but not the long-term increase [34] . Here we show that although M2 itself promotes acute latency establishment , its H2-Kd-restricted CD8+ T cell epitope is a major negative determinant of the long-term viral load . The reduction in latency associated with M2 expression required both its CD8+ T cell epitope and an appropriate host restriction element . Thus , host MHC class I polymorphisms interact with viral latency gene expression to determine the steady state gamma-herpesvirus load .
Our starting point were the observations that MuHV-4 M2 knockouts show an acute latency deficit in both BALB/c and C57BL/6 mice , but an elevated long-term latent load only in BALB/c mice [34] , [36]; and that mutating only the M2 amino acid residues critical for its interactions with Vav and Fyn [37] , [38] reproduces the acute latency deficit but not the long-term increase [37] . We hypothesized that a lack of H-2Kd-restricted epitope presentation might contribute significantly to the M2 knockout phenotype . We tested this further by comparing M2− ( vM2FS; a previously described M2 frame shift mutant [36] ) and M2+ ( vWT ) viral loads in different H2d and non-H2d mice ( Figure 1 ) . At 14 days post-infection , both H2d and non-H2d mice showed an M2-dependent latency deficit , consistent with M2 having an important role in acute latency amplification , when viral evasion limits CD8+ T cell function [7] , [14] , [22] . But by 50 days post-infection , when reactivatable wild-type virus was barely detectable in BALB/c mice ( H2d ) , M2− virus titres were maintained and now exceeded those of the wild-type . In contrast , long-term vWT titres in C57BL/6 mice ( H2b ) were equivalent to those of the M2 mutant; DBA/2 mice ( H2d ) were similar to BALB/c; FVB-N ( H2q ) were similar to C57BL/6; and B6 . C mice , where the H2d locus has been backcrossed onto a C57BL/6 background , were similar to BALB/c . M2 expression therefore increased the acute latent load independent of H2 type and reduced the long-term latent load in an H2-restricted manner: low long-term latency levels correlated with the H2d haplotype . Residues 84–92 of M2 contain its H-2Kd-restricted T cell epitope , GFNKLRSTL [26] . We tested whether the recognition of this epitope could explain the H2-restricted difference in long-term M2−/M2+ latent loads by mutating its anchor residues to alanines , either the phenylalanine at position 85 ( vM2F85A ) or the leucine at position 92 ( vM2L92A ) . We reverted the vM2F85A mutant in two ways: first conventionally , by restoring position 85 to phenylalanine ( vM2F85AR ) , and second by re-introducing the GFNKLRSTL epitope ectopically at the M2 C-terminus ( vM2F85AEPI ) . All these viruses showed an otherwise intact M2 locus , normal in vitro growth and normal replication in infected lungs ( Figure 2A–C ) . Intracellular IFN-γ staining of CD8+ T cells from infected mice ( Figure 2D ) showed that either anchor residue mutation prevented the generation of H-2Kd-GFNKLRSTL-specific CD8+ T cells , and that re-introducing the epitope into its ectopic site restored the response . The acute ( d14 ) latency titres of the anchor residue mutants in BALB/c mice were indistinguishable from the wild-type ( Figure 3 ) . Thus , there was no evidence that the point mutations affected M2 function . This was consistent with neither residue 85 nor residue 92 being crucial for the M2 Vav/Fyn interaction [37] , [38] . The C-terminal GFNKLRSTL epitope insertion also had no appreciable impact on latency establishment . At d14 post-infection , the impact of epitope presentation is limited by viral evasion . But in contrast to these normal acute titres , the long-term titres of the anchor residue mutants were increased , like those of the vM2FS mutant , while those of the vM2F85AR , vM2F85AEPI viruses were low , like those of the wild-type . Thus , the presence of a presentable epitope in M2 had a major impact on the long-term viral load . The increase in long-term viral load following CD8+ T cell epitope disruption implied that CD8+ T cell function helps to set this load in BALB/c mice . To confirm this , we depleted CD8+ T cells from vWT infected BALB/c mice by injection of anti-CD8 monoclonal antibody ( MAb ) . Importantly , depletion was initiated at 11 days post-infection , which is after the resolution of lytic infection but prior to the peak H-2Kd-GFNKLRSTL-specific CD8+ T cell response . The last MAb injection was performed at d19 . Latent loads were analysed at d21 post-infection ( Figure 4 ) . The variability in titer between depleted mice probably reflected incomplete depletion , as post-infection depletions are often less efficient than pre-infection ( our unpublished data ) , and the efficacy of depletion in individual mice infected with vWT or vM2F85AEPI correlated with viral load . Nevertheless , mice infected with the anchor residue mutants had significantly higher splenic latent loads than mice infected with epitope-expressing viruses before depletion , and not after CD8+ T cell depletion . CD8+ T cells were therefore responsible for the low latent loads of vWT and vM2F85AEPI . In situ hybridization for viral tRNA expression , a marker of lymphoid colonization [45] , [46] ( Figure 5A–B ) , showed similar results to the explant co-culture assays . Thus , the wild-type signal was high acutely ( d14 ) but low long-term; the vM2FS mutant had a low acute signal but was higher at later times; the anchor residue mutants had high signals both acutely and long-term; and re-introducing the GFNKLRSTL epitope reduced the long-term signal back to wild-type levels . The vM2F85A and vM2L92A mutants had both more viral tRNA+ germinal centres ( Figure 5B ) and bigger viral tRNA+ germinal centres ( Figure 5A ) , consistent with the idea that disrupting CD8+ T cell recognition of M2 allowed more extensive proliferation of latently infected B cells . We have previously correlated the higher long-term latent loads of M2− MuHV-4 in BALB/c mice with increased frequencies of viral genome+ germinal centre B cells [36] . We tested whether this applied also to the anchor residue mutants by subjecting flow cytometrically sorted germinal centre B cells to limiting dilution , PCR-based viral genome detection ( Figure 5C ) . The frequency of viral genome+ B cells was higher for the wild-type than for the vM2FS mutant at 14 days post-infection , and lower at 50 days post-infection; the vM2F85A and vM2L92A mutants showed high frequencies of viral genome+ B cells at both time points; and the vM2F85AR ( data not shown ) and vM2F85AEPI revertants were similar to the wild-type . Moreover , even at 133 days post-infection 1 in 19 GC B cells carried M2F85A DNA . At this time the frequency of M2F85AR DNA+ GC B cells was 1 in 5274 . Thus , the colonization of germinal centre B cells matched the total viral load in the spleen , with early colonization depending on M2 function and late colonization depending on T cell epitope presentation . M2 is positionally homologous to the KSHV K1 . In so far as both modulate B cell antigen receptor signalling [37] , [38] , [41] , [42] , they are also functionally homologous . Thus , it might be expected that MuHV-4 and KSHV share a latency program where M2 or K1 accounts for much of the presentable latent antigen . There is indirect evidence that this matters for K1: DNA sequence comparison between KSHV strains suggests that K1 has been positively selected for amino acid diversity [47] . A comparison of MuHV-4 with a closely related herpesvirus recovered from a shrew [48] shows the same phenomenon: M1 , M3 , M4 and ORF4 have non-synonymous to synonymous mutation ratios of 0 . 20–0 . 27 , while M2 has a ratio of 1 . 01 ( Andrew Davison , personal communication ) . In order to gain more direct evidence for M2 immune selection , we co-infected BALB/c mice with epitope+ and epitope− viruses , and quantified by real-time PCR viral genome loads in germinal centre B cells using virus-specific primers ( Figure 6 ) . For co-infection experiments , aliquots of the same viral preparations were used to formulate viral mixes that contained equal amounts of infectious units of each virus of interest , i . e . 5×103 PFU of each viral genotype , as determined by plaque assay . At d14 post-infection , there was little difference between vWT and vM2F85A , but by d50 vM2F85A accounted for >95% of the viral genomes . vM2F85AEPI and vM2F85A gave a similar result , the proportion of vM2F85A genomes increasing with time , while mixed vM2F85A and vM2L92A loads remained equivalent . Thus , viruses lacking M2 epitope presentation contributed disproportionately to long-term host colonization in BALB/c mice . The advantage of epitope null viruses over vWT was H2d-dependent , since it was not observed in C57BL/6 mice co-infected with vWT and vM2F85A . Here , vWT accounted for the majority of MuHV-4 genomes at all times .
Gamma-herpesvirus infections encompass complex combinations of cell types , anatomical sites , viral gene expression patterns and immune effector functions . This has made elusive a comprehensive understanding of how host immunity and viral evasion interact . Nevertheless , a consensus picture is now emerging . The long-term latent viral load is a key outcome , since it correlates with virus shedding [1] and probably also with disease . Maintained episomes replicate in step with normal cell division [49] . However , compensating for reactivation-associated latent genome loss requires a more complex program of virus-driven lymphoproliferation . This opens up another front between host immunity and viral evasion . Lytic reactivation could itself potentially re-seed latency , and this seems to be important in B cell-deficient mice [50] . However , these mice lack both the major MuHV-4 latency reservoir and virus-specific antibody , and consequently have an infection quite different to that of wild-type mice . The severe latency deficiency of viruses lacking episome maintenance [51] , [52] argues that as with EBV [53] , MuHV-4 persistence in immunocompetent hosts depends on lymphoproliferation . The efficiency with which proliferating B cells present CD8+ T cell targets must therefore also be important . The protection of B cells expressing M2 by K3 is probably only partial . The data presented here show that the impact of CD8+ T cell immunity can depend on a single epitope , consistent with epidemiological evidence of epitope selection in the EBV EBNA-3 [54] , [55] and the KSHV K1 [47] . The importance of a single latency epitope for MuHV-4 contrasts with lymphocytic choriomeningitis virus infection , where removing an immunodominant CD8+ T cell target simply brings out subdominant epitopes [56] . This may reflect that lymphocytic choriomeningitis virus does not suppress MHC class I-restricted antigen presentation , making the pool of possible epitopes larger . C57BL/6 mice illustrate what can happen when classical CD8+ T cell recognition of a key MuHV-4 target fails . Rather than overt disease , a back-up mechanism of non-classical Vβ4+CD8+ T cell recognition comes into play [17] , [57] . The higher latent loads of C57BL/6 mice despite massive CD8+Vβ4+ T cell expansion suggest this mechanism is not particularly efficient , and in C57BL/6 mice infected with K3-deficient MuHV-4 [7] or in BALB/c mice infected with the wild-type [17] CD8+Vβ4+ T cell expansion is minimal , presumably because classical recognition takes over . However , non-classical recognition appears to provide a safety net when host genetics or viral evasion limit normal antigen presentation . Unlike the fairly consistent and predictable effects of an attack on the cellular antigen presentation machinery or a silencing of viral transcription/translation , the interaction between viral epitope loss and host MHC class I diversity creates unstable and hard-to-predict outcomes . For example , the most pathogenic virus variant may be quite different between different out-bred hosts . Epitope selection may even allow gamma-herpesviruses contracted from close relatives to establish higher average latent loads than those from MHC class I-incompatible strangers . The data presented here argue that small variations in key viral latency genes can have major impacts on pathogenesis , and must therefore be considered in any attempt to understand the infection of individual hosts .
NIH-3T3-CRE cells [7] were grown in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% fetal bovine serum , 2 mM glutamine , 100 U/ml penicillin and 100 µg/ml streptomycin . Baby hamster kidney cells ( BHK-21 ) were cultured in Glasgow's modified Eagle's medium supplemented as above plus 10% tryptose phosphate broth . Murid gammaherpesvirus 4 ( MuHV-4 ) strain 68 was used in this study [58] . To prepare viral stocks , low multiplicity infections ( 0 . 001 PFU per cell ) of NIH-3T3-CRE or BHK-21 cells were harvested after 4 days and titred by plaque assay [59] . The vM2FS [36] , vM2F85A , vM2L92A and vM2F85AEPI viruses were derived from BAC-cloned MuHV-4 [60] . The M2F85A and M2L92A mutations were generated by overlapping PCR: A4353G , A4354C substituted alanine for phenylalanine at M2 position 85 , and A4332G , A4333C substituted alanine for leucine at position 92 . Mutated genomic fragments were inserted into a HinDIII-E genomic clone [61] cloned in pST76K-SR shuttle plasmid [60] using BlnI ( nt 3908 ) and XhoI ( nt 5362 ) restriction sites . To generate vM2F85AEPI , a genomic HinDIII/XhoI fragment ( nt 4029–5362 ) was PCR-amplified from an M2F85A template and cloned into pSP72 ( Promega ) . Genomic co-ordinates 3846–4029 were then amplified , using the primer 5′-AAAAAGCTTAGGGGATTCAATAAACTTAAGTCGACGTTATAACAGTGAAGGTGCTAACGCAGAA-3′ and cloned into the same vector as a BglII/HindIII fragment , thereby attaching the amino acid residues KLRGFNKLRSTL to the M2F85A C-terminus . This construct was again subcloned into the HinDIII-E shuttle plasmid using BlnI/XhoI restriction sites . A vM2F85A revertant virus ( vM2F85AR ) was generated using a wild-type HinDIII-E genomic clone . All PCR-derived regions were sequenced to confirm the integrity of the mutations . Each HinDIII-E shuttle plasmid was transformed into DH10B E . coli containing the wild type MuHV-4 BAC . Following recombination , mutated BAC clones were identified by DNA sequencing . The integrity of each BAC was confirmed by restriction digestion with BamHI and EcoRI . All viruses were reconstituted by transfecting BAC DNA into BHK-21 cells using FuGENE 6 ( Roche Molecular Biochemicals ) . The loxP-flanked BAC cassette was then removed by viral passage through NIH-3T3-CRE cells . The integrity of each reconstituted virus was checked by PCR of viral DNA across the HinDIII-E region . The stability of the introduced mutations was confirmed by viral DNA sequencing across M2 , both prior to infection and using viruses recovered from infected mice . 6- to 8-week old BALB/c , C57BL/6 , DBA/2 , FVB-N ( Instituto Gulbenkian de Ciência , Portugal ) and B6 . C . H2d mice ( kindly provided by C . Penha-Gonçalves , Instituto Gulbenkian de Ciência , Portugal ) were inoculated intranasally with 104 PFU of MuHV-4 under isofluorane anaesthesia . At different days post-infection , lungs or spleens were removed for post-mortem analysis . Titres of infectious virus were determined by plaque assay of freeze-thawed tissue homogenates on BHK-21 cells . Latent virus loads were quantified by explant co-culture of freshly isolated splenocytes with BHK-21 cells . Plates were incubated for 4 ( plaque assays ) or 5 ( explant co-culture assays ) days , then fixed with 4% formal saline and counterstained with toluidine blue for plaque counting . MuHV-4 infected BALB/c mice were depleted of CD8+ T cells by 5 intraperitoneal injections of 200 µg of monoclonal antibody YTS 169 . 4 [62] . Blood samples or splenocytes from depleted or control mice were stained with APC-conjugated anti-CD8α and phycoerythrin-conjugated anti-CD4 ( BD Pharmingen ) and analysed on a FCAScan Flow Cytometer using CellQuest software ( Becton Dickinson Immunocytometry systems ) . Spleen cells ( 1–2×106 ) were stimulated ( 6 h , 37°C ) with 1 µM GFNKLRSTL ( M284–92 ) or HYLSTQSAL ( EGFP200–208 ) peptides ( SIGMA-Genosys , Haverhill , UK ) in RPMI supplemented with 10% fetal bovine serum , 2 mM glutamine , 100 U/ml penicillin , 100 µg/ml streptomycin , 50 µM 2-mercaptoethanol , 10 U/ml recombinant murine IL-2 ( PeproTech , UK ) and 10 µg/ml Brefeldin A . The cells were then washed in PBS/10 µg/ml Brefeldin A , blocked with anti-CD16/32 mAb , stained with APC-conjugated anti-CD8a ( BD Pharmingen ) , washed twice , fixed in 2% paraformaldehyde ( 30 min , 4°C ) , washed once , permeabilized with 0 . 5% saponin , washed once , stained with a phycoerythrin-conjugated anti-interferon-gamma mAb ( BD Pharmingen ) and washed twice . All cells were analysed on a BD FACSCanto Flow Cytometer using FACSDiva software ( BD Biosciences ) . The frequency of MuHV-4 genome-positive germinal centre B cells was determined by limiting dilution and real-time PCR [37]: B220+PNAhigh B cells were recovered from pools of five spleens using a BD FACSAria Flow Cytometer ( BD Biosciences ) and serially two fold diluted . Eight replicates of each dilution were analysed by real time PCR ( ABI Prism 7000 Sequence Detection System , Applied Biosystems ) . The primer/probe sets were specific for the MuHV-4 ORF65 gene ( 5′ primer: GCCACGGTGGCCCTCTA; 3′ primer: CAGGCCTCCCTCCCTTTG; probe: 6-FAM-CTTCTGTTGATCTTCC–MGB ) . Samples were subjected to a melting step of 95°C for 10 min followed by 40 cycles of 15 s at 95°C and 1 min at 60°C . Real-time PCR data was analysed on the ABI Prism 7000 software . The purity of sorted cells was always greater than 97 . 5% . In situ hybridization with a digoxigenin-labelled riboprobe encompassing MuHV-4 vtRNAs 1–4 and microRNAs 1–6 was performed on formalin-fixed , paraffin-embedded spleen sections [46] . Probes were generated by T7 transcription of a pEH1 . 4 ( Roche Molecular Biochemicals ) . Positive follicles were scored using a Leica DM 5000B microscope . Splenic germinal centre B cells ( B220+PNAhigh ) were obtained from pools of three spleens using a BD FACSAria Flow Cytometer ( BD Biosciences ) and lysed overnight in 0 . 45% Tween-20 , 0 . 45% NP-40 , 2 mM MgCl2 , 50 mM KCl , 10 mM Tris-HCl pH = 8 . 3 and 0 . 5 mg/ml Proteinase K . Individual viral genomes ( vM2F85A/vWT; vM2F85A/vM2L92A; vM2F85A/vM2F85AEPI or vM2F85AEPI/vWT ) were quantified by real time PCR ( RotorGene 6000 5-plex HRM , Corbett Research ) , using a labeled probe specific for M2 and a common primer plus mutant-specific primer . vM2F85A/vWT viral mix: probe- 6-FAM-CATGGGGACTTTAACGTCGACCTAAGTT-TMR; common primer-GGTTAACTTCTTCAGGACTTGGTACA; M2F85A specific primer-TCCTAAAACCATAAGAAGGGGAGC; WT specific primer-TTTCCTAAAACCATAAGAAGGGGATT; vM2F85A/vM2L92A viral mix: probe- 6-FAM-TCCCCTTCTTATGGTTTTAGGAAAGCGA-TMR; common primer-CATCCCTCAGGAAATAAAAACAGTTC; M2F85A specific primer-GGCTTCCATGGGGACTTTAA; M2L92A specific primer-GCTTCCATGGGGACTTTGC; vM2F85A/vM2F85AEPI or vM2F85AEPI/vWT viral mixes: probe- 6-FAM-CCCCATGAACCCTGAGATACGTCTTCCT-TMR; common primer-TGGCTCGACTGACAGTCCAGA; M2F85A and WT specific primer ACCTAAGTTTATTGAATCCCCTAAGC; M2F85AEPI specific primer-GTCGACCTAAGTTTATTGAATCCCCT ( all primers and probes from TIBMolbiol ) . Samples were subjected to a melting step of 95°C , 5 min followed by 45 cycles of 15 s at 95°C and 45 s at 65°C . The wild-type , M2F85A , M2L92A or M2F85AEPI HinDIII-E shuttle plasmids were used as templates to derive standard curves . Real-time PCR data was analysed using Rotor-Gene 6000 Series Software . Data comparisons between different infection groups were performed using Student's t-Test , Friedman's Test or the Kruskal-Wallis Test as appropriate . For limiting dilution analysis 95% confidence intervals were determined as previously described [36] .
|
Persistent viruses present a major challenge to the immune response . Gamma-herpesviruses are a prime example , and the archetypal family member , Epstein-Barr virus ( EBV ) , has been studied for many years . A major unanswered question with EBV is why long-term virus loads—a key pathogenesis outcome—vary so widely between individuals . As most EBV studies are necessarily descriptive , the murid gamma-herpesvirus MuHV-4 provides an important focus of pathogenesis research . Here , we used MuHV-4 to address what determines long-term gamma-herpesvirus loads . We find a major role for a single MHC class I–restricted latency epitope . This reflects that latency-associated viral immune evasion and transcriptional silencing create a unique setting , in which the pool of possible epitopes is small enough for epitope loss to have a significant impact on viral fitness . Our data suggest that polymorphisms in viral latency genes and in host HLA class I together determine long-term viral loads .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"virology/persistence",
"and",
"latency",
"immunology/immunity",
"to",
"infections",
"virology/animal",
"models",
"of",
"infection"
] |
2008
|
A Single CD8+ T Cell Epitope Sets the Long-Term Latent Load of a Murid Herpesvirus
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.