Datasets:
BioLaySumm
/
BioLaySumm2025-PLOS

Dataset card Data Studio Files Community
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
Diagnostic guidelines for Visceral Leishmaniasis ( VL ) in the East African region are complex . Patients meeting the VL clinical case definition should be tested by rK39 rapid diagnostic test ( RDT ) followed by the Direct Agglutination Test ( DAT ) or tissue aspiration if RDT-negative . Otherwise , RDT-positive patients should be started on VL treatment . We evaluated how this guideline is adhered to by assessing the routine clinical practice in a university hospital in North-West Ethiopia . Retrospective record analysis was done for all patients who had an rK39-RDT done at University of Gondar ( UoG ) Hospital between June 2012 and June 2013 . We described the diagnostic work-up performed and the proportion initiated on VL treatment by test result . From a total of 928 patients tested , 308 ( 33 . 2% ) were rK39 RDT-positive . Spleen or bone marrow aspiration was done for 237 ( 77 . 2% ) RDT-positive patients . Of these , 165 were confirmed parasitologically , yielding a positive predictive value of 69 . 6% . Only 126 ( 20 . 3% ) of the 620 patients with a negative rK39 test underwent further testing by tissue aspiration , of which 22 ( 17 . 5% ) were also parasitology positive . HIV test results were available for 570 ( 61 . 4% ) patients and 36 ( 6 . 3% ) were HIV-infected . Of the 187 parasitologically confirmed patients , 182 ( 97 . 3% ) were started on VL treatment . A negative rK39 test was often not followed by further testing and a positive rK39 test result was followed by tissue aspiration in three out of four cases . Further research is required to understand why the diagnostic work-up did not comply with the guidelines , including evaluating adherence to the VL clinical case definition and quality of rK39-RDT testing . Visceral leishmaniasis ( VL ) is a vector borne , fatal disease caused by the protozoan parasite Leishmania . While Leishmania donovani is the etiologic agent in the East-African and Indian regions where transmission is anthroponotic , L infantum is the agent in the Mediterranean and Latin-American endemic regions , where transmission is zoonotic . It is an obligate intracellular organism that targets the reticuloendothelial system . The infection may initially remain dormant , manifesting in the presence of depressed cell mediated immunity at a later time . VL typically manifests with prolonged fever , weight loss , organomegaly and anemia [1] . Next to Sudan and South Sudan , Ethiopia has the highest VL burden in East-Africa . Moreover , North-West Ethiopia has the highest HIV co-infection rate in VL patients in the world , reaching up to 30–40% of those affected by VL [2 , 3] . Early diagnosis and treatment is one of the pillars of VL control . With regard to VL diagnostics , significant progress has been made over the last decade [4] . Rapid diagnostic tests ( RDT ) have been developed , most notably the rK39 RDT , which is cheap and easy to perform . While the diagnostic performance of this test was found to be consistently high in the Indian sub-continent , its sensitivity was suboptimal in the East-African region [5–19] . However , the RDTs have their well-known limitations , as they can also be positive in asymptomatically infected persons and in previously treated cases . Because antileishmanial antibodies wane only slowly after treatment , the rK39 RDT remains positive for several months to years after treatment , making the test useless to diagnose VL relapse in a patient who presents again with fever after an initial cure . According to the current World Health Organization ( WHO ) and Ethiopian guidelines , in endemic areas clinically suspected patients ( fever > 2 weeks , splenomegaly and weight loss ) without history of VL should undergo testing with rK39 RDT . If positive , they should be treated for VL [20 , 21] . Indeed , the probability of VL in those who fulfill the clinical case definition and are rK39 positive was generally found to be high ( a positive predictive value ( PPV ) > 90% at referral centers ) [17] . However , in areas where the sensitivity of rK39 is less than 90% like in East-Africa [5] , a negative test result should be followed by another test . This can be the direct agglutination test ( DAT ) or parasitological evaluation on tissue samples ( WHO TRS 2010 ) [21] . The negative predictive value ( NPV ) of the rK39 test in east Africa was found to be 81% as compared to 95% in the Indian sub-continent [17] . The VL and the HIV guidelines also specify that all VL suspects should be tested for HIV since it affects the indicated diagnostic work-up and treatment [20 , 22] . The evidence-base of these diagnostic algorithms builds on several well-conducted high quality studies [23–25] that have led to national and international guidelines [20 , 21] . There is , however , surprisingly limited data on how these tests and algorithms are actually used in routine clinical practice settings . For instance , not adhering to the VL case definition could influence the overall performance of the algorithms . Second , quality assurance can be challenging in resource poor and remote VL endemic regions , with no guarantee of the quality of the RDTs circulating within national programs and their performance . We performed a retrospective assessment on how the RDTs are used in clinical practice in a teaching hospital in Ethiopia . We used routinely collected data to assess the rK39 RDT utilization practices and alignment with recommendations in the guidelines . Amongst all individuals referred for rK39 testing in a VL research and treatment center in North-West Ethiopia , we describe the diagnostic work-up performed in relation to the actual rK39 RDT test result ( and other key clinical factors ) , the proportion tested for HIV and the proportion initiated on VL treatment . Ethical clearance for the retrospective analysis of the routinely collected data was obtained from the UoG Institutional Review Board . The data was collected and handled maintaining the confidentiality of the subjects without including their names . The Leishmaniasis Research and Treatment Center ( LRTC ) was founded in 2005 and is situated on the premises of the University of Gondar ( UoG ) Hospital in North-West Ethiopia . It is part of the Leishmaniasis East Africa Platform ( LEAP ) , an international clinical research network that includes investigators from Sudan , Kenya and Uganda . LEAP conducts collaborative clinical trials aimed at improving the management of leishmaniasis patients in the region . Good Clinical Practice and Good Clinical Laboratory Practice compliant clinical trials have been conducted at the LRTC center since 2005 with the support of the Drugs for Neglected Diseases initiative ( DNDi ) . In addition to the clinical trials , the center also provides free diagnostic and treatment services for all types of Leishmania patients coming to the UoG Hospital . Patients may directly come to the center or they may be referred to the center either from the different units of the hospital ( medical outpatient service , emergency unit and medical wards ) or from other health institutions ( health centers and district hospitals ) in the catchment area . The center has its own laboratory where most of the necessary investigations for Leishmania diagnosis and follow up are conducted . These include complete blood count , rK39 RDT , blood chemistry and microscopy of tissue aspirations . The rK39 RDT is done only in the laboratory of the center , and not in any of the other laboratories in the hospital . However , requests can be made from any of the departments and units in the hospital . There is no standardized rK39 RDT test request form; hand-written requests are made using different lab request formats . The main hospital of UoG has a general clinical laboratory , an HIV care/ART laboratory , an emergency services laboratory and a teaching and training laboratory . Patients suspected of VL are evaluated at the different units of the hospital ( outpatient departments , emergency unit , medical and pediatric wards ) and referred to the LRTC for the necessary procedures and laboratory tests related to leishmaniasis . Quite a number of patients present directly to the LRTC . Unless there is a history of previous VL ( relapse ) , rK39-RDT testing ( DiaMed-ITLEISH-Bio-Rad Laboratories ) will be the initial diagnostic test , with additional testing according to the RDT result and clinical evaluation . The white cell count , hemoglobin and platelet counts are systematically done , HIV testing and counseling is offered to all as part of provider initiated testing and counseling ( PITC ) . Some of the patients may be referred to the center after some or all of the laboratory work up is done while others undergo the clinical evaluation and all laboratory tests at the center . After the results of the rK39 tests are collected from the LRTC laboratory , physicians requesting the test from the different departments and units will make their decisions concerning the patients . This may be followed by further work-up for other diseases or linking to the LRTC for further confirmation of VL or VL treatment . The final treatment decision for VL is made by LRTC physicians . The confirmatory test used for VL diagnosis is tissue aspiration either from the spleen or bone marrow . While the spleen is the primary choice for its better yield , bone marrow aspiration will be done in case of increased bleeding tendency , small or non-palpable spleen , presence of ascites or pregnancy . Giemsa stained smears are examined under high power oil immersion microscopy and parasite load grading performed according to the WHO guidelines [21] . All VL cases are admitted for treatment , mostly to the LRTC , while non-VL patients will be referred to the other units of the hospital for their respective medical condition . Occasionally , often on patient request , VL patients may be referred to other VL treatment centers . Anti-leishmanial drugs are dispensed from the research center pharmacy to patients admitted to the center as well as to the other wards of the hospital . We conducted a retrospective record analysis of routinely collected data including all patients who had an rK39-RDT done at the laboratory of the LRTC of UoG hospital for suspicion of primary VL between June 2012 and June 2013 . Data were sourced from different registers of LRTC . This included the laboratory register , in which all individuals undergoing rK39 testing were entered and all laboratory tests were recorded; the ward patient registration book , recording all cases evaluated at the center; and the pharmacy dispensing log . Additional data was collected from the hospital laboratory register to complete the missing data for patients referred to the center after some investigations ( hematology and HIV tests ) at the hospital laboratory . A standard format was prepared to collect the following information on all individuals that were tested with the rK39 test: rk39 test result , age , sex , full blood count result , tissue aspiration site and result , HIV status and treatment decision for VL . Data were entered into an excel spread sheet and transferred to STATA 12 for analysis . PPV and NPV of the rK39 RDT were calculated comparing with the tissue aspiration results . Between June 2012 and June 2013 , a total of 928 rK39 RDT tests were performed for clinical suspicion of VL . Most patients tested were male ( 89 . 2% ) , with a median age of 25 years ( IQR 20–30 ) . No data were available on the proportion of those 928 complying with the clinical case definition , because this was not routinely recorded . Out of 668 with a complete peripheral blood count result , 74 ( 11 . 1% ) had a normal profile; 354 ( 53% ) pancytopenia; 142 ( 21 . 3% ) bicytopenia and 98 ( 14 . 7% ) monocytopenia ( Table 1 ) . Leukopenia ( white blood cell count <4500/μl ) was found in 67 . 7% ( 457/675 ) , anemia ( hemoglobin <11g/dl ) in 74 . 1% ( 501/676 ) and thrombocytopenia ( platelet count < 150 , 000/μl ) in 75 . 1% ( 503/670 ) . In total , 308 ( 33 . 2% ) had a positive rK39 RDT test result . Hematological abnormalities were more common in individuals testing rK39 positive ( Table 1 ) . Of the 308 rK39-positive individuals , 237 ( 77 . 2% ) underwent tissue aspiration ( 139 from bone marrow and 225 from spleen ) , yielding a positive result in 165 ( 69 . 6% ) ( PPV 72 . 6% from bone marrow and 67 . 1% from spleen ) . There were 71 ( 32 . 8% ) rK39-positive patients who did not undergo a tissue aspiration . Of the 620 individuals with a negative rK39 test result , 126 ( 20 . 3% ) underwent tissue aspiration , which was positive in 22 ( 17 . 5% ) ( Fig 1 and Table 2 ) . As many of the patients were referred for the test from different units outside of the LRTC , the subsequent diagnostic efforts made and ( non-VL ) treatments prescribed were not known for most of these patients from the available LRTC registers used . HIV test results were documented for 570 of the patients and 36 ( 6 . 3% ) were HIV positive . The rK39-RDT was positive in 15 ( 41 . 7% ) of the HIV positive patients , and 168 ( 31 . 5% ) of the HIV negative group . Overall , 57 . 5% of the rK39-positive cases received VL treatment . Of the rK39 negative cases , 4 . 4% received VL treatment . There were a total of five parasitologically-confirmed cases that did not receive VL treatment at the LRTC . Three patients were referred to other treatment centers in the region; two patients disappeared after VL diagnosis . RDTs have enabled an easy diagnosis of several diseases and have allowed to scale up treatment and control programs . However , there are concerns related to the high volume of tests used , the interpretation of results and quality assurance in the routine setting . This is one of the first studies evaluating VL testing practices by RDT in a routine clinical setting . While HIV test results and the subsequent actions for the rK39 negative patients were not available for a substantial proportion of individuals , linkage of the parasitologically confirmed VL cases to their treatment was good . The most striking finding was that the diagnostic work-up did not comply with the national guidelines [20]: a negative rK39 test was most of the time not followed by a second test for leishmanasis . On the other hand , a positive rK39 test result was followed by tissue aspiration in three out of four cases . The need for a parasitologically confirmed diagnosis in clinical trial protocols implemented at this clinical research center may partly explain the latter , but not the first scenario . Recent data indicate a sensitivity of 87 . 2% ( 82 . 5–90 . 8% ) and specificity of 96 . 4% ( 93 . 3–98 . 1% ) of rK39 RDT testing ( DiaMed-ITLEISH-Bio-Rad Laboratories ) in East-Africa [19] . However , this was a case-control study where a negative DAT-result was used to define the negative controls; and the sensitivity and specificity may hence be overestimated . A recent Cochrane review estimated the pooled sensitivity of the rK39-RDT at 85% ( 95% CI 75 to 93% ) and the pooled specificity at 91% ( 95% CI 80 to 97% ) for the East African region ( 18 ) . In most reported diagnostic studies from the region the prevalence of VL ( pre-test probability ) among clinical suspects was relatively high , ranging between 40 and 70% [17 , 25 , 26] . Consequently , the probability of VL in those testing positive in the rK39 test ( the post-test probability of VL or the PPV ) was typically above 90–95% [18] . In this regard , the probability of VL of 69 . 6% in those with a positive rK39 test found in this study was much lower than expected . Even if we would consider all the 71 rK39-positive patients that had not undergone tissue aspiration as true VL cases , the positive predictive value would still only be 76 . 6% . Providing clear-cut answers as to why diagnostic practices in this hospital differed so much from recommendations on the guidelines will require additional research , and the same issues should address at other levels of the health system . However , we would like to present the following considerations—either related to the study population undergoing rK39 testing , or related to the rK39 test performance—that could potentially explain the findings , and which could be operating in parallel . First , the VL diagnostic algorithm stipulates that testing should only be performed in individuals meeting the VL case definition [20] . More lenient testing practices ( i . e . not strictly adhering to the VL case definition ) would lead to a drop in VL pre-test probabilities and automatically also in post-test probabilities . It could also be that some patients with a repeat episode of fever after successful VL treatment might have been included in our series , which would reduce test specificity . Especially in a teaching hospital setting where rK39 testing can be requested from many different services and department , and by different types of health care providers , it is difficult to control strict adherence to the recommended indications for testing . Due to the lack of a standardized laboratory request form for the rK39 test , ( that should include essential clinical information ) , it is difficult to ascertain if the subjects tested with the rK39 really qualified or not in this retrospective analysis . All subjects to whom rK39 test is requested do not undergo reassessment at LRTC and subsequent linkage of the rK39 negative patients to the center is low . The test performance also depends on the characteristics of the population being tested and the spectrum and severity of underlying conditions [19 , 27] . The patient population in our study differed from those included in diagnostic accuracy study settings , especially in the sense that we hypothesize that the VL clinical case definition was not strictly applied . A lower prevalence of VL in the population undergoing rK39 testing would automatically result in higher negative predictive values in those testing rK39 negative , possibly leading physicians to abstain from doing a follow-up test as per VL guidelines . However , this assumption cannot be verified and requires further study . A declining incidence of VL within Ethiopia may be an alternative explanation for a possibly lower pretest probability of VL , but there are so far no such indications at the national and regional level . Based on the monthly reports at the LRTC , the VL case load in the hospital has been stable over the last five years . The pattern of disease in the referral hospital setting may be another factor to consider . The region is also endemic for malaria , schistosomiasis , chronic viral hepatitis that mimic VL in their clinical presentation . Given the variety of complex and difficult to diagnose cases and atypical clinical manifestations seen at this health care level , physicians may request more tests to get additional laboratory diagnostic clues . The effect of HIV—affecting rK39 RDT test performance—could not be clearly assessed in our study , due to the small sample size and missing data on HIV status . Besides less strict selection of the patients undergoing testing in routine clinical practice as compared to study settings , there might also be issues with the diagnostic performance of the tests used or with the test execution [27] . In many resource limited VL endemic regions , there is a lack of proper quality control and assurance systems covering all stages from evaluation of test performance , procurement to transportation and storage of the test kits used . While differences in performance between different brands have been well-documented , lot-to-lot variations could occur as well . Finally , national systems for training and supervision of rK39 test execution are non–existent . We are aware that many alternative explanations or contributing factors might exist for the negative tissue aspirations in rK39-positive individuals , requiring further attention . For instance , the context of scarcity of drugs might encourage physicians to obtain a higher diagnostic certainty before starting VL treatment . Although only very few of the patients included in this study were actually enrolled in an ongoing clinical study , routine clinical practice in our center for patients directly presenting at the LRTC might still be influenced by the fact of being a clinical trial site . The tissue aspiration practices cannot be attributed to secondary financial gain by the clinic or the lab as these services are done for free . Finally , the suboptimal sensitivity of tissue aspiration might have contributed to the low PPV found in this study . Although substantial proportions ( 40% ) of the aspirates were from bone marrow punctures , the diagnostic yield of spleen and bone marrow aspiration was comparable in this study . Last but not least , we should also verify a possible quality issue with parasitology readings at LRTC , as no test can ever be considered error-proof . Importantly , the findings in our study do not necessarily reflect the reality at primary health care setting where midlevel health workers practice . At health center level tissue aspiration is not done and there is no other test available for further work-up . There are also limited supplies of diagnostics and medicine . Pending further research , the following recommendations emerge from our study . First , adherence to the VL case definition during request for Leishmania tests should be assessed and enhanced [20 , 21] . Continuing medical education and supervision is indicated for health care professionals at all levels and hospital services to ensure a strict adherence to the VL case definition when using this RDT as a basis for a therapeutic decision . The design of a request form for rk39 RDT with tick boxes for the different components of the case definition could also be useful . Second , quality assurance systems should be put in place that guarantee the quality of the diagnostic devices delivered at the health care facility , and the quality of the test execution and interpretation . Experiences of quality control and quality assurance systems designed for other diseases such as malaria can be adapted for leishmaniasis as well [19 , 27] , and those should include all tests , RDT as well as microscopy . Finally , national VL programs should establish data collection tools , allowing regular monitoring of testing practices and linkage to treatment , including treatment outcome monitoring . Such tools could function as an additional safety-check allowing the rapid identification of poorly performing RDTs . This is particularly important in settings where national VL control programs are still relatively weak and dependent on external actors for funding and/or delivery of tests . Moreover , there are now many different brands of rK39 RDT tests being marketed [19] , as well as new RDTs based on different antigens ( rK26 , rK28 ) . Not all of these have been appropriately evaluated in field settings or specific geographical regions . In one of the major teaching hospitals in North-West Ethiopia , the diagnostic practices in VL do not fully comply with the national recommendations [20 , 21] in that a second leishmania test is not done for most of the rK39 negative suspects . The national treatment guidelines have a provision for a second test where rK39 tests are negative in highly suspected patients . Unfortunately a second test that can be deployed in the treatment centers is hard to find , as the DAT is costly and hard to deploy for routine testing . In addition , a positive rK39 test was followed by tissue aspiration in the majority of cases; and among those undergoing this procedure , three in ten were found to be parasite negative . This high rate of negativity by parasitology among rK39 positives can partly be explained by a false positive rK39 as well as by the inherent problems of microscopy . Further research is required to understand whether our findings in this reference center are generalizable to other levels of the health services . Besides evaluations of adherence to the VL case definition to guide rK39 testing , quality control measures for the rK39-RDT as well as parasitology should be established to ensure that VL diagnosis is carried out to the best interest of the patients . Availability of quality-assured rK39 for use in the national program and correct use and execution at the point of care are issues that deserve due attention .
The introduction of RDTs is one of the major advancements in leishmaniasis control programs . While the variability in performance from one endemic region to the other is well recognized , the utilization of these RDTs in the routine clinical setting has not been evaluated to date . In this study , we showed that the RDT use in routine practice setting has large deviations from the guidelines . Clinical suspicion of VL should be based on the full criteria of case definition as is recommended in the guidelines and the presence of the criteria should be checked by clinicians before requesting the RDT . Not respecting these clinical criteria can lead to low pretest probability of the disease and eventually low positive predictive value of the test used . Additionally , introducing regular monitoring activities with training and quality assurance system for leishmania RDTs is very important .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[]
2015
Impact of the Use of a Rapid Diagnostic Test for Visceral Leishmaniasis on Clinical Practice in Ethiopia: A Retrospective Study
Visceral leishmaniasis is an important parasitic disease of the developing world with a limited arsenal of drugs available for treatment . The existing drugs have significant deficiencies so there is an urgent need for new and improved drugs . In the human host , Leishmania are obligate intracellular parasites which poses particular challenges in terms of drug discovery . To achieve sufficient throughput and robustness , free-living parasites are often used in primary screening assays as a surrogate for the more complex intracellular assays . We and others have found that such axenic assays have a high false positive rate relative to the intracellular assays , and that this limits their usefulness as a primary platform for screening of large compound collections . While many different reasons could lie behind the poor translation from axenic parasite to intracellular parasite , we show here that a key factor is the identification of growth slowing and cytostatic compounds by axenic assays in addition to the more desirable cytocidal compounds . We present a screening cascade based on a novel cytocidal-only axenic amastigote assay , developed by increasing starting density of cells and lowering the limit of detection , and show that it has a much improved translation to the intracellular assay . We propose that this assay is an improved primary platform in a new Leishmania screening cascade designed for the screening of large compound collections . This cascade was employed to screen a diversity-oriented-synthesis library , and yielded two novel antileishmanial chemotypes . The approach we have taken may have broad relevance to anti-infective and anti-parasitic drug discovery . The protozoan parasites of the genus Leishmania are the causative agents of leishmaniasis , a group of diseases that is prevalent in 98 countries and 3 territories with approximately 1 . 3 million new cases occurring annually . There are estimated to be 20 , 000 to 40 , 000 deaths per year [1] . Leishmaniasis occurs in three main forms . Visceral leishmaniasis is the most severe form , where the parasites migrate to the internal organs , particularly the spleen and liver , resulting in death if untreated [2] . Cutaneous and mucocutaneous leishmaniasis are dermal infections which carry lower mortality but are highly disfiguring . The associated social stigmatization can have significant negative effects on psychological well-being [3 , 4] . Currently , treatment of visceral leishmaniasis is limited to a few drugs used either in monotherapy or in combination . For reasons still to be understood , some of the treatments show a lack of clinical efficacy in certain geographic regions , notably in East Africa . In India and Nepal the emergence of resistance to antimonial therapy also limits the treatment options . Additionally , the occurrence of moderate to severe adverse effects , the existence of contraindications such as pregnancy in the case of miltefosine , the relatively high cost of the treatments as well as the logistical complexity related to the storage and use of drugs in the endemic regions further restrict the use of amphotericin B , miltefosine , antimonials and paromomycin . There is therefore an urgent need for new and better drugs to address treatment needs [5–7] . High-throughput screening of diverse compound sets in phenotypic assays has proved an effective way of discovering new start points for drug discovery [8 , 9] . To facilitate screening of large compound collections against Leishmania , axenic amastigotes have been used as a surrogate for the disease-causing intracellular form [10–12] . Axenic amastigotes are thought to be a more relevant model of the human lifecycle stages of leishmania infection in comparison to the promastigote [13–17] . Several assays using axenic stages of Leishmania have been developed to the scale and robustness appropriate for library screening . However , we and others have questioned the relevance of these assays due to the poor translation of many axenic hit molecules into the physiologically more relevant but more complex intracellular Leishmania assays [10 , 18–22] . The poor confirmation rate of promastigote and axenic amastigote active compounds in intracellular assays could result from the many differences between the assays , including the localisation of intracellular parasites in the more difficult to access parasitophorous vacuole , differences in pH and composition of the growth media , stage-specific differences such as alternate energy pathway usage [16 , 23 , 24] , etc . Insights derived from our work with Trypanosoma brucei [25] point to another potential factor: many published axenic assays are likely to report not only cytocidal compounds but also growth slowing and cytostatic compounds . The latter two types of compounds are unlikely to show activity in intracellular assays which , due to the slow replication rate of the intracellular parasites and sensitivity of high content readers , tend to only report cytocidal molecules . In this paper we report the development and validation of a novel axenic amastigote Leishmania assay that reports only cytocidal molecules . This new assay was validated using a diverse compound library previously screened in our historic axenic format and in the intracellular assay [18] . We also used the new axenic assay as a primary screening platform to screen a diversity-oriented synthesis library [26] . We identified two new chemical series with antileishmanial activity and show that the new axenic assay has a significantly improved translation to the intracellular assay . We propose that the novel axenic assay provides Leishmania drug discovery efforts with an improved high-throughput platform for the screening of large compound libraries as it does not suffer from the high false positive rates seen in other axenic assays and provides both higher throughput and better robustness than intracellular assays . Leishmania donovani BOB cells ( LdBOB ) used in this study are a cloned line from strain MHOM/SD/62/1S-CL2 [13] . Cultures were maintained as described previously [18] . White , clear-bottom assay plates ( 384-wells ) were obtained from Greiner ( historic axenic assay ) and Corning ( novel axenic assay ) . Echo plates were obtained from LabCyte . Amphotericin B , Dimethyl Sulfoxide ( anhydrous , ≥ 99 . 9% , DMSO ) and Resazurin were obtained from Sigma . The small diverse library used in this study contained 15 , 667 compounds , dissolved at 10 mM in DMSO and stored under low-oxygen and low-humidity conditions . The design of this diverse library is described elsewhere [27] . The Diversity-Orientated Synthesis ( DOS ) Informer Set ( 9 , 907 compounds ) is a subset of the Broad Institute’s diversity-oriented synthesis library , which comprises approximately 100 , 000 structurally diverse small molecules that combine the complexity of natural products and the efficiency of high-throughput synthesis [26 , 28 , 29] . These small molecules have a higher ratio of sp3-hybridized atoms and stereocentres relative to compounds found in conventional screening collections [30] . The compounds are structurally diverse , from >30 different individual libraries and >250 unique scaffolds . Moreover , where possible , all stereoisomers have been individually synthesized , providing rich stereo-structure-activity relationship ( SSAR ) data directly from primary screens and facilitating rapid prioritization and optimization of hit compounds . Compounds were dispensed into 384-well assay plates by acoustic dispensing ( LabCyte ECHO ) . For potency determinations , ten-point one in three dilution curves were generated , with a top concentration of 50μM . Potencies are reported as pEC50 ( -LOG ( EC50[M] ) ) . All data was processed using IDBS ActivityBase . Raw data was converted into percent inhibition through linear regression by setting the high inhibition control as 100% and the no inhibition control as 0% . For primary single concentration screening in the novel axenic assay we introduced a static control as 0% inhibition ( signal at time of assay start ) , which is explained in more detail in the results section . Potency plates were normalised to DMSO control ( 0% effect ) and 2μM amphotericin B control ( 100% effect ) . Quality control criteria for passing plates were as follows: robust z’ ≥ 0 . 5 , signal to background > 3 , % coefficient of variation for 0% inhibition controls < 15 . The formula used to calculate robust z’ is 1- ( ( 3 x ( 1 . 4826 x MAD [0% inhibition controls] + 1 . 4826 x MAD [100% inhibition controls] ) ) / ( Median [0% inhibition controls]—Median [100% inhibition controls] ) , with MAD the median absolute deviation . The formula used for signal to background is: Median [0% inhibition controls] / Median [100% inhibition controls] . Curve fitting was carried out using the following 4 parameter logistic equation: y = A + ( B - A ) / ( 1 + ( ( 10C ) / x ) D ) , where A =% inhibition at bottom , B =% inhibition at top , C = 50% effect concentration ( EC50 ) , D = slope , x = inhibitor concentration and y =% inhibition . For compounds with low activity and poor definition of the curve top , B was fixed to 100 . For the determination of the reference compound panel potency , all experiments were carried out with a minimum of three independent repeats . Serial dilutions of LdBOB cells with defined cell concentrations were made . Each of the resulting concentrations was dispensed into at least 24 wells of a 384-well plate ( 50 μl and 25 μl per well for determining the LoD of the historic and novel axenic assay respectively ) , the rest of the wells contained media only ( blank ) . The read-out for the historic axenic assay was performed by adding resazurin at 0 . 05 mM final concentration followed by incubation for 4 h at 37°C and 5% CO2 . Fluorescence intensity was then measured using a Perkin Elmer Victor 3 plate-reader ( excitation 528 nm , emission 590 nm ) . The read-out for the novel axenic assay was carried out by adding BacTiter-Glo ( Promega ) ( volume equal to culture media volume ) to each well and the luminescence was immediately read in a Victor 3 plate-reader . A linear regression was fitted and the LoD was derived as the number of cells equal to the mean signal of the blank wells plus 3 times the standard deviation . The values were determined from 4 independent experiments LoDs are reported as LoD +/- Standard Deviation ( StDev ) . The limit of detection ( LoD ) of the resazurin readout is 4 , 390 ± 1 , 980 cells per well; N = 4 ( Fig 1A ) . The starting density ( 250 cells per well ) of the historic axenic assay is therefore 18-fold below the limit of detection ( modelled in Fig 1C ) . Consequently , compounds that either stop or inhibit growth below 4 , 390 cells within the 72 hour incubation period , but which do not kill the parasites , will look identical to truly cytocidal compounds . The limit of detection of the BacTiter-Glo readout is substantially lower at 190 ± 110 cells per well; N = 4 ( Fig 1B ) . While this is below the 250 cell starting density of the historic axenic assay , it does not give a sufficiently robust signal window to differentiate truly cytocidal compounds . To achieve this , we increased the starting cell density 80-fold to 20 , 000 cells/well . We also changed the data analysis method by modifying the normalisation procedure so that 0% inhibition reflects no growth of the parasites . To obtain the no growth control measurement , 20 , 000 cells were seeded in plates ( the starting assay density ) immediately before addition of the BacTiter-Glo reagent . Amphothericin B , a known fast and cidal acting compound , was used to define the 100% effect . The technology switch from resazurin to BacTiter-Glo , cell number increase and change in normalisation combine to produce a robust screening protocol with a signal window of 105-fold of the LoD . This large window allows clear differentiation between cytocidal and growth slowing / static compounds ( Fig 1D ) . A comparison of the main parameters that differ between both axenic assays is shown in Table 1 . During assay validation of the single concentration screening format , 155 384-well plates were screened . The assay proved to be robust with an average robust Z-factor of 0 . 88 ( ± 0 . 04 StDev ) and a Signal to Background ratio of 18 . 7 ( ± 4 . 6 StDev ) . Robustness was further confirmed by the absence of significant effects on assay performance when varying assay starting day , cell stock used or cell passage number used ( Fig A-C in S1 Text ) . Two independent replicates of a set of 35 compounds in potency mode demonstrated very good reproducibility ( R2 = 0 . 95 , Fig D in S1 Text ) ) . Amphotericin B was included as a control on each plate , and its potency was reproducible from run to run over an extended time period , further confirming the robustness of the assay ( average pEC50 of 6 . 9 ( ± 0 . 1 StDev ) ) . Overall the performance indicators support that the assay presented here is suitable for high-throughput screening . We previously screened a diverse compound set with around 16 , 000 compounds in the historic non-cidal axenic assay as well as in an intramacrophage assay [18] . The axenic assay identified 378 compounds that were active at 3μM ( 2 . 4% hit-rate ) , with only a relatively small proportion ( 83 compounds , 22% of the hits ) also active in the intracellular assay ( at 50μM ) . We reported that this high false-positive rate was the main impediment to using the axenic assay as a primary screening platform and proposed that a significant part of the high false-positive rate might be due to the identification of growth slowing and cytostatic compounds . To evaluate this , we re-screened this set of compounds in the novel axenic assay ( at 15μM ) and analysed the data using the intramacrophage assay as the gold standard ( at 50μM ) ( Fig 2 ) . The hit criteria and controls used for each assay are summarised in Table A in S1 Text . In contrast to the high hit-rate seen in the historic assay the novel assay only reported 138 hits ( 0 . 9% hit-rate ) , in spite of screening at a 3-fold higher concentration . Relative to the historic axenic assay , a much larger fraction of the novel axenic assay hits showed activity in the intracellular assay ( 49% of novel axenic hits are active in intracellular assay versus 22% for the historic assay ) . The analysis also showed that only 22% of the compounds active in the historic axenic assay were also active ( i . e . cidal ) in the novel assay , indicating that a large fraction of the compounds identified by the historic assay do not actually kill the parasites . A significant number of compounds ( 280 ) showed activity in the intracellular assay at 50μM but were inactive in the novel axenic assay screen at 15μM . However , a large fraction of these compounds ( 170 ) owe their intracellular assay activity to toxicity against the THP-1 host cells . As such , they should be considered false positives in the intracellular assay rather than false negatives in the novel cidal axenic assay . As a higher screening concentration was used in the intracellular assay we sought to determine whether this difference could account for the lack of axenic activity for the false negatives . To do so we determined the potency of 110 non-toxic , apparent false-negative compounds in the novel axenic and INMAC assays using the same top concentration ( 50 μM ) . A small number of compounds were inactive when tested in potency mode in the intracellular assay ( 13 ) which brings the tally for false positives to 183 . Just under half of the compounds tested ( 43 ) now showed activity in the novel axenic assay . The remaining false-negative compounds had weak activity in the intracellular assay with 51 compounds having a pEC50 < 5 ( EC50>10μM ) and 3 compounds with a pEC50 between 5 and 5 . 4 . We determined the potency of a panel of 36 selected control compounds in the novel axenic and intramacrophage assays ( Table B in S1 Text and Fig 3 ) . This panel includes drugs used in the field for the treatment of visceral leishmaniasis ( amphotericin B , miltefosine and paromomycin ) , representative compounds of chemical classes currently under preclinical or clinical development including nitroimidazoles and oxaboroles as well as a few earlier stage hits . The majority of these compound series show activity in animal models . In total , 21 compounds showed activity in both the novel axenic assay and the intracellular assay , and 7 compounds were inactive in both assays at the concentrations tested . With the exception of two compounds ( miltefosine and VL-2098 ) , the compounds that were active in both assays exhibited a good correlation between the two Leishmania assays ( R2 = 0 . 81 ) . Of the 8 compounds that were only active in one of the two assays most ( 5 ) were weakly active ( pEC50<5 ) . There were examples of compounds showing higher potency in the cidal axenic assay ( disulfiram and VL-2098 ) than the intracellular assay and vice versa ( miltefosine and amodiaquine ) . However , all these compounds would have been defined as active using the cidal axenic assay i . e . they would not have been missed . Our results show that the novel axenic assay is more suitable for high-throughput screening than both our historic axenic assay and the intracellular assay . Relative to the historic axenic assay it detects many less false positives , and relative to the intracellular assay it provides much higher throughput and robustness ( summarised in Table 2 ) . We therefore propose a new Leishmania hit-discovery screening cascade that uses the novel axenic assay as primary single point assay ( Fig 4 ) . In a semi-manual format , a single scientist can run 60 plates per week giving a capacity to screen 20 , 000 test compounds weekly . The throughput can be scaled further using additional automation or scientists . Hits from the primary screen may be generically toxic and should be followed up with potency determinations in both the novel axenic assay and a mammalian counterscreen ( we use HepG2 cells ) . Compounds with suitable activity and a selectivity window can then be taken forward to the intracellular assay for further confirmation of activity . In order to assess the utility of the screening cascade , a diversity-oriented synthesis library of 9 , 907 compounds was screened at 25 μM using the novel axenic assay . Compounds that showed a cytocidal effect ( i . e . less cells at assay end point compared to assay start point ) were retained as primary hits ( 72 compounds , hit-rate: 0 . 7% ) . The primary hits were next profiled using ten-point dose response curves in the novel axenic assay , in a HepG2 toxicity counterscreen and in the intracellular Leishmania assay ( screening cascade shown in Fig 5A ) . The confirmation rate in the axenic potency assay was good ( 62 compounds , or 87% of hits show cidality in retest ) and of the confirmed cidal compounds , 57 ( 92% ) showed at least some level of activity in the intracellular assay ( percent inhibition at top concentration >50 ) . The potencies obtained in the intracellular assay correlated reasonably well with the axenic data as shown in Fig 5B . Two chemical series in particular showed promising activity in the dose-response retests ( Fig 6 ) . An additional 56 related analogues from the DOS library were cherry-picked and tested in dose in the novel axenic , intracellular , and cytotoxicity assays . Stereochemical SAR ( SSAR ) analysis of all 8 possible isomers of BRD6650 ( Series 1 , Fig 6 Panels A and B ) showed that the RSS stereoisomer was the most potent ( intracellular pEC50 = 5 . 7 ) , followed by the SSR stereoisomer ( intracellular pEC50 = 5 ) . While the SAR around BRD6650 was limited to the compounds tested in the original screening collection , substitution on the phenyl group at the C-3 position of the azetidine ring was tolerated . Both electron-donating and electron-withdrawing groups were active , and diverse functionality was tolerated such as the aryl-alkyne ( BRD1184 ) , pyridyl ( BRD9157 ) , and cyclohexenyl ( BRD5744 ) groups . Minor variation on the urea functionality was also tolerated . SSAR analysis of Series 2 compounds ( e . g . BRD2647 , Panels C and D on Fig 6 ) , showed that the RSS stereoisomer was again the most potent stereoisomer ( intracellular pEC50 = 5 . 2 ) . The RRS stereoisomer was less potent; the SRR and SSR stereoisomers , which were not tested at dose , were inactive in the initial HTS assay . The other four stereoisomers for this compound were not available . The C-3 phenyl group of BRD2647 could be varied with a number of ortho substituents without loss of potency , and variation on the urea was also tolerated . The need to identify new candidates entering drug development for visceral leishmaniasis is high considering the low number of compounds currently progressing through the R&D pipeline for this disease [32] . In many disease areas phenotypic screening of large diversity-oriented compound collections has been successful at identifying novel active starting points [8 , 9 , 33 , 34] . For leishmaniasis , finding new start points is hampered by the lack of predictive , high-throughput compatible assays . When compounds are identified , they suffer from the intrinsically high attrition rate associated with early-stage discovery and lead-optimisation programmes [35] . It is therefore critically important to find new approaches to fill the leishmaniasis pipeline . The intramacrophage amastigote Leishmania model is currently considered the gold standard for in vitro drug susceptibility characterization [17 , 18 , 22 , 36 , 37] . Running this assay at a sufficiently high throughput to screen medium- or large-size compound collections is technically as well as financially challenging due to the complexity and limited robustness of this type of assay . While axenic parasites ( promastigotes or amastigotes ) can be used readily in high-throughput assays , the relatively poor translation into the intracellular assay limits their usefulness [17 , 18 , 21 , 38] . In this work we set out to develop a new Leishmania in vitro screening cascade that is suitable for high-throughput screening so that large compound collections can be accessed . As a primary screening platform we developed a new high-throughput axenic amastigote assay with significantly improved predictivity of intramacrophage activity . We have achieved this by altering the assay conditions so that only cytocidal compounds are identified . Our results show that the main reason for the poor correlation between the old axenic assay and our intracellular assay is assay setup related , rather than the result of differences in biology between the different stages ( ~ 80% of hits identified in our historic axenic assay were not cytocidal ) . While we cannot be sure this is the case for other published axenic assays as detection limits are rarely reported , we expect this may apply broadly , as a side-effect of using fast growing organisms in combination with relatively low sensitivity read-outs such as resazurin . In view of the low hit rate for Leishmania that we and others have found , there is a concern around potential false negatives . While these clearly exist , many are actually toxic to the host cell in the intracellular assay and can be considered intracellular assay false positives . We have shown that the root cause for approximately half of the non-toxic false-negative compounds that we tested is the difference in concentration that we used to screen in the axenic compared to the intracellular macrophage assay ( Fig 2 ) . As most of the remaining false-negatives are weak in the intracellular assay it is possible that they may show activity in the novel axenic assay when tested at higher concentrations . Interestingly , there are a number of compounds that are active in the novel axenic assay and not active in the intracellular assay . These may represent compounds that hit targets that are not essential intracellularly , perhaps as a result of differences in energy metabolism or supply of essential metabolites . However , our preliminary analysis suggests that these may cluster in a physicochemical space that may preclude their passage through cell membranes under physiological pH . Such compounds may represent novel start points for drug discovery , if their permeability can be improved while maintaining activity . Testing a panel of previously identified antileishmanials showed good correlation between the novel axenic assay and the intramacrophage assay for most compound classes . The aminoquinolines ( mefloquine & amodiaquine ) were only active in the intracellular assay , presumably due to their charge at acidic pH , resulting in lack of permeability in the novel axenic assay ( pH 5 . 5 ) and lysosomotropic accumulation in the intramacrophage assay [39] . Surprisingly , miltefosine was around ten-fold more active in the intracellular assay relative to the novel axenic assay . We and others have previously observed similar activity in axenic and intramacrophage models for miltefosine [17 , 18] . Miltefosine is actively taken up by Leishmania [40 , 41] , and therefore it is plausible that the potency difference is due to the much higher cell density in the novel axenic assay compared to our historic assay ( 80-fold higher ) , as this could result in lower miltefosine concentrations in the cells and hence lower potency . Disulfiram was only active in the novel axenic assay . Due to technical differences between the two assays there is a higher chance of compounds precipitating in the intracellular assay , and this could have been a problem with disulfiram as its aqueous solubility is poor and in addition this compound is unstable in serum [42] . The activity seen for VL-2098 in the novel axenic assay is in line with recently published data [43] , however the activity against intracellular amastigotes is lower than expected from the publication . A second sample of this compound was tested and gave the same results . A potential explanation for this discrepancy is the poor solubility of this compound [44] . Overall the results from this panel of compounds shows that there is a good correlation between our novel axenic assay and our intramacrophage assay , and that almost all compounds with proven antileishmanial activity show activity in the novel axenic assay , supporting its use as a primary screening assay . We propose that the novel axenic assay is the most suitable primary assay for a screening cascade aimed at accessing large compound collections as it combines the throughput and robustness of axenic assays with good predictivity of intramacrophage activity . Hits from the single-point screen should be tested in potency mode both in the axenic assay and in a human cell counterscreen to rule out any toxicity ( ideally targeting a 10-fold or higher selectivity window between the two assays ) . Finally , the active and selective compounds should be tested in the intracellular model ( cascade shown on Fig 4 ) . We validated the use of this cascade by screening a set of ~10 , 000 DOS compounds and identified two active chemical series of interest . This screen confirmed further that the novel axenic assay is a suitable primary screening platform as 92% of the confirmed actives also showed intracellular activity . The two most interesting hit series incorporate a densely functionalized azetidine ring system each containing three stereogenic elements . While azetidine-related systems such as β-lactams have played an important role in drug discovery , the fully reduced form such as those in our two series of interest have been significantly less studied . These libraries were prepared as part of a collection of skeletally diverse azetidine-based scaffolds . An important feature of this compound collection is that it contains stereochemical diversity and in our experience , Stereochemical Structure Activity ( SSAR ) analysis of a hit compound can give an indication on how selective and specific its interaction is with its molecular target [45] . While two stereoisomers of BRD6650 were active , the majority of the stereoisomers tested had at least a 9-fold drop in potency indicating a selective and specific interaction with a molecular target . The two series of interest differ in terms of their substituents at R1 and R2 and also the substitution pattern of the phenyl group at C3 which is para in series 1 but ortho in series 2 . Synthesis of additional analogues to further investigate SAR and profiling of the current leads in in vitro ADME/PK would allow us to assess the potential for these series for therapeutic development . In summary , we have developed and validated a novel axenic Leishmania assay that specifically identifies compounds with a cytocidal mode of action . This new assay has been profiled in terms of its ability to predict activity in the intracellular amastigote Leishmania stage which is currently considered the gold standard for in vitro drug screening for visceral leishmaniasis . We have demonstrated that this new assay is suitable for primary screening and provides a useful method of triaging compounds to significantly reduce the number of compounds to be profiled and confirmed for activity through the more complex intracellular assays . Additional new active hit series against Leishmania donovani are expected to be identified from several ongoing high-throughput screening campaigns incorporating the novel axenic assay as a primary screening tool . Overall , our findings show that assay conditions play a significant role in the nature of the compounds identified ( i . e . growth slowing vs cidal ) and demonstrate that appropriate axenic assays , particularly in the context of large-scale drug-discovery , can be relevant and of great value .
New drugs for visceral leishmaniasis , are urgently required as existing drugs have serious shortcomings including toxicity and drug resistance . This disease is caused by parasites from the Leishmania family which live inside human cells . Screening large collections of chemicals ( >100 , 000 ) to identify compounds that kill parasites has been used to identify new start points for drug discovery . It is complex and expensive to look at such numbers using intracellular parasites . To circumvent this , many groups screen using parasites adapted to grow outside human cells ( axenic forms ) . However , the established protocols identify growth slowing compounds as well as compounds that kill parasites . Cytocidal compounds are better start points for drug discovery . Here we present a screening cascade based on a modified axenic Leishmania assay adapted to identify compounds that kill parasites . We show that these compounds have a higher probability of being active against intracellular parasites . This new screening cascade was used to screen a compound collection and led to the identification of two new chemical series with antileishmanial activity . Their activity was confirmed against intracellular parasites . They are potential candidates for further drug development . The approach we have taken may have broad relevance to anti-infective and anti-parasitic drug discovery .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[]
2015
Development and Validation of a Novel Leishmania donovani Screening Cascade for High-Throughput Screening Using a Novel Axenic Assay with High Predictivity of Leishmanicidal Intracellular Activity
Inverted repeats capable of forming hairpin and cruciform structures present a threat to chromosomal integrity . They induce double strand breaks , which lead to gross chromosomal rearrangements , the hallmarks of cancers and hereditary diseases . Secondary structure formation at this motif has been proposed to be the driving force for the instability , albeit the mechanisms leading to the fragility are not well-understood . We carried out a genome-wide screen to uncover the genetic players that govern fragility of homologous and homeologous Alu quasi-palindromes in the yeast Saccharomyces cerevisiae . We found that depletion or lack of components of the DNA replication machinery , proteins involved in Fe-S cluster biogenesis , the replication-pausing checkpoint pathway , the telomere maintenance complex or the Sgs1-Top3-Rmi1 dissolvasome augment fragility at Alu-IRs . Rad51 , a component of the homologous recombination pathway , was found to be required for replication arrest and breakage at the repeats specifically in replication-deficient strains . These data demonstrate that Rad51 is required for the formation of breakage-prone secondary structures in situations when replication is compromised while another mechanism operates in DSB formation in replication-proficient strains . Long palindromic sequences ( inverted repeats ∼100 bp or more each without a spacer or with a short spacer ) present a threat to both prokaryotic and eukaryotic genome stability . In E . coli , long palindromes placed on plasmids are frequently excised and cause cell inviability when introduced to chromosome [1] . In yeast , they have been shown to drastically induce ectopic and allelic recombination and a variety of gross chromosomal rearrangements ( GCRs ) including deletions , translocations and gene amplification [2]–[9] . Long inverted repeats were demonstrated to undergo frequent deletions and induce gene conversion and intra-chromosomal recombination in mice [10]–[12] . Palindromic sequences have been found in the vicinity of chromosomal breakpoints of translocations in humans and are implicated in the pathogenesis of diseases . For example , palindromic AT-rich repeats ( PATRRs ) have been shown to induce both non-recurrent and recurrent translocations; the latter could result into Emanuel syndrome [13]–[18] . Palindrome-mediated large deletions and interchromosomal insertions are causative factors of several types of εγδβ thalassemia [19] and X-linked congenital hypertrichosis syndrome , respectively [20] . Also , palindromes are abundant in cancer cells and are associated with DNA amplification in colon and breast cancer , medulloblastoma and lymphoma [21]–[28] . Palindromic sequences can form hairpin and cruciform structures due to their intrinsic symmetry [1] . Formation of these aberrant structures has been considered to be responsible for the genetic instability associated with this sequence motif . Hairpins occurring on the lagging strand can interfere with DNA replication and be attacked by structure-specific nucleases leading to DSBs . In E . coli , hairpins formed during DNA replication at long palindromic repeats are cleaved by the SbcDC nuclease [29]–[33] . Similarly , in S . pombe , the nuclease activity of the Mre11/Rad50/Nbs1complex ( Mre11/Rad50 is the homolog of SbcDC ) was implicated in the generation of breaks at palindromes [8] , [34] . However , Casper et al . ( 2009 ) showed that in S . cerevisiae , the Mre11 complex is not involved in breakage at a large inverted repeat consisting of two Ty1 elements with a ∼280 bp spacer in strains where DNA polymerase α was down-regulated [35] . We previously demonstrated that in S . cerevisiae , the Mre11/Rad50/Xrs2 complex does not initiate DSBs at closely spaced Alu inverted repeats ( Alu-IRs ) but is required along with Sae2 for processing breaks that have hairpin termini [5] . This disparity in the Mre11 complex's effect on DSB generation at palindromic sequences might be attributed to the difference in the formation of stable hairpins during replication and the inability of this complex to cleave hairpins with large loops . This conjecture , however , remains to be experimentally proven . These observations also point out the existence of an Mre11-independent pathway in generating DSBs at palindromic sequences . We proposed that in yeast , Alu-IR-mediated hairpin-capped breaks can result from the resolution of cruciform structures in which a putative nuclease cleaves symmetrically at the base of the two hairpins [5] . Cruciform resolution on plasmid in yeast was shown to be dependent on the structure-specific endonuclease Mus81/Mms4 [36] , although chromosomal fragility at inverted repeats was not influenced by this complex [5] . Cruciform formation and resolution were also proposed to be the triggering events for translocations at PATRRs in human sperm cells [37]–[39] . Recently , in a plasmid transfection assay , the GEN1 nuclease was implicated in cruciform resolution in HEK293 cells , and the resultant hairpin-capped breaks were further processed by Artemis for DSB repair [40] . Whether this mechanism operates in PATRR-mediated chromosomal translocations remains to be established . Although the formation of hairpin and cruciform structures is deemed as the key initiation event for fragility at inverted repeats , the pathways that predispose eukaryotic cells to or provide protection against chromosomal breaks are still not well defined . Previously , deficiencies in Pol1 , Pol3 and Rad27 proteins responsible for synthesis of the lagging strand during DNA replication were found to augment instability at inverted repeats [3] , [7] , [9] . However , it is unknown if fragility is exclusively confined to deficiencies in lagging strand synthesis . In addition , it is important to identify mechanisms that facilitate or prevent instability of imperfect IRs that contain a spacer ( quasi-palindrome ) and are not fully homologous to each other , since these repeats prevail over perfect palindromes in the human genome [9] , [41] . In this study , we carried out an unbiased genome-wide screen aimed at identifying the genetic factors controlling fragility of homologous and divergent Alu-quasi-palindromes in yeast . Using 12 bp-spaced Alu-IRs with either 100% or 94% homology between the two repeats , we analyzed the effects of deletions of around 4800 non-essential genes and downregulation of 800 essential genes on quasi-palindrome-mediated GCRs . In addition to defects in lagging strand synthesis , we found that deficiencies in proteins involved in replication initiation and leading strand synthesis , replication pausing checkpoint pathway , the Sgs1-Top3-Rmi1 dissolvasome , proteins involved in Fe-S cluster biogenesis or telomere maintenance augment breakage and GCRs induced by Alu-IRs . Replication block and fragility at inverted repeats in replication-deficient strains were abrogated upon deletion of RAD51 , indicating an unexpected role for homologous recombination in the formation of cruciform structure at palindromic repeats when replication is compromised . We systematically analyzed the effect of more than 6000 mutations on Alu-IR-mediated fragility using a genome-wide screen in the yeast S . cerevisiae ( Figure 1 and Figure S1 ) . The screen's scheme is based on the approach developed in Tong et al . , 2001 [42] with modifications . In the query strains , a quasi-palindrome consisting of two 320 bp Alu elements in inverted orientation with a 12 bp spacer was placed telomere-distal to the counterselectable marker CAN1 on the left arm of chromosome V . The two Alu elements were either 100% or 94% homologous ( 100% Alu-IRs or 94% Alu-IRs ) . Breakage at the Alu-IRs and loss of the 40 kb telomere-proximal fragment results in canavanine-resistant colonies . The tester strains included a complete set of 4786 deletion mutations for non-essential genes ( YKO strains ) and two sets of 842 essential genes whose expression is either regulated by the doxycycline-repressible promoter ( yTHC strains ) or decreased due to mRNA perturbation ( DAmP strains ) . An hphMX cassette was positioned telomere-proximal to the Alu-IRs , providing a marker for selecting the presence of the repeats during the screening and the testers were marked by a kanMX cassette . The schematics for combining the left arm of chromosome V containing the fragile motifs and the mutations have been previously applied to study instability of the trinucleotide GAA/TTC repeats and are described in detail in Zhang et al . , 2012 [43] . Briefly , the query strains were crossed with each tester strain to get diploids , which then underwent sporulation . Haploids containing both the Alu-IRs and the mutation of interest were replica plated to canavanine-containing medium . Mutants with augmented repeat-induced GCRs exhibited increased number of canavanine-resistant papillae compared to the wild-type strains . Since the rate of canavanine-resistant colonies occurring due to GCR in the wild-type strain carrying 100% Alu-IRs is 10-fold higher ( 5×10−5 ) than in the strongest mutator Δmsh2 ( 6×10−6 ) , the screen specifically identified hyper-fragility mutants . We verified the effect of the identified mutants by recreating the hyper-fragile alleles in strains with the ADE2 gene inserted between CAN1 and Alu-IRs that allows differentiation of GCRs from mutations based on the color of canavanine-resistant clones [6] ( Figure 1 ) . To create the mutant alleles , the kanMX cassette was used to knockout non-essential genes and a tetO7 repressible promoter was used to replace the natural promoters of essential genes and regulate their expression [44] . The essential genes under the control of tetO7 promoter will be referred to as TET-ORFs in the following text . 38 mutants that exhibit a hyper-fragility phenotype in strains containing either 100% or 94% homologous Alu-IRs were identified from the screen ( Table S1 ) . 17 mutants belonged to the YKO collection , 17 mutants were uncovered from the yTHC collection and 4 mutants were identified from the DAmP collection . The mutants could be grouped into six classes of genes coding for the dissolvasome and proteins involved in replication , Fe-S cluster biogenesis , checkpoint response , telomere maintenance and DSB repair . Previously , it has been shown that downregulation of or mutation in the DNA polymerases α and δ causes increased instability of inverted repeats [3] , [5] , [7] . Consistently , we found that TET-POL1 and TET-POL3 strains destabilize both 100% and 94% Alu-IRs and exhibit 11- to 20-fold higher fragility than the wild-type strains . This screen also revealed that downregulation or deletion of other key components of the DNA replication pathway , namely , the origin recognition complex ORC , the DNA helicase Mcm2-7 , the DNA primase complex , the leading strand synthesis polymerase ε , the single-strand binding protein RPA , the polymerase sliding clamp PCNA , the clamp loader RFCs or the endonucleases Dna2 and Rad27 participating in Okazaki fragment maturation , induce fragility at Alu-IRs . Deficiencies in these proteins caused a 3- to 15-fold and a 3- to 34-fold increase in GCR rates for 100% Alu-IRs and 94% Alu-IRs , respectively . We also observed a 5- to 9-fold elevation of GCRs in strains carrying the defective replication checkpoint surveillance complex , Mrc1-Tof1-Csm3 . This result prompted us to test if Mec1 , which is recruited to stalled replication forks and phosphorylates Mrc1 in response to DNA replication stress [45] , [46] , senses inverted repeat-mediated replication impediment . Since Δmec1 is lethal , we assessed the effect of Δmec1 in Δsml1 background . We found that Δmec1Δsml1 but not Δsml1 led to a 5-fold increase in GCRs . These data demonstrate that intact replication machinery and replication checkpoint are required to prevent palindrome instability . Moreover , secondary structure formation and breakage are not only restricted to defects in lagging strand synthesis since fragility is also increased in strains where Polε and Mcm2-7 complex were downregulated . Besides the replication checkpoint surveillance mutants , the screen also revealed that GCRs mildly increase ( 2- to 4-fold ) in Δrad17 , Δmec3 , Δddc1 and Δrad24 mutants deficient in DNA damage checkpoint signaling [47] . As discussed below , this effect could be explained by the improved recovery of the broken chromosome when checkpoint activation is impaired . The third group of mutants that amplify Alu-IRs fragility included members of the cytosolic iron-sulfur protein assembly targeting complex . TET-YHR122W led to a 3- and 8-fold increase in GCRs in 100% and 94% Alu-IRs , respectively . Yhr122w was shown to physically interact with Cia1 and Mms19 in the biogenesis of Fe-S clusters in various DNA repair and replication proteins [48] , [49] . We found that disruption of MMS19 led to an 18- and 14-fold increase in GCRs in strains containing 100% and 94% Alu-IRs , respectively . This is also consistent with our previous finding that Δmms19 causes an increase in Alu-IR-induced homologous recombination [9] . The screen revealed that deletion of SGS1 , the RecQ helicase homolog implicated in the dissolution of branched DNA structures and unwinding of CTG/CAG hairpins [50] , [51] , caused a 10- and 7-fold elevation in GCRs in 100% and 94% repeats-containing strains . Sgs1 interacts with Rmi1 and Top3 to form the dissolvasome complex [52] . Consistently , we found that deletion of RMI1 and of YLR235C that partially overlaps with TOP3 also led to hyper-fragility ( Table 1 and Table S1 ) . Our data suggest potential roles of Sgs1-Rmi1-Top3 in influencing palindrome stability through unwinding the hairpin or cruciform structures formed by the repeats . The fifth group of hyper-fragile mutants consisted of TET-TEN1 , TET-STN1 and TET-CDC13 . The Ten1-Stn1-Cdc13 complex is involved in telomere maintenance and protection [53] . Downregulation of Ten1 resulted in a 3-fold elevation of fragility ( Table 1 ) . The TET-CDC13 strain demonstrated a similar increase in the level of arm loss . Notably , the closest telomere is about 40 kb away from the location of the inverted repeats . In another study , we found that downregulation of Ten1-Stn1-Cdc13 also predisposes the triplex-forming GAA/TTC repeats to breakage and expansions [43] . Taken together , these data suggest among other possibilities that this complex plays a role in helping replication machinery to move through difficult regions . Previously , we demonstrated that the Mre11-Rad50-Xrs2 complex and the Sae2 protein are required to open hairpins to initiate DSB repair at inverted repeats [5] . We also showed that in Δmre11 mutants , GCR rates increased likely due to the inability of mutants to hold DSB ends together and open the hairpin termini , which therefore increase the probability of formation of dicentric chromosomes [6] . Predictably , the screen identified Δmre11 and Δrad50 as hyper-fragile mutants with a 10- and ∼44-fold increase in GCRs induced by homologous and homeologous Alu-IRs , correspondingly . This group therefore encompasses mutants that do not impact secondary structure formation and breakage , but rather increase probability of arm loss and recovery of the broken chromosome . In the wild-type strain , DSBs induced by Alu-IRs have covalently-closed hairpin termini [5] . To determine if the nature of breaks in the identified hyper-GCR mutants was similar to the wild-type strain , we characterized DSB intermediates in a subset of mutants . In addition , estimation of the level of breaks provides a way to distinguish between mutants that facilitate formation or enhance stability of the secondary structures and mutants that increase the loss of the acentric DSB fragment ( e . g . mrx mutants ) or improve the recovery of the broken chromosome . We compared the levels of chromosomal breaks in the wild-type strain containing 100% Alu-IRs with a subset of mutants from each group described in the previous section ( Figure 2 and Figure S2 ) . DSB detection was carried out in Δsae2 strains to prevent the opening of the hairpins and the resection of the broken fragments [5] . The lethality of Δsgs1Δsae2 can be rescued by the deletion of HDF1 [54] . Therefore , the effect of Δsgs1 on DSB formation was assessed in the Δsgs1Δsae2Δhdf1 triple mutant . DSBs were analyzed with a telomere-distal probe upon AflII digestion or a telomere-proximal probe using BglII digestion of chromosomal DNA embedded in agarose plugs . Upon AflII or BglII digestion , DSBs occurring inside the repeats were expected to be 1 . 3 kb or 3 . 3 kb , respectively . We also anticipated the appearance of inverted dimers that are double the size of the DSB intermediates ( 2 . 6 kb or 6 . 6 kb , correspondingly ) . These molecules resulting from replication of hairpin-capped breaks were previously detected in the wild-type strains [5] . No DSBs were observed in the presence of Sae2 in both wild-type and mutant strains , likely due to hairpin opening and robust resection of the breaks . However , DSBs were readily detected in Δsae2 background . In TET-POL3 , TET-POL2 , Δcsm3 , Δsgs1Δhdf1 ( Figure 2 ) , Δmms19 , TET-TEN1 ( Figure S2 ) and Δsml1Δmec1 ( Figure S4 ) mutants , there was a 2- to 15- fold increase in breaks in comparison with wild-type strains when the telomere-proximal or the telomere-distal fragments were probed , indicating that these mutations increase fragility at Alu-IRs by either facilitating secondary structure formation or stabilizing the structures . It is important to note that no increase in DSBs were detected in the Δsae2Δhdf1 and Δsae2Δsml1 mutants ( Figure S3 and Figure S4 ) indicating that the increase in fragility is due to deficiencies in Sgs1 and Mec1 , accordingly . In Δrad17 , the amount of breaks was comparable to the wild-type strain , suggesting that DNA damage checkpoint-deficient mutants provide conditions for better recovery of the broken chromosomes , rather than affecting the formation and/or stability of the secondary structures . It is important to note that besides DSBs we could also detect dimers and no other intermediates were observed . The dependence of DSB detection on Δsae2 and the existence of dimers suggest that breaks in hyper-fragile mutants might contain hairpin termini similar to those in wild-type strains . To test the premise of hairpin-capped breaks in the mutants experimentally , the DSB fragments in TET-POL3Δsae2 were analyzed via neutral/alkaline two-dimensional ( 2D ) gel electrophoresis ( Figure 3 ) . We found that the 1 . 3 kb telomere-distal DSB fragment migrated as a 2 . 6 kb single-stranded DNA ( ssDNA ) fragment in the alkaline gel . Similarly , the 3 . 3 kb telomere-proximal DSB fragment migrated as a 6 . 6 kb ssDNA fragment under denaturing conditions . No additional bands ( e . g . those corresponding to nicked hairpins ) were seen , indicating that Alu-IRs generate covalently-closed hairpin-capped breaks in both TET-POL3 and wild-type strains . The symmetry of the breaks and the presence of covalently-closed hairpins at the DSB termini suggest that the final steps in breakage in mutants and wild-type are the same and include cruciform formation and resolution . The screen revealed that mutants deficient in the DNA replication pathway comprise the major group that augments fragility at Alu-IRs . Analysis of DSB intermediates indicated that cruciform resolution is the likely scenario for fragility in these mutants ( Figure 2 , 3 ) . Generation of ssDNA due to replication defects in the leading or lagging strands might provide optimal conditions for the formation of hairpins , not cruciforms . We hypothesized that a deficiency in the DNA replication can lead to formation of the cruciform structure through template switching when the fork stalls at a hairpin . In another screen for factors that channel replication stress into fragility , we identified Rad51 , a key protein in homologous recombination . In the Δrad51 background , the GCR rates of both TET-POL3 and TET-POL2 mutants decreased almost to the wild-type level ( Table 2 ) . Consistent with the reduction in GCRs , the amount of DSBs and inverted dimers in TET-POL3Δsae2 and TET-POL2Δsae2 significantly decreased upon deletion of RAD51 . Notably , lack of Rad51 does not affect GCR rates or DSB formation in the wild-type strains ( Table 2 and Figure 4 ) . Consistently , deletion of RAD54 , the auxiliary protein for strand invasion during recombination , in wild-type and TET-POL3 mutant had a similar effect on fragility ( Table S2 and Figure S5 ) indicating that the involvement of homologous recombination in the induction of fragility is specific to conditions when replication is compromised . To gain better insight into the mechanism underlying Alu-IR-induced fragility , we monitored replication progression through 100% homologous repeats in the wild-type strain and the replication-deficient mutant TET-POL3 using 2D gel electrophoresis and Southern hybridization . While replication progression was not hampered at the quasi-palindrome in the wild-type strain , the TET-POL3 mutant demonstrated a robust fork arrest at the repeats . The fact that the replication block in TET-POL3 is completely removed upon deletion of RAD51 ( Figure 5 ) argues for Rad51-mediated template switching as the signal for replication pausing . These data , along with the observation that Δrad51 suppresses DSB formation in replication deficient strains , support the scenario where an attempt to bypass hairpin structures during compromised replication via Rad51-dependent template switching promotes the formation of cruciform structures behind the replication fork . These structures are further attacked by nucleases , resulting in DSBs ( Figure 6 ) . Although DSB formation in other hyper-fragile mutants in Δrad51 background was not analyzed , the fact that the GCR levels in these strains decreased as compared to their RAD51 counterparts strongly suggests that the same mechanism of break formation operates in these mutants ( Table S2 ) . Overall , these data reveal an important role of homologous recombination in promoting DSB formation at inverted repeats , specifically in replication-deficient mutants . Inverted repeat-induced GCRs can be augmented in mutants that either influence secondary structure metabolism or alter repair of the broken chromosome . Previous studies from our lab have demonstrated that Alu-IRs-induced DSBs have hairpin termini that are opened by the Mre11 complex and Sae2 to initiate resection [5] . Unprocessed hairpin-capped molecules lead to the formation of acentric and dicentric inverted chromosomes . Detailed analysis of GCR events showed that dicentric chromosomes are stabilized as a result of breakage in anaphase , followed by resection and repair preferentially via break-induced replication with non-homologous chromosomes . It is important to note that DSB resection that precedes the healing of the broken chromosome activates checkpoint signaling and is manifested as cells arrested in G2/M [6] . Previously , we found that GCR rates are elevated in mrx mutants . This increase is not due to frequent DSBs at Alu-IRs , but rather a result of more efficient formation of dicentric chromosomes and loss of the broken acentric fragments . Consistently , Δmre11 , Δrad50 , and Δxrs2 were identified in this genome-wide screen as hyper-fragile mutants ( Table 1 and Table S1 ) . Another group of mutants that do not increase breakage but amplify GCR rates are those defective in DNA damage checkpoint signaling ( Δrad17 , Δmec3 , Δddc1 , Δrad24 ) ( Table 1 , Table S1 and Figure S2 ) . It is conceivable that in the absence of checkpoint activation after dicentric breakage , the rate of resection is decreased [55] and the broken chromosomes are replicated and segregated together to the daughter cells for several generations [56] , [57] , which improves their chances for repair . The mutants identified in the screen that increase DSB formation and GCRs at Alu-IRs are deficient in DNA replication , replication-pausing checkpoint surveillance , Fe-S cluster biogenesis , telomere maintenance and protection , or the function of the Sgs1-Rmi1-Top3 dissolvasome . As discussed below , the impact of deficiencies in these different processes on fragility can be explained by an increase in the probability of formation or stability of secondary structures during replication . The screen revealed that depletion of the major components of the replication fork responsible for synthesis of both leading and lagging strands increases Alu-IR-induced fragility . Our results are consistent with previous findings that mutations in the DNA polymerases α and δ promote excision of IRs and IRs-induced recombination and rearrangements [2] , [3] , [35] , [38] , [58] . It is possible that deficiencies in the synthesis of either the leading or lagging strand can lead to the generation of extensive single-stranded regions , thereby creating ideal conditions for the formation of hairpin structure , the initial event in Alu-IRs fragility ( Figure 6 ) . In replication-proficient mutants mismatches strongly suppress the fragility potential of inverted repeats which should be expected if cruciform extrusion is the initial step in breakage . However , in replication-deficient strains where transient hairpin structure probably precedes cruciform formation , mismatches in the inverted repeats are expected to have a lower impact on the formation of hairpin structure due to the presence of single stranded regions . This might explain the higher relative increase in fragility at imperfect repeats in comparison with repeats without heterology , for example in TET-RFA2 , TET-POL2 and TET-PRI1 mutants ( Table 1 ) . Interestingly , downregulation of the helicase Mcm2-7 and ORC also led to hyper-fragility at the repeats . Although the MCM helicase is a part of the replication machinery , it travels ahead of the fork , therefore generation of ssDNA due to depletion of this helicase is unlikely . The effect of deficiencies in MCMs and ORC on fragility might be the consequence of the inability of the closest origin ( ARS507 ) to fire since amounts of both protein complexes are important for regulating the timing of origin activation [59] . Replication forks traveling longer distances from the remote origins might be less processive and more prone to collapse upon encountering replication barriers . Downregulation of MCMs and ORC also increases instability at another fragile motif in yeast , the triplex-forming GAA/TTC repeats [43] , indicating that this phenomenon might be universal in situations when the replication fork passes through difficult regions . Consistent with this assertion , in human cell lines that have different replication landscapes , fragility at FRA3B and FRA16D sites depends on the distance the replication fork travels [60] . Alternatively , increased fragility in mutants for MCMs and ORC might be due to the assembly of a hampered replisome that lacks components required for leading or lagging strand synthesis . Deletion of MMS19 and downregulation of YHR122W , genes encoding proteins involved in Fe-S cluster biogenesis [48] , [49] , were also found to induce hyper-fragility at Alu-IRs . Recently , it has been shown that Mms19 and Yhr122W along with Cia1 , are required for the transfer of Fe-S clusters to various proteins including polymerase δ DNA primase and Dna2 [48] , [49] , deficiencies in which were identified to augment fragility in the screen . The presence of the Fe-S clusters in the polymerases α and ε [61] and the fact that these proteins interact with Mms19 [48] also makes them likely substrates for the CIA targeting complex . The effect of mutation in this pathway on Alu-IRs-mediated fragility can therefore be attributed to the impaired maturation of the DNA replication machinery . The deficiencies described above are expected to create optimal conditions for the formation of a hairpin that impedes replication progression . The hairpin might be formed at lower frequencies in replication-proficient cells as well . In both replication-proficient and -deficient strains , the secondary structure or the arrested fork might trigger the activation of checkpoint response required to recruit proteins to remove the hairpin and promote replication restart ( Figure 6 ) . The fact that deficiency of Mec1 and the Mrc1-Tof1-Csm3 complex leads to hyper-fragility implicates these proteins as possible sensors of secondary-structure-imposed replication arrest . However , the Mrc1-Tof1-Csm3 complex is also required to coordinate the Mcm2-7 helicase and DNA polymerase activities [53] , [62]–[64] , therefore , we cannot completely rule out that deficiencies in this complex affect the integrity of the replisome as well . It seemed reasonable to suggest the existence of helicases recruited to remove hairpins at the arrested fork . Indeed , the screen identified the Sgs1-Top3-Rmi1dissolvasome . Although Δsgs1 does not affect the stability of short CAG/CTG repeats ( less than 25 repeats ) , it increases the contraction and fragility rate of long CAG/CTG repeats ( 70 repeats ) , indicating that longer hairpins might be better substrates for Sgs1 activity [50] , [65] . In addition , the Sgs1-Top3-Rmi1 complex is involved in the dissolution of double Holliday junctions [66] . Hence , it is probable that this complex also irons out long hairpins formed by Alu-IRs during replication . An interesting group of mutants that destabilize Alu-IRs include TET-TEN1 , TET-CDC13 , and TET-STN1 . The Cdc13-Stn1-Ten1 ( CST ) complex is involved in protection of chromosome ends , telomerase recruitment and telomere replication . Hyper-fragility at inverted repeats due to deficiencies in this complex can be explained by the sequestration of the Tof1-Mrc1-Csm3 complex from the replisome to the single-stranded regions at uncapped telomeres [67] , [68] . Alternatively , this complex which is structurally similar to RPA [69] may facilitate replication progression through the hairpin . Dewar and Lydall , ( 2012 ) proposed that in mammalian cells the CST complex which is distributed throughout the genome [70] , aside from its role in telomere metabolism , facilitates replication through difficult regions . Taking into account that downregulation of the CST complex also increases GAA/TTC-mediated fragility and expansions [43] and the physical interaction of this complex with Polα [71] , [72] , it is reasonable to suggest that the role of CST in DNA replication might be evolutionarily conserved . In wild-type strains carrying inverted repeats , the deduced mechanism of breakage is cruciform-resolution by a putative nuclease that cuts symmetrically at the base of the two hairpins . This generates two hairpin-capped molecules that are present in equimolar ratios [5] . Since replication is a polar process , in replication-deficient strains , a nuclease attack on the accumulated hairpins or stalled replication fork would be expected to produce DSB intermediates different from those induced in the wild-type strains . Anticipated intermediates would include nicked hairpins , branched structures , or asymmetrical hairpin-capped breaks . Somewhat unexpectedly , we found that in the TET-POL3 strain in the absence of Sae2 , the DSB intermediates were structurally identical to the replication-proficient strains: only covalently-closed hairpin-capped breaks and inverted dimers resulting from replication of the DSBs were detected . Accumulation of hairpin-capped intermediates on both sides of the break indicates that cruciform-resolution is the predominant pathway for fragility under replication stress . Since deletion of SAE2 leads to stabilization of hairpin-capped breaks in all mutants analyzed , we propose that this mechanism operates not only in the TET-POL3 strain , but also in other hyper-GCR mutants identified in the screen . Based on our finding that deletion of RAD51 or RAD54 strongly decreases GCRs and breaks in replication-deficient strains ( Table 2 , Figure 4 , Table S2 and Figure S5 ) , we proposed that cruciform formation and resolution can result from the action of the homologous recombination machinery on intermediates present at the stalled replication fork . Consistent with this conjecture , replication arrest observed in TET-POL3 was also dependent on Rad51 . We cannot completely rule out the possibility that homologous recombination proteins facilitate the hairpin formation . However , taking into account that Rad51 forms nucleoprotein filaments that are essential for the invasion step of homologous recombination and that Rad54 promotes strand exchange [73] , we favor the explanation that Rad51 along with other components of the homologous recombination machinery promotes template switching when the replication fork encounters the hairpin structure . Synthesis of the hairpin-forming sequence on the unperturbed strand and reannealing of this newly synthesized DNA might allow formation of a cruciform structure which is resolved by a putative nuclease to give rise to hairpin-capped DSBs ( Figure 6 ) . In this case , the replication stalling observed in TET-POL3 would reflect the accumulation of arrested forks in response to template switching rather than inhibition of DNA synthesis by the hairpin structure . Rad51 was found to be present at unperturbed and stalled replication forks [74]–[77] , and the involvement of recombination proteins in the fork restart and bypass of DNA lesions via template switching has been demonstrated in several studies [78]–[84] . Here , we show that the attempt of homologous recombination proteins to bypass the secondary-structure barrier may be detrimental and culminate in breaks and GCRs . We also observed that in the TET-POL2 mutant there was a stronger Rad51-dependent increase in dimer formation than in the accumulation of DSBs indicating that at least in the situation where the leading strand synthesis is compromised template switch might culminate in the inversion of the replication fork . This pathway previously described in S . pombe [80] , [81] , that does not generate DSBs might operate in parallel with the cruciform resolution pathway . It is important to note that the Rad51 effect is specific in situations where replication is compromised . In replication-proficient strains , breaks and GCRs are not affected by Rad51 status , indicating that another mechanism for cruciform-formation exists . It is possible that in wild-type strains a homologous recombination-independent template switching mechanism leading to fragility operates , or that the cruciform formation is unrelated to replication . The latter hypothesis is supported by our recent finding that hairpin-capped breaks in the wild-type strain preferentially occur in G2 phase of the cell cycle ( Sheng et al . , in preparation ) . Based on this study , we propose that in the human population , the carriers of hypomorphic alleles for the BLM-hTOPOIIIα-hRMI1-hRMI2 dissolvasome and proteins involved in DNA replication , replication-pausing checkpoint surveillance , Fe-S cluster biogenesis , telomere maintenance and protection might be susceptible to inverted repeat-induced breaks and carcinogenic GCRs . Importantly , the status of these proteins determines the stability of imperfect repeats with a spacer and divergent arms that are present in the human genome [41] , [85] . At the same time , it is likely that homologous recombination can trigger chromosomal breakage at secondary structure-forming fragile sites and AT-rich palindromic sequences under conditions of replication stress . This detrimental role of homologous recombination in promoting chromosomal instability might contribute towards the development of diseases associated with fragile motifs . Homologous recombination-mediated chromosomal breakage and rearrangements might operate at secondary structure-forming fragile sites and AT-rich palindromic sequences under replication stress . This detrimental role of homologous recombination in promoting genome instability might contribute towards the development of diseases . yTHC , DAmP and YKO collections were purchased from Open Biosystems . All other strains in this study are derivatives of BY4742 ( Open Biosystems ) . The genotype of the query strains for the screen is: MATα , Δura3 , Δleu2 , Δhis3 , Δlys2 , rpl28-Q38K , Δ mfa1::MFA1pr-HIS3 , V34205::lys2::Alu-IRs , V29617::hphMX . The 100% or 94% homologous inverted Alus were inserted into the LYS2 gene via the pop-in and pop-out method as previously described [9] . The detailed construction of the query strain can be found in Zhang et al . , 2012 [43] . The effect of mutant alleles identified from the screen was verified in derivatives of YKL36 that carries the GCR assay and has the following genotype: MATa , Δbar1 , Δtrp1 , Δhis3 , Δura3 , Δleu2 , Δade2 , Δlys2 , V34205::ADE2 , lys2::Alu-IRs . To create the mutant strains , in the case of non-essential genes , the target gene was disrupted by the kanMX4 cassette [86]; in the case of essential genes , the repressible tetO7 promoter construct was PCR-amplified [44] from pCM225 ( Euroscarf ) and was used to replace the natural promoter of the gene to create the TET-alleles ( Table S3 ) . In strains used for DSB analysis , SAE2 was disrupted by TRP1 . For construction of the Δsgs1Δhdf1Δsae2 triple mutant , SGS1 was disrupted by the kanMX4 cassette , and HDF1 was knocked out by the hphMX cassette [87] ( Table S3 ) . To study the effect of RAD51 on Alu-IRs-mediated fragility , RAD51 was replaced by a hisG-URA3-hisG cassette [88] . The screen was carried out as described in Zhang et al . , 2012 [43] . Yeast cells were grown on YPD plates for 3 days . For each strain , a minimum of 14 independent colonies were taken to perform fluctuation test to estimate GCR rates . Appropriate dilutions of cells were plated on YPD and canavanine-containing plates to determine the GCR frequency . The GCR rates were calculated using the formula μ = f/ln ( Nμ ) as described in Drake , 1991 [89] . 95% confidence intervals were calculated as described in Dixon , 1969 [90] . The canavanine-containing plates used for tests were made from arginine-drop out medium with low amount of adenine ( 5 mg/L ) and supplemented with L-canavanine ( 60 mg/L ) . Yeast cells from overnight cultures were embedded into 0 . 8% low-melting agarose plugs at a concentration of 24×108 cells/ml . The plugs were treated with 1 . 5 mg/ml lyticase for 3 hr , followed by overnight 1 mg/ml proteinase K treatment . For restriction digestion of the DNA , the plugs were washed twice with 1 X TE buffer ( 10 mM Tris-Cl [pH 8 . 0] , 0 . 1 mM EDTA ) for 30 min , treated with 1 mM PMSF for 1 hr , washed with distilled water for 1 hr and equilibrated with restriction buffer for 20 min . Each plug ( ∼40 µl ) was digested with 50 units of AflII or BglII for 16 hr . Digested plugs were loaded in a 1% ( AflII digestion ) or 0 . 7% ( BglII digestion ) agarose gel , respectively , and run in 1 X TBE for 18 hr . The gels were treated with 0 . 25 N HCl for 20 min , alkaline buffer ( 1 . 5 M NaCl , 0 . 5 M NaOH ) for 30 min and neutralization buffer ( 1 . 5 M NaCl , 1 M Tris [pH 7 . 5] ) for 30 min . The gels were then transferred in 10 X SSC to charged nylon membrane for 2 hr through a Posiblotter ( Stratagene ) . Southern hybridization was carried out using P32-labeled LYS2-specific probes at 67°C overnight . DNA membranes were washed twice for 15 min each in buffer containing 0 . 1% SDS and 1% SSC and the signals were detected by the typhoon phosphoimager ( GE Healthcare Life Sciences ) . The hybridization signals were quantified using ImageJ software ( NIH ) . Yeast plugs were prepared and digested as described above . Neutral/neutral and neutral/alkaline gel analysis was performed as previously described with small modifications [5] , [91] . In the first dimensional gel electrophoresis , the plugs were loaded in a 1% ( AflII digestion ) or 0 . 7% agarose ( BglII digestion ) gel , respectively , and run for 18 hr in 1 X TBE . The gel slices containing the bands of interest were then cut out for the second dimensional gel electrophoresis . For neutral/neutral gel analysis , the gel slices were loaded in 1% ( AflII digestion ) or 0 . 7% ( BglII digestion ) agarose gel made in 1 X TBE , run in 1 X TBE for 18 hr at 1 . 7 V/cm and then processed for Southern hybridization . For neutral/alkaline gel , the gel slices were treated with 10 mM EDTA for 30 min , 5 mM EDTA for 30 min and embedded in agarose gel made in buffer containing 50 mM NaCl , 1 mM EDTA . Next , the gels were soaked in 5 X alkaline buffer for 30 min , 1 X alkaline buffer ( 50 mM NaOH , 1 mM EDTA ) for 30 min and cooled down in 1 X alkaline buffer at 4°C for 15 min . The gels were then run in 1 X alkaline buffer at 0 . 7 V/cm for 40 hr at 4°C and processed for Southern hybridization . 2D gel analysis was carried out as previously described in Brewer and Fangman , 1987 [92] . Overnight yeast cultures were synchronized in G1 with alpha factor ( 50 µg/107 cells ) at OD600 = 0 . 8 . 2 µg/ml doxycycline was added to the cultures to downregulate Polδ in the case of TET-POL3 and TET-POL3Δrad51 strains . Cells were then released into fresh YPD . 50 min after release , wild-type , TET-POL3 , TET-POL3Δrad51 strains were harvested and their genomic DNA samples were prepared as described in Friedman and Brewer , 1995 [93] . For the first dimensional gel electrophoresis , AflII digested DNA samples were loaded in a 0 . 4% agarose gel and run in 1 X TBE at 1 . 7 V/cm for 22 hr . For the second dimensional gel electrophoresis , gel slices containing bands of interest were cut out and loaded into a 1 . 2% agarose gel supplemented with 0 . 3 mg/ml ethidium bromide . The gels were run in 1 X TBE at 6 V/cm for 11 hr . Gels were then processed for Southern hybridization . Images were quantified using ImageQuant TL software ( GE Healthcare Life Sciences ) .
Inverted repeats are found in many eukaryotic genomes including humans . They have a potential to cause chromosomal breakage and rearrangements that contribute to genome polymorphism and the development of diseases . Instability of inverted repeats is accounted for by their propensity to adopt DNA secondary structures that is negatively affected by the distance between the repeats and level of sequence divergence . However , the genetic factors that promote the abnormal structure formation or affect the ability of the repeats to break are largely unknown . Here , using a genome-wide screen we identified 38 mutants that destabilize imperfect human inverted Alu repeats and predispose them to breakage . The proteins that are required to maintain repeat stability belong to the core of the DNA replication machinery and to the accessory proteins that help replication fork to move through the difficult templates . Remarkably , when replication machinery is compromised , the proteins involved in homologous recombination promote the formation of secondary structures and replication block thereby triggering breakage at the inverted repeats . These results reveal a powerful pathway for the destabilization of chromosomes containing inverted repeats that requires the activity of homologous recombination .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[]
2013
Genome-Wide Screen Reveals Replication Pathway for Quasi-Palindrome Fragility Dependent on Homologous Recombination
Auxin is a major developmental regulator in plants and the acquisition of a transcriptional response to auxin likely contributed to developmental innovations at the time of water-to-land transition . Auxin Response Factors ( ARFs ) Transcription Factors ( TFs ) that mediate auxin-dependent transcriptional changes are divided into A , B and C evolutive classes in land plants . The origin and nature of the first ARF proteins in algae is still debated . Here , we identify the most ‘ancient’ ARF homologue to date in the early divergent charophyte algae Chlorokybus atmophyticus , CaARF . Structural modelling combined with biochemical studies showed that CaARF already shares many features with modern ARFs: it is capable of oligomerization , interacts with the TOPLESS co-repressor and specifically binds Auxin Response Elements as dimer . In addition , CaARF possesses a DNA-binding specificity that differs from class A and B ARFs and that was maintained in class C ARF along plants evolution . Phylogenetic evidence together with CaARF biochemical properties indicate that the different classes of ARFs likely arose from an ancestral proto-ARF protein with class C-like features . The foundation of auxin signalling would have thus happened from a pre-existing hormone-independent transcriptional regulation together with the emergence of a functional hormone perception complex . Charophytes diverged from chlorophyte algae more than a billion years ago ( y . a . ) and led to land plants emergence around 450 million y . a . [2–6] . “Early divergent” clades display a range of body complexity going from unicellular algae in Mesostigmatophyceae and Chlorokybophyceae , to multicellular filaments in Klebsormidiophyceae ( Fig 1; S1 Fig ) [7 , 8] . “Late divergent” clades include Charophyceae and Coleochaetophyceae that share features with land plants ( S1 Fig ) , [9 , 10] but also Zygnematophyceae , that despite their simple structure are considered sisters to land plants according to recent phylogenetic studies [11 , 12] . Given the importance of the phytohormone auxin in plant development , the acquisition of its signalling pathway allowing for auxin-dependent changes in transcription is thought to have been a milestone in the water-to-land transition [2] . In land plants , this signalling pathway , called the Nuclear Auxin Pathway ( NAP ) , relies on three main protein families: TIR1/AFB ( Transport Inhibitor Response 1/Auxin Signalling F-box ) co-receptors , Aux/IAA transcriptional repressors ( Auxin/Indole-3-Acetic Acid Protein ) and ARF ( Auxin Response Factors ) Transcription Factors ( TFs ) [13 , 14] . ARFs have been classified into three evolutive classes , A , B and C . Class A includes activator ARFs whereas classes B and C contain repressor ARFs [1] . ARF interaction with DNA is mediated by their B3 domain ( B3ARF ) . Such domain is also present in ABI3 ( Abscisic Acid insensitive 3 , B3ABI3 ) and RAV ( Related to ABI/VP1 , B3RAV ) plant TFs but with different DNA binding specificities [15 , 16] . In the ARF family , the B3 domain is embedded in a larger N-terminal DNA Binding Domain ( DBD ) that includes a Dimerization Domain ( DD ) . As dimers , ARFs bind double AuxREs ( Auxin Response Elements ) sites arranged in three possible orientations: Direct Repeat ( DR ) , Everted Repeat ( ER ) and Inverted Repeat ( IR ) ( S2 Fig ) [2 , 17–19] . In charophyte algae and the bryophyte Marchantia polymorpha , the B3RAV and B3ARF domains are often associated with a C-terminal PB1 oligomerization domain , a landmark of most ARF TFs in higher plants but that was lost by RAV TFs from tracheophytes [2 , 20] . This shared B3 + PB1 domain composition led to the initial hypothesis that ARF could have arisen from RAV [21] . In the NAP , the PB1 domain mediates ARF homo-oligomerization and hetero-oligomerization with Aux/IAAs [22] . Under low auxin concentrations , Aux/IAAs bind activator ARFs through their PB1 domain [23–26] and recruit TOPLESS ( TPL ) /TOPLESS-RELATED co-repressors , leading to the formation of a repressor complex on regulatory sequences of auxin-responsive genes [27] . When auxin levels increase , the hormone-mediated interaction between Aux/IAA and TIR1/AFB leads to Aux/IAA proteasomal degradation , unlocking activator ARFs and inducing transcription [28 , 29] . Most class B and C ARF members have limited interaction capacities with Aux/IAAs [30–32] and are proposed to regulate auxin transcriptional responses in an auxin-independent manner , possibly by competitive binding with class A ARFs on DNA regulatory sequences [33 , 34] . Since some of class B and C ARFs can interact directly with TPL , formation of co-repressor complexes was proposed as another possible mechanism for transcriptional repression of auxin target genes [34–36] . The NAP was established at the beginning of land plants history . In the bryophyte M . polymorpha for example , the 3 families of NAP proteins are present ( with one member of each ARF class ) as well as the TPL co-repressor [37 , 38] . Recent studies showed the existence of two ARF subfamilies in charophytes , class C and class A/B [20] , but the absence of functional TIR1/AFB and Aux/IAAs suggested that a fully functional NAP did not exist before land plants [2 , 20 , 37 , 39–41] . How these ancestral components evolved to form the land plants NAP remains an open question . Through the structural , biochemical and phylogenetic characterisation of a proto-ARF from an early divergent charophyte we set a scenario of how the co-option of ancestral mechanisms of transcriptional control possibly led to the evolution of hormone signalling pathways in plants . To understand the evolution of ARFs , we first characterized the biochemical properties of proto-ARFs and closely related proto-RAVs from early divergent charophytes . We searched for B3 homologues in charophyte transcripts databases ( OneKp and Marchantia . info ) [12 , 44] and classified them as B3RAV or B3ARF , depending on the residues signature of their predicted DBDs ( S1 Table ) [45] . B3RAV domains were frequently associated with an APETALA2 ( AP2 ) domain and/or PB1 domains in the basal charophyte Mesostigma viride and all later clades ( Fig 1; S2 Table ) . M . viride also has an ARF homologue ( GBSK01006108 . 1 ) devoid of a PB1 domain [2] . Its DBD was reliably modelled as an ARF ( 100% confidence with AtARF1 [46 , 47] ) , but it lacks most residues involved in the interaction with AuxREs ( S3 Fig ) and thus does not qualify as a functional ARF . The proto-ARF of the earliest diverging green Charophyte algae with predicted functional B3ARF and PB1 domains was found in C . atmophyticus . Other ARF homologues were also present in all later diverging clades ( Fig 1; S3 Fig ) . We determined the properties of “ancestral” RAV and ARF proteins , focusing on K . nitens proto-RAV ( containing predicted AP2 , B3RAV and PB1 ) ( KnRAV , kfl00094_0070 ) and C . atmophyticus proto-ARF ( CaARF , AZZW-2021616 ) . The predicted B3 domains of KnRAV and CaARF display the signature residues typical of B3RAV and B3ARF , respectively ( S1 Table; S3 and S4 Figs ) suggesting that their divergent DNA binding specificities were already established in charophytes . To test this hypothesis , we characterized the binding of their DBD against the canonical DNA binding sites identified in angiosperms for ABI3 , RAV and ARF TFs . KnRAV specifically bound the AP2/B3RAV bipartite element described for Arabidopsis thaliana RAV TFs ( Fig 2A ) [48] . CaARF interacted strongly with double AuxRE sites ( DR or ER , Fig 2B ) but not with a single AuxRE site suggesting that the DBD of CaARF binds DR and ER motifs as a dimer without the help of the Middle Region ( MR ) and the PB1 domain . Altogether , these results confirm that RAV and ARF DNA binding preferences were established in basal charophytes and maintained along evolution . Next , we studied the oligomerization capacity of their PB1 domain . Based on AtARF5 PB1 structure [23 , 47] , the PB1 domains of KnRAV and CaARF were modelled as type I/II PB1 with electrostatic oligomerization potential ( Fig 2C and 2D ) . Molecular weight determination of KnRAV-PB1 and CaARF-PB1 by Size Exclusion Chromatography combined with Multi-Angle-Light Scattering ( SEC-MALLS ) experimentally validated that both domains form oligomeric complexes ( Fig 2E and 2F ) but with a lower oligomeric potential than AtARF5-PB1 ( S3 Table ) . Charophycean algae therefore appear to possess proto-RAV and proto-ARF proteins with oligomerization potential and diverging DNA binding specificities ( Fig 2A and 2B; S3 and S4 Figs ) . To further characterize the biophysical properties of proto-ARFs , we determined the predicted structure of CaARF DBD and showed that it was reliably modeled ( 99% confidence; Phyre 2 ) with AtARF1 and AtARF5 DBDs [46 , 47] except for an additional disordered region in CaARF present within the DD ( Fig 3A ) . Similar disordered regions were found as a characteristic feature of all class C ARFs ( Fig 3B and 3C; S3 and S7 Figs ) . In agreement with this , our phylogenetic studies position CaARF within clade C ( S5 Fig ) . Such insertions are expected to modify class C DNA binding compared to A and B ARFs . We tested this hypothesis using ER motifs with different spacing ( ER4-9 ) . Unlike Arabidopsis AtARF2 ( class B ) and AtARF5 ( class A ) that largely prefer ER7/8 motifs ( Fig 3D and 3E ) , CaARF showed promiscuous binding to ER4-9 but did not interact with a single AuxRE motif ( Fig 3D–3F ) confirming its interaction with ER motifs as a dimer . Arabidopsis class C AtARF10 behaves similarly to CaARF ( Fig 3G ) . This shows that CaARF has a relaxed DNA specificity allowing binding to ER binding sites with various distances between the monomeric sites and that this specificity was maintained in class C ARF along plants evolution . The presence of a specific disordered region ( Fig 3A–3C; S3 and S7 Figs ) in class C ARF DBDs suggests a possible role in their relaxed specificity , that remains to be tested . As mentioned before , certain land plants ARF proteins have the capacity to interact directly or indirectly with the TPL co-repressor [35 , 36 , 50] . We wondered when in evolution this interaction was first established . Direct TPL-recruitment usually involves two different amino acid regions in the Middle Region ( MR ) of repressor ARFs: the EAR-motif ( ERF-associated Amphiphilic Repression motif with LxLxL sequence or its LxLxPP variant ) and the BRD domain ( B3 Repression Domain with the K/RLFG sequence ) [35 , 36] , the BRD domain also being found in RAV proteins . CaARF-MR presents two potential repression regions with an EAR-like motif ( LPLLPS , similar to LxLxPP ) and a BRD domain ( KLFG ) . Since TPL EAR-interacting-region ( TPL N-terminal , TPL-N ) is extremely conserved between charophytes and land plants [49 , 51] ( Fig 3H; S8 Fig; S4 Table ) , we used A . thaliana TPL-N ( AtTPL202 ) to assay the TPL/CaARF interaction . CaARF interacted with AtTPL202 in co-purification assays and this interaction was lost with AtTPL202-F74A , mutated in the hydrophobic EAR peptide binding groove ( Fig 3I ) [49] . Moreover , mutations in CaARF KLFG ( CaARF-L523S/F524S ) or LPLLPS ( CaARF-L492A/L493A ) weakened the interaction with AtTPL202 , indicating that both sites might participate to TPL-N recruitment . The binding of the BRD domain of CaARF differs from that of the RAV1 of A . thaliana which interacts with the C-terminal part of TPL [52] , suggesting different TPL recruitment mechanisms for these two protein families . The presence of similar TPL-recruitment sequences in proto-ARFs of different charophytes clades ARFs ( S5 Table ) suggests that they might also interact with TPL . The present biochemical characterization of CaARF , a proto-ARF from an “early divergent” charophyte , identifies this protein as class C ARF , in agreement with our phylogenetic analyses ( S5 Fig ) . Mutte et al . ( 2018 ) proposed the existence of two ARF classes in “late divergent” charophytes , C and A/B , deriving from a common ancestor that diverged in an ancient charophyte clade [20] . Based on phylogenetic analyses showing that class C ARF is sister to classes A and B , and on the identification of a M . viride sequence classified as a class A/B , Flores-Sandoval et al . ( 2018 ) proposed a similar scenario where the divergence between classes A and B and class C occurred prior to the diversification of extant streptophytes [40] . This plausible scenario , built before the identification of class C ARFs in “early divergent” charophytes , is based on an unusual M . viride sequence that does not exhibit the conserved ARF DNA binding residues ( S3 Fig ) , and implies repeated loss of class A/B ARFs from Chlorokybophyceae to Coleochaetophyceae ( S6 Fig ) . Further identification of class C ARFs in the “early divergent” charophytes ( Klebsormidiophyceae [2] and Chlorokybophyceae ( this work ) ) and the presence of both classes C and A/B in the “late divergent” C . orbicularis suggest a second and more parsimonious scenario in which class A/B ARF members come from an ancestral proto-ARF , belonging to class C or class C-like that existed before the emergence of “late divergent” charophytes ( S6 Fig ) . This hypothesis implies only a few class C ARF gene losses in some Klebsormidiophyceae , Coleochaetophyceae and Zygnematophyceae species . Still , all these scenarios need to be taken with caution as they are based on transcriptomic datasets and could be challenged when genomic sequences become available . When comparing C and A/B clades we found a disordered region within the predicted DD of ancestral and land plants clade C ARFs that is not present in clade A/B neither in land plants clades A and B . We speculate that during the duplication event leading to A/B emergence from clade C , the loss of this disordered sequence occurred . The DNA interaction experiments presented in this manuscript suggest that this event might have contributed to the acquisition of a more restricted DNA specificity of class A and B ARFs for ER motifs . Apart from the similar behaviour observed for CaARF and AtARF10 when binding to DNA , ancestral clade C ARFs already presented PB1 oligomerization potential and interaction with the co-repressor TPL . The conservation of these properties along evolution is consistent with experiments conducted on Marchantia showing partial complementation of the loss of function MpARF3 by class C AtARF10 [40] . Moreover , these biochemical facts are instructive on several aspects of the evolution of the NAP in plants . First , proto-ARFs being able to interact with AuxREs supports that the NAP could have co-opted sets of genes already regulated by ARFs in charophytes , as suggested in other studies [2 , 20 , 39 , 40] . In this context , the emergence of the A/B clade with a different DNA binding behaviour could have allowed to target a more specific set of genes . Second , proto-ARF interaction with TPL provides functional evidence for a role for class C ARFs as transcriptional repressors . Putative TPL interaction motifs are also present in proto-RAV and most proto-ARFs across charophytes , which includes class A/B ARFs . The capacity to recruit TPL co-repressors could thus be an ancestral property of RAV and ARF TFs . From these observations , we propose ARFs recruitment of co-repressor complexes to AuxREs promoter elements as a primitive and conserved mechanism predating the NAP . The absence of a functional TIR/AFB-Aux/IAA co-receptor [2 , 20 , 41] indicates that this primitive system was auxin-independent . These observations are consistent with a series of experiments in Marchantia showing that auxin-responsive genes show similar transcriptional responses in WT and MpARF3 mutants [20 , 40] . Alongside the diversification of ARF DNA binding specificity , emergence of the auxin perception complex in the first land plants turned ARFs-regulated genes into auxin-responsive genes through ARF-Aux/IAA-TIR/AFB interactions evolution ( Fig 4 ) . Our work thus allowed proposing a scenario where the evolution of the binding specificity of an ancestral TF together with the emergence of a functional hormone perception complex create a hormone signalling pathway . This scenario offers a better understanding of how hormone signalling pathways can evolve from pre-existing mechanisms of transcriptional regulation independent of any hormone signalling . Potential homologs of the NAP components were searched by sequence homology to the corresponding NAP proteins from M . polymorpha . Blasts were done using different databases: OneKp , PlantTFDB and Marchantia . info . Due to the lack of proteomic data in charophyte organisms , we carried out tblastn . Each transcript was then translated using Expasy Translate tool . Sequences resulting from this search were classed using protein sequence alignments and phylogenetical studies . Protein sequences alignments were done with Multialin ( http://multalin . toulouse . inra . fr/multalin/ ) and ESPrit ( http://espript . ibcp . fr/ESPript/ESPript/ ) online tools . Phylogenetic analyses were conducted using predicted DBDs from charophyte proto-ARFs and DBDs belonging to A . thaliana and M . polymorpha ARFs . Phylogenies were done with MEGA and Phylogeny . fr software using Maximum likelihood algorithm . Protein structure modelling was done with Phyre2 online tool [47] . Three-dimensional structures were visualized with PyMOL software ( www . pymol . org ) . cDNA sequences coding for potential ancestors and the corresponding mutants were constructed as synthetic DNA ( Thermofisher ) . KnRAV and CaARF ( full-length , fragments ( CaARF-DBD ( residues 1–421 ) , CaARF-PB1 ( residues 644–750 ) , KnRAV-DBD ( residues 256–523 ) , KnRAV-PB1 ( residues 724–798 ) ) or mutants ) coding sequences were cloned into pETM40 plasmid ( EMBL ) that contains a MBP-tag in the N-terminal region except for PB1 domains from both proteins that were cloned into pETM11 ( EMBL ) that confers a N-terminal His-tag . KnRAV and CaARF specific domains were isolated by PCR from synthetic cDNA sequences ( S6 Table ) . Full-length ARF2 , ARF5 and ARF10 were cloned into pHMGWA vectors ( Addgene ) containing N-terminal His-MBP-His tags . All proteins were expressed in Escherichia coli BL21 strain . Bacteria cultures were grown with the appropriate antibiotics at 37°C until they achieved an OD600nm of 0 . 6–0 . 9 . Protein expression was induced with isopropyl-β-D-1-thyogalactopiranoside ( IPTG ) at a final concentration of 400 μM at 18°C overnight . Bacteria cultures were centrifuged , and the pellets were resuspended and sonicated in the buffers indicated in S7 Table . After centrifugation , soluble fractions of KnRAV , KnRAV-DBD , CaARF , CaARF-DBD and CaARF mutants were loaded on Dextrin-Sepharose ( GE Healthcare ) column previously equilibrated in buffer A ( S7 Table ) . After column washing , proteins were eluted in buffer A with maltose 10 mM ( S6 Table ) . PB1 domains of KnRAV and CaARF as well as full-length proteins ARF2 , ARF5 and ARF10 were purified on Nickel-Sepharose ( GE Healthcare ) columns previously equilibrated in the appropriate buffers ( S6 Table ) . After protein binding , columns were washed with 30 mM imidazole to remove all proteins non-specifically bound to the column . Proteins were eluted in the corresponding buffer containing 300 mM imidazole ( S6 Table ) . His-tags of PB1 domains were cleaved by TEV protease ( 5% w/w ) overnight at 4°C followed by incubation at 20°C for 2 h for SEC-MALLS experiments . AtTPL202 and mutants were purified as explained in Martin-Arevalillo et al . , 2017 [49] . Following purification step , all proteins were dialyzed for 15 h at 4°C in their purification buffers , frozen in liquid nitrogen and conserved at -80°C until used . DNA probes were artificially designed based on the DNA binding site for each TF ( S8 Table ) ( Eurofins ) . Oligonucleotides for the sense strand were designed with an overhanging G in 5’ that allows the labelling of the DNA ( S8 Table ) . Annealing of the oligonucleotides and Cy5-labelling of the probes were performed as described in Stigliani et al . , ( 2019 ) [19] . Electrophoretic Mobility Shift Assays ( EMSA ) , were done on native 2% agarose gels prepared with TBE buffer 0 . 5X . Gels were pre-run in TBE buffer 0 . 5X at 90 V for 90 min at 4°C . Protein-DNA mixes contained Salmon and Herring Sperm competitor DNA ( final concentration 0 . 07 mg/ml ) and labelled DNA ( final concentration 20 nM ) in the interaction buffer ( 20 mM HEPES pH 7 . 8; 50 mM KCl; 100 mM Tris-HCl pH 8 . 0 , 2 . 5% glycerol; 1 mM DTT ) . Mixes were incubated in darkness for 30 min at 4°C and next loaded in the gels . Gels were run for 1 hour at 90 V at 4°C in TBE 0 . 5X . DNA-protein interactions were visualized with Cy5-exposition filter ( Biorad ChemiDoc MP Imaging System ) . For protein-protein interaction analyses , complexes between potential interaction partners were first formed by mixing MBP-tagged CaARF ( wt or mutants ) ( 90 μg ) with His-tagged AtTPL202 ( and mutants ) ( 70 μg ) in CAPS 20 mM pH 9 . 6; Tris-HCl 100 mM pH 8; NaCl 50 mM; TCEP 1 mM buffer for 1 h at 4°C . Complexes formed were fixed through the MBP tag to Dextrin-Sepharose columns previously equilibrated with CAPS 20 mM pH 9 . 6; Tris-HCl 100 mM pH 8; NaCl 50 mM; TCEP 0 . 1 mM buffer . After incubation of the complexes with Dextrin-Sepharose for 30 min at 4°C , nonspecific interactions were removed by a washing step with the same buffer . Protein complexes were eluted with 200 μl of the same buffer containing 10 mM of maltose . MBP was used as control for unspecific interactions . The eluted fractions were analysed by SDS-page polyacrylamide gel electrophoresis 12% . Molecular weights were determined by Size-Exclusion Chromatography-Multi Angle Light Scattering ( SEC-MALLS ) on an analytical Superdex-S200 increase ( GE Healthcare ) connected to an in-line MALLS spectrometer ( DAWN HELEOS II , Wyatt Instruments ) . Analytical size exclusion chromatography was performed at 25°C at a rate of 0 . 5 mL/min for untagged PB1 domains resulting from TEV cleavage . Untagged KnRAV-PB1 MW determination was carried out in CAPS 100 mM pH 9 . 6; TCEP 1mM buffer , whereas Tris-HCl 20 mM pH 8; TCEP 1 mM was used for untagged CaARF-PB1 and AtARF5 . The refractive index measured with in-line refractive index detector ( Optirex , Wyatt Instruments ) was used to follow the differential refractive index relative to the solvent . Molecular masses calculation was done with the Debye model using ASTRA version 5 . 3 . 4 . 20 ( Wyatt Instruments ) and a theoretical dn/dc value of 0 . 185 mL/g .
Plants transition from water to land was determining for the history of our planet , since it led to atmospheric and soil condition changes that promoted the appearance of other life forms . This transition initiated around 1 billion years ago from a Charophyte algae lineage that acquired features allowing it to adapt to the very different terrestrial conditions . Land plants coordinate their development with external stimuli through signalling mechanisms triggered by plant hormones . Therefore , evolution of these molecules and their signalling pathways likely played an important role in the aquatic to terrestrial move . In this manuscript we study the origin of auxin signalling , a plant hormone implicated in all plant developmental steps . Our studies suggest that out of the three families of proteins originally proposed to trigger auxin signalling in land plants , only one existed in Charophyte ancestors as a likely transcriptional repressor independent of auxin . We show that despite millions of years of evolution , this family of proteins has conserved its biochemical and structural properties that are found today in land plants . The results presented here provide an insight on how hormone signalling pathways could have evolved by co-opting a pre-existing hormone-independent transcriptional regulatory mechanism .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "&", "methods" ]
[ "organismal", "evolution", "brassica", "dna-binding", "proteins", "hormones", "plant", "science", "model", "organisms", "plant", "hormones", "experimental", "organism", "systems", "sequence", "motif", "analysis", "plants", "research", "and", "analysis", "methods", "arabidopsis", "thaliana", "sequence", "analysis", "sequence", "alignment", "bioinformatics", "proteins", "animal", "studies", "biological", "databases", "algae", "biochemistry", "plant", "biochemistry", "eukaryota", "plant", "and", "algal", "models", "sequence", "databases", "plant", "evolution", "database", "and", "informatics", "methods", "biology", "and", "life", "sciences", "auxins", "evolutionary", "biology", "organisms" ]
2019
Evolution of the Auxin Response Factors from charophyte ancestors
The computation represented by a sensory neuron's response to stimuli is constructed from an array of physiological processes both belonging to that neuron and inherited from its inputs . Although many of these physiological processes are known to be nonlinear , linear approximations are commonly used to describe the stimulus selectivity of sensory neurons ( i . e . , linear receptive fields ) . Here we present an approach for modeling sensory processing , termed the Nonlinear Input Model ( NIM ) , which is based on the hypothesis that the dominant nonlinearities imposed by physiological mechanisms arise from rectification of a neuron's inputs . Incorporating such ‘upstream nonlinearities’ within the standard linear-nonlinear ( LN ) cascade modeling structure implicitly allows for the identification of multiple stimulus features driving a neuron's response , which become directly interpretable as either excitatory or inhibitory . Because its form is analogous to an integrate-and-fire neuron receiving excitatory and inhibitory inputs , model fitting can be guided by prior knowledge about the inputs to a given neuron , and elements of the resulting model can often result in specific physiological predictions . Furthermore , by providing an explicit probabilistic model with a relatively simple nonlinear structure , its parameters can be efficiently optimized and appropriately regularized . Parameter estimation is robust and efficient even with large numbers of model components and in the context of high-dimensional stimuli with complex statistical structure ( e . g . natural stimuli ) . We describe detailed methods for estimating the model parameters , and illustrate the advantages of the NIM using a range of example sensory neurons in the visual and auditory systems . We thus present a modeling framework that can capture a broad range of nonlinear response functions while providing physiologically interpretable descriptions of neural computation . Sensory perception in the visual and auditory systems involves the detection of elemental features such as luminance and sound intensity , and their subsequent processing into more abstract representations such as “objects” that comprise our perception . The neuronal computations performed during such sensory processing must be nonlinear in order to generate more complex stimulus selectivity , such as needed to encode the conjunction of multiple sensory features [1]–[3] as well as to develop invariance to irrelevant aspects of the raw sensory input [4] , [5] . While these computations can appear inscrutably complex , they are necessarily constructed from the underlying neural circuitry , which exhibits several well-known and relatively straightforward nonlinear properties . Nevertheless , characterizations of sensory neurons still typically rely on the assumption of linear stimulus processing , which is often implicit in standard approaches such as spike-triggered averaging and – more recently – generalized linear models ( GLMs ) [6]–[8] . While such descriptions can often provide good predictions of the neuronal response [9]–[11] , they necessarily leave out the nonlinear elements of neuronal processing that likely play a major role in building the sensory percept . Unfortunately , the space of possible nonlinear models is not bounded . While one might be inclined to incorporate details of the system and circuitry in question , more complicated models require more data for parameter estimation , and often involve poorly behaved or intractable optimization problems . As a result , practical nonlinear modeling approaches must make assumptions that limit the space of functions considered by restricting to a defined set of nonlinear interactions . Several different approaches have been developed in this regard . The most common is to identify a low dimensional “feature space” to which the neuron is sensitive , with the assumption that its firing rate depends on a nonlinear function applied only to these stimulus features . Prominent examples of this approach include spike-triggered covariance ( STC ) analysis [12] , [13] , which uses the covariance of the stimuli that elicit spikes , and information-theoretic approaches such as maximally informative dimensions ( MID ) analysis [14] and iSTAC [15] . With the subspace determined , other methods can be used to estimate a nonlinear mapping between the projection of the stimulus onto this low dimensional feature space and the firing rate [13]–[19] . A second general approach is to assume the form of nonlinearities present , most commonly based on a second-order approximation of the nonlinear stimulus-response relationship , as with the Wiener-Volterra expansion [20]–[24] , and more recent versions cast in a probabilistic context [25]–[28] . This category might also encompass neural network approaches , which characterize the stimulus-response relationship in terms of a set of fixed nonlinear basis functions , using either generic network elements [29] , [30] or more specific nonlinear models of upstream sensory processing [31] , [32] . A final commonly used approach assumes that relevant nonlinearities can be captured by directly augmenting the linear model to account for specific response properties , such as the addition of refractoriness to account for neural precision [7] , [8] , [33]–[35] , feedback terms that account for adaptation to contrast [36]–[38] , and other nonlinearities to capture response properties such as sensitivity to stimulus intensity and local context [25] . Here , we present a probabilistic modeling framework inspired by all of these approaches , the ‘Nonlinear Input Model’ ( NIM ) , which limits the space of nonlinear functions by assuming that nonlinearities in sensory processing are dominated by spike generation , resulting in both rectification of the inputs to the neuron , as well as rectification of the neuron's output . By assuming a neuron's inputs are rectified , the NIM implicitly describes neuronal processing as a sum over excitatory and inhibitory inputs , which is increasingly being seen as an important factor in sensory processing [39]–[43] . The NIM expands directly on the GLM framework , and is able to utilize recent advances in the statistical modeling of neural responses [7] , [8] , [17] , [44] , [45] , including the ability to model spike-refractoriness [7] , [8] and multi-neuron correlations [8] , [44] . As we show here , this results in a parsimonious nonlinear description of a range of neurons in both the visual and auditory systems , and has several advantages over previous approaches . Because of its relatively simple model structure , parameter estimation is well-behaved and makes efficient use of the data , even when the number of relevant inputs is large and/or the stimulus is high-dimensional . Importantly , because its form is based on an integrate-and-fire neuron , model selection and parameter estimation can be guided by specific knowledge about the inputs to a given neuron , and the elements of the resulting model can often be related to specific physiological predictions . The NIM thus provides a powerful and general approach for nonlinear modeling that complements other methods that rely on more abstract formulations of nonlinear computation . Perhaps the greatest success of linear models is in the retina , where it has been used primarily to describe the spike responses of retinal ganglion cells ( RGCs ) [10] , [11] , [16] . For a given RGC , estimating the components of the linear model typically involves measuring its spiking response to a noise stimulus , and then computing the average stimulus that preceded its spikes: the spike-triggered average ( STA ) . The STA linear filter can produce very good response predictions for typical RGCs under stationary stimulus conditions , but clearly fails for ON-OFF cells ( commonly found in rodents ) , which respond to both increases and decreases of light intensity [46]–[49] . This failure for ON-OFF cells occurs simply because the STA identifies only a single stimulus dimension , and averages out the opposing stimulus features that evoke ON and OFF responses . To explore this situation , we construct a basic model of an ON-OFF RGC , which receives separate ON and OFF inputs ( Fig . 1A ) . If these two inputs were to combine linearly , their effect would be identical to that of a single input generated by the sum of the two stimulus filters , i . e . , ( s·kON ) + ( s·kOFF ) = s· ( kON+kOFF ) = s·kSUM . Here the stimulus s at a particular time is represented as a vector ( which in general includes time-lagged elements to account for stimulus history ) such that the operation of a linear filter k is given by a dot product . Because of the averaging implicit in linear processing , a nonlinear transformation must be applied to each input in order to enable the model ON-OFF neuron to respond to both types of stimuli: i . e . , f ( s·kON ) +f ( s·kOFF ) . These f ( . ) are taken to be rectifying functions ( Fig . 1A ) , as seen experimentally [50] , [51] , and as modeled in [52]–[54] . As a result , the response of the neuron to increases or decreases of luminance is dominated by the ON or OFF pathways respectively ( Fig . 1B ) , producing a response that is selective to both ON and OFF stimulus dimensions . As expected , the STA ( Fig . 1C ) for this neuron does not match either the ON or OFF stimulus filters , but rather reflects their average . Thus , this is a clear example where nonlinear characterization is necessary to capture the RGC's stimulus selectivity . One such approach that has been applied to ON-OFF cells is spike-triggered covariance ( STC ) analysis [49] , [55] , which identifies stimulus dimensions along which the variance of the spike-triggered ensemble is either increased or decreased relative to the stimulus distribution [12] , [13] . For the example neuron in Fig . 1 , STC analysis identifies a stimulus dimension along which the variance of the spike-triggered ensemble is expanded ( Figs . 1D , E ) . While neither the STA nor STC filters correspond to the true ON or OFF filters , together they define a stimulus subspace that contains the true filters ( Fig . 1E ) . Given the dimensionality reduction achieved in determining the STC subspace ( or with other subspace identification methods ) , it is possible in principle to completely characterize the neural response function , i . e . , r = F[k1·s , k2·s] . In two-dimensions , such as in this example , this nonlinear mapping from the subspace to a firing rate can be estimated non-parametrically [18] , [56] given enough data , and potentially approximated in higher dimensions [13] , [15] , [18] , [19] , [57] . However , even if accurate estimation of this nonlinear mapping were possible , such functions are difficult to interpret , even when arising from the conjunction of simpler components . For example , in our simulated ON-OFF RGC , neither the STA/STC filters themselves nor the measured nonlinear mapping make it clear that the response is generated from separate inputs with relatively straightforward nonlinearities . This example thus motivates the modeling framework that we present here , the Nonlinear Input Model ( NIM ) , which describes a neuron's stimulus processing as a sum of nonlinear inputs , following the structure of the generative model shown in Fig . 1 . Below , we first present procedures for estimating the parameters of the NIM before demonstrating its ability to recover the inputs to the ON-OFF RGC , as well as its application to a range of other simulated and measured data from both visual and auditory brain areas . The computational challenges associated with parameter estimation are a significant barrier to the successful development and application of nonlinear models of sensory processing . In the standard linear-nonlinear ( LN ) model , the neuron's response is modeled by an initial stage of linear stimulus filtering , followed by a static nonlinear function ( “spiking nonlinearity” ) that maps the output to a firing rate ( Fig . 2A ) . The more recent adaptation of probabilistic models based on spike train likelihoods , such as in the Generalized Linear Model ( GLM ) [6]–[8] , allows for integration of other aspects of neuronal processing into the linear stimulus-processing framework , and can be used to model nonlinear stimulus processing through predefined nonlinear transformations [31] , [32] , [58] . Importantly , this approach also provides a foundation for parameter estimation for the NIM . A principal motivation for the NIM structure is that if the neuronal output at one level is well described by an LN model , downstream neurons will receive inputs that are already rectified ( or otherwise nonlinearly transformed ) . Thus , we use LN models to represent the inputs to the neuron in question , and the neuron's response is given by a summation over these LN inputs followed by the neuron's own spiking nonlinearity ( Fig . 2B ) . Importantly , this allows us to account for the rectification of a neuron's inputs imposed by the spike-generation process . The NIM can thus be viewed as a ‘second-order’ generalization of the LN model , or an LNLN cascade [59] , [60] . Previous work from our lab [45] cast this model structure in a probabilistic form , and suggested several statistical innovations in order to fit the models using neural data [45] , [61] , [62] . Here , we present a general and detailed framework for NIM parameter estimation that greatly extends the applicability of the model . This model structure has also been suggested for applications outside of neuroscience in the form of projection pursuit regression [63] , including generalizations to response variables with distributions from the exponential family [64] . The processing of the NIM is comprised of three stages ( Fig . 2C ) : ( a ) the filters ki that define the stimulus selectivity of each input; ( b ) the static ‘upstream’ nonlinearities fi ( . ) and corresponding linear weights wi which determine how each input contributes to the overall response; and ( c ) the spiking nonlinearity F[ . ] applied to the linear sum over the neuron's inputs . The predicted firing rate r ( t ) is then given as: ( 1 ) where s ( t ) is the ( vector-valued ) stimulus at time t , x ( t ) represents any additional covariates ( such as the neuron's own spike history ) , and h is a linear filter operating on x . Note that equation ( 1 ) reduces to a GLM when the fi ( . ) are linear functions . The wi can also be extended to include temporal convolution of the subunit contributions to model the time course of post-synaptic responses associated with individual inputs [45] , as well as ‘spatial’ convolutions to account for multiple spatially distributed inputs with similar stimulus selectivity [65] . Since equivalent models can be produced by rescaling the wi , and fi ( . ) ( see Methods ) , we constrain the subunit weights wi to be either +/−1 . Because we generally assume the fi ( . ) are rectifying functions , the wi thus specify whether each subunit will have an ‘excitatory’ or ‘inhibitory’ influence on the neuron . Parameter estimation for the NIM is based on maximum likelihood ( or maximum a posteriori ) methods similar to those used with the GLM [6]–[8] . Assuming that the neuron's spikes are described in discrete time by a conditionally inhomogeneous Poisson count process with rate function r ( t ) , the log-likelihood ( LL ) of the model parameters given an observed set of spike counts Robs ( t ) is given ( up to an overall constant ) by: ( 2 ) To find the set of parameters that maximize the likelihood ( eq . 2 ) , we adapt methods that allow for efficient parameter optimization of the GLM [7] . First , we use a parametric spiking nonlinearity given by F[x] = αlog[1+exp ( β ( x-θ ) ) ] , with scale α , shape β , and offset θ . Other functions can be used , so long as they satisfy conditions specified in [7] . This ensures that the likelihood surface will be concave with respect to linear parameters inside the spiking nonlinearity [7] , and in practice will be well-behaved for other model parameters ( see Fig . S1; Methods ) . Because it is straightforward to estimate the linear term h , and the wi are constrained to be +/−1 , the upstream nonlinearities fi ( . ) and the stimulus filters ki are the key components that must be fit in the NIM . While it is typically not feasible to optimize the likelihood with respect to both sets of parameters simultaneously , an efficient strategy is to use block coordinate ascent [66] , alternating between optimizing the ki and fi ( . ) , in each case holding the remaining set of parameters constant ( see Methods ) . ‘Linear’ parameters , such as h and θ , can be optimized simultaneously during either ( or both ) optimization stages . While the set of ‘upstream nonlinearities’ fi ( . ) can be represented as parametric functions such as rectified-linear or quadratic functions ( see Methods ) , a powerful approach is to represent them as a linear combination of basis functions φj ( . ) such as piecewise linear “tent” basis functions , i . e . , fi ( g ) = Σj aijφj ( g ) [17] , [45] . In doing so , estimation of the upstream nonlinearities reduces to estimating linear parameters aij inside the spiking nonlinearity , with a single global optimum of the likelihood function for a given set of stimulus filters ki . For a fixed set of upstream nonlinearities , the stimulus filters ki can be similarly optimized , although the resulting likelihood surface will not in general be convex because the ki operate inside the upstream nonlinearities . Nevertheless , we have found that in practice their optimization is well-behaved and that local minima can be avoided with appropriate optimization procedures ( Fig . S1; see Methods ) . Furthermore , it is straightforward to evaluate the likelihood function and its gradient with respect to the ki analytically ( see Methods ) , allowing for efficient gradient-based optimization . Thus , optimal parameter estimates for the NIM can be determined efficiently , even for models with large numbers of parameters ( see examples below ) . The time required for filter estimation ( typically the most time-consuming step ) scales approximately linearly with the experiment duration , the dimensionality of the stimulus , and the number of model subunits ( Fig . S2 ) . This is very favorable compared with alternative nonlinear modeling approaches such as MID [14] , which require using simulated annealing and quickly becomes intractable as the number of filters and/or stimulus dimensions is increased . Furthermore , because the NIM provides an explicit probabilistic model for the neuronal spike response , regularization of the model components can be incorporated without adversely affecting the behavior of the optimization problem [7] ( see Methods ) . This is particularly important when optimizing high-dimensional spatiotemporal filters and/or models with many inputs , which are both discussed further below . Likewise , as with other probabilistic modeling approaches – but not those relying on spike-triggered measurements [67] – the model can be optimized using data recorded with natural stimulus ensembles ( containing complex correlation structure , and non-Gaussian distributions ) without introducing biases into the parameter estimates . The NIM thus provides a nonlinear modeling framework in which large numbers of parameters can be efficiently estimated using data recorded with arbitrarily complex stimulus ensembles . In addition to this flexibility , the NIM provides model fits that are more directly interpretable due to its physiologically motivated model structure . To illustrate these advantages , below we first apply the NIM to the example ON-OFF RGC from Fig . 1 , and then demonstrate its wide applicability on recorded and simulated neurons in several different sensory areas . Returning to the example ON-OFF RGC ( Fig . 1 ) , the NIM is a natural choice given that its structure matches that of the simulated neuron . Using the estimation procedures described above , the NIM is able to successfully capture the true stimulus selectivity of its individual inputs ( Fig . 3A ) , including the ‘upstream nonlinearities’ associated with each input , as well as the form of the spiking nonlinearity ( see Methods ) . This example thus illustrates the core motivation behind the NIM of modeling a neuron's stimulus processing in terms of rectified neuronal inputs . While the structure of the simulated RGC neuron in this example may appear to be a convenient choice , its form is consistent with other models of ON-OFF processing [48] , [53] , and with models of RGCs in general [54] , [68] . Thus , to understand the advantages and disadvantages of the NIM structure , it is useful to compare it with the dominant alternative approach for describing nonlinear stimulus processing: “quadratic models” . Such models have recently been cast in an information-theoretic context [15] , [27] , [28] , as well as in the form of an explicit probabilistic model [26] which has been referred to as the ‘Generalized Quadratic Model’ ( GQM ) . The GQM can be viewed as a probabilistic generalization of STA/STC analysis [26] and of the second-order Wiener-Volterra expansion [20] . The GQM can also be written in the form of a NIM where the upstream nonlinearities fi ( . ) are fixed to be linear or squared functions: ( 3 ) where kL is a linear filter , and the M squared filters ki generally provide a low-rank approximation to the quadratic component C [26] . In this sense , the probabilistic framework described here is easily extended to encompass quadratic models , providing a means for direct comparison between different nonlinear structures . For the ON-OFF RGC , the GQM finds one linear and two quadratic filters , all of which are contained in the two-dimensional subspace identified by STC analysis , meaning that the GQM filters are also linear combinations of the true ON and OFF filters ( Fig . 3B ) . Note that while two filters are sufficient to span the relevant stimulus subspace , the third GQM filter provides an additional degree of freedom to capture the best quadratic approximation to the underlying ‘neural response function’ ( Fig . 3E ) . Although in this example the resulting quadratic function cannot completely capture the form of the response function constructed from rectified inputs , we note that it still provides a good approximation , as shown by only modest reductions in model performance compared to the NIM ( Fig . 3F ) . However , as expected from a second-order Taylor series expansion , such an approximation breaks down further from the “origin” of the subspace . Thus , the quadratic approximation will typically be less robust for stimuli with heavy-tailed distributions such as those associated with natural stimuli [69]–[71] . To illustrate this point we performed simulations of the same ON-OFF RGC presented with white noise stimuli having a Student's t-distribution , where the tail thickness was controlled by varying the number of degrees of freedom ( Fig . 3G ) . The improved performance of the NIM over the GQM is indeed substantially enhanced for stimulus distributions with heavier tails ( Fig . 3H ) . We also verified similar effects for a range of simulated neurons ( data not shown ) . We emphasize that one of the key advantages of the NIM over previously described methods is that it provides a more interpretable picture of stimulus processing as a sum of rectified neuronal inputs . As we demonstrate through several examples below in both the visual and auditory systems , it appears that sensory computation by neurons will often adhere to this general form , which is motivated primarily by physiological , rather than mathematical , considerations . One of the main advantages of the NIM structure is the ability to specifically model the effects of inhibitory inputs , which are increasingly being shown to have a large impact on neuronal processing in many sensory areas [72]–[74] . Indeed , the NIM generates predictions of the functional tuning of excitation and inhibition , and provides insight into their role in sensory processing . To demonstrate this , we apply the NIM to example neurons from visual and auditory areas . We first consider an example cat LGN neuron during the presentation of natural movies [45] , [75] , [76] . Accurate characterization of LGN processing poses substantial challenges for previous nonlinear approaches , due to the high temporal resolution of LGN responses in this context [77] combined with the large number of spatial dimensions of the stimulus . As a result , previous nonlinear applications have either utilized lower temporal resolutions [78] , [79] or parametric models of the spatial processing [38] , [45] , [80] . The methods described here allow for ( appropriately regularized ) spatiotemporal receptive fields ( STRFs ) of LGN neurons to be fit at sufficiently high resolution , using natural movies . We find that the response of the example LGN neuron consists of an excitatory receptive field that is delayed relative to the linear STRF ( Fig . 4A ) , along with a second , more delayed ‘suppressive’ receptive field ( Fig . 4B ) , corresponding to putative inhibitory input . Unlike in previous studies , the tractability of the fitting procedures used here allows for high spatial and temporal resolution of the putative inputs ( Fig . 4B ) , as well as the application of sparseness and smoothness regularization ( see Methods ) . By comparison , the GQM identifies similar STRFs , but has worse performance ( Fig . 4C ) , as well as a different nonlinear structure and resulting physiological interpretation ( Fig . S3 ) . Next we consider an example neuron from zebra finch area MLd , as the animal is presented with conspecific bird songs [81]–[83] . These neurons respond to specific frequencies of the song input , and hence their stimulus selectivity can be characterized by a linear spectrotemporal receptive field ( STRF ) [9] , which can be recovered in an unbiased manner using maximum-likelihood estimation ( Fig . 5A ) [84] despite the presence of higher order correlations in the stimulus . Application of the NIM to this example neuron again recovers both an excitatory and a temporally delayed suppressive component ( Fig . 5B ) . The description of the neuron's stimulus tuning provided by the NIM is closely related to that given by the linear model , but instead of identifying positive and negative domains of the linear STRF as excitatory and suppressive , these effects are segregated into different nonlinear processing subunits , each individually rectified . The separate excitatory and suppressive inputs provide a more accurate description of the underlying stimulus processing than a single linear STRF , as demonstrated by the significantly improved model performance of the NIM compared with the LN model ( Fig . 5C ) . As with the LGN example , the GQM identifies similar excitatory and suppressive filters as the NIM , but again provides a less physiologically interpretable description of the underlying computation ( Fig . S4 ) , and has comparable , if slightly reduced , performance ( Fig . 5C ) . Thus far we have only considered cases where the neuron's response is described by a NIM with a small number of inputs , consistent with simpler stimulus processing in sub-cortical areas . In contrast , in the visual cortex , even V1 ‘simple cells’ can exhibit selectivity to large numbers of stimulus dimensions [57] , [62] . Further , the dominant model of V1 ‘complex cells’ is the nonlinear “Energy Model” [10] , [85]–[87] , which posits quadratic stimulus processing that results in the response representing the amount of local , oriented , band-pass “stimulus energy” . The Energy Model has been broadly tested [10] , [87] , and is well supported by previous nonlinear modeling approaches [13] , [26] , [27] , [57] , [88] . While the Energy Model provides a functional description of stimulus processing for V1 complex cells , it is less clear how such stimulus selectivity is constructed , and how it is related to V1 simple cell processing . Here we demonstrate that the NIM can describe both simple and complex cell processing as a sum of rectified inputs , providing a basis for a unified description of visual cortical neuron computation [62] . We first consider two simulated V1 neurons in order to demonstrate the capacity for such a unified description , before applying the NIM to experimental data . We generate simulated data using a one-dimensional white-noise bar stimulus aligned with the simulated neurons' preferred spatial orientation ( Fig . 6A ) , which is a common , relatively low-dimensional , stimulus used in nonlinear characterizations of V1 neurons [57] , [88] , [89] . The first simulated neuron's response is constructed as a sum of six rectified direction-selective inputs ( Fig . 6B ) , consistent with the structure of the NIM , while the second neuron's response is constructed from four such inputs processed by a squaring nonlinearity , similar to the standard Energy Model of V1 complex cells [85] . For the neuron with rectified inputs , the NIM fitting procedure is indeed able to identify the true underlying stimulus filters and the form of the rectifying upstream nonlinearities ( Fig . 6C ) . Additionally , while the optimal number of filters can be determined using the cross-validated model performance , the identified stimulus filters , and the resulting model performance itself , are relatively insensitive to specification of the precise number of model subunits ( Fig . S5 ) . This demonstrates the ability of the NIM to robustly identify even relatively complex stimulus processing , in cases where such processing arises from a sum of rectified inputs . Furthermore , as with the ON-OFF RGC example above ( Figs . 1 and 3 ) , STC analysis of this simulated V1 neuron can identify the appropriate stimulus subspace , although not the true underlying filters ( Fig . 6D ) . Because of the high dimensionality of the resulting subspace , however , it is more difficult to estimate the mapping from the subspace to the neuronal response compared with the ON-OFF example . The lack of alignment between the STA/STC filters and the true filters further complicates a straightforward interpretation of the estimated function . By comparison , the GQM identifies filters with characteristics that more closely resemble those of the true input filters ( e . g . , more localized , fewer lobes ) . The improved performance of the GQM compared with an STC-based model ( Fig . 6E ) highlights the greater power and flexibility of a probabilistic modeling framework , particularly the importance of regularization . Nevertheless , the GQM filters still reflect non-trivial linear combinations of the true filters , as with the STC filters ( Fig . 6E , bottom ) . Of course , one would expect the NIM to outperform other models when the generative model is composed as a sum of rectified inputs . In a second simulated example , however , we illustrate the flexibility of the NIM in capturing other neural response functions . The second simulated neuron is constructed from four direction-selective inputs that are squared and summed together to generate a quadratic response function ( Fig . 6F ) . The NIM is still able to identify the true generative model using pairs of rectified inputs with equal but opposite input filters to represent each quadratic filter ( Fig . 6G ) . This representation is certainly not the most efficient in this case , as the GQM is able to identify the correct filters ( Fig . 6I ) using fewer parameters and a more straightforward estimation procedure . These two simulated V1 examples thus illustrate the potential tradeoffs between the NIM and GQM . On the one hand , the NIM provides a more flexible framework that can capture a broader range of nonlinear stimulus processing . In fact , any response function can in principle be represented with this structure [90] . The NIM structure is also more appropriate for explicitly modeling neuronal inputs , and thus allows for more plausible physiological interpretation of its components . On the other hand , the GQM can capture the nonlinear mapping up to second order more efficiently , and identifies the relevant stimulus subspace robustly . This suggests the potential for combining these approaches when investigating complex neuronal processing , such as by using the GQM to identify the relevant stimulus subspace and provide initial estimates of the number and properties of NIM filters , followed by application of the NIM framework ( see Methods; Figs . S1 , S3 ) . While the simulated examples above allowed for model comparisons when the neurons' response functions were known , they also provide a foundation for understanding model fits to real V1 data . We first consider a V1 neuron recorded from an anesthetized macaque in the context of similar one-dimensional white noise stimuli [57] . While this neuron has a clear STA and is considered a simple cell by classical measures , STC analysis identifies two excitatory and six suppressive stimulus dimensions ( based on inspection of the eigenvalue spectrum ) in addition to the STA ( Fig . 7A ) . In this case , the GQM identifies similar filters to STC analysis , although the application of smoothness and sparseness regularization allows it to resolve more realistic stimulus filters ( Fig . 7B ) , and produce significantly improved model performance ( Fig . 7D ) compared to an STC-based model ( see Methods ) . We also fit a NIM with six excitatory and six suppressive stimulus filters , where the number of filters was selected based on cross-validated model performance ( Fig . 7C; see also Fig . S5 ) . As expected , these 12 filters span a stimulus subspace that is largely overlapping with the subspace identified by the GQM . However , the additional stimulus filters , and the inferred upstream nonlinearities associated with each subunit , allow the NIM to capture additional aspects of the neural response function that significantly improve the cross-validated model performance relative to the quadratic models ( Figs . 7D , E ) . We also note that the NIM appears to identify a more consistent set of stimulus filters than the quadratic models . Similar comparisons also come to light in when applying the models to V1 complex cells , even in the most demanding stimulus contexts . To illustrate this , we consider an example V1 neuron recorded from an anesthetized cat presented with natural and naturalistic stimuli ( Fig . 8A ) [62] , [91] . Because the stimuli are sequences of two-dimensional images , the required spatiotemporal stimulus filters span two dimensions of space and one dimension of time ( Fig . 8B ) , resulting in a very large number of parameters associated with each subunit . Nevertheless , the parameters of the GQM and NIM can be estimated directly utilizing appropriate regularization ( see Methods ) . The GQM estimated for this neuron is comprised of a pair of excitatory , direction-selective squared filters , as well as a weaker , non-direction-selective linear filter ( Fig . 8C ) . This characterization reflects the neuron's spatial-phase invariance , and is thus consistent with an Energy Model description . While such selectivity suggests that this neuron would be ideally suited for a quadratic model , the NIM ( Fig . 8D ) significantly outperforms both the GQM and a whitened STC-based model [9] , [58] , [62] ( Fig . 8E ) . The NIM identifies four rectified excitatory inputs that share similar spatial tuning and direction selectivity , but with different spatial phases ( Fig . 8D ) . This description is similar to that provided by the quadratic terms of the GQM , but by identifying the nonlinearities associated with each of these inputs individually , the NIM has additional flexibility that results in improved performance ( Fig . 8E ) . This suggests that a description of complex cells using physiologically plausible inputs ( in the form of the NIM ) may be a viable alternative to the Energy Model . The improved performance of the NIM is also likely due , at least in part , to the heavy-tailed distribution associated with the naturalistic movie stimuli ( as described above , Figs . 3G , H ) . Thus , the application of the NIM to V1 neurons further illustrates the generality of the method , and specifically emphasizes its ability to capture substantially more complex stimulus processing , with large numbers of inputs . We note that because cortical neurons are several synapses removed from receptor neurons , a cascade model with a longer chain of upstream LN components might be more appropriate , although existing methods could not be used for parameter estimation with such a model . The ability of the NIM to capture a given neuron's stimulus processing thus relates to the extent to which the upstream neurons themselves can be approximated by LN models . In cases where this assumption is not appropriate , one can apply a fixed nonlinear transformation to the stimulus resembling the response properties of upstream neurons [31] , [32] , thus allowing the problem to be cast into a more general NIM framework . We have presented a physiologically inspired modeling framework , the NIM , which extends several recently developed probabilistic modeling approaches . Specifically , the NIM assumes a form analogous to an integrate-and-fire neuron , whereby a neuron receives a set of rectified excitatory and inhibitory inputs , each of which is assumed to process the stimulus linearly . The parameters can be estimated robustly and efficiently , and the resulting model structure is able to capture a broader range of neural responses than previously proposed probabilistic methods . Importantly , the physiologically inspired model structure of the NIM also allows for greater interpretability of the model fits , as the components of the model take the form of stimulus-driven excitatory and inhibitory inputs . The NIM thus provides a framework for connecting nonlinear models of sensory processing directly with the underlying physiology that can be applied in a range of sensory areas and experimental conditions . As described above , the key parameters in the NIM are the stimulus filters ki and the set of coefficients aij representing the upstream nonlinearities fi ( . ) . While these parameters cannot generally be optimized simultaneously , a powerful approach is to use block coordinate ascent [66] and alternate between optimizing the filters ki , and upstream nonlinearities fi ( . ) , holding the remaining parameters fixed in each iteration . The parameters of the spiking nonlinearity function F[x; α , β , θ] = αlog[1+exp ( β ( x-θ ) ) ] can also be estimated iteratively , or as a final stage after convergence of the ki and fi ( . ) ( which we find is typically sufficient ) . Note that the parameter β is not generally identifiable in the model ( being degenerate with the coefficients aij of the upstream nonlinearities ) , but joint estimation of α and β after the other model parameters are fixed allows for a more precise final fit to the spiking nonlinearity function . Thus , at each stage of the fitting procedure we have the problem of maximizing a ( penalized ) log-likelihood function with respect to some subset of parameters , while holding a remaining set of parameters fixed . In all cases , we use a standard line search strategy to locate an optimum of the likelihood function given some initial values for the parameters . Because we are often optimizing very high-dimensional parameter vectors ( specifically when optimizing the ki ) , we use a quasi-Newton method with a limited-memory BFGS approximation of the inverse Hessian matrix [92] to determine the search direction . This code is implemented in the Matlab function “minFunc” , provided by Mark Schmidt ( available at http://www . di . ens . fr/~mschmidt/ ) . When using sparseness ( L1 ) regularization we utilize the Matlab package “L1General” , also provided by Mark Schmidt . When optimizing the coefficients aij of the upstream nonlinearities we additionally enforce a set of linear constraints ( described below ) , and in such cases we utilize Matlab's constrained optimization routine “fmincon” . A Matlab implementation of the NIM parameter estimation routines described here is available from our website: ( http://www . clfs . umd . edu/biology/ntlab/NIM/ ) Optimization of the filters can be accomplished efficiently by analytic calculation of the log-likelihood gradient with respect to the ki , which is given by: ( 4 ) where the ‘internal generating function’ , F'[ . ] and fi' ( . ) are the derivatives of F[ . ] and fi ( . ) with respect to their arguments , and sm ( t ) is the mth element of the stimulus at time t . While the likelihood surface is not generally convex with respect to the ki , the optimization problem is well-behaved in practice . We note that while the derivatives of the fi ( . ) are discontinuous ( piece-wise constant ) when using the tent-basis representation ( eq . 6 below ) , gradient-based optimization methods still provide robust results , in particular because we use regularization to enforce smooth fi ( . ) such that the contribution of the discontinuities to the overall log-likelihood gradient is negligible in practice . To diagnose the presence of undesirable local maxima , and to identify the global optimum of the likelihood function , we use repeated random initializations of our optimization routine ( Fig . S1 ) . In some cases , such as the ON-OFF RGC example ( Figs . 1 , 3 ) , this approach reveals that the choice of initial values for ki does not affect the identified local optimum . In other cases , the likelihood surface will contain more than one distinct local maximum , although usually only a small number . For example , when optimizing the filters for the example MLd neuron ( Fig . 5 ) we found two distinct local optima of the likelihood function . For models with large numbers of subunits , the filter optimization remains well-behaved , generally identifying a relatively small number of local optima that correspond to similar models ( Fig . S1 ) . This procedure can be greatly sped up by initially optimizing the filters in a low-dimensional stimulus subspace , rather than in the full stimulus space . Such subspace optimization has been previous used in conjunction with STC analysis to identify the relevant stimulus subspace [15] , [62] , [93]; however the GQM provides a means of generalizing the robust subspace identification properties of STC analysis to arbitrary non-Gaussian stimuli , and in cases where regularization is important . With a low-dimensional subspace identified the filters of a NIM can be rapidly optimized , and many filter initializations can be tested . We begin the NIM fitting with its upstream nonlinearities fi ( . ) initialized to be threshold-linear functions: ( 5 ) and perform initial estimation of the filters . While other rectifying functions can be used , the use of scale-invariant functions such as this one has the advantage that the effect of the upstream nonlinearity is independent of the scale of the filter . After estimating the ki , we then estimate the fi ( . ) nonparametrically , as a linear combination of a set of piecewise linear basis functions fi ( g ) = Σj aijφj ( g ) [17] , [45] , while holding the ki fixed . These basis functions are given by: ( 6 ) These piecewise linear functions are particularly useful as they provide a set of localized basis functions , requiring only that we choose a set of ‘grid points’ xk . These points can be selected by referencing the distribution of the argument of fi ( . ) , i . e . , p ( gi ) where gi ( t ) = ki•s ( t ) , either at n-quantiles of p ( gi ) , or at uniformly spaced points across the support of p ( gi ) . In order to encourage interpretability of the model subunits as ‘neural inputs’ , we constrain the fi ( . ) to be monotonically increasing functions by using a system of linear constraints on the aij during optimization . Because the model is invariant to shifts in the ‘y-offset’ of the fi ( . ) ( which can be absorbed into the spiking nonlinearity function ) , we add the additional set of constraints that fi ( 0 ) = 0 to eliminate this degeneracy . Furthermore , changes in the upstream nonlinearities can influence the effective regularization of the ki , by altering how each ki contributes to the model prediction . As a result , the coefficients aij are rescaled after each iteration so that the standard deviation of each subunit's output is conserved . This ensures that the upstream nonlinearities do not absorb the scale of the ki . An important advantage of explicit probabilistic models such as the NIM is the ability to incorporate prior knowledge about the parameters via regularization . Because each of the filters ki often contains a large number of parameters , regularization of the filters is of particular importance , as discussed elsewhere in the context of the GLM [9] , [58] , [84] , [94]–[97] , as well as other nonlinear models [26] . Such regularization can impose prior knowledge about the smoothness [9] , [26] , [94] , sparseness [58] , [84] , [94]–[96] , and localization [62] , [97] of filters in space , frequency and time . We consider several different forms of regularization in the examples shown , to encourage the detection of smooth filters with sparse coefficients . Specifically , we add a general penalty term of the form: ( 7 ) to the equation for the log-likelihood ( equation 2 ) , where Ls and Lt are the discrete Laplacian operators with respect to spatial ( or spectral ) and temporal dimensions respectively , and λiLs , λiLt and λis are hyperparameters which determine the strength of spatial and temporal smoothness , and sparseness regularization , respectively . Other types of regularization , such as those that encourage localized filters [62] , [97] , as well as approximate Bayesian techniques for inferring hyperparameters [26] , [58] , [94] , [96] could be incorporated as well , although we do not do so here . Because we also expect the upstream nonlinearities fi ( . ) to be smooth functions , we incorporate penalty terms when estimating the parameters of the fi ( . ) . Because we represent the fi ( . ) as linear combinations of localized tent basis functions: fi ( . ) = aijφj ( . ) , we can encourage smooth fi ( . ) by applying a penalty of the form: λiL∥Laij∥2 to the set of coefficients aij corresponding to a given fi ( . ) , where L is again the one-dimensional discrete Laplacian operator . In general , the hyperparameters can be inferred from the data using Bayesian techniques [94] , or estimated using a ( separate ) cross-validation data set . Both methods can be time-consuming , however , and in practice we find that similar results can be achieved by ‘manually’ tuning the hyperparameters to produce filters ki and upstream nonlinearities fi ( . ) with the expected degree of smoothness/sparseness . To demonstrate that our results were not overly sensitive to the selection of hyperparameters , we compare the NIM and GQM fit to the example V1 neuron from Fig . 8 using a range of regularization strengths ( Fig . S6 ) . To evaluate model performance , we use k-fold cross-validation , in general taking the log-likelihood as a performance metric . The likelihood has the advantage over related measures such as R2 in that it does not require repeated stimulus presentations to estimate , and thus can be applied to most data sets . It can also capture goodness-of-fit when spike history terms are incorporated [35] . Subtracting the log-likelihood of the null model ( that predicts a constant firing rate , independent of the stimulus ) provides a measure of the information carried by the spike train about the stimulus , in units of bits per spike [45] , [98] . This measure is also directly related to the more traditional measure of deviance , which compares the log-likelihood of the estimated model to that of the ‘saturated’ model . In order to provide a more direct connection to standard measures of model performance based on repeated presentations of a stimulus , we also computed the ‘predictive power’ of the models for the simulated ON-OFF RGC ( Fig . 3F ) , which is defined as the fraction of ‘explainable’ variance accounted for by the model [99] . Due to the lack of sufficient repeat trial data for our recorded data examples we could not compute this measure in those cases , however qualitatively similar results would be expected . While selection of the optimal number of excitatory and suppressive subunits can be performed using standard model selection techniques , such as nested cross-validation , this choice can also often be guided by the specific application . Importantly , we find that the subunits identified by the NIM , as well as its performance , are generally robust towards precise specification of the number of excitatory and suppressive subunits , with ‘nearby’ models typically providing a very similar characterization of the neurons' stimulus processing ( Fig . S5 ) . This robustness is further aided by the incorporation of sparseness regularization on the filters , where the filters of extraneous subunits tend to be driven to zero . The procedure of testing a series of NIMs with different subunit compositions can again be substantially facilitated by optimizing the filters in a low-dimensional stimulus subspace , such as identified by STC or GQM analysis ( Fig . S5 ) . In order to simulate the response of an ON-OFF RGC , we generated a Gaussian white noise process sampled at 15 Hz ( such as a luminance-modulated spot stimulus ) , which was then filtered using separate ON- and OFF-like filters ( Fig . 1A ) . These filter outputs were then rectified using functions of the form f ( x ) = log ( 1+exp ( b1x ) ) , summed together and the resulting signal was passed through a spiking nonlinearity of the form F[x] = alog ( 1+exp ( b2 ( x-c ) ) ) . This conditional intensity function was then used to generate a set of spike times . To generate heavy-tailed stimulus distributions ( Figs . 3G , H ) , we sampled white noise from a Student's t-distribution with a range of values for the degrees of freedom to control the tail thickness . The data were simulated at a temporal resolution of 8 . 3 ms , and model filters were represented at a lower resolution of 33 ms , with a length of 1 s . For the GQM and NIM we incorporated smoothness regularization on the filters , and for the NIM we also incorporated smoothness regularization on the upstream nonlinearity coefficients aij . To identify the STA/STC subspace depicted in Figs . 1 and 3 , we performed STC analysis after projecting out the STA . For comparison with the NIM , we also created a simple model based on the STA and STC filters , using a GLM-based optimization of linear coefficients on the outputs of the STA filter and the squared outputs of the STC filters , similar to previous work [58] . Note that in order to maximize performance when estimating STC-based models , we did not project out the STA before computing the STC filters . Data for the LGN example were recorded extracellularly from an anaesthetized and paralyzed cat by the Alonso Lab [45] , [75] , [76] . The stimulus consisted of 800 seconds of a 32×32 pixel natural movie , refreshed at 60 Hz , which was recorded from a camera mounted on top of a cat's head [100] . A 17×17 pixel patch of the movie was cropped around the receptive field , detected via STA at the optimal latency , and the movie was up-sampled by a factor of six , to produce a temporal resolution of 2 . 8 ms . Ten-fold cross-validation was used for evaluating model performance . Each filter was represented by space-time separable center and surround components , and thus consisted of two sets of spatial and temporal filters [101] . Temporal filters were represented with 30 equally spaced tent basis functions , with grid points ranging from 0 to −240 ms . For the LN model , the spatial filters were initialized as Gaussian functions with the same center as the STA and different widths ( 1 pixel for the center and 6 pixels for the surround ) . For the GQM and NIM , both excitatory and suppressive filters were initialized to be the same as the optimal linear filters . In the filter optimization stage , the spatial and temporal filters were optimized alternately until convergence of the log-likelihood . Both the GQM and the NIM were fit using smoothness regularization for the spatial and temporal kernels , and sparseness regularization for the spatial kernels . For the NIM , we also used smoothness regularization on the aij when estimating the upstream nonlinearities . Data for the songbird auditory midbrain example were provided by the Theunissen lab through the CRCNS database [83] , and details of experimental methods can be found in [81] , [82] . The example neuron was recorded extracellularly from the zebra finch mesencephalicus lateralis dorsalis ( MLd ) . Stimuli consisted of 20 different conspecific bird songs , each lasting 2–4 sec , and each presented 10 times . These 20 songs were then divided into 5 equal groups for five-fold cross-validation . The raw sound waveforms were preprocessed by computing the spectrogram using a short-time Fourier transform to produce a stimulus matrix X ( t , f ) , representing the power of the audio signal at frequency f and time t . We used a time resolution of 2 ms , and 20 uniformly spaced frequency bins , ranging from 250 Hz to 8 kHz . For estimating spectrotemporal filters , we used 20 time lags . Thus , each filter was represented by 400 parameters . Filter estimates were regularized using sparseness and smoothness penalties , where the smoothness penalty utilized the spectrotemporal Laplacian ( with equal weighting in the frequency and time dimensions ) . The simulated V1 neurons shown in Fig . 6 were constructed as LNLN models ( Fig . 2C ) . The stimulus filters were spatial Gabor functions that were amplitude- and phase-modulated in time ( i . e . , direction-selective ) . Stimulus filters were identical up to a spatial translation , and were weighted by a spatial Gaussian envelope . The filter outputs were then passed through a set of static nonlinear functions ( either x2 , or log ( 1+exp ( b1x ) ) ) , before being summed together , and passed through the spiking nonlinearity ( again , of the form alog ( 1+exp ( b2 ( x-c ) ) ) to generate a conditional intensity function . Spike times were simulated in response to binary random bar stimuli [57] , using a time resolution of 10 ms , and 24 bar positions . Both the GQM and NIM were fit using a sparseness penalty on the filters . For the NIM , we also used smoothness regularization on the aij when estimating the upstream nonlinearities . To measure how well the estimated model filters matched the true filters , we represented the model filters as linear combinations of the true filters . The V1 neuron shown in Fig . 7 was recorded from an anesthetized macaque [57] . The stimuli ( refreshed at 100 Hz ) consisted of random arrays of black and white bars covering the neuron's classical receptive field , and oriented along its preferred orientation . Full experimental details can be found in [57] . Spatiotemporal filters were represented by 16 ‘pixels’ and 14 time lags . For model evaluation we used ten-fold cross-validation . Model fitting was analogous to that described for the V1 simulations ( Fig . 6 ) . STC-based models were constructed as described above for the simulated ON-OFF RGC . The V1 neuron shown in Fig . 8 was recorded from an anesthetized cat [91] . The stimuli consisted of natural and naturalistic movies at various contrasts , including noise processes with pink spatial and white temporal statistics , pink temporal and white spatial statistics , pink temporal and pink spatial statistics , and natural movies recorded with a ‘cat cam’ [100] . The mean luminance across all stimuli ( 15 different stimuli , each lasting 2 minutes ) was the same . The raw movies were 64×64 pixels and were sampled at 50 Hz . These raw movies were spatially down-sampled and cropped to produce 20×20 pixel patches that were individually mean-subtracted . Model performance was evaluated using five-fold cross-validation , with cross-validation sets constructed by taking 20% of the data from each stimulus type . For all analysis we used 8 time lags to construct spatiotemporal filters ( each described by 8×20×20 = 3200 parameters ) . For STA/STC analysis we first whitened the stimulus by rotating into the principal component axes and normalizing each dimension to have unit standard deviation [9] . Because the stimulus covariance matrix for natural stimuli has many eigenvalues close to zero , we avoided amplifying noise associated with these low-variance dimensions by using a pseudoinverse of the covariance matrix , effectively discarding the n lowest variance dimensions of the stimulus [9] , [95] . In addition to removing biases due to pairwise correlations in the stimulus , this method effectively imparts a prior favoring spatiotemporally smooth filters , since the lowest variance dimensions of natural stimuli have high spatial and temporal frequencies . We retained 500/3200 of the stimulus dimensions for STA/STC analysis . To estimate filters of the LN model , GQM and NIM , we used sparseness regularization , as well as penalties on the ( two-dimensional ) spatial Laplacian at each time lag . To display the three-dimensional spatiotemporal filters we plot the time slice of each filter containing the most variance across pixels ( ‘best time slice’ ) , as well as the projection of the filter onto a spatial axis orthogonal to the neuron's preferred orientation ( ‘space-time projection’ ) [62] . The preferred orientation was determined by fitting a two-dimensional Gabor function to the best time slice for each filter , and taking the ( circular ) average of the individual Gabor orientations across all filters .
Sensory neurons are capable of representing a wide array of computations on sensory stimuli . Such complex computations are thought to arise in large part from the accumulation of relatively simple nonlinear operations across the sensory processing hierarchies . However , models of sensory processing typically rely on mathematical approximations of the overall relationship between stimulus and response , such as linear or quadratic expansions , which can overlook critical elements of sensory computation and miss opportunities to reveal how the underlying inputs contribute to a neuron's response . Here we present a physiologically inspired nonlinear modeling framework , the ‘Nonlinear Input Model’ ( NIM ) , which instead assumes that neuronal computation can be approximated as a sum of excitatory and suppressive ‘neuronal inputs’ . We show that this structure is successful at explaining neuronal responses in a variety of sensory areas . Furthermore , model fitting can be guided by prior knowledge about the inputs to a given neuron , and its results can often suggest specific physiological predictions . We illustrate the advantages of the proposed model and demonstrate specific parameter estimation procedures using a range of example sensory neurons in both the visual and auditory systems .
[ "Abstract", "Introduction", "Results", "Methods" ]
[ "auditory", "system", "visual", "system", "mathematics", "computational", "neuroscience", "single", "neuron", "function", "statistics", "sensory", "systems", "biology", "neuroscience", "statistical", "methods" ]
2013
Inferring Nonlinear Neuronal Computation Based on Physiologically Plausible Inputs
Polymerase mu ( Polμ ) is an error-prone , DNA-directed DNA polymerase that participates in non-homologous end-joining ( NHEJ ) repair . In vivo , Polμ deficiency results in impaired Vκ-Jκ recombination and altered somatic hypermutation and centroblast development . In Polμ−/− mice , hematopoietic development was defective in several peripheral and bone marrow ( BM ) cell populations , with about a 40% decrease in BM cell number that affected several hematopoietic lineages . Hematopoietic progenitors were reduced both in number and in expansion potential . The observed phenotype correlates with a reduced efficiency in DNA double-strand break ( DSB ) repair in hematopoietic tissue . Whole-body γ-irradiation revealed that Polμ also plays a role in DSB repair in non-hematopoietic tissues . Our results show that Polμ function is required for physiological hematopoietic development with an important role in maintaining early progenitor cell homeostasis and genetic stability in hematopoietic and non-hematopoietic tissues . In higher eukaryotes , DNA double strand breaks ( DSB ) are repaired through two main pathways: homologous recombination [1] and non-homologous end-joining ( NHEJ ) ( reviewed [2]–[4] ) . Although NHEJ predominates during the G0/G1 and early S phases of the cell cycle [5] , [6] , both mechanisms can act in coordination [7]–[9] . Polymerase mu ( Polμ ) is an error-prone DNA-directed DNA polymerase belonging to the PolX family , with a high amino acid similarity to TdT [10]–[13] and , like TdT , Polμ has terminal transferase activity [11] . Unlike TdT , Polμ is expressed in many tissues ( liver , kidney , lung , brain , testis ) , although it is especially abundant in lymphohematopoietic organs [11] , [14] . Polμ interacts with components of the NHEJ repair pathway ( Ku 70/80 ) and is up-regulated after the induction of DSB by γ-irradiation , a known clastogen [15] . Double immunostaining for Polμ and phosphorylated γ-H2AX , a modified histone found at DSB sites , shows that Polμ is recruited to these sites [15] . In vitro , Polμ participates in specific NHEJ reactions that require DNA polymerase activity [15]–[17] . Analysis of a Polμ knockout mouse model has identified a specialized function for this enzyme during V ( D ) J recombination [18] . Specifically , Polμ is required for correct recombination of the immunoglobulin κ light chain during B cell development , and its deficiency results in shorter , non-productive Vκ-Jκ junctions and hence lymphocyte cell death at the transition from PreB to Immature B cell stage [18] . Because V ( D ) J recombination is essentially a modified end-joining reaction ( reviewed in [19] ) , the participation of Polμ in this process demonstrates its requirement for at least a subset of NHEJ reactions in vivo . This role of Polμ has also been demonstrated in vitro [16] . The expression pattern of Polμ [11] , [14] suggests that it might participate in NHEJ in other tissues , especially in non-B cell hematopoietic populations . In recent years it has been demonstrated that several DNA repair pathways play a clear role in the maintenance of the hematopoietic system and of hematopoietic stem cell ( HSC ) number and function during aging [20] , [21] . For example , in Ercc1-deficient mice hematopoiesis is impaired and bone marrow cellularity and hematopoietic progenitor cell ( HPC ) numbers are reduced [22]; in mice deficient in Brca2 , the reconstitution capacity of HSC and progenitor cells is reduced [23] . Hypomorphic Rad50 mutant mice have growth defects and are predisposed to cancer and progressive HSC failure [24] . Irradiated Parp−/− mice are myelosuppressed , display extensive hemorrhaging and show altered extramedullary hematopoiesis during the recovery phase [25] . Mice deficient either in XPD ( from the nucleotide excision repair pathway ) or Ku80 ( NHEJ ) show reduced HPC numbers in bone marrow . Although the numbers of HSC in these mice is unaltered , HSC reconstitution potential during aging is reduced [21] . A similar phenotype is observed in another mouse model of NHEJ deficiency ( Lig4Y288C mice; [20] ) , and together these reports strongly suggest that HSC accumulate mutations during aging that impair their function and highlight the role of the different repair pathways in maintaining hematopoietic homeostasis [20] , [21] . We have investigated the role of Polμ in hematopoiesis and general DNA repair . Polμ−/− mice have reduced numbers of most peripheral blood populations and show impaired hematopoiesis as a result of the reduced numbers of hematopoietic progenitors . Polμ−/− hematopoietic and non-hematopoietic cells accumulate DSB , both spontaneously and after γ-irradiation , indicating a role for Polμ in general DSB repair . The generation of Polμ−/− mice has been described previously [14] . Mutant and wildtype ( wt ) mice were bred in our specific pathogen-free facilities and were routinely screened for pathogens . Most experiments were carried out with animals in the mixed original ( 129/Balb/c ) background . Where indicated , experiments were carried out in the C57BL/6 background . B6 . SJL mice , were purchased from The Jackson Laboratory ( Bar Harbor , Maine ) , and housed at the CIEMAT animal facility . All experiments were performed according to Spanish and European regulations for the use and treatment of experimental animals , with the approval of the Centro Nacional de Biotecnología ( CNB ) and the Fundación Centro Nacional de Investigaciones Cardiovasculares ( CNIC ) animal ethics committees . Femurs and tibias were removed from mice , and bone marrow was extracted by complete flushing with PBS under sterile conditions . Spleens were retrieved and disaggregated in sterile PBS . In each case , tissue extracts were incubated in 0 . 85% NH4Cl to lyse erythrocytes . Cells were counted with a hemocytometer and used in subsequent assays . Blood was collected in a capillary tube from the retro-orbital sinus and transferred to EDTA-coated tubes . For hematology , 25 µl were analyzed in an MS9 machine ( Kemia , Madrid , Spain ) . For flow cytometry ( FCM ) , whole blood was incubated with the corresponding antibodies , fixed , and erythrocytes lysed with OptiLyse C ( Immunotech ) . Cells were analyzed in a Cytomics FC 500 cytometer ( Beckman Coulter ) . Bone marrow cells were stained with antibodies in ice-cold PBS , 0 . 5% BSA , 2 mM EDTA . Antibodies used were anti-CD19-PE , -CD19-SPRD , -CD3-FITC , -CD4-SPRD , -CD8-PE , -CD11b ( Mac1 ) -PE , -Gr1-FITC , -B220-FITC , -B220-PE , -IgM-PE , -Ter119-PE , -CD41-FITC , ( BD-Pharmingen ) . Cell cycle status was monitored by estimating total cell DNA content in a flow cytometer after staining 70% methanol fixed cells with 10 µg/ml propidium iodide and 10 µg/ml RNase A . For bone marrow progenitor analysis , bone marrow cells were extracted from femurs and tibias , counted , and stained in ice-cold PBS , 0 . 5% BSA plus 2 mM EDTA . Except for samples for detection of CMPs , GMPs and MEPs , samples were blocked with FcBlock ( anti-CD16/CD32 , BD Pharmingen ) . Samples were stained with the following antibodies: biotin-conjugated lineage antibody cocktail ( anti-CD3e , -CD11b , -B220 , -Ly6G/Ly6C and -TER-119 ) , anti-CD117 ( c-kit ) ( FITC or PE ) , anti-CD16/32-FITC , anti-CD135-PE , anti-Sca1-PE-Cy7 ( all from BD Pharmingen ) ; anti-CD127-Biotin , anti-CD127-PE ( eBiosciences ) ; anti-CD34-PE-Cy5 ( Biolegend ) , and Streptavidin-Pacific Blue ( Invitrogen ) . To exclude dead cells Hoechst 33582 was added to every sample . All samples were acquired with a BDFASCanto II cytometer ( BDBiosciences ) . At least 1 million cells were acquired for each sample . Data were analyzed with FlowJo software ( Tree Star , Inc . ) and fluorescent-minus-one controls were used to define gates [26] . For histology , femurs or tibias were formalin-fixed , decalcified and paraffin-embedded . Sections ( 5 µm ) were hematoxylin-eosin stained . Photographs were acquired with a Leica DM RB microscope fitted with a DP 70 digital camera ( Olympus ) and linked to DP controller software ( Olympus ) . Splenocytes from wt or Polμ−/− animals were cultured in RPMI 1640 medium ( Biowhitaker ) supplemented with 10% FCS , non-essential amino-acids , 2 mM L-glutamine , 1 mM sodium pyruvate , 10 U/mL penicillin/estreptomycin , 10 µg/ml LPS ( Sigma ) and 10 ng/ml IL-4 ( Preprotek Inc . ) . After 72 hours , cultures were gamma irradiated at 8 Gy . Cells were collected after 1 , 3 and 6 hours; cell were collected by centrifugation and processed for protein extraction . To evaluate proliferation during stimulation with SCF and IL-3 , bone marrow cells ( 106 ) were seeded in Iscove's modified Dulbecco medium ( IMDM ) supplemented with 10% fetal calf serum ( Gibco-BRL ) , 10% conditioned medium from IL-3-producing WEHI-3B cells , and 50 ng/ml human SCF ( Stem Cell Technologies ) . For long-term bone marrow cultures ( LTBMC ) , whole bone marrow suspensions were seeded in Methocult 5300 ( Stem Cell Technologies ) with 10−6 M hydrocortisone ( Sigma ) , and incubated ( 33°C , 5% CO2 ) for 1–4 weeks as indicated , with a weekly culture medium change . Mouse embryonic fibroblasts ( MEF ) were grown in Dulbecco's Modified Eagle's Medium ( DMEM ) supplemented with 10% FCS , 2 mM L-glutamine , 10 mM HEPES , 1 mM sodium pyruvate , non-essential amino acids , and 50 µg/ml gentamycin ( all from BioWhittaker ) . Cell cultures were maintained at 37°C , 95% humidity and 5% CO2 . For MEF growth assays , embryos were minced with a razor blade and incubated in 0 . 25% trypsin , 0 . 1% EDTA for 30 min at 37°C . The cell suspension was plated in DMEM , 10% FCS containing antibiotics . Confluent cultures were split and passage cultures plated at 1×106 cells on 10-cm tissue culture plates ( Falcon ) . Cells were trypsin passaged every 3–4 days . Cumulative cell growth was calculated with the formula PD = log ( nf−ni ) log2 , where PD is population doubling and ni and nf are the initial and final number of cells . Cells were cultured in the appropriate differentiation medium: Methocult M3630 for PreB colonies; M3534 for CFU-G+M colonies; and M3231 with 10% WEHI-3B-conditioned medium , SCF ( 50 ng/ml ) and erythropoietin ( 6 U/ml ) for BFU-E colonies ( all from Stem Cell Technologies ) . Colonies were counted 7 days after seeding . When indicated , colony area was measured by analysis of photomicrographs with ImageJ ( http://rsb . info . nih . gov/ij/ ) . Bone marrow mononuclear cells ( BMNC ) were recovered from male wt or Polμ−/− C57BL/6 mice ( which express the CD45 . 2 surface antigen in hematopoietic cells ) as described above . Cells were prepared from four animals per genotype . The Polμ−/− or wt cells ( 2×106 ) were mixed with 2×106 competitor bone marrow mononuclear cells from female B6 . SJL mice ( which express the CD45 . 1 surface antigen in hematopoietic cells ) . The mix was injected into the tail veins of female previously irradiated B6 . SJL recipients ( see below ) . Repopulation was monitored over 4 months post-transplant by FCM to detect the percentage of CD45 . 1+ and CD45 . 2+ cells in peripheral blood . Four months post-transplant , animals were sacrificed and the relative population of CD45 . 1+ and CD45 . 2+ cells was analyzed in all hematopoietic organs . In parallel , the number of male cells was assessed by Y-chromosome FISH [27]- and Y chromosome-specific PCR [28] . The number of repopulating units ( RU ) was determined as described [29] . Briefly , number of donor RU = ( percentage of donor cells×number of competitor RU ) / ( 100-percent of donor cells ) . We have assumed that 1×105 competitor cells contain 1 RU . Comet assays were performed with the Cometassay kit ( Trevigen ) . Comet images were obtained for at least 50 cells per condition and analyzed with TriTek CometScore ( http://www . tritekcorp . com ) . For γ-irradiation survival assays of myeloid clonogenic forming units ( CFU-C ) , bone marrow cells were seeded as for clonogenic assays and immediately irradiated ( 0–8 Gy ) in a Cesium Mark1 irradiator ( Shepherd Associates ) . The number of surviving CFU-C colonies was scored 7 days post-irradiation . For γ-irradiation survival assays with MEF , 3×102 cells were seeded in 10 cm2 culture plates ( Corning Inc . ) and irradiated ( 0–8 Gy ) after 24 hours; Crystal Violet stained colonies were counted after 7 days . To evaluate differential sensitivity to irradiation , mice were exposed to γ-irradiation ( 4–10 Gy , single dose ) and closely monitored throughout the experimentation period . Animals were killed at the first appearance of signs of poor health . Recipients of bone marrow transplants were pre-irradiated with a total dose of 10 . 2 Gy ( two doses of 5 . 1 Gy , 3 h apart ) . For in vivo DNA damage experiments , mice were irradiated with a single dose of 5Gy and killed after 1 , 3 or 6 hours . All irradiations were carried out in a Cesium Mark1 irradiator ( Shepherd Associates ) . Cell suspensions were deposited on superfrost slides ( Menzel-Glaser ) for 5 min . Adhered cells were fixed in 4% paraformaldehyde ( 20 min , 4°C ) , washed three times in PBS containing 0 . 1% Tween ( PBST ) and blocked ( 1 h ) in PBST/10% BSA . Preparations were incubated overnight with anti-phosphorylated γ-H2AX antibody ( Upstate ) in PBST/1% BSA . After washing with PBST and staining with secondary antibody ( 40 min ) , preparations were washed and mounted , with DAPI , in VectaShield mounting medium ( Vector ) . Confocal images were acquired on a Fluoview FV1000 Olympus microscope using Fluoview version 1 . 4 acquisition software . TIFF images were analyzed with Image J . Spleen and bone marrow populations were scored for the number of γ-H2AX positive cells and the number of γ-H2AX foci per nuclei . Nine-month-old wt or Polμ−/− mice were irradiated ( 5 Gy ) and bone marrow cell suspensions prepared after a 6 h recovery period , to allow time for in vivo DNA repair . Cells ( 2×106/ml ) were cultured in Myelocult 5300 medium ( Stem Cell ) supplemented with 20% FCS , 10% WEHI cell-conditioned medium ( IL-3 ) and antibiotics . Metaphase spreads were prepared four days after irradiation . For telomere in situ hybridization ( Tel-FISH ) , metaphase cells were hybridized with a telomeric Cy3- or FITC-labeled PNA- ( CCCTAA ) 4 probe ( Applied Biosystems ) essentially as described [30] , except that post-hybridization washes ( 3×10 min ) were performed in PBST at 50°C . FISH images were captured with a Nikon 80I microscope fitted with a 100×1 . 3 NA planfluor objective and an Olympus DP digital camera . Between 40 and 80 metaphase spreads were scored for chromosomal aberrations ( chromosome and chromatid breaks , dicentrics , rings and Robertsonian-like chromosomes: see examples in Figure S5 ) . For non-histone proteins , extracts were prepared with RIPA lysis buffer ( 150 mM NaCl , 1% NP-40 , 0 . 5% deoxycholate , 0 . 1% SDS , 50 mM Tris , pH 8 . 0 ) . Acidic extraction of histones was carried out as follows . Tissue or pelleted cells were mechanically disaggregated in 0 . 25 M HCl , incubated at 4°C with agitation for 14 h , centrifuged ( 13000 rpm , 10 min , 4°C ) , and further incubated for 4 h in 8 volumes of acetone . After centrifugation ( 5000 rpm , 5 min ) , pellets were washed with acetone and dried in a SpeedVac . Extracts were incubated in 0 . 25 M HCl for 2 hours at 4°C and centrifuged at 13000 rpm prior to quantification and loading on SDS-PAGE gels . After transfer to PDVF membranes , blots were probed with the following antibodies: mouse monoclonal anti-p21 ( Santa Cruz Biotechnology , sc-6246 , 1∶1000 ) , mouse monoclonal anti-beta actin ( Abcam , ab8226-100 , 1∶5000 ) , rabbit polyclonal anti-H2AX-P ( Upstate , 07-164 , 1∶1000 ) and rabbit polyclonal anti-H3 ( Abcam , ab8226-100 , 1∶5000 ) . Wt or Polμ−/− bone marrow cells were irradiated at 4 or 8 Gy . Cells ( 1×106 ) were incubated ( 1 h ) with 5 µM DCFDA ( Molecular probes ) in reduced-serum OPTIMEM ( Gibco ) and analyzed by FCM . DCFDA fluorescence was analyzed by measuring the mean Fl-1 fluorescence intensity of at least 10000 cells from three animals . Unless indicated all analyses were by Student's t test . *:p<0 . 05; **:p<0 . 01; ***:p<0 . 001 . Polμ is widely expressed , but is especially abundant in lympohematopoietic tissues and cells [11] , [14] . Polμ is known to participate in DSB repair reactions in vitro [15]–[17] , but the only in vivo role assigned so far is in Vκ-Jκ recombination during B cell development [18] . To investigate whether Polμ has a function in other hematopoietic lineages , we first studied peripheral blood ( PB ) cell populations in Polμ−/− mice . Concurring with the reduction in B cell numbers associated with abnormal Vκ-Jκ recombination , [18] , the B cell count in Polμ−/− mice was 4 . 9-fold ( p<0 . 001 ) lower than in wt controls ( Figure 1A; CD19+ cells ) . Monocyte numbers were also reduced ( 2 . 2-fold , p<0 . 01 , Figure 1A; Mac1+ Gr1− cells ) , as were neutrophil numbers , though here the difference was not statistically significant ( p<0 . 09 , Figure 1A; Mac1+Gr1+ cells ) . The only cell type analyzed that was apparently unaffected was the T cell lineage ( CD3+ cells ) . Hematological analysis revealed moderate thrombocytopenia , with 403×109 platelets/ml in Polμ−/− mice , compared with 663×109 platelets/ml in wt counterparts ( p<0 . 01 , Figure 1B ) ; this was associated with increased bleeding times ( Figure S1 ) Thus peripheral blood populations other than B lymphocytes are altered in Polμ−/− mice . To assess whether the defective blood population profile originated in bone marrow ( BM ) , we compared BM cell numbers between wt and Polμ−/− mice . Our results show that BM cellularity was 1 . 5-fold lower in Polμ−/− mice ( p<0 . 001 , Figure 1C ) . Immunohistochemistry confirmed that Polμ−/− BM contained fewer cells than wt BM , as shown by the dilated endothelial bone marrow sinusoids ( Figure 1D ) . Therefore the most likely cause of the thrombocytopenia and reduced myeloid cell numbers in peripheral blood in Polμ−/− mice is a defect in BM hematopoiesis . To investigate the function of Polμ during hematopoiesis , we analyzed wt and Polμ−/− ( C57BL/6 ) BM cells by flow cytometry ( FCM ) . The B lymphoid and myeloid compartments were significantly reduced in Polμ−/− BM ( p<0 . 05 ) , and there was a more moderate reduction in erythroid cell numbers ( Figure 1E ) . To determine whether the reduction in B cell numbers was exclusively due to the previously described Vκ-Jκ defect [18] , we used FCM to analyze B cell development in the bone marrow of Polμ−/− or wt C57BL/6 mice ( see Figure 1F for representative plots and gating strategies ) . With the exception of PreB cells ( B220+IgM−CD43−CD25+ cells ) , all BM B cell populations were significantly reduced in Polμ−/− BM . PreB cell frequency was increased in Polμ−/− BM ( Figure 1H ) , indicating an accumulation at the PreB stage , probably due to a blockade at the PreB-to-immature ( IgM+ ) cell transition . Similar results were obtained with Polμ−/− animals in the mixed background ( not shown ) . These results suggest that low B cell numbers in Polμ−/− mice are not only the result of deficient Vκ-Jκ recombination but also reflect altered hematopoiesis prior the specific defect during Vκ-Jκ recombination described previously [18] , suggesting that Polμ deficiency affects most hematopoietic lineages in bone marrow . Hematopoiesis is a clonal process in which immature progenitors give rise to more committed progenitors . We therefore used colony forming unit ( CFU ) assays to determine total progenitor numbers in wt and Polμ−/− BM ( Figure 2A–D ) . The numbers of PreB and myeloid progenitors ( CFU-PreB and CFU-C ) in Polμ−/−mice were 2 . 2- and 1 . 3-fold lower than in wt ( p<0 . 001 , Figure 2A , B ) . Megakaryocytic and erythroid progenitors ( CFU-Mk and BFU-E ) were also affected , though to a lesser extent ( Figure 2C , D ) . Importantly , whereas granulomonocyte and monocyte progenitors ( CFU-GM and CFU-M ) were reduced in Polμ−/− BM ( correlating with the peripheral blood monocyte deficiency ) , granulocyte progenitors ( CFU-G ) were not , thus explaining the absence of neutropenia ( Figure 1A ) . These results show that Polμ deficient mice have lowered numbers of committed progenitors and suggest that lack of Polμ differentially affects lineage-specific progenitors . To examine this in more depth , we next quantified hematopoietic lineage precursors by FCM . The populations measured were common lymphoid progenitors ( CLP ) [31] , defined as Lin−IL7R−Sca1+c-kit+ cells , ( Figure 2E ) ; common myeloid progenitors ( CMP ) , myelomonocytic progenitors ( GMP ) and megakaryocyte/erythroid progenitors ( MEP ) [32] ( Figure 2F ) ; hematopoietic stem cells were defined as Flt3− cells after gating on Lineage−c-kit+Sca1+ [33] ( Figure 2G ) . Although FCM analysis revealed that the frequency of these populations was little altered in Polμ−/− mice ( Table S1 ) , myeloid progenitors and stem cells were less abundant in the BM of Polμ−/− mice and some populations were more severely affected ( Figure 2E–G ) ; for example , HSC numbers were 39% lower in Polμ−/− BM ( p<0 . 01 ) . We then analyzed engraftment potential of Polμ−/− HSC in a competitive repopulation study [29] . In this experiment 2×106 BMNC ( bone marrow nucleated cells ) from male wt or Polμ−/− C57BL/6 mice were mixed with 2×106 competing cells from female B6 . SJL mice ( 1∶1 ) , and the mixed population transplanted into lethally irradiated female B6 . SJL recipients ( see Figure 2H for experimental design ) . The percentage of donor and competitor cells was determined 16 weeks later and used to calculate the number of repopulating units ( RU ) in the donor BM . No differences were detected in either the frequency of ( Figure 2I ) or absolute ( Figure 2J ) RU numbers between wt and Polμ−/− bone marrow . These results indicate that although Polμ deficiency reduces the numbers of most hematopoietic progenitors and stem cell populations the defect can be rescued ( at least partially ) , by transplantation into a healthy stroma ( Figure 2I ) . Additionally , clonogenic analysis detected very few progenitors in Polμ−/− spleen ( Figure S2 ) , indicating that Polμ−/− mice do not compensate for the BM deficit with extramedullary hematopoiesis . The low progenitor numbers in Polμ−/− mice could result either from impaired differentiation or from defective progenitor proliferation or self renewal capacity . To investigate this , we examined the proliferation of Polμ−/− hematopoietic stem and progenitor cells in vitro . In clonogenic differentiation assays , colonies of Polμ−/− progenitors were significantly smaller ( 50% smaller for CFU-C , p<0 . 05; 60% smaller for CFU-PreB , p<0 . 05 ) and contained fewer cells than wt colonies , indicating defective self-renewal or proliferative capacity ( Figure 3A ) . This was confirmed by growing BMNC in conditions that promote myeloid progenitor expansion ( see methods ) . After 4 days , wt BMNC cultures contained 2 . 3-fold more myeloid progenitors ( p<0 . 001 ) and 2 . 8-fold more erythroid progenitors ( p<0 . 01 ) than Polμ−/− cultures ( Figure 3B ) . We then assayed BMNC in long-term bone marrow cell cultures ( LTBMC , see methods ) . Two weeks after initiation , Polμ−/− cultures contained 35% fewer cells than wt ( p<0 . 01 , Figure 3C ) . Cell cycle analysis revealed an increased incidence of cell death in Polμ−/− cultures ( Figure S3 ) . Together with the bone marrow reconstitution experiments , these results suggest that Polμ−/− bone marrow stromal cells fail to efficiently support hematopoiesis and that Polμ−/− hematopoietic progenitor numbers do not expand in vitro as efficiently as WT progenitors . One possible cause of reduced Polμ−/− hematopoietic precursors numbers is impaired DSB repair . To test this , we stained nucleated cells from bone marrow and spleen for phosphorylated γ-H2AX , a marker of DNA double strand breaks [34] . Compared with wt , Polμ−/− bone marrow had a significantly higher number of DSB per nucleus ( 3 . 8 fold , p<0 . 05; Figure 4 A , B ) ; in Polμ−/− spleen cells ( mostly non-proliferating ) a smaller increase was observed that did not reach significance ( Figure 4A ) . This suggests that Polμ is required for DSB repair in cycling cells . We confirmed the higher DSB incidence in Polμ−/− bone marrow cells by comet assay , which revealed significantly longer comet tails in Polμ−/− BM cells ( p<0 . 001; Figure 4C , D ) . Mouse embryonic fibroblasts ( MEF ) derived from mice deficient in NHEJ proteins show a highly reduced proliferative capacity compared to wt controls [35]–[37] . To assess the roles of Polμ in DNA repair and proliferation in other tissues , we analyzed wt and Polμ−/− MEF . Consistent with the proposed role of Polμ in NHEJ , our results show that Polμ−/− MEF showed reduced growth after long-term culture and they enter senescence prematurely ( Figure 4E ) . In addition , primary Polμ−/− MEF were genomically unstable compared with wt MEF . In particular , Polμ−/− MEF showed a 4-fold increase in chromosomal aberrations , with a striking 16-fold increase in the level of radial configurations and a 4 . 5-fold increase in chromosomal breaks , signatures of deficient DNA repair and increased radiosensitivity ( Table 1 and Figure 4F; see Figure S5 for examples of chromosome aberrations ) . These results demonstrate that lack of Polμ results in DSB accumulation in bone marrow and probably in connective tissue . This suggests that proliferation deficiency by Polμ−/− cells is due to accumulation of unrepaired DSB or a general delay in DSB repair , which may lead to cell cycle arrest and cell death [38] . To test this in vivo , we irradiated Polμ−/− and wt mice ( 9 Gy ) and analyzed survival over time . Fifteen days post-irradiation all Polμ−/− mice were dead , compared with only 40% of wt mice ( p<0 . 001 Log rank test; Figure 5A ) . Most Polμ−/− mice died between days 9 and 11 ( Figure 5A ) , and hematologic analysis revealed extreme neutropenia ( not shown ) , indicating hematopoietic failure . The impact of Polμ deficiency on the response to irradiation damage is also illustrated by the higher radiosensitivity of Polμ−/− progenitors ( Figure 5B; p<0 . 01 ) . Bone marrow cells from wt and Polμ−/− mice showed G2 accumulation after irradiation ( 5Gy ) , suggesting that the G2/M cell cycle checkpoint is functional in Polμ−/− cells ( Figure S4A ) . Western blot detected increased p21 accumulation in irradiated Polμ−/− splenocytes and even in non-irradiated cells ( Figure S4B ) , suggesting activation of the G1 cell cycle checkpoint . Irradiation of Polμ−/− MEF ( 2–8 Gy ) significantly reduced survival compared with similarly treated wt MEF ( 2 to 123-fold; p<0 . 05 . Figure 5C ) . Histological analysis of irradiated mice showed vacuolar degeneration in liver ( where we detect intense γ-H2AX accumulation after irradiation; Figure S4C ) , hemorrhage and congestion in the lung , tubular degeneration in kidney , and tubular atrophy in testis ( Figure 5D ) . An alternative explanation for increased tissue damage in irradiated Polμ−/− mice might be increased production of reactive oxygen species ( ROS ) . We therefore measured intracellular ROS levels in irradiated bone marrow cells ( 4 and 8 Gy ) by staining with DCFDA ( Di-cloro-fluorescein diacetate: a marker of intracellular peroxides ) . ROS accumulation was unaffected by Polμ−/− deficiency ( Figure 5E ) , indicating that ROS production does not contribute to the radiosensitivity of Polμ−/− animals or cells . These data thus show that the requirement for Polμ in DNA repair extends to tissues outside the hematopoietic compartment ( Figure 5C , D ) . To examine the effect of Polμ deficiency on DNA repair in the hematopoietic system more closely , we analyzed the frequency of phosphorylated γ-H2AX foci in bone marrow and spleen after whole animal γ-irradiation ( 5Gy ) . In both genotypes 91% of all BM cells were γ-H2AX+ , showing that these cells are more susceptible to γ-irradiation . In contrast , the proportion of phospho-γ-H2AX positive cells in Polμ−/− splenocytes ( 95% ) was significantly higher than in wt cells ( 27%; p<0 . 001 ) . Further , the number of γ-H2AX foci per cell was significantly increased in Polμ−/− cells compared with wt ( 1 . 7-fold for BM cells and 4 . 8-fold for splenocytes , p<0 . 001; Figure 6A–B ) . Changes in γ-H2AX phosphorylation were confirmed by western blot of γ-irradiated ( 8 Gy ) spleen B cells; as predicted , Polμ−/− cells show higher amounts of phospho-γ-H2AX , and these levels are sustained for longer ( Figure 6C ) . We confirmed reduced DSB repair in Polμ−/− BM cells by comet assay . Before cell extraction , irradiated mice ( 5Gy ) were left to recover for 3 hours to allow in vivo DNA repair ( see Methods ) . BM cells from these mice were separated by single cell electrophoresis under alkaline conditions , to measure the relative levels of DNA breaks; comet tail moment was 2-fold higher in Polμ−/− cells compared with wt cells ( p<0 . 001; Figure 6D ) , indicating reduced DNA repair in Polμ−/− cells . The effect of the impaired DNA repair on genetic stability was investigated by telomere-directed fluorescent in situ hybridization ( Tel-FISH ) and DAPI staining of control or irradiated ( 5Gy ) wt and Polμ−/− BM metaphase cells . Figure 6E shows representative results from irradiated cells . These experiments detected a higher level of genomic instability in Polμ−/− cells both before and after irradiation . Non-irradiated Polμ−/− samples demonstrated only a 0 . 5-fold increase in chromosome breaks compared with controls ( 32 . 6 vs . 22 breaks per 100 cells ) . After irradiation , the total number of aberrations was 3-fold higher in Polμ−/− cells ( 1101 vs . 364 per 100 cells in Polμ−/− and wt; ( p<0 . 0003 ) . Compared with wt , chromosome breaks in irradiated Polμ−/− cells were increased 2 . 7 fold ( p<0 . 0014 ) , radial configurations 4 . 8 fold ( p<0 . 05 ) , and end-to-end fusions ( dicentrics , Robertsonial-like fusions and rings ) 3 . 5 fold ( p<0 . 0003 ) ( Figure 6F , Table 1 ) . Supplemental Figure 5 shows example images of the chromosomal alterations analyzed . These results confirm that Polμ is required for DSB repair in vivo in hematopoietic and non-hematopoietic systems , particularly in response to genotoxic stress . Polμ is a gap filling , error-prone DNA polymerase that , in association with the NHEJ core machinery , promotes microhomology use during DSB repair [17] , [39] . Polμ deficiency does not affect mouse development or generate a severe phenotype in adults [14] . This is probably because its function can be replaced or compensated , at least in part , by other DNA polymerase activities [14] , [17] . The most prominent phenotype reported to date in Polμ−/− mice is partial impairment of immunoglobulin Vκ-Jκ recombination [18] , associated with alterations in the peripheral B cell compartment [18] . Polμ−/− mice are nonetheless able to mount an almost normal humoral immune response [14] , [40] . Here we show that Polμ is required for DSB repair in vivo and that this defect confers radiosensitivity , promotes genetic instability and leads to deficient hematopoiesis . Polμ deficiency alters many peripheral blood cell populations , affecting not only the B cell compartment [14] , but also the myeloid lineages ( Figure 1A ) . Polμ−/− bone marrow showed a severe reduction ( ∼40% ) in total bone marrow cell numbers ( Figure 1C , D ) . This reduction affects all hematopoietic lineages as demonstrated by flow cytometry ( Figure 1 ) . Interestingly , B cell development is affected prior Vκ-Jκ recombination ( Figure 1F–H ) . All committed hematopoietic progenitors were reduced in Polμ−/− BM; the hematopoietic defect was traced back to the hematopoietic stem cell , since FCM revealed a reduction in the number of HSC ( Figure 2G ) . Interestingly , transplant of Polμ−/− bone marrow ( together with competitor WT cells ) into irradiated recipients rescued ( at least partially ) the hematopoietic defect , suggesting that Polμ−/− stroma does not completely support hematopoiesis , alternatively , the co-transplantation of hematopoietic wt cells might provide factors that support the development of Polμ−/− HSC . Polμ−/− progenitors do not expand in vitro as efficiently as WT controls , suggesting that lack of Polμ results in cell death or reduced proliferation . Taken together these results suggest that hematopoietic cells in the bone marrow of Polμ−/− mice are reduced due to a combination of extrinsic ( defective bone marrow stroma ) and intrinsic ( reduced DNA repair and expansion potential ) factors . We demonstrate that hematopoietic Polμ−/− cells contain more DSB than their wt counterparts , as shown by the increase in the basal level of phospho-γ-H2AX positive cells and comet assay results ( Figure 4A–D ) and increased basal chromosomal breakage ( Table 1 ) . The reduced DSB repair capacity of hematopoietic Polμ−/− cells is more evident after genotoxic insult , demonstrating increased genomic instability ( Figure 6; Table 1 ) . Our results also show that Polμ is required for DNA repair in other tissues; including connective tissue ( MEF ) ( Table 1 and Figure 4E–F ) , liver and kidney ( Figure 5B ) . Lack of Polμ results in radiation-induced cell death at doses that are non-lethal for wt mice ( Figure 5A ) . However , immunohistochemistry failed to demonstrate any gross tissue or organ damage outside the hematopoietic system in non-irradiated Polμ−/− mice ( not shown ) . This suggests that Polμ deficiency may be compensated for or tolerated in organs in which progenitor cell proliferation is slower than in the hematopoietic system or in MEF . Alternatively , it may indicate that - in some tissues - Polμ function is only required under stress conditions ( for example , irradiation ) , which require faster repair . This is further supported by the fact that although most hematopoietic lineages are reduced in Polμ−/− BM , different subsets are affected to different extents . For example although CFU-GM and CFU-M progenitors are reduced in Polμ−/− BM , CFU-G progenitors are not . Another example is the reduction of the HSC population by 39% ( p<0 . 001 ) , while the CLP is unaffected . Whether the less-affected cell types have less DNA damage or use different DNA repair mechanisms remains to be investigated . To our knowledge this is the first report to show increased radiosensivity in Polμ deficient mice and cells . A previous study reported a relative lack of radiosensivity in Polμ−/− MEF [18] . This contrasts with our data , which strongly suggest that Polμ is necessary for the repair of double strand breaks . Polμ-deficient MEF showed reduced growth curves and increased basal and radiation-induced levels of chromosomal aberrations; phenotypes similar to those previously reported for Ku80−/− and LigIV−/− MEF [30] , [35] , [36] , [37] , [41]–[45] . Apart from the different genetic background of the two Polμ−/− mouse models , there are other possible explanations for this discrepancy . First , the gene targeting strategy is different between the two Polμ−/− mouse models . In our model , exons 2 to 4 are deleted , which eliminates the BRCT domain and completely abrogates protein expression in Western blot analysis [14] . In the knockout generated by Bertocci and coworkers , exons 6 to 11 are deleted , eliminating the core of the polymerase activity . Although these mice do not express full-length Polμ mRNA , no information is provided regarding protein expression [18] , [40] . This is important , since a truncated Polμ BRCT domain with biological activity may be expressed in this mouse model . An alternative explanation is that the two studies use different experimental procedures to detect radiosensitivity . We γ-irradiated MEF cultures plated at clonogenic dilutions ( 300 cells in 10 cm2 dishes , see Methods ) and calculated survival by scoring colonies . In contrast , [18] performed cell survival assays at much higher cell densities and used X-rays as the agent of DNA damage; survival was assessed by trypan blue exclusion . Differences in the sensitivities of these approaches might account for the differences observed in MEF survival in the two models . The role of DNA repair in hematopoietic homeostasis has become increasingly clear in recent years [20]–[23] . Knockout models of many DNA repair enzymes are characterized by either reduced stem cell function or alterations in specific hematopoietic populations . Moreover , since stem cells are protected from apoptosis [20] , [21] , probably because of their quiescent state [46] , deficiency in DNA repair enzymes allows mutations to accumulate in these cells . Conversely , more committed progenitors , which cycle more rapidly , would be more sensitive to cell cycle arrest or apoptosis , thus explaining the observed reduction in the numbers of progenitor cells in Polμ-deficient mice . Thus the survival of HSC comes at the price of accumulated mutations; two recent reports show that mice deficient in DNA repair maintain expanded stem cell pools , but that these pools have a reduced differentiation potential during aging [20] , [21] . Polμ deficiency , as reported for other NHEJ factors such as Ku80 [21] and Ligase IV [20] , targets the hematopoietic system and HSC . There are , however , striking differences in the phenotypes observed in Ku80 and LigaseIV mice and Polμ−/− mice . In the first two cases , HSC were unaffected in younger mice but were progressively exhausted during aging . Polμ unlike Ku80 or Ligase IV only participates in a small number of NHEJ reactions that require an error-prone DNA polymerase to generate or promote microhomology pairing between the DNA ends of the DSB [15]–[17] and Polμ−/− mouse has reduced HSC numbers in the bone marrow . These suggest that different components of the NHEJ system will have different roles in hematopoiesis or that some of the DSB generated in hematopoietic cells require Polμ for their complete and efficient repair , and cannot be repaired by compensatory mechanisms with sufficient speed or efficiency to prevent cell death or premature senescence . Additionally , the whole body irradiation experiments presented here show that DSB-repair defects and cell death in Polμ−/− mice are not restricted to the hematopoietic system but are widespread , demonstrating a role for Polμ in DSB repair in other tissues . Our results thus show that Polμ is required for genomic stability and DNA repair , through its participation in DSB repair in hematopoietic stem and progenitor cells as well as in non-hematopoietic cells .
Double-strand breaks ( DSB ) in DNA are a highly deleterious type of genetic damage , potentially causing genomic rearrangements or cell death if unrepaired . DSB can be triggered by environmental factors ( such as electromagnetic radiation or clastogenic chemicals ) or normal cell metabolism . The main mechanism of DSB repair in mammals is thought to be the non-homologous end-joining ( NHEJ ) pathway . Our article describes how DNA polymerase mu ( Polμ ) , a recently identified component of the NHEJ machinery , is required for hematopoiesis—the process that generates and maintains the correct balance of the millions of blood cells needed to sustain life and defend against infection . Hematopoietic stem cells ( HSC ) divide asymmetrically , yielding another HSC and a progenitor cell . These progenitors proliferate and differentiate , their progeny eventually generating mature blood cells . In mice in which Polμ is genetically eliminated , we found that hematopoietic progenitors proliferate slowly and are functionally impaired . The incidence of DSB in hematopoietic cells from these mice is increased , suggesting that reduced DNA repair may be the cause of the hematopoietic defects . DNA damage was also increased in tissues unrelated to hematopoiesis , including liver , kidney , lung , and mouse embryonic fibroblasts . Thus , these results demonstrate that Polμ plays an important role in general DSB repair in many cell lineages .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results", "Discussion" ]
[ "molecular", "biology/dna", "repair" ]
2009
Altered Hematopoiesis in Mice Lacking DNA Polymerase μ Is Due to Inefficient Double-Strand Break Repair
The spleen is one of the main affected organs in canine visceral leishmaniasis ( CVL ) . Disorganization of the splenic white pulp ( SWP ) has been associated with immunosuppression and disease progression . This study aims to assess structural and cellular changes in the splenic extracellular matrix of dogs with CVL , correlating these changes with the parasite load and clinical signs . Splenic fragments were collected from 41 naturally infected animals for parasite load quantification by quantitative PCR , histopathological analysis and immunohistochemistry for CD3+ , CD4+ , and CD8+ T cells; CD21+ B cells; Ki-67+ , IFN-γ+ , and IL-10+ cells; and the MMP-9 and ADAM-10 enzymes . Laminin , collagen and fibronectin deposition were also evaluated . The animals were grouped according to the level of SWP organization . SWP disorganization was accompanied by a reduction in the quantity of lymphoid follicles/mm2 ( p > 0 . 0001 ) . Animals with moderate to intense SWP disorganization showed more clinical signs ( p = 0 . 021 ) , higher laminin ( p = 0 . 045 ) and collagen deposition ( p = 0 . 036 ) , higher MMP-9 expression ( p = 0 . 035 ) and lower numbers of CD4+ T cells ( p = 0 . 027 ) in the spleen than the animals with organized SWP . These data suggest that splenic structure and function are drastically altered and compromised during CVL . In Brazil , visceral leishmaniasis is caused by Leishmania infantum , and the domestic dog is the main urban reservoir[1] . Additionally , dogs have been used as models for the study of human disease[2] because disease progression in dogs is similar to that in humans[3] . Concerning the immune response in canine visceral leishmaniasis ( CVL ) , susceptible animals have a remarkably attenuated humoral and cellular immune response against the parasite , resulting in the appearance of clinical signs[2] . Although the cytokine expression profile has also been associated with disease resistance or progression to CVL[4] , the data regarding cytokine expression in the spleen remain controversial[1 , 5 , 6] . The spleen is the key organ responsible for removing damaged and senescent cells from blood circulation . The adaptive immune responses against the pathogens captured from the blood begin in this organ[7] . The composition of the spleen is based on two structures: the red pulp and the white pulp , an area rich in T and B lymphocytes . Additionally , cells that are highly specialized in antigenic presentation are found in this organ . The white pulp exhibits an organization in which it is wrapped around central arterioles , which are branches of the artery[7] . Multiple molecules , namely , laminin , fibronectin and collagen , are important proteins that constitute the basal and interstitial membranes of this organ and are fundamental to the structuring and maintenance of the white pulp , contributing to the upkeep of cellular compartmentalization . These proteins also contribute to the cellular communication and transport of cytokines and chemokines through structures known as sinusoids and conduits , which comprise type IV collagen , laminin and reticular fibroblasts[8–10] . Differences in the localizations of extracellular matrix ( ECM ) molecules in the spleen suggest that the ECM plays a role in the compartmentalization of immune cells to their respective niches , and clearly , future analyses of ECM animal models require the consideration of possible immunological defects[10] . The spleen is among the main organs affected by CVL and is responsible for the immune response to systemic infection; however , knowledge on the pathogenic mechanisms involved is limited . Splenic microarchitecture disruption in infected dogs has been associated with fibrosis and disease progression[11 , 12] . The increased deposition of laminin , fibronectin and collagen in the livers of naturally infected dogs , with a concomitant decrease in T lymphocytes and an increase in the parasite load , has been described elsewhere[13] . Still , the structural changes of the splenic ECM are accompanied by reductions in the mRNA expression levels of chemokines and their receptors[14] . These authors have suggested that the failure to express these molecules may lead to deficient leukocyte migration , hampering the development of an efficient immune response . Although metallopeptidases ( MMP-2 and MMP-9 ) , which are enzymes that degrade ECM compounds , have already been demonstrated in the sera of naturally infected dogs[15] , their activities in tissues , particularly in the spleens of infected dogs , have not been evaluated . We hypothesized that changes in the organization of splenic white pulp ( SWP ) could be accompanied by alterations in ECM molecule deposition , metallopeptidase activity and lymphocyte distribution , thereby affecting the immune response and resulting in failure to control the parasite , especially during the final stages of the disease . Thus , the aim of the present study was to examine alterations to splenic ECM compounds through the analysis of collagen , laminin and fibronectin deposition and metallopeptidase expression ( MMP-9 , ADAM-10 ) in groups of animals showing different degrees of SWP organization . Additionally , we aimed to describe and compare the distributions and quantities of T and B lymphocytes in their segregated areas , proliferation markers ( Ki-67 ) and cytokine expression ( IFN-γ and IL-10 ) in these groups . Herein , the disorganized SWP from dogs naturally infected with L . infantum showed a higher deposition of ECM compounds , namely , laminin and collagen , and higher MMP-9 in situ expression than the organized SWP . These alterations were associated with reductions in lymphocyte niches and the frequency of CD4+ cells . The animals included in this study were L . infantum-naturally infected dogs that were destined for euthanasia as recommended by the politics of the Brazilian Ministry of Health and after all owners of the dogs had provided formal written consent . The samples were collected during necropsies conducted by veterinarians from the Laboratório de Pesquisa Clínica em Dermatozoonoses em Animais Domésticos ( LAPCLIN-DERMZOO–INI/FIOCRUZ ) approved by Comitê de Ética em Uso de Animais ( CEUA—Fiocruz ) under the license LW-54/13 and according to the Brazilian Law 11794/08 and Sociedade Brasileira de Ciência em Animais de Laboratório ( SBCAL ) . Forty-one dogs from Barra Mansa- ( Rio de Janeiro , Brazil ) , obtained from a convenience sampling , diagnosed with CVL were included in this study . Later , the animals were grouped according to the level of SWP organization . The primary diagnosis was performed by Centro de Controle de Zoonoses da Secretaria Municipal de Barra Mansa using the CVL DPP test ( Bio-Manguinhos-FIOCRUZ ) as the screening test and ELISA ( EIE–Leishmaniose Visceral Canina Bio-Manguinhos ) as the confirmatory test . In addition , the infection was confirmed by parasite isolation , and all animals were associated with positive cultures in NNN/Schneider's biphasic medium at 25°C . The animals included in the study were negative for the most common hemoparasitoses with Fuller Kit ( Eurovet—Babesia canis ) and Snap 4DX Plus Test ( IDEXX -Anaplasma/Ehrlichia/Lyme/Heartworm ) . All dogs with coinfections ( hemoparasitosis and leishmaniasis ) were excluded from this study . Clinical evaluations were performed by two veterinarians according to the clinical score adapted elsewhere[16 , 17] . Shortly , six common clinical signs ( dermatitis , onychogryphosis , keratoconjunctivitis , the body condition , alopecia and lymphadenopathy ) were scored on a semiquantitative scale from 0 ( absent ) to 3 ( severe ) . Body condition was classified as 0 ( Ideal: easily palpable ribs with minimal fat around , waist easily seen from above and evident abdominal silhouette ) , 1 ( Thin: ribs easily palpable and may be visible without palpable fat , the top of the lumbar region is visible , prominent pelvic bones , silhouette of waist and abdomen apparent ) , 2 ( Very thin: Ribs , lumbar and pelvic bones are easily visible , absence of palpable fat , evidence of other bony prominences and minimal muscle loss ) and 3 ( Extremely thin: All bony structures are prominent and in evidence at a distance , no discernible body fat and severe loss of muscle mass ) . None of the evaluated animals showed obesity . The sum of the values was used to determine the final clinical classification as low ( 0–2 ) , moderate ( 3–6 ) or high ( 7–18 ) score . After clinical examination , euthanasia was conducted by veterinarians from the LAPCLIN-DERMZOO by intravenously administering 1 . 0 ml/kg 1 . 0% Thiopental ( Thiopentax , Cristália ) . After the absence of a corneal reflex due to deep anesthesia was detected , 10 ml of 19 . 1% potassium chloride was intravenously administered ( Isofarma ) . During necropsy , tissue samples from the spleen were collected into sterile DNA-free polypropylene tubes and frozen at -20°C until further use . Total DNA was extracted from approximately 10 mg of each spleen sample . The DNA extraction was performed using the QIAmp DNA Mini kit ( Qiagen , CA , USA ) , which included an initial phase of digestion with 20 μl of proteinase K ( 20 mg/mL ) for 1 hour at 56°C . The DNA was dissolved in 50 μl of tris EDTA buffer ( AE buffer ) and quantified with the NanoDrop spectrophotometer ( Thermo Scientific , USA ) . The parasite loads in the spleen samples were estimated by quantitative PCR ( qPCR ) as previously described by Cavalcanti and collaborators[17] . HPRT primers ( Supplementary S1 Table ) were used to normalize the concentrations of canine DNA in each sample . To quantify the parasite load , primers ( S1 Table ) described by Prina and collaborators[18] were used to amplify a product corresponding to the small subunit ribosomal RNA ( ssrRNA , multicopy gene ) . The qPCR reactions were run with the Step One equipment ( Applied Biosystems , Molecular Probes , Inc . ) using the detection system Power SYBR Green Master Mix ( Applied Biosystems , Molecular Probes , Inc . ) . Total purified DNA ( 100 ng ) in 2 μl was added to a final PCR reaction volume of 20 μl containing 1X Power SYBR Green and 300 nM of each primer for HPRT or 500 nM of each primer for ssrRNA PCR . qPCR was performed with an activation step at 95°C for 10 minutes , followed by 40 cycles of denaturation and annealing/extension ( 95°C for 15 seconds , 60°C for 1 minute and 68°C for 30 seconds ) . A melt curve stage was performed for each specific amplification analysis ( 95°C for 15 seconds , 60°C for 1 minute and 95°C for 15 seconds ) . All reactions were performed in duplicate for each target , and both targets were run in the same plate for the same sample . Quantifications of peripheral blood mononuclear cells ( PBMC ) from non-infected dogs and L . infantum promastigotes ( MCAN/BR/2007/CG-1 ) were performed using a cell counter ( Z1 COULTER COUNTER , Beckman Coulter , Fullerton , CA , USA ) , and total DNA was extracted from 1 . 0 x 106 PBMCs and 1 . 0 x 107 promastigotes . Standard curves for HPRT and ssrRNA genes were prepared using serial 10-fold dilutions from 10−2 to 107 of total purified DNA . The animals were grouped as high or low parasite burden ( cutoff: 5 . 5 x 104 parasites/106 host cells ) as described by Cavalcanti et al[17] . The spleen fragments were fixed in 10% formalin , embedded in paraffin and cut into 5-μm-thick sections that were mounted on microscope slides . The sections were stained with hematoxylin and eosin and examined by light microscopy ( Nikon Eclipse E400 –Tokyo , Japan ) . Structural changes to spleen lymphoid tissues were analyzed as previously described[19] . The parameters analyzed were as follows: the presence of granulomas and the degree of the structural organization of the white pulp ( 1 , organized–the presence of a distinct periarteriolar lymphocyte sheath , germinal center , mantle zone and marginal zone; 2 , slightly disorganized–the presence of either hyperplastic or hypoplastic changes leading to a loss in the definition of any region of the white pulp; 3 , moderately disorganized–evident white pulp with poorly individualized or indistinct regions; and 4 , intensely disorganized–a follicular structure that was barely distinct from the red pulp and T cell areas ) . The number of lymphoid follicles was determined as lymphoid follicles/mm2 of tissue in tissues sections using Nikon ACT-1 software version 2 . 62 ( Nikon ) . Well-formed granulomas were characterized by the presence of epithelioid cells and well-established limits . Collagen deposits were quantified in picrosirius red-stained tissue using ImageJ 1 . 48v software ( NIH , USA ) . Since the parameters evaluated did not differ infected dogs presenting organized SWP from those of uninfected dogs , animals presenting organized SWP were considered as the reference for comparisons between groups . Immunohistochemistry was based on the method reported by Morgado and collaborators[20] . Briefly , spleen fragments frozen in OCT resin ( Sakura ) were cut into 3–5 μm-thick sections and mounted onto microscope slides ( silanized slides; DakoCytomation , Carpinteria , CA , USA ) . The slides were fixed in acetone PA ( Merck , Darmstadt , Germany ) and subjected to hydration , endogenous peroxidase blockage ( peroxidase blocking reagent; Dako ) and nonspecific staining blockage ( 0 . 4% BSA; Sigma , USA ) . The specimens were then incubated with primary antibodies directed against CD3+ , CD4+ , CD8+ , CD21+ , IFN-γ+ , IL-10+ ( Serotec ) , Ki-67+ ( eBioscience ) , laminin , fibronectin , ADAM-10 ( Abcam ) or MMP-9 ( Serotec ) , followed by incubation with the Labelled Polymer ( Envision System-HRP , Dako ) . Aminoethyl carbazole ( AEC kit; Invitrogen ) was used as the substrate-chromogen system , and the slides were counterstained with Mayer’s hematoxylin ( Sigma ) . The slides were examined under a light microscope ( Zeiss ) , and the number of marked cells/mm2 tissue in the red pulp was determined . Two blinded readers evaluated the slides . Laminin and fibronectin deposits were quantified using ImageJ 1 . 48v software ( NIH , USA ) , and the results are presented as the percentage of the marked area . Primary antibody suppression was used as the negative control reaction . The SPSS program for Windows , version 16 ( SPSS Inc . , Chicago , IL , USA ) was used for statistical analysis . The Kolmogorov-Smirnov test was used to evaluate the Gaussian distributions of the variables . The data were analyzed using Student’s t-test for variables with parametric distributions and the Mann-Whitney test for variables with nonparametric distributions . Correlations were determined using Spearman’s rank correlation coefficient . The data are reported as the mean and standard error of the mean ( SEM ) or and the median and minimum and maximum values . Non-numerical data were analyzed in 2 x 2 contingency tables with Fisher’s exact test using Prism software ( GraphPad Prism version 6 . 01 ) . P-value < 0 . 05 was considered significant . Considering the 41 animals diagnosed with CVL included in this study , the main clinical signs observed were onychogryphosis ( 79 . 4% ) , followed by dermatitis ( 67 . 6% ) , loss of body condition ( 41 . 2% ) , alopecia ( 38 . 2% ) , lymphadenomegaly ( 23 . 5% ) and keratoconjunctivitis ( 23 . 5% ) ( S1 Fig and S2 Table ) . The animals were classified into 3 groups based on the clinical score: 1-low score ( N = 13 , 31 . 7% ) ; 2-moderate score ( N = 15; 36 . 6% ) ; and 3-high score ( N = 13; 31 . 7% ) . Based on histopathological analysis of the spleen , the white pulp was classified as organized SWP ( OR , n = 3 , 7 . 31% and Fig 1A ) , slightly disorganized SWP ( SD , n = 8 , 19 . 51% and Fig 1B ) , moderately disorganized SWP ( MD n = 15 , 36 . 6% and Fig 1C ) and intensely disorganized SWP ( ID; n = 15 , 36 . 6% and Fig 1D ) . We observed that the higher the level of SWP disorganization was , the smaller the number of lymphoid follicles would be ( P-value = 0 . 0001 ) ( Fig 1E ) . The animals with MD-ID SWP showed higher clinical scores ( median 6 . 0; range 0 to 12 . 0 ) than the animals with OR-SD SWP ( clinical score: median 1 . 0; range 0 to 8 . 0 P-value = 0 . 021 ) ( Fig 1F ) . The analysis of the granuloma number/mm2 of splenic tissue revealed that only 6 dogs showed well-formed granulomas ( n = 6; median 3 . 83 × 106; minimum-maximum 4 , 6100–20 , 300 , 000 parasites/106 cells ) , and this finding was accompanied by a higher parasite burden compared to the samples that did not show well-formed granulomas ( n = 35; median 5 . 26 × 105; minimum-maximum 145–6 , 970 , 000 parasites/106 cells , P-value = 0 . 039 ) . Most of the dogs analyzed in this study ( n = 30 , 73 . 2% ) showed moderate to intense SWP disorganization . We hypothesized that this disorganization was associated with a modification to ECM molecule expression . In this context , the expression of laminin , fibronectin and collagen were analyzed by immunohistochemistry ( Fig 2A–2M ) . Fibronectin deposition was more abundant in the red pulp and in the intrafollicular than laminin deposition ( Fig 2C and 2D , respectively ) . In general , the deposition of laminin was less intense in the intrafollicular region than in the red pulp ( Fig 2H ) . In the MD-ID group , the laminin deposition assumed an irregular distribution . For example , in the follicle delimitation , laminin deposition was discontinuous and less intense in the animals with moderately to severely disorganized SWP than in the animals with organized to slightly disorganized SWP . We did not observe differences in fibronectin deposition when the groups were compared ( Fig 2M ) . Higher laminin and collagen deposition in dogs with moderately to intensely disorganized SWP were observed ( Fig 2O and 2P and S3 Table , P-value = 0 . 045 and 0 . 036 , respectively ) . To determine whether the alterations in laminin deposition were accompanied by alterations in fibronectin deposition , we correlated these data and observed a positive correlation between laminin and fibronectin deposition ( P-value = 0 . 043; r2 = 0 . 393 and S2A Fig ) . We investigated whether MMP-9 and ADAM-10 , two peptidases associated with matrix remodeling , could be altered in animals presenting different degrees of SWP disorganization by immunohistochemistry ( Fig 3C and 3D ) . MMP-9 expression was higher in animals with MD-ID SWP than in animals with OR-SD SWP ( P-value = 0 . 035 ) ( Fig 3E and S3 Table ) . No significant differences were observed for ADAM-10 expression ( P-value>0 . 05 ) ( Fig 3F and S2 Table ) . To examine the lymphocyte subpopulations of the disorganized spleen , CD4+ , CD8+ and CD21+ cells were quantified in the red pulp ( Fig 4A–4J ) since no individual lymphoid structure can be observed in intensely disorganized tissue . The CD4+ T cell number was lower in the MD-ID animals than in the OR-SD animals ( P-value = 0 . 027 , Fig 4K and S3 Table ) . No differences in the numbers of CD8+ and CD21+ cells were evident between the groups ( Fig 4L and 4M and S3 Table ) ( P-value>0 . 05 ) . To verify whether associations between the cellular profile and extracellular molecule or metallopeptidase alterations were present , we correlated these data and observed positive correlations between collagen deposition and CD21+ cells ( P-value = 0 . 008 and r2 = 0 . 418; S2B Fig ) and between CD21+ cells and MMP-9 expression ( P-value = 0 . 010 and r2 = 0 . 401; S2C Fig ) . We also evaluated IFN-ɣ+- and IL-10+-expressing cells . Although no differences were evident between groups , a correlation between IFN-ɣ+ and IL-10+ expression was observed ( P-value = 0 . 0074 and r2 = 0 . 412 ) ( Fig 4N and 4O and S2D Fig ) . Because the dogs were naturally infected and because the parasite load and splenic histopathological alterations were noted as markers of infection/disease progression , as described elsewhere[5 , 12] , we examined the associations between the parasite load and splenic histopathological alterations . According to Cavalcanti[17] , we grouped the dogs as follows: 1-Low parasite load and organized SWP ( LOW/OR ) , 2- Low parasite load and disorganized SWP ( LOW/DS ) and 3-High parasite load and disorganized SWP ( HIGH/DS ) ( Fig 5 ) . When these groups were compared , a low lymphoid follicle number ( Fig 5A ) and high laminin deposition ( Fig 5B ) were detected even in dogs with low parasite loads . Collagen deposition ( P-value = 0 . 041 ) and MMP-9 expression ( P-value = 0 . 035 ) were higher in the LOW/DS and HIGH/DS groups than in the LOW/OR group ( Fig 5D and 5E ) . No differences in ADAM-10 expression were evident between the groups ( Fig 5F ) . Furthermore , no differences in fibronectin deposition or in the numbers of CD8+ , CD21+ and Ki-67+ cells/mm2 were evident ( P-value>0 . 05 ) ( Fig 5C and 5H–5J ) . However , a lower CD4+ cells/mm2 red pulp value was observed in LOW/DS dogs than in LOW/OR dogs ( Fig 5G ) . In addition , the number of CD4+ cells/mm2 red pulp of the HIGH/DS animals was similar to that of the LOW/DS animals ( P-value>0 . 05 ) . However , the reduction in size and the atrophy in both the lymphoid periarteriolar sheath and lymphoid follicles in the spleens from dogs with high parasite loads and similar numbers of CD4+ cells in the red pulp favor the hypothesis that SWP disorganization contributes to difficulties in cell migration ( Fig 5G , the raw data are shown in the S4 Table ) . Herein , alterations in histology , ECM compounds , metallopeptidase expression and CD4+ cell quantity were detected in the spleens of dogs naturally infected with Leishmania infantum . SWP microarchitecture disorganization was observed in most of the animals ( N = 30 , 73 . 2% ) . The disorganized SWP showed a reduction in the size and number of lymphoid follicles and the periarteriolar sheath , as well as a low number of CD4 + lymphocytes in the red pulp , suggesting that T and B lymphocytes were not migrating to their specific sites and/or were beginning apoptosis . A reduction in spleen cellularity has already been demonstrated in a murine model of experimental infection with L . infantum[21] . The spleen is a secondary lymphoid organ that contains segregated areas of T and B lymphocytes , which favors the antigenic presentation and activation of these cells . An absence of or reduction in lymphocytes in their respective areas can lead to an activation deficit that contributes to the failure to control the parasite load . We observed low numbers of CD4 + cells in the animals with moderate to intense SWP disorganization . In CVL , the spleen shows a marked reduction in the gene expression levels of several cytokines , chemokines and their receptors[16 , 17] . In particular , a reduction in CXCL13 is associated with atrophy and lymphoid follicle disorganization[6] . We observed that the majority of animals with low parasite loads already had SWP disorganization and low numbers of CD4+ T cells , with concomitantly high laminin expression . Dogs with disorganized SWP , even those with low parasite loads , showed high laminin , collagen and MMP-9 expression . This high expression of ECM molecules can be caused by the intense inflammation occurring at the beginning of infection[5 , 15] . The increase of MMP-2 and MMP-9 in the sera of dogs with visceral leishmaniasis has been previously demonstrated[15 , 22] . Thus , in our study , we found that MMP-9 expression was higher in dogs with disorganized SWP than in dogs with organized SWP . Moreover , MMP-9 expression was more intense in dogs with high parasite burdens and disorganized SWP . In this context , as observed in cardiomyopathy , the expression levels of metalloproteinases , such as MMP-9 , MMP-3 and MMP-10 , produced by activated macrophages[23] can be positively regulated during this initial inflammatory process , leading to alterations in the degradation of collagen , laminin and fibronectin , to facilitate the migration of inflammatory cells to the site of infection , consequently modifying the organization of the SWP . MMP-9 plays a role in the orientation of cell migration , with ECM degradation mainly affecting type IV collagen , vascular remodeling and the inflammatory process[15 , 24 , 25] . The increased deposition of ECM components has been described in the livers of dogs[13] and in the murine thymus and lymph node[8] and has been correlated with an increased parasite load , suggesting that the deposition of fibronectin and laminin might be responsible for the success of the infection[13] . Fibronectin degradation can generate peptide fragments , which bind to receptors in macrophages , hampering the activation and function of these cells . Furthermore , the changes in the deposition of these components may reflect the inflammatory and degenerative processes , which is ultimately important in the dissemination of the parasite[19] . On the other hand , we cannot exclude the possibility that the parasite has direct effects on the ECM molecules . In this context , the direct effect of L . amazonensis promastigotes on an in vitro collagen matrix has been previously demonstrated[26] . When an animal is inoculated , the promastigotes are exposed to the dermis , which comprises ECM , growth factors and resident cells[27] . To establish the infection , promastigotes must overcome the obstacles presented by the dermal ECM , which may affect the tissue and thereby contribute to pathogenesis[27 , 28] . However , herein , we evaluated dogs with chronic infections , and the effects of amastigotes on ECM remain unknown . In the liver , laminin and fibronectin expression were correlated with the parasite burden and disease progression in dogs naturally infected with L . infantum ( syn . L . chagasi ) , suggesting that these molecules were important to the invasion of Leishmania parasites[13] . We found an association between high laminin expression and ECM remodeling . Moreover , our results showed that high laminin expression was evident even when the parasite load was still low , suggesting that laminin could play roles in all stages of infection . Another metallopeptidase evaluated in this study was ADAM-10 , and ADAM-10 expression was detected in all evaluated groups . No differences in ADAM-10 expression were observed between the studied groups , which suggests that although this disintegrin is involved in type IV collagen degradation[23 , 25] , in the lymphoid follicle ontogeny[29 , 30] and in the regulation and maintenance of the lymphoid architectures of secondary lymphoid organs[29 , 31] , ADAM-10 expression is not associated with SWP disorganization in the animals included in this study . Animals with disorganized SWP presented high laminin deposition and low amounts of CD4+ cells . In thymic tissue , O'campo and collaborators[32] have reported that laminin is associated with the premature migration of CD4+ T and CD8+ T lymphocytes . In this context , CD4+ cells are leaving the spleen via mechanisms not yet known . In dogs with high parasite loads , we observed that the number of CD4+ cells in the red pulp was similar to that in dogs with organized white pulp; however , atrophy of the lymphoid periarteriolar sheath and lymphoid follicles was evident . We suggest that CD4+ cells in advanced infection may not have been able to migrate to their specific compartments within the white pulp due to the low expression levels of chemokines , cytokines and their receptors[6 , 14 , 21 , 33 , 34] or due to alterations in the distributions of splenic conduits . Additional experiments should be performed to confirm this hypothesis . Granuloma formation , which is an efficient immune response for controlling parasite burden in visceral leishmaniasis , is one parameter for evaluations of the cell-mediated immune response[35] . In the present study , we observed the presence of well-formed granulomas only in 6 animals; however , in murine experimental infections , a marked infectious granulomatous reaction , which involved Kupffer cells , in the liver led to parasite load control[36 , 37] . Herein , the parasite burdens in the dogs with well-formed granulomas were higher than in the dogs in which granulomas were not observed . These data suggest a delay/deficit in the formation of an efficient immune response by the evaluated dogs . In acute phase of experimental Leishmania infantum-infection in macaque model , there was a strong Th1 response , and parasite load could be controlled in blood , bone marrow , lymph nodes and liver but not in the spleen where the parasite burden remained constantly [38] . During the chronic phase , the immune response converted to an IL-10-dominated environment in the spleen [38] . Herein , the studied dogs were naturally infected , and probably were in the chronic phase , which could explain the absence of differential IL-10 expression when the groups were compared . Altogether , the data presented herein support the conclusion that the disorganization of the splenic microarchitecture is a frequent alteration in the evaluated dogs , suggesting that the splenic structure and function are drastically altered and compromised during CVL . These alterations to ECM compounds and immune cells might consequently lead to immunosuppression and severe disease .
Infected dogs play important roles in the transmission of visceral leishmaniasis . These dogs are considered reservoirs of parasites in urban areas and fail to mount an efficient anti-Leishmania immune response . However , the specific immunosuppression profile is not completely understood . In our report , we evaluate and discuss the morphophysiological alterations in the spleens of dogs with visceral leishmaniasis . We found an association between extracellular matrix alterations and a failure to control the parasite load . We suggest a role for these alterations in hindering an immune response that is otherwise able to control the parasite load , thereby leading to disease progression . Our research contributes to the current knowledge of the immunopathology of canine visceral leishmaniasis .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[ "medicine", "and", "health", "sciences", "immune", "physiology", "immune", "cells", "spleen", "granulomas", "immunology", "vertebrates", "parasitic", "diseases", "dogs", "animals", "collagens", "mammals", "parasitic", "protozoans", "protozoans", "leishmania", "cellular", "structures", "and", "organelles", "animal", "cells", "proteins", "extracellular", "matrix", "immune", "response", "biochemistry", "leishmania", "infantum", "eukaryota", "cell", "biology", "physiology", "biology", "and", "life", "sciences", "cellular", "types", "amniotes", "organisms" ]
2018
Morphophysiological changes in the splenic extracellular matrix of Leishmania infantum-naturally infected dogs is associated with alterations in lymphoid niches and the CD4+ T cell frequency in spleens
Mice lacking the type I interferon receptor ( IFNAR−/− mice ) reproduce relevant aspects of Crimean-Congo hemorrhagic fever ( CCHF ) in humans , including liver damage . We aimed at characterizing the liver pathology in CCHF virus-infected IFNAR−/− mice by immunohistochemistry and employed the model to evaluate the antiviral efficacy of ribavirin , arbidol , and T-705 against CCHF virus . CCHF virus-infected IFNAR−/− mice died 2–6 days post infection with elevated aminotransferase levels and high virus titers in blood and organs . Main pathological alteration was acute hepatitis with extensive bridging necrosis , reactive hepatocyte proliferation , and mild to moderate inflammatory response with monocyte/macrophage activation . Virus-infected and apoptotic hepatocytes clustered in the necrotic areas . Ribavirin , arbidol , and T-705 suppressed virus replication in vitro by ≥3 log units ( IC50 0 . 6–2 . 8 µg/ml; IC90 1 . 2–4 . 7 µg/ml ) . Ribavirin [100 mg/ ( kg×d ) ] did not increase the survival rate of IFNAR−/− mice , but prolonged the time to death ( p<0 . 001 ) and reduced the aminotransferase levels and the virus titers . Arbidol [150 mg/ ( kg×d ) ] had no efficacy in vivo . Animals treated with T-705 at 1 h [15 , 30 , and 300 mg/ ( kg×d ) ] or up to 2 days [300 mg/ ( kg×d ) ] post infection survived , showed no signs of disease , and had no virus in blood and organs . Co-administration of ribavirin and T-705 yielded beneficial rather than adverse effects . Activated hepatic macrophages and monocyte-derived cells may play a role in the proinflammatory cytokine response in CCHF . Clustering of infected hepatocytes in necrotic areas without marked inflammation suggests viral cytopathic effects . T-705 is highly potent against CCHF virus in vitro and in vivo . Its in vivo efficacy exceeds that of the current standard drug for treatment of CCHF , ribavirin . Crimean-Congo hemorrhagic fever virus ( CCHFV ) is a negative-strand RNA virus belonging to the genus Nairovirus of the family Bunyaviridae . The virus is endemic in Africa , Asia , southeast Europe , and the Middle East . Hyalomma ticks transmit the virus to humans , wildlife , and livestock . Humans may also be infected by contact with infected livestock . Human-to-human transmission occurs mainly in the hospital setting . In humans , the virus causes a febrile illness that may be associated with hemorrhage , liver necrosis , shock , and multiorgan failure . Further hallmarks of the disease are increased levels of serum aspartate and alanine aminotransferase ( AST and ALT , respectively ) , thrombocytopenia , and disseminated intravascular coagulopathy . The average case fatality rate is 30–50% , but may be higher in nosocomial outbreaks [1]–[5] . The pathophysiology of the disease is poorly understood . Endothelial and liver cell damage , induction of proinflammatory cytokines , and dysregulation of the coagulation cascade are thought to play a role [3]–[8] . Studies on the pathophysiology of Crimean-Congo hemorrhagic fever ( CCHF ) have been hampered by the lack of an appropriate animal model , as no mammal with fully functional immune system has been described so far — except humans — that develops disease upon infection . The first animal model was neonatal mouse [9] . Recently , two transgenic mouse models for CCHF have been described , first , mice lacking the signal transducer and activator of transcription 1 ( STAT1−/− mice ) and second , mice lacking the type I ( alpha/beta ) interferon receptor ( IFNAR−/− mice ) [10]–[12] . Both knockout mice are defective in the innate immune response , die rapidly from CCHFV infection , and reproduce relevant aspects of human CCHF . Surrogate models for CCHF employ IFNAR−/− mice infected with Dugbe or Hazara virus [13] , [14] , two CCHFV-related nairoviruses that are not known to cause disease in human . Work with these models can be carried out at biosafety level ( BSL ) -2 , while work with infectious CCHFV requires BSL-4 facilities . In the present study , we aimed at characterizing the pathological changes in the liver of CCHFV-infected IFNAR−/− mice in more detail . Furthermore , we employed this model to evaluate the antiviral efficacy of ribavirin , arbidol , and T-705 ( favipiravir ) against CCHFV in vivo . These drugs are either in clinical use or in an advanced stage of clinical testing . Ribavirin inhibits CCHFV replication in cell culture [15] and is administered to CCHF patients , though its clinical benefit is not proven and discussed controversially [16]–[19] . It shows beneficial effects in the neonatal and STAT1−/− mouse models [9] , [10] . Ribavirin currently is the only drug available for treatment of CCHF . Arbidol is a broad-spectrum antiviral showing activity against a range of RNA viruses in vitro and in vivo , most notably influenza A virus [20]–[24] . In Russia and China , the drug is in clinical use primarily for prophylaxis and treatment of acute respiratory infections including influenza . Arbidol is assumed to act via hydrophobic interactions with membranes and virus proteins , thus inhibiting viral fusion and entry [25]–[27] . T-705 is a potent inhibitor in vitro and in animal models of influenza virus , phleboviruses , hantaviruses , arenaviruses , alphaviruses , picornaviruses , and norovirus [28]–[35] . Following conversion to T-705-ribofuranosyl-5′-triphosphate , it presumably acts as a nucleotide analog that selectively inhibits the viral RNA-dependent RNA polymerase or causes lethal mutagenesis upon incorporation into the virus RNA [36]–[40] . T-705 ( favipiravir ) is currently in late stage clinical development for the treatment of influenza virus infection . This study was carried out in strict accordance with the recommendations of the German Society for Laboratory Animal Science under supervision of a veterinarian . The protocol was approved by the Committee on the Ethics of Animal Experiments of the City of Hamburg ( Permit no . 44/11 ) . All efforts were made to minimize the number of animals used for the experiments and suffering of the animals during the experiments . All staff carrying out animal experiments has passed an education and training program according to category B or C of the Federation of European Laboratory Animal Science Associations . The animal experiments in this study are reported according to the ARRIVE guidelines [41] . A total of 162 mice were used for this study and all mice were included in the analysis . CCHFV strain Afg-09 2990 had been isolated in 2009 in our laboratory from a patient with a fatal course of infection [42] and passaged 2 times before it has been used in this study . The virus stock was grown on Vero E6 cells , quantified by immunofocus assay ( see below ) , and stored at −70°C until use in in vitro and in vivo experiments . Ribavirin ( CAS no . 36791-04-5; PubChem CID 37542 ) was obtained from MP Biomedicals ( order no . 02196066 ) , arbidol hydrochloride ( CAS no . 131707-23-8; PubChem CID 131410 ) from Waterstone Technology , USA ( order no . 49823 ) , and T-705 ( favipiravir; CAS no . 259793-96-9; PubChem CID 492405 ) was custom synthesized by BOC Sciences , Creative Dynamics , USA . The compounds were dissolved in dimethyl sulfoxide ( DMSO ) at a concentration of about 10 mg/ml and stored at −20°C . Final DMSO concentration in the cell culture supernatant was 0 . 1% . Vero E6 cells were grown in Dulbecco's Modified Eagle's Medium ( DMEM ) ( PAA Laboratories ) supplemented with 5% fetal calf serum ( FCS ) and streptomycin/penicillin and seeded at a density of 4×104 cells per well of a 24-well plate at 1 day before infection . Cells were inoculated with CCHFV at a multiplicity of infection ( MOI ) of 0 . 01 in the BSL-4 laboratory . The inoculum was removed after 1 h and replaced by fresh medium complemented with different concentrations of compound . For arbidol experiments , cells were additionally pretreated with arbidol 18 h before infection . Concentration in cell culture supernatant of infectious virus particles was measured 2–4 days post infection ( p . i . ) by immunofocus assay . Cell growth and viability under compound treatment was determined by the 3- ( 4 , 5-dimethylthiazol-2-yl ) -2 , 5-diphenyl-2H-tetrazoliumbromide ( MTT ) method as described [43] . A sigmoidal dose–response curve was fitted to the data using Prism GraphPad 6 . 0 ( GraphPad Software ) . The inhibitory concentrations that reduced the virus titer by 50% , 90% , and 99% ( IC50 , IC90 , and IC99 , respectively ) and the cytotoxic concentrations that reduced cell growth by 50% and 90% ( CC50 and CC90 , respectively ) were calculated from the sigmoidal functions . For analysis of combinations of two drugs , an 8×8 concentration matrix was tested . Drugs x and y were tested in the concentrations c = 0; IC90/8; IC90/4; IC90/2; IC90; IC90•2; IC90•4; IC90•8 in all possible combinations ( cx , cy ) . The IC90 values were derived from the prior single-drug experiments . The drug combination data were analyzed using the Bliss independence drug interaction model [44] . This model is defined by the equation Exy = Ex+Ey− ( Ex•Ey ) , where Exy is the additive effect of drugs x and y as predicted by their individual effects Ex and Ey . Ex = ( Vobs ( 0 , 0 ) −Vobs ( cx , 0 ) ) /Vobs ( 0 , 0 ) and Ey = ( Vobs ( 0 , 0 ) −Vobs ( 0 , cy ) ) /Vobs ( 0 , 0 ) , where Vobs ( cx , cy ) is the observed , i . e . experimentally determined virus titer for ( cx , cy ) . In analogy to the MacSynergy II program [44] , [45] , which evaluates antivirus data according to the Bliss independence model , a three-dimensional approach was used to identify areas where observed effects are greater ( synergy ) or less ( antagonism ) than those predicted by Exy . To this end , the ratio between predicted virus titer Vpred ( cx , cy ) = Vobs ( 0 , 0 ) • ( 1−Exy ) and observed virus titer Vobs ( cx , cy ) was calculated for each drug combination ( cx , cy ) . A ratio >1 indicates synergy ( i . e . for ( cx , cy ) the virus titer predicted for additive effect is higher than the experimentally determined virus titer ) , a ratio <1 indicates antagonism ( i . e . for ( cx , cy ) the virus titer for predicted additive effect is lower than the experimentally determined virus titer ) . IFNAR−/− mice ( 129Sv background ) [46] were bred in the Specific Pathogen Free animal facility of the Bernhard-Nocht-Institute . Six to twelve-week-old female animals ( weight median 20 g , range 15–24 g ) were used for all experiments , except of nine males that were used for determination of the lethal virus dose . A group size of 5 animals was expected to provide sufficiently accurate estimates of survival rate , viremia , and clinical chemistry parameters . It allows to detect an 80% difference in survival rate between control and treatment group with p ( alpha ) = 0 . 05 and power ( 1 – beta ) = 0 . 8 . Experimental groups were age-matched . Three to five animals of a group were kept together in a conventional cage without enrichments . They had ad libitum access to food and water . Infection experiments with CCHFV were performed in the animal facility of the BSL-4 laboratory with artificial light/dark cycles . Three to ten animals per group ( depending on whether organ collection was planned ) were infected by intraperitoneal ( i . p . ) injection with 0 . 3 to 104 focus forming units ( FFU ) of CCHFV in 100 or 200 µl DMEM containing 2% FCS . The mode of administration was chosen to facilitate comparability with previously described CCHF mouse models [10] , [11] . After infection , mice were monitored daily for signs of disease , and body weight and rectal body temperature were measured using thermometer BIO-TK8851 with BIO-BRET-3 rectal probe for mice ( Bioseb , France ) . Animals with severe signs of disease such as seizures , bleeding , abdominal distention , diarrhea , agony , or weight loss of >15% within 2 days were euthanized . Blood samples of 30–80 µl per animal were drawn by tail vein puncture in intervals of 1–4 days over a period of 14 days ( ≤5 blood drawings in total ) for clinical chemistry and viremia measurement . For organ collection , when criteria for euthanasia were fulfilled , and at the end of the experiment , animals were euthanized with an isoflurane overdose followed by cervical dislocation . Organs were collected after death or at day 3 p . i . from 2–3 animals that have been randomly chosen from experimental groups with 7–10 animals , and analyzed for infectious virus titer and histopathological changes . Experiments were not replicated . Ribavirin was administered once daily by the i . p . route . A stock of 10 mg/ml in 0 . 9% NaCl was prepared before each application . Animals received a ribavirin dose of 100 mg/ ( kg×d ) ( 200 µl for a 20-g mouse ) or 200 µl of 0 . 9% NaCl as a placebo . The ribavirin dose that is fatal to 50% of mice ( LD50 ) is 220 mg/ ( kg×d ) [47] . Treatment was commenced 1 h p . i . and continued until death or day 8 . Arbidol was administered once daily per os using a stomach probe . Suspensions of 15 or 30 mg/ml in 0 . 5% methylcellulose was prepared before each application . Animals received an arbidol dose of 75 or 150 mg/ ( kg×d ) ( 100 µl suspension for a 20-g mouse ) or 100 µl of 0 . 5% methylcellulose as a placebo . Treatment was commenced 1 day before infection and continued until death or day 8 . In vivo toxicity of arbidol was evaluated in 26 uninfected animals treated with 0 , 25 , 75 , 150 , 300 , or 600 mg/ ( kg×d ) for 8 days . No toxic effects were observed in this dose range . T-705 was administered twice daily per os using a stomach probe . Suspensions of 0 . 75 , 1 . 5 , 3 , or 30 mg/ml in 0 . 5% methylcellulose were prepared daily . Animals received a T-705 dose of 7 . 5 , 15 , 30 , or 300 mg/ ( kg×d ) ( 100 µl suspension twice daily for a 20-g mouse ) or 100 µl of 0 . 5% methylcellulose twice daily . Treatment was commenced 1 h p . i . or later and continued until death or day 8 . Infectious virus particles in blood and organ samples were determined by immunofocus assay . Organ samples were homogenized in 500 µl DMEM–2% FCS using Lysing Matrix D ( MP Biomedicals ) in a beat mill . Vero cells in 24-well plates were inoculated with 200 µl of serial 10-fold dilutions of sample . The inoculum was removed after 1 h and replaced by a 1%-methylcellulose–DMEM–6% FCS overlay . After 5 days of incubation , cells were fixed with 4% formaldehyde in phosphate-buffered saline ( PBS ) , washed with water , and permeabilized with 0 . 5% Triton X-100 in PBS . After washing and blocking with 10% FCS in PBS , infected cell foci were detected with CCHFV nucleoprotein ( NP ) -specific monoclonal antibody A4 [48] . After washing , cells were incubated with peroxidase-labeled anti-mouse IgG . Foci were visualized with tetramethylbenzidine and counted . Virus-specific antibodies in blood were detected by immunofluorescence assay ( IFA ) using cells infected with CCHFV strain Afg-09 2990 as an antigen . Mouse serum was inactivated for 1 h at 60°C and tested at a dilution of 1∶20 . Serum samples were diluted 1∶10 or higher , if required , in 0 . 9% NaCl and analyzed for AST and ALT activity by using commercially available colorimetric assay kits at 25°C ( detection limit for undiluted serum is 2 . 25 U/l for AST and 2 . 65 U/l for ALT ) ( Reflotron , Roche Diagnostics ) . Parameters were measured for individual animals . Lung , kidney , heart , spleen , brain , and liver were collected , fixed in 4% formaldehyde in PBS , and embedded in paraffin using a Leica ASP300 S tissue processor and a Leica EG1160 embedding station ( Leica ) . Sections ( 4 µm ) were stained with hematoxylin–eosin ( H&E ) or processed for immunohistochemistry ( IHC ) . IHC sections were stained using the Ventana BenchMark XT automated staining system ( Ventana Medical Systems ) and Cell Conditioning solution 1 or 2 ( Ventana ) for 30–60 min . Sections were incubated with primary antibodies directed against B cell marker B220 ( 1∶400; eBioscience ) , apoptosis marker cleaved caspase-3 ( 1∶100; R&D Systems ) , T cell marker CD3 ( 1∶100; Dako ) , myeloid-lineage cell ( e . g . macrophage ) marker Iba-1 ( 1∶2 , 000; Wako Chemicals ) , inducible nitric oxide synthase ( iNOS ) expressed by activated monocyte-derived cells ( 1∶50; Abcam ) , and cell proliferation marker Ki67 ( 1∶250; Abcam ) for 1 h . Primary antibodies were detected with anti-mouse IgG , anti-rabbit IgG , or anti-rat IgG Histofine Simple Stain MAX PO immuno-enzyme polymer ( Nichirei Biosciences ) and stained with 3 , 3′-Diaminobenzidine ( DAB ) substrate using the ultraView Universal DAB Detection Kit ( Ventana ) . Cells were counterstained with hematoxylin . IHC with primary antibodies directed against CCHFV NP ( monoclonal antibody A4 [48] , 1∶500 ) and neutrophil marker Ly6G ( 1∶1 , 000; BD Bioscience ) was performed manually . Sections were boiled in citrate buffer ( pH 6 ) for 1 h and incubated with antibody at 4°C overnight . Primary antibodies were detected with anti-mouse IgG Histofine Simple Stain AP or anti-rat IgG Histofine Simple Stain MAX PO immuno-enzyme polymer and stained with Fast Red ( Roche ) or DAB ( Sigma-Aldrich ) substrate , respectively . Mayer's hematoxylin solution was used for counterstaining . Sections were coverslipped with Tissue Tek mounting medium ( Sakura Finetek ) . Statistical analysis was performed with GraphPad 6 . 0 ( GraphPad Software ) . Unpaired groups were compared with the two-tailed Mann–Whitney U test for continuous parameters and with two-tailed Fisher's exact test for frequencies . Survival curves were compared with the log-rank ( Mantel–Cox ) test . Before testing antivirals in the IFNAR−/− mouse model , we aimed at determining the optimal infection dose for CCHFV strain Afg09-2990 and characterizing the disease caused by this particular strain . To this end , IFNAR−/− mice were infected i . p . with 0 . 3 , 1 , 3 , 10 , 100 , 1 , 000 , and 10 , 000 FFU . Animals died from the infection even after inoculation with 0 . 3 FFU ( inoculum , died/infected: 0 . 3 FFU , 4/5; 1 FFU , 5/5; 3 FFU , 4/5; 10 FFU , 6/8; 100 FFU 13/13; 1 , 000 FFU , 8/8; 10 , 000 FFU , 3/3 ) . This indicates that only a few infectious virus particles of CCHFV strain Afg09-2990 are sufficient to initiate a productive infection . A lethal outcome was consistently observed with ≥100 FFU . Therefore , the model was further characterized for the inoculation doses 10 , 100 , 1 , 000 , and 10 , 000 FFU ( Fig . 1 ) . Animals infected with 100 or 1 , 000 FFU died between days 3 and 6 , while animals infected with 10 , 000 FFU uniformly died at day 2 . Before death , animals lost about 15% of body weight ( Fig . 1 ) . At day 2 , the mean AST and ALT values were around 300 U/l and 100 U/l , respectively , in animals inoculated with 100–1 , 000 FFU . Both values were higher in the 10 , 000 FFU group ( AST 1 , 600 U/l and ALT 500 U/l ) ( Fig . 1 ) . AST and ALT elevations indicated cell damage , in particular of liver cells . At day 2 , virus titer in blood ranged from below detection lime in the 10 FFU group , via 3 log10 FFU/ml in the 100 and 1 , 000 FFU groups , up to 5 log10 FFU/ml in the 10 , 000 FFU group ( Fig . 1 ) . At day 3 , virus was found in all organs analyzed ( spleen , kidney , liver , heart , lung , and brain ) at titers ranging from 4–7 log10 FFU/g irrespective of the inoculation dose ( Fig . 1 ) . The maximum virus concentration was found in liver . As the inoculation with ≤10 FFU was not uniformly lethal and the inoculation with 10 , 000 FFU leaves only 2 days between infection and lethal outcome for therapeutic intervention , further experiments were conducted with a dose of 100 or 1 , 000 FFU . Lung , heart , kidney , brain , liver and spleen of CCHFV-infected IFNAR−/− mice were collected at day 3 and assessed on H&E-stained sections . Virus distribution in all organs and inflammatory response in liver were visualized by IHC . Naïve IFNAR−/− mice served as a control . The antiviral activity of ribavirin , arbidol hydrochloride , and T-705 against CCHFV strain Afg09-2990 was tested in Vero E6 cells . All three compounds were able to suppress virus replication by 3–4 log10 units at concentrations of ≥10 µg/ml ( Fig . 4 ) . IC50 and IC90 values ranged from 0 . 6–2 . 8 µg/ml and 1 . 2–4 . 7 µg/ml , respectively . IC99 values ranged from 2 . 0–9 . 5 µg/ml . Cell toxicity in the test range as measured by MTT test was only evident for arbidol hydrochloride ( ribavirin CC50>32 µg/ml; arbidol hydrochloride CC50 8 . 3 µg/ml , CC90 20 µg/ml; T-705 CC50>15 µg/ml ) ( Fig . 4 ) . In conclusion , all three compounds showed a potent antiviral effect against CCHFV Afg09-2990 in cell culture . Arbidol displayed toxicity with a therapeutic index of about 10 . Ribavirin was tested in comparison to a placebo group receiving the vehicle ( 0 . 9% NaCl solution ) ( Fig . 5 ) . Both groups of IFNAR−/− mice were infected with 100 FFU CCHFV . Although one animal survived after treatment , ribavirin did not significantly increase the survival rate ( p = 0 . 4 ) . However , the drug prolonged the time to death ( median 3 vs . 6 days for placebo vs . ribavirin , p = 0 . 0007 ) , reduced the levels of AST ( p = 0 . 001 ) and ALT ( p = 0 . 006 ) at day 2 , reduced the virus titer in blood at day 2 ( p = 0 . 0007 ) , increased the weight at day 2 and 3 ( p = 0 . 03 and p = 0 . 002 , respectively ) , and reduced the terminal virus concentration in all organs when compared to placebo at day 3 ( p<0 . 001 separately for each organ ) . Histopathological analysis of organs collected at day 3 from ribavirin-treated mice revealed only small disseminated foci of necrosis; most of the liver parenchyma resembled naïve mice . Markedly reduced hepatocellular necrosis correlated with low numbers of apoptotic hepatocytes ( cleaved caspase-3 ) , T-cells ( CD3 ) , B-cells ( B220 ) , and activated monocyte-derived cells ( iNOS ) . Virus antigen-positive cells ( NP ) were significantly reduced in liver and spleen compared to untreated or placebo-treated mice ( data not shown ) . However , ribavirin-treated mice that succumbed to infection on days 4–9 showed extensive bridging hepatocellular necrosis at the time of death ( Fig . 2 ) . Like in untreated mice , the necrosis was accompanied by presence of numerous Iba-1-positive macrophages ( Kupffer cells ) , showing enlarged cell bodies and focal clustering , and iNOS-expressing activated monocyte-derived cells ( Fig . 3 ) . Both alterations are suggestive for strong monocyte/macrophage activation . However , in contrast to untreated mice , virus antigen was hardly detectable in liver tissue of the treated mice ( Fig . 2 ) , consistent with the low virus titer in all organs ( Fig . 5 , bottom ) . Thus , ribavirin reduces CCHFV load and delays disease progression , but it does not prevent terminal liver necrosis , monocyte/macrophage activation , and lethal outcome in the IFNAR−/− mouse model . Arbidol hydrochloride [75 and 150 mg/ ( kg×d ) ] was tested in comparison to a placebo group receiving the vehicle ( 0 . 5% methylcellulose ) [Fig . 5 and data not shown for 75 mg/ ( kg×d ) ] . Both groups were infected with 1 , 000 FFU CCHFV . Mice were pretreated one day before inoculation . However , the drug changed neither survival rate and survival time , nor any of the other parameter measured . Even reducing the inoculation dose to 10 FFU had no effect when compared to the historical control group . T-705 was tested in comparison to a placebo group receiving the vehicle ( 0 . 5% methylcellulose ) ( Fig . 6 ) . All groups were infected with 100 FFU CCHFV . Initially , a high dose of T-705 [300 mg/ ( kg×d ) ] was tested . The drug was administered from day 0 to day 8 . Placebo-treated animals died between day 3 and 4 . At day 3 , they showed weight loss of nearly 20% , increase in body temperature up to 40°C , AST values of 1 , 200–51 , 000 U/l , and ALT values of 260–6 , 700 U/l . All animals of the treatment group survived the infection and showed no signs of disease . Virus was detected neither in blood nor in the organs throughout the observation period ( Fig . 6 and data not shown ) . Histopathology and IHC at day 3 revealed largely normal liver tissue with absence of virus antigen and inflammatory cells ( Figs . 2 and 3 ) . To determine the efficacy of the drug at an advanced stage of the infection , time-of-addition experiments were performed ( Fig . 6 ) . Treatment with a high dose of the drug was commenced 1 day or 2 days after virus inoculation and continued until day 8 . Survival was 100% in both groups and animals showed hardly any signs of disease . Only if treatment started 2 days p . i . , minor changes in weight , temperature , and AST were seen at day 3 . Virus remained undetectable in blood and organs in both time-of-addition groups throughout the observation period . To provide evidence for infection of the animals in the T-705 treatment groups , the development of CCHFV-specific antibodies was measured 21 days p . i . Only 1/10 ( 10% ) of the animals treated post-exposure , but 10/10 ( 100% ) of the animals treated from day 1 or 2 p . i . developed antibodies , indicating that virus replication under post-exposure treatment with T-705 was even not sufficient to elicit antibodies . To define the lowest effective dose of T-705 , animals received 30 , 15 , or 7 . 5 mg/ ( kg×d ) T-705 or 0 . 5% methylcellulose as a placebo ( Fig . 7 ) . Treatment was commenced 1 h p . i . and continued until death or day 8 . All animals of the 30 and 15 mg/ ( kg×d ) treatment groups survived and showed hardly any signs of disease . Virus was detected at low level only in blood of one animal of the 15 mg/ ( kg×d ) treatment group at day 11 ( Fig . 7 ) . A dose of 7 . 5 mg/ ( kg×d ) did not prevent a lethal outcome , although it prolonged the time to death ( p = 0 . 0007 ) , and reduced the levels of AST ( p = 0 . 0007 ) and ALT ( p = 0 . 004 ) at day 3 . Taken together , T-705 is highly efficient against CCHFV in the IFNAR−/− mouse model . Ribavirin is currently in clinical use for treatment of CCHF [16]–[19] . Therefore , it is important to know if T-705 could be given in combination with ribavirin and how both drugs interact . First , the antiviral activity of 64 combinations of ribavirin and T-705 was determined in cell culture . The 8×8 concentration matrix was designed around the IC90 values of both drugs as determined above . Infectious virus particles were measured 3 days p . i . by immunofocus assay and cell viability was determined by the MTT method . The dose–response surface demonstrates that combinations of ribavirin and T-705 exhibit strong antiviral effects with suppression of virus replication by >5 log units ( Fig . 8A ) . Possible antagonistic or synergistic effects were evaluated using the Bliss independence model in analogy to the algorithms of the MacSynergy II program [44] , [45] . This analysis revealed clear synergistic effects when the drugs were combined in concentrations around their IC90 . In this area of the matrix , the experimental virus titer was up to 2 log units lower than the titer predicted according to the Bliss independence model for additive effect ( Fig . 8B ) . The MTT test did not reveal drug toxicity over the whole matrix ( Fig . 8B ) . To test the effects of drug combination in vivo , animals received a T-705 dose of 30 or 7 . 5 mg/ ( kg×d ) in combination with a ribavirin dose of 100 mg/ ( kg×d ) . The 30 mg/ ( kg×d ) T-705 dose , which is protective upon single-drug administration , was chosen to test if addition of ribavirin interferes with T-705 efficacy . To explore if combination of two sub-effective doses may result in an effective treatment , 7 . 5 mg/ ( kg×d ) T-705 and 100 mg/ ( kg×d ) ribavirin were co-administered . None of the parameters in the 30 mg/ ( kg×d ) T-705 plus 100 mg/ ( kg×d ) ribavirin group ( Fig . 9 , left ) was statistically significantly different from the parameters of the 30 mg/ ( kg×d ) T-705 single-drug group ( Fig . 7 ) . On the other hand , the combination of a 7 . 5 mg/ ( kg×d ) dose of T-705 with a 100 mg/ ( kg×d ) dose of ribavirin ( Fig . 9 right ) improved the survival rate compared to single-drug treatments ( Figs . 5 and 7 ) , although the increase did not reach statistical significance ( p = 0 . 08 for T-705+ribavirin vs . T-705 alone , and p = 0 . 07 for T-705+ribavirin vs . ribavirin alone; two-tailed Fisher's exact test ) . In conclusion , T-705 and ribavirin exert synergistic effects according to the Bliss independence model when combined in concentrations around their IC90 in vitro . Co-administration of both drugs in the animal model suggests that a combined treatment yields beneficial rather than adverse effects . In this study , we have used IFNAR−/− mice as an in vivo model to evaluate the efficacy of antivirals against CCHFV . Main pathological alteration in mice infected with the recently isolated CCHFV strain Afg09-2990 was acute hepatitis with extensive necrosis , reactive proliferation of hepatocytes , mild to moderate inflammatory response , and morphological signs of monocyte/macrophage activation . CCHFV-infected and apoptotic hepatocytes were found in the necrotic areas . Ribavirin , arbidol hydrochloride , and T-705 were active against CCHFV Afg09-2990 in cell culture . However , arbidol hydrochloride was inactive in vivo . Ribavirin was partially active , while T-705 was highly efficient in the mouse model . The latter drug was effective even if the window for therapeutic intervention was less than 2 days . Three mouse models for CCHF have been described in the past: the neonatal mouse model , STAT1−/− mice , and IFNAR−/− mice [9]–[12] . The latter models take advantage of the defect in the innate immune response , which apparently is essential to protect mice from productive CCHFV infection . In all three models , mice are dying from the infection within a few days . We prefer to work with IFNAR−/− mice , as the genetic defect concerns only the interferon type I signaling , while STAT1 deficiency prevents the upregulation of genes due to a signal by either type I or type II interferons and neonatal mice are immunologically tolerant ( neonatal tolerance ) . IFNAR−/− mice are highly susceptible to CCHFV infection . The inoculum sufficient to initiate a productive infection is very low — 0 . 3 FFU — which presumably corresponds to just a few infectious virus particles . This is consistent with experiments in AG129 mice lacking interferon type I and type II receptors , in which an inoculum of 0 . 1 FFU of lymphocytic choriomeningitis virus was sufficient to infect the animals [49] . It has been shown very recently that the IFNAR−/− model mimics hallmarks of human CCHF disease [12] . These findings are extended here by a more detailed immunhistopathological analysis of the liver . Histopathology , virus load measurement , antigen staining in various organs , and the measurement of AST and ALT demonstrate that the liver is the major target organ of CCHFV . Although virus was found in all organs , the titer in liver is the highest and exceeds 7 log10 FFU/g in some experiments . The higher virus titer in liver compared to other organs may explain why the IHC analysis for virus antigen revealed clearly positive cells only in the liver , while the signals in other organs were weak or absent . In exceptional cases , AST and ALT values reached 10 , 000 U/l and 1 , 000 U/l , respectively , demonstrating massive liver cell damage . However , while ALT is specific for this organ , AST is also present at high level in the heart , skeletal muscle , kidneys , brain , and red blood cells [50] . Therefore , the high AST/ALT ratio may also indicate extrahepatic cell damage . Overall , the histological and biochemical findings are compatible with the diagnosis of a fulminant liver damage . Our findings are also in agreement with the pathological observations in human CCHF; hepatocellular necrosis with hyperplastic and hypertrophic Kupffer cells and mild or absent inflammatory cell infiltrates is the prominent histopathological finding in humans [3] . Two new aspects , which may provide some clues as to the pathophysiology of CCHF , are noteworthy . First , in the necrotic lesions of the liver , both CCHFV-infected hepatocytes and apoptotic hepatocytes clustered . This may suggest that CCHFV-infected cells undergo apoptosis and necrosis . As the inflammatory response was only mild to moderate , a direct cytopathic effect of the virus on hepatocytes may be involved in the induction of apoptosis and necrosis . Importantly , this hypothesis has been raised in early IHC studies on humans with CCHF as well [3] . Secondly , activated Iba-1-positive macrophages and activated monocyte-derived cells expressing iNOS were found in the liver . These cells may play a crucial role in the strong proinflammatory immune responses following CCHFV infection , as demonstrated by significant increases of serum proinflammatory cytokines and chemoattractant molecules in IFNAR−/− and STAT1−/− mice , as well as in humans [6]–[8] , [10] , [12] . In this study , we have employed the IFNAR−/− mouse model for testing antivirals against CCHFV in vivo . Ribavirin is the standard treatment in human CCHF — although its clinical efficacy is not proven [16]–[19] — and shows beneficial effects in the neonatal and STAT1−/− mouse models [9] , [10] . Therefore , we first evaluated the IFNAR−/− mouse model using this drug . The survival time was prolonged , while the survival rate was not increased , which largely corresponds to the results of the high-dose challenge experiments in STAT1−/− mice [10] . Despite the delay in disease progression and the reduction in virus load in blood and organs as evidenced by virus titration and IHC , ribavirin was not able to prevent the lethal pathophysiological cascade . Importantly , the development of terminal liver necrosis with marked monocyte/macrophage activation in the virtual absence of virus in the organ demonstrates that host pathways , once they are triggered by the virus , mediate pathology and death irrespective of the presence of the trigger . The second compound tested was arbidol hydrochloride , a broad-spectrum antiviral drug in clinical use against flu [20]–[24] . Arbidol hydrochloride efficiently suppressed CCHFV in cell culture . However , no beneficial effects in the IFNAR−/− mouse model were observed . The drug was administered via the same route but at higher dose than in previous studies that showed beneficial effects against influenza A , coxsackie B , and hantaan virus in mice [20] , [22] . The compound had significant toxicity at higher concentrations in cell culture . It is conceivable that its antiviral effect in vitro is at least partially attributable to general cell toxic effects that are not detected in the MTT assay used to assess cell viability . Therefore , it might be that the in vitro data overestimate the true antiviral effect of the drug against CCHFV . In addition , arbidol is extremely hydrophobic , which may reduce its oral bioavailability in mice . It might be worth testing arbidol in pharmaceutical formulations with enhanced solubility in future [51] . An important observation in our study is the strong antiviral effect of T-705 against CCHFV in cell culture and in the IFNAR−/− mouse model . This compound has been shown to be highly active against a range of viruses in vitro and in vivo , including orthomyxoviruses , arenaviruses , and bunyaviruses of the genera hantavirus and phlebovirus [28]–[34] . Therefore , its activity against CCHFV , a bunyavirus of the genus nairovirus , is not unexpected . However , in view of the low or lacking potency of ribavirin and arbidol in vivo — both of which have almost the same IC50 and IC90 values than T-705 — the high in vivo potency of T-705 is surprising . The IC50 for CCHFV is 5–30 times lower than the IC50 values for other bunyaviruses [35] , which may indicate that this virus is particularly sensitive to T-705 . Even if it was given 2 days before the expected time of death , the animals survived and hardly showed signs of disease . If given immediately post-exposure , the drug suppresses virus replication below the level required to elicit antibodies . We could reduce the dose by a factor of 20 [from 300 to 15 mg/ ( kg×d ) ] with the drug still showing post-exposure efficacy . The mode of action of T-705 against CCHFV is not known . In analogy to other segmented negative strand viruses , T-705-ribofuranosyl-5′-triphosphate may be incorporated into the nascent RNA strand and inhibit further strand extension or induce lethal mutagenesis [36]–[40] . How ribavirin acts against CCHFV is still not known , although the drug is in clinical use since decades . Several mechanisms have been proposed for other viruses: it may be incorporated into the virus RNA causing lethal mutagenesis [52] , interfere with capping [53] , inhibit the viral RNA polymerase [54] , [55] or inhibit the host cell enzyme inosine monophosphate dehydrogenase ( IMPDH ) resulting in reduced GTP levels [56]–[58] . T-705-ribofuranosyl-5′-monophosphate was 150 times weaker than ribavirin-5′-monophosphate in its IMPDH inhibitory effect , suggesting that IMPDH is not a major target enzyme for T-705 [39] , [40] . Given that the mode of action of both drugs is poorly understood , it is difficult to predict how they interact . However , as ribavirin is the standard drug for treatment of CCHF [16]–[19] , a ribavirin/T-705 combination treatment would be an obvious option in clinical practice . Our experiments suggest that both drugs do not act in an antagonistic manner in vitro and in vivo . According to the Bliss independence model there is even evidence for synergistic interaction in vitro and the experiments in the animal model point to a beneficial rather than adverse interaction in vivo . In conclusion , our data hold promise for clinical efficacy of T-705 or ribavirin/T-705 combination treatment in human CCHF .
Crimean-Congo hemorrhagic fever ( CCHF ) is endemic in Africa , Asia , southeast Europe , and the Middle East . The case fatality rate is 30–50% . Studies on pathophysiology and treatment of CCHF have been hampered by the lack of an appropriate animal model . We have employed CCHF virus-infected transgenic mice , which are defective in the innate immune response , as a disease model . These mice die from the infection and show signs of disease similar to those found in humans . First , we studied the liver pathology in the animals , as hepatic necrosis is a prominent feature of human CCHF . Secondly , we used the model to test the efficacy of antiviral drugs that are in clinical use or in an advanced stage of clinical testing . Besides ribavirin , the standard drug for treatment of CCHF , we tested arbidol , a drug in clinical use against respiratory infections , and T-705 , a new drug in clinical development for the treatment of influenza virus infection . While ribavirin and arbidol showed some or no beneficial effect , respectively , T-705 was highly efficacious in the animal model . These data hold promise for clinical efficacy of T-705 in human CCHF .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results", "Discussion" ]
[ "antimicrobials", "medicine", "and", "health", "sciences", "pathology", "and", "laboratory", "medicine", "emerging", "viral", "diseases", "crimean-congo", "hemorrhagic", "fever", "tropical", "diseases", "microbiology", "animal", "models", "model", "organisms", "emerging", "infectious", "diseases", "population", "modeling", "neglected", "tropical", "diseases", "research", "and", "analysis", "methods", "animal", "models", "of", "infection", "viral", "hemorrhagic", "fevers", "infectious", "diseases", "mouse", "models", "infectious", "disease", "modeling", "pathogenesis", "virology", "microbial", "control", "biology", "and", "life", "sciences", "viral", "diseases", "antivirals", "computational", "biology" ]
2014
Evaluation of Antiviral Efficacy of Ribavirin, Arbidol, and T-705 (Favipiravir) in a Mouse Model for Crimean-Congo Hemorrhagic Fever
The mechanism by which a complex auditory scene is parsed into coherent objects depends on poorly understood interactions between task-driven and stimulus-driven attentional processes . We illuminate these interactions in a simultaneous behavioral–neurophysiological study in which we manipulate participants' attention to different features of an auditory scene ( with a regular target embedded in an irregular background ) . Our experimental results reveal that attention to the target , rather than to the background , correlates with a sustained ( steady-state ) increase in the measured neural target representation over the entire stimulus sequence , beyond auditory attention's well-known transient effects on onset responses . This enhancement , in both power and phase coherence , occurs exclusively at the frequency of the target rhythm , and is only revealed when contrasting two attentional states that direct participants' focus to different features of the acoustic stimulus . The enhancement originates in auditory cortex and covaries with both behavioral task and the bottom-up saliency of the target . Furthermore , the target's perceptual detectability improves over time , correlating strongly , within participants , with the target representation's neural buildup . These results have substantial implications for models of foreground/background organization , supporting a role of neuronal temporal synchrony in mediating auditory object formation . Attention is the cognitive process underlying our ability to focus on specific components of the environment while ignoring others . By its very definition , attention plays a key role in defining what foreground is , i . e . , an object of attention , and differentiating it from task-irrelevant clutter , or background [1]–[5] . In the visual modality , studies have shown that figure/ground segmentation is mediated by a competition for neural resources between objects in the scene [2] , [6] , [7] . This competition is biased in favor of different objects via top-down attention as well as behavioral and contextual effects that work to complement or counteract automatic bottom-up processes . An intricate neural circuitry has been postulated to take place in this process spanning primary visual , extrastriate , temporal , and frontal cortical areas [7]–[19] . In the auditory modality , however , there have been a limited number of studies that attempted to explore the neural underpinnings of attention in the context of auditory stream segregation , and the mechanisms governing the extraction of target sounds from a background of distracters [20]–[28] . It is largely unknown how top-down ( e . g . , task-driven or context-dependent ) and bottom-up ( e . g . , acoustic saliency or “pop-out” ) attentional processes interact to parse a complex auditory scene [29] , [30] . In a simultaneous behavioral and neurophysiological study using magnetoencephalography ( MEG ) , we illuminate this interaction using stimuli shown in Figure 1A , consisting of a repeating target note in the midst of random interferers ( “maskers” ) . This design generalizes paradigms commonly used in informational masking experiments , [31] , which explore how listeners' ability to perceive an otherwise salient auditory element is strongly affected by the presence of competing elements . For these stimuli , the ability to segregate the target note depends on various acoustic parameters , including the width of the spectral protection region ( the spectral separation between target and masker frequencies ) . We adapt classic informational masking stimuli to the purposes of this study by randomly desynchronizing all background maskers throughout the duration of the trial , making the target note the only regular frequency channel in the sequence ( with repetition rhythm of 4 Hz ) . The informational masking paradigm has been shown to invoke similar mechanisms to those at play in classic stream segregation experiments [32] , [33] , both in the systematic dependence of performance on the size of masker–target spectral separation , as well as the improvement of performance over time over the course of few seconds . While maintaining the same physical stimulus , we contrasted the performance of human listeners in two complementary tasks: ( 1 ) a “target task” in which participants are asked to detect a frequency-shifted ( ΔF ) deviant in the repeating target signal; and ( 2 ) a “masker task” in which participants are asked to detect a sudden temporal elongation ( ΔT ) of the masker notes . Crucially , attention is required to perform either task , but the participants' attention must be focused on different sound components of the acoustic stimulus in each case . Additionally , all the stimuli are diotic ( identical in both ears ) , averting any confounding effects of spatial attention . The effect of spectral protection region width on the performance of both tasks is illustrated in Figure 1B . In the left panel , it can be seen that the detectability of the target becomes easier with increasing protection region ( significantly positive slope; bootstrap across participants , p<10−4 ) , a result that is in line with previous hypotheses of streaming that correlate the ease of target detection with the frequency selectivity of neurons in the central auditory system [34]–[38] . In contrast , the same manipulations of protection region do not substantively affect masker task performance ( right panel ) ( not significantly different from zero; bootstrap across participants , p>0 . 3 ) . The masker task , designed to divert attentional resources away from the target , involves a more diffuse attention to the spectrally broad and distributed masker configuration; and compared to the target task , reflects a different top-down bias in the way the same stimulus is parsed . The behavioral performance was unchanged whether tested under purely psychoacoustic or neural recording conditions ( no significant difference; unpaired t-test; target task: t = −0 . 75 , p = 0 . 46; masker task: t = 0 . 09 , p = 0 . 93 ) . For the neural recordings , we used the stimuli with the eight-semitones spectral protection region because they roughly matched the behavioral performance across tasks ( d-prime for both is approximately equal to three ) . The target task is not at ceiling with the chosen protection region , hence still engaging participants' selective attentional processes . Depending on listeners' attentional focus , the percept of an auditory target in a complex scene is differentially mirrored by the responses of neurons in auditory cortex . Using the high temporal resolution of MEG , we measure the neural responses to this stimulus paradigm in 14 human participants . Figure 2A reveals that , during the performance of the target task , the target rhythm emerges as a strong 4-Hz component in the neural signal of an individual participant . In contrast , during the masker task , the cortical response entrained at 4 Hz is noticeably suppressed in comparison ( Figure 2A , right panel ) . This differential activation is strong evidence of the modulatory effect of task-dependent attention on the neural representation of a single acoustic stimulus , much like visual attention [39] , [40] . Additionally , this attentional effect on the neural signal is not just momentary but is sustained over the duration of the trial ( steady state ) . This attentional effect is confirmed in the population of 14 participants ( Figure 2B ) , with an average normalized neural response of 20 . 9 in the target task and 8 . 3 in the masker task: a gain of more than two and a half for neural phase-locked , sustained activity when participants' attention is directed towards the repeating note ( individually , 11 out of 14 participants showed a significant increase: paired t-test , p<10−4 ) . Direct correlation between the target task neural response and target task behavior is not observed , but as shown below , changes in a participant's target neural response are significantly correlated with changes in the participant's behavioral responses . The MEG magnetic field distributions of the target rhythm response component , examples of which are shown in the inset of the graphs in Figure 2A , reveal the stereotypical pattern for neural activity originating separately in left and right auditory cortex . The neural sources of all the target rhythm response components with a sufficiently high signal-to-noise ratio originate in auditory cortex [41] . The neural source's mean displacement from the source of the auditory M100 response [42] was significantly different ( two-tailed t-test; t = 2 . 9 , p = 0 . 017 ) by 13 . 8±4 . 9 mm in the anterior direction , for the left auditory cortex only ( no significant differences were found in the right hemisphere due to higher variability there ) . The displacement was not statistically significant in the remaining directions ( 3 . 2±3 . 5 mm lateral; 11 . 3±6 . 4 mm superior ) ; the goodness of fit for these sources was 0 . 51±0 . 05 ( artificially reduced in accordance with [43] ) . Assuming an M100 origin of planum temporale , an area of associative auditory cortex , this is consistent with an origin for the neural response to the target rhythm in Heschl's gyrus , the site of core auditory cortex including primary auditory cortex , and a region known to phase-lock well to 4-Hz rhythms [44] . The neural response change at the target rate of 4 Hz is highly significant ( bootstrap across participants , p<10−4 ) ( Figure 3A ) . In contrast , there is no significant change in normalized neural response at other frequencies , whether at frequencies nearby ( one frequency bin on either side of 4 Hz ) or distant ( alpha , theta , and low gamma band frequencies sampled with approximately 5-Hz spacing up to 55 Hz ) . This demonstrates that this feature-selective sustained attention modulates the cortical representation of the specific feature , but not general intrinsic rhythms , whether in the same band or other bands . Changes in response phase coherence across channels were also assessed at the same frequencies ( Figure 3B , sample participant in Figure 3C ) . This analysis focuses on the distant channel pairs with enhanced phase coherence at each specific frequency . Only the phase coherence at the target rate shows a significant enhancement ( bootstrap across participants , p = 0 . 002 ) , further demonstrating that change from one form of attention to another does not modulate general intrinsic rhythms . This 30% enhancement is distributed across channel pairs , revealing increased phase coherence both within and across hemispheres . We also observe a task-dependent hemispheric asymmetry in the representation of the neural response at the target rate . During the target task , the left hemisphere showed a greater normalized neural response than the right hemisphere ( bootstrap across participants , p = 0 . 001 ) ; during the masker task , the right hemisphere showed a greater normalized neural response than the left hemisphere ( bootstrap across participants , p = 0 . 04 ) ( Figure 3D ) . Together with the behavioral demands of the task , the bottom-up saliency of a target note contributes to both the neural response and participant performance . A close examination of the physical parameters of the stimulus reveals that the frequency of the target note affects the audibility of the repeating rhythm , with higher-frequency targets popping out more prominently than their lower-frequency counterparts . This variation in the pop-out sensation may be explained by the contours of constant loudness of human hearing showing an approximately 5-dB increase over the target note range 250–500 Hz [45] , because our stimuli were normalized according to their spectral power , not loudness . We exploit this physical sensitivity of the auditory system and determine the effect of this target pop-out on the neural and behavioral performances in both target and masker tasks . Figure 4A ( orange line ) confirms that behavioral performance in the target task is easier for higher-frequency targets ( >350 Hz ) than for lower frequencies ( t-test; t = −3 . 3 , p = 0 . 002 ) . Correlated with this trend is an increased neural response to the target for higher frequencies compared to lower frequencies ( red line ) ( increase not statistically significant alone ) . Conversely , the masker task shows a trend of being oppositely affected by the physical saliency of the target note despite its irrelevance for task performance ( approaching significance; t-test , t = 1 . 8 , p = 0 . 08 ) . On the one hand , the neural power is increased for high-frequency targets reflecting their increased audibility ( dark-blue line ) ( though not statistically significant alone ) . On the other hand , as the target becomes more prominent , the participants' performance of the background task deteriorates , indicating a distraction effect caused by the presence of the repeating note ( light-blue line ) . Additionally , phase coherence is significantly enhanced for high-frequency targets over low-frequency targets only during the target task ( bootstrap across participants , p<10−3 ) ( Figure 4C ) . This result confirms that the physical parameters and acoustic saliency of a signal can interfere with the intended attentional spotlight of listeners and effectively deteriorate task performance [46] , [47] , both neurally and behaviorally . In order to establish the correspondence within participants between the neural and behavioral responses under both task conditions in a parametric way , we quantified the slope ( converted into an angle ) relating the normalized neural signal with the listener's d-prime performance on a per-participant basis . The average slope angle for the target task is 55 . 1° , i . e . , a positive slope , demonstrating the positive correlation between the two measures . Bootstrap analysis confirms this; Figure 4B , left panel , illustrates both the bootstrap mean of 55 . 3° ( green line ) and the 5th to 95th percentile confidence limits ( gray background ) , all with positive slopes . Analysis of the masker task also demonstrates the anticorrelation trend between the neural and behavioral data , with an average slope angle of −36 . 3° shown in yellow . The bootstrap analysis also confirms this; Figure 4B ( right panel ) shows that the 5th to 95th confidence intervals ( gray background ) yield a robust negative slope with a bootstrap mean of −37 . 6° ( green line ) . The perceptual detectability of the regular target rhythm improves over time , following a pattern that is highly correlated with the neural buildup of the signal representation . Consistent with previous findings of buildup of auditory stream segregation [24] , [35] , [48] , [49] , participants' performance during the target task improves significantly over several seconds as shown in Figure 5A ( solid orange line ) ( bootstrap across participants , p<10–4 ) . This similarity suggests that target detection is mediated by top-down mechanisms analogous to those employed in auditory streaming and object formation [50] . These streaming buildup effects tend to operate over the course of a few seconds , and cannot be explained by attentional buildup dynamics reported to be much faster or much slower in time [51] , [52] . Moreover , the neural response to the target rhythm also displays a statistically significant buildup ( Figure 5A , dashed red line ) ( bootstrap across participants , p = 0 . 02 ) closely aligned with the behavioral curve , and consequently , decoupled from the actual acoustics . The remarkable correspondence between these two measures strongly suggests that the enhanced perception of the target over time is mediated by an enhancement of the neural signal representation , itself driven by an accumulation of sensory evidence mediated by top-down mechanisms . No such neural buildup of the neural response to the target rhythm is present for the masker task . The MEG magnetic field distributions of the target rhythm response component in Figure 5A ( insets ) , showing the stereotypical pattern of neural activity originating separately in left and right auditory cortex , illustrate the changing strength of the neural activity over time in an individual participant . We confirm the correlation within participants between the psychometric and neurometric curves over time by running a bootstrap analysis on a per-participant basis . As expected , the slope correlating the d-prime and neural response curves for each participant yield a mean positive slope angle of 34 . 3°; bootstrap across participants shows a mean of 32 . 7° , with the 5th to 95th confidence intervals falling within the upper-right quadrant ( Figure 5B ) . We also note that the subsegments over which the neural buildup is measured are required to span several rhythmic periods ( at least three; see Figure 6A ) . There is no buildup using intervals with shorter durations , despite sufficient statistical power . ( This can be shown via the data plotted in the dashed curve in Figure 6A . The normalized responses in the range 3 . 5 to 4 . 5 are elements of an F ( 2 , 180 ) distribution , corresponding to p-values in the range 1 . 5% to 3 . 5% . ) This implicates temporal phase coherence ( in contrast to spatial phase coherence ) as critical to the buildup of the neural target representation . That is , the power in each period is not increasing , but the power integrated over several periods is increasing . This can only occur if the phase variability decreases with time , i . e . , the neural buildup is due to a buildup in temporal phase coherence rather than power . As noted above , the subsegments , or windows , over which the neural buildup is measured are required to span at least three rhythmic periods , since there is no buildup observed using intervals with shorter durations . Figure 6A illustrates this buildup for both the three-cycle and one-cycle cases . The requirement of a longer time window shows that the buildup is not merely due to increased power at 4 Hz , since in that case , a window of one rhythmic period would also show buildup . This in turn implies that temporal phase coherence ( in contrast to spatial phase coherence ) is critical to the buildup of the neural target representation . This is further demonstrated by a quantitative model . Typical simulated response profiles generated by the model are shown in Figure 6B . The horizontal axis in the model is not increasing time , but decreasing variability of the distribution of phase of the 4-Hz signal ( i . e . , the phase of the signal has greater variability initially and gets more regular as one proceeds along the axis ) . In the three-cycle window case , the buildup is pronounced , but not in the one-cycle window case . Note that the model does not attempt to emulate the downturn at the end of the experimental curve , nor does it attempt to emulate the rate at which the buildup occurs as a function of time ( which would assume a linear decrease in temporal phase coherence over time ) . The model results show that buildup can be due to increasing temporal coherence , and not due to increasing power . The neural noise , representing the stochastic firing patterns of the neurons underlying the MEG signal , is also required for the model's agreement with the data . A slight rise in the model's one-cycle window case may be seen , but it is not due to power ( which never changes ) , rather it is due to increased coherence over trials , which is a weak side effect of increased temporal coherence . This study's novel experimental paradigm builds on previous work in stream segregation using simpler stimuli [34] , [35] , [53] , [54] , but ( 1 ) using a richer stimulus design and ( 2 ) keeping the physical parameters of the stimulus fixed while manipulating only the attentional state of the listeners . One major finding is that auditory attention strongly modulates the sustained ( steady-state ) neural representation of the target . Specifically , sustained attention correlates with a sustained increase in the time-varying neural signal , in contrast with onset transients [22] , [55] or nonspecific , constant ( “DC” ) [56]–[58] effects of attention on auditory signals . The location of the modulated neural representation is consistent with core auditory cortex , hence supporting current evidence implicating neuronal mechanisms of core auditory cortex in the analysis of auditory scenes [35] , [59]–[61] . Furthermore , this modulation of neural signal is significantly distant from the source of the M100 and so cannot be explained as simply a train of repeated M100 responses . This steady-state increase in the signal strength is specific to the frequency of the target rhythm , and is additionally complemented by an enhancement in coherence over distant channels , reflecting an increased synchronization between distinct underlying neural populations . This attentional effect ( in both power and phase ) appears exclusively at the target frequency and is absent not only from other frequency bands whose intrinsic rhythms and induced response might show attentional changes , but even from adjacent frequency bins , which argues against any theory of neural recruitment or redistribution of energy at the low-frequency spectrum . Therefore , our findings argue that processes of attention interact with the physical parameters of the stimulus , and can act exclusively to enhance particular features to be attended to in the scene , with a resolution of a fraction of a hertz . Our analysis focuses on steady-state components of feature-based analysis , hence , complementing event-based analyses that relate temporal components of the recorded potential to specific mechanisms of feature-based attention [62]–[66] . Second , the data reveal that enhanced acoustic saliency ( driven by bottom-up processes ) , which causes an increase in perceptual detectability , also correlates with an increase in the sustained power and coherence of the neural signal . In this case , the increase in neural signal occurs regardless of the task being performed , but with different behavioral consequences: in the target task , it leads to an increase in performance , but in the masker task , a decrease ( via interference ) . This outcome allows us to give different explanations of this “attentionally modulated” neural change: as a marker of object detectability during the first task , but as a neural correlate of perceptual interference during the second task . Third , the data show a left-hemisphere bias in the cortical representation of the target , for the target task , suggesting a functional role of the left hemisphere in selective attention , consistent with previous findings in visual [67] and auditory [61] , [68] modalities . This bias may also be due to a left-hemisphere bias specific to Heschl's gyrus ( the location of core auditory cortex ) , for slow rhythmic tone pips ( without a masker background ) , as seen in [69] , [70] . In contrast , for the masker task , the hemispheric bias in cortical representation of the ( now nonattended ) target is reversed to the right , and might be simply due to the nature of the attentional demands of the task ( more diffuse attention to the global structure of the sound ) , or to the right-hemispheric bias of steady-state responses when attention is not specifically directed to the rhythm [71] . It also appears , for both tasks , that the deviant detection itself is not guiding the lateralization of the response , running counter to that of Zatorre and Belin [72] , since the task/deviant requiring spectral change detection shows a left-hemisphere bias , and the task/deviant requiring temporal change detection shows a right-hemisphere bias . Finally , this study offers the first demonstration of the top-down–mediated buildup over time of the neural representation of a target signal that also follows the same temporal profile of the buildup based on listeners' detectability performance in the same participant . Using the current experimental paradigm , we are able to monitor the evolution in time of attentional processes as they interact with the sensory input . Many studies overlook the temporal dynamics of the neural correlates of attention , either by using cues that prime participants to the object of attention ( thereby stabilizing attention before the onset of the stimulus ) , or by explicitly averaging out the buildup of the neural signal in their data analysis ( focusing instead on the overall contribution of attention in different situations , and not monitoring the dynamics by which the process builds up ) . Our findings reveal that even though the sensory target signal is unchanged , attention allows its neural representation to grow over time , closely following the time course of the perceptual representation of the signal , within participants . Together , these findings support a view of a tightly coupled interaction between the lower-level neural representation and the higher-level cognitive representation of auditory objects , in a clear demonstration of auditory scene segregation: the cocktail party effect [73] . Our experimental paradigm allows both task-driven ( top-down ) and stimulus-driven ( bottom-up ) processes to guide perception . For listeners performing the target task , the target rhythm is the attended auditory object , a foreground stream to be separated from a noisy background . The masker task , requiring the listener to reverse the role of the foreground and background , allows the contrasting situation to be considered under otherwise identical acoustical conditions . This permits a controlled de-emphasis of the auditory role of the target rhythm , without the need for a “passive” listening condition under which the amount of the listener's attention is lessened , but actually unknown , and strongly variable across participants . The data suggest that new models of attention may be required , based on temporally coherent or locally synchronous neural activity rather than neural amplification [74] . The buildup of neural responses over time is seen only when integrated over several periods of the target rhythm , but not for individual periods . This result is difficult to explain using standard models of attention that rely solely on gain-based changes , or even on gain/spectral-sensitivity hybrid models [75] , [76] . Instead , a more plausible theory of neural mechanisms underlying the role of top-down attention in the buildup of perceptual streams would involve top-down projections acting in conjunction with the physical stimulus as regulators or clocks for the firing patterns of neuronal populations in auditory cortex . Another conceivable mechanism for this increase in temporal coherence may arise from general sharpening of temporal tuning , which would work for auditory streams far more complex than the regular stream presented here . The neural underpinnings of this bottom-up/top-down interaction are likely to mediate changes in the response profiles of cortical neurons , via mechanisms of synaptic and receptive field plasticity which have been shown to be gated by attention; whereby attention plays a crucial role in shifting cortical circuits from one state to another depending on behavioral demands [77]–[80] . We speculate that temporal patterns of neuronal firings are crucial in any scene segregation task to resolve the competition between attended and unattended objects , hence , delimiting the cognitive border between different streams . Overall , a significant outcome of this study is that it not only demonstrates a strong coupling between the measured neural representation of a signal and its perceptual manifestation , but also places the source of this coupling at the level of sensory cortex . As such , the neural representation of the percept is encoded using the feature-driven mechanisms of sensory cortex , but shaped in a sustained manner via attention-driven projections from higher-level areas . Such a framework may underlie general mechanisms of “scene” organization in any sensory modality . Nine participants ( six males; mean age 29 y , range 24–38 y ) participated in the psychoacoustic study . Eighteen participants ( 11 males; mean age 27 y , range 21–49 y ) participated in the MEG study . Three participants took part in both studies . Among the 18 participants in the MEG study , four participants were excluded from further analysis due to an excess of nonneural electrical artifacts or an inability to perform the tasks , leaving 14 participants ( eight males; mean age 27 y , range 21–49 y ) . All participants were right handed [81] , had normal hearing , and had no history of neurological disorder . The experiments were approved by the University of Maryland Institutional Review Board , and written informed consent was obtained from each participant . Participants were paid for their participation . The stimuli were generated using MATLAB ( MathWorks ) . Each trial was 5 . 5 s in duration with 8-kHz sampling . Every trial contained one target note , repeating at 4 Hz , whose frequency was randomly chosen in the range 250–500 Hz in two semitone intervals . The background consisted of random tones at a density of 50 tones/s , uniformly distributed over time and log-frequency ( except for the spectral protection region ) . The frequencies of the random notes were randomly chosen from the five-octave range centered at 353 Hz , in two semitone intervals , with the constraint that no masker components were permitted within a four , or eight , or 12 semitone around the target frequency ( the spectral protection region half-width ) . This random sampling of masker frequencies ensures a minimum spectral distance of two semitones between maskers , and keeps the probability of harmonically related maskers minimal . Masker and target tones were 75 ms in duration with 10-ms onset and offset cosine ramps . All masker tones were presented at the same intensity as the target tone . Fifteen exemplar stimuli were generated for each of the four condition types: null condition ( no deviants ) ; target condition ( one target deviant per stimulus ) ; masker condition ( one masker deviant per stimulus ) ; and combined condition ( one target deviant and one masker deviant , at independent times , per stimulus ) . Each target deviant was the displacement of a target note ( upward or downward ) by two semitones from the target frequency . Each masker deviant was a single 500-ms time window in which all masker tones were elongated from 75 ms to 400 ms . The temporal location of the deviant ( for both target and masker tasks ) was randomly distributed along the 5 . 5-s trial duration , with timing as indicated by the behavioral buildup curve in Figure 4 . In the psychoacoustic experiment , participants were presented with 180 stimuli ( three protection regions×four conditions×15 exemplars ) per task . The progression from one trial to the next was initiated by the participant with a button-press . In the MEG experiment , only the eight-semitone spectral protection region half-width was used , giving 60 stimuli ( 1 protection region × 4 conditions × 15 exemplars ) per task . The interstimulus intervals ( ISIs ) were randomly chosen to be 1 , 800 , 1 , 900 , or 2 , 000 ms . For each task , the participants were presented with three blocks , repeating the ensemble of 60 stimuli three times ( totaling 180 stimuli ) . Participants were allowed to rest after each block , but were required to stay still . The identical stimulus ensemble ( including identical ISIs in the MEG case ) was presented for both target and masker tasks . Depending on the task being performed , participants were instructed to listen for the presence of a frequency deviant in the target rhythm ( target task ) or a duration deviant in the masker ( masker task ) ; each task deviant was present in exactly half the trials . A model simulation to illustrate a mechanism of neural buildup was implemented in MATLAB ( MathWorks ) . The model simulates MEG responses by generating 4-Hz signals whose phase is a random variable with constant mean , additionally corrupted by additive Gaussian white noise . This noise represents neural variability inherent in the neural processing mechanisms underlying the MEG signal and is critical to the model ( external magnetic field noise had already been removed from the data by active filtering [41] and is not modeled ) . The normalized response power was calculated by the same method as in the experiment: concatenating 50 signals and normalizing the power at 4 Hz by the average power in the 3–5 Hz band , averaging the 20 best channels ( out of 100 simulated auditory channels ) , and then averaging that over simulation runs . This was done for five different distributions of the phase random variable with standard deviations ranging from 1/60 to 1/12 of a cycle . The model's level of Gaussian white noise was obtained by the biological requirement that the model's normalized response for the three-cycle window , with highest jitter , match the data's; the buildup rate as a function of decreased variability , as well as the average level of the normalized response for the one-cycle window , follow automatically . The simulated response results depend weakly on the number of channels simulated . The experimental MEG system records from 157 channels , but not all are strongly auditory . The results shown here use 100 simulated auditory channels .
Attention is the cognitive process underlying our ability to focus on specific aspects of our environment while ignoring others . By its very definition , attention plays a key role in differentiating foreground ( the object of attention ) from unattended clutter , or background . We investigate the neural basis of this phenomenon by engaging listeners to attend to different components of a complex acoustic scene . We present a spectrally and dynamically rich , but highly controlled , stimulus while participants perform two complementary tasks: to attend either to a repeating target note in the midst of random interferers ( “maskers” ) , or to the background maskers themselves . Simultaneously , the participants' neural responses are recorded using the technique of magnetoencephalography ( MEG ) . We hold all physical parameters of the stimulus fixed across the two tasks while manipulating one free parameter: the attentional state of listeners . The experimental findings reveal that auditory attention strongly modulates the sustained neural representation of the target signals in the direction of boosting foreground perception , much like known effects of visual attention . This enhancement originates in auditory cortex , and occurs exclusively at the frequency of the target rhythm . The results show a strong interaction between the neural representation of the attended target with the behavioral task demands , the bottom-up saliency of the target , and its perceptual detectability over time .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "neuroscience/cognitive", "neuroscience", "neuroscience/sensory", "systems" ]
2009
Interaction between Attention and Bottom-Up Saliency Mediates the Representation of Foreground and Background in an Auditory Scene
Simultaneous analysis of genetic associations with multiple phenotypes may reveal shared genetic susceptibility across traits ( pleiotropy ) . For a locus exhibiting overall pleiotropy , it is important to identify which specific traits underlie this association . We propose a Bayesian meta-analysis approach ( termed CPBayes ) that uses summary-level data across multiple phenotypes to simultaneously measure the evidence of aggregate-level pleiotropic association and estimate an optimal subset of traits associated with the risk locus . This method uses a unified Bayesian statistical framework based on a spike and slab prior . CPBayes performs a fully Bayesian analysis by employing the Markov Chain Monte Carlo ( MCMC ) technique Gibbs sampling . It takes into account heterogeneity in the size and direction of the genetic effects across traits . It can be applied to both cohort data and separate studies of multiple traits having overlapping or non-overlapping subjects . Simulations show that CPBayes can produce higher accuracy in the selection of associated traits underlying a pleiotropic signal than the subset-based meta-analysis ASSET . We used CPBayes to undertake a genome-wide pleiotropic association study of 22 traits in the large Kaiser GERA cohort and detected six independent pleiotropic loci associated with at least two phenotypes . This includes a locus at chromosomal region 1q24 . 2 which exhibits an association simultaneously with the risk of five different diseases: Dermatophytosis , Hemorrhoids , Iron Deficiency , Osteoporosis and Peripheral Vascular Disease . We provide an R-package ‘CPBayes’ implementing the proposed method . Given the summary statistics for a SNP across multiple traits , CPBayes estimates two different measures evaluating overall pleiotropic association and an optimal subset of associated traits underlying a pleiotropic signal . The evidence for aggregate-level pleiotropic association is given by the local false discovery rate ( locFDR ) and the Bayes factor ( BF ) . Let H0 denote the global null hypothesis of no association with any trait and H1 denote the global alternative hypothesis of overall association with at least one of the traits . Then locFDR is defined as the posterior probability of H0 being true conditioned on the summary statistics and BF is the ratio of the likelihood of summary statistics under H1 versus H0 . A small value of locFDR ( e . g . , 0 . 05 ) or a large value of Bayes factor ( e . g . , BF > 1 ) indicates an evidence for the overall pleiotropic association . We estimate the locFDR and BF based on the MCMC posterior sample of model parameters obtained by Gibbs sampling . The subset of traits which is selected as the set of non-null traits most often in the MCMC posterior sample is defined as the maximum a posteriori ( MAP ) estimate of the optimal subset of associated traits . CPBayes explicitly accounts for possible correlation between the effect estimates across traits . To estimate the correlation structure of the effect estimates , we use two different approaches—one based on the number of overlapping cases and controls between studies and another based on genome-wide summary statistics across traits . The correlation formulae based on the number of shared cases and controls ( Eq 6 in Material and methods ) is accurate when the genetic variant and environmental covariates are not associated with the traits of interest . However in real data , environmental covariates are expected to be associated with the primary traits . In such scenario , the GW summary statistics based approach is more robust with respect to estimating a reasonably accurate correlation structure . CPBayes also provides some more insights into a pleiotropic signal , e . g . , marginal trait-specific posterior probability of association ( PPAj ) , direction of associations , credible interval of true genetic effects , etc . Detailed description of CPBayes and a brief outline of ASSET , BH0 . 01 and GPA are provided in the “Material and methods” section and supporting information . We compare CPBayes and ASSET with respect to correctly detecting a signal of pleiotropy and the accuracy of selection of non-null traits underlying a pleiotropic signal in various simulation scenarios . We consider multiple case-control studies with or without shared controls [7] and a cohort study where the data on multiple disease states are available for a group of individuals [31 , 32] . First , we specify the simulation model to generate the phenotype and genotype data . After computing the summary statistics based on the simulated data , we assume that only the summary-level data are available . For case-control studies with overlapping subjects or a cohort study , we estimate the correlation structure of summary statistics based on Eq 6 ( Material and methods ) . For non-overlapping case-control studies , we consider a separate group of 7000 cases and 10000 controls in each study . For overlapping case-control studies , we consider a distinct set of 7000 cases in each study , and a common set of 10000 controls shared across all the studies . For each disease , we assume an overall disease prevalence of 10% in the whole population . While simulating the genotype data for multiple case-control studies , we assume the standard logistic model of disease probability conditioning on the genotype: P ( case | G ) = exp ( α + β G ) 1 + exp ( α + β G ) , where G is the genotype at the SNP of interest coded as the minor allele count ( 0 , 1 , 2 ) . We assume that the SNP is in Hardy-Weinberg equilibrium ( HWE ) . Let A ( minor ) and a be the two alleles at the SNP and p = P ( A ) . Under HWE , the genotype probabilities are: P ( AA ) = p2 , P ( Aa ) = 2p ( 1 − p ) , P ( aa ) = ( 1 − p ) 2 . Given the log odds ratio ( β ) and disease prevalence , we compute the probability of observing each genotype conditioned on the case/control status using Bayes theorem . Based on these conditional probabilities , we simulate the genotypes in cases and controls . For the cohort study , we consider 15000 individuals . First , we generate the genotype data at a quantitative trait locus ( QTL ) and continuous multivariate phenotype data using the simulation model in Majumdar et al . [31] that was adopted from Galesloot et al . [32] . Then we dichotomize each continuous phenotype ( liability ) to case-control status subject to an overall disease prevalence of 10% . We describe the simulation model in more details in S1 Text in supporting information . We emphasize that the simulation models used here are general in nature and independent of the modeling assumptions underlying CPBayes which are also very general and only require that the sample sizes of the participating GWAS should be sufficiently large to satisfy the standard asymptotic properties ( Material and methods ) . These simulation models were also used in Bhattacharjee et al . [7] , Galesloot et al . [32] , Majumdar et al . [31] . For overlapping case-control studies and a cohort study when the summary statistics are expected to be correlated , a combined strategy of CPBayes is implemented ( Material and methods ) . If the effect estimates are strongly correlated and a majority of the traits are associated with the risk locus ( non-sparse scenario ) , the Gibbs sampler underlying CPBayes may sometimes be trapped in a local mode of the posterior distribution of model parameters . To increase robustness for correlated summary statistics , CPBayes considers a joint strategy combining its uncorrelated and correlated versions . First , we implement CPBayes considering the correlation structure of the effect estimates . If the selected subset of non-null traits have the smallest univariate association p-values among all the traits , we accept the results; otherwise , we employ CPBayes assuming that the effect estimates are uncorrelated and accept the results obtained . To investigate the performance of CPBayes using real data , we analyzed multiple traits in the large Northern California Kaiser Permanente “Resource for Genetic Epidemiology Research on Adult Health and Aging” ( GERA ) cohort obtained from dbGaP [dbGaP Study Accession: phs000674 . v1 . p1] . We also analyzed this dataset using ASSET for an empirical comparison of the methods . We restricted our analysis to 62 , 318 European-American individuals , who constitute more than 75% of the dbGaP data . We tested 657 , 184 SNPs genotyped across 22 autosomal chromosomes for their potential pleiotropic effects on 22 case-control phenotypes in the GERA cohort ( S8 Table ) . Note that in the dbGaP data , the cancers are collapsed into a single variable ( any cancer ) . Therefore , we could only use an overall cancer categorization even though the genetic architecture is likely heterogeneous across different cancers . The phenotypes are correlated modestly with a maximum absolute value of correlation as 0 . 36 observed between Hypertension and Dyslipidemia . Before our analysis , we undertook the following QC steps . First , we removed individuals with: over 3% of genotypes missing; any missing information on covariates ( described below ) ; genotype heterozygosity outside six standard deviations; first degree relatives; or discordant sex information . This left us with 53 , 809 individuals . Next , we removed SNPs with: MAF < 0 . 01; 10% or more missingness; or deviation from HWE at a level of significance 10−5 . This leaves 601 , 175 SNPs that were tested for pleiotropic association by CPBayes and ASSET . We adjusted the analysis for the following covariates: age , gender , smoking status , BMI category and 10 principal components of ancestry ( PCs ) . We tested the single-trait association for each of 22 phenotypes by a logistic regression of the case-control status on the genotype incorporating the same set of adjusting covariates . We used SNP-trait effect estimates ( log odds ratios ) and their standard errors in CPBayes and ASSET . As the summary statistics are correlated here , we used the combined strategy of CPBayes and the correlated version of ASSET . Since we have environmental covariates in the GERA study , we estimated the correlation matrix of the effect estimates under the null using the GW summary statistics data [34] . First , we extracted all of the SNPs for which the trait-specific univariate p-value across 22 traits are > 0 . 1 . This ensures that each SNP is either weakly or not associated with any of the 22 phenotypes . Then we selected a set of 24 , 510 independent SNPs from the initial set of null SNPs by using a linkage disequilibrium ( LD ) threshold of r2 < 0 . 01 ( r: the correlation between the genotypes at a pair of SNPs ) . Finally , we computed the correlation matrix of the effect estimates as the sample correlation matrix of β ^ 1 , … , β ^ 22 across the selected 24 , 510 independent null SNPs . We also considered different SNP filtering thresholds and compared the resulting correlation matrices . We provide numerical results demonstrating that the estimated matrix was not sensitive to our primary choice of the thresholds . We also give numerical results indicating that the estimated correlation matrix based on the sample overlap counts ( Eq 6 ) may be biased . These numerical results and their interpretation are provided in more detail in S9 Table and S1 Text . We apply the conventional GW level of statistical significance 5 × 10−8 for ASSET . For CPBayes , a pre-fixed significance threshold of locFDR needs to be considered here . While concluding that a locFDR threshold of 5% indicates good evidence of association , selecting the most promising pleiotropic variants may require a more stringent threshold ( as with the frequentist p-value threshold 5 × 10−8 ) . Liley and Wallace [12] also highlighted this point and suggested using a more stringent threshold of conditional false discovery rate ( cFDR ) than the nominal levels ( e . g . , 0 . 05 or 0 . 01 ) . They used a FDR and cFDR cut-off in the order of 10−5 or 10−6 to make the analysis analogous to using a stringent threshold of p-value ( 5 × 10−8 ) . We applied various thresholds of locFDR and detected 610 ( locFDR < 10−2 ) , 537 ( locFDR < 10−3 ) , 523 ( locFDR < 10−4 ) , 481 ( locFDR < 10−5 ) , 442 ( locFDR < 10−6 ) , 417 ( locFDR < 10−7 ) and 380 ( locFDR < 10−8 ) SNPs , respectively . We note that locFDR is not used as extensively as p-value in practice . Unlike the commonly used p-value threshold 5 × 10−8 , different thresholds of locFDR ( or other FDR related measures ) have been used in different applications [8 , 9 , 12] . Since Liley and Wallace [12] used a FDR and cFDR cut-off in the order of 10−5 or 10−6 , here we report the results based on a similar threshold of locFDR as 10−6 which detected 442 SNPs . Of note , locFDR , FDR and cFDR are distinct by definition . Theoretically , locFDR is an upper bound of FDR [27] . ASSET detected 394 SNPs based on the p-value threshold 5 × 10−8 . We note that CPBayes and ASSET identified a common set of 322 SNPs based on these chosen significance thresholds . Many of the associated SNPs are expected to be in LD . It is challenging to report independent pleiotropic variants since the LD pattern across a chromosome is irregular and converting the conditional analysis approach used in single phenotype GWAS to select independently associated variants in multi-phenotype context is difficult . Hence for the sake of convenience , we undertook the following simplified approach to identify correlated LD blocks . For CPBayes ( ASSET ) on each chromosome , we first chose the associated SNP that has the minimum locFDR ( ASSET p-value ) and created a LD block around it using a threshold of r2 = 0 . 25 . Then we implement the same strategy on the remaining set of associated SNPs to identify the next LD block , and so on . A major limitation of this approach is that defining such discrete LD blocks may not be on par with the irregular LD pattern across a chromosome and choosing an appropriate threshold of r2 is also difficult . For CPBayes , 442 GW associated SNPs comprised 59 LD blocks , and for ASSET , 394 GW associated SNPs comprised 30 LD blocks . For each of 394 SNPs detected by ASSET , the optimal subset of non-null traits always included more than one trait . Within each LD block detected by ASSET , we chose the SNP associated with the maximum number of traits as the lead SNP of the block . If multiple SNPs in a LD block satisfy this criterion , the one with minimum p-value of pleiotropic association was selected . We present the results for the lead SNPs detected by ASSET only on chromosome 1 and 2 ( to save space ) in S11 Table . CPBayes selected more than one trait for 93 among 442 SNPs . Within each LD block identified by CPBayes , we chose the SNP associated with the maximum number of phenotypes as the lead SNP of the block . If multiple SNPs satisfy this criterion , we chose the one having the minimum locFDR . In addition , if every SNP in a LD block is associated with one trait , we chose the SNP which provided the maximum number of traits having the marginal trait-specific posterior probability of association ( PPAj ) > 20% ( these traits were termed important phenotypes ) . Again , if multiple SNPs satisfy this criterion , we chose the one having the minimum locFDR . We note that the strategy of choosing the lead SNP in a LD block identified by CPBayes and ASSET are similar but technically different . As CPBayes locFDR and ASSET p-value are not directly comparable , it is very difficult to contrast the two methods with respect to the power of detecting aggregate-level pleiotropic association in a real data application . However , it makes sense to contrast the selection of non-null traits at a pleiotropic variant detected by both the methods . In Table 3 , we present the results for the independent pleiotropic SNPs at which CPBayes selected at least two phenotypes . At some of the lead SNPs detected by CPBayes , some phenotypes produced a non-negligible value of PPAj but were left out of the optimal subset of non-null traits . In Table 4 , we list these SNPs and the corresponding important phenotypes having a PPAj > 20% . In S10 Table , we report the independent SNPs at which CPBayes selected one trait . In the tables for CPBayes , we present PPAj and the direction of association ( genotype was coded as the number of the wild allele ) for the selected phenotypes ( Material and methods ) . In all the tables for CPBayes and ASSET , we also provide the trait-specific univariate association p-values . Many of the pleiotropic variants detected by CPBayes and ASSET are already reported in the NHGRI-EBI GWAS catalog . For example , rs6025 at 1q24 . 2 ( Table 3 ) has been associated with inflammatory bowel disease and venous thromboembolism; rs10455872 at 6q25 . 3 ( Table 3 ) has been associated with myocardial infarction , response to statins ( LDL cholesterol change ) , coronary artery disease , and aortic valve calcification; rs1410996 at 1q31 . 3 ( Table 4 ) is reported to be associated with Post bronchodilator FEV1/FVC ratio in COPD , End-stage coagulation and Age-related Macular Degeneration; rs4506565 at 10q25 . 2 ( Table 4 ) has been associated with Fasting glucose-related traits and Type 2 Diabetes . For a majority of the SNPs detected by ASSET , the subset of non-null traits included many phenotypes that have large univariate association p-values . For example , rs77394225 at 1q31 . 3 was detected by both the methods ( S10 and S11 Tables ) ; ASSET selected 10 traits , 9 of which had univariate p-value ≥ 0 . 3 and one ( Macular Degeneration ) had p-value = 6 . 04 × 10−16 ( S11 Table ) . In contrast , CPBayes only selected Macular Degeneration ( S10 Table ) . This suggests that CPBayes selects only those phenotypes with strong genetic associations , while ASSET may select many more traits with lower specificity as seen in our simulation study . For the independent pleiotropic SNPs identified by both the methods , for contrast’s sake , we applied BH0 . 01 . At rs77394225 , BH0 . 01 only selected Macular Degeneration which is consistent with CPBayes . Among the nine independent pleiotropic SNPs on chromosome 1 and 2 detected by ASSET , BH0 . 01 selected more than one trait only for two SNPs , whereas ASSET selected multiple traits for each of them ( S11 Table ) . This again indicates lower specificity of ASSET . CPBayes detected six independent pleiotropic SNPs that were associated with at least two phenotypes ( Table 3 ) . For example , at rs6025 ( 1q24 . 2 ) , it selected a maximum of 5 phenotypes: Dermatophytosis , Hemorrhoids , Iron Deficiency , Osteoporosis and Peripheral Vascular Disease , which have univariate p-values equal to 0 . 0018 , 0 . 0014 , 0 . 0004 , 0 . 0002 and 6 . 81 × 10−14 , respectively ( Table 3 ) . Interestingly , this SNP was positively associated with Dermatophytosis , Hemorrhoids and Iron Deficiency , but negatively associated with Osteoporosis and Peripheral Vascular Disease ( Table 3 ) . rs10455872 at 6q25 . 3 appeared to be the lead SNP for both the methods in their corresponding LD block on chromosome 6 . In Fig 4 , we contrast the selection of traits at rs10455872 between the methods . CPBayes selected Cardiovascular Disease , Dyslipidemia and Peripheral Vascular Disease ( Fig 4 ) . ASSET selected these three traits and five more phenotypes with large univariate p-values indicating weak genetic effects ( Fig 4 ) . CPBayes detected two independent pleiotropic SNPs in the chromosomal region 6p21 . 32: rs13211628 which was associated with Asthma , Cancers and Dyslipidemia , and rs3957148 which was associated with Asthma , Type 2 Diabetes and Macular Degeneration ( Table 3 ) . The r2 value between rs13211628 and rs3957148 was 0 . 03 . We provide a circos plot in Fig 5 presenting 23 pair-wise trait-trait pleiotropic association signals detected by CPBayes . It shows that CPBayes detected rs7601401 at 2p16 . 1 associated with Osteoarthritis and Abdominal Hernia; rs687289 at 9q34 . 2 associated with Dyslipidemia , Type 2 Diabetes and Peripheral Vascular Disease . We also present forest plot for some of the independent pleiotropic signals ( at rs6025 , rs13211628 , rs10455872 , rs3957148 , rs687289 ) detected by CPBayes in S11–S15 Figs . For these six independent pleiotropic SNPs , BH0 . 01 selected the same subset of traits as CPBayes . For CPBayes , the marginal trait-specific posterior probability of association ( PPAj ) provides a better insight into the relative strength of association between a pleiotropic variant and the selected non-null traits . For example at rs6025 , PPAj for Dermatophytosis , Hemorrhoids , Iron Deficiency , Osteoporosis and Peripheral Vascular Disease are 67% , 71% , 97% , 94% and 100% , respectively ( Table 3 ) . This implies that the association with Peripheral Vascular Disease is the strongest among the five selected phenotypes . At some of the GW significant SNPs detected by CPBayes , it produced a non-negligible value of PPAj for some traits; but these traits were left out of the optimal subset of non-null traits ( Table 4 ) . For example at rs4506565 ( 10q25 . 2 ) , CPBayes only selected Type 2 Diabetes , but Dyslipidemia also produced a PPAj of 52% . Thus , even though the effect of rs4506565 on Dyslipidemia was not strong enough to make it into the optimal subset , a further consideration of the pleiotropic effect of rs4506565 on Type 2 Diabetes and Dyslipidemia looks promising . We observed similar pattern across the other pleiotropic variants listed in Table 4 . We note that the combined strategy of CPBayes used the uncorrelated version only for one SNP among all the 601 , 175 SNPs analyzed . We have proposed a Bayesian meta-analysis approach CPBayes for pleiotropic association analysis based on summary-level data . It simultaneously evaluates the evidence of aggregate-level pleiotropic association and estimates an optimal subset of traits associated with the risk locus under a unified Bayesian statistical framework . The method is implemented by Gibbs sampling designed for both uncorrelated and correlated summary statistics . We have conducted an extensive simulation study and analyzed the large GERA cohort for evaluating the performance of CPBayes . An appealing feature of CPBayes is that , in addition to locFDR , Bayes factor , and an optimal subset of non-null traits , it simultaneously provides other interesting insights into an observed pleiotropic signal . For example , it estimates a trait-specific posterior probability of association ( PPAj ) , the direction of association , posterior mean/median and the credible interval of the unknown true genetic effect across traits . PPAj quantifies the marginal probability of each trait being associated with a pleiotropic variant . As demonstrated in the real data application , even if CPBayes does not select a phenotype in the optimal subset of non-null traits , PPAj for the phenotype may not be negligible . This may help an investigator to better explain a pleiotropic signal . One can also define the optimal subset of associated traits as {Yj: PPAj > p} , where p can be chosen as 0 . 5 ( known as the median model ) , or other values . Moreover , the joint posterior probability of association for a particular subset of traits can be calculated . Such flexibility in making inference on pleiotropy is mainly due to the MCMC construction underlying CPBayes . CPBayes selects the non-null traits underlying a pleiotropic signal with higher accuracy than ASSET . CPBayes performs the selection probabilistically through updating the latent association status by MCMC . ASSET selects that subset of traits as non-null which maximizes the observed value of a weighted linear combination of the normalized univariate association statistics corresponding to the phenotypes belonging to a subset . So given the summary statistics , ASSET does not select the non-null traits probabilistically based on the distribution of the summary statistics . For ASSET , even if a trait having a small genetic effect contributes a little to a pleiotropic signal , it is included in the optimal subset of associated traits . But CPBayes considers only those traits as non-null which substantially contribute to a pleiotropic signal . For example in the real data application , at rs10455872 , allergic rhinitis had an estimated odds ratio 1 . 05 ( univariate association p-value 0 . 07 ) . ASSET included this phenotype in the optimal subset of non-null traits as it might contribute to the overall signal of pleiotropic assocition . However , CPBayes estimated PPAj for this trait as 2 . 4% ( S13 Fig ) . So its contribution to the pleiotropic signal was not necessarily null , but CPBayes did not include it in the optimal subset of non-null traits as the effect was weak . We also note that ASSET is based on the framework of a fixed effects meta-analysis and assumes that the effects in a given direction ( positive/negative ) have the same size . But we observed in our real data application that , in a given direction , the effects of a variant across phenotypes may often be heterogeneous . CPBayes allows for heterogeneity simultaneously in the direction and size of the effects . S2 Table summarizes the key features of CPBayes and ASSET . While assessing the selection accuracy , we have placed more emphasis on specificity than sensitivity . This was because a higher sensitivity at the expense of a lower specificity can lead to a false selection of too many traits as non-null . CPBayes consistently maintained a very good level of specificity while offering a good level of sensitivity across a wide range of simulation scenarios . While CPBayes produced a limited number of independent pleiotropic SNPs associated with more than one phenotype in the analysis of GERA cohort , these pleiotropic signals seem very promising . The subsets of non-null traits selected by BH0 . 01 in the GERA cohort were consistent with CPBayes but not with ASSET which indicates that the non-null traits for a pleiotropic variant selected by CPBayes may be more reliable than ASSET . Here we note that , BH0 . 01 only facilitates the selection of associated traits underlying a pleiotropic signal but can not test for the evidence of overall pleiotropic association as CPBayes and ASSET . While CPBayes and ASSET estimate the measure of aggregate-level pleiotropic association and subset of non-null traits simultaneously under the same framework , BH0 . 01 has to be implemented in a separate step for a GW significant pleiotropic variant . Hence , CPBayes is a substantially more complete statistical tool for pleiotropy analysis than BH0 . 01 . While evaluating the selection accuracy of different approaches by simulations , we simultaneously obtained the measures of overall pleiotropic association provided by CPBayes and ASSET across 500 replications . We present various summary measures of these in some selected simulation scenarios ( to save space ) and corresponding brief interpretation in S3 , S4 and S5 Tables and S1 Text . We also carried out simulations for 50 traits . Since ASSET is computationally very slow for 50 traits due to an extremely large number of possible subsets of traits , we only implemented CPBayes and BH0 . 01 . CPBayes performed consistently well similarly as for smaller number of traits . See S1 Text and S6 and S7 Tables for more details . Note that the continuous spike inherits the assumption that a SNP contributes to the variation of all traits under consideration , and the distinction is made between a negligible and a significant contribution . In contrast , the Dirac spike assigns the null effects explicitly to zero . We conducted a simulation study ( see S1 Text and S10 Fig ) to compare the continuous spike and Dirac spike . We found that the continuous spike offers better accuracy in the selection of non-null traits than the Dirac spike . The continuous spike is also computationally much faster ( 2-3 times ) than the Dirac spike . Hence , we adopted the continuous spike for constructing CPBayes . In a related work , Han and Eskin [35] proposed a modified random effects meta-analysis for combining heterogeneous studies coupled with a Bayesian approach to provide a better interpretation of an observed signal of aggregate-level association . They investigated how to combine heterogeneous genetic studies across different populations/ethnicities . However , they did not address how to account for a possible correlation between the summary statistics while selecting the most important studies underlying an observed signal of aggregate-level association . Moreover , they assumed that the non-null effects are similar across studies which is less likely to hold in the context of pleiotropy . Hence we compared CPBayes with ASSET and GPA . We note that the CPBayes locFDR ( Bayes factor ) and ASSET p-value are not directly comparable . In our simulation study , we adopted the strategy outlined by Coram et al . [33] to compare CPBayes and ASSET with respect to the number of true positives detected while maintaining the FDR at a pre-specified threshold . Another possible approach to quantify the false positive rate could be the permutation-based strategy suggested by Servin and Stephens [36] . However in our context , such an approach is computationally too expensive as it requires the MCMC underlying CPBayes to be implemented for each permuted dataset . Of note , we did not conduct a replication study and all the pleiotropic association signals obtained from the Kaiser cohort are reported based on the analysis only in the discovery sample . Even though we demonstrated CPBayes only for binary traits , we note that CPBayes can also be applied to non-binary traits . We carried out simulations for continuous traits ( distributed as multivariate normal ) , and observed a similar pattern between the performance of CPBayes and ASSET as for binary traits . Another useful approach to pleiotropy analysis is MultiPhen [37] which can accommodate general types of traits . However , we chose ASSET as the main competing method because it simultaneously provides an optimal subset of associated traits along with a measure of aggregate-level pleiotropic association , which provides a direct comparison to CPBayes . In contrast , MultiPhen does not facilitate the simultaneous selection of the optimal associated traits underlying a pleiotropic signal . Of note , MultiPhen requires individual level phenotype and genotype data . In the analysis of GERA cohort , we assumed that the correlation matrix of estimated effect sizes is the same across SNPs , for each of which , we converted this correlation matrix to its covariance matrix by incorporating its standard error across traits . Hence , if the standard error vector varies across SNPs , the covariance matrix of the effect estimates also varies . We did some simulations to assess this assumption of constant correlation matrix . Consider the simulation framework of a cohort study with five case-control phenotypes designed to evaluate the partial ROC curves for CPBayes . Here the continuous traits ( liability ) underlying the binary traits were generated at random for every SNP following the simulation model in S1 Text; hence the binary traits dataset varied across SNPs . So the correlation matrix of effect estimates obtained by Eq 6 also varied across SNPs . We implemented CPBayes for each SNP using the corresponding effect estimates’ correlation matrix obtained by Eq 6 which is accurate for the null SNPs as there are no environmental covariates here . In the simulated data for a single risk SNP , we anticipate that the real correlation matrix of the effect estimates will be very close to the sample overlap correlation matrix ( Eq 6 ) . Because , a single risk SNP usually explains a very small proportion of total heritability for a complex trait and the genetic correlation between two traits due to a single risk SNP is expected to be very small . Next we estimated a constant correlation matrix as the sample correlation matrix of the observed effect estimates for the null SNPs ( following the GW strategy in GERA cohort analysis ) , and implemented CPBayes for all the SNPs using this constant correlation matrix . We observed that the partial ROC curve for CPBayes obtained by using varying correlation matrices and constant correlation matrix across SNPs had almost the same AUC . We also found by simulations that using the constant correlation matrix provides a very similar selection accuracy compared to using varying correlation matrix across SNPs . We repeated these simulations for a cohort study with five continuous phenotypes distributed as multivariate normal . Since the phenotypes are normally distributed , the real correlation matrix of effect estimates can be analytically computed . Again , CPBayes performed very similarly using constant and varying correlation matrices . If a set of binary traits are measured on separate independent group of individuals , one would expect the summary statistics across traits be independent . However , if these traits are mutually exclusive because of competing risks , risk SNPs for one trait may be underrepresented among the cases of the other trait , leading to a correlation in the summary statistics , especially for high penetrance variants . For common and complex traits , competing risks may not result in mutual exclusivity , and may lead to very limited correlation among subjects from independent samples . Moreover , since such traits arise from many different low risk SNPs , any corresponding correlation among summary statistics would also be limited . Taken together , we expect that this would have negligible impact on our assessment of pleiotropy , although separate studies may consider this possibility further . Since CPBayes individually analyzes each SNP using marginal summary statistics across traits , it can not distinguish between pleiotropy and co-localization . Any marginal SNP-level meta analysis approach including ASSET also has this limitation . We analyzed 22 traits in the GERA cohort . Ideally one would only include genetically correlated traits in a pleiotropy analysis to maximize the power of a multi-trait approach . However , determining a priori which traits are genetically related can be challenging . This could be based on the literature or estimated from one’s own data . In the latter situation , one must be cognizant of potential bias due to empirically determining genetic correlations . To date , most pleiotropy analyses have focused on a small set of context-specific traits ( e . g . , psychiatric disorders , cancers ) . Expanding to larger numbers of disparate traits may provide important insights to shared biological mechanisms . In general , the selection of traits to consider for pleiotropy analyses can be based on co-heritability ( genetic correlation ) analyses , existing literature , and biological / clinical expertise . In future work , we aim to investigate whether the computing speed of CPBayes can be increased by using a variational Bayes approach or by using an optimization technique ( e . g . , EM algorithm or its variants ) instead of using MCMC , while preserving the efficiency of the method . Also , we want to explore how to relax the assumption in CPBayes that s j 2 is a reasonably accurate estimate of σ j 2 which requires a larger sample size to be satisfied . In summary , CPBayes is an efficient Bayesian meta-analysis approach to simultaneously analyze pleiotropy for two or more traits . It has a strong theoretical foundation and allows for heterogeneity in both the direction and size of effects . One can implement it for both cohort data and separate studies of multiple phenotypes having non-overlapping or overlapping subjects . In addition to parameters of primary interest ( e . g . , the measures of overall pleiotropic association , the optimal subset of associated traits ) , it provides other interesting insights into a pleiotropic signal ( e . g . , the trait-specific posterior probability of association , the direction of association , the credible interval of unknown true genetic effect across traits ) . It is computationally feasible and a user-friendly R-package ‘CPBayes’ is provided for general use . The continuous spike and slab prior in our context [19 , 20] is described as follows: for j = 1 , … , K , βj|zj , τ , di n d ˜ ( 1−zj ) ×N ( 0 , τ2 ) +zj×N ( 0 , ( τd ) 2 ) ;τ>0 , 0<d<1 , ( τd ) 2>τ2P ( zj=1|q ) =q;P ( zj=0|q ) = ( 1−q ) ;0<q<1q|c1 , c2~Beta ( c1 , c2 ) ;d|e1 , e2~Beta ( e1 , e2 ) ( 1 ) The latent variable zj denotes the association status of Yj . When zj = 0 , βj ∼ N ( 0 , τ2 ) , and when zj = 1 , β j ∼ N ( 0 , ( τ d ) 2 ) , where ( τ d ) 2 > τ 2 . The usefulness of such a formulation is that τ can be set small enough so that , if zj = 0 , |βj| would probably be very small to safely be considered as zero ( Yj is not associated with the SNP ) , and d can be chosen sufficiently small ( so 1 d > > 1 ) such that , if zj = 1 , βj can be considered as non-zero ( Yj is associated with the SNP ) . The proportion of traits having a non-null genetic effect is denoted by q . For simplicity and reduction in computational cost , we consider τ as fixed . We choose e1 = e2 = 1 which correspond to the uniform ( 0 , 1 ) distribution . The parameter d is updated in a given range so that the slab variance ( τ d ) 2 ( i . e . , variance of non-null effects across traits ) varies in a pre-fixed range . We describe the continuous spike and slab prior in the context of modeling pleiotropy with diagrams in S1 and S2 Figs . The Dirac spike and slab prior in current context [18 , 20] is given by: for j = 1 , … , K , β j | q , b i . i . d . ˜ ( 1 − q ) × δ { 0 } ( β j ) + q × N ( 0 , b 2 ) q | c 1 , c 2 ~ Beta ( c 1 , c 2 ) , 0 < q < 1 ( 2 ) Here , δ{0} ( βj ) = 1 if βj = 0 , and δ{0} ( βj ) = 0 if βj ≠ 0 . So under no association , βj = 0 . The proportion of associated traits is given by q . To perform a fully Bayesian analysis , we implement MCMC by the Gibbs sampling algorithm to generate posterior samples of the model parameters based on which we draw the statistical inference for pleiotropy . We derive the Gibbs samplers for both uncorrelated and correlated summary statistics . Here we describe the inference procedure for the continuous spike . The Gibbs sampling algorithm for the continuous spike ( Algorithm 1 ) is outlined later in this section , and the algorithm for Dirac spike ( Algorithm S1 ) is stated in supporting information ( S1 Text ) . The mathematical derivation of the full conditional posterior distributions underlying the Gibbs samplers are also given in supporting information ( S1 Text ) . Let {β ( i ) , Z ( i ) , q ( i ) , d ( i ) ; i = 1 , … , N} denote N posterior samples of ( β , Z , q , d ) obtained by MCMC after a certain burn-in period . We have used a burn-in period of 5000 and MCMC sample size of 15000 in our simulation study and the real data application . First , we want to test the global null hypothesis of no association ( H0 ) against the global alternative hypothesis of association with at least one trait ( H1 ) . Since , for the continuous spike , the latent association status distinguishes between an association being null or non-null , we set H0: z1 = … = zK = 0 ( Z = 0 ) versus H1: at least one of z1 , … , zK = 1 ( Z ≠ 0 ) . The summary statistics across traits can be correlated due to overlap or close genetic relatedness among subjects across different studies . For case-control studies , Zaykin and Kozbur [40] and Lin and Sullivan [41] derived a simple formula of correlation among β ^ 1 , … , β ^ K . For k , l ∈ {1 , … , K} and k ≠ l , c o r r ( β ^ k , β ^ l ) = ( n k l ( 11 ) n k ( 0 ) n l ( 0 ) n k ( 1 ) n l ( 1 ) + n k l ( 00 ) n k ( 1 ) n l ( 1 ) n k ( 0 ) n l ( 0 ) ) / n k n l ( 5 ) Here n k ( 1 ) , n k ( 0 ) , and nk ( or n l ( 1 ) , n l ( 0 ) , and nl ) denote the number of cases , controls , and total sample size for the study of Yk ( or Yl ) ; n k l ( 11 ) and n k l ( 00 ) denote the number of cases and controls shared between the studies of Yk and Yl . Let n k l ( 10 ) be the number of overlapping subjects that are cases for Yk but controls for Yl; similarly , let n k l ( 01 ) be the number of shared subjects that are controls for Yk but cases for Yl . Here , the above formula can be generalized to: c o r r ( β ^ k , β ^ l ) = ( n k l ( 11 ) n k ( 0 ) n l ( 0 ) n k ( 1 ) n l ( 1 ) − n k l ( 10 ) n k ( 0 ) n l ( 1 ) n k ( 1 ) n l ( 0 ) − n k l ( 01 ) n k ( 1 ) n l ( 0 ) n k ( 0 ) n l ( 1 ) + n k l ( 00 ) n k ( 1 ) n l ( 1 ) n k ( 0 ) n l ( 0 ) ) / n k n l ( 6 ) This formula is accurate when none of the phenotypes Y1 , … , YK is associated with the SNP and environmental covariates . An alternative strategy [5 , 34] is based on using GW ( genome-wide ) summary statistics data to estimate the correlation structure , which is useful when the environmental covariates are associated with the phenotypes . For continuous and normally distributed traits , the correlation matrix of effect estimates under the null is the phenotypic correlation matrix . But , its calculation requires individual level phenotype data across multiple traits . For general type of traits ( e . g . non-normal continuous traits , count phenotypes ) , a standard formula of correlation between the effect estimates may be difficult to derive . Such a formula may also require information only available from individual-level phenotype data . For example , the effect estimates’ correlation formula for binary traits requires the number of cases , controls in each study and the number of overlapping cases and controls between studies . In such scenarios , since the genome-wide ( GW ) effect estimates corresponding to multiple traits will be available in the pleiotropy analysis , the GW summary statistics based approach can be applied irrespective of the type of traits without requiring any individual-level phenotype data . So , the GW summary statistics based approach is useful to estimate the correlation structure of effect estimates in various scenarios including correlated non-binary traits . For strongly correlated summary statistics , when a majority of the traits are associated with the risk locus ( non-sparse scenario ) , the Gibbs sampler can sometimes be trapped in a local mode rather than the global mode of the posterior distribution due to possible multi-modality of the posterior distribution of model parameters . We observed this pattern in our simulation study . It may result in an incorrect selection of associated traits , reducing the robustness of CPBayes . We noticed that , in such a scenario , if the summary statistics are assumed to be uncorrelated , the MCMC does not get trapped in a local mode and moves around the global mode . But ignoring the correlation can give a lower ( larger ) Bayes factor ( locFDR ) and sensitivity of the selected traits . Hence , for correlated summary statistics , we combine the correlated and the uncorrelated versions of CPBayes as follows . First , we execute CPBayes considering the correlation among β ^ 1 , … , β ^ K . Let A denote the selected subset of non-null traits that contains K1 traits . Let B denote the subset of K1 traits that have the smallest univariate association p-values . If A and B match , we accept the results; otherwise , we implement CPBayes assuming that β ^ 1 , … , β ^ K are uncorrelated and accept the results obtained . Note that , if A is empty , we accept the results provided by the correlated version of CPBayes . In this combined strategy , we induce a frequentist sense of selection . Because , majority of the multiple testing procedures reject the null hypotheses for which smallest univariate p-values are obtained . However , we note that in the analysis of the GERA cohort , the combined strategy used the uncorrelated version very few times . The reason is that the non-sparse scenario may not occur frequently in real data . It is straightforward to observe that the Dirac spike can be obtained from the continuous spike by first setting τ = 0 and τ d = b in Eq 1 , and then integrating out the latent variables Z from the model . We note that , the latent association status ( Z ) could only be used in the model for the continuous spike . For the Dirac spike , the inclusion of Z in the model makes the corresponding MCMC reducible , and hence non-convergent to its stationary distribution ( details not provided ) . Also , for the continuous spike , the full conditional posterior distributions of z1 , … , zK are independent which leads to an explicit estimation of the locFDR/Bayes factor based on the MCMC sample . But , for the Dirac spike , the explicit estimation of the locFDR/Bayes factor appears to be very difficult in the correlated case , because the full conditional posterior distributions of β1 , … , βK are not independent for correlated summary statistics . Bhattacharjee et al . [7] introduced an elegant subset-based meta analysis method ASSET to analyze pleiotropy . While regressing kth phenotype Yk on genotype G , let β ^ k , s k be the estimates of the association parameter and its standard error , k = 1 , … , K . Adopting the framework of a fixed-effects meta analysis , for a subset of traits A , ASSET defines the Z statistic as: Z ( A ) = ∑ k : Y k ∈ A π k ( A ) Z k , where Z k = β ^ k s k , and π k ( A ) is an appropriate weight associated with Yk belonging to A . For example , if there are K separate GWAS for Y1 , … , YK with the kth study having a sample size nk , one can consider π k ( A ) = n k ∑ k ∈ A n k . The global association of a SNP with at least one trait is measured by the test-statistic: Z m a x = m a x A | Z ( A ) | , where the maximization is taken across all possible A . In addition to the p-value of global association , ASSET also offers an optimal subset of non-null traits that are associated with the SNP , which is essentially the subset of traits that constructs Zmax . For more details , see Bhattacharjee et al . [7] . Benjamini and Hochberg [30] introduced a sequential procedure that controls the expected FDR in multiple hypothesis testing . Majumdar et al . [31] demonstrated that the BH procedure is a simple but efficient strategy to select non-null traits underlying a pleiotropic signal . For each individual risk SNP associated with at least one trait , we applied the BH procedure to the univariate association p-values for the phenotypes under consideration with the level of FDR as 0 . 01 which was suggested by Majumdar et al . [31] . We refer it as BH0 . 01 . Here we state the Gibbs sampling algorithm for the continuous spike described in Eq 1 . It is a desirable practice to provide the MCMC with a good initial value of the model parameters for faster convergence to its stationary distribution . Hence , we apply BH0 . 01 on the univariate association p-values of K traits and assign zj = 1 if Yj is found to be significantly associated , otherwise set zj = 0; j = 1 , … , K . We also choose an initial value of q as the proportion of non-null traits detected by BH0 . 01 ( the boundary situations of no/all non-null traits are taken care of appropriately ) . Define Σ 2 = d i a g ( τ 1 2 , … , τ K 2 ) ( a diagonal matrix with diagonal elements τ 1 2 , … , τ K 2 ) , where τj = τ if zj = 0; and τ j = τ d if zj = 1; j = 1 , … , K . So β|Z∼MVN ( 0˜ , Σ2 ) . Let Σ1 = S . Also , let β−j = ( β1 , … , βj−1 , βj+1 , … , βK ) , and Z−j = ( z1 , … , zj−1 , zj+1 , … , zK ) . Algorithm 1 Gibbs sampling for continuous spike in correlated case 1: Start: 2: Assign the initial values of Z and q as described above . 3: loop: 4: Simulate β from its full conditional posterior distribution: β | Z , q , d , β ^ ∼ MVN [ ( Σ 1 - 1 + Σ 2 - 1 ) - 1 Σ 1 - 1 β ^ , ( Σ 1 - 1 + Σ 2 - 1 ) - 1 ] . 5: For j = 1 , … , K , update zj using the full conditional posterior probability: P ( z j = 0 | Z - j , β , q , d , β ^ ) = 1 1 + ratio j , where ratio j = q 1 - q d exp [ - β j 2 2 τ 2 ( d 2 - 1 ) ] . 6: Let k 1 = ∑ j = 1 K z j , k0 = K − k1 . Update q using q | β , Z , d , β ^ ∼ Beta ( c 1 + k 1 , c 2 + k 0 ) . 7: We assume that e1 = e2 = 1 . Update d from its full conditional posterior distribution in a fixed range so that the slab variance ( τ d ) 2 varies in a given range ( v0 , v1 ) ; let the corresponding range of d be given by: d0 < d < d1 . If k 1 = ∑ j = 1 K z j > 0 , then d = y 2 C , where C = 1 2 τ 2 ∑ j : z j = 1 β j 2 , and y follows a truncated ( 2 C d 0 2 < y < 2 C d 1 2 ) central χ k 1 + 1 2 distribution . If k1 = 0 , d is updated from the truncated ( d0 < d < d1 ) Beta ( 1 , 1 ) distribution . 8: goto loop until all the MCMC iterations are finished . We note that , d can be updated using the truncated central χ2 distribution as long as the second shape parameter of its Beta prior ( e2 ) is 1 . If the summary statistics are uncorrelated , step 4 of Algorithm 1 is modified as: for j = 1 , … , K , update βj by sampling from its full conditional posterior distribution: β j | β - j , Z , q , d , β ^ ∼ N ( σ j 2 s j 2 β ^ j , σ j 2 ) , where 1 σ j 2 = 1 s j 2 + 1 τ j 2 . All the other steps remain the same as in the Algorithm 1 .
Genome-wide association studies ( GWAS ) have detected shared genetic susceptibility to various human diseases ( pleiotropy ) . We propose a Bayesian meta-analysis method CPBayes that simultaneously evaluates the evidence of overall pleiotropy while determining which traits are pleiotropic . This approach investigates pleiotropy using GWAS summary statistics and allows for overlapping subjects across traits . It performs a fully Bayesian analysis and offers a flexible inference . CPBayes also provides additional information about a pleiotropic signal , such as the trait-specific posterior probability of association and the credible interval of unknown true genetic effects . Using computer simulations and an application to a large GWAS cohort , we demonstrate that CPBayes can offer improved accuracy compared to the existing subset-based meta-analysis approach ASSET . We provide a user-friendly R-package ‘CPBayes’ for general use of this approach .
[ "Abstract", "Introduction", "Discussion", "Materials", "and", "methods" ]
[ "genome-wide", "association", "studies", "medicine", "and", "health", "sciences", "simulation", "and", "modeling", "diabetes", "mellitus", "research", "design", "dyslipidemia", "endocrine", "disorders", "genome", "analysis", "type", "2", "diabetes", "research", "and", "analysis", "methods", "case-control", "studies", "endocrinology", "peripheral", "vascular", "disease", "genetic", "loci", "metabolic", "disorders", "phenotypes", "genetics", "biology", "and", "life", "sciences", "genomics", "vascular", "medicine", "computational", "biology", "vascular", "diseases", "human", "genetics" ]
2018
An efficient Bayesian meta-analysis approach for studying cross-phenotype genetic associations
Major human pathologies are caused by nuclear replicative viruses establishing life-long latent infection in their host . During latency the genomes of these viruses are intimately interacting with the cell nucleus environment . A hallmark of herpes simplex virus type 1 ( HSV-1 ) latency establishment is the shutdown of lytic genes expression and the concomitant induction of the latency associated ( LAT ) transcripts . Although the setting up and the maintenance of the latent genetic program is most likely dependent on a subtle interplay between viral and nuclear factors , this remains uninvestigated . Combining the use of in situ fluorescent-based approaches and high-resolution microscopic analysis , we show that HSV-1 genomes adopt specific nuclear patterns in sensory neurons of latently infected mice ( 28 days post-inoculation , d . p . i . ) . Latent HSV-1 genomes display two major patterns , called “Single” and “Multiple” , which associate with centromeres , and with promyelocytic leukemia nuclear bodies ( PML-NBs ) as viral DNA-containing PML-NBs ( DCP-NBs ) . 3D-image reconstruction of DCP-NBs shows that PML forms a shell around viral genomes and associated Daxx and ATRX , two PML partners within PML-NBs . During latency establishment ( 6 d . p . i . ) , infected mouse TGs display , at the level of the whole TG and in individual cells , a substantial increase of PML amount consistent with the interferon-mediated antiviral role of PML . “Single” and “Multiple” patterns are reminiscent of low and high-viral genome copy-containing neurons . We show that LAT expression is significantly favored within the “Multiple” pattern , which underlines a heterogeneity of LAT expression dependent on the viral genome copy number , pattern acquisition , and association with nuclear domains . Infection of PML-knockout mice demonstrates that PML/PML-NBs are involved in virus nuclear pattern acquisition , and negatively regulate the expression of the LAT . This study demonstrates that nuclear domains including PML-NBs and centromeres are functionally involved in the control of HSV-1 latency , and represent a key level of host/virus interaction . Herpes simplex virus type 1 ( HSV-1 ) , a major human pathogen , is a persistent human neurotropic virus and a model of long-term interaction between a host cell and a parasite . HSV-1 establishes a long-term latent infection in neurons of the trigeminal ( or Gasserian ) ganglia ( TG ) of the peripheral nervous system , from which it reactivates periodically to replicate and spread [1] . The establishment of latency is dependent on a sequence of physiological and molecular events involving the host immune system , the cellular antiviral response , and the ability of the virus to initiate a latent gene expression program . Latent HSV-1 dsDNA genomes localize in the nucleus of the host neuron where they remain as multi-copy chromatinized episomes , which do not integrate into the host-cell genome [2] , [3] . During latency , HSV-1 lytic gene expression is strongly repressed; although some lytic transcripts could be detected at low level , by highly sensitive techniques [4]–[6] . The latency-associated transcript ( LAT ) locus is the only gene to be highly expressed throughout the persistent stage , from establishing latency to reactivation [7] . LAT is a noncoding RNA , synthesized as an 8 . 3-kb polyadenylated , unstable primary transcript , and is rapidly processed into a stable 2-kb intron lariat and several microRNAs [8]–[11] . LAT expression has been linked to several aspects of the latency process , including neuron survival , viral genome chromatin status , lytic gene expression , number of latently infected neurons , and efficiency of reactivation in animal models [2] , [3] , [10] , [12]–[17] . Although LAT appears to regulate latency and reactivation , several studies have shown that LAT is probably expressed only in a subset of latently infected neurons , implying that latency is intrinsically a heterogeneous event [18]–[23] . The heterogeneity of HSV-1 latency has also been observed at the level of the viral genome copy number in individual neurons , which has been directly correlated with reactivation probability , suggesting that it is a functionally significant parameter [19] , [20] , [24] . How these variable parameters impact on the biology of the latent virus and the reactivation process remains unclear . Moreover , host-cell factors and the cellular environment can be anticipated to also account for the variability of latency and for determining the ability of HSV-1 to reactivate . Therefore , the study of latency requires experimental approaches in which the heterogeneity can be fully assessed with regard to viral genome features , viral gene expression , and host-cell nuclear components . In situ fluorescence-based strategies offer such a possibility , through a multi-parametric reading of a cell population at the single-cell level . The mammalian cell nucleus is a highly organized compartment containing the chromosomes and several nuclear domains , which reflect the various molecular activities taking place in the nucleus . Numerous studies reported that the position of a gene within the nucleus is correlated with its transcriptional status [25] , [26] . The predetermined nuclear positions of genetic loci within the nuclear architecture are key determinants of gene expression , together with transcription factors and epigenetic chromatin modifications [25] , [27] , [28] . Nuclear structures known to influence gene expression include the nuclear envelope , telomeres , centromeres and pericentromeres , and nuclear domains such as promyelocytic leukemia ( PML ) nuclear bodies ( NBs , also called ND10 ) , transcription factories , polycomb group complexes , and the nucleolus [26] , [29]–[34] . Among these nuclear domains , PML-NBs are proteinaceous structures that reorganize in response to various cellular stressors [33] , [35] , [36] . PML-NBs provide a nuclear environment that can be associated with transcription of cellular genes [32] , [37] , [38] . However , PML-NBs contain repressor proteins such as HP1 , ATRX , and hDaxx [39] , [40] , which have an inhibitory effect on transcription and replication of RNA and DNA viruses , supporting the silencing activity of PML-NBs [41] , [42] . In cultured infected cells , the association of PML-NB with genomes of several viruses , including HSV-1 , has led to the hypothesis that PML-NBs may operate as a nuclear relay for innate host-cell defense mechanisms , blocking replicative infection by creating an environment unfavorable for viral gene expression [32] , [42]–[50] . However , how nuclear domains impact in vivo on the biology of persistent viruses such as HSV-1 and whether they may intervene in the latency process , in particular in the acquisition of essential parameters involved in latency maintenance and reactivation , is currently unknown . In this study , we took advantage of a physiologically well-characterized mouse model of HSV-1 infection , to develop an efficient fluorescent in situ hybridization ( FISH ) approach for detecting HSV-1 genomes during latency in neurons from infected mouse TG . Using a high-resolution visualization technique , we described the intra-nuclear distribution of the latent HSV-1 genome in neurons , and correlated HSV-1 patterns with LAT expression . We found that HSV-1 genomes were non-randomly associated with two nuclear domains , PML-NBs and centromeres . Using infected PML knockout ( KO ) mice , we showed that PML/PML-NBs influence viral genome distribution and negatively regulate expression of LAT . Finally , we demonstrated that HSV-1 genomes associated with PML-NBs or centromeres were negative for the expression of LAT . The lack of an efficient in situ detection method of viral genomes has been a major technical limitation to the study of herpes virus infection and disease both in animal models and human samples . Detection of HSV-1 genomes by FISH in latently infected mouse tissues has remained unsuccessful despite attempts of many groups [51] . To determine the intra-nuclear organization of the multiple copies of HSV-1 and its influence HSV-1 gene expression , we developed a DNA-FISH protocol and applied it to an established lip-inoculation mouse model in which HSV-1 establishes significant latency in the TG ( Figure 1A; [52] ) . Infected and mock-infected mice were sacrificed at 28 d . p . i . , a time point at which HSV-1 latency is known to be fully established [51] , and the TGs were cryo-sectioned . Our FISH protocol efficiently detected latent HSV-1 genomes in mouse neuronal tissues ( Figure 1D; see Materials and Methods for details ) . The DNA-FISH probes recognized a 90 kb region of the viral genome , excluding the LAT locus ( named hereafter “HSV-1 genome probes , ” Figure 1B ) . Importantly , our protocol did not include a signal amplification procedure and thus is well suited for the study of intra-nuclear organization by high-resolution microscopy . Signal specificity was assessed through several control experiments , including FISH analysis of mock-infected mice , FISH with control probes without HSV-1 sequences ( Figure 1C ) , and a comparison between our probe and a commercially available probe ( Figure S1 ) . In TG sections from infected mice sacrificed at 28 d . p . i . , the FISH signal for HSV-1 DNA was observed only in the nuclei of neurons , where it appeared as a dotted pattern comprising spots of various numbers , sizes , and intensities ( Figure 1D and S1 ) . The presence of the HSV-1 genome in neurons and not in satellite cells is consistent with the results of previous in situ PCR and single-cell PCR studies [20] , [24] . Two main intra-nuclear patterns were observed: a single , bright , round spot ( termed “single” ) and numerous spots of non-uniform size and shape ( termed “multiple” ) ( Figure 1D–E ) . Both patterns were observed in different inoculation models ( lip , eye , and whisker pads ) , animals ( mice of different inbred strains and rabbits ) , and viral strains ( SC16 , 17syn+ , KOS/M , and McKrae ) , indicating that they are characteristic of neurons latently infected with HSV-1 . The occurrence of each pattern was variable ( Figure 1E ) , suggesting that the strain of virus and or the route of infection may affect how the viral genome accumulates in neurons . In addition to these two primary patterns , a single spot accompanied by one or two smaller spots ( termed “single+” ) and a multiple pattern that filled the nucleus ( termed “super-multiple” ) were less frequently observed ( Figure S1 ) . The presence of multiple genome spots in a large proportion of infected neurons is consistent with the results of earlier single-cell quantitative PCR ( qPCR ) analyzes [19] , [20] . We further confirmed that the sizes of the spots observed by FISH were consistent with the presence of several copies per spot ( Table 1 ) . The spot in the single pattern was 0 . 80±0 . 14 µm wide ( n = 48 ) , a size similar to that of in vitro-induced quiescent genomes [50] , which were estimated by qPCR to contain four to five copies of the genome . The spots of the multiple pattern varied from 0 . 40 to 3 µm in diameter ( more in the case of large aggregates ) . We measured the sizes of individual FISH-detected HSV-1 genomes in in vitro-infected cells to define a reference . Single-copy parental genomes entering the nucleus appeared as spots that were 0 . 51±0 . 08 µm wide ( n = 76 ) , which was similar to the width of isolated spots within the multiple pattern ( 0 . 44±0 . 07 µm; n = 51 ) , indicating that these spots may represent single genomes . Based on this analysis , it can be predicted that the single-spot pattern contains more than one copy of the genome and that in the multiple-spot pattern , the genome can be either isolated or aggregated . We next used serial sectioning to explore whether HSV-1 established latency with a topographical preference within the TG . Neurons shown by FISH to be positive for HSV-1 were distributed all along the TG ( one of the three sections analyzed ) , without any enrichment along the antero-posterior axis ( Figure 1F ) . Similarly , the frequencies of the single and multiple patterns were equivalent throughout the TG ( one mouse is shown as an example in Figure 1G ) . The frequency varied from section to section , and no reproducible pattern could be detected in a group of six mice . Overall , these results show that the HSV-1 latent genome in mouse neuronal tissues can be detected by FISH , with sufficient efficiency and quality for the analysis of intra-nuclear distribution . During latency in neuron nuclei , the HSV-1 genome is present as multiple copies , as shown previously using other methods [15] , and adopts a non-random intra-nuclear organization . Data from several groups suggest that LAT is expressed in a fraction of infected neurons during latency . We set up a dual RNA/DNA-FISH assay based on tyramide signal amplification ( TSA ) technology to co-detect LAT transcripts and HSV-1 genomes ( Figure 1B and 2A ) . We challenged the sensitivity of our RNA FISH method by using up to 20 times the amount of probe ( 1000 ng/assay instead of 50 ng ) and increasing the TSA time . We failed to detect neurons with weak LAT signals , indicating that our test efficiently detected LAT-expressing neurons . In mice at 28 d . p . i . , 18 to 31% of the HSV-1 DNA-containing neurons were positive for the 2-kb LAT RNA , thus confirming the results previously obtained by different approaches [18] , [20] . Notably , fewer than 10% of neurons with a single pattern were positive for 2-kb LAT ( Figure 2B ) . In contrast , 40 . 3±9 . 5% of the multiple-pattern neurons expressed LAT , suggesting that the multiple pattern reflects conditions favorable for LAT transcription ( Figure 2B ) . A reciprocal analysis showed that 83 . 0% of LAT-positive neurons contained the HSV-1 genome in a multiple pattern ( Figure 2C ) . This suggests that the organization of the HSV-1 genome in a multiple pattern is necessary , but not sufficient , to support LAT transcription . These data demonstrate that transcription of the LAT locus is linked to the intra-nuclear pattern of the viral genome . The distribution of latent HSV-1 in neuron nuclei did not bear any resemblance to the patterns of known nuclear domains , and it remained unclear whether the viral genome associated with particular structures . As HSV-1 gene expression has been shown to involve viral chromatin [2] , [3] , we first focused on nuclear structures that are known to control cellular gene transcription through heterochromatin domains: the nuclear envelope , telomeres , centromeres , and pericentromeres . HSV-1 latent genomes were rarely found at the periphery of the nucleus , thus excluding a preferential association with the nuclear envelope . The association of the HSV-1 genome with telomeres and centromeres was assessed by dual-color DNA-FISH . No co-localization of the HSV-1 genome with telomeres was observed when assessing single or multiple patterns ( Figure 3A ) . In mouse cells , the centromeres are positioned at the surface of pericentromeric aggregates ( also called chromocenters ) , which are commonly detected by Hoechst staining [53] ( Figure 3A and S2 ) . By dual DNA-FISH , performed using specific probes to detect minor and major satellites , we confirmed that the organization observed in the cultured cells was similar to that in TG neurons , and that heterochromatic aggregates detected by DNA staining represent pericentromeres ( Figure S2 ) . Latent HSV-1 co-localized with centromeric repeats in 10 . 1±0 . 8% of the single-pattern neurons and in 39 . 4±8 . 8% of the multiple-pattern neurons ( Figure 3B ) . In contrast , the frequency of association with pericentromeres remained low for both single- and multiple-pattern neurons ( 4 . 51±3 . 4% and 8 . 43±4 . 9% , respectively , n = 1 , 249 neurons in two mice ) . Within individual neuron nuclei , only a subset of HSV-1 FISH spots was associated with centromeres , showing that they are not the only residence sites of latent genomes . An immuno-FISH staining of the centromeric protein ( CENP ) -A , which is essential for the stability and functionality of the centromere [54] further confirmed the localization of a subset of HSV-1 genome onto the centromeres ( Figure 3C–D ) . Additionally , these data demonstrate that the HSV-1-associated centromeric loci are likely to be functional centromeres . The association of HSV-1 genomes with centromeres did not appear to be an artifact caused by a strong HSV-1 signal in the multiple pattern for the following reasons: ( i ) the centromere and HSV-1 signals were largely co-localized ( Figure 3A and 3D , bottom right image ) ; ( ii ) the positioning of HSV-1 genomes adjacent to pericentromeres did not increase concomitantly with an increase in HSV-1 signal density ( Figure 3B and S2 ) ; ( iii ) and co-detection of HSV-1 and telomeres did not result in signal co-localization , even though each cell contained twice as many telomeres as centromeres and telomeres are proximal to centromeres in acrocentric mouse chromosomes . Because the single HSV-1 pattern did not frequently coincide with centromeres , and because a tight interplay between HSV-1 and PML-NBs exists in vitro , we developed an immuno-FISH approach to analyze whether PML-NBs could be involved in HSV-1 latency . In non-infected tissues , PML was detected by immuno-FISH in both non-neuronal cells and neurons . Neurons contained 1–10 PML spots , although a subpopulation of neurons did not display any detectable signal in the nucleus ( Figure 4A ) . In mice at 28 d . p . i . , a qualitative assessment of the number of PML-NBs in infected neurons did not reveal any obvious change , by comparison with uninfected neurons . In latently infected mice , PML protein invariably associated with single-pattern HSV-1 genomes , whereas it associated with HSV-1 genomes in only 61% of multiple-pattern neurons ( n = 201 neurons ) . In the latter case , only some HSV-1 spots were associated with PML , revealing heterogeneity among the genomes regarding their association with PML . To determine whether HSV-1 genomes associated with bona fide PML-NBs , immuno-FISH and 3D microscopy were used to detect two stable signature components of PML-NBs , ATRX and Daxx . Both ATRX and Daxx were found to be associated with single-pattern HSV-1 genomes ( Figure 4B ) ; in multiple-pattern genomes , ATRX and Daxx co-localized with at least one HSV-1 genome focus , consistent with the observed frequency of the association of these genomes with PML . A triple-labeling experiment confirmed that PML and Daxx ( Figure 4C ) or PML and ATRX ( not shown ) simultaneously associated with the HSV-1 genome . Careful inspection of PML-NBs associated with HSV-1 revealed that PML protein had a ring-like shape , with HSV-1 genome in its center ( Figure 4D ) . The presence of HSV-1 DNA within PML-NBs was intriguing because PML-NBs have been generally found to be devoid of nucleic acids and to be localized adjacent to or within 2 µm of genomic loci [37] , [55] . High-resolution 3D confocal microscopy confirmed that in the case of the HSV-1 latent genome , the DNA was clearly inside the PML ring ( Figure 4D ) , and that PML was wrapped around the viral genome . This organization was also observed during the early phase of mouse infection ( Figure 5C ) and in vitro in cells infected with replication-defective HSV-1 [50] . Thus , our observations show that PML assembles around HSV-1 genomic DNA , forming an atypical DNA-containing PML-NB ( DCP-NB ) . PML-NB reorganization and co-localization with HSV-1 genomes were observed as early events in lytic infection in cultured cells [43] , [45] , [56] , [57] , raising the possibility that the association we observed during latency could be initiated early during the establishment of latency ( the acute phase ) . In immuno-FISH analyzes performed on sections from mice sacrificed at 6 d . p . i . , we observed that the PML protein signal within PML-NBs was stronger , and the PML-NBs were generally larger and more numerous in acute-phase tissues compared with latently infected and non-infected tissues ( Figure 5A ) . Increases in the PML signal were observed in both neurons and accessory cells and , importantly , were restricted to infected TG ( Figure 5A–B and S3; see Materials and methods for details ) , demonstrating that the increase in the PML signal resulted from the on-going infection . The increase in the PML signal could be attributable to the recruitment of nucleoplasmic PML ( which accounts for 90% of nuclear PML; [36] into PML-NBs , or to an increase in the overall amount of PML . Western blotting of whole TG from mice at 6 d . p . i . , showed increases in total PML and PML isoform levels in acutely infected TG compared with non-infected TG ( see Materials and methods ) and TG from non-infected mice ( Figure 5B ) , demonstrating that the change in PML protein pattern results from an increase in total cellular PML protein and not only from a more efficient recruitment of nucleoplasmic PML into PML-NBs . These data support the stimulation of PML expression during the acute phase of HSV-1 infection , probably as a result of IFN pathway activation [41] . PML and HSV-1 formed 1–12 DCP-NBs per infected neuron nucleus , and most of them also contained ATRX , and Daxx ( Figure 5C ) . Thus HSV-1 genome and PML patterns were significantly different from those observed in latently infected neurons . We conclude that acute infection provokes a PML response , leading to the formation of HSV-1 DCP-NBs , and that the association between PML and the viral genome is initiated during the very early stages of the latency process . The above observations raised the possibility that the HSV-1/PML interaction may play a role in the formation of the latent HSV-1 patterns . To address this , we quantified HSV-1 latent genome patterns in latently infected PML-deficient mice . Both PML+/− and PML−/− mice displayed a significant decrease in the number of single-pattern neurons and a concomitant increase in the number of neurons with the super-multiple pattern ( Figure 5D–E ) . These data show that PML protein and/or PML-NBs influence the intra-nuclear pattern adopted by the viral genome within latently infected neurons . Additionally , based on the number of viral genome foci detected within individual neurons , we conclude that in absence of PML/PML-NBs , the number of genome copy in latent TGs is higher , suggesting that PML/PML-NBs play a role in limiting the number of viral genomes that establish latency . The above observations establish strong links between LAT expression , HSV-1 intra-nuclear distribution , and the association of the HSV-1 genome with PML-NBs and centromeres . This raised the possibility that the association of the HSV-1 genome with PML-NBs and centromeres may regulate LAT transcription . In support of this hypothesis , in single-pattern neurons , HSV-1 is systematically associated with PML-NBs , and LAT RNA is rarely present . To test whether association of genomes with PML NBs in multiple pattern cells had any effect on LAT expression in those cells , we performed a triple labeling experiment to simultaneously detect the HSV-1 genome , 2-kb LAT RNA , and PML/centromeres . Preliminary observations suggested that the 2-kb LAT signal was not correlated with the association of the HSV-1 genome with PML-NBs or centromeres . To confirm this , we traced LAT expression and the association of HSV-1 with PML and centromeres for each neuron across the entire TG of one mouse . The data clearly showed that there was no correlation ( Table 2 ) . This suggests that within a nucleus containing multiple copies of the HSV-1 genome , the association of some of the HSV-1 genomes with PML-NBs or centromeres does not have a dominant negative effect on the expression of LAT from the other copies of the viral genome . An individual neuron contains a heterogeneous population of HSV-1 genomes ( “free” or associated with a nuclear domain ) . Thus , the transcriptional status of these genomes may also be heterogeneous . To explore this possibility , we utilized the primary ( nascent ) 8 . 3-kb LAT transcript ( Figure 1B ) as a marker of the site of active transcription , in order to identify genomes that were being transcribed . Figure 6A illustrates the co-detection of HSV-1 genomes ( red ) , the nascent 8 . 3-kb LAT transcript ( blue ) , and the stable 2-kb LAT RNA ( as a control ) . The nascent LAT RNA appeared as a set of large dots ( 1 to 7 per nucleus ) , each dot being associated with at least one HSV-1 genome spot ( Figure 6 ) . Such dotted pattern has been previously observed by ISH using peroxidase and alkaline phosphatase staining [58] . This suggests that in a single neuron , LAT can be transcribed from several copies of the HSV-1 genome . Notably , these neurons also contained several HSV-1 spots that were not associated with any 8 . 3-kb LAT RNA signal . Although we cannot exclude the possibility that these genomes are transcribed at a level below the sensitivity of our FISH method , the data suggest that they are not transcribed . Overall , these results show that only a fraction of the HSV-1 genomes within a single infected neuron are significantly transcribed and that the transcriptional status of HSV-1 genomes is highly heterogeneous in individual neurons . We next analyzed whether LAT is actively transcribed from PML-NB-associated latent HSV-1 genomes . In mouse TG sections , we detected the HSV-1 genome , its associated nascent LAT RNA product , and PML-NBs by a triple-labeling approach . In the LAT-expressing neurons , the nascent 8 . 3-kb LAT RNA was never associated with viral genomes that co-localized with PML-NBs ( Figure 6B , bottom panel ) . We paid particular attention to 8 . 3-kb LAT-positive neurons with the single and single+ patterns , and observed that the LAT positive neurons were all neurons with a single+ pattern , and that the genome that was transcribed was the one not associated with PML . The larger HSV-1 genome spot surrounded by PML protein was never associated with an 8 . 3-kb LAT RNA spot ( Figure 6B , top and middle panels ) . These data further support the idea that PML-NBs repress transcription of the associated HSV-1 genome . However , in the 8 . 3-kb LAT-positive neurons with the multiple pattern , several non-transcribed HSV-1 genomes were not associated with PML-NBs . It is likely that 8 . 3-kb LAT is not transcribed from many of the genomes and that factors other than PML-NBs also regulate LAT transcription . We extended the analysis to centromere-associated HSV-1 genomes . Similarly to the findings for PML-NBs , centromere-associated viral genomes were never adjacent to the nascent 8 . 3-kb LAT RNA , suggesting that centromeres may also inhibit LAT transcription ( Figure 6C ) . If PML were to inhibit LAT transcription from the HSV-1 genome , one would expect to see an increase in LAT expression in a PML-deficient background . The 2-kb LAT RNA-FISH analysis of latently infected PML+/− and PML−/− mice revealed that the percentage of 2-kb LAT-positive neurons was higher in PML−/− mice compared with wild-type and heterozygous mice ( Figure 6D ) . The increase in the LAT-positive neuron percentage was not simply related to greater numbers of neurons with multiple/super-multiple pattern . Indeed , within this category of neurons , LAT expression was twofold higher in PML−/− animals ( Figure 6E ) . These data demonstrate that PML/PML-NBs play a role in the regulation of LAT expression and support their transcriptional repressor activity . Here , we report the structural and functional interactions of the genomes of a persistent virus , HSV-1 , with the host-cell nuclear environment . Our data reveal two new features of the viral genomes that characterize the latency state . First , the intra-nuclear distribution of the latent genome is not random and correlates with viral gene expression , and second , the host-cell nuclear domains play a role in viral genome pattern acquisition and in the control of viral gene expression . Thus , the interaction between the viral genomes and host-cell nuclear components represents a new level of host–virus interaction , which is likely to participate in the process of latency and reactivation . The ability to explore the cell host–virus interaction is of outmost importance in our understanding of persistent viral infections because regulation of the latent HSV-1 genome relies mainly on cellular components . A substantial benefit of the in situ immuno-DNA/RNA-FISH developed in this study lies in the simultaneous detection of the viral genome , virally encoded transcripts , and cellular components in the same cell . This will enable us to address important issues of cell host–virus interactions in tissues obtained from physiologically infected animal models but also in emerging in vitro HSV-1 latency models [59] . The FISH approach provides high-resolution individual cell data without sacrificing the access to a global view of the virus and of the host-cell population . This appears as a major advantage given that HSV-1 latency is highly heterogeneous . FISH and immuno-FISH will be essential assets to study latency and will complement the currently used biochemical and molecular approaches . We confirmed that LAT was expressed in a fraction of infected neurons and that viral copy numbers varied among neurons . Based on the HSV-1 genome pattern and our estimate of genome copy number per FISH spot , the single and multiple patterns likely represent the low-copy and high-copy virus genome-containing neurons , respectively , identified by contextual analysis [19] . Additionally , we found that the HSV-1 latent genomes were heterogeneously distributed within neuron nuclei and preferentially associated with PML-NBs and centromeres . LAT expression is positively correlated with HSV-1 genome pattern and negatively correlated with its association with PML-NBs and centromeres , demonstrating that the intra-nuclear distribution of HSV-1 genomes is a major feature of the latency process . LAT detection was almost exclusively associated with the multiple genome pattern , demonstrating that LAT expression is restricted to neurons with high viral genome copy numbers . Single-cell contextual analysis has also revealed that a high genome copy number per neuron is associated with a higher probability of virus reactivation [14] , suggesting that this parameter may be a key aspect of latent genome status . Importantly , the various copies of HSV-1 within a single nucleus are not transcriptionally equivalent , with LAT being transcribed only from a subset of genomes . This suggests that latent HSV-1 genomes are comparable to , and behave like , multi-allelic cellular genes , raising the possibility that only a subset of these genomes are susceptible to sustain full reactivation ( i . e . , to reach expression of lytic genes ) . The dotted pattern observed with the 8 . 3 kb LAT probe was previously reported [58] , and was proposed to be sites of early processing of the LAT transcript . Our data confirm this hypothesis by demonstrating that these clouds of LAT primary RNA are associated with HSV-1 genomes . PML-NBs are probably the most thoroughly studied nuclear domains in the context of virus infection for their involvement in the innate antiviral response and in the interferon ( IFN ) response pathway . Our data from the acute phase and from PML−/− mice support a role for PML-NBs in limiting the extent of viral replication during acute-phase , and thus the number of HSV-1 genomes that establish latency in each neuron . These data are consistent with the known role of PML-NBs , through the activity of several of their major components such as PML , Sp100 , Daxx , ATRX , and small ubiquitin-like modifier ( SUMO ) protein , as repressors of HSV-1 onset of lytic infection in cultured cells [57] , [60]–[63] . We provide a clear demonstration that PML expression increased in vivo in acutely infected mouse TG , and that the PML protein , through the formation of HSV-1-containing PML-NBs , associated with the HSV-1 genome during the early phase of latency . Additionally , we showed that the HSV-1 genomes remain associated with PML-NBs during latency in over 80% of infected neurons , suggesting that PML-NBs play an antiviral role influencing latency and probably reactivation . PML-NBs have been proposed to create a specific local nuclear environment by concentrating proteins and hosting biochemical reactions within the PML shell . PML-NBs reorganize in response to various stressors , potentially to relocate their activity at selected nuclear sites [64] . The reorganization of the PML-NBs resulted in the formation of new PML-NBs around the HSV-1 genome at early stages of latency , thus altering the immediate nuclear environment of the incoming viral genome . Importantly , we showed that PML-NBs remain associated with viral genomes long after replicative infection has ceased , indicating that maintenance of this particular type of DNA-containing PML-NB ( DCP-NB ) requires neither on-going viral replication nor the associated antiviral and IFN signaling pathways . The pattern of both HSV-1 genome and PML-NBs are dramatically different between acute phase and latency , indicating that a profound remodeling of these patterns takes place during establishment of latency . Ongoing studies will provide pattern analysis at intermediate time between 6 d . p . i . and 28 d . p . i . The formation of DCP-NBs can be seen as a response to the presence of chromatinized foreign DNA [42] , or more broadly , the presence of pathology-associated abnormal chromatin [65] , [66] . Moreover , PML-NBs repress the synthesis of the LAT primary transcript through their association with HSV-1 genomes , from which microRNAs are produced [10] . The atypical assembly of a PML-NB around a genetic locus may thus be considered a distinct form of PML-NB controlling the expression of noncoding RNA in pathological situations . Indeed , PML-NBs assemble around pericentromeric satellite sequences and telomeres , two cellular loci known to give rise to noncoding RNA [65] , [66] . We showed that HSV-1 genomes are also associated with host neuron centromeres during latency . Bishop and colleagues previously showed that foreign DNA delivered by polyomavirus-like particles was localized to centromeres [42] . Our data from a biologically relevant context and an in vivo model support the idea that centromeres represent docking sites for virus genomes . Centromeres and the adjacent pericentromeres are among the best-characterized nuclear domains that silence nearby genes [29]–[31] . Consistent with this , we showed that centromere-associated HSV-1 genomes did not express LAT RNA . We want to emphasize that the association with HSV-1 occurs at the centromere itself , which distinguishes the current set of data from most other published data related to associations of cellular genes with pericentromeres [67]–[75] . Only a subset of HSV-1 genomes within a nucleus is found associated with centromeres , indicating that this association is not the main mechanism repressing transcription of latent genomes . Of note , both PML-NBs and centromeres ( because of their proximity with pericentromeres ) are enriched in ATRX and Daxx . In addition , hDaxx has been shown to co-localize with centromeres in human cells [76] . This raises the possibility that both nuclear domains exert their repressive effect on HSV-1 transcription through common factors [62] . Interestingly , HSV-1 has developed strong “anti-centromere” activity through the combined activities of the viral E3 ubiquitin ligase ICP0 protein [77] and the proteasome . In cultured cells , ICP0 induces the degradation of at least 10 CENPs , which results in the alteration of centromeric chromatin and destabilization of the centromeres [78]–[81] ( S . Gross and P . Lomonte , personal communication ) . The biology of HSV-1 does not favor ICP0-induced centromere destabilization , prompting the mitotic arrest of infected cells [82] , [83] . Indeed , HSV-1 is able to replicate independently of the cell cycle [84] , and the lytic cycle does not depend on cell arrest at the mitotic phase . This suggests that HSV-1 targets centromeres not to control their effect on chromosome segregation , but rather to control an activity more relevant of differentiated , non-dividing cells . On the other hand , it is suspected that ICP0 , which does not bind DNA and is not a transcription factor per se [85] , inhibits the activities of numerous repressive nuclear factors in order to favorably modify the nuclear environment to stimulate the virus replicative cycle , at both the onset of a new infection and during the course of reactivation [60] , [86]–[93] . ICP0 is known to be essential for full reactivation of HSV-1 in latently infected quiescent cells [94]–[96] . We therefore propose that centromeres , although they seem to act as repressors during latency , may offer a favorable nuclear environment for transcriptional events during reactivation , providing their protein composition and structure are modified by ICP0 . This agrees with data showing that centromere/pericentromere regions are sites of intense transcriptional activity following the exposure of cells to a variety of stressors such as heat shock , UV , and heavy metals , which potentially induce HSV-1 reactivation [97]–[99] . On-going work should provide evidence to support this hypothesis . For animal experiments performed in France: all procedures involving experimental animals conformed to ethical issues from the Association for Research in Vision and Ophthalmology ( ARVO ) Statement for the use of animals in research , and were approved by the local Ethical Committee of UPR-3296-CNRS , in accordance with European Community Council Directive 86/609/EEC . All animals received unlimited access to food and water . For animal experiments performed in the USA: animals were housed in American Association for Laboratory Animal Care-approved housing with unlimited access to food and water . All procedures involving animals were approved by the Children's Hospital Animal Care and Use Committee and were in compliance with the Guide for the Care and Use of Laboratory Animals . Wild-type HSV-1 strains SC16 and 17syn+ were used . Stocks were generated in rabbit skin cell monolayers , and viral titers were determined as described previously [100] . Briefly , six-week-old inbred female BALB/c mice ( Janvier Breeding , Le Genest Saint Ile , France ) , were inoculated with 106 PFU of the SC16 virus , injected into the upper-left lip of the mice . Mice were observed daily for clinical signs of ocular infection from 0 to 28 d . p . i . The sided inoculation of the lip results in an asymmetrical infection , which is characterized by an extremely low load of virus on the right TG compared to the left TG . Thus , data presented in this study were collected on the left TG , except in Figure 3A and 3D , as indicated [4] . Data presented in this study were collected from the left TG , except for those shown in Figure 3A and 3D . For the 17syn+/eye model , inoculation was performed as described previously [101] . Briefly , prior to inoculation , mice were anesthetized by intra-peritoneal injection of sodium pentobarbital ( 50 mg/kg of body weight ) . A 10 µL drop of inoculum containing 105 PFU of 17syn+ was placed onto each scarified corneal surface . This procedure results in ∼80% mice survival and 100% infected TG . PML wild-type , heterozygous , and knockout mice were obtained from the NCI Mouse Repository ( NIH , http://mouse . ncifcrf . gov; strain , 129/Sv-Pmltm1Ppp ) [102] . Genotypes were confirmed by PCR , according to the NCI Mouse Repository guidelines . Primers: Frozen sections of mouse TG were performed as previously described [100] . Mice were anesthetized at 6 or 28 d . p . i . , and before tissue dissection , mice were perfused intra-cardially with a solution of 4% formaldehyde , 20% sucrose in 1× PBS . The whole head , or individual TG were prepared as previously described , and 10 µm frontal sections were collected in three parallel series , and stored at −80°C . DNA-FISH probes were Cy3 labeled by nick-translation as described previously [103] . Briefly , cosmids 14 , 28 and 56 [104] comprising a total of ∼90 kb of HSV-1 genome ( see Figure 1A ) were labeled by Nick translation ( Roche Diagnostic ) with dCTP-Cy3 ( GE Healthcare ) , and stored in 100% formamide ( Sigma-Aldrich ) . The DNA-FISH procedure was adapted from Solovei et al . [105] . Frozen sections stored at −80°C were thawed , rehydrated in 1× PBS and permeabilized in 0 , 5% Triton X-100 . Heat based unmasking was performed in 100 mM citrate buffer , and sections were post-fixed using a standard methanol/acetic acid procedure , and dried for 10 min at RT . DNA denaturation of section and probe was performed for 5 min at 80°C , and hybridization was carried out overnight at 37°C . Hybridization mix contained 30 ng of each probe in 10% dextran , 1× denhardt , 2XSSC , 50% formamide . Sections were washed 3×10 min in 2XSSC and 3×10 min in 0 . 2XSSC at 37°C , and nuclei were stained with Hoechst 33258 or ToPro3 ( Invitrogen ) . All sections were mounted under coverslip using Vectashield mounting medium ( Vector Laboratories ) and stored at +4°C until observation . Frozen sections were treated as described for DNA-FISH up to the antigen-unmasking step . Tissues were then incubated for 24 h with the primary antibody ( diluted at 1/100 ) . After three washes , secondary antibody ( 1/200 ) was applied for 1 h . The secondary antibodies ( Invitrogen ) were either AlexaFluor-conjugated ( PML , CENP-A ) , or HRP conjugated ( ATRX and Daxx ) , which were subsequently detected by enzymatic amplification according to manufacturer's guideline ( TSA , Invitrogen ) . Following immunostaining , the tissues were post-fixed in 1% PFA , and DNA-FISH was carried out from the methanol/acetic acid step onward . RNA-FISH probe labeling and RNA-FISH procedures were performed as described previously [4] . Biotinylated single-strand RNA probes were prepared by in vitro transcription ( Ambion ) using plasmids pSLAT-2 , pSLAT-4 and pSLAT-6 as template ( see Figure 1 ) ( Kind gift of S . Efstathiou , University of Cambridge , UK ) . Frozen sections were treated as described for DNA-FISH up to the antigen-unmasking step using solutions containing 2 mM the RNAse inhibitor Ribonucleoside vanadyl complex . The sections were pre-hybridized in 50% formamide/2× SSC and hybridized overnight at 65°C with 50 ng of RNA probe a 50% formamide buffer . Sections were washed in50% formamide/2× SSC at 65°C , and in 2× SSC at room temperature . Detection was performed using streptavidin-HRP conjugate , followed by TSA amplification ( Invitrogen ) with an AlexaFluor 350 conjugated substrate , according to the manufacturer's guidelines . The DNA-FISH procedure was performed starting from the methanol/acetic acid post-fixation step . The following primary antibodies were used: anti-mouse PML ( mAb3739 , Millipore ) , anti-mouse CENP-A ( rabbit mAb C51A7 , Cell Signaling Technologies ) , anti-ATRX H-300 ( Santa Cruz Biotechnology ) , anti-Daxx M-112 ( Santa Cruz Biotechnology ) , anti-pan-HSV-1 ( LSBio ) , and anti-mouse actin ( Sigma-Aldrich ) . All secondary antibodies were Alexa Fluor-coupled and were raised in goats ( Invitrogen ) . HRP-coupled secondary antibodies were provided with the TSA kit ( Invitrogen ) . Observations and most image collections were performed using an inverted Cell Observer microscope ( Zeiss ) with a Plan-Apochromat ×100 N . A . 1 . 4 objective and a CoolSnap HQ2 camera from Molecular Dynamics ( Ropper Scientific ) . When indicated , images were collected on a Zeiss LSM 510 confocal microscope using a Plan-Apochromat ×63 N . A . 1 . 4 objective , except for those shown in Figure 2E , which were collected using a Zeiss LSM 780 microscope . Line scans , 3D projections , and surface rendering were performed using AIM and Zen software ( Zeiss ) . TGs were collected at 6 or 28 d . p . i . and snap-frozen . Frozen tissues were ground , thawed in lysis buffer ( 10 mM Tris-EDTA , pH 8 . 0 ) containing a protease inhibitor cocktail , and briefly sonicated . Protein extracts were homogenized using QiaShredders ( Qiagen ) . Protein concentration was estimated by the Bradford method . Extracted proteins were analyzed by Western blotting using anti-mouse PML antibody ( mAb3739 , Millipore ) [106] .
After an initial lytic infection , many viruses establish a lifelong latent infection that hides them from the host immune system activity until reactivation . To understand the resurgence of the associated diseases , it is indispensable to acquire a better knowledge of the different mechanisms involved in the antiviral defense . During latency , viral genomes of nuclear-replicative viruses , such as herpes simplex virus type 1 ( HSV-1 ) , are stored in the nucleus of host cells in a non-integrated form . Latency establishment is associated with a drastic change in HSV-1 gene expression program that is maintained until reactivation occurs . The last two decades of research has revealed that the functional organization of the cell nucleus , so-called nuclear architecture , is a major factor of regulation of cellular genes expression . Nonetheless , the role of nuclear architecture on HSV-1 gene expression has been widely overlooked . Here we describe that the genome of HSV-1 selectively interacts with two major nuclear structures , the promyelocytic nuclear bodies ( PMLNBs or ND10 ) and the centromeres . We provide evidence supporting that these nuclear domains directly influence the behavior of latent viral genomes and their transcriptional activity . Overall , this study demonstrates that nuclear architecture is a major parameter driving the highly complex HSV-1 latency process .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "cellular", "structures", "cellular", "stress", "responses", "viral", "classification", "microbiology", "host-pathogen", "interaction", "animal", "models", "model", "organisms", "dna", "viruses", "cell", "nucleus", "animal", "models", "of", "infection", "biology", "mouse", "cell", "biology", "viral", "persistence", "and", "latency", "immunity", "virology", "innate", "immunity", "neurons", "cellular", "types", "molecular", "cell", "biology" ]
2012
HSV-1 Genome Subnuclear Positioning and Associations with Host-Cell PML-NBs and Centromeres Regulate LAT Locus Transcription during Latency in Neurons
CA1 pyramidal neurons receive hundreds of synaptic inputs at different distances from the soma . Distance-dependent synaptic scaling enables distal and proximal synapses to influence the somatic membrane equally , a phenomenon called “synaptic democracy” . How this is established is unclear . The backpropagating action potential ( BAP ) is hypothesised to provide distance-dependent information to synapses , allowing synaptic strengths to scale accordingly . Experimental measurements show that a BAP evoked by current injection at the soma causes calcium currents in the apical shaft whose amplitudes decay with distance from the soma . However , in vivo action potentials are not induced by somatic current injection but by synaptic inputs along the dendrites , which creates a different excitable state of the dendrites . Due to technical limitations , it is not possible to study experimentally whether distance information can also be provided by synaptically-evoked BAPs . Therefore we adapted a realistic morphological and electrophysiological model to measure BAP-induced voltage and calcium signals in spines after Schaffer collateral synapse stimulation . We show that peak calcium concentration is highly correlated with soma-synapse distance under a number of physiologically-realistic suprathreshold stimulation regimes and for a range of dendritic morphologies . Peak calcium levels also predicted the attenuation of the EPSP across the dendritic tree . Furthermore , we show that peak calcium can be used to set up a synaptic democracy in a homeostatic manner , whereby synapses regulate their synaptic strength on the basis of the difference between peak calcium and a uniform target value . We conclude that information derived from synaptically-generated BAPs can indicate synapse location and can subsequently be utilised to implement a synaptic democracy . CA1 pyramidal neurons receive numerous synaptic inputs across their extensive dendritic tree , with synapses located up to hundreds of micrometres from the soma [1] . Due to electrotonic filtering , a distal synapse evokes a smaller EPSP at the soma than a proximal synapse of equal synaptic strength and is therefore less effective at generating somatic action potentials [2] . In CA1 pyramidal neurons , synaptic scaling overcomes this inequality with larger synaptic conductances at distal Schaffer collateral synapses than at proximal synapses [3]–[5] . This makes the amplitude of a synaptic response at the soma independent of its dendritic location , a phenomenon known as ‘dendritic democracy’ [6] . It is not clear what cues synapses may use to establish this distance-dependent scaling along the dendrites but internal activity-dependent signalling by the neuron may provide this information . A likely candidate is the backpropagating action potential ( BAP ) , which decreases in amplitude and arrives later as it travels further along the apical shaft [7] , [8] . BAPs activate voltage-gated calcium channels , causing transient , local increases in calcium concentrations at dendritic spines [9] , [10] . Previous experiments , including ours , have measured BAPs that were induced artificially in neurons via somatic current injection [7]–[10] . However , in vivo action potential generation occurs via synaptic stimulation distributed across the dendritic tree , which could evoke a different spatiotemporal pattern of voltage and calcium concentration at spines . Furthermore , AP propagation speed decreases in dendrites with a smaller diameter , such as distal and oblique dendrites [11] . Thus the relationship between BAP features and distance may vary across different branches of the dendritic tree and depend on previous synaptic activity . The two available stimulation methods in slice experiments , extracellular stimulation and glutamate uncaging , are not yet able to elicit a physiologically-realistic , synaptically-evoked BAP . The first method , extracellular stimulation , stimulates both glutamatergic and GABAergic axons and requires an artificially large tetanus to induce an AP . The second method , synapse stimulation by laser-induced glutamate uncaging near a spine , requires scanning two-photon laser microscopes that currently are only able to uncage glutamate at ca . 10 spines within a 5 ms time window . This is insufficient to elicit an AP at the soma of CA1 pyramidal cells , which requires many simultaneously-activated synapses . In addition , measuring voltage in dendritic branches with voltage sensitive dyes is difficult , due to their limited signal-to-noise ratio and toxicity [12] . Although direct patch clamp recordings have been made at the apical shaft [3] , [13] , this is not yet possible for the thin oblique dendrites , where the majority of the spines are located [14] . Therefore we took a modelling approach to investigate whether the calcium and voltage signals associated with synaptically-evoked BAPs contain sufficient information to predict synapse location . We added spines to a well-established CA1 pyramidal neuron model that contains both active and passive properties distributed across a detailed morphology and that has been verified by combined dendritic and somatic recordings [15] . In addition , we used a range of CA1 morphology reconstructions , so as to exclude potential morphology-specific simulation results . We investigated whether features of the voltage and calcium signals , namely their peak , integral and time of onset , could be used as predictors for synaptic location . A good distance predictor should not only contain reliable distance information but should also give consistent results for different types of stimulation . Importantly , the predictor should be a suitable candidate for homeostatic scaling of synaptic strength . This implies that the value of the predictor should respond to changes in synaptic strength , enabling the system to self-organise into a state of synaptic democracy . Under in vivo-like conditions of synaptic stimulation , in non-scaled CA1 pyramidal neurons , we find that the peak value of calcium transients , but not membrane potential , integral values or onset latencies , is strongly correlated with distance and EPSP attenuation . Interestingly , setting one peak calcium target for all spines and homeostatically regulating synaptic strength on the basis of peak calcium resulted in synaptic democracy . Thus , calcium signals in spines induced by synaptically-evoked action potentials contain distance-dependent information across the CA1 dendritic tree that can be used to set up a synaptic democracy . A previous morphologically-realistic compartmental model of a hippocampal CA1 pyramidal cell was modified to include Schaffer collateral spines across the dendritic tree [15] , [16] . In short , the multi-compartment model includes calcium buffering and the following ionic currents: a voltage-gated sodium current ( INa ) , a potassium delayed rectifier current ( IKDR ) , a fast inactivating , A type potassium current ( IA ) , a hyperpolarisation-activated mixed cation current ( Ih ) , a LVA T-type calcium current ( ICaT ) , a HVA R-type calcium current ( ICaR ) , a HVA L-type calcium current ( ICaL ) , a calcium-dependent potassium current ( IAHP ) and a slowly inactivating potassium current ( Im ) [15] . To make the model consistent with our calcium imaging data ( Fig . 1B , C ) , we had to set the density of L-type calcium currents in the proximal apical shaft ( first 50 µm from soma ) equal to the density in the distal dendrites; this had little effect on the backpropagation of APs in the model . In addition , we applied the model to two other CA1 pyramidal cell morphologies , based on a Neurolucida reconstruction of biocytin-filled neurons . Since no full axon reconstruction was available for these morphologies , we used the axon reconstruction described in the original morphology . The model is implemented in NEURON [17] and the code for all simulations in this paper is available from the ModelDB database ( accession number 144490 http://senselab . med . yale . edu/senselab/modeldb ) . To model synaptic input , spines were distributed at random over the dendrites of the stratum radiatum dendritic section , based on distribution patterns for adult CA1 pyramidal neurons [18] . For each stimulation condition , the simulation was repeated 100 times with a new distribution of synapses . Peak , integral and delay-to-peak ( defined as the time from synapse stimulation until the peak signal ) were measured for voltage and calcium signals in spines . To ensure sufficient data points per stimulation to determine a reliable mean per synapse , only synapses that were activated 10 or more times were analysed . Spines were simulated using separate compartments for the neck ( diameter 0 . 2 µm , length 1 . 0 µm ) and spine head ( diameter 0 . 4 µm , length 0 . 2 µm ) [10] . Synaptic NMDA and AMPA receptors and R-type calcium channels were located on the spine head ( see below for detailed description ) . Apart from these receptors and channels , the spines had only passive conductances , the membrane resistance being 10 kΩcm2 and the intracellular resistance 50 Ωcm ( the same as in the oblique dendrites and proximal apical shaft; see [15] ) . Calcium entered into the spine and dendrites through activation of synaptic glutamate receptors and voltage-gated calcium channels . Based on experimental evidence [10] and modelling results [19] , we assumed that there was no diffusion of calcium through the spine neck . Accumulation , buffering and extrusion of calcium in the spine head were modelled using first order kinetics [20]: ( 1 ) where [Ca2+]i is the intracellular calcium concentration , [Ca2+]i , 0 = 70 nM is the resting concentration of calcium , ICa is the total calcium current through the NMDA , AMPA and R-type channels , τCa = 12 ms is the calcium pump extrusion time constant , κ = 20 is the buffer capacity [10] , v is the volume of the spine head and F is Faraday's constant . The time course of NMDA currents were modelled as a sum of exponentials , , with rise time constant τ1 = 1 . 7 ms and decay time constants τ2 = 68 ms and τ3 = 444 ms . The time constants derive from excised patch recordings at 22°C [21] , corrected for the simulation temperature of 34°C using a Q10 of 3 [22] . The peak NMDA conductance was gNMDA = 45 pS , based on a peak conductance of 70 pS measured in spines in CA1 cells [23] and allowing for a 40% reduction due to the steady-state calcium-dependent NMDA receptor inactivation [24] . Voltage-dependent block was modelled as an instantaneous process , with the fraction of unblocked channels being given by Vargas-Caballero and Robinson [25]: ( 2 ) where R is the molar gas constant and T is the temperature in Kelvin . The NMDA receptor passed both a nonspecific ion current IM , NMDA and a calcium current ICa , NMDA given by: ( 3 ) where ENMDA = 0 mV is the NMDA reversal potential and VCa is the effective calcium driving force given by the Goldman-Hodgkin-Katz current equation , ( 4 ) with [Ca2+]e being the extracellular calcium concentration of 2 mM . The ratio of IM , NMDA to ICa , NMDA derives from the ratio of calcium to caesium permeability in hippocampal CA1 and CA3 [23] . The AMPA conductance was modelled by a dual exponential with a rise time constant of 0 . 2 ms and a decay time constant of 5 ms [26] . The maximum AMPA conductance , gAMPA , was 200 pS for all synapses in the non-scaled simulations . This value was based on experimentally measured EPSCs at the apical shaft and the AMPA channel reversal potential [3] . Nonspecific ion and calcium flow through the AMPA channels was modelled in the ratio 99 . 8%∶0 . 2% [23] . The current through the R-type calcium channel was given by: ( 5 ) where gCa , R approaches the slope conductance of the channel for large negative voltages . The kinetics of the R-type channel were taken from the recordings at 22°C [27] and scaled to 34°C using a Q10 of 3 typical of ion channels [28] . The conductance was ( 6 ) where m and h are Hodgkin-Huxley state variables obeying first order kinetics . Their steady-state values were ( 7 ) with V in mV . The time constants ( in ms ) of the state variables were ( 8 ) where T is temperature in degrees Celsius . Based on a unitary conductance of 17 pS [27] and 10 channels per spine [29] , we took to be 170 pS . With these parameter values , the peak calcium concentration in a spine in response to a somatically-induced BAP was around 1 µM , which is well within the range measured experimentally [10] . Synaptic inputs were modelled at subthreshold ( 190 synapses activated ) and suprathreshold ( 240 synapses activated ) levels as bursts of synchronous Schaffer collateral activity . To simulate spike jitter that occurs during sharp-wave or theta rhythms , synapses were activated in an asynchronous pattern , randomly drawn from a 10 ms time window [30] . During a single synaptic stimulation episode , inputs were activated only once . The attenuation at a synapse is defined as , where is the amplitude of the EPSP at the synapse and the corresponding amplitude measured at the soma . The synapses were scaled by multiplying each synapse's conductance in the unscaled simulations by its EPSP attenuation and then dividing by the mean EPSP attenuation of all synapses . In developing neurons , postsynaptic calcium regulates AMPA trafficking and expression in spines [31] , [32] . To investigate whether this can be used to set up a synaptic democracy , we carried out simulations in which the AMPA conductances of activated synapses were adjusted based on the peak calcium levels in the spines . At the start of the simulation , all synapses had the same AMPA conductance ( gAMPA = 200 pS ) . For each simulation run , a different set of 240 synapses was activated to induce a BAP . In each run , the AMPA conductance of an activated synapse was updated according to ( 9 ) where gAMPA , i ( r ) and [Ca2+]i ( r ) are the AMPA conductance and peak calcium at synapse i at run r; [Ca2+]T is the target peak calcium value , which we set at 47 . 0 µM , the median of the peak calcium in the scaled synapses simulation described above; and k = 0 . 1 determines the speed with which the AMPA conductance changes . Thus , in each run , the AMPA conductance changes depending on the difference between the target peak calcium and the current peak calcium; 500 runs were sufficient to create stable synaptic strength in all spines . To quantify the predictive power of each feature x for synapse distance or attenuation y , we determined the least squares fit of the distance or attenuation to a straight line or an exponential . In each case the significant fit which gave the higher R2 value was accepted . All animal use was approved by the Animal Welfare Committee of the VU University Amsterdam . Young adult male Wistar rats ( P28–P42 ) were decapitated and brain removed in ice cold slice solution containing ( in mM ) :110 choline chloride , 11 . 6 Na-ascorbate , 3 . 10 Na-pyruvate , 2 . 50 KCL , 1 . 25 NaH2PO4 , 7 MgCl2 , 0 . 50 CaCl2 , 10 glucose , 26 NaHCO3 [33] . 300 µm horizontal hippocampal slices were cut using a LEICA VT1000S vibratome . Slices were transferred to a holding chamber containing artificial cerebrospinal fluid ( aCSF ) containing ( in mM ) : 125NaCl , 3 KCl , 1 . 2 NaH2PO4 , 10 glucose and 26 NaHCO3 , and heated at 34°C for 20 minutes before storing at room temperature until recording started . All recordings were made in 32°C aCSF . Whole cell patch-clamp recordings were made from CA1 pyramidal cells using 2 . 5–4 . 5 MΩ glass pipettes filled with intracellular solution containing ( in mM ) : 154 K-gluconate , 1 KCl , 10 HEPES , 4 Mg-ATP , 4 K2 phosphocreatine , 0 . 4 GTP . In some experiments , 0 . 2% biocytin was added for morphological verification and K-gluconate was adjusted to 148 mM . Pipettes were filled with intracellular solution containing Alexa-594 ( 80 µM ) and the calcium dye , fluo-4 ( 200 µM ) ( Molecular Probes , Invitrogen ) . Series resistance was not allowed to exceed 20 MΩ and was monitored throughout the recording . Fluorescent dyes were allowed to diffuse into the cell for 20 minutes before measurements began . Dendrites were line-scanned bidirectionally at a frequency of 8 kHz , at various distances from the soma , using a LEICA RS2 two-photon laser scanning microscope with a 63× objective and a Ti∶Sapphire laser tuned to 830 nm excitation . Action potentials were elicited in the soma by a 50 ms current pulse . Relative fluorescence changes are given as the percentage change of Fluo-4 fluorescence from baseline relative to the stable , voltage-independent Alexa-594 fluorescence as described before [34] . Three traces were averaged per distance . A double exponential was fitted to the signal to determine the rise , decay and peak amplitude of the fluorescence signal . Fits were regarded as significant when A was significant with a 99% confidence interval . At the end of the experiment , a z-stack was made of the neuron , to reconstruct the dendritic tree . Using the open source program ImageJ [35] , the region of the dendritic tree scanned was traced back to the soma in 3D based on fluorescence to determine the actual distance travelled by the BAP . In some experiments , the recorded cell filled with biocytin was fixed in 4% paraformaldehyde at the end of the experiment , and processed immunohistochemically with chromogen 3 , 3′diaminobenzidine tetrahydrochloride using the avidin–biotin–peroxidase method . Two different CA1 pyramidal neuron dendritic morphologies were selected for manual reconstruction using Neurolucida ( MicroBrightField ) . Neurolucida reconstructions were directly imported into NEURON and are publicly available together with the code of the model . We measured BAP-induced calcium currents experimentally , using multi-photon calcium imaging in hippocampal horizontal slices ( P28–P42 ) . Calcium concentrations were measured in the dendritic apical shaft at different distances after a BAP in CA1 hippocampal cells was induced by somatic current injection ( Fig . 1A ) . The BAP-induced amplitude of the calcium signal decreased with distance along the apical shaft up to 400 µm , in accordance with previous reports in mature CA1 hippocampal neurons ( Fig . 1B , C , [7] , [8] ) . This suggests that information about distance , required to set up a dendritic democracy , could be provided by calcium concentration . However in vivo , action potentials are evoked by synaptic inputs rather than by current injected directly into the soma . Synaptic activity in the dendrites , which alters local dendritic excitability , affects propagation of the resultant BAP into the same dendritic region . Because slice experiments do not allow investigation of synaptically-evoked BAPs , we adapted a realistic morphological and electrophysiological model of a CA1 pyramidal cell ( Fig . 1D ) [15] , [16] . Stimulating the model at the soma by current injection , mimicking the experimental data described above , showed a similar inverse relationship between local calcium influx at the apical shaft and dendritic distance to that found in the experiment ( Fig . 1H ) . We therefore proceeded to use the model to test the effect of synaptically-induced action potentials . To test what effect synaptic stimulation could have on backpropagation , BAPs were generated via synchronous synaptic activation across the model CA1 neuron's dendrites ( Fig . 2A ) . Voltage and calcium concentration were measured in the spines . Synaptically-induced BAPs generated a markedly different pattern of spatial-temporal kinetics than somatically-induced BAPs ( Fig . S1 ) , with a smaller and lower range of peak calcium values for somatic induction alone compared with synaptic stimulation ( Fig . S1M ) . With synaptically-induced BAPs , synaptic stimulation caused an initial depolarisation and influx of calcium into activated spines via AMPA receptors and partially-unblocked NMDA receptors ( Fig . 2B–F ) . Arrival of the BAP at the spine head produced an additional increase in local membrane potential and calcium concentration via further unblocking of NMDA receptors and opening of R-type voltage-gated calcium channels ( Fig . 2B–F ) . Because R-type channels have a high threshold for activation , the calcium influx through R-type channels is more prominent in proximal synapses than in distal synapses , in which the calcium influx is mediated mainly through NMDA receptors ( Fig . 2F ) . We next tested whether features of the voltage or calcium signals could provide a distance measure to the spines under these in vivo-like stimulation conditions . We looked at the peak value , the integral and the delay-to-peak of the calcium and voltage signals . Surprisingly , peak voltage induced by the synaptically-driven BAP was not a good correlate of synaptic distance , accounting for only 24% of the variance ( Fig . 2G ) . Although the peak voltage decreased with distance along the apical shaft , in agreement with experimental observations [13] , [36] and model predictions for passive dendrites [11] , [37] , there was no good overall correlation of peak voltage with distance . This can be explained by a difference in BAP spread into the apical and oblique dendrites . In the apical dendritic spines , peak voltage decreased with path distance from the soma but this relationship was reversed when the BAP entered the oblique dendritic spines ( Fig . 2G ) . Consequently , for a given BAP amplitude , there was a considerable range of potential synapse distances , e . g . an amplitude of 70 mV only localised the path distance between synapse and soma to within a range of 200–850 µm . In contrast , peak calcium levels in the stimulated spines were good correlates of soma-synapse distance , with an exponential fit explaining 64% of the variance ( Fig . 2J ) . Unlike peak voltage , peak calcium concentrations in oblique dendritic spines did not increase strongly relative to those in spines on the apical shaft . Besides the peak values , we also looked at the integrals of the voltage and calcium signals in the spines . The integral of membrane voltage showed a weaker correlation with distance ( R2 = 0 . 47 , Fig . 2H ) than peak calcium but substantially higher than peak membrane voltage . Distal synapses showed lower integral values than proximal synapses . Intriguingly , the integral of membrane potential was relatively constant within an oblique dendrite , while synapses along the apical shaft showed strong distance-dependent modulation of the membrane potential integral . The calcium integral showed a similar distance-dependent pattern , although with a higher correlation ( R2 = 0 . 56 , Fig . 2K ) . We further tested whether the time delay from onset of synaptic stimulation to BAP arrival at the spine was a good correlate for soma-synapse distance . Both delay-to-peak of BAP voltage and delay-to-peak of calcium were only weakly correlated with synaptic location ( R2 = 0 . 37 and 0 . 33 , Fig . 2I , L , respectively ) . Thus , for synaptically-stimulated BAPs , peak calcium at the spine is the best correlate of distance from the soma . CA1 pyramidal neurons exhibit a wide variety of morphologies , all characterised by a thick apical shaft with small oblique dendrites at the sides . To test whether morphological variation in CA1 pyramidal cells would influence our simulation results , we repeated the simulation with the two other morphologies described in Fig . 3 . Across the different morphologies , peak calcium was consistently moderately or strongly correlated with distance for all CA1 morphologies ( R2 = 0 . 55 and 0 . 67 compared to R2 = 0 . 66 of the original morphology ) ( Fig . 3 ) . Therefore , morphological variation did not affect our main finding: namely , peak calcium levels in the spines after a synaptically-induced BAP can consistently provide distance information to the synapse . We also tested our model with a two-fold increase of the AMPA conductance ( gAMPA = 400 pS instead of 200 pS ) , which yielded similar results ( Fig . S2 ) . Hippocampal pyramidal neurons exhibit highly synchronised firing patterns during cholinergic activation , with synaptic inputs phase-locked to the local network oscillation [38] . Within this phase-locked period , there is asynchronous jitter in the onset of synaptic currents [30] , [39] . In between these bursts , cells are silent and receive only subthreshold activation [40] ( i . e . activation that does not trigger an action potential at the soma and consequently no BAP ) . Therefore , we tested whether the calcium and voltage features correlated with dendritic distance when synapses were activated either asynchronously or under subthreshold conditions . As with synchronous synaptic stimulation , peak voltage in asynchronously-activated spines was weakly correlated with distance ( R2 = 0 . 23 , Fig . 4B ) . Both delay-to-peak voltage and delay-to-peak calcium also showed moderate or weak correlation with distance ( Fig . 4D , R2 = 0 . 53 , Fig . 4H , R2 = 0 . 32 ) . Integral voltage and integral calcium continued to contain a moderate amount of distance information ( Fig . 4C , R2 = 0 . 50 , Fig . 4G , R2 = 0 . 60 ) . However , these correlations were still lower than peak calcium , which remained the strongest correlate with dendritic distance ( Fig . 4F , R2 = 0 . 65 ) . To investigate subthreshold activation , the number of activated synapses was reduced so that a BAP was no longer generated at the soma . Peak voltage continued to convey no accurate distance information ( Fig . 4J , R2 = 0 . 12 ) . Peak calcium was not informative anymore and the range of calcium concentration dropped to 26–32 µM compared to 30–90 µM in the BAP case ( Fig . 4N , R2 = 0 . 03 and 2J ) . Voltage and calcium integrals maintained their spatiotemporal pattern , although with lower correlations ( Fig . 4K , R2 = 0 . 35 , Fig . 4O , R2 = 0 . 25 , compared with Fig . 2H , K ) . The delay between the onset of the stimulus and reaching peak calcium concentration was highly correlated with distance , albeit in the opposite direction from that in suprathreshold activation ( R2 = 0 . 87 , Figs . 4P and 2L ) . The delay-to-peak voltage showed no distance discrimination after 400 µm ( Fig . 4L ) . Thus , at subthreshold stimulation , delay-to-peak calcium but not actual peak calcium was the best correlate of distance . Why is BAP-induced peak calcium consistently the best correlate of soma-synapse distance ? Following synapse activation , the initial calcium influx into spines is mediated by AMPA receptors and NMDA receptors that have been partially unblocked by the synaptically-induced depolarisation ( Fig . 2B–F ) . Arrival of the BAP at the spine head causes an additional increase in local membrane potential and calcium influx , via further unblocking of NMDA receptors ( Fig . 2B–F ) . NMDA receptors act as coincidence detectors [41] and require , besides synaptic stimulation , postsynaptic depolarisation , in this case provided by the BAP , to unblock the channel further . This calcium increase due to the BAP is responsible for the distance dependency of peak calcium . Since the calcium influx is dependent on membrane potential ( see Eqns . 3–7 , Materials and Methods ) , and the BAP amplitude decreases with distance in the apical shaft ( Fig . 2G , Fig . 5A ) , in agreement with experimental findings [13] , [36] , there is less BAP-induced calcium influx at distal synapses in the apical shaft than at proximal synapses ( Fig . 2C , J ) . Without a BAP , as in subthreshold stimulation , the NMDA channels remained closed and peak calcium is no longer informative about distance ( Fig . 4N ) . Although the amplitude of the BAP decreases with distance in the apical shaft , peak voltage is not an accurate correlate of synaptic distance for all synapses ( Fig . 2G ) . This is due to the BAP amplitude increasing again when the BAP enters the thin oblique dendrites ( Fig . 2G , Fig . 5A ) . Because of the tapering and sealed end of the obliques , peak voltage increases towards the distal end of the oblique dendrites [42] ( Fig . 2G , Fig . 5A ) , similar to effects seen in sealed ends of axons [43] . Simultaneously with the increase in peak voltage , however , the width of the voltage signal decreased in the oblique dendrites ( Fig . 5B ) , so that the integral of voltage remained relatively constant ( Fig . 2H , Fig . 5C ) . This narrow voltage signal in the thin oblique dendrites shown by our simulations is in agreement with findings from voltage-sensitive dye studies in thin basal dendrites [44] . The different behaviour of the integral of voltage in the obliques , as compared with peak voltage , is why the integral of voltage gives a stronger overall correlation with distance ( Fig . 2H , Fig . 5C ) . Biologically , the integral of voltage is read out by calcium [41] , [45] . Due to the slow time constants involved in calcium extrusion ( see Eqn . 1 ) and calcium influx through NMDA channels ( see Eqns . 2–4 ) , the calcium concentration effectively reflects the integral of voltage . Peak calcium therefore also correlated well with synaptic distance ( Fig . 2J ) . A combination of lower time constants for the NMDA channel and a quickening of the calcium extrusion with and without the elimination of R-type channels abolished the distance-dependent correlation with peak calcium ( Fig . 5D , E , F , right panel ) whilst the correlations with peak voltage and voltage integral were not affected ( Fig . 5F , left and centre panels ) . Although the R-type voltage-gated calcium channels , which have fast kinetics , contribute to the calcium influx , especially in the proximal synapses , the influx through the NMDA receptors alone is sufficient to create the distant-dependency of peak calcium , as can be seen when the R-type channels are removed ( Fig . 5E , Fig . S3 ) . From a functional perspective , a more important measure than path distance of a synapse from the soma is the amount of attenuation EPSPs undergo en route to the soma . Although a major factor influencing EPSP attenuation is distance , there are also other factors involved such as dendritic diameter , the activation state of the dendrite and low threshold voltage-gated channels . To test whether attenuation can also be predicted by BAP features , we measured EPSP attenuation for each synapse , defined as the difference in EPSP amplitudes at the synapse and soma , divided by the EPSP amplitude at the soma ( see Materials and Methods ) . As expected it increased with distance from the soma , though the rate of increase with distance reduced in the smaller oblique branches ( Fig . 6A ) . Attenuation depended on the BAP features in the same way as soma-dendritic distance . Again , peak calcium had a strong predictive power ( R2 = 0 . 72 , Fig . 6E ) while peak voltage and delay-to-peak voltage and calcium were moderately to strongly-correlated ( Fig . 6B , D , G ) . This demonstrates that BAP-induced calcium levels can predict not only soma-synapse distance but also the more physiologically-relevant EPSP attenuation . The integrals of calcium and voltage were also strongly correlated with EPSP attenuation ( Fig . 6C , R2 = 0 . 59 , Fig . 6F , R2 = 0 . 65 ) due to low variation in attenuation along each oblique dendrite ( Fig . 6A ) . Peak calcium is correlated with distance and could potentially be used by the synapse as distance indicator in setting up a synaptic democracy . For a democracy to be established in a self-organising manner , the feature upon which synaptic strengths are scaled should in itself be responsive to changes in synaptic strength , otherwise synaptic strengths would not stabilise . Therefore , we investigated whether peak calcium , which correlated with distance in non-scaled synapses , is influenced by synaptic strength . To test this , we scaled the strength of synapses , defined as the size of the synapse's AMPA conductance , according to EPSP attenuation ( Fig . 6 ) , to mimic synaptic democracy and then investigated the resulting distance-dependency of peak calcium ( Fig . 7 ) . While peak calcium showed a clear distance-dependency in the non-scaled simulations , this became less prominent in the scaled scenario ( Fig . 7D , Fig . 2J , R2 = 0 . 48 compared to R2 = 0 . 64 ) , indicating that peak calcium was influenced by synaptic strength and reflected in significantly different distributions of peak calcium ( Kolmogorov-Smirnov test , p<0 . 001 ) . The other features showed similar or lower correlations with distance as the non-scaled simulations ( Fig . 7A–C , E , F , Fig . 2 ) . Thus , peak calcium is influenced by synaptic strength and could therefore potentially be used to regulate synaptic strength . Interestingly , two recent experimental studies showed that peak calcium can regulate the amount of AMPA receptors in a homeostatic manner . Decreased levels of postsynaptic calcium resulted in an increased amount of AMPA receptors via retinoic acid [32] . Conversely , increased levels of activity at individual synapses resulted in a decreased amount of AMPA receptors , in an NMDA-dependent manner [31] . We investigated whether homeostatic scaling governed by peak calcium levels could produce a synaptic democracy . Based on the results of the scaled synapse simulations described above ( Fig . 7D ) , we set the target peak calcium value ( see Eqn . 9 ) for all spines at 47 . 0 µM . Initially , all synapses had the same strength ( i . e . , AMPA conductance ) . After each simulation run , the synaptic strength of each synapse was increased or decreased depending on whether the peak calcium value in the spine was lower or higher , respectively , than the target peak calcium level ( Fig . 8A , Eqn . 9 ) . Note that in this simulation , synapses were updated once they were activated regardless of whether the response was sub- or suprathreshold ( Fig . S4 ) . After 500 runs , the synaptic strengths had reached a stable arrangement in which the distal synapses had higher synaptic strengths than the proximal ones ( Fig . 8 C , D , Fig . S4 ) . Why was this pattern of synaptic strengths produced ? When peak calcium in a spine is below the target value , the AMPA conductance of that synapse is increased ( Fig . 8A ) . The next time the synapse is activated and a BAP arrives , the higher AMPA conductance leads to a higher postsynaptic depolarisation and consequently a higher calcium influx via the NMDA channels . During subsequent synaptic activation , the AMPA conductance is further increased until the peak calcium is on average at the target value . The opposite changes occur when the peak calcium starts below the target value . Since the target peak calcium is the same for all synapses , and peak calcium decreases with distance from the soma ( with unscaled synapses ) , distal synapses end up with higher AMPA conductances than proximal synapses . To test whether the homeostatic scaling had produced a true synaptic democracy , we looked at the EPSP size at the soma . With all synapses having the same strength , EPSP amplitude at the soma depended on the distance of the synapse ( Fig . 8E ) . With the homeostatically-scaled synapses , the EPSP size measured at the soma was mostly distance-independent ( Fig . 8F ) , indicating the establishment of a synaptic democracy . A minority of synapses , namely the most proximal synapses , were not scaled . The density of R-type calcium channels is relatively high for proximal synapses , so that the calcium influx through these channels already resulted in calcium concentrations above the target value . For these spines , the AMPA conductance continued to decrease , as the peak calcium continued to remain above the target value . These spines ( 2 . 9% of total spines ) converged towards zero synaptic strength and as a consequence , gave rise to very low EPSP values at the soma ( Fig . 8 C , D , F ) . However , for the majority of CA1 pyramidal synapses , homeostatic scaling based on peak calcium concentrations is able to produce a synaptic democracy . In both our model and experiments , the peak calcium concentration evoked by somatically-generated action potentials decays with distance from the soma , suggesting that calcium signals could hold distance information for the synapse . To determine whether this was true for synaptically-induced BAPs , when the dendritic tree has a different level of excitability , we employed a computational model . The model predicts that for synaptically-driven BAPs , peak calcium concentration is the best correlate of synapse distance from the soma , both for synchronous and asynchronous synapse stimulation . This finding is robust for different dendritic morphologies . Furthermore , the functional measure of attenuation of the EPSP signal is also predicted well by peak calcium concentration . Importantly , homeostatic scaling of synaptic strength based on one target level of peak calcium for all synapses resulted in proximal synapses having lower strengths than distal synapses . Thus , peak calcium could be used for distant-dependent scaling of synaptic strengths and the establishment of a synaptic democracy . The computational model used in this paper has multiple advantages over previous models and experiments exploring AP backpropagation . Firstly , due to technical limitations in experiments , backpropagation is always studied by direct stimulation of the soma to generate APs . In this situation , dendritic excitability is at resting state before the BAP arrives . In reality , dendrites are first locally activated by synaptic inputs , which in turn elicit an AP that backpropagates into the dendrites [6] . Synaptic activation has significant effects on local dendritic ion channels , modifying conductance states or even inactivating channels [13] , [46] ( Fig . S1 , see also discussion below ) . Our model is synaptically-driven , with synapses dispersed across different branches , and takes into account the excitability due to local dendritic inputs . The initial synaptic stimulation changes membrane potentials in the spine via voltage-gated channels and induces calcium influx . The subsequent BAP causes an additional influx of calcium , mediated by R-type channels and NMDA channels that are unblocked by the initial spine activation ( Fig . 2 ) . All these effects do not occur with somatic stimulation alone ( Fig . S1 ) , which suggests that the results from BAP experiments previously described in the literature might differ from what occurs during in vivo synaptic stimulation . Secondly , in our study , in contrast to previous studies [47] , [48] , measurements are made in the spines , which are modelled as separate compartments with distinct channel composition from the dendrites . This is important biologically because the spine neck compartmentalises many biochemical signalling processes within the spine head [10] , [19] and the spine head is likely to have a more direct influence on the local synaptic response than the dendritic shaft . Thirdly , in the model both calcium and voltage signals can be recorded simultaneously from multiple spines distributed across apical and oblique regions of the dendritic tree . Experimentally , this is currently not possible because it requires multi-photon microscopy that is fast enough to switch to and measure multiple spines within a millisecond time window . In addition , currently available voltage sensitive dyes have toxicity issues with low signal-to-noise ratio , and direct electrophysiological measurements in the spines are not possible . Lastly , in a model it is possible to incorporate adaptive changes in synaptic strength to investigate the development from a neuron with equal synapse strengths towards a neuron with scaled synapses . In experimental conditions this progression in time cannot be followed within the same cell . Our model incorporates the current information available in the literature on ion channel densities and kinetics . This information comes from dendritic patch clamp recordings and electronmicroscopy immuno-labelings [15] . However , the original model of Poirazi et al . ( 2003 ) was optimised for a single morphology . To ensure that our results are not dependent on a unique morphology , we repeated the simulations for multiple morphologies ( Fig . 3 ) . The same main result was found for all morphologies tested , suggesting that our conclusions for peak calcium can be generalised to the general CA1 pyramidal cell population . In addition , the results were robust to different values of the AMPA conductance ( Fig . S2 ) . Note that our model had only excitatory synaptic inputs and did not include other types of input , such as GABAergic inhibition [49] , [50] . The effects of these inputs need to be addressed in another study . Synaptic democracy could be beneficial for memory encoding in CA1 pyramidal neurons . With a synaptic democracy in place , an effective synaptic strength can be encoded equally well at a distal synapse as at a proximal synapse , so that a neuron will have many equivalent synapses at its disposal . Another type of large pyramidal cell , the layer 5 neocortical pyramidal cell , does not have synaptic democracy [51] . This may be understood from a functional perspective: layer 5 pyramidal cells receive axonal inputs from different layers , each impinging at different but highly localised areas of the dendrite . These inputs convey different types of information and need to be processed differently , either via classical EPSP summation at the soma or via NMDA or calcium spikes in the dendrites [49] , [52] . In contrast , CA1 pyramidal neurons receive axonal inputs from a single layer , CA3 , and the inputs are widely dispersed over the dendrite [1] . If all these inputs need to be treated equally , CA1 neurons would benefit from a synaptic democracy . Distance-dependent synaptic scaling was first shown in slice preparations using combined dendritic and somatic patch recording [3] . London and Segev [53] challenged whether the synaptic scaling observed in vitro could give rise to a true synaptic democracy in vivo , pointing out that slice preparations have artificially low levels of synaptic background activity . Their computational model showed that the scaling seen in vivo [3] would be insufficient to counteract the attenuation when in vivo like background activity is present . However , in contrast with our model , their model did not contain active channels , which , crucially , affect EPSP attenuation , and they used only subthreshold synaptic stimulation . In our model , background activity is present in the form of suprathreshold synaptic stimulation that triggers a BAP . The conclusions from our model simulations should be regarded as a hypothesis , which could be tested experimentally once the field has developed a suitable technique to stimulate many synapses simultaneously to elicit an action potential . The most promising technique for doing this is glutamate uncaging combined with a piezo-controlled laser system that can quickly jump from one spine to the next to stimulate and measure hundreds of spines simultaneously [54] . We tested whether a few activated spines together with a somatically-induced action potential would be able to mimic a synaptically-induced BAP in an experimental setting . Compared with a synaptically-induced BAP , both a somatically-induced BAP and a combination of a somatically-induced BAP and synaptic stimulation produced a different voltage pattern across the dendrites ( Fig . S1 ) . Peak calcium is lower and the spread of membrane depolarisation into the oblique dendrites is different . In contrast , synaptic stimulation combined with a somatically-induced BAP would be a valid way to mimic the calcium signals in spines after a synaptically-induced BAP ( Fig . S1D , I ) . The calcium signal in spines induced by a somatically-evoked BAP is only a fraction of what is induced by a synaptically-evoked BAP , with some distal synapses even lacking calcium influx altogether ( Fig . S1M ) . However , note that a somatically-evoked BAP alone combined with the stimulation of a few synapses will lead to a great variability in spine calcium concentrations , depending upon the precise timing of the BAP relative to synapse activation across the dendritic tree . In general , the timing and amplitude of signals are employed in a wide range of biological processes , such as protein synthesis and gene regulation [31] , [55] . For neurons , membrane voltage and calcium concentration are important signals , which can open ion channels , evoke action potentials , trigger signalling cascades and activate protein synthesis [32] , [56] . We therefore investigated which aspects of the calcium and voltage signals , namely their peak , integral and timing , could be used to establish a distance-dependent synaptic scaling ( synaptic democracy ) . Surprisingly , we found that the peak voltage of the AP as it backpropagated into the dendritic tree was not a strong predictor of synapse location . The main reason for this low predictability is the differential propagation along the apical shaft and secondary/tertiary dendrites . In the apical shaft , BAP amplitude decreases with distance , in agreement with our experimental data and that of others [7] , [42] , [47] , but increases when entering the oblique dendrites [57] . The time delay between synapse activation and BAP arrival , important for spike timing-dependent plasticity ( STDP ) , has been hypothesised to be a signal for creating a synaptic democracy [58] . We observed that the delay-to-peak calcium is a predictor for synapse location only in subthreshold conditions ( Fig . 4P ) and is weakly correlated with distance in suprathreshold conditions . In addition , using delay-to-peak calcium or delay-to-peak voltage as a distance indicator would be problematic because with both these features the correlation with distance is in opposite directions for sub- and suprathreshold conditions ( Fig . 2L , Fig . 4P , Table 1 ) . This implies that a scaling process based on time-delay features would receive conflicting information with alternate subthreshold and suprathreshold inputs . It was previously proposed that a reverse STDP rule is required to set up synaptic democracy , followed by a switch to classical STDP rules in the adult CA1 pyramidal cell [58] . However , many studies show that the hippocampus uses classical STDP rules at CA3-CA1 synapses in both juvenile and adult stages [59]–[61] . Thus , no experimental evidence exists to-date for a developmental switch . Moreover , reverse STDP rules only at juvenile stages would make it difficult to accommodate new spines , which are regularly created in the adult neuron [62] . These new spines would not be scaled and would therefore behave differently from the other spines . Another modelling study showed that a distance-dependent STDP rule could also result in synaptic democracy [63] . However this mechanism would still require an internal signalling of distance to the spines to set the distance dependence learning rules . Our study , in contrast , has the same learning rule for all synapses . For non-scaled synapses , peak calcium in spines showed the strongest correlation with distance ( Figs . 2J , 3 , 4F ) , with higher calcium concentrations in proximal synapses than in distal synapses , suggesting that this feature could be used for distance-dependent synaptic scaling . The high correlation of peak calcium with distance can be explained by the integrative properties of calcium , reflecting the integral of voltage rather than peak voltage ( Fig . 5 ) . Peak calcium was also strongly correlated with EPSP attenuation . However , when the neuron received only subthreshold activation ( i . e . no BAP was generated ) , peak calcium was much lower and did not show distance dependence ( Fig . 4N , Table 1 ) . Importantly , using one target calcium level for all synapses and a homoeostatic scaling of AMPA conductances based on peak calcium resulted in a neuron with a self-organising form of synaptic democracy . Homeostatic scaling of synapses based on peak calcium is an attractive and biologically plausible mechanism for creating a synaptic democracy . Firstly , calcium has been shown to regulate protein transcription , protein modulation and protein insertion [55] . Recently , two studies have shown that AMPA receptor expression at the synapse is homeostatically regulated by calcium [31] , [32] . Reduced levels of postsynaptic calcium stimulate the production of retinoic acid , which in turn increase AMPA conductance [32] . Conversely , increased levels of activity in individual synapses of hippocampal cultures resulted in decreased AMPA receptor expression in a NMDA-dependent manner [31] . Secondly , homeostatic scaling takes place on a time scale of days , allowing synaptic changes governed by mechanisms such as STDP to occur on a shorter time scale . Thirdly , the system is dynamic and the same scaling rule can be used for all synapses independently of distance , so that newly-created synapses can scale themselves within an existing synaptic democracy . In summary , our synaptically-driven model suggests that peak calcium levels in the spines are a strong predictor of the distance of a synapse from the soma and the level of attenuation its EPSP undergoes . It is robust to varying levels of activity , different dendritic morphologies and applies to both larger apical dendrites and smaller distal and oblique dendrites . Our results show that a form of homeostatic synaptic self-regulation , in which the synapse can utilise the BAP-induced peak calcium to adjust its strength , results in a synaptic democracy , where all synapses are equally heard at the soma .
Neurons receive information from other neurons via hundreds of contacts ( synapses ) spread across their dendritic branches . Input signals from synapses propagate along a dendrite to the cell body ( soma ) , where the neuron decides whether or not to produce an action potential . Signals that travel further decay more . Were all synapses equally strong , a synapse far from the soma would have less influence on the decision than a synapse close by . However , neurons in the hippocampus , which are involved in learning and memory , have synapses far from the soma that are stronger than those close by , so that all synapses have an equal voice ( “synaptic democracy” ) . But how can a synapse “know” how far it is from the soma ? Using a computational model of a hippocampal neuron , we show that the action potential , which propagates from the soma back into the dendrites , contains information with which synapses can estimate their somatic distance . Specifically , the calcium concentration at the synapse , which is modulated by the backpropagating action potential , decreases with distance from the soma . We show that when the strength of a synapse is adapted in a self-organising manner based on calcium concentration , synaptic democracy is obtained .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results", "Discussion" ]
[ "cellular", "neuroscience", "computational", "neuroscience", "biology", "neuroscience" ]
2012
Spine Calcium Transients Induced by Synaptically-Evoked Action Potentials Can Predict Synapse Location and Establish Synaptic Democracy
After loading with live Leishmania ( L ) amazonensis amastigotes , mouse myeloid dendritic leucocytes/DLs are known to undergo reprogramming of their immune functions . In the study reported here , we investigated whether the presence of live L . amazonensis amastigotes in mouse bone marrow-derived DLs is able to trigger re-programming of DL lipid , and particularly neutral lipid metabolism . Affymetrix-based transcriptional profiles were determined in C57BL/6 and DBA/2 mouse bone marrow-derived DLs that had been sorted from cultures exposed or not to live L . amazonensis amastigotes . This showed that live amastigote-hosting DLs exhibited a coordinated increase in: ( i ) long-chain fatty acids ( LCFA ) and cholesterol uptake/transport , ( ii ) LCFA and cholesterol ( re ) -esterification to triacyl-sn-glycerol ( TAG ) and cholesteryl esters ( CE ) , respectively . As these neutral lipids are known to make up the lipid body ( LB ) core , oleic acid was added to DL cultures and LB accumulation was compared in live amastigote-hosting versus amastigote-free DLs by epi-fluorescence and transmission electron microscopy . This showed that LBs were both significantly larger and more numerous in live amastigote-hosting mouse dendritic leucocytes . Moreover , many of the larger LB showed intimate contact with the membrane of the parasitophorous vacuoles hosting the live L . amazonensis amastigotes . As leucocyte LBs are known to be more than simple neutral lipid repositories , we set about addressing two related questions . Could LBs provide lipids to live amastigotes hosted within the DL parasitophorous vacuole and also deliver ? Could LBs impact either directly or indirectly on the persistence of L . amazonensis amastigotes in rodent skin ? Leishmania spp are protozoan parasites that are transmitted in the dermis of the mammalian host by blood-feeding sand flies . Once in the dermis of the mouse , the metacyclic promastigotes enter , both macrophages and dendritic leucocytes ( DLs ) where they differentiate into amastigotes within the parasitophorous vacuoles ( PV ) . Though present only in low numbers in both skin and skin-draining lymph nodes , DLs have been recognized to play a central role in initiating and regulating the immune processes that take place in these two coupled tissues [1] , [2] , [3] , [4] . More recently , the DL functional repertoire has been extended to cover metabolic functions such as the targeting of neutral lipids to lipid bodies ( LBs ) [5] . Briefly , not only have these cytosolic LBs been recognized in cell lineages other than the adipocyte lineage [6] , [7] , they are also known to perform functions beyond passive lipid storage and lipid homeostasis [8] , [9] , [10] . For instance , these LBs , which are composed of neutral lipids , i . e . triacyl-sn-glycerol ( TAG ) and cholesteryl esters ( CE ) , have been observed in both non-myeloid cell lineages and myeloid leucocyte lineages . Human fibroblasts hosting Toxoplasma gondii [11] , [12] , or human erythrocytes hosting Plasmodium falciparum [13] , historically and currently are under investigation to decipher LB biogenesis and functions [14] . These investigations have shown that tissue macrophages hosting parasitic microbes such as Mycobacterium tuberculosis [15] , [16] and Trypanosoma cruzi [17] contain higher numbers of LBs . They also showed that the LBs in mammalian myeloid leucocytes have functions that range from the synthesis of inflammatory mediators [18] to the cross-presentation of phagocytosed antigens [19] and the delivery of lipids to the live parasitic intracellular microorganisms hosted by these myeloid leucocytes [16] . Altogether , these findings suggest that the biogenesis of these cytosolic LBs reflects recourse , by these parasitic intracellular microorganisms , to harbor host cell-derived lipids for their full development . As mentioned above , although macrophages are the dominant myeloid leucocytes hosting cell-cycling L . amazonensis ( L . am ) amastigotes , the DLs hosting live L . am amastigotes are known to contribute to sustained remodeling of mouse dermis in at least some inbred strains as niches where small amastigote populations persist . Interestingly , while L . am-hosting C57BL/6 , C3H/He and BALB/c skin ( ear pinna or footpad ) shows uncontrolled L . am amastigote burden and progressive skin damage [20] , [21] , [22] , [23] , [24] , [25] , [26] , L . am-hosting DBA/2 mouse ear pinna shows all the features of natural rodent ear pinna , namely rapid control of the amastigote population [20] in a process that not only prevents the development of skin damage but could also account for the persistence of transmissible amastigotes to the blood-feeding adult female sand flies . Given the low frequency of live amastigote-hosting DLs in both skin and skin-draining lymph nodes , we decided to begin our comparative analysis with C57BL/6 and DBA/2 bone marrow-derived DLs generated from GM-CSF-responsive progenitors . Once generated , the DL cultures were maintained at 34°C ( mouse skin temperature ) and further exposed or not to L . am amastigotes that had been carefully collected from nude mouse footpads inoculated with L . am promastigotes . Twenty-four hours later , both amastigote-free DLs and DLs hosting live L . am amastigotes were sorted from the four distinct cultures by flow cytometry in Biosafety Level 2 ( BSL2 ) containment . This procedure yielded live L . am amastigote-hosting DLs and control DLs , both at high purity . These four distinct sorted DL populations were then subjected to a comparative high content analysis of their transcriptional profiles in order to detect any coordinated transcriptional profiles that could account for LB biogenesis and dynamic features [27] . Of the many transcripts that were seen to be upregulated , we focused on those coding for molecules that ensure i ) the transport/uptake of non esterified fatty acids ( NEFA ) and the uptake of cholesterol , ii ) the esterification of uptaken fatty acids and cholesterol iii ) and those transcripts that regulate LB biogenesis and dynamic features . Finally , in this paper , we briefly discuss how cytosolic LBs that autonomously emerge in DLs hosting live L . am amastigotes could both directly and indirectly contribute - in damage-free rodent skin - to the sustained presence of L . am intracellular amastigotes that are transmissible to blood-feeding female sand flies . Six-week-old female DBA/2 , C57BL/6 and Swiss nu/nu mice were purchased from Charles River ( Saint Germain-sur-l'Arbresle , France ) and housed in our A3 animal facility in compliance with the relevant guidelines in force at Institut Pasteur which is a member of the “Comité d'Ethique pour l'Expérimentation Animale” ( CEEA ) - Ile de France . Animal housing conditions and the procedures used in the work described herein were approved by the “Direction des Transports et de la Protection du Public , Sous-Direction de la Protection Sanitaire et de l'Environnement , Police Sanitaire des Animaux” under number B75-15-28 in accordance with the Ethics Charter of animal experimentation that includes appropriate procedures to minimize pain and animal suffering . HL , TL and GM are authorized to perform experiments on vertebrates ( licences: HL , 75-1550; GM , 75-331; TL , 75-717 issued by the Paris Department of Veterinary Services , DDSV ) and were responsible for all the experiments conducted personally or under their supervision as governed by the laws and regulations relating to the protection of animals . DsRed2 transgenic L . am were generated as previously described [27] . Wild-type or DsRed2-transgenic L . am strain LV79 ( WHO reference number MPRO/BR/72/M1841 ) amastigotes were prepared as previously reported [27] . Briefly , footpad skin hosting proliferative amastigotes was sampled in Swiss nude mice and amastigote populations further extracted and prepared for addition to DL cultures . DLs were differentiated from bone marrow cells of 6-week-old DBA/2 or C57BL/6 mice according to a method previously described [28] . Briefly , bone marrow cells were seeded at 4×106 cells per 100 mm diameter bacteriological-grade Petri dish ( Falcon , Becton Dickinson Labware , Franklin Lakes , NJ ) in 10 ml of Iscove's modified Dulbecco's medium ( IMDM; BioWhittaker Europe , Verviers , Belgium ) supplemented with 10% heat-inactivated foetal calf serum ( FCS; Dutscher , Brumath , France ) , 1 . 5% supernatant from the GM-CSF-producing J558 cell line , 50 U/ml penicillin , 50 µg/ml streptomycin , 50 µM 2-mercaptoethanol and 2 mM glutamine . Cultures were incubated at 37°C in a humidified gas phase with 18% O2 and 5% CO2 . On day 6 , suspended cells and loosely adherent cells were harvested using 1% Versene ( EDTA ) ( Seromed ) pre-warmed at 37°C and cultured in complete IMDM supplemented with 10% primary culture supernatant . On day 10 , cells were harvested using 1% Versene ( EDTA ) pre-warmed at 37°C and distributed in hydrophobic 6-well plates ( Greiner , St Marcel , France ) at a concentration of 9×105 cells/ml in 3 ml complete IMDM . On day 14 , DL-containing plates were switched to 34°C and live L . am amastigotes were added - at a ratio of 5 amastigotes for 1 DL - or not for a further 24 hours in complete medium [29] , or in complete medium supplemented with 200 µM oleic acid loaded onto bovine serum albumin ( BSA ) [30] . For further analyses by flow cytometric and epifluorescence microscopy ( EFM ) , DLs were carefully detached after 5 minutes of incubation in 1% Versene ( EDTA ) solution at 34°C and resuspended at 4°C in Dulbecco's PBS with 2% FCS ( PBS-FCS ) . All experimental procedures were performed according to BSL2 practices . Cells were first incubated in PBS-FCS supplemented with 10% heat-inactivated donkey serum for 5 minutes . The cells were then incubated for 30 minutes in PBS-FCS containing 0 . 2 g/ml anti-MHCII monoclonal antibodies ( mAbs ) ( M5/114 ) conjugated to PE-Cy5 ( eBioscience , San Diego , USA ) . After two washes , cells were re-suspended at 5×106 cells/ml in PBS containing 3% FCS and 1% supernatant from the GM-CSF producing J558 cell line [31] . Cell aggregates were dissociated on a 70 µm filter ( Falcon ) , and placed on ice pending cell sorting as previously described [27] . Cell sorting was performed on a FACSAria ( BD Biosciences , San Jose , CA ) fitted with fully sealed sample injection and sort collection chambers operating under negative pressure . After staining with M5/114 , DLs were selected by BD FACSDiva software ( BD Biosciences ) . PE-Cy5 and DsRed2 fluorescence was collected through 695/40 and 576/26 band pass filters , respectively . FSC and SSC were displayed on a linear scale and used to discard cell debris [27] . Live amastigote-hosting DLs were sorted by selecting cells expressing both surface MHC II molecules and DsRed2 fluorescence . Sorting conditions included a sheath pressure of 70 Psi , a flow rate of 7 and the use of a 70 µm nozzle tip . Cells were collected at 4°C in polypropylene tubes ( BD Biosciences ) previously coated with FCS ( overnight at 4°C ) . Sorted cells were immediately used for further studies [27] . Total RNA was extracted from MHC II+ DLs ( RNeasy+ Mini-Kit , Qiagen ) and its quality and concentration was determined using a NanoDrop ND-1000 micro-spectrophotometer ( Kisker , http://www . kisker-biotech . com ) and an Agilent-2100 Bioanalyzer ( Agilent , http://www . chem . agilent . com ) . RNA Integrity Number ( RIN ) scores were determined for each sample ( RNA Integrity Numbers ≥7 . 5 ) providing an objective and standardised measure of RNA quality on a scale of 1 to 10 ( with a value of 10 corresponding to the highest quality ) [32] , [33] . Altogether , 200 ng of total RNA per sample were processed , labelled and hybridized to Affymetrix Mouse Gene ST 1 . 0 arrays , following Affymetrix Protocol ( http://www . affymetrix . com/support/downloads/manuals/expression_analysis_technical_manual . pdf ) . Three Biological replicates per condition were run as described previously [20] . After hybridization , the arrays were stained and scanned at 532 nm using an Affymetrix GeneChip Scanner 3000 which generated individual CEL files for each array . Gene-level expression values were derived from the CEL file probe-level hybridization intensities using the model-based Robust Multichip Average algorithm ( RMA ) [34] . RMA performs normalization , background correction and data summarization . An analysis was performed using the LPE test [35] ( to identify significant differences in gene expression between parasite-free and parasite-harbouring DLs ) , and a p-value of p<0 . 05 was considered as significant . Estimated false discovery rate ( FDR ) was calculated using the Benjamini and Hochberg approach [36] in order to correct for multiple comparisons . A total of 1 , 340 probe-sets showing significant differential expression were input into Ingenuity Pathway Analysis software v5 . 5 . 1 ( http://www . ingenuity . com ) , to perform a biological interaction network analysis . The symbols of the modulated genes are specified in the text ( fold change [FC] values between brackets ) , while their full names are given in additional file 1 . MIAME-compliant data are available in the GEO database http://www . ncbi . nlm . nih . gov/geo/ accession GSE . Flow cytometric ( FCM ) analyses were performed on freshly detached cells . LBs were stained by adding 4 , 4-difluoro-1 , 3 , 5 , 7-tetramethyl-4-bora-3a , 4a-diaza-s-indacene-8-propionic acid , succinimidyl ester , or BODIPY 493/503 ( Molecular Probes , Life technologie SAS , Saint Aubin , France ) to a concentration of 2 µg/ml in PBS for 30 minutes at 34°C . After careful washings in PBS , the presence of MHC II molecules was determined by staining . The uptake of fluorescent palmitic acid ( 4 , 4-difluoro-5 , 7-dimethyl-4-bora-3a , 4a-diaza-s-indacene-3-hexadecanoic acid , BODIPY FL C16 , Molecular Probes ) was analysed by incubating DLs with 0 . 05 µM probe for 30 minutes at 34°C . Cells were washed as described above and the presence of MHC II molecules was detected by immune-staining . Briefly , DLs were incubated with the immuno-staining agent for 15 minutes in PBS-FCS supplemented with 10% heat-inactivated donkey serum . The DLs were then incubated in PBS containing 10% FCS and 0 . 01% sodium azide ( NaN3 ) in the presence of antibodies directed against surface antigens . Extracellular staining procedures were performed with mAbs directed against MHC II ( M5/114 clone ) and CD36 ( clone 72 . 1 ) conjugated to PE-CY5 and PE , respectively . Biotinylated antibodies directed against CD11c ( HL3 ) and IgG control ( B81-3 clone ) were purchased from eBioscience and were used at 0 . 5 µg/ml . Streptravidin conjugated to Phycoeythrin/PE ( Molecular Probes ) was added ( 1 . 5 µg/ml ) to detect binding of the biotinylated antibodies to their respective molecular targets . Intracellular amastigotes in PE MHC II-stained DLs were co-detected by intracellular immunophenotyping using 2A3-26 mAb [27] after adding Cytofix/Cytoperm solution ( BD Biosciences ) . Intracellular staining of amastigotes in PE MHC II-stained DLs was performed after fixation in PBS containing 1% paraformaldehyde ( PFA ) for 20 min at 4°C . DLs were then washed in Perm/Wash solution from the BD Cytofix/Cytoperm Plus Kit ( BD Biosciences ) and incubated for 30 min at 4°C with 5 µg/ml of Alexafluor 488- conjugated 2A3-26 mAb which was shown to strictly bind to the L . amazonensis amastigote [29] . BODIPY FL C16 and BODIPY 493/503 were used both to measure fatty acid ( FA ) uptake by DL and detect the presence of LBs in DL cytosol . After detachment , DLs were centrifuged on poly-L-lysine-coated glass coverslips at 1000 rpm for 5 minutes then incubated at 34°C for 1 hour . FA uptake and localization was analysed after a short 30-minute incubation with BODIPY FL C16 ( 0 . 05 µM in PBS ) at 34°C . Samples were then washed and fixed in 2% PFA , surface MHC II molecules were stained as described above . LBs were stained in PBS containing 2 µg/ml BODIPY 493/503 for 30 minutes at 34°C . Samples were then washed in PBS and fixed in 2% PFA for 20 minutes . The presence of surface MHC II molecules was then detected as described above using biotinylated M5/114 antibodies ( e-Bioscience ) revealed by streptavidin Texas red ( Pierce , Thermo Scientific , Rockford , USA ) . Finally , samples were mounted on glass slides with Mowiol ( Calbiochem , Darmstadt , Germany ) containing 5 µg/ml Hoechst 33342 ( Molecular Probes ) : as the latter is incorporated into DNA it stains the nuclei of both the host cells and the amastigotes . Images were acquired on an upright Zeiss Axioplan 2 microscope controlled by Zeiss Axiovision 4 . 4 software . DLs intended for conventional TEM analyses were cultured for 24 hours post contact with amastigotes or not , then carefully detached . After washing in PBS , DLs were fixed overnight at 4°C in 2 . 5% glutaraldehyde buffered with 0 . 1M sodium cacodylate buffer , pH 7 . 2 . DLs were then washed three times in the same buffer , and were post-fixed in 2% osmium tetroxide ( OsO4 ) ( Merck , Germany ) in 0 . 1M sodium cacodylate buffer , pH 7 . 2 , for 1 hour . After dehydration in graded series ethanol , they were embedded in Epon 812 mixture and subjected to heat polymerization . Thin sections were then cut on a Leica-Microsystems UC7 ultramicrotome , collected on 200 mesh formvar-coated cupper grids and stained with uranyl acetate and Reynold's lead citrate , then observed on a JEM 1200 EXII electron microscope ( Jeol , Tokyo , Japan ) using a megaview camera ( Olympus , Soft imaging systems , Münster , Germany ) . Samples for high pressure freezing and freeze substitution were inactivated by aldehyde fixation as described above . Cells were then taken up in capillary tubes ( Leica , Vienna ) as described [37] . The filled tube was separated by clamping into segments shorter than 2 mm and was then placed in the 200 µm deep cavity of a brass planchette , Type A ( Agar Scientific , Stanstad , UK ) filled with 1-hexadecen . The flat side of the complementary Type B planchette closed the filled planchette and it was frozen in an HPM 010 unit ( BalTec , now Abra Fluid AG , Widnau , Switzerland ) . Freeze substitution was performed in an osmium ( 2% OsO4 ) acetone medium with 1% of water being added [38] . Gently pressing a pre-cooled forceps ( No . 5 , Dumont , Switzerland ) in liquid nitrogen introduced small fractures into the hexadecen and thus allowed the substitution mix access to the sample . Substitution was carried out at −90°C for 48 hours and at −60°C and −30°C for 8 hours each in a freeze substitution device ( Leica , Vienna , Austria ) . Samples were then incubated for 1 hour on ice in 2% OsO4 in dry acetone , followed by 1 h at room temperature in the same solution . Samples were thereafter washed with dry acetone and embedded stepwise in EPON . Two-sided Student's paired t-tests were used to compare flow cytometry experiments ( 4<n<6 ) . A Mann-Whitney test was used to compare ear thickness measurements and number of parasites per DL . DLs were derived from GM-CSF-responsive progenitors ( otherwise known to be present in the bone marrow cell suspensions prepared from the femurs of C57BL/6 and DBA/2 mouse inbred strains ) as previously described for DLs derived from BALB/c mouse bone marrow [27] . More than 97% of the cells in these cultures expressed CD11c in parallel to CD11a and CD11b ( data not shown ) . Two different cell subsets were defined by flow cytometry ( FCM ) analysis . The first did not express surface MHC II molecules , and consequently was not considered as bona fide DLs ( subset 1 , figure S1A ) . The second did express surface MHC II molecules , albeit at various intensities . These are DL phenotypic signatures per se ( subset 2 , figure S1A ) . The latter features were thus used for all subsequent DL FCM analyses , as well as for DL sorting . The cultures were either left in medium alone ( ctrl condition ) or places in contact with live L . am amastigotes at a ratio of 5 amastigotes per cell ( L . am condition ) and either processed or analysed 24 hours later . Combined with the DL features mentioned above , transgenic L . am amastigotes expressing fluorescent DsRed2 protein were used for the specific detection and sorting of all DLs harbouring live L . am amastigotes , without further fixation and permeabilization . Thus , ctrl DLs and DLs hosting live DsRed2 L . am amastigotes were sorted based on their expression of MHC II molecules and MHC II molecules plus DsRed2 fluorescence , respectively ( Figure S1B ) . This sorting strategy constitutes a major improvement over all previously published methods for an in-depth comparative characterization of ctrl DLs versus DLs harbouring live DsRed2 Leishmania amastigotes ( Figure S1B ) [27] . A genome-wide transcriptional analysis based on Affymetrix technology was performed on sorted DLs from both mouse strains . It highlighted that DLs hosting live L . am amastigotes showed only minor transcriptional modifications compared to ctrl DLs , and this both in term of frequency and magnitude . In fact , out of 28 , 853 mouse genes , only 858 and 932 were captured with differential expression at the 5% significance level in C57BL/6 and DBA/2 DLs , respectively . These numbers correspond to only 1 . 6% and 1 . 8% of genes modulated in C57BL/6 and DBA/2 DLs harbouring live L . am amastigotes . In the absence of any signatures indicative of de novo FA synthesis , the modulation of two transcripts , namely i ) phosphatidic acid phosphatase type 2B ( ppap2B ) - known to code for a non-specific phosphatase located in plasma membrane lipid rafts and caveolae - and ii ) mgl - known to convert monoacyl sn glycerol ( MAG ) into glycerol and NEFAs - was interpreted as signatures accounting for NEFA delivery in DLs hosting L . am+ also written L . am+DLs ( Table 1 and figure S2 ) . Therefore , any subsequent protein translation from the coordinated up-regulation of both i ) scavenger receptor-coding transcripts and ii ) NEFA-esterifying enzyme-coding transcripts , NEFA could be directed from the plasma membrane either to the cytosol or to the cortical Endoplasmic Reticulum ( ER ) where it could be enzymatically converted into chemically inert TAGs and cholesterol esters . The latter are known to be the core lipids stored in more or less numerous LBs , especially if lipolysis machinery is prevented from operating in these otherwise dynamic cytosolic LBs . LCFA uptake was promoted in L . am+DLs by the concerted up-regulation of transcripts encoding for key surface molecules ( Table 1 ) . First , LCFA uptake may be promoted by an increase in transcripts encoding for fatty acid translocase ( cd36 ) , resulting in higher levels of CD36 in DLs harboring L . am amastigotes . This higher CD36 expression was confirmed at the protein level by FCM analyses ( n = 4 independent experiments ) . The cell surface of L . am+ DLs in mice of both inbred strains showed a higher quantities of CD36 than L . am− DLs ( MFI CD36 for L . am− vs L . am+ DLs: 36 . 5+/−23 . 0 versus 65 . 2+/−51 . 7 for C57BL/6 DLs , and 125 . 9+/−103 . 4 versus 157 . 4+/−78 . 6 for DBA/2 DLs , see figure 1 illustrating one of the 3 experiments ) . Second , the concomitant up-modulation of transcripts encoding for CD36 and caveolin-1 – one of the structural proteins of caveolae where CD36 can be targeted - may account for the increased LCFA uptake by live amastigote-hosting DLs . Interestingly in DBA/2 DLs , LCFA uptake and transport could also have been promoted by the up-regulation of the fatty acid transporter FATP1 ( slc27a1 ) that probably facilitates LCFA uptake coupled with an esterification step ( Table 1 ) . L . am+ DLs also showed up-modulation of transcripts encoding for the intracellular lipid chaperones FABP4 and FABP5 ( Figure S2 and table 1 ) which are known to be involved in the uptake and metabolism of intracellular LCFA . The up-modulation of these transcripts probably resulted from up-modulation of the nuclear receptor PPAR-gamma ( pparγ ) ( Table 1 ) : PPAR-γ transcription factor is known to promote the expression of CD36 , caveolin 1 , FATP1 and FABP4 , and each of these may further increase LCFA transport/uptake . To determine whether DLs hosting live amastigotes show increased LCFA uptake , ctrl and L . am+ cultures were loaded for 30 minutes with fluorescent palmitic acid ( Bodipy FL-C16 ) , and analysed by FCM ( Figure S3A ) and EFM ( Figure S3B ) . Live L . am+ ( DsRed2+ ) DLs showed higher fluorescence of BODIPY FL-C16 than L . am− ( DsRed2− ) DLs . Interestingly , EFM analyses also revealed that LCFA were able to reach the amastigotes ( Figure S3B insert ) , indicating that live amastigotes in their PV are capable of rapidly scavenging LCFA from their host cell . Glycerolipids are the main structural and functional constituents of Leishmania membranes , with TAG being the second most prevalent lipid class in axenically cultured Leishmania promastigotes [39] . Our Affymetrix-based analyses showed that these enzymes were subject to transcriptional modulations that may promote several pathways involved in diacyl-sn-glycerol ( DAG ) homeostasis and TAG generation ( Table 2 , figure 2 ) . We then analysed the pathways that lead to DAG production by considering two sites of potential production , namely , i ) plasma membrane and ii ) endoplasmic reticulum . DAG production may be promoted in the plasma membrane by up-modulation of non-specific phosphatidic acid phosphatase type 2B ( ppap2b ) and phospholipases d ( pld3 , pld4 , figure 2A1 , table 2 ) . These enzymes may act sequentially in lipid rafts to generate DAG from phosphatidylcholine , as shown previously for pld2 and ppap2b [40] . Whether or not the DAG produced here in the plama membrane level can rapidly participate in TAG generation will need further analysis of the presence or not of cortical ER , i . e . , ER contacting the DL plasma membrane . Non-cortical ER is not expected to be a site of DAG production for first , although we cannot exclude the possibility that MAG conversion may be favoured by up-modulation of dgat1 in C57BL/6 DLs ( Table 2 ) , MAG esterification into DAG ( MAG pathway ) is not promoted ( Figure 2A2 , table 2 ) . Second , in contrast to the modulation of the transcripts accounting for DAG conversion to TAG - that will be further documented in the next section - no up-regulation was noted for either of the transcripts coding for the enzymes involved in the de novo sn-glycerol-3-phosphate pathway ( Figure 2A3 ) , or the transcripts coding for the enzymes involved in DAG conversion to other molecules such as MAG , PC , phosphatidic acid or phosphatidylethanolamine . Not only may increased TAG generation in the caveolae [41] be sustained by increased NEFA uptake and re-esterification , but the up-regulation of dgat1 transcripts ( Figure 2B , table 1 ) may also account for the DAG conversion to TAG . We also noted up-regulation of the transcripts encoding for ADRP ( perilipin-2 ) , which is known to coat the surface of lipid droplets and protect TAG from the action of lipases . By contrast , we noted the up-modulation of abhd5 transcripts ( Figure 2B , table 2 ) . These encode for a lysophosphatidic acid acyltransferase which is able to increase 20-fold the hydrolase activity of atgl [42] . Consequently , a certain amount of TAG could be hydrolysed in L . am+ DLs to balance NEFA re-esterification . The increased dgat1 transcripts observed in C57BL/6 DLs suggests that such TAG hydrolysis occurs since this enzyme would appear to be involved in the re-esterification of hydrolysed TAG by exogenous FA [43] . This complex process may lead to a remodelling of TAG by preferring substrates such oleoyl-CoA . In C57BL/6 DLs , the down-modulation of lipa transcripts is expected to prevent any hydrolysis of the TAG stored in LBs ( see next section ) . As Leishmania perpetuation is known to be strictly reliant upon cholesterol provided by the host organism , we set about investigating whether cellular cholesterol uptake , trafficking and the complex systems regulating cholesterol delivery may be modified in DLs hosting live L . am amastigotes . Mammalian cells acquire cholesterol from three sources: i ) from de novo synthesis in the ER , iii ) from low-density lipoprotein ( LDL ) -derived cholesterol and iii ) from the cholesterol/cholesteryl ester ( CE ) cycle . Our data showed that , in contrast to BALB/c mouse bone marrow-derived macrophages hosting proliferating L . am amastigotes [44] , DLs harbouring live L . am amastigotes did not show any fluctuations in the many enzyme-coding transcripts involved in the de novo biosynthesis of cholesterol . The only exception to this was mvd in DLs of C57BL/6 origin ( Table 3 ) . We noted that cholesterol hydroxylase , which is known to convert cholesterol to 25-hydroxycholesterol/25-HC , was up-regulated . It should here be noted that 25-HC can act on the SCAP/SREBP complex - involved in cholesterol synthesis - by suppressing the SREBP maturation process [45] , [46] , and that any increase in 25-HC might prevent cholesterol synthesis . By contrast , we also noted that transcripts encoding for molecules involved in the uptake and transport of exogenous cholesterol and cholesterol storage were up-modulated ( Table 3 ) . For instance , L . am+ DLs showed concerted up-modulation of multiple surface receptors ( Table 3 ) that may promote the uptake of cholesterol , either specifically or after LDL endocytosis . Cholesterol esterification in L . am+ C57BL/6 DLs appeared to be promoted through increased transcription of cytosolic thiolase acat2 ( Figure 3 , Table 3 ) . This increased CE generation may also be promoted by the higher FA content resulting from increased FA uptake and transport , as described previously . The absence of any modulation of aadacl transcript encoding for cholesteryl ester hydrolase ( CEH ) , and the down-modulation of lipa and lipe in L . am+ C57BL/6 and DBA/2 DLs , respectively , ( Table 3 ) suggested that CE hydrolysis is prevented . No transcriptional signatures were detected except the increased transcription of cav-1 that could account for cholesterol delivery to cell surface caveolae and promote its efflux out of the L . am+ DLs . We noted increased transcriptional expression of genes coding for LB surface proteins , in particular cav-1 , cav-2 , stom , abhd5 and adrp . Since no evidence was found of any promotion of de novo synthesis of cholesterol and TAG , we cultured DLs hosting or not hosting live L . am amastigotes in the presence of oleic acid – a mono-unsaturated LCFA - bound to BSA and compared LB number and size in L . am+ versus L . am− DLs . Briefly , DLs were places in contact with DsRed2-LV79 amastigotes for 3 hours and cultured in the presence of 200 µm oleate/BSA for the next 21 hours . Thirty minutes before FCM analyses of unfixed control ( Ctrl ) and L . am+ DLs , a fluorescent lipophilic probe - BODIPY 493/503 dye - was added to the cultures: its incorporation into the highly hydrophobic ester core of LBs allowed the latter to be evidenced by FCM or EFM as green organelles ( Figures 4 and 5 ) . Parasitized DLs were detected by the orange fluorescence of intracellular DsRed2+ amastigotes . Figure 4A shows the outcome of a representative FCM experiment performed on C57BL/6 mice , and Figure 4B the collective results of 7 independent experiments . Similar results were obtained for DBA/2 DLs ( data not shown ) . White histogram ( Figure 4A1 ) represents background mean fluorescence emitted by ctrl DLs ( here mean fluorescence intensity ( mfi ) was 214 ) . This value was increased ( Figure 4A3 , mfi = 716 , black histogram ) in the oleate-free DLs containing DsRed2+ amastigotes fraction . A statistical analysis showed that the quantity of neutral lipids was significantly higher in L . am+ than in both L . am− DLs from the same culture ( Figure 4A3 , grey histogram ) , and ctrl DLs ( Figure 4A1 , white bars ) ( p<0 . 01 and p<0 . 01 , figure 4B ) . In the oleate only-treated sample ( Figure 4A2 , white histogram ) , mfi reached 454 , proof that the neutral lipids content was higher than in ctrl DLs ( Figure 4A1 ) ) . This increase was significant ( white bars , p<0 . 04 , figure 5B ) . When exposed to live amastigotes plus oleate ( Figure 4A4 ) , DLs contained an even higher neutral lipids content: L . am+ DLs gave a higher mfi ( figure 4A4 , black histogram: 1014 ) than both L . am− DLs from the same culture ( Figure 4A4 , grey histogram: 443 ) and DLs treated by oleate alone ( Figure 4A2 , white histogram: 454 ) . This increase was significant ( p<0 . 005 ) in both cases , and also significant when compared to L . am+ DLs from the oleate-free sample ( black bars , p<0 . 04 , figure 4B ) . Altogether , these findings indicate that the presence of intracellular amastigotes was associated with increased neutral lipids contents in oleate-treated and untreated DLs . However , since amastigotes have their own LBs , this increased fluorescence could be due to LBs from the host , from the parasite , or from both , and FCM analyses cannot differentiate between them . We therefore conducted apotome analyses of our DL samples by EFM ( Figure 5 ) . No LBs were seen in control cultures ( image 1 ) , but in all types of DL cultures , i . e . in the presence of L . am ( image 2 ) , oleate only ( image 3 ) or L . am plus oleate ( image 4 ) . LBs from oleate-free L . am+ DLs were located in the periphery relative to the amastigote nucleus . LBs from oleate-treated L . am+ DLs were located similarly or could also be found far from the parasites . LBs differed in size and number depending on culture conditions: they were most numerous in the oleate-treated L . am+ DLs , followed by oleate-treated DLs , and oleate-free L . am+ DLs . These EFM analyses were unable to determine precisely whether the LBs were of host or parasite origin . To overcome these FCM and EFM limitations , cell cultures were analysed by TEM imaging . Osmium tetroxide is highly reactive with unsaturated FA and can therefore be used to reveal the presence of lipid esters derived from the LB core [47] . Since our transcriptomic analysis focused specifically on host cell response , we analysed the number , size and precise location of LBs in the cytoplasm of DLs . These LBs were of low electron density i . e . light grey and uniform in appearance . LBs were observed in only 9 . 7% of ctrl DLs ( Figure 6A and 6B1 ) and in roughly the same proportion of DLs cultured with L . am ( 6 . 8% ) ( Figure 6A and 6B1 ) , which was consistent with the lack of any promotion of de novo TAG and cholesterol synthesis suggested by the transcriptomic analysis . Altogether , these findings indicate that the increased BODIPY staining obtained in FCM and EFM analyses of ctrl and L . am+ DLs was due to LBs from the parasites , not the host . By contrast , significant numbers of LBs were detected in the cytoplasm of DLs when oleate was added to both L . am− and L . am+ samples ( Figure 6A , 6B1 ) . LBs were detected at a higher frequency in oleate-treated L . am+ DLs ( Figure 6A , 6B1 ) . The total number ( Figure 6B1 ) and area ( Figure 6B2 ) of LBs in host cell cytoplasm were significantly higher under this condition . These observations indicate that the reprogramming of lipid metabolism in DLs harbouring L . am amastigotes may be phenotypically reflected 24 hours post-infection in the presence of oleic acid . Similar results were obtained for DBA/2 DLs ( Figure S4 ) . Interestingly , a high number of the resulting LBs were seen to be in intimate contact with the membrane of PVs ( Figure 7A ) and parasites ( Figure 7B , 7C ) . In conclusion , our study shows that neutral lipid metabolism was rapidly reprogrammed in GM-CSF responsive mouse DLs hosting live L . am amastigotes . Whatever the origin - C57BL/6 or DBA/2 mouse - of the bone marrow from which the DLs were generated , this DL reprogramming showed similar features at both the transcriptional and morphological levels , with many LBs being detected in DL cultures to which oleic acid had been added . Parasitized DLs showed coordinated transcriptional modulations that correlated in part to pparγ up-regulation and promoted the generation and storage of neutral lipids: TAG and cholesteryl esters . The generation of these lipids was singular since not derived from de novo synthesis but from increased import of key constituents , i . e . FAs and cholesterol , from the extracellular milieu , and up-modulation of transcripts involved in their ( re- ) esterification , such as TAG and CE ( Figure 8 ) . When live L . am amastigote-hosting DLs were exposed to oleate , LBs were located in close proximity to PV , and some established close contacts with the PV membrane . No direct fusion of LB phospholipid monolayer with the PV bilayer membrane was evidenced by the methods used in our study . But in a Leishmania we can speculate that LBs store neutral lipids that could be further scavenged by L . am amastigotes otherwise shown to be bound to the PV membrane . These LBs could constitute an essential source of both triacylglycerol and cholesterol . As a precursor of major phospholipids such as phosphatidylcholine , phosphatidylethanolamine and phosphatidylserine , TAG from host DL may not only be hydrolysed to provide the live amastigotes with DAG , but could also be important for the synthesis of their key membrane components . CE from host DL may be used to provide the live amastigotes with cholesterol and CE since it is known that the cholesterol present in Leishmania parasites is not a product of de novo sterol biosynthesis , but is derived from the host ( see for review: [48] ) . Moreover , FAs could be used by amastigotes to produce energy via FA β-oxidation , a process that is known to occur in Leishmania and involves several putative enzymes that have been detected by sequencing of the L . major genome [49] . β-oxidation is particularly pronounced in amastigotes versus promastigotes , the former relying on FA and amino acids as their main sources of energy [50] . FA like oleate can be involved in the synthesis of polyunsaturated FA ( PUFA ) through elongases and desaturases , as evidenced in L . major ( for review: [51] ) . FA and cholesterol can be used in L . am amastigotes for energy and lipid storage , through TAG and CE synthesis , and to generate cytosolic LBs . This de novo glycerolipid synthesis involves the activity of several enzymes such as G-3-P acyltransferase ( Lm GAT ) described in L . major [52] , and GPAT activity expressed in different Leishmania species [52] , [53] . A recent report has suggested a novel mechanism by which exogenous lipids can affect DC function . Indeed , the increased uptake of FAs- that leads to TAG accumulation in LBs - has been shown to reduce antigen processing and presentation to effector T cells [5] . Interestingly , in vitro DLs hosting L . am show altered responsiveness to exogenous stimuli , impaired differenciation and migration , and a low capacity to prime naive CD4+T cells [20] . It is not clear how the accumulated lipid could interfere with antigen handling in DCs ( for review , [54] ) . Further investigations should be conducted to determine whether LB-loaded DLs hosting live L . am amastigotes can prime and re-activate regulatory T lymphocytes that are reactive to unique peptides delivered from persistent amastigotes . It will be also important to identify the mechanism by which FAs and other lipids might affect DL function and to determine how lipid accumulation relates to and links with membrane , cytosolic and nuclear sites of action of FAs and other lipids on DLs . Altogether , once loaded with LBs , the DLs hosting live L . am amastigotes , may indirectly promote the amastigote-driven remodeling of rodent skin as a dynamic niche where two L . am developmental stages durably co- persist: i ) those undergoing “controlled” proliferation and those pre-adapted to blood-feeding female sand flies , namely , the next host population upon which L . am perpetuation is dependent . Pharmacological agents and transgenic mice are now available to clarify the direct role played by these metabolic active DL organelles in L . am replication and/or in the persistence of non-cell-cycling amastigotes .
Once they have gained entry to mammals , live Leishmania ( L ) amazonensis amastigotes are known to subvert both macrophages and dendritic leucocytes ( DLs ) as host cells . These L . amazonensis amastigotes then may or may not proliferate in these two phagocytic leucocyte lineages , but in both cases the otherwise versatile differentiation program of these lineages is known to be rapidly remodeled . Here , we describe the rapid reprogramming of C57BL/6 and DBA/2 mouse bone marrow-derived DLs , with a special focus on cytosolic lipid bodies ( LBs ) that are known to store neutral lipids such as triacyl-sn-glycerol ( TAG ) and cholesteryl esters ( CE ) . After extracting RNA from carefully sorted amastigote-free DLs and L . amazonensis amastigote-hosting DLs , an Affymetrix-based analysis clearly showed a singular and coordinated increase in DL transcripts involved in ( i ) long-chain fatty acid uptake , transport and esterification to TAG and ( ii ) cholesterol uptake and esterification to cholesteryl esters . Oleic acid was added to check that neutral lipid metabolism was both rapidly increased and reprogrammed in amastigote-hosting DLs . It should be noted that the LBs in live amastigote-hosting DLs were more numerous , and that the largest of these LBs were in contact with live amastigote- hosting parasitophorous vacuoles . We further discuss these findings in the context of live L . amazonensis amastigote-rodent host interactions .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results", "and", "Discussion" ]
[ "eukaryotic", "cells", "molecular", "cell", "biology", "immune", "cells", "antigen", "processing", "and", "recognition", "immunity", "gene", "expression", "parasitology", "immunology", "host-pathogen", "interaction", "biology", "genomics", "microbiology", "cellular", "types", "immune", "response" ]
2013
Reprogramming Neutral Lipid Metabolism in Mouse Dendritic Leucocytes Hosting Live Leishmania amazonensis Amastigotes
The waggle dance of honey bee ( Apis mellifera L . ) foragers communicates to nest mates the location of a profitable food source . We used solid-phase microextraction and gas chromatography coupled with mass spectrometry to show that waggle-dancing bees produce and release two alkanes , tricosane and pentacosane , and two alkenes , Z- ( 9 ) -tricosene and Z- ( 9 ) -pentacosene , onto their abdomens and into the air . Nondancing foragers returning from the same food source produce these substances in only minute quantities . Injection of the scent significantly affects worker behavior by increasing the number of bees that exit the hive . The results of this study suggest that these compounds are semiochemicals involved in worker recruitment . By showing that honey bee waggle dancers produce and release behaviorally active chemicals , this study reveals a new dimension in the organization of honey bee foraging . More than fifty years ago , Karl von Frisch demonstrated through a series of elegant experiments that the waggle dance of honey bees uses symbolic communication to convey information about a subject that is both spatially and temporally removed from the receiver of the signal [1] . The waggle dance is therefore a unique animal signal that exhibits several of the important properties of true language , which are generally attributed only to “advanced” organisms such as marine mammals , nonhuman primates , and humans . Honey bees are themselves quite advanced , however; a honey bee society consists of a large group of individuals of overlapping generations , living permanently together with cooperative care of young and a division of labor . To coordinate the complex interactions among the members of such a society , a sophisticated system of communication is necessary . The role of the waggle dance in this sophisticated system of communication is primarily to direct the colony's foraging effort toward nectar- and pollen-producing flowers . Successful foragers perform the dance within the nest to recruit other bees to a profitable food source . By closely following a dancer , potential recruits acquire information about the location and richness of the advertised food source [1 , 2] . However , despite our considerable knowledge of the information contained in the dance , we still do not understand how dancers attract and convey information to recruits in the darkness of the hive [2] . Airborne sounds [3 , 4] , substrate vibrations [5 , 6] , and tactile cues [7] seem to play some role in attracting recruits to waggle dancers or conveying dance information , but each of these modalities appears to be neither necessary nor sufficient for recruitment [2] . Another modality that may be involved in waggle dance communication is olfaction . Evidence to date that olfaction plays a role in honey bee waggle-dance communication is limited to odors acquired from the environment at or en route to the floral food source [2 , 8 , 9] . ( These odors are thought to serve as cues that allow foraging recruits to pinpoint the food source advertised by the waggle dancer . ) However , the production of olfactory signals by the waggle dancers themselves could attract recruits and convey dance information . Pheromones and other semiochemicals are used frequently by honey bees and other social insects to coordinate the activities of colony members [10 , 11] and hence may be expected to help organize the vital task of foraging . Furthermore , evidence suggests that closely related Hymenopterans , such as bumblebees , use pheromones within the nest to organize foraging activity [12] . Whereas bumblebees do not communicate via waggle dances , olfactory-based foraging communication may be an ancestral trait and thus present in honey bees . A preliminary study suggested indeed that the scent of active honey bee foragers could encourage other bees to forage [13] . The goal of this study was to investigate whether waggle dancers produce and release into the air chemical compounds that distinguish them from other foragers . We addressed this first goal by using solid phase microextraction ( SPME ) and gas chromatography coupled with mass spectrometry ( GC/MS ) . If these distinguishing compounds are semiochemicals or pheromones , then adding these compounds into the hive will affect the behavior of foraging bees . To test whether the compounds are behaviorally active , we measured foraging activity before and after we injected the volatilized compounds into the hive . Our results show that honey bee waggle dancers produce four characteristic volatile compounds that increase foraging activity . We discovered four conspicuous compounds in the airspace surrounding dancing bees that were not present in the airspace surrounding nondancing bees ( Figure1 ) . We obtained similar results for the airspace surrounding waggle dancers on an artificial swarm as compared with the air over an area of the swarm with no waggle dancers . From extracts made of three intact waggle dancers immersed for 1 min into 250 μl of hexane , we identified the four substances that generated peaks 1 , 2 , 3 , and 4 as marked in Figure1 to be Z- ( 9 ) -tricosene , tricosane , Z- ( 9 ) -pentacosene , and pentacosane , respectively . All four compounds were present in significantly higher amounts on the abdomens of waggle dancers than in either ( a ) nondancing foragers that returned from the same unscented nectar source , or ( b ) nonforaging bees [Figure 2 , one-way analysis of variance ( ANOVA ) for each compound , degrees of freedom ( df ) = 2 , 49 , p < 0 . 001 for all compounds; Tukey HSD for unequal n , p < 0 . 0005 for all compounds , experiment-wise α = 0 . 05] . Intriguingly , although only marginally significant , waggle dancers that danced more vigorously ( i . e . , appeared to perform waggle runs at higher rates and with more exaggerated movement of the abdomen as assessed subjectively by one person , n = 5 bees ) tended to emit more of all four compounds than less vigorous dancers did ( n = 13 ) ( Figure 3 , Mann-Whitney U test , p = 0 . 05 , 0 . 07 , 0 . 13 , and 0 . 08 for peaks 1 , 2 , 3 , and 4 , respectively ) . Nondancing foragers and nonforaging bees did not emit different amounts of these compounds ( Tukey HSD for unequal n , p = 0 . 076 , 0 . 074 , and 0 . 400 , for peaks 2 , 3 , and 4 , respectively ) with the exception of Z- ( 9 ) -tricosene ( p < 0 . 001 ) . To test whether the waggle-dance–specific substances affect behavior , we injected onto the dance floor a gaseous mixture of the three commercially available compounds dissolved in hexane [hereafter called TTP ( Z- ( 9 ) -tricosene , tricosane , pentacosane ) solution] . TTP trials lasted 32 min , during which we recorded the number of bees exiting the hive each minute , and during which we made two injections: Injection 1 , which consisted of 50 μl of pure hexane , was made during minute 1 , and Injection 2 , which consisted of 50 μl of TTP solution , was made during minute 16 . To control for solvent and treatment , we performed “Hexane trials” , in which Injection 2 , like Injection 1 , consisted of 50 μl of pure hexane . Experiments were done with two colonies , C1 and C2 , using one colony at a time . We performed a total of 20 trials ( 10 TTP trials and 10 Hexane trials ) with C1 , and a total of 28 trials ( 15 TTP trials and 13 Hexane trials ) with C2 . Only one trial was performed per day , and all trials were performed at the same time of day . Because the compounds originate from waggle dancers under natural conditions , we ascertained that at least one dancer was present during each trial . Injection of TTP increased the number of bees exiting the hive ( Figure 4 ) . The normalized mean number of bees exiting the hive during minutes 25–32 differed significantly between TTP trials ( i . e . , Injection 2 = TTP solution ) and Hexane trials ( i . e . , Injection 2 = hexane ) ( two-sample T-test on normally distributed sample groups with equal variances; Colony 1: T = −3 . 29 , p = 0 . 004 , df = 18; Colony 2: T = −2 . 25 , p = 0 . 033 , df = 26 ) . This difference between TTP trials and Hexane trials was not observed following Injection 1 , which consisted of hexane for both types of trial ( Colony 1: T = −0 . 61 , p = 0 . 551 , df = 18; Colony 2: T = −0 . 47 , p = 0 . 643 , df = 26 ) . We also observed within TTP trials an increase in the number of bees exiting the hive following Injection 2 , as compared with the number exiting following Injection 1 , but this was significant for only one colony ( two-sample T-test; Colony 1: T = −1 . 63 , p = 0 . 121 , df = 18; Colony 2: T = −2 . 66 , p = 0 . 013 , df = 28 ) . This increase could possibly be the result of circadian foraging patterns , but this would not explain the lack of such an effect during Hexane trials ( two-sample T-test; Colony 1: T = 0 . 83 , p = 0 . 415 , df = 18; Colony 2: T = −0 . 68 , p = 0 . 502 , df = 24 ) , which were conducted at the same time of day as TTP trials . We have identified in this study four compounds that are unique to waggle-dancing bees and that are behaviorally active . This waggle-dance scent originates from the waggle dancers themselves; it is not acquired from the environment while foraging , nor is it a byproduct of a bee's age or task , because it is emitted in only minute quantities by nondancing foragers returning from the same food source . The prominent compounds of the waggle-dance scent are the cuticular hydrocarbons Z- ( 9 ) -tricosene , tricosane , Z- ( 9 ) -pentacosene , and pentacosane . Hydrocarbons such as these are produced subcutaneously and are not stored within a gland [14] . The waggle-dance scent is therefore best classified as a semiochemical or nonglandular pheromone . Waggle dancers could raise the levels of the compounds through increased synthesis and/or increased release onto the epicuticle . The compounds could then be released passively into the air , perhaps assisted by the relatively high body temperature of the waggle dancers . The chemical nature and source of the compounds of the waggle-dance scent differ from those of the bumblebee foraging recruitment pheromone [15] . The main compounds of the bumblebee recruitment pheromone were identified as eucalyptiol , ocimene , and farnesol , which are terpene derivatives . These compounds are produced in the bees' tergal glands . Their different chemistry and source suggest different evolutionary origins for the bumblebee foraging recruitment pheromone and the waggle-dance scent of honey bees . Hence , it is unlikely that the waggle-dance scent of honey bees has evolved from the bumblebee foraging recruitment pheromone . The compounds of the waggle-dance scent have been identified in earlier studies on honey bees . Z- ( 9 ) -tricosene , tricosane , Z- ( 9 ) -pentacosene , and pentacosane have been previously identified in hexane washes of the cuticles of foraging-age worker bees [16] and have been shown to be perceptible to honey bee workers [17] . More recent work has shown that , among many others , the four compounds are present in the air surrounding foragers at a feeder station [18] . However , with the exception of demonstration of a possible kin-recognition function of Z- ( 9 ) -tricosene [19] , past studies did not link specific compounds with specific behaviors of honey bees . In insects other than honey bees , however , the individual compounds of the waggle-dance scent have been linked to specific behaviors . For example , tricosane , pentacosane , and Z- ( 9 ) -pentacosene are compounds of a pheromone that foragers of the social wasp Vespula vulgaris use to lay and follow terrestrial trails [20] . Compounds of the waggle-dance scent are also well-known sex attractants in other insects , such as flies [14] . The waggle-dance scent may have a similar marking or attracting function , which the context and results of our experiments link to recruitment behavior . Honey bees produce the scent when they perform waggle dances , both in the hive and on swarms , and in both contexts , workers are recruited to the advertised site . In the more common context of foraging , a general measure of recruitment is the number of foragers that leave the hive . Our results show that injection of the waggle-dance scent onto the dance floor increased the number of bees that left the hive . These bees can be assumed to be foragers , because only foragers leave the hive without noticeably hesitating , performing orientation flights , or gathering at the hive entrance . Our experiments show that the waggle-dance scent increases the number of foragers leaving the hive , but the exact mechanism underlying the effect is still unclear . Given the function of the waggle-dance scent in other hymenoptera , we propose that the waggle-dance scent , which in our experiments was fanned onto the dance floor , attracts potential recruits to the dance floor , thereby increasing the probability of encounters between potential recruits and dancers , and finally the number of recruits . Under natural conditions , the scent would originate from the dancers themselves , thus the odor plume would mark not only the dance floor , but the individual dancers . This should enable recruits to locate the dancers themselves , which could enhance recruitment efficiency . Recruits could even seek out more vigorous dancers , who typically advertise especially profitable food sources [21] , and who seem to emit higher concentrations of the scent . Hence , the waggle-dance scent may mark and attract recruits to successful foragers , and thus help to rapidly focus the colony's foraging effort on the most profitable food sources . Besides marking successful dancers , it is feasible that the waggle-dance scent facilitates the transfer of information from the dancer to recruits . The spatio-temporal pattern of a dancer's odor plume could , for example , indicate the length of a waggle run , and thus provide information to a recruit , even if she lost contact with the dancer herself . This hypothesis is supported by the observation that a mechanical model of a waggle dancer recruits bees to a food source only after the model has touched a waggle dancer ( H . E . Esch , personal observation ) . Whereas the waggle-dance signal is likely a signal intended for new recruits , two other groups of bees , namely foragers already devoted to a food source and in-hive receiver bees , could glean cues from the waggle-dance scent . Foragers that are already devoted to a food source do not readily follow new dances if their source becomes unavailable , but rather wait for it to replenish [8] . Because the waggle-dance scent does not seem to identify specific food sources , it can provide only limited information to these foragers . However , high concentrations of the scent could alert them to generally good foraging conditions . This could be useful at the beginning of daily foraging or if foraging can be resumed after a spell of bad weather . This mechanism may have been responsible for the effect observed in an earlier preliminary study , in which the number of foragers visiting an empty feeder increased following exposure to the air from a foraging colony [13] . In our experiments , however , the increase in foragers was not likely caused by already-devoted foragers for three reasons . First , experiments were done well after the time that colonies started daily foraging . Second , external conditions such as the weather were remarkably stable , which made strong fluctuations in the numbers of already-devoted foragers unlikely . Third , we did not record a conspicuous drop in the number of bees that left the hive after the initial increase ( Figure 4 ) ; if the increase would have been due mostly to already-devoted foragers , we would have expected a quick drop to original levels once these foragers found that there was no change in food-source profitability . However , it is possible that such a drop could be hidden by the more substantial numbers of newly recruited foragers . Receiver bees are the second group of bees that could glean cues from the waggle-dance scent . Receiver bees unload nectar from newly returned foragers , which relieves foragers of the time-consuming search for empty storage cells . A high concentration of the waggle-dance scent would indicate a high demand for receiver bees , and could help attract receiver bees to the dance floor . The tremble dance , which is performed by successful foragers that perceive a shortage in receiver bees [22] , may help to additionally spread the scent to potential receiver bees . Bees were kept indoors in four-frame observation hives at the Carl Hayden Bee Research Center in Tucson , Arizona , United States . To test whether waggle dancers produce specific chemical compounds , we performed experiments with two colonies , using first one , then the other colony . Foragers were trained to an artificial feeder [1] ∼100 m from the hive that offered nonscented sugar water . Foragers visiting the feeder were marked on the thorax with powdered paint and thus could be recognized in the hive . We used a SPME fiber ( 65 μm polydimethylsiloxane/divinylbenzene; Supelco; http://www . sigmaaldrich . com/Brands/Supelco_Home . html ) to sample chemicals . After sampling , the SPME fiber was injected into a GC , CP-3800 ( Varian; http://www . varian . com ) coupled to a MS ( Saturn 2200 Ion Trap , Varian ) . We compared the chemical profile of the air over dancing bees to that over nondancing bees , and we also compared the chemical profile of waggle dancers' abdomens to those of both nondancing foragers and nonforaging bees . For air samples , we exposed the fiber for 5 consecutive min approximately 2 cm above the surface of the comb . For abdomen profiles , we briefly touched the fiber to the tip of the abdomen of an individual bee [23–26] . Waggle dancers were sampled during a waggle run , shortly after they began dancing . Sampled bees were observed from the moment they entered the hive until the sample was taken . It is possible that we classified as nondancing foragers bees that danced before observation began ( e . g . , in the entrance tunnel to the hive ) or that danced after SPME sampling , but this would bias our results in only a conservative direction . After SPME sampling , the fiber was desorbed for 3 min in the GC/MS , and compounds separated on a Varian VF-5MS 30 m × 0 . 25 mm inner diameter ( ID ) column with an injector temperature of 250 °C , and a column temperature of 40 °C for 5 min , which was then ramped at 50 °C/min to 150 °C , followed by a ramp at 15 °C/min to 260 °C with a 4 . 5-min hold; flow rate was 1 ml/min . The MS was operated in electron ionization mode at 150 eV . Tentative identification of the peaks was made by comparing MS fragment patterns with spectra from the National Institute of Standards and Technology ( NIST ) 98 and Wiley library databases . Chemical ionization with acetonitrile was used to determine molecular weights and to assign double-bond position through derivatization of the double bond and formation of characteristic addition compounds [27 , 28] . The identities of peaks 1 , 2 , and 4 were further confirmed by the production of identical retention times and fragment patterns in both electron and chemical ionization modes when compared with chemical standards . TTP trials consisted of 32 min during which we made two injections: Injection 1 , 50 μl of pure hexane , during minute 1; Injection 2 , 50 μl of TTP solution , during minute 16 . Hexane trials were similar to TTP trials except that Injection 2 , like Injection 1 , consisted of 50 μl of pure hexane . The TTP solution contained Z- ( 9 ) -tricosene , tricosane , and pentacosane , each diluted 1:100 in hexane and mixed at a ratio of 1:2:3 , respectively ( this ratio produced chromatograms with peak heights that approximately matched those from samples of waggle dancers ) , and then further diluted 1:10 in hexane . To volatilize the liquid TTP solution , we injected it into a heated glass tube with 0 . 5-cm ID . Immediately after injection , we inserted into the tube a fan to blow the vaporized solution through the tube and into a funnel ( 10-cm diameter ) positioned 1 cm above the comb surface of a colony's dance floor . The identical method was used to volatilize and deliver hexane only during Hexane trials ( Injection 1: hexane , Injection 2: hexane ) . To avoid contamination between TTP and Hexane trials , we used separate equipment for TTP solution and hexane . The temperature of the gaseous mixture arriving on the dance floor was maintained between 35 °C and 40 °C . At least one waggle dancer was present on the dance floor during each trial . We conducted each trial on a different day between 14 July and 8 October 2004 . To avoid seasonal effects , we randomized the type of trial ( TTP or Hexane ) performed each day . To reduce the effect of time of day , trials for a colony started during the same hour every day . Trials for the first colony started between 1200 and 1300 , and for the second colony between 1030 and 1130 . To further account for day and time effects , the data for each trial were normalized by dividing each 1-min count by the average number of bees exiting the hive per minute during the 10 min immediately preceding the trial . To compare the presence of the four compounds on the abdomens of waggle dancers with either nondancing foragers that returned from the same unscented food source or nonforaging bees , we used a one-way ANOVA for each compound using Box-Cox transformed data , and a Tukey HSD for unequal n , with data for both colonies pooled . To compare the effect of injection of TTP solution with injection of hexane , we used two-sample T-tests on normally distributed sample groups with equal variances . To account for day and time effects , the data for each trial were normalized by dividing each 1-min count by the average number of bees exiting the hive per minute during the 10 min immediately preceding the trial .
A honey bee colony consists of many thousands of individuals , all of which help to perform the work that allows their colony to thrive . To coordinate their efforts , honey bees have evolved a complex communication system , no part of which is more sophisticated than the waggle dance . The waggle dance is unique , because it exhibits several properties of true language , through which a forager communicates the location and profitability of a food source to other bees in the darkness of the hive . The information coded in the dance has been extensively researched , but we still do not understand how information is actually transferred from the dancing bee to the receivers of the message . Because information is often transferred by scent in honey bee colonies , we investigated whether waggle dancers produce a scent that distinguishes them from foragers that do not dance . We found that dancers produce four hydrocarbons that distinguish them from nondancing foragers , and that , when blown into the hive , increase foraging activity . These results show that waggle-dancing bees produce a unique scent that affects the behavior of their fellow foragers . We discuss likely meanings of this olfactory message and its potential role in waggle-dance communication .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "evolutionary", "biology", "ecology", "chemical", "biology", "insects" ]
2007
The Scent of the Waggle Dance
A working knowledge of the proximate factors intrinsic to sterile caste differentiation is necessary to understand the evolution of eusocial insects . Genomic and transcriptomic analyses in social hymenopteran insects have resulted in the hypothesis that sterile castes are generated by the novel function of co-opted or recruited universal gene networks found in solitary ancestors . However , transcriptome analysis during caste differentiation has not been tested in termites , and evolutionary processes associated with acquiring the caste are still unknown . Termites possess the soldier caste , which is regarded as the first acquired permanently sterile caste in the taxon . In this study , we performed a comparative transcriptome analysis in termite heads during 3 molting processes , i . e . , worker , presoldier and soldier molts , under natural conditions in an incipient colony of the damp-wood termite Zootermopsis nevadensis . Although similar expression patterns were observed during each molting process , more than 50 genes were shown to be highly expressed before the presoldier ( intermediate stage of soldier ) molt . We then performed RNA interference ( RNAi ) of the candidate 13 genes , including transcription factors and uncharacterized protein genes , during presoldier differentiation induced by juvenile hormone ( JH ) analog treatment . Presoldiers induced after RNAi of two genes related to TGFβ ( Transforming growth factor beta ) signaling were extremely unusual and possessed soldier-like phenotypes . These individuals also displayed aggressive behaviors similar to natural soldiers when confronted with Formica ants as hypothetical enemies . These presoldiers never molted into the next instar , presumably due to the decreased expression levels of the molting hormone ( 20-hydroxyecdysone; 20E ) signaling genes . These results suggest that TGFβ signaling was acquired for the novel function of regulating between JH and 20E signaling during soldier differentiation in termites . The complex society of eusocial insects includes sterile castes with phenotypes specialized for individual social tasks . To clarify the mechanisms associated with acquiring sterile castes is a fundamental goal in the evolutionary biology of eusocial insects . In hymenopteran species ( bees , ants and wasps ) , many sociogenomic studies have been conducted in the last decade ( reviewed in [1 , 2] ) based on a vast quantity of genomic and transcriptomic information . These studies suggested that a co-option of the universal gene network in the solitary ancestor was involved in the evolution of sterile castes [3–7] . However , in termites , which are distantly related to hymenopteran social insects , there is still no evidence to support the importance of a universal gene network for the evolution of sterile castes , largely because genetic profiles during caste differentiation have not been clarified . One of the most important species for addressing this issue is the damp-wood termite Zootermopsis nevadensis , because genomic and transcriptomic information is available [8] , gene function analysis has been successfully applied [9–11] , and sterile caste differentiation can be observed under natural conditions [9 , 12] as well as under artificial hormone treatment [10] . In contrast to hymenopteran eusocial insects , in termites the soldier is regarded as the first acquired permanently sterile caste [13] . Soldiers exhibit unique species-specific morphology , which is formed through developmental processes that include double molts ( worker—presoldier and presoldier—soldier ) . Since presoldiers cannot revert to workers , this stage is necessary for soldier-specific morphological changes ( e . g . cuticular formation [10] ) . These processes are not observed in other insects , and are accompanied by extraordinary phenotypic modifications similar to metamorphosis in holometabolous insects [14] . Soldier differentiation is regulated by juvenile hormone ( JH ) , and high JH titer levels in workers trigger soldier differentiation [15] . Indeed , it was reported that JH receptor gene ( Methoprene tolerant: Met ) expression was involved in soldier-specific morphogenesis during soldier molts in Z . nevadensis [9] . Moreover , the molting hormone ( 20-hydroxyecdysone; 20E ) is also crucial for soldier differentiation . Ecdysone receptor gene ( EcR ) expression was also involved in presoldier molt and soldier-specific cuticular pigmentation in Z . nevadensis [11] . However , regulatory mechanisms of a crosstalk between JH and 20E to generate highly specialized phenotypes are still a mystery . One possibility is that it is mediated by the universal gene network commonly possessed by other hemimetabolous insects , such as wnt and TGFβ ( Transforming growth factor beta ) signalings , both of which have multiple roles for embryogenesis and hormonal regulation . Especially , TGFβ signaling should be a focus , because it was reported as a regulator of JH synthesis in the cricket , Gryllus bimaculatus [16] . Moreover , TGFβ signaling was involved in the neuronal development and metamorphosis in Drosophila melanogaster and the German cockroach , Blattella germanica [17 , 18] . To clarify the intrinsic mechanism of soldier differentiation mediated by hormonal changes , comparative transcriptome analysis of the head during each molting process under natural conditions will be most effective , because both JH and ecdysone biosynthetic organs ( corpora allata and molt gland , respectively ) exist in the insect head . In this study , transcriptome analysis ( RNA-seq ) was performed in Z . nevadensis to detect an unknown regulatory factor involved in crosstalk between JH and 20E during soldier differentiation . In an incipient colony of this species , the oldest 3rd instar larva ( No . 1 larva ) molts into a presoldier , whereas the next-oldest 3rd instar larva ( No . 2 larva ) molts into the next instar ( 4th instar larva ) . Gut-purged individuals ( those with the elimination of gut contents before the molt ) are always observed 3 days prior to molt . Therefore , the time frame of worker , presoldier and soldier molts can be determined by records of the day of molt or gut-purge [9 , 12] . Note that because individuals developed beyond the 2nd instar function as workers , they are referred to as workers in this study . Based on the RNA-seq analysis , some important candidate genes were identified , all of which showed increased expressions in the head before the presoldier molt . The function of these genes and relationships with JH and/or 20E during soldier differentiation were confirmed by RNA interference ( RNAi ) and qPCR analysis . Based on the results obtained , we discuss how termite soldiers are differentiated with marked morphological modification via the double molting processes . The expression levels of large numbers of genes ( 14 , 204/15 , 876 genes , 89 . 5% of all genes described in the model OGSv2 . 2; [8] ) were observed by the present RNA-seq data from 3 molting processes ( Fig 1A ) . MDS ( multi-dimensional scaling ) of all expression patterns were similar to one another during the 3 molting processes ( Fig 1B ) . Namely , similar expression patterns were observed in the same developmental stage ( pGP , GP0 , GP3 , M0 or M3 ) , not in each molting process . Note that the expression pattern of the M3 library ( three days after the molt ) in the presoldier molt ( M3p; Fig 1A ) was similar to that of the pGP library ( pre-gut-purging ) . This was due to the nearly equal sampling points of the former and the pGP library in the soldier molt ( pGPs; Fig 1A ) . To narrow down the crucial genes of presoldier development , significantly highly expressed genes were investigated by a comparison with the same developmental stage ( pGP , GP0 and GP3 ) . Highly expressed genes in the presoldier molt compared with those in worker and soldier molts ( i . e . pGPp > pGPw and pGPp > pGPs , GP0p > GP0w and GP0p > GP0s , GP3p > GP3w and GP3p > GP3s ) were 28 genes in total ( Fig 1C , S1 Table ) . Moreover , 103 genes were more highly expressed in the presoldier molt than those in the worker molt and equally expressed compared with soldier molt ( i . e . pGPp > pGPw and pGPp = pGPs , GP0p > GP0w and GP0p = GP0s , GP3p > GP3w and GP3p = GP3s; Fig 1C , S1 Table ) . Consequently , a total of 131 genes were identified , but these numbers were below 1% of all genes observed in this study . Expression patterns of these genes were divided into 4 clusters by cummeRbund ( Fig 1D ) . Expression patterns of 77 out of 131 genes were shown to be related to the molting process ( Cluster 2–4 ) . On the other hand , expression patterns of the remaining 54 genes were not related to the molting process , and might be different during each molt ( Cluster 1 ) . Based on the annotation of these 54 genes by the FlyBase and nr database in NCBI , 3 transcription factors , 1 JH binding protein and some uncharacterized protein genes were observed ( S2 Table ) . The following functional analysis was performed in these target genes . For the screening of genes crucial to soldier differentiation , RNAi treatments of some genes were performed during presoldier molt . A total of 13 candidate genes , including transcription factors ( Znev_04641 , Znev_05644 and Znev_11299 ) , JH binding protein ( Znev_03428 ) and function-unknown hypothetical genes in D . melanogaster ( total 9 ) , were selected for the following functional analysis . There is a possibility that these genes have a crucial role for soldier differentiation , because they may have an important role in JH-dependent gene expression changes . Due to the limited numbers of soldier-destined individuals in the incipient colonies ( basically only one individual in each colony; [12] ) , functional analysis was conducted using the JH analog ( JHA ) -induced presoldier differentiation experiments . The knockdown of 8 out of 13 focal genes did not have any effects on the individuals , and normal presoldiers were induced by JHA treatment ( S3 Table ) . RNAi of three genes ( Znev_05644 , Znev_15631 and Znev_16430 ) had a lethal effect in JHA treated workers ( S3 Table ) . On the other hand , RNAi of Znev_04641 ( transcription factor SOX-11-like isoform X3; hereafter called ZnSox11 ) and Znev_01548 ( uncharacterized protein gene ) resulted in noteworthy phenotypes after the presoldier molt induced by the JHA treatment compared with the GFP controls ( Fig 2A ) . Namely , both RNAi-treated individuals possessed soldier-like phenotypes with well-tanned cuticle despite being the 1st molting stage from workers ( in this case , the 7th instar larvae ) . Moreover , behavioral analysis showed that biting frequency in response to the presence of the ant F . japonica was significantly higher than that of the GFP controls and similar to that of natural soldiers ( Fig 2B , S4 Table ) . In the GFP controls , the 2nd ( i . e . , induced soldier ) molts were observed within 30 days after the 1st molt ( molting rates: 53 . 8% , n = 14/26 ) . However , the 2nd molts were never observed in both ZnSox11 and Znev_01548 RNAi-treated individuals within the same periods ( n = 0/25 and 0/27 , respectively ) . The color nature values ( HSB color model ) of head capsules of both ZnSox11 and Znev_01548 RNAi-treated individuals were intermediate between presoldiers ( = GFP RNAi-treated individuals ) and soldiers ( S1 Fig , S5 Table ) . The double knockdown of ZnSox11 and Znev_01548 caused phenotypes similar to single knockdown of each gene ( S1 Fig , S5 Table ) . High expressions of both genes before the presoldier molt in an incipient colony were confirmed by the real-time qPCR using head samples ( 3 biological replicates ) ( Fig 2C , S6 Table ) . Moreover , expression levels of both genes in the whole body were inhibited by RNAi treatment of JH receptor gene , Methoprene tolerant ( Met ) , during artificial presoldier molting processes ( Fig 2D , S6 Table ) . Based on the blast search , we observed that Znev_01548 contained TGFβ ( Transforming growth factor beta ) propeptide domain ( E-value = 1 . 71e-4: NCBI , S2 Fig ) . RNAi of ZnSox11 and Znev_01548 significantly inhibited the same gene expression levels in heads of JHA-treated individuals compared with the GFP control ( Fig 3A , S6 Table ) . The expression levels of soldier characteristic genes , including ZnTro and ZnLac2 , were activated by RNAi treatment of two genes , especially after the molt , compared with the GFP control ( Fig 3B , S6 Table ) . Furthermore , expression of 20E signaling genes after the molt was also influenced by the RNAi treatments ( Fig 3C , S6 Table ) . Expression levels of ZnEcR in the head just after the JHA-induced presoldier molt were significantly inhibited by Znev_01548 RNAi compared with the GFP control . Expression levels of ZnE75 were also significantly reduced by RNAi of both genes ( Fig 3C , S6 Table ) . On the other hand , expressions levels of ecdysone synthetic genes were not decreased by either RNAi treatment ( S3 Fig , S6 Table ) . RNA-seq analysis demonstrated that expression patterns of all mapped genes were essentially similar in three molting processes ( worker , presoldier and soldier molts ) under natural conditions . These results suggest that soldier differentiation is not related to the large gene expression changes exhibited during postembryonic larval molts . It has been proposed that co-option of a few genes in the influential gene network drives evolution of novel traits , including the sterile caste of hymenopteran social insects [5 , 6] . Similar expression patterns among each of the molts observed here may support that the co-option of a few key genes is also involved in the formation of termite soldiers . Comparative transcriptome analysis before each molt ( pGP , GP0 or GP3 ) provided 131 genes as significantly highly expressed in the head during presoldier differentiation . However , many genes ( 77 genes , 58 . 8% ) exhibited similar expression patterns through each molting process regardless of their phenotypic differences ( Fig 1D ) . There is a possibility that the remaining genes ( 54 genes , 41 . 2% ) possess important functions for soldier differentiation , including a specific morphogenesis and double molting system . Within 54 genes identified ( S2 Table ) , we especially focused on the transcription factors ( Znev_04641 , Znev_05644 and Znev_11299 ) , JH binding protein ( Znev_03428 ) and function-unknown hypothetical genes ( total 9 ) . Functional analysis using the artificial presoldier induction method showed that similar phenotypic effects were caused by ZnSox11 and Znev_01548 RNAi . Knockdown of these genes in the JHA-treated workers ( 7th instar larvae ) resulted in the decrease of developmental stages ( just a single molt ) and activation of soldier-specific morphogenesis with developed mandibles and well-tanned cuticle formations ( Fig 2A ) . The Sox11 gene belongs to SOX ( Sry-related HMG box ) gene family , which has multiple functionalities including embryogenesis , sex determination , and neurogenesis both in vertebrates and invertebrates [19 , 20] . Although the function of Sox11 is unclear in insects , it is involved in the regulation of embryonic development in mammals [21 , 22] , probably through its relationship with TGFβ signaling [21] . Znev_01548 was a function-unknown gene , but interestingly it contained TGFβ propeptide domain , functioned as configuration of LAP ( latency-associated peptide ) region in the TGFβ ligand precursor protein to form homodimer with TGFβ binding protein in mammals [23] . TGFβ signaling is well known as a universal gene network of many biological functions in metazoans ( reviewed in [24] ) . Importantly , it is involved in the regulation of JH biosynthesis in flies and crickets [25 , 16] , and the activation of a 20E receptor gene ( EcR ) in crickets [17] . Present results are not inconsistent with these evidences . Namely , both ZnSox11 and Znev_01548 expression increased after JHA treatment , and was decreased by the knockdown of JH receptor gene , Met ( Fig 2D , S6 Table ) . Moreover , knockdown of both genes inhibited the expression of the 20E signaling gene E75 ( also EcR in the case of Znev_01548 RNAi ) ( Fig 3 , S6 Table ) . One possibility of the function of TGFβ signaling during soldier differentiation is a mediation between JH ( high JH titer levels ) and 20E ( EcR and/or 20E signaling activation ) . The high EcR expression in the head just after the presoldier molt was necessary for the molt into a soldier from a presoldier in Z . nevadensis [11] . Consequently , TGFβ signaling may modulate the double molting processes via the regulation of EcR and/or 20E signaling gene expression changes during soldier differentiation ( Fig 4 ) . Because double knockdown of ZnSox11 and Znev_01548 did not enhance the soldier-like phenotype , there may be a rate-limiting process in TGFβ signaling during soldier differentiation . Further expression and function analysis of other TGFβ signaling members would be required to clarify this possibility . Moreover , expression levels of both ZnSox11 and Znev_01548 also fluctuated during the worker molt ( Fig 2C , S6 Table ) . To know whether TGFβ signaling has another role during the worker molt , RNAi analysis should be performed . All termite species with the soldier caste possess the presoldier stage and the double molting system during soldier differentiation [26] . The presoldier is not engaged in any tasks in the colony and is regarded as solely an intermediate stage for the soldier caste . Therefore , the double molting system may be necessary to generate soldier-specific defensive morphology . However , the present RNAi experiments produced the soldier-like phenotypes with only a single molt from workers , which possibly could be engaged in defensive tasks like natural soldiers . There is a possibility that these individuals with incomplete phenotypes are the first acquired soldiers during the course of termite evolution . The acquisition of JH dependent 20E regulatory system mediated by the TGFβ signaling might then allow presoldier morphogenesis and complete soldier differentiation via the double molting process . To follow this hypothesis , expression and function of other genes in TGFβ signaling should be analyzed in detail , especially in the cockroach lineages which are closely related to termites . We performed the RNA-seq analysis in the head during caste differentiation under natural conditions in Z . nevadensis . Two important genes related to TGFβ signaling were detected as highly expressed genes before the presoldier molt . Functional analyses demonstrated that these genes were involved in the formation of presoldier-specific phenotypes and the activation of the next soldier molt . Based on the gene expression analyses , TGFβ signaling may be involved in the mediation between JH and 20E signalings during termite soldier differentiation . Further analyses will elucidate the role of TGFβ signaling for the double molting system required for soldier differentiation in modern termite species . The results of this study will provide new cutting edge data for the discussion of sterile caste evolution in eusocial insects . Several mature colonies of Z . nevadensis were collected from laurel forests in Hyogo Prefecture , Japan , in May 2013 . All colonies were kept in plastic cases at approximately 25°C in constant darkness until the emergence of newly molted alates . In accordance with previous studies [12 , 27] , 50 incipient colonies were founded by mating male and female alates in 40 mm plastic dishes , and these colonies were kept at approximately 25°C in constant darkness . According to the developmental schedule previously described in an incipient colony [9 , 12] , individuals were collected at the following developmental stages during soldier differentiation ( 3rd instar—presoldier—soldier ) and worker molt ( 3rd instar - 4th instar; Fig 1A ) ; pre gut-purging ( pGP ) , 0 days after gut-purging ( GP0 ) , 3 days after gut-purging ( GP3 ) , 0 days after molting ( M0 ) and 3 days after molting ( M3 ) . For RNA interference ( RNAi ) experiments , 7th instars engaged in worker tasks in the mature colony were sampled from a colony collected in Hyogo Prefecture in May 2015 . Individuals used for RNA extraction were immersed immediately in liquid nitrogen and stored at −80°C . For RNA-sequencing analysis during each molt ( worker , presoldier and soldier molts ) , total RNA was extracted from heads of three individuals using SV Total RNA isolation kit ( Promega Madison , WI , USA ) . The amounts of RNA and DNA in each sample were quantified using a Qubit fluorometer ( Life Technology , Eugene , OR , USA ) , and the quality of RNA was validated using an Agilent 2100 bioanalyzer ( Agilent Technologies , Palo Alto , CA , USA ) . For real-time qPCR analysis to validate the results based on the RPKM ( Reads Per Kilobase of exon model per Million mapped reads ) values ( see below ) , total RNA of each developmental stage was extracted from heads using ISOGEN ( NipponGene , Tokyo , Japan ) . Biological triplicates derived from three different heads were prepared . For real-time qPCR analysis under the RNAi assay of candidate genes , total RNA of 7th instars was extracted from the whole body using ISOGEN . The extracted RNA was purified with RNase-free DNaseI ( Takara , Shiga , Japan ) for removing genomic DNA . Total RNA ( 500 ng ) was used for cDNA synthesis using a TrueSeq sample preparation kit ( Illumina , San Diego , USA ) according to the low throughput protocol . The amount of compound cDNA library was quantified using the qPCR method with a Library Quantification Kit—Illumina/Universal ( KAPA Biosystems , Wilmington , USA ) . Each cDNA library ( 50 μL , 2 nM ) was used for RNA-sequencing , which was performed by single-end 100 bp sequence with the next-generation sequencer Hiseq 2000 ( Illumina ) . One library was prepared in each developmental stage and total 15 libraries were sequenced . All of the reads have been deposited in the DDBJ Sequence Read Archive ( DRA ) database under accession numbers DRA006300 . The quality of all reads was checked by the FastQC program [28] , and low quality reads were removed by SolexaQA software [29] . Adapter sequences were removed using the Cutadapt program [30] . Cleaned reads were mapped to genome sequence data ( gene model OGSv2 . 2; [8] ) using Tophat v2 . 0 . 8 software with Bowtie2 v2 . 1 . 0 . 0 [31] . MDS plots of all expressed genes ( 14 , 204 genes ) were constructed based on the Jensen-Shannon distances by the cummeRbund package [32] using the RPKM values obtained . Expression rates were calculated by the cuffdiff command in Cufflinks software [32] . Highly expressed genes in the head before the presoldier molt were selected by comparison between the same developmental stages described above ( pGP , GP0 and GP3 ) in two molts ( worker and presoldier molts , presoldier and soldier molts ) . According to the expression patterns during molts , highly expressed genes in the head were clustered using the cummeRbund package [32] . Gene similarity searches were conducted using the blastx algorithm with blast+ package ( version 2 . 3 . 0 ) [33] against the FlyBase ( in case of no hits , non-redundant ( nr ) database ) ( performed on 15 May 2017 ) of GenBank in the NCBI server . We set an E-value threshold of ≤1e-4 and bit score threshold of ≥ 40 for BLAST hits . According to the methods of Saiki et al . ( 2014 ) [34] , filter paper was treated with 0 ( for control ) or 10 μg JHA ( pyriproxyfen; Wako , Osaka , Japan ) dissolved in 400 μl acetone and placed in a 90 mm petri dish with ten 7th instar individuals . All petri dishes were kept in an incubator at 25°C in constant darkness and checked for dead or newly molted individuals every 24 hours . For the expression analysis of candidate genes during the presoldier molt induced by the JHA treatment , individuals were sampled at the following 8 points; 10 , 7 , 4 and 1 day before the 1st ( i . e . induced presoldier ) molt , and 0 , 3 , 5 and 7 days after the 1st molt . Note that when the individuals were treated with JHA , the gut-purged individuals were observed approximately 7 days before the 1st molt ( see results ) . Double-strand RNA ( dsRNA ) of candidate genes was generated by the partial cDNA sequences amplified by the gene-specific primers ( S3 Table ) using T7 RNA polymerase with a MEGA script T7 transcription kit ( Ambion , Austin , TX , USA ) . According to previous reports [9–11 , 34 , 35] , GFP was selected as a control gene and dsRNA was generated using the GFP vector pQBI-polII ( Wako , Osaka , Japan ) . Specific primers for the 13 target genes were designed from genome sequence data [8] using Primer3 plus software [36] ( S3 Table ) . Specific primers of the JH receptor gene , Methoprene tolerant ( ZnMet; Znev_9570 ) were designed according to the previous study ( S3 Table; [9] ) . The amount of compound dsRNA was quantified by a NanoVue spectrophotometer ( GE Healthcare Bio-Sciences , Uppsala , Sweden ) , and 500 ng of dsRNAs were injected into the side of the thorax of individuals using a Nanoliter 2000 microinjector ( World Precision Instruments , Sarasota , FL , USA ) at 24 hours after the JHA treatment ( target genes except for ZnSox11 and Znev_01548 , n = 10; ZnSox11 , n = 30; Znev_01548 , n = 30; double knockdown of ZnSox11 and Znev_01548 , n = 10 ) or simultaneously with JHA treatment ( ZnMet , n = 5 ) . Phenotypes of the molted individuals were observed 7 days after the 1st molt . The 2nd ( i . e . induced soldier ) molting rates were measured at day 30 after the 1st molt . For the gene expression analysis under the RNAi treatment , the JHA-treated individuals were collected at 10 , 7 , 4 and 1 day before the 1st molt , and 0 , 3 , 5 and 7 days after the 1st molt . According to the methods of Ishikawa et al . ( 2012 ) [37] , aggression levels of individuals were quantified by the frequency of biting against the wood ant Formica japonica as a hypothetical enemy . One RNAi-treated individual or natural soldier was placed in a 40 mm plastic dish with moistened filter paper , and a F . japonica placed on the end of a toothpick was presented for 3 min . The number of times that F . japonica was bitten were counted . These experiments were replicated 6–9 times using different individuals of both Z . nevadensis and F . japonica . To quantify the color of the head capsule , RNAi treated individuals and natural soldiers were preserved in the FAA solution ( ethanol : formalin : acetic acid = 16 : 6 : 1 ) for 24 hours and stored in 70% ethanol . These fixed individuals were observed using a SZX10 stereomicroscope and 3CCD digital camera XD250-2D ( Olympus , Tokyo , Japan ) . From the captured images of their head capsules , average color properties of the head capsule area were detected using the color picker of Adobe Photoshop CS software 7 . 0 ( Adobe Systems Inc . , San Jose , CA , USA ) . The color properties of each individual were evaluated as the average values of ten points randomly chosen for the HSB ( hue angle , saturation , brightness ) color model [c . f . 11 , 39] . Kruskal-Wallis test and Steel-Dwass tests were performed for comparison among the RNAi-treated individuals and natural soldiers using the statistical software Mac Statistical Analysis ver . 2 . 0 . ( Esumi , Tokyo , Japan ) . The cDNA was synthesized using a High Capacity cDNA Reverse Transcription Kit ( Applied Biosystems ) . The expression level of each gene was quantified using THUNDERBIRD SYBR qPCR Mix ( TOYOBO , Osaka , Japan ) and a MiniOpticon Real-Time System detection system ( Bio-Rad , Hercules , CA , USA ) . For the RNAi-treated 1st molting individuals ( JHA-induced soldier-like individuals; see results ) , expression levels of the following 2 genes ( Troponin and Laccase2 ( Lac2 ) ; hereafter , soldier characteristic genes ) were analyzed to check whether these individuals were genetically similar with the natural soldiers . Troponin , muscle formation gene was highly expressed in soldiers of Reticulitermes flavipes [38] . In the present study , highly Troponin expression levels were observed in soldier heads after the molt ( M3 library , RPKM value = 13479 . 6 ) compared with worker heads ( M3 library , 130 . 9 ) . Lac2 was involved in the cuticular tanning during soldier differentiation and high expression levels were observed in heads of R . speratus [39] and Z . nevadensis soldiers [10] . Specific primers for qPCR of the following target genes , ZnSox11 ( Znev_04641 ) , Znev_01548 , Troponin ( ZnTro; Znev_06448 ) and three ecdysone synthetic genes ( Neverland: ZnNvd; Znev_04416 , Shroud: ZnShr; Znev_16529 , Spook: ZnSpo; Znev_04417 ) were designed from genome sequence data [8] using Primer3 plus software [36] . The qPCR primers of Lac2 ( ZnLac2; Znev_17433 ) , Ecdysone receptor ( ZnEcR; Znev_13925 ) and Ecdysone-inducible protein E75 ( ZnE75; Znev_11406 ) were previously described [10 , 11] . The following six house-keeping genes were used as endogenous controls for constitutive expression: EF1-alfa ( DDBJ/EMBL/GenBank Accession No . AB915828 ) , beta-actin ( No . AB915826 ) , NADH-dh ( No . AB936819 ) and three ribosomal protein genes ( RS49: GeneID: KDR21989 , RPS18: KDR22651 and RPL13a: KDR22610 ) [8 , 10] . The most appropriate gene was evaluated using GeNorm [40] and NormFinder [41] , and EF1-alfa was selected in all real-time qPCR analyses performed in this study ( S4 Table ) . Relative expression levels of the target genes were calculated by adopting the standard curve method . Expression levels were calculated using biological triplicates . Prior to the use of ANOVA , we performed the Browne-Forsythe test on the variance equality . Post-hoc tests among each developmental stage were conducted using Scheffe’s F test . These statistic analyses were performed using the statistical software Mac Statistical Analysis ver . 2 . 0 .
The acquisition of a sterile caste is a key step in animal eusocial evolution . The soldier is the first acquired permanently sterile caste in termites , which are distantly related to hymenopteran eusocial insects ( bees , ants and wasps ) . However , the evolutionary background to acquiring the soldier caste is still largely unclear . Here we performed transcriptome analysis of heads during worker and soldier caste differentiation under natural conditions in Zootermopsis nevadensis , for which the whole-genome sequence is available . Soldiers differentiate from workers via a double molting processes , and the intermediate stage is called a presoldier . Based on the comparison among molting stages , some presoldier-specific highly expressed genes , including transcription factors and several uncharacterized protein genes , were identified . RNAi of two genes presumably involved in the TGFβ signaling resulted in the formation of presoldiers possessing soldier-like phenotypes ( e . g . well-tunned cuticle ) without further molting into soldiers . Expression levels of 20E signaling genes were negatively affected by the RNAi treatment of these genes . This study provides a novel insight into the hormonal control for termite caste differentiation .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "methods" ]
[ "invertebrates", "medicine", "and", "health", "sciences", "rna", "interference", "gene", "regulation", "regulatory", "proteins", "dna-binding", "proteins", "animals", "organisms", "physiological", "processes", "transcription", "factors", "epigenetics", "zoology", "molting", "genetic", "interference", "proteins", "gene", "expression", "termites", "insects", "arthropoda", "animal", "physiology", "biochemistry", "rna", "signal", "transduction", "eukaryota", "cell", "biology", "nucleic", "acids", "physiology", "genetics", "tgf-beta", "signaling", "cascade", "biology", "and", "life", "sciences", "cell", "signaling", "biological", "defense", "mechanisms", "signaling", "cascades" ]
2018
TGFβ signaling related genes are involved in hormonal mediation during termite soldier differentiation
As well as their importance to nutrition , fatty acids ( FA ) represent a unique group of quorum sensing chemicals that modulate the behavior of bacterial population in virulence . However , the way in which full-length , membrane-bound receptors biochemically detect FA remains unclear . Here , we provide genetic , enzymological and biophysical evidences to demonstrate that in the phytopathogenic bacterium Xanthomonas campestris pv . campestris , a medium-chain FA diffusible signal factor ( DSF ) binds directly to the N-terminal , 22 amino acid-length sensor region of a receptor histidine kinase ( HK ) , RpfC . The binding event remarkably activates RpfC autokinase activity by causing an allosteric change associated with the dimerization and histidine phosphotransfer ( DHp ) and catalytic ATP-binding ( CA ) domains . Six residues were found essential for sensing DSF , especially those located in the region adjoining to the inner membrane of cells . Disrupting direct DSF-RpfC interaction caused deficiency in bacterial virulence and biofilm development . In addition , two amino acids within the juxtamembrane domain of RpfC , Leu172 and Ala178 , are involved in the autoinhibition of the RpfC kinase activity . Replacements of them caused constitutive activation of RpfC-mediated signaling regardless of DSF stimulation . Therefore , our results revealed a biochemical mechanism whereby FA activates bacterial HK in an allosteric manner , which will assist in future studies on the specificity of FA-HK recognition during bacterial virulence regulation and cell-cell communication . Quorum-sensing is a process that bacterial cells communicate with each other to elicit specific physiological responses , including virulence against hosts [1 , 2] . How single-celled bacteria detect and respond to population density is a fundamental question in studying quorum sensing . Previous studies have reported that a number of chemicals , such as acylated homoserine lactones , peptides , quinolones , and small molecular fatty acids ( FA ) , were implicated in the “bacterial languages” that are usually recognized by bacterial sensor histidine kinases ( HK ) to elicit quorum sensing [1 , 3] . Typically , HK and its cognate response regulator ( RR ) constitute a two-component signal transduction system ( TCS ) , the predominant detection-response mechanism in prokaryotic cells . The N-terminal input region of an HK detects specific stimuli , and an invariant histidine residue within its C-terminal dimerization and histidine phosphotransfer ( DHp ) domain is autophosphorylated . The HK then modulates the phosphorylation level of the cytoplasmic RR by its phosphotransferase or phosphatase activity . Eventually , the RR uses its C-terminal output domain to regulate gene expression or cellular behavior [4] . In the past three decades , the basic biochemical processes of protein phosphorylation and dephosphorylation during TCS regulation have been well documented , however , as the first event to trigger the cell-cell communication , only a few of ligand-HK interactions were experimentally investigated [5 , 6] . Therefore , how HK recognizes various signals , especially in non-model bacteria , remains incompletely studied [7] . The major difficulty in studying HK-ligand interactions is the fact that the majority of HK are membrane-bound proteins with various hydrophobic helices [8] . Usually HK are enzymatic inactive in solutions containing detergents . Traditional strategies that express soluble truncated HK by deleting input regions and transmembrane helices then prevent the investigation of ligand-HK interactions [9 , 10] . In addition , HK are quite difficult to be crystalized so that the high resolution , three-dimensional structure of full-length HK is usually unavailable . This impedes the understanding of the structural mechanism of ligand-HK interaction [5] . Furthermore , signals like FA are hydrophobic molecules whose dissolution require organic solvents , these properties made the HK-ligand interactions difficult to be measured by commonly used biophysical methods , such as surface plasmon resonance ( SPR ) and isothermal titration calorimetry ( ITC ) [11 , 12] . Therefore , multi-disciplinary approaches based on extensive genetic analysis are needed to investigate FA-HK relationships . In bacteria , diffusible signal factor ( DSF ) is a special family of signaling FA molecules . Various DSF-family members ( such as DSF , BDSF , CDSF , SDSF , etc . ) have been found to control quorum sensing and virulence in a number of bacteria , including plant pathogen Xanthomonas spp . and human pathogens Pseudomonas aeruginosa , Stenotrophomonas maltophilia , and Burkholderia cepacia [13 , 14] . These FAs take part in inter-species and inter-kingdom communications between bacteria and other organisms , including bacteria , fungi and host plants [15–21] . The first identified molecule of this family , DSF , was found in the Gram-negative bacterium Xanthomonas campestris pathovar ( pv . ) campestris , causal agent of black rot disease of cruciferous plants which encode approximately 52 HKs [22] . DSF is a medium-chain FA with a cis-11-methyl-dodecenoic acid structure [23] . Previous studies have revealed that extracellular DSF stimulates a TCS , RpfC-RpfG , to control bacterial virulence and quorum-sensing [24 , 25] . Of them , RpfC encodes a putative hybrid-type HK with multiple phosphorylation sites , while its cognate RR , RpfG , was the first HD-GYP domain-containing protein proved to have the phosphodiesterase activity to hydrolyze second messenger c-di-GMP into GMP [26] . At low bacterial cell density , RpfC binds to and represses the DSF synthase RpfF by a receiver ( REC ) domain , preventing production of DSF [27] . At high cell density , high concentration of extracellular DSF activates RpfC-RpfG system to degrade c-di-GMP , releasing the suppression of c-di-GMP on a global transcription factor Clp that controls the expression of multiple virulence factors [15 , 28 , 29] . However , because of the afore-mentioned technical difficulties in studying membrane-bound HK-FA interactions , whether DSF is the ligand of RpfC and biochemical mechanism of RpfC activation is unknown . Recently , two PAS-domain-containing HKs , RpfS and RpfR , were shown to bind DSF in X . campestris and BDSF in Burkholderia cepacia , respectively [30 , 31] . However , both RpfS and RpfR are soluble , cytoplasmic proteins without transmembrane helices , they are probably cytoplasmic receptors of intracellular DSF and unlikely cell-surface receptors to sense extracytoplasmic stimuli . Therefore , how the bacterial pathogen detects DSF dispersed in external environment and triggers cell-cell communication remains an opening question . In this work , we show that DSF directly binds to a short N-terminal sensor of RpfC , elevating levels of RpfC autophosphorylation . A group of amino acid residues within the region adjoining to membrane were indispensable to DSF-RpfC binding and autokinase activation . DSF-RpfC interaction resulted in allosteric change in the DHp and catalytic ATP-binding ( CA ) regions , which may promote in trans phosphorylation of RpfC . In addition , substitutions of two amino acids within the juxtamembrane domain of RpfC caused constitutive activation of the HK . Our data revealed the biochemical mechanism responsible for the interaction between HK and FA , and provided insight into bacterial signaling during cell-cell communication . RpfC belongs to a group of hybrid-type of HK with sensing mechanisms associated with membrane-spanning helices [8] . The putative secondary structure of RpfC has two characteristics different from the prototypical HKs ( Fig 1A ) : Firstly , the signal input region of RpfC contains five hydrophobic TM helices and a putative 22-amino acid ( aa ) -length , periplasmic sensor at the most front end of its N-terminus . Secondly , there is a short juxtamembrane domain ( 16 aa-length ) , rather than a HAMP linker ( about 50 aa-length ) , connects the input region to DHp-CA domains . In addition , RpfC also contains a C-terminal histidine phosphotransfer ( HPt ) domain and a REC domain ( Fig 1A ) . The enzymatic activity of RpfC has never been investigated before . To biochemically confirm that RpfC is a HK , a truncated , soluble RpfC protein ( RpfCΔinput ) lacking the N-terminal input region ( including sensor and TM domains ) was obtained and purified . However , RpfCΔinput did not exhibit any detectable autokinase activity ( Fig 1B ) , suggesting that the input region is critical for maintaining enzymatic activity . To address this question , we obtained a full-length RpfC protein ( RpfCFL ) with a C-terminal His6 epitope tag . Two membrane-embedded forms of RpfCFL , liposome and inverted membrane vesicle ( IMV ) , were reconstructed and purified . As shown in Fig 1C and 1D , both forms of RpfCFL exhibited clear autokinase activity , making it possible to enzymatically investigate the mechanism of RpfC activation . To determine if DSF affects the enzymatic activity of RpfC , DSF was added to reaction mixtures containing the liposome or IMV forms of RpfCFL . As shown in Fig 1C and 1D , the level of RpfCFL-P phosphorylation approximately doubled compared with the control . Kinetic analyses of the IMV and liposome forms of RpfCFL showed an increase in the phosphorylation level of the IMV form at 30 s post DSF addition , whereas a similar increase was not detected until 2 min for the liposome form . This difference might be caused by variation in the phospholipid compositions of the IMV and liposome forms , which would affect autokinase activity . In addition , dose-response analysis of DSF on the activity of RpfC revealed that addition of 0 . 5 μM DSF was sufficient to elicit a detectable increase in the level of RpfCFL-P ( Fig 1E and 1F ) . This concentration agrees with the previously reported minimal bioactive concentration of DSF ( approximately 0 . 5 μM ) that required for eliciting cell-cell communication [23] . Increasing the DSF concentration resulted in a logarithmic increase in the RpCFL-P level , and RpfCFL-P levels tapered off as they neared 20 μM ( Fig 1F ) , suggesting that the system had reached saturation point . RpfC is a hybrid histidine kinase that contains additional HPt and REC domains ( Fig 1A ) . To exclude the possibility that the elevation of RpfCFL-P levels was caused by a change in DSF-dependent phosphoryl transfers from the DHp domain to these domains , the conserved phosphorylation sites within the HPt and REC domains were independently replaced [RpfCH657A and RpfCD512V] . The IMV forms of the two recombinant RpfC proteins were used in the phosphorylation assay . As shown in S1 Fig , neither of the amino acid replacement affected the DSF-dependent elevation of RpfC autokinase activity . Taken together , these findings provide direct biochemical evidences to demonstrate a long-term supposition that RpfC is an HK whose autokinase activity can be activated by the ligand DSF . Because membrane-bound HK usually employ periplasmic sensors and TM helices to detect signals , mutagenesis was used to identify regions critical for detecting DSF . A series of in-frame rpfC deletion mutants in a ΔrpfF genetic background that lost the capability to synthesize endogenous DSF were constructed . These constructs include a mutant with a deletion of the putative short sensor region ( rpfCΔsensor ) , four mutants with their TM regions deleted in pairs to maintain the general topology of the protein ( rpfCΔTM1-2 , rpfCΔTM2-3 , rpfCΔTM3-4 , rpfCΔTM4-5 ) , and a mutant with the input regions ( sensor and TM regions ) completely deleted ( rpfCΔinput ) . Western blotting analyses revealed that deletions of the TM2–TM3 or input region caused instability of RpfC , whereas deletion of the short sensor region resulted in increase of the cellular levels of RpfC protein ( S2A Fig ) , indicating the presence of a negative autoregulatory loop mediated by the sensor . As deletion of rpfF completely eliminated the synthesis of endogenous DSF , these double mutants were used to determine whether the different input regions were functional in sensing exogenously added DSF . Since the DSF-RpfC regulated , extracellular protease ( EXP ) activity can be directly observed without staining during bacterial growth [32] , it was selected to be measured as a representative phenotype of the DSF-RpfC regulation . As compared with the positive control ( ΔrpfFΔrpfC-rpfC ) , EXP activities of all these mutants were severely decreased , most likely because they lost the ability to detect exogenous DSF stimulation ( Fig 2A ) . To quantify the effect of DSF perception in these mutants , we used pHM2 vector to construct a biosensor ( PengXcc-GUS ) by fusing β-glucuronidase gene ( GUS ) to the promoter of engXcc ( XC_0639 ) , which encodes an extracellular endoglucanase specifically regulated by the RpfC-RpfG system [15 , 23] . After providing the biosensor in trans to these mutants , GUS activity assay revealed that the transcription level of engXcc was reduced to 13 . 9–18 . 6% of that of the control ( Fig 2B ) . Similarly , when DSF was added to the bacterial cultures , the ability of these mutants to form biofilms and produce extracellular polysaccharides ( EPS ) was significantly decreased ( Fig 2C and 2D ) , exhibiting deficiencies in sensing DSF . To further investigate the regulatory function of the input region of RpfC in virulence , the same deletions as described above were generated in a wild-type ( WT ) background . Phenotypic characterization revealed that bacterial virulence ( Fig 3A and S3A Fig ) , EXP production ( Fig 3C ) , biofilm formation ( Fig 3D ) , and EPS production ( Fig 3F ) of these mutants were all significantly reduced compared with the control . Quantification of the PengXcc-GUS activity ( provided in trans by a recombinant pHM2 vector in each strain ) showed that engXcc expression levels in these mutants decreased to 6 . 2–7 . 9% of the WT level ( Fig 3B ) . In addition , as rpfC negatively modulates DSF synthesis , to measure the DSF production of the mutants , each strain was spotted onto a NYG-milk plate in close proximity to a ΔrpfF mutant deficient in endogenous DSF production . As shown in Fig 3E , deletion of the sensor region did not release the suppression of DSF production , while deletion of the transmembrane regions ( TM12-TM45 ) moderately decreased the inhibition of DSF synthesis by RpfC , since the rpfF mutant exhibited higher EXP activity to degrade milk . Deletion of the entire input domain ( sensor and TM regions ) resulted extensive secretion of EXP , suggesting that the RpfC-mediated inhibition of DSF synthesis is remarkably eliminated . Taken together , mutational analyses suggested that the sensor and TM regions of RpfC are all involved in DSF detection and regulation of bacterial virulence . Of these , the function of the RpfC sensor appears to be particularly important as its deletion resulted in an increase in the cellular amount of RpfC protein , and had no effect in suppression of DSF-regulated EXP production as the other regions of input domain . The following analysis therefore mainly focused on the possible interaction between the RpfC sensor and DSF . We proposed that the RpfC sensor contains the amino acids essential for sensing the DSF signal . Multiple-Alignment of the RpfC sensor sequences from orthologs of close-relative bacteria of X . campestris pv . campestris showed that 15th to 22nd amino acids are highly conserved in species belonging to the Xanthomonadaceae family ( Fig 4A ) , indicating this region is critical in function . Thereafter , alanine-scanning mutagenesis was used to identify essential amino acids in recognizing DSF . A full-length rpfC sequence was first amplified and then inserted into the pHM1 vector . Next , 19 non-Ala codons , except the initial Met1 residue , within the sensor region were individually point mutated into the Ala codon . The two indigenous Ala ( Ala16 and Ala21 ) were also individually replaced into Val ( Fig 1A ) . These 21 recombinant vectors were then individually electroporated into a ΔrpfFΔrpfC double mutant strain containing a GUS reporter fused to the promoter region of engXcc on the bacterial chromosome . Western blotting analysis revealed that apart from RpfCK2A , whose cellular amount was relatively low , all recombinant RpfC proteins were stable ( S2B Fig ) . An exogenous DSF stimulation assay was then used to compare the engXcc transcription levels of these mutants to that of the positive control strain containing a plasmid-borne rpfC . As shown in Fig 4B , in the absence of exogenous DSF , replacements of S3A , L8A , R11A , D17A and Q22A caused slightly but significant changes in PengXcc activity ( P ≤ 0 . 05 ) . In the presence of exogenous DSF , three replacements , D17A , S18A , and Q22A , caused significant reductions in the PengXcc activity ( levels 22 . 9–32 . 9% of the WT level ) . The R15A and E19A substitutions resulted in significant but intermediate reductions in the activation of the PengXcc ( 39 . 0–66 . 1% of the WT level ) . In contrast , the S3A replacement caused a stable , significantly increase in the PengXcc activity ( to 110 . 0% of that of the control , Fig 4B ) . The above result revealed that substitutions in the amino acids from Arg15 to Gln22 ( except Ala16 , His20 and Ala21 ) of RpfC sensor resulted in deficiencies in sensing DSF . Consequently , further amino acid substitution analyses were conducted , which includes construction of recombinant strains containing a plasmid-borne RpfCR15K , RpfCR15H ( Fig 4C ) , RpfCD17E , RpfCD17N ( Fig 4E ) , RpfCS18T , RpfCS18C ( Fig 4I ) , RpfCE19D , RpfCE19Q ( Fig 4F ) , RpfCQ22E , and RpfCQ22N ( Fig 4H ) in the genetic background of ΔrpfFΔrpfC double mutations . Under DSF stimulation , RpfCR15K strain has similar PengXcc activity to that of the control ( 110% level ) in sensing DSF ( Fig 4C ) , suggesting that the positively charged , polar residues with similar side chains ( R or K ) are important in this location . RpfCS18T substitution , which naturally occurred in several species belonging to the genus Xanthomonas ( Fig 4A ) , significantly increased the PengXcc activity to the 139% level of the control ( Fig 4I ) , implying that the hydroxyl oxygen in the side chain of Ser ( S ) or Thr ( T ) is essential to sense DSF . However , all the other substitutions resulted in significant decreases of the PengXcc activity , indicating that these amino acids in the RpfC sensor region are essential and evolutionarily fixed to these locations ( Fig 4C–4I ) . In addition , recombinant strains containing RpfCA16G , RpfCA16D ( Fig 4D ) , RpfCA21G and RpfCA21D ( Fig 4G ) were also constructed . PengXcc activity assay showed that RpfCA16G and RpfCA21G replacements did not impact the DSF perception ( Fig 4D and 4G ) . If the nonpolar Ala in the two sites were replaced by the negatively charged , polar Asp that disrupts the native conformation of the sensor region , the PengXcc activities were significantly decreased to 78% and 52% of the control , respectively ( Fig 4D and 4G ) . Since the RpfCS18T substitution increased the PengXcc activity under DSF treatment and is the naturally occurred variation ( Fig 4A and 4I ) , we further constructed 16 recombinant strains in the genetic background of ΔrpfFΔrpfC double mutant . Each strain contains a replacement of Ser18 of RpfC to one of the other 16 amino acids besides afore-mentioned Ala , Thr , and Cys . Quantification of PengXcc activity showed that except the RpfCS18T replacement , all the other substitutions led to significantly decrease in the engXcc expression ( Fig 4I ) , supporting the view that a Ser/Thr residue in this site is critical in sensing DSF . To genetically evaluate the biological roles of the identified amino acids in sensing DSF , phenotypes of the recombinant bacterial strains ( in the background of ΔrpfFΔrpfC double mutant ) were examined following addition of exogenous DSF . Compared with the rpfC complementation strain , the strain containing RpfCS3A replacement showed similar levels of EPS production , biofilm formation and EXP production ( Fig 5A–5C ) , while strains containing RpfCR15A , RpfCD17A , RpfCS18A , RpfCE19A , and RpfCQ22A replacements exhibited significant decreases in EXP and EPS production , as well as in biofilm formation ( Fig 5A–5C ) . Strains containing a RpfCS18T substitution had similar or increased levels as the control in EXP activity , EPS production and biofilm formation ( Fig 5D , 5E and 5F ) . Collectively , these genetic analyses suggest that five amino acids of Arg15 , Asp17 , Ser18 , Glu19 , and Gln22 adjoining to the transmembrane helices of RpfC are essential in detecting DSF , which is in parallel to the fact that they are highly conserved in bacterial evolution . Besides DSF perception , the roles of these essential amino acids in controlling bacterial virulence were also analyzed . Recombinant bacterial strains were constructed in the ΔrpfC background , each contains a pHM1 vector to produce a RpfC derivate with an amino acid replacement . Western blotting analysis revealed that RpfC protein amounts of these strains were stable ( S2J and S2K Fig ) . Plant inoculation assays showed that R15A , D17A , S18A , E19A and Q22A substitutions resulted in substantial attenuation in virulence against host cabbage B . oleraceae ( Fig 5G and S3B Fig ) , while the S18T replacement did not affect bacterial virulence ( Fig 5G and S3C Fig ) . In addition , the production of one of the major virulence factors of X . campestris , EPS , was also significantly decreased in all of the tested strains except of the one containing RpfCS18T replacement ( Fig 5H and 5J ) . These data strongly suggested that deficiency in DSF binding and perception resulted in deficiencies in bacterial virulence . It is noticeable that S3A substitution also resulted in remarkably decrease in virulence , albeit it caused slight hypersensitivity in detecting DSF as afore-mentioned ( Fig 5G and S3B Fig ) . This result suggests that besides DSF perception , S3A is involved in additional , unknown function in regulating virulence . To detect a possible direct DSF-RpfC interaction , microscale thermophoresis ( MST ) was used because this technique is superior in studying membrane-bound receptors that are embedded in liposomes or nanodiscs with diverse ligands such as fatty acids and metals [33 , 34] . As shown in Fig 6A , DSF bound to the RpfCFL liposome with a dissociation constant ( Kd ) of 0 . 82 ± 0 . 12 μM , which represents a stronger binding affinity than those of the RpfS-DSF and RpfR-DSF interactions [31 , 35] . However , DSF didn’t bind to the truncated RpfCΔsensor liposome ( Fig 6B ) and soluble RpfC protein without input region ( RpfCΔinput , Fig 6C ) . Thermal shift assay ( TSA ) using differential scanning fluorimetry was also employed to measure the DSF-RpfC interaction . As shown in Fig 6E , during thermal denaturation , addition of DSF resulted in the significant increase of melting temperature ( Tm ) of RpfCFL liposome from 60 . 27°C to 64 . 27°C ( 5 μM DSF vs . 10 μM RpfC ) or 66 . 27°C ( 10 μM DSF vs . 10 μM RpfC ) , strongly supporting a direct binding between DSF and RpfC . When the RpfCΔsensor liposome and soluble RpfCΔinput were applied in TSA , no thermal shift was detected after DSF stimulation ( Fig 6F and 6G ) . These MST and TSA results suggest that DSF binds to the sensor region of RpfC . To directly detect the sensor-DSF interaction , the sensor peptide fused with a glutathione S-transferases ( GST ) tag was successfully obtained by a pGEX6P-1 expression system , and this sensor peptide was purified by on-column cleavage together with size exclusion chromatography to remove the GST tag . MST analysis revealed that DSF bound to the sensor peptide with a similar affinity of the full-length RpfC liposome ( Kd = 0 . 14 ± 0 . 04 μM , Fig 6D ) . Since this sensor peptide only contains a tryptophan ( Trp7 ) , its autofluorescence is quite low and not applicable in TSA , circular dichroism spectra ( CD ) analysis was used to measure the effect of DSF stimulation on the secondary structure of sensor peptide . CD analysis showed that after addition of DSF , the secondary structure of the short peptide was remarkably changed . The ratio of strand was gradually elevated along with the increase of DSF concentration ( Fig 6H ) , while DSF stimulation did not have recognizable impact on the secondary structure of GST ( Fig 6I ) . We further measured the impacts of substitutions of identified essential amino acids on the binding affinity with DSF . RpfCFL with corresponding replacements of amino acids in the sensor region were purified , assembled into liposomes , and their interactions with DSF were measured by MST . As shown in S4 Fig , substitutions of RpfCR15A , RpfCD17A , RpfCS18A , RpfCE19A and RpfCQ22A completely eliminated the binding between DSF and RpfC liposomes . In contrast , the RpfCS3A replacement decreased the Kd value to 0 . 613 ± 0 . 34 μM ( S4A Fig ) , indicating that the RpfCS3A substitution slightly enhanced the DSF-RpfC interaction . Collectively , the above results experimentally demonstrated that DSF directly binds to the sensor domain of HK RpfC . Since the structures of RpfC and its orthologs remain unclear , limited proteolysis together with shotgun mass spectrometry were used to assess the conformational changes of RpfC involved in detecting DSF . During the analysis , the non-hydrolyzable ATP analog adenosine 5′ ( β , γ-imido ) triphosphate ( AMP-PNP ) was added as a mimic for nucleotide binding . As shown in Fig 7A and S5 Fig , following addition of DSF to the reaction mixture , the patterns and amounts of most of the degraded RpfCFL liposome fragments were similar to those of the control ( DSF minus ) . However , DSF stimulation repeatedly caused a large accumulation of a protein fragments ( 30 kDa , including those in the S3A substitution , S5A Fig ) . Nano-LC-MS/MS analysis revealed that the band represented the DHp-CA region of RpfC ( from 192nd to 474th aa . ) . Proteolysis of RpfCFL liposomes with D17A , S18A , E19A , and Q22A replacements revealed similar degradation footprints , regardless of the presence or absence of DSF ( Fig 7B and S5 Fig ) . These results suggest that the binding of DSF to the RpfC sensor caused conformational changes in the HK associated with the DHp-CA region . To investigate whether substitutions of the identified essential amino acids influence the ability of the RpfC autokinase to react to DSF , phosphorylation levels of the liposomes of the RpfCFL derivatives were compared to that of the WT . As shown in Fig 7C , stimulation of the RpfCS3A substitution using a physiological concentration of DSF ( 0 . 75 μM ) resulted in hypersensitivity in detecting DSF , with the RpfC-P level remarkably increased under stimulation . RpfC with R15A , D17A , S18A , and E19A replacements exhibited substantially decreased autokinase activities , regardless of the absence or presence of DSF . The Q22A replacement did not affect RpfC autophosphorylation in the absence of DSF , but decreased DSF sensing , as shown by the significant decrease in RpfC-P levels under DSF treatment ( Fig 7C ) . In addition , the autophosphorylation level of RpfCS18T was increased under DSF stimulation , albeit that the increase was slightly lower than that of the WT RpfC ( Fig 7C ) . These results suggest that the identified essential amino acids , especially those located in the region adjoining to membrane , play important roles in RpfC autophosphorylation and DSF perception . RpfC contains a 16 aa-length , short juxtamembrane domain between the transmembrane helices and DHp-CA domains ( Fig 1A ) . The juxtamembrane domain is highly conserved among RpfC orthologs from close-relatives of X . campestris pv . campestris ( Fig 8A ) , implying that it has an important role in activating the RpfC autokinase after ligand perception . Alanine-scanning mutagenesis was again used to analyze the function of this region . As shown in Fig 8B , under stimulation of near-saturated concentration of DSF , 13 amino acid substitutions ( in the background of ΔrpfFΔrpfC-rpfC background ) gave rise to significantly decrease of the capability to sense DSF that is quantified by PengXcc activities . In the absence of DSF , although 12 replacements also caused significant decrease in the background PengXcc activities , it is noticeable that two substitutions , RpfCL172A and RpfCA178D , exhibited significant elevation of PengXcc activities ( Fig 8B ) . When a low concentration of DSF ( 1 μM ) was applied in treatment , both strains with RpfCL172A or RpfCA178D substitution also show constitutive activation of RpfC ( Fig 8C ) . Especially , the PengXcc activity of the strain containing RpfCA178D replacement in the absence of DSF treatment even increased to the similar level to that of the DSF stimulation ( Fig 8C ) . The phenotypes of the bacterial strains were also characterized . Without DSF stimulation , recombinant bacterial strains containing RpfCL172A and RpfCA178D , which were constructed in the genetic background of ΔrpfCΔrpfF double mutation , produced more biofilm , EPS , and EXP towards the levels of the positive control strain ( ΔrpfFΔrpfC-rpfC ) under DSF stimulation ( Fig 8D , 8E and 8G ) . For example , in the absence of DSF treatment , RpfCL172A and RpfCA178D replacements caused significantly increases of bacterial EPS production to 259% and 247% levels of that of the control strain , respectively , similar to the EPS amounts that were generated under DSF stimulation . Since RpfCL172A and RpfCA178D substitutions constitutively activated the RpfC-regulated processes regardless of the presence or absence of DSF , we reasoned that mutations in the codons of Leu172 or Ala178 suppress the phenotypic deficiencies caused by mutations in the codons of essential amino acids of the sensor region in detecting DSF . To challenge this , we selected the Ala178 site for further genetic epistatic analysis . Three double mutants in the background of ΔrpfFΔrpfC-rpfCA178D were constructed by point mutating the codon of Asp17 , Ser18 , or Gln22 . PengXcc activity assay revealed that all of these double mutations ( rpfCD17A-A178D , rpfCS18A-A178D and rpfCQ22A-A178D ) significantly suppressed the deficiency in the engXcc expression that is caused by the point mutations in the codons of essential amino acids ( Fig 8F ) , regardless of DSF stimulation . In addition , although both deletion of rpfF gene and point mutation of the codons of essential amino acids sensing DSF caused serious decrease in bacterial virulence , RpfCA178D substitution suppressed the deficiency by recovering bacterial virulence level toward that of the positive control ( Fig 8H and S3D Fig ) . The above genetic analysis suggests that the Leu172 and Ala178 are involved in autoinhibition of RpfC kinase activity without DSF stimulation . In vitro autokinase assay then revealed that the autophosphorylation levels of the recombinant RpfCL172A and RpfCA178D proteins are remarkably higher than that of the WT RpfC in the absence of DSF treatment , exhibiting a constitutive activating state ( Fig 8I ) . Collectively , these results support a view that the juxtamembrane domain of RpfC inhibits its autokinase activity when the concentration of DSF is low . However , DSF perception by the sensor region releases this inhibition to activate the RpfC autophosphorylation . During this process , Leu172 and Ala178 in the juxtamembrane domain play critical roles in inhibiting the RpfC activity . How a ligand interacts with HK is one of the fundamental questions in studying bacterial quorum sensing . Here , we provided enzymological , genetic and biophysical evidences to demonstrate that a HK of X . campestris , RpfC , is a bona fide membrane-bound receptor that directly binds a fatty acid signal , DSF . The results confirmed a long-held hypothesis regarding cell-cell communication in phytopathogenic bacteria [23 , 25 , 36] . DSF binds with high affinity to a 22-amino acid sensor region in the N-terminal of RpfC ( Figs 1 and 6 ) . The binding of DSF causes allosteric change associated with the DHp-CA domain of RpfC , which facilitates RpfC autophosphorylation ( Fig 7 ) . Systematic mutational investigation together with biochemical analysis identified six essential residues in the DSF-RpfC interaction ( Figs 4 and 5 ) . Of these , five amino acids located in the region adjoining to membrane are indispensable for DSF-RpfC binding ( Figs 4 and 5 ) , while the Ser3 of the RpfC sensor region is functionally unique , as replacement of this residue resulted in slight hypersensitivity in detecting DSF . In addition , two point mutations ( rpfCL172A and rpfCA178D ) in the coding sequence of RpfC juxtamembrane domain effectively suppressed the phenotypic deficiencies caused by mutations in the sensor coding region , which is due to the constitutive activation of RpfC autokinase . These results support a molecular model ( Fig 9 ) that the juxtamembrane domain inhibits the autokinase activity of RpfC when the extracellular concentration of DSF is low . However , when the DSF concentration increases along with the rise of bacterial population , DSF binds to the sensor region of RpfC and activates the HK by releasing this inhibition . To our best knowledge , this work provides the first experimental evidence to support a direct membrane-bound HK-FA interaction during bacterial quorum sensing and regulation of virulence . In addition to their roles in nutrition , FA acts as important signaling molecules in both eukaryotes and prokaryotes . For example , in animals , free FA signals are detected by G-protein-coupled receptors ( GPCR ) or receptor tyrosine kinases ( RTK ) [37] . A recent structural study revealed that FA ligands of an human insulin secretion modulator , GPR40 , bind to hydrophilic/positively charged residues in various docking sites that are formed by the characteristic seven-TM helices of GPCR [38] . In addition , complex FA , such as cholesterol , inhibits the autokinase activity of human RTK epidermal growth factor receptor ( EGFR ) by influencing membrane heterogeneity-mediated transmembrane signal transduction [39] . In the present work , MST and TSA analyses revealed that DSF binds RpfC liposome with a relatively strong affinity ( Kd = 0 . 82 ± 0 . 12 μM , Fig 6A ) . This binding affinity is higher than those of the DSF-RpfR ( 1 . 37 μM , measured by ITC ) and DSF-RpfS ( 1 . 40 μM , by ITC ) interactions [31 , 35] , while both RpfR and RpfS are cytosolic , soluble proteins without transmembrane helix . In vitro autokinase assay showed that 0 . 5 μM DSF was sufficient to activate RpfC autokinase ( Fig 1 ) . This concentration is very close to the reported minimum DSF concentration in eliciting cell-cell communication [23] , and in the physiological range of extracellular DSF , which approximately ranged from 0 . 002 to 27 . 4 μM dependent on the growth stages of bacterial population [27 , 40 , 41] . Our data suggest that the N-terminal short sensor region of RpfC plays a fundamental role , if not exclusive , in directly binding DSF ( Fig 6D and 6H ) . In the sensor region , five amino acids adjoining to membrane of the RpfC are highly conserved in the bacteria belonging to the Xanthomonadaceae family ( Fig 4A ) . It is likely that these amino acids form a primary docking site for DSF . This hypothesis is supported by the following data: 1 ) Replacements of these residues completely dissociated the DSF-RpfC interaction ( S4 Fig ) . 2 ) This region contains hydrophilic or charged residues , especially Ser18 which can be functionally substituted by a Thr residue and favors hydrogen bond formation , most possibly with the carboxyl group of DSF . 3 ) Replacement of these residues completely eliminated the DSF-triggered , conformational change associated with DHp-CA domain ( Fig 7 ) . 4 ) Amino acid replacements , especially of Ser18 and Gln22 , decreased the level of autophosphorylation of RpfC in response to DSF ( Fig 7 ) . The above results suggest that the short sensor region of HK likely acts as a “hook” to catch DSF when it is diffusely passing through the periplasm . In a previous report , ITC analysis revealed that DSF binds to the PAS_4 domain of RpfS of X . campestris [31] . Collectively , it suggests that the binding sites of DSF to proteins are diverse and need to be investigated further . It is noteworthy that the function of Ser3 within the RpfC sensor appears unique . This residue is highly conserved in all sequenced Xanthomonas species , whereas in other closely related bacterial species , Ala , Asp , Asn and Lys , respectively , occupy this position ( Fig 4A ) . In the presence of exogenous DSF , S3A replacement caused hypersensitivity of RpfC in DSF detection: the binding affinity increased by approximately 20% compared with WT RpfC ( S4 Fig ) , and the autokinase activity were also elevated ( Fig 7C ) . However , as with the other essential residues , S3A replacement also attenuated virulence ( Fig 5G and S3B Fig ) , suggesting that the Ser3 plays an additional role in regulating virulence , or DSF-RpfC binding affinity is subtly optimized during evolution so that any abnormality is detrimental to bacterial fitness . The role of Ser3 in binding DSF remains unclear . In previous study on the autoinducer CAI-I and HK CqsS in Vibrio cholerae [5] , it revealed the length of hydrocarbon chain of autoinducer is critical in the ligand-HK interaction . The length of the FA chain is also a critical parameter in the activity of DSF-family compounds [23] , therefore , one possible function of Ser3 is to act as a “ruler” to help suitable FA molecules gain entry into the RpfC docking site . However , further evidence is needed to clarify this hypothesis . How a ligand activates a membrane-bound HK remains an important question . To date , several structural mechanisms for HK activation have been proposed [42] , including scissor blade [43] , piston-like change [44] , or retortion triggered by ligand binding mechanisms [45] . In these models , the movement of CA domains is at the center of in trans phosphorylation of the homodimer of HK . Our data revealed that DSF stimulation resulted in conformational change associated with DHp-CA domain ( Fig 7A ) . In addition , DSF-sensor interaction released the juxtamembrane domain-mediated autoinhibition on the RpfC kinase activity . Among them , Leu172 and Ala178 are two critical amino acids since replacement of them resulted in constitutive activation of RpfC . These results support a view that DSF acts as an allosteric activator of RpfC by releasing the autoinhibition of its juxtamembrane region . Further investigation is necessary to obtain the high resolution structure of the DSF-RpfC complex to dissect the structural mechanism of RpfC activation , which is also meaningful to study the specificity in sensing different DSF-family signals . All bacterial strains and recombinant vectors used in this work are listed in S1 Table . Xanthomonas campestris pv . campestris ( Xcc ) 8004-derived strains and wildtype strain ( WT ) grew at 28°C in NYG medium ( tryptone 5 g L-1 , yeast extract 3 g L-1 , glycerol 20 g L-1 , pH 7 . 0 ) or 210 medium ( sucrose 5 g L-1 , casein enzymatic hydrolysate 8 g L-1 , yeast extract 4 g L-1 , K2HPO4 3 g L-1 , MgSO4·7H2O , 0 . 3 g L-1 , pH 7 . 0 ) . E . coli DH5α was used as the host for construction of all recombinant vectors . E . coli BL21 ( DE3 ) strain was used as the host for expressing recombinant proteins with pET30a vector ( Novagen , USA ) . Appropriate antibiotics were added when needed as following concentrations: kanamycin ( 50 μg ml-1 ) ; spectinomycin ( 150 μg ml-1 ) ; ampicillin ( 100 μg ml-1 ) and rifamycin ( 25 μg ml-1 ) . Xcc 8004 and E . coli electro-competent cells were prepared by extensively washing bacterial cells three times with ice-cold glycerol ( 10% ) . Transformation condition of both X . campestris pv . campestris and E . coli cells was set as 1 . 8 kV cm-1 , 25 μF and 200 Ω and conducted in a Bio-Rad Pulser XCell ( Bio-Rad , USA ) . HPLC purified diffusible signal factor ( DSF , CAS No . 677354-23-3 , purify > 90 . 0% ) was purchased from Sigma Aldrich ( USA ) and used in different concentrations as indicated in different experiments . If not specially mentioned , general molecular biology techniques , including PCR , DNA ligation , enzyme restriction , western blotting , etc , were according to the protocols in Molecular Cloning[46] . All in-frame deletion ( markerless ) mutants of rpfC and double mutant of rpfC-rpfF were constructed using suicide vector pK18mobsacB[47] by a homologous , double cross-over method . Briefly , the 5' and 3' genomic sequences of a targeted region were amplified using the primers listed in S2 Table , and correct PCR products were ligated into suicide vector pK18mobsacB . The recombinant pK18mobsacB vector was electroporated into competent cells of Xcc 8004 to generate single-crossover mutants by selection on NYG plates containing kanamycin . Afterwards , single-crossover mutants were cultured in NYG medium ( antibiotic-free ) for 1–2 hours and then grew on NYG plates containing 10% sucrose to select second-round homologous cross-overs . Correction of candidate bacterial mutants ( resistant to 10% sucrose but sensitive to kanamycin ) was verified by PCR and subsequent sequencing . To genetically complement the ΔrpfC mutant , a full-length rpfC gene was amplified using primers listed in S2 Table , ligated into the broad-host vector pHM1 [48] , and electroporated into E . coli DH5α to generate the recombinant vector . This vector was then extracted from E . coli DH5α and electroporated into the ΔrpfC or ΔrpfCΔrpfF mutant , in which transcription of full-length rpfC was under the control of a PlacZ promoter . Besides the initial residue Met1 , the N-terminal sensor region of RpfC contains 21 residues , with two of them being Ala . To conduct alanine-scanning mutagenesis , full-length rpfC coding sequence was amplified by PCR and inserted into a pGEM T-easy vector ( Promega , USA ) , and Easy Mutagenesis System ( TransGen Biotech , China ) was used to construct point mutations according to the manufactory’s manual . Coding sequences of 19 non-Ala residues were mutated into Ala , respectively , and Ala16 and Ala21 were mutated into Val , respectively . The point mutation was confirmed by sequencing . These inserts with corresponding point mutations were cut by restriction enzymes , purified , and ligated into broad host vector pHM1 [48] . Their expressions were under the control of a PlacZ promoter . Recombinant vectors were then electroporated into ΔrpfC or ΔrpfCΔrpfF mutants as needed . Primers used to create these mutants are listed in S2 Table . To construct a biosensor using expression level of engXcc , which is subject to the control of RpfC-RpfG , as a parameter to estimate DSF-RpfC interaction , 5′ promoter region of engXcc ( 254 bp ) was amplified by PCR , transcriptionally fused with a gusE gene to create PengXcc-GUS reporter insert ( with native gusE Shine-Dalgarno to drive protein translation ) . For different purposes , the reporter sequence was provided in trans or integrated into bacterial chromosome . For in trans complementation , this reporter sequence was cloned into a pHM2 vector ( no promoter upstream multiple cloning site ) , which was then electroporated into ΔrpfC or ΔrpfCΔrpfF mutants for GUS expression analysis . For alanine-scanning mutagenesis , the PengXcc-GUS reporter sequence was integrated into the chromosomal locus of engXcc by homologous , double cross-over via a recombinant pK18mobsacB vector ( pKengXcc-GUS ) . For GUS activity assay , bacterial strains were cultured and adjusted to OD600 = 0 . 1 , then grew without DSF or with 10 μM DSF for about 9 hours . Cells were collected by centrifugation ( 12 , 000 g , 10 min at 4°C ) , and immediately frozen in liquid nitrogen . GUS extraction buffer ( 50 mM sodium phosphate [pH 7 . 0] , 5 mM DTT , 1 mM EDTA [pH 8 . 0] ) was added to resuspend the cells and then these bacterial cells were lysed by sonication . The mixture was centrifuged ( 12 , 000 g , 10 min at 4°C ) and the supernatant was used for GUS activity assay . Levels of GUS expression were quantified by its activity using 4-methylumbelliferyl ß-D-glucuronide ( 4-MUG , purchased from Sigma Aldrich , USA ) as a substrate . A standard curve was prepared by diluting the 4-MU stock solution . The fluorescence of samples and standard curve solutions were measured using an excitation wave-length of 360 nm and an emission wave-length of 460 nm . Protein concentrations of supernatants were measured using Coomassie brilliant blue G-250 Protein Assay ( Bio-Rad , USA ) with BSA as a standard . For each experiment , at least three independent repeats were conducted for calculating the parameters . Plant inoculation and virulence assay were conducted as previously described[49] . In brief , six-week-old cabbage cultivar Brassica oleraceae cv . Jingfeng 1 was used as host plants . WT strain of Xcc and sterile 10 mM MgCl2 were used as positive and negative controls , respectively . All bacterial strains were cultured overnight in NYG medium containing appropriate antibiotics . Cells were collected , washed by 10 mM MgCl2 , and the concentrations were adjusted to OD600 = 0 . 4 before inoculating into plant leaves using sterile scissors . After inoculation , the plants were kept in a greenhouse at 25°C–30°C and relative humidity >80% . Lesion length was scored 10 days after inoculation , and virulence level was scored semi-quantitatively as follows: 0 , no visible effect; 1 , limited chlorosis around the cut site; 2 , chlorosis extending from the cut site; 3 , blackened leaf veins , death , and drying of tissue within the chlorotic area; 4 , extensive vein blackening , death , and drying of tissue . Assay of extracellular polysaccharides production ( EPS ) was conducted according to previous study[50] . Bacterial strains were cultured at 28°C in NYG medium until OD600 = 0 . 4 . If necessary , DSF with appropriate concentration was added and bacteria were cultured for 75–96 hours before EPS production measurement . Quantification of biofilm development was conducted by classic method of crystal violet staining and according to previous study [51] . Bacterial strains were grown at 28°C in NYG medium until OD600 = 1 . 0 , and 200 μl culture were inoculated into a 96 well plate ( Costar , USA ) , cultured for 12 hour before quantification . To test the effect of DSF on the formation of biofilm , bacterial strains were cultured and adjusted to OD600 = 0 . 1 . Then , those bacteria strains with appropriate concentration of DSF were grown for about 9h at 28°C and 200 μl culture were inoculated into 96 well plate as mentioned above . Estimate of extracellular protease ( EXP ) were conducted on NYG-milk plate as described previously [52] . If needed , 3 . 5 μL DSF ( 30 μM or 10 μM ) was added near the bacterial colony . C-terminal His6-tagged recombinant proteins were expressed by constructing corresponding recombinant pET30a ( Novagen ) vectors that were electroporated into E . coli BL21 ( DE3 ) strain ( S1 Table ) . Primers used to generate these constructs are listed in S2 Table . His6-tagged proteins were expressed and purified using affinity chromatography with Ni-NTA agarose beads ( Novagen , USA ) , according to manufacturer′s instructions . Purified proteins were concentrated using Centricon YM-10 columns ( Millipore ) and the elute buffer was changed into storage buffer for further use ( 50 mM Tris-HCl , pH 8 . 0 , 0 . 5 mM EDTA , 50 mM NaCl and 5% glycerol ) . Preparation of inverted membrane vesicles ( IMV ) containing full-length RpfC was according to the protocol of our previous study with minor modification[53] . Briefly , after sonication , cell debris of E . coli BL21 ( DE3 ) was abandoned by 6 , 000 g and the membrane containing full-length RpfC in supernatant was collected by ultracentrifugation at 60 , 000 g at 4°C for 60 min . After ultracentrifugation , the membrane was washed in high-salt buffer ( 20 mM sodium phosphate , pH 7 . 0; 2 M KCl; 10% glycerol; 5 mM DTT; 1 mM PMSF ) twice . Finally , the membranes were resuspended in 0 . 5 ml storage buffer ( 20 mM Tris-HCl , pH 7 . 5; 10% glycerol ) for autokinase assays . Liposomes were reconstituted as described by previous study[54] . Briefly , IMV of RpfC was obtained as above-mentioned , and dissolved in suspension buffer ( 20 mM phosphate , 500 mM NaCl , 20 mM imidazole , pH 7 . 4 ) to an approximate concentration of 10 mg ml-1 for preparation of liposomes . 900 μl IMV suspension with 100 μl 10% n-Dodecyl-ß-D-maltoside ( DDM ) were mixed by “end-over-end” mixing for 45 min at 4°C . The supernatant was collected by centrifugation 50 , 000 g for 30 min and purified by Ni-affinity chromatography . The Ni-NTA beads ( Novagen , USA ) were pre-equilibrated with 5 volumes of binding buffer ( 20 mM phosphate , 500 mM NaCl , 20 mM imidazole , pH 7 . 4 , 0 . 1% DDM ) . Then the solubilized IMV and pre-equilibrated beads were mixed and incubated at 4°C for 30 min . After the mixing , the supernant was removed and the deposition was washed with binding buffer until the absorbance at OD280 nm returned to base line . Finally , 100 μl of elution buffer ( 20 mM phosphate , 500 mM NaCl , 250 mM imidazole , pH 7 . 4 , 0 . 1% DDM ) was added to elute the purified RpfC-His6 . For embedding RpfC by liposome , 10 mg liposomes ( Avanti Polar Lipids ) were dissolved into 1 ml buffer ( 50 mM Tris-HCl pH7 . 5 , 10% glycerol , 0 . 47% Triton-100 ) . Then purified RpfC-His6 in elution buffer was added . The mixture was stirred at 4°C for 45 min . The final ratio of phospholipids to protein was about 10:1 ( w/w ) . Bio-Beads ( beads: detergent = 10:1 , Bio-Rad , USA ) were added to remove the detergent and the solution was stirred gently at 4°C overnight . Residual detergent was removed completely by addition of Bio-Beads after further incubation for 2 hours . The Bio-Beads were pipetted off and liposomes containing RpfC-His6 were gathered by centrifugation with 200 , 000 g , 4°C for 30 min . The RpfC liposome was resuspended in final buffer ( 20 mM Tris-HCl , pH 7 . 5 , 10% glycerol ) and stored at 80°C until used . in vitro phosphorylation assay was conducted as described in our previous study [53] . For autophosphoylation , RpfC liposomes or IMV were incubated with 100 μM ATP containing 10 μCi [γ-32P]ATP ( PerkinElmer , USA ) in autophosphorylation buffer ( 50 mM Tris-HCl , pH 7 . 8 , 2 mM DTT , 25 mM NaCl , 25 mM KCl , 5 mM MgCl2 ) for indicated time ( 28°C ) . If necessary , DSF was added into the mixture 20 min before addition of ATP . The reaction was stopped with 6 × SDS-PAGE loading buffer . The phosphorylated proteins were separated by 12% SDS-PAGE . After SDS-PAGE electrophoresis , gels were separated from back glass plate and placed in a Ziploc bag and exposed to a phosphor screen for 1 hour . The screen was scanned with a PhosphorImage system ( Amersham Biosciences , USA ) at 50 μM solution . If necessary , signal intensity was measured by Quantity One software ( Bio-Rad ) . RpfC sensor fused with a GST tag was obtained and purified according to the GST Gene Fusion System Handbook ( Amersham Biosciences ) with GST Resin ( TRANS ) . In order to acquire sensor peptide with GST tag cleaved off , on-column cleavage procedure was conducted . In brief , lysate of recombinant E . coli BL21 ( DE3 ) strain was mixed with pre-equilibrated GST Resin with PBS buffer for ten minutes before loading into column . The column was washed by PBS buffer and resuspened with PreScission buffer ( 50 mM Tris-HCl , pH 8 . 5 , 150 mM NaCl ) . Following the injection of 2 units PreScission Protease ( GenScript ) , the column was sealed and placed on a rotator at 4°C . After 10 h of digestion , the flow fractions were collected , which contains preliminary sensor peptide with the GST tag being moved . If necessary , the sensor peptide was purified again by the GST Resin to get rid of uncleaved sensor-GST fusion protein . Eventually , purified sensor peptide was obtained by using size exclusion chromatography with column Superdex 75 10/300 GL ( GE Healthcare ) , stored under -80°C before use . DSF was mixed with purified proteins or liposomes of RpfC to a final concentration of 0 μM , 5 μM or 10 μM respectively in the reaction buffer ( 50 mM Tris-HCl pH 7 . 8 , 25 mM NaCl , 100 mM KCl ) . The protein concentration was 0 . 5–1 μg/μl . The TSA was performed by a Prometheus NT . 48 nanoDSF device with a temperature gradient of 20–95°C , 1°C /min . Unfolding transition points were determined according to the changes of intrinsic tryptophan fluorescence at 330 nm , 350 nm . The ratio of fluorescence and the melting temperature ( Tm ) were calculated by the NT Melting Control software ( NanoTemper Technologies ) . Binding reactions of RpfC to DSF was measured by microscale thermophoresis in a Monolith NT . Label Free ( Nano Temper Technologies GMBH , Germany ) instrument which detects changes in size , charge and conformation induced by binding . RpfC liposomes were collected with centrifugation of 200 , 000 g for 40 min and resuspended in MST buffer ( 50 mM Tris-HCl pH 7 . 8 , 150 mM NaCl , 10 mM MgCl2 , 0 . 05% Tween-20 ) to an approximate concentration of 0 . 1 μM . A range of concentration of DSF ( range from 0 . 06 μM to 2 mM ) in assay buffer ( 50 mM Tris-HCl pH 7 . 8 , 150 mM NaCl , 10 mM MgCl2 , 0 . 05% Tween-20 , 5% methanol ) was incubated with RpfC liposomes ( 1:1 , v/v ) for 10 minutes . The sample was loaded into the NT . Label Free standard capillaries and measured with 20% LED power and 40% MST power . Purified sensor protein was dissolved in reaction buffer ( 50 mM Tris-HCl pH 8 . 5 , 150 mM NaCl , 0 . 1% Tween-20 ) to a final concentration as 8 μM . Dilute DSF from 0 . 0122 μM to 25 μM in buffer ( 50 mM Tris-HCl pH 8 . 5 , 150 mM NaCl , 0 . 25‰ methanol ) . Different concentrations of DSF and sensor protein ( 1:1 , v/v ) were mixed and loaded into NT . Label Free standard capillaries . The label free MST assay was performed with 20% LED power and 40% MST power . KD Fit function of the Nano Temper Analysis Software Version 1 . 5 . 41 was used to fit curve and calculate the value of dissociation constant ( Kd ) . The limited proteolysis experiments were performed at 0°C with 1 . 4 μg RpfC liposomes in a reaction buffer containing 50 mM Tris-HCl , pH 8 . 0 , 100 mM NaCl , 2 mM DTT and 1 . 13 mM AMP-PNP . Trypsin was added to a final concentration of 0 . 018 μg μl-1 to degrade RpfC liposome . Aliquot was removed at indicated time and the reaction was stopped by 5 × SDS loading buffer ( 250 mM Tris-HCl pH 6 . 8 , 10% ( w/v ) SDS , 0 . 5% ( w/v ) bromophenol blue , 50% ( v/v ) glycerol , 25 mM PMSF ) . Samples were separated by 12% SDS-PAGE gel and protein bands were detected by silver staining . The sequence of the different peptide fragments were determined by a nanoLC-MS/MS with Orbitrap Fusion system ( Thermo scientific , USA ) . To determine if DSF has impact on RpfC sensor conformation change , CD analysis was carried out on a Chirascan CD Spectrometer ( Applied Photophysics , UK ) , with 10 mm pathlength and 1 nm bandwidth . Sensor protein with GST tag cleaved off was dissolved in buffer ( 50 mM Tris-HCl pH 8 . 5 , 150 mM NaCl ) to 60 μM . Dilute DSF with Sensor protein to a series of concentration ( 0 μM , 100 μM and 500 μM ) . CD wavelength scans were collected between 200 nm-260 nm . The spectra data were analyzed on the http://dichroweb . cryst . bbk . ac . uk website with Contin-LL method [55] .
Besides roles in nutrition , lipids also function as important signals in the regulation of prokaryotic and eukaryotic cells . In bacteria , fatty acids are part of the language of cell-cell communication known as quorum sensing for a decade . However , how bacteria detect these signals and regulate virulence remains elusive . Here , we provide multiple evidences to show that a full-length receptor histidine kinase , RpfC , directly binds to a fatty acid-based signal factor using a short sensor region . This binding event stimulates RpfC autokinase activity by triggering conformational change in its catalytic region , which is critical in regulating bacterial quorum sensing and virulence . Our results confirm a long-outstanding assumption in cell signaling of phytobacteria , and provide a technical pipeline to analyze fatty acid-receptor interactions .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "methods" ]
[ "bacteriology", "biofilms", "exopolysaccharides", "vesicles", "chemical", "compounds", "xanthomonas", "campestris", "microbiology", "organic", "compounds", "mutation", "plant", "science", "amino", "acid", "substitution", "amino", "acids", "plant", "pathology", "cellular", "structures", "and", "organelles", "bacteria", "proteins", "chemistry", "liposomes", "plant", "bacterial", "pathogens", "recombinant", "proteins", "biochemistry", "bacterial", "biofilms", "point", "mutation", "polysaccharides", "cell", "biology", "organic", "chemistry", "plant", "pathogens", "genetics", "biology", "and", "life", "sciences", "physical", "sciences", "xanthomonas", "glycobiology", "organisms" ]
2017
Fatty acid DSF binds and allosterically activates histidine kinase RpfC of phytopathogenic bacterium Xanthomonas campestris pv. campestris to regulate quorum-sensing and virulence
Patterned expression of many developmental genes is specified by transcription factor gene expression , but is thought to be refined by chromatin-mediated repression . Regulatory DNA sequences called Polycomb Response Elements ( PREs ) are required to repress some developmental target genes , and are widespread in genomes , suggesting that they broadly affect developmental programs . While PREs in transgenes can nucleate trimethylation on lysine 27 of the histone H3 tail ( H3K27me3 ) , none have been demonstrated to be necessary at endogenous chromatin domains . This failure is thought to be due to the fact that most endogenous H3K27me3 domains contain many PREs , and individual PREs may be redundant . In contrast to these ideas , we show here that PREs near the wing selector gene vestigial have distinctive roles at their endogenous locus , even though both PREs are repressors in transgenes . First , a PRE near the promoter is required for vestigial activation and not for repression . Second , only the distal PRE contributes to H3K27me3 , but even removal of both PREs does not eliminate H3K27me3 across the vestigial domain . Thus , endogenous chromatin domains appear to be intrinsically marked by H3K27me3 , and PREs appear required to enhance this chromatin modification to high levels at inactive genes . The patterns of chromatin histone modifications differ between cell types , reflecting the activity of genes for developmental programs . Tri-methylation of the lysine-27 residue of histone H3 ( H3K27me3 ) typically marks extended chromatin domains , leading to chromatin compaction and epigenetic gene silencing that is maintained as cells differentiate [1 , 2] . Histone methylation is thought to be initiated at discrete regulatory elements called Polycomb Response Elements ( PREs ) within domains . These elements bind multiple DNA-binding factors , recruiting the PRC1 and PRC2 complexes , including the Polycomb chromatin factor and the E ( z ) histone methyltransferase , respectively [3 , 4] . Transgenes carrying PREs are sufficient to silence reporter genes and to nucleate new H3K27me3 domains [5–7] . However , the function of PREs in their endogenous domains is less clear . Deletion of PREs from the homeobox gene cluster BX-C have limited defects in gene silencing [8–10] , but no reduction of histone methylation of this domain . While multiple PREs within the BX-C domain may be redundant , deletion of all mapped PREs near the invected and engrailed genes have no effect on methylation of the locus , and it remains unknown how histone methylation is maintained [11] . Genomic mapping has identified regions where both the PRC1 and PRC2 complexes colocalize , and regions where each complex is found separately . Only about one-half of all Polycomb binding sites are within H3K27me3 domains , and thousands of additional sites are located near the promoters of active genes [12 , 13] , where they may modulate gene expression by holding RNAPII at paused promoters . Here , we characterize the in vivo roles of two PREs near the vestigial gene . While these two PREs are silencers in transgene assays , targeted mutations reveal that the promoter PRE is required for full gene expression . Using a new efficient method for genomic mapping of chromatin factors , we demonstrate that methylation across the domain remains in the absence of both PREs . Our results reveal that PREs stimulate but are not necessary for domain methylation . To profile chromatin domains in different tissues , we used a chromatin mapping strategy that tethers micrococcal nuclease at factor binding sites . In the CUT&RUN procedure [14] , unfixed cells are soaked with a factor-specific antibody , which binds to chromatin . Next , a protein-A-micrococcal nuclease ( pA-MNase ) fusion protein is soaked in , binding to the chromatin-bound antibody . Activation of the tethered MNase by adding calcium then cleaves exposed DNA around the binding sites of the targeted factor . Sequencing of the cleaved DNA fragments thus maps the location of the targeted chromatin protein . CUT&RUN obviates the need to work with chromatin preparations or to optimize affinity recovery of chromatin particles , and works efficiently with small numbers of cells [15 , 16] . To implement CUT&RUN for tissue samples , we simply dissected brains and wing imaginal discs from ten larvae , lightly permeabilized the whole tissues with digitonin , and sequentially incubated the tissues with antibody to H3K27me3 and then with pA-MNase . MNase was then activated and finally the cleaved DNA was isolated , subjected to Illumina paired-end sequencing , and mapped to the Drosophila dm6 genome assembly . We similarly mapped the Polycomb protein , which binds at Polycomb Response Elements ( PREs ) . H3K27me3 domains in larval tissues have been previously mapped by Chromatin Immunoprecipitation [17] . Profiles of H3K27me3 distribution generated by CUT&RUN using substantially less material were similar ( Pearson’s r = 0 . 94 ) . Both methods reveal changes in chromatin methylation that correspond to tissue-specific changes in gene expression . For example , the ANTENNAPEDIA-COMPLEX ( ANTP-C ) cluster of homeobox genes are encompassed in a H3K27me3 domain in larval brains , consistent with the predominant silencing of this cluster in this tissue [18] ( Fig 1A ) . In contrast , in wing imaginal discs where Antp is transcribed , chromatin over most of this gene is depleted for H3K27me3 . Interestingly , some histone methylation remains across the 3’ exons of Antp , which indicate that a shorter isoform of the Antp gene may be transcribed in this tissue . Notably , histone methylation is not completely eliminated from the transcribed Antp gene ( Fig 1A ) . To quantify changes in chromatin landscapes between tissues , we measured the read count coverage at 125 chromatin domains ( listed in S3 Table ) with high H3K27me3 in larval brain and wing disc samples . Most chromatin domains are similarly methylated between these tissues , but a small number of domains have lower read counts for H3K27me3 in wing discs compared to larval brains ( Fig 1B ) . One group of domains have low levels of histone methylation in larval brains and lose methylation in wing discs , however , genes in these domains are not expressed in either tissue . A second set of domains encompass genes that are expressed in wing discs but not in brains , including apterous ( ap ) , nubbin ( nub ) , vestigial ( vg ) , and Drop ( Dr ) ( Fig 1B , blue ) , and histone methylation across these domains is lower in wing discs . Histone methylation is not eliminated , as 25–50% of H3K27me3 levels remains even when domain genes are transcribed , and this is noticeably greater than background levels at random regions outside of domains ( red in Fig 1B , Fig 1C ) . Thus , activation of these genes is accompanied by reduction–but not loss–of the H3K27me3 modification . We focused on the vg gene locus as a simple model . The vg gene is required for wing determination , is the only gene in a 32 kb H3K27me3 chromatin domain ( Fig 1A ) . The vg gene is inactive in brain tissues , and is expressed only in the pouch of wing imaginal discs . This domain is heavily methylated in larval brains , but reduced to ~50% across the domain in wing discs ( Fig 1A and 1C ) . This is consistent with loss of H3K27me3 when the vg gene is activated , but wing discs are a mixture of cells with and without vg expression . We therefore isolated vg-expressing cells to profile the chromatin status of the active gene . We made a transgene construct containing the vg Quadrant enhancer [19] , and the GAL4 transcriptional activator , and used this to drive expression of GFP in the wing pouch . We then used FACS to isolate GFP-positive cells and profiled these cells by CUT&RUN . As expected , profiles of these cells show reduced histone methylation across wing-specific genes like Antp ( Fig 1A , Supplementary Information ) . We found that H3K27me3 across the vg domain is reduced to ~20% of it’s levels in non-expressing cells , but remains ~4-fold more methylated than background levels across the genome ( Fig 1A and 1C ) . Thus , while gene activation is associated with elimination of the H3K27me3 modification , a small amount of methylation remains across many activated domains . Two potential PREs within the vg domain have been identified by sequence motifs [20] and by chromatin profiling [12 , 13 , 21] . The first region , which we term the proximal PRE ( pPRE ) , is located 300 bp downstream of the mapped Transcriptional Start Site ( TSS ) of the vg gene , which was previously mapped by primer extension [22] . The second distal PRE ( dPRE ) is located ~25 kb downstream ( Fig 2A ) . We used CUT&RUN to map the Polycomb chromatin protein in larval tissues , and found that this protein is bound at both of these PREs in larval brains ( Fig 1A ) . Surprisingly , Polycomb binding is detectable at both PREs in both wing discs and even in FACS-isolated vg-expressing cells ( Fig 1A ) . We quantified the amount of Polycomb and found that Polycomb is retained to similar levels at the pPRE of the active vg gene , and at reduced levels at the vg dPRE ( Fig 1D ) . These levels are clearly above background levels at a non-target promoter or an insulator element . We also observed retention of Polycomb at the promoters of the Dr and nub genes , which are active in wing pouch cells ( Fig 1D ) . The promoter of the ap gene which changes methylation between brain and wing disc samples has very low levels of Polycomb in repressed brain samples , and is indistinguishable from background in wing pouch cells ( Fig 1D ) . These results indicate that in some cases Polycomb remains present at de-repressed genes . To test the silencing effects of these PREs , we created transgenes including these two regions , and integrated these at the same landing site in the Drosophila genome . We then tested if these transgenes induce pairing-sensitive silencing ( PSS ) , a diagnostic feature of PREs where a transgene reporter gene is silenced in homozygous animals [23] . Previous studies showed that a dPRE-containing transgene will cause PSS [24] , and a similar transgene with the 1 kb dPRE region integrated at the landing site also causes PSS ( Fig 3 ) . We found that a transgene with the 300 bp pPRE region also shows strong PSS , demonstrating that the pPRE is also a silencing element . These PREs can interact with each other and cause silencing , as animals heterozygous for a pPRE transgene in one landing site and a dPRE transgene on the homolog also show PSS ( Fig 3 ) . CUT&RUN for Polycomb defined a 200 bp segment where Polycomb binds near the vg promoter ( Fig 2A ) . We used high-resolution mapping by native ChIP in Drosophila S2 cells to precisely define binding sites for three juxtaposed Polycomb-bound sites , one of which is also bound by the Pleiohomeotic ( PHO ) transcription factor ( S1 Fig ) . Deletion of one of these peaks from the pPRE transgene alleviates PSS ( Fig 3 ) , thus this sequence is required for reporter silencing . We analyzed the dPRE similarly . We found that while a transgene including the dPRE induces PSS , deletion of the Polycomb-bound site within the dPRE alleviates this silencing ( Fig 3 ) . We conclude that the Polycomb-bound sites in both the pPRE and dPRE elements are required for transgene silencing . To measure silencing at the endogenous vg locus , we integrated reporter genes by gene-targeting near the pPRE and near the dPRE . The promoter of the engrailed gene is active in the posterior half of the wing imaginal disc [25] , including part of the expression domain of vg in the wing pouch . We used an engrailed-GAL4 ( en-GAL4 ) transgene [26] to drive expression of GAL4-dependent UAS-YFP and UAS-RFP reporters in the wing disc . Control reporter gene insertions produce RFP and YFP throughout the posterior half of the wing disc ( Fig 4A ) . In contrast , UAS-YFP reporters inserted in the vg domain are silenced throughout most of the wing disc , with reduced expression only within the posterior part of the wing pouch ( Fig 4A ) . Insertions near the dPRE show similar reduced expression in the wing pouch and silencing in the rest of the wing disc . Thus the vg domain appears to be packaged in repressed chromatin in most of the wing disc , but in derepressed chromatin in the wing pouch ( Fig 2A ) . We confirmed that silencing in the wing disc is mediated by chromatin by expressing a dominant-negative H3 . 3K27M mutant histone to reduce chromatin levels of H3K27me3 [27] in the posterior half of the wing disc . Indeed , expression of the mutant histone derepressed the GFP reporter gene throughout the posterior half of the wing disc ( Fig 4A ) . This demonstrates that the vg domain is in two chromatin states in the wing imaginal disc: a silenced configuration , and a derepressed configuration in wing pouch cells where vg is normally expressed . However , expression of the mutant histone does not derepress expression of the vg gene itself ( Fig 4B ) . Thus , the silenced configuration appears to only affect the inserted reporter gene . To define the function of the PREs near the vg gene , we deleted each element from the endogenous locus ( see Methods ) . Precise breakpoints for each of the recovered deletions were determined by Sanger sequencing , and tested against each other and against previously characterized vg alleles ( Fig 2B and 2C; S1 Table; S2 Table ) . We used the vgnw allele—which deletes part of the coding region of the gene—as a null allele . Ectopic expression of vg converts legs and eyes into wing-like structures [28]; thus deletion of silencing elements should derepress the vg gene and transform non-wing tissues . However , we observed no such transformations in animals homozygous for deletion of the pPRE or of the dPRE . Deletion of both PREs from the vg domain did not transform non-wing tissues . Thus , there appears to be no role for Polycomb silencing in limiting vg expression . Surprisingly , we found that deletions of the pPRE reduce expression of the vg gene . Wing development is sensitive to the amount of Vg protein , and reduction in vg expression results in progressive notching and deletion of the wing . We found that animals carrying the vgCL1 or vg13A deletions are viable , but have severely reduced wings ( Fig 5A ) . A smaller deletion ( vgCZ ) which deletes only one of the Polycomb-binding sites within the pPRE has a more limited effect . The vgCZ allele gives no phenotype as a homozygote , but in combination with a null allele adults have notched wings which characteristic of weak vg alleles . We also recovered a similarly weak allele ( vgCL2C ) that is a single-base pair C-to-T substitution in the pPRE . This substitution lies precisely at the center of the major Polycomb-bound site , in a sequence similar to consensus motifs for Sp1 transcription factors which direct Polycomb binding [29] . Each of these mutations are recessive and viable alleles , and thus are distinct from deletions of the vg promoter , which are recessive lethal alleles [22] . Thus , the pPRE appears to positively contribute to vg expression . In contrast , animals with a deletion of the dPRE ( the vgR5 allele ) are viable with normal wings , either as a homozygote ( Fig 5A ) . Animals lacking both PREs are wingless like the pPRE single mutant ( Fig 5A ) . The adult wing blade differentiates from the pouch of the larval wing imaginal disc where vg is expressed , and we therefore imaged Vestigial protein in wing discs of larvae . Animals carrying the pPRE deletion vgCL1 have small wing discs where the wing pouch is reduced , no Vestigial protein is detectable , and the central stripe of Wingless ( Wg ) signalling ligand that marks the edge of the future wing blade is absent ( Fig 5B ) . The smaller pPRE deletion vgCZ and the point mutant vgCL2C have more limited effects: Vg is produced , but with occasional gaps in wing discs ( Fig 5B , yellow arrowheads ) . These gaps are often associated with gaps in the central Wg stripe , consistent with the notching of adult wings . In contrast , the dPRE deletion vgR5 has no extra staining or defects of Vg , and the central Wg stripe is continuous . Finally , animals carrying both the pPRE and dPRE double deletions have small wing discs similar to the single pPRE deletion , with no Vg protein detected in wing discs ( Fig 5B ) . We conclude that the pPRE is required for expression of the vg gene in wing discs , but the dPRE is not . We then tested if Polycomb factors are required at the pPRE for vg expression . The vgCL2C basepair substitution is a recessive allele , and animals heterozygous for this mutation have no phenotype . However , in combination with a heterozygous Pc3 allele , vgCL2C/+ animals have deformed wings ( Fig 6A ) . The wings of vgCL2C/vgnw; Pc3/+ animals have more enhanced wing notching , while control Pc3/+ siblings have normal wings ( Fig 6A; Table 1 ) . Wing discs from vgCL2C/vgnw; Pc3/+ larvae show gaps in Vg staining and gaps in the central Wg stripe , consistent with the adult notching ( Fig 6B ) . Similarly , mutations in the RING1b homolog Sce enhances the phenotype of vgCL2C/vgnw animals ( Fig 6A ) . These effects suggest that PRC1 components bound at the promoter in active cells ( Fig 1D ) positively influence vg expression . Finally , combining Polycomb mutations with the vgR5 deletion show no wing defects ( Table 1 ) , demonstrating that the genetic interaction between Polycomb and the pPRE allele is specific . Regulatory structures are present in the 5’UTRs of some transcripts . However , it is unlikely that the mutations we created affect an mRNA function , because they coincide precisely with the chromatin features of the pPRE . Further , the single-base pair mutation vgCL2C is enhanced by Polycomb mutations , supporting the idea that it affects the function of the chromatin element , not an mRNA function . It is unusual for PRC1 to be implicated in transcriptional activity , but there are examples . In one case in the mouse midbrain , Polycomb is required to bring enhancers to the meis2 gene promoter before meis2 is expressed , but then is not required after induction [30] . To test if the vg pPRE is similarly required before activation of the vg gene or if the pPRE is required in cells expressing vg , we generated cells homozygous for a pPRE deletion from heterozygous cells by FLP recombinase-mediated mitotic recombination at different times in development [31] . The pPRE/+ heterozygous animals have no wing defects , but FLP expression produces animals with a range of defects in the wing blade , ranging from notches in the wing margin to complete loss of one wing ( Fig 6C ) , implying that pPRE mutant clones lose vg expression whenever they are induced . Together , these results indicate that the pPRE is continually required to maintain expression of the vg gene . In transgenes , a PRE is required to nucleate and maintain a Polycomb-regulated domain by recruiting PRC1 and PRC2 complexes [6 , 7] . We tested if histone methylation of the vg domain depends on the pPRE or on the dPRE . We profiled the chromatin of wildtype and PRE deletion mutants in larval brains , where the vg gene is not active . H3K27me3 levels are high across the vg domain in wildtype larval brains ( Fig 7 ) . However , there is no reduction in H3K27me3 across the vg domain In animals lacking the pPRE . In contrast , histone methylation is reduced to ~45% of wildtype levels in animals lacking the dPRE ( Fig 7 ) . Finally , histone methylation is reduced to ~20% wildtype levels when both PREs deleted . These results indicate that the dPRE is predominantly responsible for histone methylation of the vg domain , although the pPRE can also contribute to domain methylation . Notably , the residual methylation across the vg domain when both PREs are deleted remains higher than background levels of H3K27me3 in the genome . The amount of residual methylation is similar to that in cells where the vg gene is active , suggesting that this is the minimal level of H3K27 methylation of this domain . It is possible that minor or cryptic PREs in a domain may direct histone methylation when major PREs are deleted [11] . We therefore profiled Polycomb binding in larval brains from wildtype and PRE deletion mutants , normalizing landscapes to peak heights in the ANTP-C domain ( Fig 7A , bottom ) . Deletion of either the pPRE or the dPRE eliminates only its peak of Polycomb binding . We observe no Polycomb binding when both PREs are deleted , suggesting that there are no alternative or cryptic PREs in the vg domain . Thus , the low level of histone methylation across the vg domain appears to be independent of Polycomb binding sites . The developmental gene vestigial is contained within chromatin that has the features of a Polycomb chromatin domain , being marked by the H3K27me3 histone modification and silencing inserted reporter transgenes . While the two PREs from this domain both act as silencers in transgenes , our results show that they have distinct roles at the endogenous locus . One PRE is primarily responsible for histone methylation of the domain , but has no effect on silencing or expression of the vg gene . Previous studies have also suggested that PREs differ in their effects . For example , one of two PREs at the dachshund locus also directs methylation of its domain [35] . In contrast , deletions of the known PREs of the engrailed locus had no effect on domain methylation [11] . Cryptic PREs or non-coding RNAs have been proposed to direct histone methylation in these situations . However , there we found that no new PREs appear in the vg domain , and while non-coding RNAs have been identified near one of the PREs in the vg domain [36] , these RNAs are deleted in our mutants . Thus , the vg domain appears to have a low level of undirected H3K27-methylation . Such domain methylation might be directed by other histone modifications . Active chromatin regions are often marked by H3K27-acetylation , which is antagonistic to H3K27 methylation [37] . Further , in Drosophila cells , the H3K27 methyltransferase E ( z ) acts globally and dimethylates ~50% of all nucleosomes [38] . Perhaps sequence features of some regions predispose unacetylated regions to accumulate H3K27-trimethylation , and these regions can then reach high levels of H3K27me when PREs are active . The second PRE in the vg domain is distinct; while it has little effect on domain methylation , it is required for normal vg expression . PRE localization near promoters is a common feature of the Drosophila genome [12 , 39] , and are well-positioned to regulate gene activity . Polycomb can silence gene expression by inhibiting transition of RNA polymerase II ( RNAPII ) to its elongating form , and this is a major step for controlling the expression of developmental genes [40 , 41] . However , ~1000 active promoters in the Drosophila genome are also bound by Polycomb [12] . While this binding has been suggested to reduce transcriptional output , loss of Polycomb results in both loss of silencing at some genes and decreased transcription at others [39] . In genome-wide studies it has been difficult to determine if downregulation is due to pleiotropic effects or a requirement for Polycomb at some genes , The vg gene is the first example of a promoter that requires a PRE for expression during development . Positive effects may be mediated by PREs looping together enhancers and promoters [30] as well as silencing elements . The effects of a specific PRE may then depend on what regulatory elements it brings to a target promoter . Indeed , PREs have been demonstrated to switch between silencing and activating states [42] , and in mammals variant Polycomb complexes have been described that activate developmental genes [43 , 44] . Our observations that the vg pPRE can both silence a reporter gene and promote expression of the endogenous gene suggests that promoters differ in their interactions with PREs , and this may be critical to integrate Polycomb regulation with developmentally-programmed enhancers . All crosses were performed at 25°C . Transgenes , mutations and chromosomal rearrangements not detailed here are described in Flybase ( http://www . flybase . org ) . The vgnw allele is a deletion of the last two exons of the vg transcript , and so we used this as a standard null allele . New alleles of vg produced in this study are described in S1 Table .
Eukaryotic genes are packaged in chromatin , and their transcription relies on activators that recruit RNA polymerases and on repressive factors . In multicellular organisms , cell types have distinct patterns of gene expression , and these patterns are controlled by by the expression of cell-type-specific transcription factors and by modulating chromatin structure . The Polycomb system is one major system for the chromatin-mediated silencing of developmental gene expression , where a histone methyltransferase marks extended chromatin domains with trimethylation of lysine-27 of the histone H3 tail ( H3K27me3 ) and forms repressed chromatin . In Drosophila , repressive regulatory elements called Polycomb Response Elements ( PREs ) are thought to nucleate histone methyltransferase binding which then spreads across these domains . In this study , we demonstrate that two PREs near the developmental vestigial gene have distinct and separable effects on gene activation and chromatin structure . Both PREs are functional repressors in transgenes , but the PRE located near the vestigial promoter is required for gene transcription . This PRE has no effect on histone methylation of the domain . The second PRE located in the middle of the chromatin domain is required for high-level H3K27me3 of the domain , but this methylation is not required to refine vestigial gene expression . Significant chromatin methylation remains when both PREs are deleted . Our findings imply that PREs near promoters may play activating roles in gene expression in the Drosophila genome . We suggest that some domains of H3K27me3 may have little consequence for correctly patterning gene expression .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[ "invertebrates", "gene", "regulation", "dna-binding", "proteins", "invertebrate", "genomics", "animals", "animal", "models", "developmental", "biology", "drosophila", "melanogaster", "model", "organisms", "methylation", "experimental", "organism", "systems", "epigenetics", "morphogenesis", "chromatin", "drosophila", "research", "and", "analysis", "methods", "chromosome", "biology", "proteins", "animal", "studies", "gene", "expression", "chemistry", "histones", "imaginal", "discs", "animal", "genomics", "insects", "arthropoda", "biochemistry", "eukaryota", "cell", "biology", "genetics", "biology", "and", "life", "sciences", "chemical", "reactions", "physical", "sciences", "genomics", "organisms", "silencer", "elements" ]
2019
Separate Polycomb Response Elements control chromatin state and activation of the vestigial gene
Hepatitis B virus ( HBV ) replicates its DNA genome through reverse transcription of a viral RNA pregenome . We report herein that the interferon ( IFN ) stimulated exoribonuclease gene of 20 KD ( ISG20 ) inhibits HBV replication through degradation of HBV RNA . ISG20 expression was observed at basal level and was highly upregulated upon IFN treatment in hepatocytes , and knock down of ISG20 resulted in elevation of HBV replication and attenuation of IFN-mediated antiviral effect . The sequence element conferring the susceptibility of HBV RNA to ISG20-mediated RNA degradation was mapped at the HBV RNA terminal redundant region containing epsilon ( ε ) stem-loop . Furthermore , ISG20-induced HBV RNA degradation relies on its ribonuclease activity , as the enzymatic inactive form ISG20D94G was unable to promote HBV RNA decay . Interestingly , ISG20D94G retained antiviral activity against HBV DNA replication by preventing pgRNA encapsidation , resulting from a consequence of ISG20-ε interaction . This interaction was further characterized by in vitro electrophoretic mobility shift assay ( EMSA ) and ISG20 was able to bind HBV ε directly in absence of any other cellular proteins , indicating a direct ε RNA binding capability of ISG20; however , cofactor ( s ) may be required for ISG20 to efficiently degrade ε . In addition , the lower stem portion of ε is the major ISG20 binding site , and the removal of 4 base pairs from the bottom portion of ε abrogated the sensitivity of HBV RNA to ISG20 , suggesting that the specificity of ISG20-ε interaction relies on both RNA structure and sequence . Furthermore , the C-terminal Exonuclease III ( ExoIII ) domain of ISG20 was determined to be responsible for interacting with ε , as the deletion of ExoIII abolished in vitro ISG20-ε binding and intracellular HBV RNA degradation . Taken together , our study sheds light on the underlying mechanisms of IFN-mediated HBV inhibition and the antiviral mechanism of ISG20 in general . Hepatitis B virus ( HBV ) infection remains a significant health threat to humans , leading to the elevated rate of severe liver diseases , such as fulminant hepatitis , fibrosis , cirrhosis , primary hepatocellular carcinoma , and other clinical complications [1] . HBV is the prototype member of hepadnaviridae family which contains a number of DNA viruses replicating their genome through reverse transcription of a viral RNA intermediate in hepatocytes . After cell infection mediated by the viral receptor , called sodium taurocholate cotransporting polypeptide ( NTCP ) [2] , the 3 . 2kb relaxed circular ( rc ) viral DNA genome in the virion enters the nucleus and transforms into a nucleosome-decorated covalently closed circular ( ccc ) DNA minichromosome . Using cccDNA as transcription template , five viral mRNAs are transcribed from different transcription initiation sites but all terminated at a same polyadenylation site , which are , 3 . 5~3 . 6 kb precore mRNA , 3 . 5 kb pregenomic ( pg ) RNA , 2 . 4 and 2 . 1 kb surface ( envelope ) mRNA , and 0 . 7kb X mRNA . Besides serving as the template for translation of viral core protein and polymerase ( pol ) , the pgRNA is also the template for reverse transcription . The pol recognizes a stem loop structure ( epsilon , ε ) at the 5’ terminus of pgRNA to recruit core proteins to encapsidate pol/pgRNA complex into nucleocapsid , where the reverse transcription takes place to yield progeny viral rcDNA [3] . Considering the importance of viral RNA in HBV life cycle , it is conceivable that HBV RNA reduction will result in a suppression of both viral DNA replication and antigen production . Therefore , hijacking HBV RNA is believed to be an important antiviral strategy for host defense against HBV . In line with this , interferons are able to directly reduce HBV RNA through both transcriptional and posttranscriptional mechanisms [4–7] . In terms of interferon-stimulated genes ( ISG ) , the tripartite motif-containing protein 22 ( TRIM22 ) and DEAD-box RNA helicase DDX3 have been shown to inhibit HBV RNA transcription [8 , 9] , and the zinc finger antiviral protein ( ZAP ) and myeloid differentiation primary response gene 88 ( MyD88 ) have been identified as host intrinsic antiviral factors against HBV through promoting the decay of HBV RNA [10 , 11] . In search of ISGs that inhibit HBV replication , we have previously found that the 20 KDa interferon-stimulated gene product , referred to as ISG20 , inhibited HBV replication primarily through reducing HBV RNA transcript levels in cell cultures [4] . This protein , which has been shown to be induced by both type I and type II IFNs , belongs to DEDDh subgroup of the DEDD exonuclease superfamily . Sequence homologies are distributed at three distinct motifs termed as ExoI ( a . a7-16 ) , ExoII ( a . a 86–101 ) and ExoIII ( a . a 147–157 ) [12] . ISG20 is a 3’ to 5’ exonuclease with the substrate preference of single-stranded RNA over double-stranded RNA or single-stranded DNA . Its enzymatic activity is mediated by the ExoII motif , as substitution of the Aspartic acid 94 ( D94 ) in the ExoII motif to Glycine ( D94G ) completely abolishes its exonuclease activity [13] . Cellular proteins may be involved in regulation of ISG20 in substrate recognition and/or binding in order to avoid any non-specific cell toxicity , however which remain to be identified . ISG20 has been shown playing an important role in host antiviral innate immune defense [14] . It is a key ISG against alphavirus replication in host antiviral defense mediated via IFN-α/β signaling [15] . Cells with overexpressed ISG20 are resistant to the infections of RNA viruses including vesicular stomatitis virus ( VSV ) , influenza virus , encephalomyocarditis virus ( EMCV ) , hepatitis A and C virus ( HAV and HCV ) , yellow fever virus , and HIV , but not to a DNA virus , adenovirus [16–18] . In the context of HBV infection , it has been reported that ISG20 was highly induced during IFN-mediated suppression of viral replication in HBV transgenic mouse hepatocyte cell lines [19] , and in the liver of acutely HBV infected Chimpanzee during viral clearance phase [20] . Interestingly , chronic hepatitis B patients who had good responses to IFN treatment showed significantly higher levels of ISG20 in their livers compared to either the patients who were non-responders to IFN treatment or the healthy controls [21] . All the above findings suggest that ISG20 may play a role in suppressing HBV in response to host immune activation or IFN-based immunotherapy . Therefore , we set out to further characterize the antiviral function and mechanism of ISG20 in innate control of HBV . We report herein that: 1 ) the endogenous ISG20 is a restriction factor for HBV replication and it can be upregulated by all three types of IFN to further inhibit HBV; 2 ) ISG20 inhibits HBV replication primarily by promoting viral RNA degradation; 3 ) ISG20-mediated HBV RNA degradation depends on its ribonuclease activity , but the enzymatically inactivated form retains antiviral activity against pgRNA encapsidation by competing with HBV pol for pgRNA binding; 4 ) ISG20 recognizes HBV RNA through directly binding to the lower stem region of epsilon and the binding domain on ISG20 is located in Exo III motif; 5 ) ISG20-epsilon binding is essential to ISG20-mediated HBV RNA degradation in cell cultures , and other cofactor ( s ) may be required for ISG20 to efficiently degrade the stem-loop region of HBV RNA . Our results thus have impacts on a better understanding of ISG20 biology and host-virus interaction , and may provide basis for host ribonuclease-based antiviral development . ISG20 is an interferon inducible protein [16] . Firstly , we assessed the basal expression of ISG20 in cell cultures and its inducibility by IFN . Using Hela cells as a positive control , the basal level of ISG20 expression was detected in both HepG2 cells and Huh7 cells by Western blot , and its expression was significantly upregulated by IFN-α stimulation ( Fig 1A ) . Furthermore , ISG20 can be highly induced by all three types of IFNs in both HepG2 cells and primary human hepatocytes ( PHH ) ( Fig 1B ) . According to a previous study on the transcriptional regulation of ISG20 , a unique interferon stimulated response element ( ISRE ) in the TATA-less human ISG20 promoter confers IRF-1-mediated constitutive transcriptional activity of ISG20 gene and responsiveness to both type I and II interferons [22] . The basal and cytokine-inducible expression of intrahepatic ISG20 indicates a potentially important role of ISG20 in host defense mechanisms against pathogen infections in hepatocytes . In an effort to determine the antiviral activity of ISG20 against HBV replication in hepatocytes , FLAG-tagged wild type ISG20 was co-transfected with pHBV1 . 3 ( HBV precore mRNA and pgRNA transcription is controlled by viral enhancer II and core promoter ( EnII/Cp ) ) or pCMVHBV ( pgRNA transcription is governed by CMV-IE promoter ) in HepG2 cells . HBV viral RNA and DNA were analyzed , as shown in Fig 2A . The result clearly demonstrated that ISG20 markedly inhibited HBV DNA replication ( middle panel ) primarily through reducing the steady state levels of 3 . 5kb viral pgRNA ( top panel ) which is the template for HBV DNA synthesis , and such effect is independent of promoter ( HBV EnII/Cp or CMV-IE ) used to transcribe the pgRNA . A similar antiviral effect of ISG20 was also observed in Huh7 cells ( Fig 2B ) . Furthermore , the ISG20-mediated pgRNA reduction is dose dependent ( S1 Fig ) and not due to a possible cytotoxicity effect caused by ISG20 overexpression ( S2 Fig ) . In addition , along with pgRNA reduction , the levels of 2 . 4 kb and 2 . 1 kb HBV surface mRNA , which share 100% sequence identity with the 3’ portion of pgRNA , were also reduced upon ISG20 expression ( Fig 2 , top panels ) . The reduction of HBV 2 . 4/2 . 1 kb RNA by ISG20 was further confirmed by co-transfection of ISG20 and plasmid pLMS expressing HBV subgenomic surface mRNA in HepG2 cells ( S3A Fig ) . Consistent with ISG20-mediated reduction of 3 . 5 kb and 2 . 4/2 . 1 kb HBV RNA , the levels of supernatant HBeAg and HBsAg were also significantly decreased ( S3B Fig ) . Considering that HBV plasmid DNA is the viral transcription template in the transfection system , it is possible that ISG20 reduces HBV RNA through destabilizing the transfected HBV plasmid DNA , or by blocking the nuclear import of input plasmid . To test these possibilities , we recovered the HBV plasmid DNA from the transfected cells by Hirt DNA extraction , followed by Dpn I digestion and Southern blot analysis . The result showed that the levels of input HBV plasmid DNA recovered from whole cells were equal in the absence or presence of ISG20 , as revealed by Dpn I-digested fragments ( S4A Fig , top panel ) . Furthermore , while ISG20 overexpression reduced HBV core DNA replication in cytoplasm , it did not affect the level of remaining cytoplasmic HBV plasmid DNA ( S4A Fig , middle panel ) , further confirming that ISG20 does not alter the stability and distribution of the input HBV plasmid . In addition , ISG20 transfection in HBV stable cell line HepDES19 cells also resulted in the reduction of HBV RNA , which were transcribed from the integrated HBV DNA ( S4B Fig ) . Collectively , these results clearly exclude any negative effect of ISG20 on the stability of HBV RNA transcription template , which is consistent with the reported inability of ISG20 to degrade double stranded DNA substrate in vitro [13] . Next , in order to determine whether ISG20-mediated down-regulation of HBV RNA via transcriptional or posttranscriptional mechanism , we first assessed the potential effect of ISG20 on viral promoter activity . As shown in S5 Fig , the luciferase reporter assays demonstrated that ISG20 did not significantly affect the activities of HBV Enhancer II/Core ( EnII/Cp ) promoter , S2 promoter , or CMV-IE promoter , but even enhanced HBV S1 promoter activity . Therefore , it is unlikely that ISG20 reduces HBV RNA through inhibiting viral promoter activity at the transcriptional level . We then compared the decay kinetics of HBV RNA in the absence and presence of ISG20 overexpression . Briefly , ISG20 expression vector or control empty vector was transfected into the inducible HBV stable cell line HepDES19 cells in the absence of tetracycline ( tet ) to induce HBV pgRNA transcription and ISG20 expression; 36 h later , tet was added back to shut down the de novo pgRNA transcription from the transgene , and after that , the levels of HBV RNA at different time points were analyzed by Northern blot ( Fig 3A ) . The representative Northern blot demonstrated that the levels of HBV RNA in HepDES19 cells overexpressing ISG20 at each time point were less than that in the control group , suggesting that ISG20 promotes HBV RNA degradation . This is consistent with the previous observations that ISG20 functions as the 3’-5’ riboexonuclease to degrade single-stranded viral RNA genomes [14 , 16] . It is worth noting that not every HBV RNA-containing HepDES19 cells were transfected by ISG20 , therefore , the actual effect of ISG20 on HBV RNA stability ought to be stronger than observed in this experiment . To further quantitatively validate the observed enhancement of HBV RNA degradation by ISG20 , we transfected HepG2 cells with tetracycline-inducible HBV expression vector pTREHBVDES and plasmid pTet-off expressing the tetracycline-responsive transcriptional activator ( tTA ) , together with either control vector or ISG20 plasmid . In this way , both HBV RNA and ISG20 coexisted in same cells . Four days later , similarly to the above experiment in HepDES19 cells , tetracycline was added back to block new round of pgRNA transcription , and the time course HBV RNA level was quantified by qPCR . As shown in Fig 3B , the regular half-life of intracellular HBV RNA was approximately 6 h , which is consistent with previous measurements [10 , 23] , but the decay kinetics of HBV RNA was much faster in cells transfected with ISG20 , of which the HBV RNA half-life was much less than 3 h , suggesting that ISG20 promotes HBV RNA decay . Hence , we conclude that ISG20-mediated HBV RNA reduction is primarily through downregulation of HBV RNA stability . There is a basal level of ISG20 expression in hepatocytes and that could be further induced by interferons ( Fig 1 ) . It was therefore of interest to assess the antiviral activity of endogenous ISG20 against HBV infection under its basal and induced levels . Firstly , we performed the test in HBV stable cell line HepDES19 [24] . As shown in Fig 4 , knock down of ISG20 expression in the absence of IFN-α treatment resulted in a marked upregulation of the steady state levels of HBV total RNA , encapsidated pgRNA , and DNA ( comparing lane 4 to lane l ) . The observed relatively greater upregulation of core DNA than viral RNA upon knock down of ISG20 might be due to the longer lifespan of DNA . When the cells were treated with IFN-α , ISG20 was upregulated , and viral RNA and DNA were reduced , in an IFN dose-dependent manner ( lanes 1–3 ) . It is worth to note that the IFN-elicited reduction of HBV RNA in HepDES19 cells was less profound than in HBV-transfected cells with ISG20 overexpression ( compared Fig 4 to Fig 2 ) . Such difference might be due to the reported weak response of hepatoma cells to IFN treatment [25] , and/or the potential unknown negative regulator ( s ) of ISG20 co-induced by IFN . Nonetheless , in the cells treated with IFN-α and ISG20 siRNA , the levels of HBV RNA were restored to the untreated control level ( comparing lanes 5 and 6 to lanes 1–3 ) , suggesting that ISG20 plays a major role in IFN-mediated reduction of HBV RNA in HepDES19 cells . However , neither viral RNA nor DNA reached their levels in cells treated with ISG20 siRNA only ( comparing lanes 5 and 6 to lane 4 ) , indicating that other ISGs are involved in IFN-mediated suppression of HBV DNA replication at multiple steps . Next , we assessed the antiviral function of ISG20 in an HBV in vitro infection system , specifically , the viral receptor NTCP reconstituted HepG2 cells [26] . The endogenous ISG20 in HepG2-NTCP12 cells was stably inhibited through lentiviral shRNA transduction . The obtained ISG20 knock down cell line and control knock down cell line were then infected with HBV and treated with or without IFN-α . As shown in Fig 5A , the basal ISG20 expression was successfully knocked down in HepG2-NTCP12-shISG20 cells ( comparing lane 3 to lane1 ) , and IFN-α significantly upregulated ISG20 expression in the HepG2-NTCP12-shcontrol cells ( comparing lane 2 to lane1 ) , but only slightly induced ISG20 in the HepG2-NTCP12-shISG20 cells ( comparing lane 4 to lane3 ) . Then the outcomes of HBV infection were analyzed . The HBcAg immunofluorescence staining demonstrated that knock down of ISG20 resulted in a higher percentage of HBV positive cells when compared to control cells without or with IFN-α treatment ( Fig 5B ) . Because that the ISG20-mediated HBV RNA degradation will lead to an overall suppression of the entire HBV life cycle in the infection system , we infer that the increased percentage of core staining-positive cell number in HBV infected HepG2-NTCP-shISG20 cells is a combinational end effect of increased viral protein expression , DNA replication , and cccDNA synthesis , plus the possible higher virus spread rate . Since HBV RNA is a more convincing marker for HBV infection and the authentic antiviral target of ISG20 , the qPCR assay of HBV total RNA from the infected cells further demonstrated that IFN-α is able to reduce HBV RNA in infected cells , and knocking down of ISG20 greatly favors HBV infection in both IFN-treated and untreated cells ( Fig 5C ) . Consistent with the observations in HepDES19 cells ( Fig 4 ) , inhibition of ISG20 significantly but partially abrogated the overall antiviral activity of IFN-α in HBV infected cells ( Fig 5B and 5C ) , presumably due to the induction of other antiviral ISGs . Collectively , the above data suggested that ISG20 serves as a host intrinsic restriction factor to limit HBV infection under both basal expression and cytokine induction . ISG20 has three putative exonuclease domains and it has been reported that a single D94G mutation in the Exo II domain completely abolished the protein’s enzymatic activity in vitro [13] . In line with this , we found that the enzymatic inactive mutant F-ISG20D94G failed to reduce HBV RNA in HepG2 cells ( Fig 6 , top panel ) , suggesting that ISG20-mediated HBV RNA decay requires its exoribonuclease activity . Interestingly , although the levels of total viral RNA and capsid remained the same , the level of encapsidated HBV pgRNA was significantly reduced under the expression of ISG20D94G mutant ( middle panel , comparing lanes 5&6 to 1&2 ) , which consequently led to the inhibition of viral DNA replication ( bottom panels ) . The inhibitory effect of ISG20D94G on pgRNA encapsidation was further confirmed with HBV genome harboring polymerase Y63D mutation , which supports HBV pgRNA binding and encapsidation but not DNA replication ( S6 Fig ) Since HBV pgRNA encapsidation requires the dynamic interactions among three viral components , specifically the pgRNA , Pol , and core protein [27] , it is possible that ISG20D94G might be antagonizing one or more of these three viral factors to block encapsidation . However , ISG20D94G did not inhibit capsid assembly ( Fig 6 , 3rd panel from the top ) , and ISG20-mediated HBV RNA degradation does not rely on viral core or pol protein ( S7 Fig ) . Based on these data , we hypothesized that ISG20 might physically bind to HBV RNA for degradation , and thus in the particular case of ISG20D94G , its association with pgRNA inhibited the binding of HBV polymerase to pgRNA and the subsequent pgRNA encapsidation . To test the above hypothesis , we first investigated the association between ISG20 and HBV RNA in cell culture . In this immunoprecipitation ( IP ) -Northern assay ( Fig 7 ) , plasmid pCMVHBVΔCΔP which supports HBV RNA transcription but does not support pgRNA encapsidation in cis due to the deletion of core and Pol expression was used as HBV RNA transcription template , and FLAG-tagged HBV Pol ( FLAG-Pol ) was used in trans as positive control for pgRNA binding . To avoid the rapid RNA degradation mediated by wild type ISG20 expression , ISG20D94G mutant was used for this assay . The upper panel of Fig 7 indicates the input HBV RNA level and FLAG-Pol or F-ISG20D94G protein levels . The lower panel of Fig 7 shows the Northern blot results after IP with FLAG antibody ( Ab ) . As expected , FLAG-Pol interacted with HBV RNA , as FLAG Ab , but not the HA Ab , could pull down the HBV RNA . Same as FLAG-Pol , ISG20D94G also specifically formed complex with HBV RNA in this IP-Northern assay , suggesting that ISG20 interacts with HBV RNA in cell cultures , though it remains unknown whether the interaction is direct or indirect . Next , we set out to investigate the potential competitive effect of ISG20 on Pol-pgRNA binding . Fixed amount of FLAG-Pol and pCMVHBVΔCΔP were co-transfected into cells together with plasmid expressing HA-tagged ISG20D94G ( HA-ISG20D94G ) in a titrated manner . Fig 8 shows the input HBV RNA and FLAG-Pol and HA-ISG20D94G protein levels ( upper panels ) and the immunoprecipitated HBV RNA levels after HA or FLAG Ab pull-down ( the lower panels ) . The HA Ab IP-Northern result demonstrated that ISG20D94G bound to HBV RNA in a dose dependent manner . In contrast , increased amount of ISG20D94G resulted in the decreased amount of HBV RNA associated with Pol , suggesting that ISG20D94G competes with HBV Pol for binding of pgRNA and thus inhibits HBV pgRNA encapsidation . It is assumed that the wildtype ISG20 may possess additional antiviral function as what ISG20D94G does in cells , but the inhibition of pgRNA encapsidation , if any , will likely be overwhelmed by its dominant primary effect on RNA degradation . The above data demonstrated that ISG20 interacts with HBV RNA . In order to map the ISG20-targeting sequences , we made use of a series of internal deletion clones and terminal deletion clones of the 3 . 5kb HBV RNA genome to express truncated RNAs and their sensitivities to ISG20-mediated RNA degradation were assessed ( Fig 9A and S8A Fig ) . Firstly , we found that all those pgRNA mutants with internal deletions ( within nt 2009-3182/1-1574 ) were sensitive to ISG20 ( S8B Fig ) . Then the terminal redundancy ( TR , nt 1820–1918 , containing ε ( nt 1849–1909 ) ) mutants with deletion of either 5’ TR ( Δ5TR ) or 3’ TR ( Δ3TR ) from pgRNA were tested . Interestingly , while the single TR deleting pgRNA remained sensitive to ISG20 ( Fig 9B , lanes 1–4 ) , the pgRNA mutant with both 5’ and 3’ TR removed ( Δ5/3TR ) became completely resistant ( Fig 9B , lanes 5–6 ) , indicating that one copy of the TR sequences is sufficient to confer the susceptibility of HBV pgRNA to ISG20-mediated RNA degradation . This is also consistent with the fact that ISG20 is also able to degrade HBV subgenomic RNAs which only contain the 3’ TR ( Fig 2 and S3 Fig ) . To further confirm that the HBV TR is the ISG20 responsive element , HBV core promoter activity reporter plasmids EnII/Cp-Luc with TR region inserted at the 5’ ( EnII/Cp-TR-Luc ) , or 3’ ( EnII/Cp-Luc-TR ) , or both 5’ and 3’ ( EnII/Cp-TR-Luc-TR ) non-translational region of luciferase gene ( Fig 9C ) were transfected into HepG2 cells with or without F-ISG20 . The luciferase activities were measured and the result demonstrated that the introduction of HBV TR into luciferase mRNA resulted in a significant reduction of luciferase gene expression under ISG20 expression ( Fig 9C ) . Therefore , we conclude that ISG20 selectively targets TR sequences of HBV RNA to initiate RNA degradation . HBV Pol recognizes and binds on a stem-loop structure in 5’ TR region of HBV pgRNA , called epsilon ( ε ) , to initiate pgRNA encapsidation and reverse transcription [29] . Since ISG20-mediated HBV RNA degradation relies on TR sequence ( Fig 9 ) and ISG20D94G competes with Pol for HBV RNA binding ( Fig 8 ) , it is therefore of immediate interest to investigate whether ISG20 also directly binds on HBV ε to exert its antiviral functions . To this end , an electrophoretic mobility shift assay ( EMSA ) was employed . As shown in Fig 10 , purified recombinant His-tagged ISG20 proteins were incubated with 5’ radiolabeled HBV ε RNA in vitro and the protein/RNA complexes were analyzed via the mobility shifting of the RNA in the native polyacrylamide gel . The proper folding of stem-loop structure of ε RNA was confirmed by recombinant Pol/ε complex formation as detected in the EMSA assay ( S9 Fig ) . Interestingly , in the absence of other cellular proteins , His-ISG20 could directly , dose-dependently bind on ε RNA , shown as the shift bands in Fig 10C ( lanes 3 , 5 , 7 ) , and the binding specificity between His-ISG20 and ε was confirmed by the His-antibody ( Ab ) /His-ISG20/ε super-complex formation detected as a super-shift in EMSA ( lanes 4 , 6 , 8 ) . Competition of ISG20 binding to the radiolabeled ε RNA by excess amount of cold ε probes indicated again that ISG20 specifically binds to HBV ε ( lanes 9 , 10 , 11 ) . It is of note that the incubation of ISG20 and ε RNA during EMSA was performed in protein/RNA binding buffer without manganese , an essential ion for the enzymatic activity of ISG20 [13] , therefore the degradation of ε RNA , if any could occur , would not be observed in this binding assay . To further dissect the substructural domain of ε targeted by ISG20 , the synthetic upper stem-loop RNA and the lower stem-loop RNA of ε and their mutants , as shown in the Fig 11A , were radiolabeled and used in gel shift assay . Interestingly , ISG20 bound to the lower stem with bulge sequence serving as loop , but not the upper stem-loop of ε in EMSA ( Fig 11B ) , which indicates that ISG20 recognizes the lower stem and/or bulge of HBV ε as its binding site . To further determine the binding position within the lower stem and bulge region , the bulge sequences were mutated or the lower stem were truncated from the bottom ( Fig 11A ) , and used as probes in EMSA . Interestingly , we found that the loop ( or bulge ) sequence is not critical for ISG20 binding as long as the lower stem is intact; however the bottom 4 base-pairs of the stem is indispensable for ISG20 binding ( Fig 11C ) . To further confirm the requirement of these 4 base-pairs for ISG20 to carry out its biological functions , we disrupted the base pairing at the bottom of ε stem-loop in HBV 2 . 1 kb surface mRNA by removing 4 nucleotides from the 3’ end of ε coding sequence , and found that the mutant RNA became resistant to ISG20-mediated RNA degradation in cell cultures ( Fig 11D ) . Collectively , these findings imply that ISG20 recognizes sequence-specific secondary structure of HBV ε for RNA binding and degradation . Since ISG20 is able to bind on its RNA substrate directly in the absence of any other cellular or viral proteins , it is therefore of interest to investigate which domain of ISG20 is responsible for RNA binding . Three distinct exonuclease motifs , namely Exo I ( a . a7-16 ) , Exo II ( a . a 86–101 ) and Exo III ( a . a 147–157 ) , have been predicted in ISG20 protein sequence from previous studies [14 , 30] ( Fig 12A ) . Based on that , the recombinant mutant ISG20 proteins with each Exo domain deletion or Exo II D94G mutation were tested for ε binding activity by EMSA . As shown in Fig 12B and 12C , in comparison to wildtype ISG20 , deletion of Exo I or Exo II did not reduce , but even increased the binding affinity of ISG20 with ε ( lanes 2–4 ) ; however ISG20 with Exo III deletion completely lost the binding activity ( lane 5 ) , indicating that the Exo III is the functional binding domain of ISG20 for its RNA substrate . In addition , the enzymatically inactive D94G mutant retains its function for ε binding ( lane 6 ) , which is consistent with the observed association of ISG20D94G with HBV RNA and subsequent inhibition of pgRNA encapsidation in cells ( Figs 6–8 ) . It is worth noting that the ΔExo I and ΔExo II deletion mutants , as well as the ISG20D94G mutant , exhibited altered mobility compared to wildtype ISG20 in the gel shift assay ( Fig 12C , lane 2 vs . lanes 3 , 4 , 6 ) . While it is predictable that the faster mobility of ΔExo I:ε and ΔExo II:ε complexes is likely due to the protein’s lower molecular weight ( MW ) , the lower shift of ISG20D94G:ε complex is probably because that such RNP complex exhibits a different structural conformation compared to the ISG20:ε complex . To confirm the function of Exo III domain for ISG20 activity , the FLAG-tagged ISG20 protein with a deletion of Exo III domain was expressed in HBV transfected cells . As shown in Fig 12D , while the wildtype ISG20 significantly reduced HBV RNA levels as expected , the ISG20ΔExoIII mutant was similar to ISG20D94G in inefficiency in reduction of HBV RNA . Therefore , we conclude that , while ISG20 requires Exo II domain for its enzymatic activity , the separate Exo III domain is responsible for RNA substrate binding of ISG20 , which is a prerequisite for RNA degradation . Lastly , we set out to characterize the enzymatic activity of ISG20 against ε RNA in vitro . The previous biochemistry study has demonstrated that ISG20 catalyzes RNA degradation in vitro in a manganese-dependent manner [13] . As shown in Fig 13 , under the reaction condition that supports ISG20 ribonuclease activity [13] , a wide spectrum endoribonuclease RNase A efficiently degraded ε RNA as a positive control ( panel A and B , lanes 1–3 ) . However , ISG20 exhibited much less efficiency to degrade the full length ε RNA substrate ( panel A and B , lanes 1 , 4–5 ) , and other two stem-loop subdomains ( US+L , LS+B ) ( panel A , lanes 6–11 ) , suggesting that ISG20 alone does not favor RNA with secondary structures for degradation . Such observation is consistent with a previous study showing that ISG20 operated poorly on double-stranded regions of RNA substrate in vitro [13] . In line with these , we found that ISG20 was able to catalyze the degradation of RNA substrate derived from the left arm of ε ( panel B , lanes 6–8 ) , which is lack of secondary structures . Furthermore , ISG20 efficiently degraded the single-stranded linear 30-mer poly ( rA ) oligo in vitro ( panel B , lane 11 ) , confirming that the recombinant ISG20 protein used in the assay was enzymatically active . As a control , RNase A was unable to degrade poly ( rA ) because it only cleaves RNA string at the 3´ end of pyrimidine ( rC or U ) residues ( panel B , lane 10 ) [31] . Since the RNA substrates were 5’ radiolabeled , the pattern of degradation intermediates generated by ISG20 is consistent with the 3’-5’ exoribonuclease activity of ISG20 ( panel B , lanes 7 , 8 , 11 ) . Collectively , the above in vitro data indicate that other host cofactor ( s ) may be required to coordinate with ISG20 for the observed efficient HBV RNA degradation in cells . In summary , we report here that ISG20 functions as a 3’-5’ exoribonuclease to degrade HBV RNA , by which serves as an intrinsic host restriction factor for HBV infection . The selectivity of ISG20 for HBV RNA is determined by a unique viral ribonucleotide string that shapes a stem-loop structure ( ε ) which exists in all HBV RNA species . ISG20 directly binds to the lower stem of ε through the C-terminal Exo III domain , and such protein-RNA interaction , possibly with the assistance of a host cofactor ( s ) , will lead to the major antiviral action of ISG20 , which is the enzymatic activity dependent HBV RNA degradation . Subsequently , viral protein translation and DNA replication will be inhibited ( Fig 14 ) . In addition , under certain circumstance that the enzymatic activity of ISG20 could be inhibited ( i . e . the ISG20D94G mutant ) , the binding of ISG20 to 5’ ε may remain antiviral by competing the binding of viral polymerase with ε and blocking the subsequent pgRNA encapsidation and reverse transcription . It is well documented that , during acute HBV infection , the activation of host T-cell responses plays a pivotal role in clearing virus from hepatocytes via a noncytolytic mechanism , which is largely attributed to the antiviral and proinflammatory cytokines produced by the infiltrating immune cells , such as IFN-γ and TNF-α [32 , 33] . To date , IFN-α is the only available immunotherapy for treatment of chronic hepatitis B [34] . Besides serving as an immunomodulatory agent to regulate the host adaptive immunity against HBV , the intracellular antiviral activity of IFN-α is mediated by an array of ISGs . In search of intrahepatic ISGs associated with immune clearance of HBV , a growing body of evidence implies that ISG20 may be involved in IFN mediated inhibition of HBV , including 1 ) the upregulation of ISG20 in HBV transgenic mouse hepatocyte cell lines in which the HBV replication was inhibited by IFN [19]; 2 ) the elevation of ISG20 level in viral clearance phase of an acutely HBV infected Chimpanzee [35]; and 3 ) the upregulation of ISG20 in chronic hepatitis B patients who responded well to IFN-α therapy [21] . Consistent with the above findings , we found that ISG20 can be highly induced in hepatocytes in vitro by all three types of interferons , and it plays a significant role in IFN-α elicited antiviral effects on HBV infection and replication ( Figs 1 , 4 and 5 ) . Considering that HBV RNA are essential genetic components for viral replication and gene expression in its life cycle , upregulation of ISG20 expression by cytokines to eliminate HBV RNA is a very economical and efficient way in host defense mechanisms . In addition , our study also expands the antiviral spectrum of ISG20 to a DNA virus . Considering that current HBV therapy with nucleoside analogues is ineffective against viral protein expression and cannot efficiently promote HBeAg or HBsAg seroconversion [34] , it is thus envisioned that the ISG20-based antiviral therapy could be developed in the future to reduce both viremia and antigenemia in CHB patients . While its cellular function remains elusive , ISG20 has been previously shown to inhibit a number of RNA viruses , including VSV , influenza viruses , HIV-1 , picornaviruses ( EMCV , HAV ) , and flaviviruses ( HCV , WNV , and bovine viral diarrhea virus ( BVDV ) ) [16–18 , 36 , 37] . The previously reported antiviral effects of ISG20 on these RNA viruses are all dependent on its exonuclease activity , though the degradation of viral RNA directly catalyzed by ISG20 has not been experimentally confirmed . However , RNA degradation might not be the only mechanism of ISG20’s antiviral action . For instance , Zhou et al found that ISG20 did not promote the degradation of a replication-defective HCV RNA genome in Huh7 . 5 cells [17]; and Qu et al observed that ISG20 inhibits Influenza A Virus replication through interaction with viral nucleoproteins [37] . Furthermore , not all the RNA viruses are susceptible to ISG20 , as severe acute respiratory syndrome coronavirus ( SARS-CoV ) has shown resistance to ISG20 in cell cultures [17] , indicating that there is viral specific factor ( s ) to determine the antiviral specificity of ISG20 . Here , we clearly demonstrated that ISG20 is able to promote HBV RNA degradation in cell cultures , and such effect is dependent on the protein’s exonuclease activity ( Figs 2–6 ) . During the preparation of this manuscript , Leong et al also reported that ISG20 inhibited HBV replication in cell cultures and in hydrodynamic injected mouse liver via exoribonuclease-dependent degradation of viral RNA , which is largely consistent with our results , but their study did not touch on the molecular mechanism for the selective targeting of HBV RNA by ISG20 [38] . In our study , we further mapped the sequence element responsible for ISG20-mediated HBV RNA degradation to ε region in HBV RNA genome ( Figs 7–9 ) , and confirmed the direct binding of ISG20 to ε in EMSA assay ( Fig 10 ) . To the best of our knowledge , this is the first report revealing that the RNA sequence/structure-specific recognition is required for ISG20 to degrade its viral RNA substrate . Such characteristic of ISG20 makes it mechanistically distinct from RNase L , a well-known antiviral ISG that , upon activation , nonspecifically destroys both viral and cellular RNA in cells [39] . In this regard , ISG20 represents a novel class of cellular antiviral ISGs , which is reminiscent of the sequence-specific antiviral RNAi system [40] , but ISG20 targets viral RNA directly without the help of small guide RNA . HBV RNAs are typical mRNAs produced by host RNA polymerase II , all of which possess 5’ cap and a 3' poly ( A ) tail similar to cellular mRNAs [41] . Despite of the sequence difference with host mRNA , the ε is a unique structure signature of HBV RNA , which serves as pgRNA packaging signal and the priming site for interacting with viral polymerase [42] . It is thus conceivable that ISG20 recognizes ε as a “non-self” target to initiate its antiviral function . In line with this , it has been recently reported that host innate sensor RIG-I binds to ε to trigger IFN-λ production [43] . In addition , we further determined that the bottom 4bp of lower stem of ε is the minimal responsive element of ISG20 , as the removal of such sequence from HBV RNA made ISG20 lose both the ε binding activity in vitro and HBV RNA reduction effect in cell cultures ( Fig 11 ) . These sequence and structure information could serve as a consensus motif to predict ISG20-responsive RNA elements in other viruses , and perhaps host RNAs as well . Despite acting as a host antiviral factor , the other potential functions of ISG20 in regulating host cellular function remain obscure . Since ISG20 targets viral mRNA for destabilization , it is plausible that they may also alter the stability of certain cellular mRNA to regulate host gene expression . For example , the observed enhancement of HBV S1 promoter activity by ISG20 may be due to the ISG20-mediated upregulation of host transcription factor ( s ) required for S1 promoter activity or downregulation of negative regulator ( s ) of S1 promoter ( S5 Fig ) . This notion is also supported by previous observations that ISG20-mediated antiviral effect on HCV was not through inhibition of HCV RNA stability or translation , indicating that ISG20 might degrade the mRNA of a host factor ( s ) essential for HCV replication [17] . More interestingly , considering that ISG20 favors short RNAs with stem-loop structures for affinity binding , it is also possible that ISG20 potentially targets host microRNA ( miRNA ) precursors ( pri-miRNA and pre-miRNA ) [44] and long non-coding RNA ( lncRNA ) [45] , which may contain ε-like stem-loop structures , to regulate miRNA processing and lncRNA stability , respectively . In addition , it is known that IFNs regulate cellular gene expression primarily at the mRNA level [46] , so it is possible that ISG20 , as a IFN-inducible RNase , may partly mediate the action of IFN through altering the host mRNA stability . We are currently conducting RNA-seq analysis to search for potential host RNA substrates of ISG20 . ISG20 has three predicted exonuclease domains ( Exo I-III ) with Exo II being validated as the major enzymatic domain [14] . We found that the catalytic site ( D94 ) in Exo II of ISG20 was essential for HBV RNA reduction , but the ISG20D94G mutant retained RNA binding activity and inhibited HBV pgRNA encapsidation through competing the binding of HBV polymerase to ε ( Figs 6–8 ) . Although such antiviral effect may be just minor in the context of wildtype ISG20-mediated HBV RNA degradation , this serendipitous finding led us to discover the binding property of ISG20 with HBV RNA ( Fig 10 ) . Interestingly , the minimal binding site on ε for ISG20 is located at the bottom 4bp of the stem-loop ( Fig 11 ) , which is dispensable for HBV polymerase to bind ε [47] , indicating that the inhibition of polymerase binding by ISG20 is due to ISG20/ε interaction-induced steric hindrance and/or conformational shift of ε . In addition , the domain of ISG20 for ε binding was mapped to the Exo III region ( Fig 12 ) , suggesting that the substrate interaction and degradation domains of ISG20 are separated . This is consistent with the reported crystal structure analysis of ISG20 , which predicted that the amino acid composition of Exo II is unfavorable for RNA binding [30] . However , no classical RNA binding motif was found in the Exo III domain or the entire ISG20 sequence by computational prediction ( http://www . rcsb . org/pdb/protein/Q96AZ6 ) . ISG20 has been shown to interact with host U1/2/3 small nucleolar RNAs ( snoRNAs ) [48] , which also have stem-loops , further indicating that ISG20 may possess a novel binding motif for such kind of double stranded RNA secondary structures including HBV ε . We have previously identified the host zinc-finger antiviral protein ( ZAP ) as a restriction factor of HBV by promoting HBV RNA degradation [10] . ZAP exerts its antiviral activity against HBV with certain features similar to ISG20 , specifically , both ZAP and ISG20 interact with HBV RNA for RNA degradation in a ε-dependent manner . Since ZAP does not possess nuclease activity [49] , our initial hypothesis was that ZAP might recruit ISG20 to ε for RNA degradation . However , co-immunoprecipitation assay did not reveal any detectable association between ZAP and ISG20 in the absence or presence of HBV RNA ( S10 Fig ) , indicating that ZAP and ISG20 work independently to target HBV ε for their antiviral functions , and host cells have evolved multiple mechanisms to recognize the “non-self” HBV ε as an antiviral sensing signal . It is worth to note that ISG20 is a 3’-5’ exoribonuclease and the ε locates near the 3’ end of all HBV RNAs and there is an additional 5’ copy in precore mRNA and pgRNA [14 , 50] , we therefore speculate that the binding of ISG20 to ε is an initial substrate sensing step , and RNA topology changes and/or ISG20 translocation may be requires for subsequent RNA degradation from 3’ end of HBV RNA . In addition , since ISG20 could single-handedly degrade poly ( rA ) oligo in the in vitro degradation assay ( Fig 13B ) , whether the poly ( A ) -specific ribonuclease ( PARN ) is required to remove the poly ( A ) tail in ISG20-mediated HBV mRNA degradation reaction awaits further investigations . Interestingly , our in vitro RNA degradation assay demonstrated that ISG20 efficiently degrades single-stranded RNA but not the ε RNA substrate ( Fig 13 ) , indicating that additional host factor ( s ) , likely a RNA helicase , is required to unwind the double-stranded region of ε for ISG20 to catalyze RNA degradation . Therefore , we have embarked on the identification of cofactors in ISG20 degradesome complex by proteomic approach . Taken together , the phenotypic and mechanistic characterizations of ISG20-mediated antiviral effect on HBV replication presented in current study not only shed light on ISG20 biology and virus-host interaction , but also provide insight into development of ribonuclease-based antiviral therapy for treatment of hepatitis B . Human hepatocyte-derived HepG2 and Huh7 cells were maintained in DMEM/F12 medium ( Mediatech ) supplemented with 10% fetal bovine serum , 100 U/ml penicillin and 100 μg/ml streptomycin . HepDES19 cells were maintained in the same way as HepG2 , but with the addition of 1 μg/ml tetracycline ( tet ) and 400 μg/ml G418 [24] . To initiate HBV replication in HepDES19 cells , tet was withdrawn from the culture medium and the cells were cultured for the indicated time . Primary human hepatocytes were purchased from Triangle Research Labs and cultured according to supplier’s protocol . HepG2-NTCP12 cells were maintained as previously described [26] . Endogenous ISG20 expression in HepG2-NTCP12 cells was stably knocked down by transduction with ISG20 shRNA lentiviral particles ( Santa Cruz Biotechnology ) per the manufacturer’s directions . The transduced cells were selected with 3 μg/ml puromycin , the antibiotics-resistant cells were pooled and expanded into cell line , namely HepG2-NTCP12 shISG20 . Control known down cell line was generated by transduction with control shRNA lentiviral particles ( Santa Cruz Biotechnology ) and designated HepG2-NTCP12 shcontrol . Control siRNA and ISG20 siRNA were purchased from Santa Cruz Biotechnology for transient knock down experiments . Recombinant human IFN-α2a , IFN-γ , and IFN-λ1 were purchased from PBL Biomedical Laboratories . HBV ( genotype D , subtype ayw ) replication competent plasmids , pHBV1 . 3 and pCMVHBV , in which the transcription of viral pgRNA is governed by authentic HBV core promoter and the human cytomegalovirus immediate-early ( CMV-IE ) promoter , respectively , were described previously [4 , 10 , 51 , 52] . The plasmid pTREHBVDES transcribes pgRNA under the control of tetracycline-inducible CMV promoter , it has been previously used together with plasmid pTet-off ( Clontech ) to generate the HepDES19 stable cell line [24] . pHBV1 . 3ΔC is a 1 . 3mer HBV plasmid with mutation of the start codon ( ATG to ATA ) of core protein open reading frame ( ORF ) [53] ( provided by Dr . Robert Lanford ) . pHBV1 . 3ΔP is a 1 . 3mer HBV plasmid with start codon mutation ( ATG to ACG ) of pol ORF to abolish pol expression but without changing the amino acid sequence of the overlapped core protein . Plasmid pCMVHBVΔCΔP contains start codon mutations of core ( ATG to CTG ) and pol ( ATG to ACG ) ORF in pCMVHBV backbone to block the expression of core and pol , respectively . pCMVHBV-Y63D bears a Y63D substitution in the reverse transcriptase domain of HBV pol gene to block DNA synthesis but still allows pgRNA encapsidation [54] ( Provided by Dr . Jianming Hu ) . Plasmid pCMV-FLAG-Pol , which expresses HBV Pol with three copies of the FLAG epitope at the N terminus by CMV promoter [55] , was provided by Dr . Wang-Shick Ryu and used in cell transfections . Plasmid pcDNA-3FHP , another HBV polymerase expression construct that expresses a 3×FLAG-tagged HBV polymerase ( HP ) under T7 promoter in pcDNA backbone [56] , was provided by Dr . Jianming Hu and used in in vitro translation of 3F-HP . Plasmid pLMS expresses 2 . 4kb and 2 . 1kb HBV surface mRNA under the control of authentic viral surface promoters ( S1 and S2 ) [57] ( provided by Dr . Youhua Xie ) . Plasmid pMS expresses 2 . 1kb HBV surface mRNA under the control of a CMV promoter , and pMSΔ4bp was constructed by removing four nucleotides from the bottom right arm of the lower stem of ε sequence in pMS . The pgRNA internal deletion clones ( pgID-1 to pgID14 ) and terminal redundancy ( TR ) deletion clones ( pg-Δ5TR , pg-Δ3TR , pg-Δ5/3TR ) , and luciferase reporter plasmids ( En/Cp-Luc , S1-Luc , S2-Luc , En/Cp-TR-Luc , En/Cp-Luc-TR , and En/Cp-TR-Luc-TR ) were constructed previously [10] . Plasmids expressing N-terminal FLAG-tagged wildtype ISG20 ( F-ISG20 ) and ISG20D94G ( F-ISG20D94G ) under the control of CMV promoter were described previously [36] . F-ISG20ΔExoIII was derived from F-ISG20 by deleting the coding sequence for ExoIII domain of ISG20 . Plasmid HA-ISG20D94G that expresses HA-tagged ISG20D94G was constructed by Q5 Site-Directed Mutagenesis Kit ( New England Biolabs ) to replace the FLAG-tag sequence in F- ISG20D94G with HA-tag . Plasmid HA-ZAPS expresses the HA-tagged ZAP-S protein in eukaryotic cells [10] . Cells seeded in the collagen coated plate were transfected with indicated plasmid ( s ) or siRNA by Lipofectamine 2000 ( Life Technologies ) according to the manufacturer’s directions . HepG2 cells in 96-well-plate were transfected with promoter reporter plasmid plus vectors expressing gene of interest by Lipofectamine 2000 . pRL-CMV Renilla luciferase reporter plasmid was cotransfected for normalization of transfection efficiency . For each transfection , empty control plasmid was added to ensure that each transfection receives the same amount of total DNA ( 200 ng/well ) . Three days after transfection , luciferase activities were measured by the dual luciferase assay kit ( Promega ) and BioTek Synergy 2 Multi-Mode Reader . HepG2-NTCP12 shcontrol and shISG20 cells were spinoculated with HBV virion particles derived from HepDE19 cells at 100 virus genome equivalent ( vge ) /cell according to a published protocol [26] . Six days post infection , cells were fixed and subjected to HBcAg immunofluorescence as described previously [26] . Total cellular RNA , Hirt DNA , cytoplasmic encapsidated HBV pgRNA and core DNA were extracted as described previously [10 , 24 , 51 , 58 , 59] . For HBV RNA qPCR assay , DNase I-treated total cellular RNA was used to generate cDNA by SuperScript III Reverse Transcriptase ( Life Technologies ) . Real-time PCR was performed with SYBR Green Master ( Roche ) and the LightCycler 96 System ( Roche ) for detecting HBV total RNA by using HBV specific primers targeting the 3’ overlapping region of all the HBV RNA species ( forward primer: 5’-ACTCTCTCGTCCCCTTCTCC-3’ , reverse primer: 5’-GGTCGTTGACATTGCAGAGA-3’ ) . The relative expression levels of HBV RNA were normalized to β-actin from the same samples . For HBV RNA Northern blot analysis , total cellular RNA or encapsidated pgRNA samples were resolved in 1 . 5% agarose gel containing 2 . 2 M formaldehyde and transferred onto Hybond-XL membrane ( GE Healthcare ) . For DNA Southern blot analysis , Hirt DNA or HBV core DNA samples were resolved by electrophoresis into a 1 . 5% agarose gel and blotted onto Hybond-XL membrane . Membranes were probed with either α-32P-UTP ( 800 Ci/mmol , Perkin Elmer ) labeled plus-strand-specific ( for Northern blot hybridization ) or minus-strand-specific ( for Southern blot hybridization ) full-length HBV riboprobe and exposed to a phosphorimager screen . Hybridization signals were quantified with QuantityOne software ( Bio-Rad ) . The cytoplasmic HBV capsid particles were resolved in native agarose gel by electrophoresis and transferred onto nitrocellulose membrane , followed by capsid enzyme immunoassay assay ( EIA ) using antibodies against core ( DAKO ) and in situ HBV DNA hybridization as described previously [60] . Whole cell lysate samples prepared by Laemmli buffer was resolved in a 4%-12% gradient SDS-PAGE and proteins were transferred onto Immobilon PVDF-FL membrane ( Millipore ) . The membranes were blocked with Western Breeze blocking buffer ( Life Technologies ) and probed with antibodies against ISG20 [16] , FLAG-tag ( Sigma , clone M2 ) , HA-tag ( Covance , clone 16B12 ) , or β-actin ( Millipore ) . Bound antibodies were revealed by IRDye secondary antibodies . The immunoblot signals were visualized and quantified with the Li-COR Odyssey system . HBeAg and HBsAg in culture fluid were detected by HBeAg ELISA kit ( Autobio Diagnostics ) and HBsAg ELISA kit ( Abazyme ) following the manufacturer’s instructions . HepG2 cells were cotransfected with pCMVHBVΔCΔP and control vector , FLAG-tagged wildtype or D94G mutant ISG20 in 60 mm dishes and maintained for 4 days . The harvested cells were lysed on ice with cell lysis buffer containing 1% NP-40 , 10 mM Tris . HCl ( pH 7 . 5 ) , 1 mM EDTA , 50 mM NaCl , 8% sucrose , and 1 U/μl of RNasin Plus RNase Inhibitor ( Promega ) . After centrifugation to remove the cell debris , the clarified cell lysates were incubated with EZview Red Anti-HA or Anti-FLAG Affinity Gel ( Sigma-Aldrich ) at 4°C for 2 h with gentle rotation . The beads were spun down and resuspended gently with rinse buffer ( 10 mM Tris . HCl ( pH 7 . 5 ) , 1 mM EDTA , 50 mM NaCl , and 1 U/μl of RNasin Plus RNase Inhibitor ) for three times at 4°C . The pelleted beads were subjected to RNA extraction with TRIzol , and protein sample preparation with Laemmli buffer . Immunoprecipitated ISG20 protein and HBV RNA were analyzed by Western blot and Northern blot assays , respectively . In terms of the RNA binding competition assay for ISG20 and HBV Pol , 2 μg of pCMVHBVΔCΔP was cotransfected with either 1 μg of control vector or 1 μg of pCMV-FLAG-Pol and with increased amount of HA-ISG20 D94G as indicated in Fig 8 . FLAG or HA immunoprecipitation was performed with the same protocol described above and the associated HBV RNAs were visualized via Northern blot using 32P radiolabeled HBV RNA specific riboprobe . The ORF of wildtype ISG20 and ISG20D94G mutant were PCR amplified from plasmid F-ISG20 and F-ISG20D94G , respectively , and cloned into pRSET A prokaryotic expression vector ( Thermo Fisher Scientific ) under the T7 promoter at BamHI and EcoRI sites to generate plasmid expressing His-tagged ISG20 ( His-ISG20 ) and ISG20D94G ( His-ISG20D94G ) . ISG20 deletion mutant His-ISG20ΔExoI , His-ISG20ΔExoII , and His-ISG20ΔExoIII were derived from His-ISG20 by using Q5 Site-Directed Mutagenesis Kit ( New England Biolabs ) per the manufacture’s protocol . The protein expression vectors were transformed into Escherichia coli BL21 ( DE3 ) pLysS Competent Cells ( Promega ) and the cells were propagated with aeration at 37°C in 0 . 5 L of SOB broth in the presence of 100 μg/ml Ampicillin to an A600 of ∼0 . 6 , followed by adding 1 mM isopropyl-1-thio-β-D-galactopyranoside ( IPTG ) to induce protein expression at 37°C for 3 hours . The induction of the recombinant proteins were detected via SDS-PAGE and Coomassie staining as compared to an uninduced control sample . The aforementioned His-tagged proteins expressed from bacteria were purified using denaturing conditions . Briefly , the bacterial pellet was resuspended in 0 . 1 M Na-phosphate , 0 . 1 M Tris-HCl , 6 M Guanidine-HCl , pH 8 . 0 with fresh 1 mM PMSF and stirred for 2 h at 4°C to lyse the cells and solubilize the proteins under denaturing condition . The cell extract was centrifuged at 12 , 000×g for 20 min . The supernatant fraction containing soluble protein was incubated in batch with PerfectPro Ni-NTA Agarose ( 5PRIME ) for 1 h at room temperature . The resin was washed once with 0 . 1 M Na-phosphate , 0 . 1 M Tris-HCl , 6 M Guanidine-HCl , pH 6 . 3 and two times with 0 . 1 M Na-phosphate , 0 . 1 M Tris-HCl , 6 M Guanidine-HCl , pH 6 . 3 with additional 20 mM imidazole . The resin was loaded into the supplied purification column , and the protein was eluted with 0 . 1 M Na-phosphate , 0 . 1M Tris-HCl , 6 M Guanidine-HCl , pH 4 . 6 and 300 mM imidazole with 1 mM PMSF . The eluted protein was dialyzed against three changes of 1×PBS , 300 mM NaCl , 5% Glycerol with freshly added 1 mM DTT and 0 . 1 mM PMSF . The soluble proteins were concentrated with Pierce concentrators ( 9KD cut-off , Thermo Scientific ) . Bio-Rad Bradford protein assay and Coomassie staining were used to measure the concentration and the purity of the proteins . The synthetic HBV ε RNA fragments shown in S1 Table were dissolved in 1×TE buffer ( DEPC treated ) to a concentration of 1 μg/μl and denatured in 80°C water bath for 5 min , followed by slow cooling down to room temperature for RNA annealing and secondary structures formation . HBV ε RNAs were [γ-32P] end-labeling by T4 polynucleotide kinase ( New England Biolabs ) and purified by Quick Spin Sephadex G25 column ( Roche ) . The indicated amount of ISG20 proteins were incubated with 100 ng 32P-radiolabeled HBV RNAs in the presence of 20 mM HEPES , 100 mM KCl , 1 mM DTT , 0 . 5 mg/ml BSA , 10% Glycerol and 0 . 05 μg/μl poly[dI-dC] at 30°C for 30 min to form nucleoprotein complexes . 1 μl of monoclonal anti-polyHistidine antibody ( Clone H1029 , Sigma ) was used for supershifting of the His-ISG20/ HBV ε complex . 1 μg , 2 μg or 4 μg of cold unlabeled HBV ε RNAs were used to compete for binding of the ISG20 protein to 100 ng radiolabeled HBV ε . The nucleoprotein complexes were separated by native 5% PAGE at 200 V in a gel buffer containing 50 mM Tris , 45 mM Boric acid , 1% ( v/v ) Glycerol for 2 h in the cold room . The gel was fixed in 10% acetic acid and 10% methanol , dried , and visualized by autoradiography . Expression vector pcDNA-3FHP was used to express 3×FLAG tagged HBV Pol ( 3F-HP ) in TnT Coupled Reticulocyte Lysate System ( Promega ) by following manufacturer’s instructions . The control IVT reaction was done with empty pcDNA3 . 1 vector ( Invitrogen ) . One microliter of control or 3F-HP IVT sample was mixed with 1 μl of 32 P-radiolabeled ε RNA ( 100 ng/μl ) and 18 μl of TMNK RNA binding buffer ( 20 mM Tris [pH 7 . 5] , 2 mM MgCl2 , 15 mM NaCl , 20 mM KCl , and 4 mM DTT ) which has been previously used for HBV pol:ε RNA binding reaction [61] . Then 1% Halt protease inhibitor cocktail ( ThermoFisher ) , 4 μl of RNasin RNase inhibitors ( Promega ) , and 6 μg of yeast tRNA were added and the binding reaction was incubated at 30°C for 1 h . The pol:ε binding complex was revealed by EMSA as described above . The ribonuclease assay was performed according to previously published reaction conditions for ISG20 [13] . Briefly , the synthetic RNA oligos ( S1 Table ) were pretreated as described above to allow secondary structure formation before purification and 5’-32P labeling . 100 ng of radiolabeled RNA substrates were incubated with RNase A ( New England Biolabs ) or purified recombinant His-ISG20 in 10 μl nuclease assay buffer ( 50 mM HEPES-KOH ( pH 7 . 0 ) , 10% glycerol , 50 mM NaCl , 1 mM MnCl2 , 0 . 01% Triton X-100 , and 1 mM DTT ) at 37°C for 15 min . The reactions were stopped with 10 μl of an 80% Formamide dye solution , and the mixtures were heated at 80°C for 3 min and then fractionated in Novex 10% TBE-Urea denaturing polyacrylamide gel ( ThermoFisher ) , and the gel was dried and subjected to autoradiography .
HBV is a DNA virus but replicates its DNA via retrotranscription of a viral RNA pregenome . ISG20 , an antiviral RNase induced by interferons , inhibits the replication of many RNA viruses but the underlying molecular antiviral mechanism remains elusive . Since all the known viruses , except for prions , have RNA products in their life cycles , ISG20 can be a broad spectrum antiviral protein; but in order to distinguish viral RNA from host RNA , ISG20 may have evolved to recognize virus-specific signals as its antiviral target . We demonstrated herein that ISG20 selectively binds to a unique stem-loop structure called epsilon ( ε ) in all HBV RNA species and degrades viral RNA to inhibit HBV replication . Because ε is the HBV pregenomic RNA packaging signal and reverse transcription priming site , the binding of ISG20 to ε , even in the absence of ribonuclease activity , results in antiviral effect to prevent DNA replication due to preventing viral polymerase binding to pgRNA . We also determined the structure and sequence requirements of ε RNA and ISG20 protein for ISG20-ε binding and antiviral activity . Such information will aid the function study of ISG20 against viral pathogens in host innate defense , and ISG20 has potentials to be developed into a therapeutic agent for viral diseases including hepatitis B .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "methods" ]
[ "medicine", "and", "health", "sciences", "pathology", "and", "laboratory", "medicine", "electrophoretic", "mobility", "shift", "assay", "gene", "regulation", "pathogens", "microbiology", "hepatitis", "b", "virus", "plasmid", "construction", "viruses", "dna", "replication", "dna", "construction", "molecular", "biology", "techniques", "dna", "research", "and", "analysis", "methods", "small", "interfering", "rnas", "rna", "structure", "gene", "mapping", "proteins", "medical", "microbiology", "gene", "expression", "microbial", "pathogens", "hepatitis", "viruses", "viral", "replication", "molecular", "biology", "biochemistry", "rna", "nucleic", "acids", "virology", "viral", "pathogens", "genetics", "interferons", "biology", "and", "life", "sciences", "non-coding", "rna", "restriction", "fragment", "mapping", "organisms", "macromolecular", "structure", "analysis" ]
2017
Interferon-inducible ribonuclease ISG20 inhibits hepatitis B virus replication through directly binding to the epsilon stem-loop structure of viral RNA
Since the outbreak of severe acute respiratory syndrome ( SARS ) in 2003 , the three-dimensional structures of several of the replicase/transcriptase components of SARS coronavirus ( SARS-CoV ) , the non-structural proteins ( Nsps ) , have been determined . However , within the large Nsp3 ( 1922 amino-acid residues ) , the structure and function of the so-called SARS-unique domain ( SUD ) have remained elusive . SUD occurs only in SARS-CoV and the highly related viruses found in certain bats , but is absent from all other coronaviruses . Therefore , it has been speculated that it may be involved in the extreme pathogenicity of SARS-CoV , compared to other coronaviruses , most of which cause only mild infections in humans . In order to help elucidate the function of the SUD , we have determined crystal structures of fragment 389–652 ( “SUDcore” ) of Nsp3 , which comprises 264 of the 338 residues of the domain . Both the monoclinic and triclinic crystal forms ( 2 . 2 and 2 . 8 Å resolution , respectively ) revealed that SUDcore forms a homodimer . Each monomer consists of two subdomains , SUD-N and SUD-M , with a macrodomain fold similar to the SARS-CoV X-domain . However , in contrast to the latter , SUD fails to bind ADP-ribose , as determined by zone-interference gel electrophoresis . Instead , the entire SUDcore as well as its individual subdomains interact with oligonucleotides known to form G-quadruplexes . This includes oligodeoxy- as well as oligoribonucleotides . Mutations of selected lysine residues on the surface of the SUD-N subdomain lead to reduction of G-quadruplex binding , whereas mutations in the SUD-M subdomain abolish it . As there is no evidence for Nsp3 entering the nucleus of the host cell , the SARS-CoV genomic RNA or host-cell mRNA containing long G-stretches may be targets of SUD . The SARS-CoV genome is devoid of G-stretches longer than 5–6 nucleotides , but more extended G-stretches are found in the 3′-nontranslated regions of mRNAs coding for certain host-cell proteins involved in apoptosis or signal transduction , and have been shown to bind to SUD in vitro . Therefore , SUD may be involved in controlling the host cell's response to the viral infection . Possible interference with poly ( ADP-ribose ) polymerase-like domains is also discussed . The SARS coronavirus ( SARS-CoV ) is much more pathogenic for humans than any other coronavirus . Therefore , protein domains encoded by the SARS-CoV genome that are absent in other coronaviruses are of particular interest , because they may be responsible for the extraordinary virulence . The most prominent such domain has been identified by bioinformatics as part of non-structural protein 3 ( Nsp3 ) of the virus and appropriately named the “SARS-unique domain” ( SUD ) [1] . With a molecular mass of 213 kDa , Nsp3 is the largest of the non-structural proteins of SARS coronavirus ( see Figure 1 ) . Comprising 1922 amino-acid residues ( polyprotein 1a/1ab residues Ala819 to Gly2740 ) , SARS-CoV Nsp3 is larger than the entire replicase of Picornaviridae . It contains at least seven subdomains [2]: An N-terminal acidic domain ( Ac , also called Nsp3a ) ; an X-domain ( also designated as ADRP , or Nsp3b ) ; the SUD ( Nsp3c ) ; a papain-like proteinase , PL2pro ( also called Nsp3d ) ; and additional domains ( Nsp3e–g ) that include a transmembrane ( TM ) region . At present , it is completely unclear whether and how the individual domains of Nsp3 interact with one another or with other components of the coronaviral replicase complex . Also , some of them possibly recognize proteins of the infected host cell [2] . In the absence of functional data on these domains , attempts have been made to derive their possible biological role from their three-dimensional structures ( see [3] for a review ) . The NMR structure of an N-terminal fragment of the acidic domain ( Nsp3a ) has revealed a ubiquitin-like fold complemented by two additional short α-helices ( [4] , PDB code 2IDY ) . NMR chemical-shift analysis suggested that these non-canonical structural elements might bind single-stranded RNA with some specificity for AUA-containing sequences , although the KD values observed are relatively high ( ∼20 µM ) . Interestingly , a second ubiquitin-like domain occurs in Nsp3 , as part of the papain-like proteinase ( PL2pro , Nsp3d , [5]; PDB code 2FE8 ) . The PL2pro cleaves the viral polyprotein after two consecutive glycine residues to release Nsp1 , Nsp2 , and Nsp3 , respectively ( The remaining cleavage reactions are performed by the coronaviral main proteinase ( Mpro; [6]–[8] ) ) . In addition to its proteolytic activities on the N-terminal third of the polyproteins , the SARS-CoV PL2pro has also been shown to be a deubiquitinating enzyme [9]–[12] . Lindner et al . [13] have shown that in addition to its proteolytic and deubiquitinating activity , the SARS-CoV PL2pro acts as a de-ISGylating enzyme . Induction of ISG15 and its subsequent conjugation to proteins protects cells from the effects of viral infection [14] , [15] . Since the ISG15 gene is induced by interferon as part of the antiviral response of the innate immune system , the de-ISGylation activity of Nsp3d could explain the suppression of the interferon response by the papain-like protease , in addition to a possible direct interaction between the PL2pro and IRF3 [16] . Among the subdomains of the Nsp3 multidomain protein , there is also the so-called “X- domain” ( Nsp3b ) , which shows structural homology to macrodomains . The latter name refers to the non-histone-like domain of the histone macro2A [17]–[19] . In animal cells , such domains are occasionally physically associated with enzymes involved in ADP-ribosylation or ADP-ribose metabolism . Because of this linkage and on the basis of sequence similarity to Poa1p , a yeast protein involved in the removal of the 1″-phosphate group from ADP-ribose 1″-phosphate ( a late step in tRNA splicing; [20] ) , it has been proposed that the coronaviral X-domains may have the function of ADP-ribose-1″-phosphatases ( ADRPs; [21] ) . The crystal structures of X-domains of SARS-CoV [22] , [23] as well as of HCoV 229E and Infectious Bronchitis Virus ( IBV ) [24] show that the protein has the three-layer α/β/α fold characteristic of the macrodomains . Embedded between the X-domain ( Nsp3b ) and the PL2pro ( Nsp3d ) , the SARS-unique domain ( SUD; Nsp3c ) fails to show sequence relationship to any other protein in the databases [1] . We have produced full-length SUD ( residues 389 to 726 of Nsp3 ) , and a more stable , shortened 264-residue version ( residues 389 to 652; henceforth called SUDcore ) , by expression in Escherichia coli . This definition of the boundaries of the SUD is based on the structural results described here . We report crystallization of SUDcore and its X-ray structure in two crystal forms , at 2 . 2 and 2 . 8 Å resolution , respectively . The structure turns out to consist of two further copies of the macrodomain , in spite of the complete absence of sequence similarity . In addition , we demonstrate that each of the subdomains binds G-quadruplexes , both in DNA and RNA fragments , and that selected mutations of lysine residues in the first subdomain , SUD-N , lead to reduced nucleic-acid binding , whereas those in the second subdomain , SUD-M , abolish it . Out of the many SUD constructs designed and tested by us , SUDcore ( Nsp3 residues 389–652 ) turned out to be relatively stable and could be crystallized ( Table 1 ) . Two crystal forms were observed under identical crystallization conditions: Form-1 crystals were monoclinic ( space group P21 , two SUDcore molecules per asymmetric unit ) and diffracted X-rays to 2 . 2 Å resolution; form-2 crystals were triclinic ( space group P1 , four SUDcore molecules per asymmetric unit ) and diffracted to 2 . 8 Å . Both structures were determined by molecular replacement ( see Materials and Methods ) . The r . m . s . deviations ( on Cα atoms ) between the models derived from the two different crystal structures are around 0 . 7 Å . The models have good stereochemistry ( Table 1 ) . 94 . 7% of the amino-acid residues are in the favoured regions of the Ramachandran plot and 4 . 6% are in allowed regions . 0 . 6% are outliers . In all six independent copies of the SUDcore monomer , residue Val611 adopts forbidden conformational angles . This residue is located in a turn described by the polypeptide chain where it leaves the subdomain interface ( see below ) and reaches the surface of the molecule . The side chain makes a hydrophobic contact across the subdomain interface and is also contacting the side chain of Phe406 of a symmetry-related SUDcore dimer in the crystal lattice in the monoclinic crystal form ( this also applies to two of the four monomers in the triclinic form ) . SUDcore exhibits a two-domain architecture ( Figure 2A ) . The N-terminal subdomain ( SUD-N ) comprises Nsp3 residues 389–517 , and the C-terminal subdomain of SUDcore contains residues 525–652 . We call the latter the “middle SUD subdomain” , or SUD-M , because full-length SUD has a C-terminal extension of 74 residues compared to SUDcore . The SUD-N and SUD-M subdomains have a similar fold and can be superimposed with an r . m . s . d . of 3 . 3–3 . 4 Å ( based on Cα positions ) ; they share 11% sequence identity ( see Figure 2C for a structural alignment ) . Of the 14 amino-acid residues identical between the two subdomains , four form a conserved Leu-Glu-Glu-Ala motif at the N-terminus of helix α4 . The linker between the two subdomains ( residues 518–524 ) has no visible electron density . This is due to elevated mobility of the linker , rather than proteolytic cleavage , since we showed by SDS-PAGE of dissolved crystals that the SUDcore polypeptide ( in the presence of β-mercaptoethanol ) has the apparent molecular mass to be expected ( ∼29 kDa; not shown ) . In addition to the linker , SUD-N and SUD-M are connected by a disulfide bond between cysteines 492 and 623 ( Figure 2B ) . Disulfide bonds are rare in cytosolic proteins , but in coronaviral Nsps , examples of such bonds have been reported [25] , [26] . The fold of each SUD subdomain is that of a macrodomain ( Figure 2A ) . Macrodomains consist of a largely parallel central β-sheet surrounded by 4–6 α-helices . The order of regular secondary-structure elements in SUD-N is βN1-αN1-βN2-αN2-βN3-βN4-αN3-βN5-αN4-βN6 , and in SUD-M βM1-αM1-βM2-αM2-βM3-βM4-αM3-βM5-αM4-βM6-αM5 . The topology of the β-strands is β1–β6–β5–β2–β4–β3 , all of which are parallel except β3 ( Figure 2A ) . Between the two subdomains , most of the secondary-structure elements are conserved with respect to their position in the three-dimensional structure , although they often differ in length . This is particularly obvious for α-helix 1 , which comprises just four residues in the N-terminal subdomain but eleven in the M subdomain . Similarly , α-helix 2 has 5 vs . 10 amino-acid residues in the two subdomains . In general , the strands of the central β-sheet appear to align better between the two subdomains than do the α-helices . Each of the SUDcore subdomains is related to the macrodomain of the histone macro2A ( [18]; PDB code 1ZR3 , molecule C; for SUD-N: Z-score 9 . 8 , r . m . s . d . 2 . 5 Å for 112 out of 184 Cα atoms , 12% sequence identity; for SUD-M: Z-score 8 . 6 , r . m . s . d . 2 . 8 Å for 115 out of 184 Cα atoms , 19% sequence identity ) . Called “X-domains” , single macrodomains are also found in alphaviruses , in hepatitis E virus , and in rubella virus , in addition to coronaviruses [27] , [28] . The SARS-CoV X-domain ( Nsp3b ) , the domain immediately preceding the SUD in Nsp3 , shares no recognizable sequence identity with SUD-N ( 12% ) or SUD-M ( 7% ) ( Figure 2C ) , but its three-dimensional structure [22] , [23] ( PDB code 2ACF , chain A ) can be superimposed onto each of the two SUD subdomains with an r . m . s . d . ( based on Cα atoms ) of 2 . 7 Å and 2 . 3 Å , respectively ( Figure 2D ) . Thus , within Nsp3 , SARS-CoV has three macrodomains aligned one after the other . In both crystal forms , SUDcore displays the same head-to-tail dimer , with the SUD-N subdomain of monomer A interacting with the SUD-M subdomain of monomer B , and vice versa . Approximately 1130 Å2 of solvent-accessible surface per monomer is buried upon dimerization ( Figure 3 ) . Due to the two-domain architecture of each monomer , the resulting four lobes give the dimer a quasi-tetrahedral shape ( Figure 3A ) . Involving ∼10 hydrogen bonds and four well-defined salt-bridges ( AspB440…ArgA554 , ArgB473…GluA619 , ArgB554…AspA440 , and GluB619…ArgA473 ) , interactions between the monomers are largely hydrophilic . As to be expected , the structures of the monomers are very similar to one another , with r . m . s . d . values ( for Cα atoms ) of 0 . 58 Å between monomers A and B of the monoclinic crystal form , and 0 . 11–0 . 37 Å between monomers A–D of the triclinic form . The structure of SUD-M alone is even better conserved between the individual copies of SUDcore . Also , the fold of the SUD-M subdomain is similar to the model of the SUD fragment 527–651 derived from NMR measurements , which was published very recently ( r . m . s . d . ∼0 . 9 Å ) [29] . The function of the coronaviral X-domain is still unclear; for some coronaviruses such as HCoV 229E and SARS-CoV , it has been shown to exhibit a low ADP-ribose-1″-phosphate phosphatase ( Appr-1″-pase , occasionally also called “ADRP” ) activity and to bind the product of the reaction , ADP-ribose [21]–[23] , [30] . However , the two subdomains of SUDcore do not bind ADP-ribose , as we have demonstrated by zone-interference gel electrophoresis ( Figure S1 ) . When we investigated possible interactions between SUD and nucleic acids by zone-interference gel electrophoresis , we found that the domain binds oligo ( G ) and oligo ( dG ) stretches with a KD of ∼1 µM , but not oligo ( dA ) , ( dC ) , or ( dT ) [31] . Single-stranded nucleotides of random sequence are only bound if they are longer than ∼15 nucleotides . Here we demonstrate that each of the two individual SUD subdomains also binds oligo ( dG ) ( Figure 4A ) . With oligo ( dH ) , where H stands for A , C , or T , but not G , only very small gel shifts , if at all , were observed . As oligo ( G ) stretches are known to form G-quadruplexes , i . e . four-stranded nucleic-acid structures formed by contiguous guanines [32] , we also examined the binding to the oligodeoxynucleotide 5′-GGGCGCGGGAGGAATTGGGCGGG-3′ , a G-rich sequence present in the bcl-2 promoter region . This oligonucleotide has been shown by NMR spectroscopy to form a G-quadruplex ( [33]; PDB code 2F8U ) . We found that both full-length SUD and SUDcore do indeed bind this oligodeoxynucleotide and that this process is enhanced by the addition of K+ ions , which are known to stabilize G-quadruplex structures ( Figure 4B ) . In agreement with the ability of SUD to non-specifically bind to oligonucleotides of >15 bases [31] , both SUD and SUDcore were found to bind the reverse-complementary sequence , but with low affinity and , more importantly , independent of K+ ions . As there is no evidence for SARS-CoV Nsp3 entering the nucleus and binding to DNA , we examined whether SUD would bind to an RNA known to form a quadruplex structure . Indeed , zone-interference gel shift experiments revealed major shifts for both SUD and SUDcore in the presence of the oligoribonucleotide 5′-UGGGGGGAGGGAGGGAGGGA-3′ , which is a protein-binding element in the 3′-nontranslated region of chicken elastin mRNA [34] and forms G-quadruplexes [35] ( Figure 4C ) . Furthermore , we observed a significant gel shift for SUDcore when we added the short oligonucleotide UGGGGU , which has also been shown to form a G-quadruplex ( [36]; PDB code 1J8G ) . This shift was also enhanced by the addition of K+ ( Figure 4D ) . Thus , SUD binds RNA ( rG ) -quadruplexes and DNA ( dG ) -quadruplexes with comparable affinity . Inspection of the structure of the SUD dimer reveals a central narrow cleft running across the dimer surface , but distinct from the monomer-monomer interface ( Figure 3C ) , which could be a binding site for another protein . In addition , there are several positively charged patches in the center of the dimer ( Figure 3B ) , and on its backside ( Figure 3C ) , which could be involved in binding to G-quadruplexes . We have prepared four sets of mutations by replacing lysine residues ( and one glutamate ) in these patches by alanines . The first two pairs of mutations , K505A+K506A ( M1 , at the end of helix αN4 ) and K476A+K477A ( M2 , in the loop between αN3 and βN5 ) , are located on the surface of the SUD-N subdomain and lead to reduced shifts with G-quadruplexes in the zone-interference gel electrophoresis experiment , both with the G-quadruplex from the bcl-2 promoter region ( Figure 5 ) and with ( dG ) 10 ( not shown ) . The second set of mutations , K563A+K565A+K568A ( M3 ) and K565A+K568A+E571A ( M4 ) are located in the loop connecting αM2 and βM3 of the SUD-M subdomain and abolish G-quadruplex binding altogether ( Figure 5 ) , again with both oligonucleotides . When the SARS-unique domain was first predicted [1] , the boundaries of the domain were set approximately at Nsp3 residues 352 and 726 . We made major efforts to produce this protein in a stable form , but with little success . Only when we used in-vitro protein synthesis , were we able to obtain small amounts of a relatively stable preparation comprising Nsp3 residues 349–726 [31] . At the N-terminus of this construct , up to eleven residues actually correspond to the C-terminus of the preceding X-domain ( Nsp3b ) . When we expressed a gene construct coding for SUD ( 349–726 ) in E . coli , we observed rapid proteolytic degradation of the N-terminal segment . The relatively stable intermediate obtained had its N-terminus at Nsp3 residue 389 . The N-terminal segment ∼359–388 is predicted to be intrinsically unfolded by several prediction programs ( not shown ) . Therefore , we assume segment 359–388 to be merely a linker between Nsp3b and SUD , and 389 to be the first residue of the latter . This assignment is justified by the observation that in our crystal structures reported here , the SUD-N subdomain is a complete macrodomain without any residues lacking at the N-terminus . Therefore , the protein corresponding to Nsp3 residues 389–726 is called “full-length SUD” here . In this communication , we describe the crystal structures at 2 . 2 Å and 2 . 8 Å resolution ( monoclinic and triclinic form , respectively ) of the core of the SARS-unique domain ( SUDcore , Nsp3 residues 389–652 ) . SUDcore turns out to consist of two subdomains , SUD-N ( Nsp3 residues 389–517 ) and SUD-M ( 525–652 ) , each exhibiting the fold of a macrodomain . The two subdomains are connected by a flexible linker ( residues 518–524 ) and a disulfide bond . Even though coronavirus replication occurs in the cytosol , where the environment is reductive , it is unlikely that the formation of this disulfide is an artifact owing to handling of the protein: As the linker between the SUD-N and SUD-M subdomains is very short ( seven residues ) , and the mutual orientation of the subdomains is fixed due to the tight dimerization , cysteine residues no . 492 and 623 will be very close to one another irrespective of the exact conformation of the linker . In fact , disulfide bonds are not uncommon in coronaviral non-structural proteins ( Nsps ) involved in RNA replication or transcription . Among others , they have been observed in HCoV-229E Nsp9 [25] and turkey coronavirus Nsp15 [26] , but in these cases , the disulfide bond connects two subunits of the homo-oligomeric proteins , whereas the occurrence in SUDcore is the first case of an intramolecular disulfide bond described in a coronavirus Nsp . Coronavirus replication in the perinuclear region of the cell is localized to double-membrane vesicles that have been hijacked from the endoplasmic reticulum or late endosomes [37]–[40] . These vesicles are around 200–350 nm in diameter and present alone or as clusters in the cytosol [38] . The milieu inside or at the surface of these vesicles is unknown , but it is well possible that it is partially oxidative . It has also been speculated [25] that formation of disulfide bonds may be a way for the coronaviral Nsps to function in the presence of the oxidative stress that is the consequence of the viral infection [41]–[43] . Our identification of two macrodomains in SUDcore brings the number of these domains in SARS-CoV Nsp3 to three . What are the functions of these modules ? The original SARS-CoV “X-domain” ( Nsp3b ) has been shown to have low ADP-ribose-1″-phosphate phosphatase ( Appr-1″-pase or “ADRP” ) activity [21]–[23] . However , this assignment is controversial . A nuclear Appr-1″-pase ( Poa1p in yeast , [20] ) is an enzyme of a tRNA metabolic pathway , but there is no evidence for coronavirus Nsp3 ever being translocated to the nucleus , and the other enzymes involved in this pathway are missing in coronaviruses ( with the exception of the cyclic 1″ , 2″-phosphodiesterase ( CPDase ) in group 2a viruses such as Mouse Hepatitis Virus , Bovine Coronavirus , and Human Coronavirus OC43 ) . Therefore , it has been proposed that the X-domain may be involved in binding poly ( ADP-ribose ) , a metabolic product of NAD+ synthesized by the enzyme poly ( ADP-ribose ) polymerase ( PARP; [23] ) . However , we have recently demonstrated that the X-domain of Infectious Bronchitis Virus ( IBV ) strain Beaudette , a group-3 coronavirus , does not have significant affinity to ADP-ribose [24] . This can be explained on the basis of crystal structures: In the X-domain ( Nsp3b ) of SARS-CoV [23] , and in that of HCoV 229E [24] , a stretch of three conserved glycine residues is involved in binding the pyrophosphate unit of ADP-ribose , whereas in the corresponding domain of IBV strain Beaudette ( but not in all IBV strains , see [44] ) , the second glycine is replaced by serine , leading to steric interference with ADP-ribose binding [24] . In the two SUD subdomains , the triple-glycine sequence is not conserved ( see Figure 2C ) , and hence , they do not bind ADP-ribose either . Neuman et al . [2] reported that full-length SUD binds cobalt ions , whereas a domain called SUD-C by these authors , which is however almost identical ( residues 513–651 ) to our SUD-M ( 525–652 ) , does not . From this , they concluded that the metal-binding activity is associated with the cysteine residues in the N-terminal subdomain . We were also able to observe binding of cobalt ions to SUDcore by following the occurrence of a peak at 310 nm in the UV spectrum , which , in contrast to the data presented by Neuman et al . [2] , could be reverted by addition of zinc ions . However , when we removed the N-terminal His-tag , this phenomenon could no longer be observed . Furthermore , we note that of the four cysteine residues in the SUD-N subdomain ( residues 393 , 456 , 492 , and 507 ) , 456 and 507 are non-accessible in the interior of the subdomain , and 492 is involved in the buried disulfide bond to Cys623; therefore , Cys393 and perhaps the solvent-exposed His423 would remain the only potential ligands for cobalt ions in SUD-N . However , these residues are >12 Å apart and thus unlikely to chelate cobalt ions . For SUD-M , a recent publication [29] reported binding to oligo ( A ) . However , we fail to observe this ( Figure 4A , lane labeled “A” ) . Instead , we have demonstrated that full-length SUD and SUDcore bind oligodeoxynucleotides and oligoribonucleotides that form G-quadruplexes . For full-length SUD and SUDcore , we had previously shown binding to oligo ( dG ) and oligo ( G ) stretches [31] , but the demonstration here of oligo ( dG ) binding to the individual SUDcore subdomains , SUD-N and SUD-M , is unexpected because their overall electrostatic properties are very different from one another: SUD-N is acidic ( pI = 5 . 3 ) , whereas SUD-M is basic ( pI = 9 . 0 ) . However , even SUD-N shows surface patches with positive electrostatics that could bind nucleic acid ( Figure 3B ) . We have used automatic docking procedures to place the G-quadruplex found in the bcl-2 promoter region ( [33]; PDB code 2F8U ) into our crystal structures . One potential binding site identified is in the cleft between the SUD-M and the SUD-N subdomains within the SUDcore dimer ( Figure S2A ) ; this binding site is spatially close to the mutations M3 and M4 , consistent with the observation that these mutations abolish binding completely . However , we have previously shown by Dynamic Light-Scattering that G-quadruplex binding leads to oligomerization of SUDcore [31] . Consequently , we have also constructed models based on the packing modes of SUDcore dimers observed in our crystal structures . One potential binding site for G-quadruplexes might be in a cleft between two consecutive SUDcore dimers as they occur in both the monoclinic and triclinic crystal forms ( Figure S2B ) , but for confirmation , any of these models will have to await crystallographic determination of the complex . In summary , our mutation experiments demonstrate an involvement of several of the many lysine residues of SUD in binding G-quadruplexes , but as it is probably extended surfaces of SUDcore oligomers that participate in this process , it is not possible to pinpoint any single amino-acid residue . The target of SUD binding could be G-quadruplexes in RNA of viral or/and cellular origin . The SARS-CoV genome contains three G6-stretches ( one on the plus-strand and two on the minus-strand ) and an additional two G5-sequences , which could perhaps form local G-quadruplexes . However , the G-stretch binding capabilities of SUD and SUDcore seem to have been optimized for recognition of longer G-rich sequences . By systematic variation of the length of oligo ( dG ) , we found that SUDcore exhibits strongest affinity ( KD ∼0 . 45 µM ) for ( dG ) 10 to ( dG ) 14 [31] . The 3′-nontranslated regions of several host-cell mRNAs coding for proteins involved in the regulation of apoptosis and in signaling pathways contain long G-stretches and could also be targets of SUD . Examples of such mRNAs are those coding for the pro-apoptotic protein Bbc3 [45] , RAB6B ( a member of the Ras oncogene family , [46] ) , MAP kinase 1 [47] , and TAB3 , a component of the NF-κB signaling pathway [48] . It is conceivable that these proteins might be targets for the virus when interfering with cellular signaling . Changes in the stability and/or translation efficiency of these mRNAs due to the binding of a viral regulatory factor could result in an altered reaction of the infected cell to apoptotic signals , or it could silence the antiviral response . The idea that coronaviral X-domains might function as modules binding poly ( ADP-ribose ) [23] received support from the observation that some macrodomains are connected with domains showing poly ( ADP-ribose ) polymerase ( PARP ) activity , i . e . in the so-called macroPARPs ( PARP-9 and PARP-14 ) [49] . There are 18 human genes for members of the PARP family; the prototype enzyme , PARP-1 , catalyzes the post-translational modification of many substrate proteins , including itself , in a multitude of cellular processes ( DNA repair , transcriptional regulation , energy metabolism , and apoptosis ) [50]–[52] . Interestingly , SUD-M and the C-terminal 74-residue subdomain ( SUD-C ) that is missing in our SUDcore construct together show a ∼15% sequence identity ( 32% similarity ) to the catalytic domain of PARP-1 . However , the three-dimensional structures of SUD-M ( this work ) and the C-terminal domain of PARP-1 [53] are different and cannot be superimposed . Another feature common between SARS-CoV SUD and PARP-1 is that the latter has recently been shown to bind to G-quadruplexes [54] , although it is generally assumed that this occurs through the DNA-binding domain rather than the catalytic domain of PARP-1 . PARP-1 and most of its family members are located to the nucleus , while PARP-4 and others predominantly act in the cytoplasm [50]–[52] . PARP-4 is incorporated into vaults , RNA-containing subcellular particles in the cytoplasm [55] . Furthermore , ZAP , a human antiviral protein comprising a C-terminal PARP-like domain devoid of catalytic activity , has been shown to exhibit antiviral activity on alphaviruses [56] , which contain an X-domain similar to that of coronaviruses [23] , [27] , [28] . In addition , ZAP contains an N-terminal zinc-finger domain , a central TiPARP ( 2 , 3 , 7 , 8-tetrachlorodibenzo-p-dioxin ( TCDD ) -inducible PARP ) domain , and a WWE domain ( a protein-protein interaction module in ubiquitin and ADP-ribose conjugation proteins ) . In fact , ZAP appears to be part of the human innate immune system and to play a role comparable to APOBEC3G in HIV infection [57] . It is possible that this group of viruses has evolved macrodomains to counteract the antiviral activity of ZAP . Indeed , macrodomains can inhibit PARPs , as has been shown for the macrodomain of the histone mH2A1 . 1 , which downregulates the catalytic activity of PARP-1 [58] . Having three macrodomains at its disposal , SARS-CoV may be much more efficient in knocking down the antiviral response of the host cell than other coronaviruses . Whether this involves a direct interaction between SUD and ZAP or another member of the PARP family , or competition for G-quadruplexes in viral or host-cell RNA , remains to be shown . Full-length SUD ( Nsp3 residues 389–726 ) and the fragment SUDcore ( Nsp3 residues 389–652 , previously called “SUDc5b” ) of SARS-CoV strain TOR2 ( acc . no . AY274119 ) were produced recombinantly in E . coli as described [31] . The coding regions for the SUD-N subdomain ( Nsp3 residues 389–524 ) and the SUD-M subdomain ( Nsp3 residues 525–652 ) were constructed by introducing an appropriate deletion into the previously described plasmid pQE30-Xa-c5b [31] using site-directed mutagenesis . Plasmids encoding SUD-N and SUD-M were prepared using primers listed in Table S1 . The coding regions for four sets of mutations of SUDcore , M1 ( K505A+K506A ) , M2 ( K476A+K477A ) , M3 ( K563A+K565A+K568A ) , and M4 ( K565A+K568A+E571A ) , were constructed by introducing appropriate mutations into plasmid pQE30-Xa-c5b [31] using site-directed mutagenesis . Plasmids encoding these mutants were prepared using primers also listed in Table S1 . All plasmids provided an N-terminal His-tag and a short linker sequence encoding a factor-Xa cleavage site . The coding regions of the expression plasmids were verified by DNA sequencing . E . coli M15 ( pRep4 ) was used as expression host for these constructs . SUD-N , SUD-M , and the mutated proteins were purified according to the same protocol as for SUDcore [31] . SUDcore displayed >95% purity in SDS-PAGE , and monodispersity in Dynamic Light- Scattering . Initial crystallization screening was performed using the sitting-drop vapor-diffusion method in 96-well Intelli-Plates ( Dunn Laboratories ) . Several commercial kits ( Sigma , Jena Bioscience ) were used for the screening . The protein concentration was 6 mg/ml . Using a Phoenix robotic system ( Art Robbins ) , drops were made of 260 nl protein and 260 nl precipitant solution . The optimized crystallization condition consisted of 20% polyethylene glycol monomethyl ether 5000 and 0 . 2 M ammonium sulfate in 0 . 1 M morpholinoethane sulfonic acid ( pH 6 . 5 ) . Plate-like crystals grew in 3–5 days , to maximum dimensions of 0 . 02×0 . 02×0 . 01 mm3 . Many SUDcore crystals had to be tested for diffraction until one yielding data to 2 . 2 Å resolution was found . The best diffracting crystals belonged to space group P21 . Under the same crystallization conditions , a second crystal form belonging to space group P1 was observed , diffracting to lower resolution of about 2 . 8 Å . Crystals were cryoprotected in reservoir solution that included 30% glycerol , and were harvested into a loop prior to flash-cooling in liquid nitrogen . All data were collected at 100 K from a single crystal each at beamline BL14 . 2 , BESSY ( Berlin , Germany ) , using an MX225 CCD detector ( Rayonics ) , or at beamline I911-2 at MAX-lab ( Lund , Sweden ) , using a Mar165 CCD detector ( Marresearch ) . Data were processed with MOSFLM [59] , and reduced and scaled using the SCALA [60] program from the CCP4 suite [61] . Crystals belonging to space group P21 had unit-cell parameters a = 46 . 36 Å , b = 68 . 55 Å , c = 94 . 21 Å , β = 99 . 17° , those belonging to space group P1 had unit-cell parameters a = 68 . 68 Å , b = 75 . 52 Å , c = 80 . 54 Å , α = 77 . 16° , β = 75 . 61° , γ = 74 . 48° . Data-collection statistics for both crystal forms are shown in Table 1 . The asymmetric unit of the P21 form contained two SUDcore monomers , giving a Matthews coefficient [62] of 2 . 5 Å3 Da−1 and a solvent content of 51% , whereas the P1 crystal form had four monomers per asymmetric unit , giving corresponding parameters of 3 . 2 Å3 Da−1 and 63% . We attempted to solve the structure by molecular replacement into the P21 form using the NMR coordinates of a subdomain comprising SARS-CoV Nsp3 residues 513–651; PDB code 2JWJ [29] , [63] ) , which is almost identical to the SUD-M subdomain of SARS-CoV Nsp3 . Using the program Phaser [64] , [65] , we could find two solutions , and the C-terminal part of SUDcore was well defined in the electron-density maps . However , for the N-terminal half , only a few segments of poly ( Ala ) chain could be built into the maps . This starting model was then refined in BUSTER-TNT [66] using Local Structure Similarity Restraints ( LSSR ) [67] as non-crystallographic symmetry ( NCS ) restraints to give R and Rfree values of 0 . 453 and 0 . 479 , respectively . The resulting 2mFo-DFc electron density was subjected to density modification using solvent flattening , histogram matching , and 2-fold NCS-averaging using DM [68] . The averaging masks were calculated and updated using the auto-correlation procedure [69] as implemented in DM . Using the automatic building program BUCCANEER [70] together with REFMAC [71] ( as implemented in the CCP4i [72] interface for CCP4 ) in an iterative procedure for 20 cycles resulted in a model for 501 residues in 10 chains ( the longest having 208 residues ) , in which 448 residues were assigned both a chemical identity and a sequential residue number , while the remaining 53 residues were modeled as poly ( Ala ) in 8 shorter chains . The R and Rfree values resulting from REFMAC were 0 . 374 and 0 . 414 , respectively . This model was refined in BUSTER-TNT , again using LSSR as NCS restraints for the common parts in the already sequenced 448 residues of the dimer , to R and Rfree values of 0 . 269 and 0 . 316 . The improved electron density was again subjected to density modification using DM as detailed above , but using a lower solvent content of 35% as well as anisotropically scaled observed amplitudes as output by BUSTER-TNT . The resulting density-modified and NCS-averaged map was then used for automatic model building using the iterative BUCCANEER/REFMAC procedure described above . This produced a model with 511 residues in 5 chains with 487 residues sequenced . The R and Rfree values from REFMAC for this model were 0 . 289 and 0 . 326 , respectively . Since the refinements in BUSTER-TNT at that point showed some problematic low correlations between Fo and Fc at low resolution , the original images collected from the P21 crystal were reprocessed using XDS [73] and SCALA , applying different high-resolution cutoffs for different segments of the collected images . Details for this dataset are given in Table 1 . Subsequent refinement of the P21 form with REFMAC , under application of weak NCS restraints , yielded a model with R = 0 . 211 , Rfree = 0 . 264 . The advanced handling of NCS restraints through LSSR in BUSTER-TNT gave a final model R = 0 . 211 and Rfree = 0 . 268 . The final model in the P21 form comprises 513 residues ( A389–A516; A524–A652; B393–B519; B526–B652 ) . Chain A of the P21 form was used for molecular replacement with the program MOLREP [74] into the P1 form . There was an unambiguous solution for four molecules in the asymmetric unit . This model was refined with BUSTER-TNT ( using LSSR for NCS restraints ) and rebuilt in Coot [75] to final values of R = 0 . 223 and Rfree = 0 . 240 . The final model of the P1 form comprises 1014 residues . The figures were made with PyMOL [76] . The zone-interference gel electrophoresis ( ZIGE ) device was adapted from Abrahams et al . [77] . ZIGE assays were performed using a horizontal 1% agarose gel system in TBE buffer ( 20 mM Tris , 50 mM boric acid , 0 . 1 mM ethylenediaminetetraacetic acid ( EDTA ) , pH 8 . 3 ) . The protein was incubated at room temperature for 30 min with different concentrations of oligodeoxynucleotides , such as ( dG ) 10 and bcl-2 promoter region ( 5′-GGGCGCGGGAGGAATTGGGCGGG-3′ ) , or oligoribonucleotides ( 5′-UGGGGGGAGGGAGGGAGGGA-3′ and 5′-UGGGGU-3′ ) . The samples were mixed with dimethylsulfoxide ( DMSO; final concentration 10% ( v/v ) ) and a trace of bromophenolblue ( BPB ) . These protein-oligonucleotide samples were applied to the small slots . Oligonucleotide with the same concentration as in the small slots was also mixed with DMSO and BPB in 1xTBE buffer and applied to the long slots of the gel ( total volume 100 µl ) . Electrophoresis was performed at 4°C for 1 h with a constant current of 100 mA . Staining was performed as outlined in [77] . Protein Data Bank: Coordinates and structure factors have been deposited with accession code 2W2G ( P21 crystal form ) and 2WCT ( P1 crystal form ) .
The genome of the SARS coronavirus codes for 16 non-structural proteins that are involved in replicating this huge RNA ( approximately 29 kilobases ) . The roles of many of these in replication ( and/or transcription ) are unknown . We attempt to derive conclusions concerning the possible functions of these proteins from their three-dimensional structures , which we determine by X-ray crystallography . Non-structural protein 3 contains at least seven different functional modules within its 1922-amino-acid polypeptide chain . One of these is the so-called SARS-unique domain , a stretch of about 338 residues that is completely absent from any other coronavirus . It may thus be responsible for the extraordinarily high pathogenicity of the SARS coronavirus , compared to other viruses of this family . We describe here the three-dimensional structure of the SARS-unique domain and show that it consists of two modules with a known fold , the so-called macrodomain . Furthermore , we demonstrate that these domains bind unusual nucleic-acid structures formed by consecutive guanosine nucleotides , where four strands of nucleic acid are forming a superhelix ( so-called G-quadruplexes ) . SUD may be involved in binding to viral or host-cell RNA bearing this peculiar structure and thereby regulate viral replication or fight the immune response of the infected host cell .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "biochemistry", "biochemistry/rna", "structure", "virology/viral", "replication", "and", "gene", "regulation", "biophysics/structural", "genomics", "biophysics/protein", "folding", "virology/virulence", "factors", "and", "mechanisms", "virology/emerging", "viral", "diseases", "biophysics/rna", "structure", "biochemistry/structural", "genomics", "biophysics" ]
2009
The SARS-Unique Domain (SUD) of SARS Coronavirus Contains Two Macrodomains That Bind G-Quadruplexes
The acquisition of the external genitalia allowed mammals to cope with terrestrial-specific reproductive needs for internal fertilization , and thus it represents one of the most fundamental steps in evolution towards a life on land . How genitalia evolved remains obscure , and the key to understanding this process may lie in the developmental genetics that underpins the early establishment of the genital primordium , the genital tubercle ( GT ) . Development of the GT is similar to that of the limb , which requires precise regulation from a distal signaling epithelium . However , whether outgrowth of the GT and limbs is mediated by common instructive signals remains unknown . In this study , we used comprehensive genetic approaches to interrogate the signaling cascade involved in GT formation in comparison with limb formation . We demonstrate that the FGF ligand responsible for GT development is FGF8 expressed in the cloacal endoderm . We further showed that forced Fgf8 expression can rescue limb and GT reduction in embryos deficient in WNT signaling activity . Our studies show that the regulation of Fgf8 by the canonical WNT signaling pathway is mediated in part by the transcription factor SP8 . Sp8 mutants elicit appendage defects mirroring WNT and FGF mutants , and abolishing Sp8 attenuates ectopic appendage development caused by a gain-of-function β-catenin mutation . These observations indicate that a conserved WNT-SP8-FGF8 genetic cassette is employed by both appendages for promoting outgrowth , and suggest a deep homology shared by the limb and external genitalia . Development of the external genitalia is a crucial aspect of mammalian evolution that enables internal fertilization , a pivotal step towards land invasion . All therian mammals including metatherians develop external genitalia around their urogenital outlets . In mice , development of the embryonic anlage of external genitalia , the genital tubercle ( GT ) , is identical in both sexes before androgen-mediated penile masculinization which occurs around embryonic day 16 . The early androgen-independent phase of GT development is achieved through coordinated growth and patterning of cloacal endoderm-derived urethral epithelium ( UE ) , mesoderm-derived para-cloacal mesenchyme ( PCM ) and the ventral ectoderm , which results in a cone-like structure with a ventral-medial positioned urethra surrounded by GT mesenchyme within an ectodermal epithelial capsule . The development of the GT as an unpaired body appendage , is often compared to that of the paired-type appendages , the limbs [1] , [2] . Despite their anatomical and functional differences , the morphogenesis of both structures appears to involve similar genetic controls . Regulatory genes/pathways including canonical WNT signaling [3] , [4] , HH signaling [5]–[7] , BMP signaling [8] , [9] and Hox genes [10]–[14] are essential for the development of both appendages . Some have suggested that the regulatory mechanisms common to both might be evolutionarily linked [11] , [15] , [16] . The first step in appendage outgrowth is the establishment of an independent proximodistal developmental axis apart from the main body axis . This process requires precise regulation from instructive signals , which often come from a distal signaling center . In addition to promoting and maintaining outgrowth , these signals also provide directional information that determines the orientation and shape of future structures . Moreover , genes required for subsequent patterning and differentiation are often regulated by the distal signaling center . For example , during limb development , the initiation and continuous outgrowth of the limb bud rely on growth factors secreted from a strip of ectodermal epithelium , termed the apical ectodermal ridge ( AER ) , positioned at the distal edge of limb bud ( Figure 1A and 1E ) . Fibroblast Growth Factors ( FGFs ) are crucial signals provided by the AER , as FGF-soaked beads can replace the AER to induce limb outgrowth [17]–[20] . Furthermore , the AER FGF signals are obligatory to maintain a positive feedback loop involving SHH and Gremlin that coordinates patterning and growth of the limb [21]–[23] . The early GT is built by two mesenchymal swellings at either side of the cloacal membrane , which is later joined by a third outgrowth anterior to the cloacal membrane ( Figure 1B–1D ) . Unlike limb , subsequent outgrowth of the GT has to accommodate a continuous extension of the cloacal endoderm , which forms the epithelial lining of the future urethral tube . This unique pattern of GT development suggests that the centrally located distal cloacal endoderm ( later the distal urethral epithelium or dUE , marked in red in Figure 1B–1D , also shown by Fgf8 in situ in Figure 1F–1H ) is a strategic place for GT outgrowth . Consistently , an earlier study demonstrated that Fgf8 is expressed in a strip of cells located at the distal-most part of the cloacal endoderm right below the ventral ectoderm ( Figure 1B inset ) . Subsequent functional analyses revealed that physically removing Fgf8-expressing cells or application of neutralizing FGF8 antibody could abolish GT growth in organ culture , and this growth inhibition could be reversed by adding FGF8-soaked beads [24] . Along with these studies , we uncovered that activity of the canonical WNT-β-catenin pathway is restricted to the Fgf8-expressing distal cloacal endoderm and later the dUE [4] . We found that abolishing β-catenin ( β-Cat ) in the cloacal endoderm caused GT agenesis/reduction , whereas ectopic activation of the same pathway resulted in GT over-development . Interestingly , the site and level of WNT signaling activity positively correlated with Fgf8 expression and the extent of outgrowth in both limbs and GT . These findings illustrated a parallel between dUE and AER signaling during appendage outgrowth . However , the exact mechanisms and functional relevance for this WNT-Fgf8 regulation remain to be elucidated . Recently , two independent studies reported a surprising observation that abolishing Fgf8 [25] , [26] , or simultaneously abolishing Fgf4 and Fgf8 [26] , in the cloacal endoderm does not affect GT outgrowth or GT-specific gene expression . These results questioned the relevance of FGF signals in external genitalia development and challenged the view that GT formation is organized and maintained by the dUE , which further suggested that the mechanisms underpinning limb and GT outgrowth are indeed divergent [2] . Herein , we adopted comprehensive genetic approaches to address the inductive signals in GT development , in comparison with that of the limb . We sought to define the role for FGF signaling and dUE-expressed Fgf8 in the GT , and explore mechanisms upstream of FGF activation in the dUE of the GT and AER of the limb . Results from our analyses revealed a remarkably conserved Wnt-Sp8-Fgf8 genetic circuitry that is crucial for proximodistal outgrowth of both paired- and unpaired-type appendages in mice . Previous studies demonstrated that inactivating one or two FGF ligands did not affect genital development [25] , [26] . We reasoned that the extensive genetic redundancy among FGF ligands might have undermined the power of these experiments . Therefore , we sought to re-evaluate the function of FGF signaling in GT development by conditionally abolishing FGF receptors . Fgfr1 and Fgfr2 are the only FGF receptors expressed in the developing GT [27] . Both Fgfr1 and Fgfr2IIIc are expressed in the para-cloacal mesenchyme ( PCM ) during GT outgrowth [27] , and Fgfr2IIIb is expressed in both the PCM and the urethral epithelium ( UE ) [27] , [28] . To abolish all FGF responsiveness in the developing GT , we employed an Msx2rtTA;tetO-Cre system which enables doxycycline ( Dox ) -inducible gene ablation in both the PCM and the UE [29] , [30] . Dox was given to pregnant females on embryonic day ( E ) 9 . 5 and E10 . 5 to induce Cre-mediated recombination of the Fgfr1f/f [31] and Fgfr2f/f [32] alleles . The phenotypes of Msx2rtTA;tetO-Cre;Fgfr1f/f;Fgfr2f/f mutant embryos [hereafter referred to as GT-Fgfr1/2-double conditional knockouts ( cKO ) , or dcKO] were analyzed by scanning electron microscopy ( SEM ) . GT-Fgfr1;r2-dcKO genital tubercles were underdeveloped compared to their littermate controls at all stages examined starting from E11 . 5 , when a clear tubercle structure could first be detected ( Figure 2B ) . At E12 . 5 , the dcKO GTs showed a clear deficiency in proximodistal outgrowth , as they were much shorter than the controls ( data not shown ) . At E15 . 5 , the mutant GTs were smaller in size , deformed ( Figure 2D ) and lacked the characteristic mesenchymal patterning present in controls ( Figure S1A–S1B ) . To further explore the cellular basis for the reduction in GT size , we analyzed proliferation and cell-death in the GT of these mutants . We performed phospho-histone H3 ( PHH3 ) staining on E11 . 0 coronal GT sections and counted the number of PHH3+ cells in a fixed area . A 28% reduction in PHH3-positive cell number in the genital mesenchyme of the dcKOs was observed ( Figure S1C-S1E , n≥10 , p = 0 . 0017 ) . We also performed TUNEL analyses but did not observe any differences in the number of apoptotic cells between control and dcKO mutant GTs ( data not shown ) . An examination of genes known to mediate genital tubercle initiation revealed alterations in normal gene expression patterns as early as E11 . 5 . Bmp4 , Wnt5a and Msx1 were expressed in the PCM in control GTs ( Figure 2E and 2G , and Figure S1F ) , and their expression was barely detectable in dcKO GTs ( Figure 2F and 2H , and Figure S1G ) . Msx2 was expressed in both the PCM and UE in controls ( Figure S1H ) . In the GT-Fgfr1;r2-dcKOs , Msx2 expression was absent in the PCM and downregulated in the UE ( Figure S1I ) . Moreover , UE expression of Shh was also downregulated in the dcKOs ( Figure 2I–2H ) . In contrast , dUE-Fgf8 expression remained unchanged in the GT-Fgfr1;r2-dcKOs ( Figure 2K–2L ) , suggesting that maintenance of Fgf8 expression is independent of FGFR1 and FGFR2 during GT development . Collectively , these data demonstrate that FGF signaling is obligatory for promoting proximodistal GT outgrowth and for maintaining genital-specific gene expression . The Msx2rtTA;tetO-Cre system mediates gene deletion in both the cloacal endoderm as well as the PCM . Therefore , it is not clear whether changes observed in the aforementioned Fgfr1;r2-dcKOs were direct results of compromised FGF responsiveness in the PCM , or secondary to loss of FGF receptors in the cloacal endoderm . Thus , we tested the requirement for FGF responsiveness in these two compartments by using a previously characterized endoderm-specific ShhEGFPCre allele [4] , [33] , and a PCM-specific Dermo1Cre allele [4] , [32] , respectively . ShhEGFPCre/+;Fgfr1f/f;Fgfr2f/f ( UE-Fgfr1;r2-dcKO ) and Dermo1Cre/+;Fgfr1f/f;Fgfr2f/f ( PCM-Fgfr1;r2-dcKO ) mutants were generated and their GTs analyzed . Intriguingly , GTs from UE-Fgfr1;r2-dcKOs did not exhibit any morphological abnormalities in the early outgrowth phase , and their size and shape were comparable to stage-matched controls ( Figure S2A–S2D ) . Histological analysis revealed normal patterning of the genital mesenchyme in UE-Fgfr1;r2-dcKOs embryos ( Figure S2E and S2F ) . The only phenotype observed in these mutants was the abnormal maturation of urethral epithelium similar to what was observed in Fgfr2IIIb mutants [28] , which will be discussed in a separate manuscript . Furthermore , regulatory genes including Msx2 , Bmp4 , Wnt5a and Fgf8 , were also properly expressed in these UE-Fgfr1;r2-dcKO embryos ( Figure S2G–S2J ) . In contrast , the GTs of PCM-Fgfr1;r2-dcKO were clearly smaller than their littermate controls ( Figure S3A–S3B ) . Further analyses revealed that the distal GT mesenchymal expression of P-ERK1/2 , a previously established FGF target gene [26] , was also downregulated in these mutants ( Figure S3C–S3D ) . In addition , PHH3 staining on E11 . 5 embryos revealed a 20% reduction in mitotic figure number in the PCM-Fgfr1;r2-dcKOs ( Figure S3E–S3G ) . Collectively , these data indicate that the main target for FGF signaling during GT outgrowth is the PCM . A similar requirement for FGF signaling in the limb mesenchyme has been described previously [34] . Fgf8 is normally expressed in the distal-posterior cloacal endoderm at E10 . 5 , and then in the dUE through E11 . 5–E14 . 5 . We sought to determine whether the PCM could respond to dUE-expressed Fgf8 in vivo . We generated a conditional Fgf8 overexpressor mouse line by knocking the Fgf8 full-length cDNA ( Accession: BC048734 ) into the ubiquitously expressed Rosa26 locus , preceded by a floxed transcriptional stop cassette ( R26Fgf8 ) . This design allows ectopic Fgf8 expression upon Cre-mediated recombination ( Figure 3A ) . We used a tamoxifen ( Tm ) -inducible ShhCreERT2 allele [4] , [33] and generated ShhCreERT2/+;R26Fgf8 gain of function ( GOF ) mutants ( UE-R26Fgf8-GOF ) , to achieve UE-specific Fgf8 overexpression upon Tm treatment at E10 . 5 . The Cre expression domain of this ShhCreERT2 allele recapitulates endogenous Shh expression , which includes all cloacal endodermal cells . Eight hours after Tm administration , we noted a clear upregulation ( arrowhead in Figure 3C ) and an anterior expansion ( arrow in Figure 3C ) of Fgf8 expression in the cloacal endoderm evidenced by whole-mount in situ hybridization . The ectopic expression in the anterior cloacal endoderm ( arrow in Figure 3C and 3E ) was the result of ectopic Fgf8 expression from the R26Fgf8 allele . Notably , a concurrently augmented Bmp4 expression was evident in the PCM ( insets in Figure 3B and 3C ) . Sixteen hours after the initial tamoxifen injection , we observed a 19% increase in mitotic index in the PCM ( Figure S4A–S4C , n = 10 , p = 0 . 009 ) . Consistently , the mutant GTs were larger than controls at E12 . 5 ( Figure S4D–S4E ) and E14 . 5 ( Figure 3H–3I ) , respectively . These findings clearly demonstrate that endodermally expressed FGF8 can mediate gene expression and promote cell proliferation in the neighboring PCM . Interestingly , we noted that endogenous Fgf8 expression was differentially regulated in the UE-R26Fgf8-GOF mutants . Sixteen hours after Tm treatment , we observed a distinct downregulation of Fgf8 expression in the GOF mutant dUE ( arrowhead in Figure 3E ) that persisted 24 ( Figure 3G ) and 48 hours after treatment ( data not shown ) . It is also noteworthy that transcription from the R26 locus was much weaker than that from the endogenous Fgf8 locus , evidenced by the faint signal in the anterior cloacal endoderm ( indicated by arrows in Figure 3C and 3E ) . Collectively , these findings demonstrate that the developing GT is sensitive to changes in FGF dosage , and a feedback loop is deployed to ensure proper signaling activity when misregulation occurs . To compare the function of FGF8 in the limb and GT , we mated the R26Fgf8 allele with a transgenic Msx2-Cre line [35] , which confers Cre expression in the forming and mature AER and the ventral limb ectoderm . As expected , the AER-R26Fgf8-GOF mutants exhibited excessive limb growth and developed extra digits ( asterisk in Figure 3J ) , an enlarged calcaneus bone ( arrowhead in Figure 3K ) , and ectopic skeletal elements ( arrows in Figure 3J–3K , and Figure S4G ) in both forelimbs and hindlimbs . These overgrowth phenotypes are similar but more severe than what has been observed in the AER-Fgf4-GOF embryos [36] , and further support the concept that FGF8 plays a pivotal role in promoting the outgrowth of both appendages . Our previous work has shown that the WNT-β-catenin signaling pathway and Fgf8 expression are tightly coupled in the distal signaling centers both in the limb and in the GT [4] . The above finding that Fgf8 was repressed by its own overexpression was in sharp contrast to our previous observation in the UE-β-Cat-GOF mutants where ectopic up-regulation Fgf8 was sustained in the UE [4] . This suggested that the canonical WNT pathway plays a key role in controlling the Fgf8 auto-regulatory feedback loop . Together , these observations suggest that the WNT-Fgf8 regulatory relationship is essential for appendage formation , and prompted us to test whether loss of Fgf8 was the critical event causing appendage reduction in embryos deficient in β-Cat in the AER and the UE . Therefore , we attempted to restore Fgf8 expression in β-Cat loss of function ( LOF ) embryos and analyze its effect on appendage formation . We generated ShhEGFPCre/+; β-Catf/f;R26Fgf8/+ ( UE-β-Cat-LOF;R26Fgf8 ) mutants and analyzed their GT development . In contrast to absence of the GT in the UE-β-Cat-LOF embryos ( Figure 4B ) , a cone-shaped tubercle structure was readily discernible in compound mutants carrying the R26Fgf8 allele ( Figure 4A–4C ) . To determine whether this rescued structure bears GT characteristics , we performed both histological and gene expression analyses . Hematoxylin and Eosin ( H&E ) stained E12 . 0 transverse GT sections showed that the morphology of the rescued GT closely resembled that of the controls with the urethra properly positioned at the ventral side of the GT ( Figure 4D ) . In addition , Hoxa13 and Hoxd13 , both markers of the GT , were expressed in the rescued genital tubercles ( Figure 4E and 4F ) . Altogether , these data indicate that restoring FGF8 rescued GT agenesis caused by β-catenin deficiency . It is noteworthy that this rescue of β-Cat-cKO by FGF8 is confined to the GT , as other caudal malformations observed in the β-Cat-cKO including failed cloaca septation and defective tail formation , were still present in the UE-β-Cat-LOF;R26Fgf8 embryos ( data not shown ) . To test whether forced Fgf8 expression can also rescue limb deficiency caused by β-Cat ablation , we generated Msx2-Cre; β-Catf/f ( AER-β-Cat-LOF ) and Msx2-Cre; β-Catf/f;R26Fgf8/+ ( AER-β-Cat-LOF;R26Fgf8 ) mutants . Consistent with a previous investigation [3] , all autopod elements , radius , and distal two-thirds of the ulna were missing from the forelimbs of AER-β-Cat-LOFs ( Figure 4I ) , whereas the humerus was thinner and lacked the deltoid tuberosity ( Figure 4I compared to G , arrow ) . In contrast , the LOF embryos with the R26Fgf8 allele developed normal humeri with the deltoid tuberosity ( Figure 4K , arrow ) . The radius was evident and the ulna was longer and thicker ( Figure 4K ) . Moreover , several small pieces of alcian blue-stained cartilage were observed distal to the ulna , indicating the presence of autopod rudiments ( arrowhead and inset in Figure 4K ) . The hindlimbs of the AER-β-Cat-LOFs were largely absent except for a small remnant of the pelvic girdle ( Figure 4J ) , whereas the R26Fgf8-expressing β-Cat-LOF embryos developed near normal pelvic girdles and femurs ( Figure 4L ) along with one or two ectopic cartilages ( asterisk in Figure 4L ) . These phenotypes were consistently observed in all AER-β-Cat-LOF;R26Fgf8 embryos ( n = 10 ) , and indicated that exogenously supplying FGF8 can partially restore distal limb structures lost in the AER-β-Cat-LOF embryos . Collectively , these data suggest that FGF8 can promote outgrowth of both the GT and the limb in the absence of canonical epithelial WNT activity , suggesting that it functions as a downstream effector of WNT signaling during limb and GT outgrowth . The positive feedback loop involving FGFs and SHH plays a critical role in appendage outgrowth , as evidenced by the down-regulation in Fgf expression in both the AER [21] , [22] , [35] , [37] and the dUE [5] , [6] of Shh-KOs , which display reductions in both appendages . We next examined whether forced Fgf8 expression could also rescue the severe appendage deficiencies caused by the absence of SHH . We generated ShhCreERT2/CreERT2;R26Fgf8/+ ( Shh-KO;UE-R26Fgf8 ) embryos , which allowed us to induce Fgf8 expression in the UE of Shh null mutants ( ShhCreERT2 allele is also a null allele ) . However , we detected neither tubercle formation ( Figure S5B ) , nor Hoxa13 or Hoxd13 expression ( Figure S5C and S5D ) in the cloacal region of these compound mutants at E12 . 5 , 48 hours after Tm treatment . We also generated Msx2-Cre; R26Fgf8/+; ShhCreERT2/CreERT2 ( Shh-KO;AER-R26Fgf8 ) mutants , to test whether sustaining Fgf8 expression in the AER of Shh mutants can restore limb development . We carefully analyzed four Shh-KO embryos carrying both Msx2-Cre and R26Fgf8 alleles , and found no evidence of more advanced limb development ( Figure S5G and S5H ) , compared to thirteen Shh-KO mutants without R26Fgf8 alleles ( Figure S5E and S5F ) . Together , these results indicated that although Shh and Fgf8 expression are interdependent during appendage outgrowth , their function in promoting appendage outgrowth is independent and non-redundant . We next analyzed GT gene expression in UE-β-Cat-LOF; R26Fgf8 mutants to further interrogate the genetic networks underlying GT development and identify downstream targets of dUE signaling . Fgf8 expression was detected in the distal cloacal endoderm in controls ( Figure 5A ) , but was absent in UE-β-Cat-LOF mutant ( Figure 5B ) at E10 . 5 . The R26Fgf8-expressing LOF embryos , on the other hand , showed very weak Fgf8 expression throughout the cloacal endoderm ( arrow in Figure 5C ) . This low-expression was consistent with what we observed in the UE- R26Fgf8-GOF embryos , suggesting that these Fgf8 transcripts were transcribed from the R26 locus . Expression of Bmp4 and Wnt5a was normally detected in the PCM ( Figure 5D and Figure S6A ) , absent in the UE-β-Cat-LOFs ( Figure 5E and Figure S6B ) , and partially restored by ectopically supplying Fgf8 expression from the R26Fgf8 allele ( Figure 5F and Figure S6C ) . The SHH pathway was also compromised in the β-Cat-LOF mutants . Shh ( Figure 5H ) and Ptch1 expressions ( Figure S6E ) were markedly decreased in the cloacal endoderm and the PCM , respectively . On the other hand , in the LOF embryos with the R26Fgf8 allele , the expression of Shh was partially restored in the UE ( Figure 5I ) and the PCM expressed Ptch1 restored to a level comparable to controls ( Figure 5F ) . Combined , these data indicate that most genes differentially regulated in the UE-β-Cat-LOF mutants were responsive to FGF8 induction , suggesting that their expression is controlled by the dUE-FGF signals . However , we found that Sp8 , a transcription factor normally expressed in the cloacal endoderm ( Figure 5J ) , was lost in the UE of UE-β-Cat-LOF mutant ( Figure 5K ) , and did not respond to FGF8 supplementation ( Figure 5L ) . Sp8 is expressed throughout the cloacal endoderm and later in the UE ( Figure 6A ) , overlapping with the Fgf8-expressing dUE during GT development . Previous studies have implicated SP8 in the transcriptional regulation of Fgf8 expression in the mouse commissural plate [38] and the chick limb ectoderm [39] . These findings prompted us to explore its role in mediating the WNT-Fgf8 pathway . We first examined Sp8 expression in the UE-β-Cat-LOF and GOF-mutants ( for GOF analyses , we used a previously established β-CatΔex3 allele which produces stabilized β-catenin upon Cre-mediated recombination [40] ) . We found that Sp8 expression was reduced in the β-Cat-LOF UE ( Figure 6B ) and increased in the β-CatΔex3 GOF UE ( Figure 6C ) . Similarly , robust Sp8 expression was also detected in the AER ( Figure 6D ) , markedly reduced in the AER- β-Cat-LOF mutants ( Figure 6E ) , and upregulated in the β-CatΔex3 GOF mutants where β-catenin activity was ectopically augmented in the limb ectoderm ( Figure 6F , inset showing ventral view of a hindlimb ) . These data indicated that Sp8 is downstream of Wnt-β-catenin signaling in the UE and the limb ectoderm . Next , we examined the GT phenotype of Sp8-null ( KO ) mutants . We found that 14/36 mutant embryos examined between E12 . 5–E15 . 5 exhibited GT agenesis ( Figure 6H ) , while the rest demonstrated a range of GT defects including deformation , hypoplasia and proximal hypospadias ( data not shown ) . Fgf8 expression was completely absent in the cloacal endoderm in all Sp8 KOs examined at E11 . 5 ( Figure 6J , n = 9 ) . Notably , these embryos also exhibited other caudal malformations such as deformed perineum , anal channel and tails , raising the concern whether the observed GT defects were secondary to earlier cloacal or neural tube malformations . Thus , we employed a conditional Sp8 null allele with the ShhCreERT2 line to generate UE-Sp8-LOF mutants , and induced Sp8 deletion by Tm treatment around the time of GT initiation ( E10 . 5 ) . GTs from the UE-Sp8-LOF mutants were smaller than their age-matched controls , especially at the distal tip ( Figure 6L ) . We also found a clear reduction in Fgf8 expression in these mutants by in situ analyses ( inset in Figure 6L ) . Consistently , the expression domains of Wnt5a ( Figure S7B ) and Msx2 ( Figure S7D ) , both maintained by FGF8 from the neighboring dUE , were reduced . Limb truncation , attributed to a failure to form the AER and consequently loss of Fgf8 expression , has been previously described in Sp8−/− embryos [41] . We generated AER-Sp8-LOFs using the Msx2-Cre line , and these mutants also showed a defect in limb outgrowth as evidenced by a loss of distal structures ( Figure 6Q–6S ) . The stylopod and zeugopod developed normally in the forelimbs , while typically only one abnormal digit formed in these mutants ( Figure 6Q–6R ) . The tibia and fibula were either missing or severely truncated , and no autopod was observed in the hindlimbs ( Figure 6S ) . Notably , these limb defects can also be partially rescued by over expression of Fgf8 . Skeleton preparation of E18 . 5 embryos showed that the AER-Sp8-LOF embryos carrying R26Fgf8 allele developed three digits in the forelimb ( Figure 6U ) , and full-length tibia and fibula along with several irregular autopod elements in the hindlimbs ( Figure 6V–6W ) . Altogether , these data indicated that Sp8 is required to maintain Fgf8 expression and appendage outgrowth in the distal signaling center of both the limb and GT . To test whether Sp8 is in the same genetic pathway with β-catenin and Fgf8 , we sought to determine whether SP8 mediates WNT-induced Fgf8 expression in the dUE and AER . We generated compound mutants using the β-CatΔex3 allele , which we have previously shown to activate Fgf8 expression in both the UE and the limb ectoderm , together with the floxed Sp8 allele and the corresponding Cre lines . We compared Fgf8 expression in embryos with one β-CatΔex3 allele and different numbers of functional Sp8 alleles by real-time RT-PCR and in situ hybridization . We found that the level of WNT-induced Fgf8 expression in the UE positively correlated with the number of functioning Sp8 alleles ( Figure 7A–7E ) . Removing one wild type Sp8 allele caused a twofold reduction in Fgf8 expression , whereas ablating both wild type Sp8 alleles reduced Fgf8 expression by more than three-fold ( Figure 7E ) . These results were further verified by Fgf8 in situ hybridization ( Figure 7A–7D ) . Similarly , deleting both Sp8 alleles abolished the ectopic Fgf8 expansion in the limb ectoderm of Msx2-Cre; β-CatΔex3/+ ( AER-β-Cat-GOF ) embryos ( Figure 7F–7H ) , and consequently attenuated the polysyndactyly phenotype caused by constitutively active canonical WNT signaling ( Figure 7I–7O ) . With two wild type Sp8 alleles , AER-β-Cat-GOF mutants developed an average of 6 . 8 digits in the forelimbs , and 6 . 1 digits in the hindlimbs ( Figure 7J and 7M ) . In thirty percent of the embryos , ectopic limb in the flank ectoderm and ventral ectoderm were detected ( Figure S8A and S8B ) . In comparison , AER-β-Cat-GOF;Sp8-null mutants only developed 4 . 9 digits in the forelimbs and 4 . 2 digits in the hindlimbs ( Figure 7K and 7N , 28% and 31% reduction compared to AER-β-Cat GOFs , respectively ) . In addition , no extra limbs were observed at ectopic locations . Collectively , these results indicate that SP8 is responsible , at least in part , for the WNT-induced activation of Fgf8 expression during appendage outgrowth . To test whether augmented Sp8 expression alone can induce Fgf8 overexpression , we generated a R26Sp8 allele to conditionally overexpress Sp8 using the same strategy as for the R26Fgf8 line ( Figure 8A ) . Mice carrying ShhEGFPCre or Msx2-Cre alleles were used to generate corresponding UE-R26Sp8-GOF and AER-R26Sp8-GOF embryos . Overexpression of Sp8 in the UE and AER was confirmed by in situ hybridization ( Figure 8C , and data not shown ) . However , unlike Fgf8- or β-Cat-GOF embryos , Sp8-GOF embryos showed normal development in both appendages ( Figure 8E and 8I ) . Fgf8 expression in the dUE and the AER of the corresponding Sp8-GOF mutant was also comparable to their wild type counterparts ( Figure S9A–S9D ) . We also crossed the R26Sp8 allele into the UE- or AER-β-Cat-LOF mutants to test whether forced Sp8 expression can bypass epithelial β-catenin to induce Fgf8 expression and initiate/maintain appendage outgrowth . We carefully analyzed six compound mutants and did not observe phenotypic rescue in either the GT ( Figure 8F ) or the limb ( Figure 8J–8M ) . In addition , no Fgf8 induction was detected in the dUE of the corresponding β-Cat-LOF mutants carrying R26Sp8 allele ( Figure 8G ) . All of these results indicate that SP8 by itself is insufficient to activate Fgf8 expression , and suggest that SP8 is a facilitator of WNT-mediated Fgf8 activation during appendage formation . Our work provides in vivo evidence that FGF signaling is indispensible for early GT outgrowth . These findings are consistent with results from organ culture studies that inhibition of FGF signaling caused an arrest in GT development [24] . Our data demonstrates that the PCM's ability to respond to an FGF signal is essential for normal early genital tubercle outgrowth , as removing FGF receptors from the PCM caused impaired cell proliferation and perturbed normal gene expression patterns which led to severe GT reduction . The obligatory role for FGF signaling during tubercle morphogenesis underscores the importance of identifying FGF ligands important for GT development . Our data are consistent with the views of Haraguchi et . al . [24] , suggesting that dUE-expressed Fgf8 plays a key role in promoting GT outgrowth . Fgf8 is expressed at the correct time and place to signal to the PCM , which expresses both Fgfr1 and Fgfr2 . This is in conflict with recent data by Seifert et . al . [25] in which they suggested that a normal GT could develop in the absence of Fgf8 expression . They did not detect FGF8 protein , which led to the conclusion that the Fgf8 mRNA may be present but not actively translated by the UE . They proposed that the ventral ectoderm may be an alternative source of other FGF ligands . In contrast , our results indicate that FGF8 protein can be made by the cloacal endoderm/UE as we demonstrated that a weakly-expressed R26Fgf8 allele can profoundly alter cell proliferation and gene expression in the neighboring PCM , and rescue GT agenesis in β-catenin mutants . In addition , our observation that the endogenous Fgf8 promoter is repressed upon forced Fgf8 overexpression revealed that the developing GT is equipped with mechanisms that can fine-tune the level of FGF8 signaling . All of these data indicate that the GT is not only responsive but also sensitive to FGF8 during normal development . The inability to detect FGF8 in the dUE by IHC is likely caused by the low expression level of Fgf8 . We found that in E10 . 5 mouse embryos , robust AER Fgf8-expression was observed 2 hours after incubation with Alkaline Phosphatase substrates following standard in situ hybridization procedures , whereas dUE-expression was not observed until 12–18 hours later . In addition , the extensive genetic redundancy among FGFs has to be carefully considered . A recent study demonstrated ectopic Fgf4 expression in the dUE of Fgf8-cKO embryos , and ectopic Fgf3 expression in the dUE of Fgf4;Fgf8-dcKO embryos [26] . Both FGF4 and FGF3 can efficiently induce mitogenic activity when paired with FGFR1 or FGFR2 in vitro [42] . These observations indicate that not only FGFs endogenous to the dUE can compensate for the loss of Fgf8 , other FGFs can also be ectopically activated to fulfill the requirement for FGF signaling . Activities from these ectopic FGFs could well explain the lack of GT phenotype in Fgf8-cKO mutants . Intriguingly , induction of Fgf3 and Fgf4 was not observed in either β-Cat- or Sp8-cKOs ( Figure S10A–S10D ) , while the dUE expression of Fgf9 was downregulated in Sp8-cKOs ( Figure S10E–S10F ) . These results suggest that the upregulation of these compensatory FGF factors also requires WNT and SP8 . Alternatively , the hypothesis that FGFs can be produced by the ventral ectoderm is plausible [25] . However , one has to keep in mind that the GT is built and patterned around the UE . Most regulatory genes expressed distally including Msx1 , Msx2 , Wnt5a and Lef1 , showed peri-dUE expression but not sub-ectodermal expression . Consistently , we have shown that the expression of these genes is orchestrated by UE-specific WNT and FGF signaling . Considering all the available evidence , we conclude that FGF8 produced by the dUE is most likely the endogenous ligand that mediates FGF responses during GT development . Notably , the GT phenotype of the UE-Fgfr1;r2-dcKO embryos is less severe than what was observed previously in Fgfr2IIIb-KOs [28] . This difference is likely attributable to the method of gene ablation since in the Fgfr2IIIb-KO embryos , Fgfr2-IIIb is abolished from not only the UE , but also from the ventral ectoderm . The underdeveloped phenotype described in Fgfr2IIIb-KO embryos reflects a deficiency from ectodermal FGF responsiveness as it can be phenocopied by conditional ablation of both Fgfr1 and Fgfr2 using the ectodermally restricted Msx2-Cre allele ( Lin et . al . , unpublished data ) . We have demonstrated a conserved WNT-SP8-FGF8 pathway in the distal signaling epithelia that functions to promote proximodistal outgrowth of both limbs and genitalia . We and others have shown that the canonical WNT pathway is a master molecular switch in the signaling epithelia during appendage formation , as epithelial WNT activation is not only necessary but also sufficient to induce Sp8 and Fgf8 expression and appendage outgrowth . FGF8 is the downstream signal output for the WNT pathway during this process as it acts directly on recipient mesenchymal cells to promote cell proliferation and establish patterns of gene expression . This genetic hierarchy has been supported by our observation that forced overexpression of Fgf8 , even at a level much lower than endogenous expression , can bypass the requirement for epithelial WNT-β-catenin signaling to activate gene expression and initiate/maintain appendage outgrowth . Notably , in both appendages , the rescue of β-catenin deficiency by ectopic Fgf8 expression is only partial . This could be due to either weak Fgf8 expression from the R26 locus compared to the endogenous level of expression , or the possibility that in addition to regulating Fgf8 expression , canonical WNT signaling is also required for the formation of the AER structure , independent of FGF signaling [20] . The regulation of Fgf8 by the canonical WNT-β-catenin signaling pathway is in part mediated by SP8 , as Sp8 expression is regulated by WNT signaling in both limb ectoderm and the UE , and is necessary for Fgf8 expression and subsequent appendage formation . However , unlike WNT activity and Fgf8 expression , the expression domain of Sp8 is not restricted to the AER and dUE but also includes the limb ectoderm [41] as well as proximal UE ( Figure 6A ) , suggesting that SP8 is a permissive but not inductive factor for the establishment of Fgf8 expression . In support of this notion , overexpressing Sp8 in the limb ectoderm or UE did not cause any perturbation in Fgf8 expression or appendage formation . In addition , forced Sp8 expression failed to rescue Fgf8 expression and appendage outgrowth in the epithelial β-Cat-LOF mutants . Notably , even with both Sp8 alleles mutated , the dUE and AER Fgf8 expression in the β-Cat-GOF mutants is still higher than in the controls . This finding apparently is counterintuitive to the obligatory role for SP8 in maintaining Fgf8 expression during normal outgrowth processes . One possible explanation is that the level of WNT signaling induced by the stabilized β-catenin protein from the β-CatΔex3 allele might be too high , and not subjected to endogenous regulatory mechanisms . This could potentially trigger ectopic events that lead to Fgf8 overexpression . It is also noteworthy that the other members of SP/KLF transcription factor family Sp6 , was upregulated in the β-Cat-GOF mutants ( Figure S11 ) . Sp6 is expressed in the cloacal endoderm and the AER in early appendage development ( Figure S11B and Figure S6A , respectively ) , and loss of Sp6 causes abnormal AER formation and Fgf8 expression [43] . Similar to Sp8 , the expression of Sp6 is also downstream of the canonical WNT signaling pathway but independent of FGFs in the AER [43] . The exact function of SP6 in regulating GT development and dUE-Fgf8 expression remains to be determined . However , it is plausible that high levels of SP6 induced by ectopic WNT signaling might compensate for the absence of SP8 in the regulation of Fgf8 expression . The exact molecular mechanisms through which Fgf8 expression is regulated by SP8 and β-catenin , and how Sp8 expression is regulated by WNT-β-catenin remain to be determined . Direct binding of the β-catenin/LEF1 complex to cis-regulatory elements within the Fgf8 promoter has been reported in dental epithelial cell lines [44] and nephron progenitors [45] , suggesting that this WNT-Fgf8 pathway also functions in the development of other organs involving epithelial-mesenchymal interactions . In our preliminary studies , we identified several novel LEF binding sites around Fgf8 and Sp8 genes ( data not shown ) . The functional relevance of these binding sites in GT and limb development will be the goal of future investigations . Although the distal signaling cascades appear to be similar in both appendages , the upstream events leading to their initiation are likely different . In the limb , the ectoderm is the site for both Wnt3 expression [3] , [46] , and the induction of canonical Wnt downstream targets as evidenced by TOPGAL activity [47] . On the other hand , the establishment of the dUE signaling center in the cloacal endoderm appears to involve signal transduction between the cloacal endoderm and the ventral ectoderm . Both TOPGAL and Fgf8 expression are confined to the endodermal cells adjacent to the genital ectoderm [4] , [24] , and the contact between ectoderm and endoderm appears to be a prerequisite for Fgf8 induction and GT initiation [25] . In support of this notion , candidate WNT ligands responsible for activating canonical WNT signaling in the dUE , Wnt3 and Wnt7a , are both expressed in the genital ectoderm [4] , [26] . Altogether , our results demonstrate extensive parallels between genetic networks regulating the outgrowth of both the limb and GT . These findings strongly support the notion initially brought up by two pivotal studies on the role of Hox genes in appendage development [10] , [11] , that mammalian GT development appears to be achieved through co-option of the limb outgrowth program , i . e . the mammalian GT and limb share deep homology [48] , [49] . Clones containing full length Fgf8 ( BC048734 ) and Sp8 ( BC082582 ) cDNAs were obtained from Invitrogen ( Carlsbad , CA ) . The cDNAs were released and subcloned into the NotI site of the pBigT vector [50] . The insert containing either Fgf8 or Sp8 cDNA and a floxed transcription stop cassette was released by PacI/AscI double digestion and cloned into pRosa26-PA [51] . The targeting construct was linearized by SwaI and subjected to electroporation ( performed by ES cell core at Washington University ) . The ES cells were screened for recombination by PCR and southern blotting . Five out of seventy-two clones were positive for recombination for R26Fgf8 allele , and six out of sixty clones were positive for R26Sp8 allele . Positive clones were expanded , karyotyped and used for blastocyst injection . For both lines , at least three chimeras were able to pass the knock-in alleles through germline transmission , and the phenotypes resulting from expression of the knock-in alleles were identical . For all experiments described in the manuscript , three embryos with the same genotype were examined if not otherwise specified . All animals were maintained according to NIH guidelines and in compliance with animal protocol approved by Washington University . Msx2rtTA [29] , Sp8f/f and Sp8−/− [41] , tetO-Cre [52] , β-Catf/f [53] , β-Catex3/ex3 [40] , ShhCreERT2 and ShhEGFPCre [33] , Msx2-Cre [35] , Fgfr1f/f [31] , Dermo1Cre and Fgfr2f/f [32] alleles were previously described . Tamoxifen treatment and doxycycline treatment were performed as previously described [30] . For all experiments , we used three independent biological samples from each genotype . Each sample contained a pool of RNA isolated from two E12 . 5 GTs with the corresponding genotype . The results were analyzed using the delta-CT method . Expression of the corresponding genes was normalized to that of housekeeping gene Rpl7 . Whole mount in situ hybridization was performed using a standard protocol . The probes were previously described [4] . Paraformaldehyde ( 4% ) -fixed embryos were dehydrated and embedded in paraffin . Five-micron transverse GT sections were generated using a standard microtome . Scanning electron microscopy analyses were performed as previously described [4] . Skeleton preparation was performed as previously described [54] . The Genbank accession numbers for genes included in the manuscript are as follows: Fgf8 ( NM_010205 . 2 ) , Sp8 ( NM_177082 . 4 ) , β-catenin ( NM_007614 . 3 ) , Fgfr1 ( NM_010206 . 2 ) , Fgfr2 ( NM_010207 . 2 ) , Shh ( NM_009170 . 3 ) .
Mammalian limbs and external genitalia are body appendages specialized for locomotion and internal fertilization , respectively . Despite their marked anatomical and functional differences , development of the limb and external genitalia appears to involve similar genetic controls , and some have suggested that regulatory mechanisms common to both might be evolutionarily linked . One essential aspect for appendage development is the establishment and maintenance of a separated proximodistal developmental axis apart from the main body axis , which is often instructed by a distal signaling epithelium . Herein , we adopted comprehensive mouse genetic approaches to investigate regulatory mechanisms underlying the distal signaling center in the limb and the GT , and uncovered a conserved genetic cassette that is utilized by both paired and unpaired appendages to establish a distal signaling center in the epithelium that mediates subsequent proximodistal outgrowth . Our results further suggested that the evolution of the external genital organ involved co-option of the same genetic program underpinning limb development .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "developmental", "biology", "embryology", "molecular", "development", "biology" ]
2013
Delineating a Conserved Genetic Cassette Promoting Outgrowth of Body Appendages
M . tuberculosis ( MTB ) species-specific antigenic determinants of the human T cell response are important for immunodiagnosis and vaccination . As hypoxia is a stimulus in chronic tuberculosis infection , we analyzed transcriptional profiles of MTB subject to 168 hours of hypoxia to test the hypothesis that upregulation by hypoxia might result in gene products being recognized as antigens . We identified upregulation of two region of difference ( RD ) 11 ( Rv2658C and Rv2659c ) , and one RD2 ( Rv1986 ) absent from commonly used BCG strains . In MTB infected persons , the IL-2 ELISpot response to Rv1986 peptides was several times greater than the corresponding IFN-γ response to the reference immunodominant ESAT-6 or CFP-10 antigens . The IL-2 response was confined to two epitopic regions containing residues 61–80 and 161–180 . The biggest population of IL-2 secreting T cells was single cytokine positive central memory T cells . The IL-2 response to live MTB bacilli lacking Rv1986 was significantly lower than the response to wild type or mutant complemented with Rv1986 . In addition , the IL-2 response to Rv1986 was significantly lower in HIV-TB co-infected persons than in HIV uninfected persons , and significantly increased during antiretroviral therapy . These findings demonstrate that Rv1986 is an immunodominant target of memory T cells and is therefore of relevance when considering the partial efficacy of currently used BCG vaccines and provide evidence for a clinical trial comparing BCG strains . Mycobacterium tuberculosis remains a formidable health problem as it is estimated to infect one-third of the world's population and causes around 1 . 5 million deaths per year [1] . Control is largely based around the partially effective vaccine Mycobacterium bovis Bacille Calmette Guérin ( BCG ) and on the early detection and treatment of infected persons with active or latent disease [2] . Study of the antigens of M . tuberculosis is therefore a priority both to improve vaccination via the selection of protective antigens , and to define immunodiagnostic candidates that enhance the specificity and sensitivity of the widely used purified protein derivative ( PPD ) based tuberculin skin test ( TST ) . A significant landmark in both respects was the discovery that a M . tuberculosis genomic region designated region of difference ( RD ) 1 is deleted from all strains of BCG and thereby partially accounts for the avirulence of the vaccine [3] , [4] . RD1 encodes a pair of co-regulated secreted proteins ( ESAT-6 and CFP-10 ) that are highly immunogenic [5] , [6]: restoration of these genes into BCG improves its vaccine efficacy [7] . Assays of the T cell interferon ( IFN ) -γ secretion in response to the combination of ESAT-6 and CFP-10 ( interferon-γ release assays , IGRA ) have been developed that have operational advantages and improve the specificity and possibly sensitivity of tuberculosis immunodiagnosis [8] . The availability of the complete sequence of M . tuberculosis also permitted further genomic characterization of various BCG strains [9] , [10] , [11] . It became apparent that , against a background of accumulating single nucleotide polymorphisms , BCG underwent sequential genomic deletions that thereby characterize various strains . The strains most commonly in use such as BCG Glaxo , Danish and Pasteur have most deletions . This led to the proposal that one of the reasons behind the partial vaccine efficacy of BCG was that it had become too attenuated to successfully mimic natural MTB infection [12] . Some empirical evidence in humans favoring this hypothesis is provided by the finding that BCG Japan induced greater cytotoxic and T helper 1 responses in infants than Danish BCG [13] . The largest difference between BCG Japan and BCG Danish is the presence of RD2 in the former but not the latter . The discovery of immunodominant antigens in M . tuberculosis has hitherto largely been based on dominance in antibody responses that are neither the basis of protection against tuberculosis nor of IGRA . A more rational approach might be to relate what is highly expressed by bacilli in vivo or in vitro ( and thereby potentially available as an antigen ) as recently investigated in bovines [14] . In humans there has been investigation of proteins encoded by genes of the dosR regulon that is induced in axenic culture by hypoxia and by nitric oxide [15] , stresses that are considered relevant to bacilli in nature [16] , [17] , [18] . Analysis of selected dosR regulated proteins confirmed the immunodominance of α-crystallin 1 ( Acr1 ) encoded by Rv2031 [19] , [20] , [21] , as well as potentially infection stage specific antigens [19] , [22] , [23] , [24] . It has been shown in vitro that the up regulation of dosR regulated genes represents an early somewhat transient response to hypoxia: upregulation of a larger group of genes characterizes the hypoxic time course occurs in cultures subject to 4–7 days ( as opposed to 2 hours ) hypoxia [25] , of which 230 are defined as the enduring hypoxic response ( EHR ) . We hypothesized that these EHR and other hypoxia-induced genes would be worthy of consideration as antigens , especially those that were species-specific . We therefore undertook a study of the immunogenicity of M . tuberculosis specific genes induced by hypoxia and other well-characterized antigens ( ESAT-6 , CFP-10 , Acr1 ) in humans with active and latent tuberculosis . Cross-reference of genes with the greatest fold induction in hypoxic culture over 7 days [25] with sequence databases revealed two species-specific ( Rv2658c and Rv2659c , both RD11 encoded ) and one partially species-specific gene ( Rv1986 , RD-2 encoded ) . The fold induction and sigA normalized transcript intensity over a time course of 168 hrs hypoxia for these genes ( and of Acr1 , CFP-10 and ESAT-6 ) are shown in Table 1 . Interestingly whilst the fold induction for the RD1 encoded genes fell , the normalized intensity remained at a similar absolute level to that of both EHR and the dosR regulated Acr1 gene . Interferon-γ ( IFN-γ ) ELISpot was performed using PBMC from 40 persons with active ( 20 ) or latent ( 20 ) tuberculosis , and IL-2 ELISpot on 13 and 14 persons in each category . Immunodominance was assessed both quantitatively ( median SFC/106 PBMC ) and by frequency of response ( >20 SFC/106 PBMC ) . CFP-10 and ESAT-6 were co-dominant for the IFN-γ response by both methods ( Figure 1A and B ) . The largest SFC response in latent disease was to CFP-10 ( 102 SFC/106 PBMC , IQR 38-444 ) . With the exception of ESAT-6 all other responses were significantly lower ( p≤0 . 007 ) . The largest response in active disease was to ESAT-6 ( 172 SFC/106 PBMC , IQR 47–423 ) . With the exception of CFP-10 all other responses were significantly lower ( p≤0 . 002 ) . Although peptide pool Rv2659c-2 was preferentially recognized by latently infected persons ( 6 SFC/106 PBMC , IQR 1–28 versus 0 SFC/106 PBMC , IQR 0–7 , p = 0 . 028 ) these responses were very modest . When analyzed by proportions , no pool was preferentially recognized by either clinical group ( Figure 1B ) . The most frequent response in the combined group ( latent plus active ) was to CFP-10 ( 36/40 , 90% ) : with the exception of ESAT-6 the proportion of persons responding to the other antigens was lower in every case ( p≤0 . 002 ) . Patients with active TB preferentially recognized pooled peptides from Rv2658c ( 7 IL-2 SFC/106 PBMC , IQR 1–23 versus 0 SFC/106 PBMC , IQR 0–6 p = 0 . 042 ) . However when analyzed by proportion , no pool was preferentially recognized by either clinical group . There was however a striking IL-2 response in both active and latent disease to Rv1986 pool 2 ( 795 SFC/106 PBMC , IQR 51–1428 in active infection; 1194 SFC/106 PBMC , IQR 862–1650 in latent infection , Figure 1C and D ) . All other antigen specific IL-2 responses were significantly lower in both latent ( p≤0 . 0007 ) and active infection ( p≤0 . 02 ) . The most frequent response in the combined group ( latent plus active ) was also to Rv1986-2 ( 24/26 , 92% ) : with the exception of ESAT-6 and CFP-10 , the proportion of subjects responding to the other antigens was lower in every case ( p≤0 . 009 ) . Epitope mapping of the individual peptide determinants of the IFN-γ response to ESAT-6 and CFP-10 have shown several regions in each molecule that can restimulate T cells with no single peptide giving rise to a response in >50% subjects [26] , [27] and similar findings are reported for other antigens of M . tuberculosis [22] . We were therefore interested to determine whether a similarly ‘degenerate’ pattern of multiple IL-2 inducing epitopes occurred in Rv1986 . PBMC from 20 persons with latent tuberculosis were assayed in the presence of 10 µg/ml of each peptide or no stimulus . A highly focused pattern of dominance was observed with peptides p61-80 ( 84 SFC/106 PBMC , IQR 56–134 ) and p161–180 ( 68 SFC/106 PBMC , IQR 49–104 ) being clearly the best recognized ( Figure 2 ) . 90% of subjects had a response >20 SFC/106 PBMC to p61–80 and 95% to p161–180 with no other peptide being recognized by >45% subjects . There were less frequent and lower magnitude responses to p151–170 perhaps suggesting the epitope core for some MHC Class II molecules may include residues common to both peptides ( 161–170 ) . Analysis of the T cells responsible for type 1 cytokine responses is critical to understand protective immunity against TB [28] . In PBMC from 5 persons with latent tuberculosis , we therefore determined the phenotype of CD4+T cells responsible for type 1 cytokine ( IFN-γ , IL-2 and TNF ) production when stimulated with the peptides of Rv1986 ( 61–80 and 161–180 ) or the combination of peptides from CFP-10 and ESAT-6 as a comparison . T cell phenotypes were defined based on the surface markers CD45RA and CD27: Central memory cells ( TCM ) are positive for CD27 and negative for CD45RA; effector memory ( TEM ) are negative for both CD27 and CD45RA and Terminally differentiated T cells ( Tdiff ) are negative for CD27 and positive for CD45RA . Single cytokine positive cells predominated overall ( Figure 3 ) . Most IL-2 derived from TCM irrespective of stimulus . The two Rv1986 peptides restimulated nearly ten times the percentage of IL-2 producing TCM cells than CFP-10 and ESAT-6 ( median: 0 . 226% CD3+CD4+ TCM vs 0 . 024% CD3+CD4+TCM , p = 0 . 055 , Figure 3 panel A and B ) . We further investigated the ability of Rv1986 to induce the secretion of other cytokines when compared to CFP-10 . We used 16-hour cell culture supernatants from 39 persons with either active ( 19 ) or latent ( 20 ) tuberculosis . Multiple cytokine secretion was assessed both quantitatively ( pg/ml , after background correction ) and by frequency of response ( >2 fold above background ) . Similar levels of cytokine responses were observed in both analyses in persons with active and latent tuberculosis ( data not shown ) , therefore the clinical groups were combined for further analysis . When analyzed quantitatively and corrected for multiple comparisons , Rv1986 pool 1 and 2 stimulated significantly higher levels of IL2sR , TNF , IL-10 , IL-13 , MIP-1 alpha and MIP-1 beta than CFP-10 , and similar levels of RANTES . Levels of IL-13 were very modest ( Figure 4A , B and C ) . Thus the Rv1986 peptides were associated with a distinct pattern of cytokine production other than IL-2 when compared to CFP-10 . We next determined whether there was any difference in the IL-2 and IFN-γ responses to live strains of MTB in which Rv1986 was intact or deleted . 13 persons with latent tuberculosis were tested ( only 9 patients for IFN-γ due to limitation in cell numbers ) . Although the overall IFN-γ SFC response to these MTB strains was much stronger than the IL-2 response , there was no significant difference in IFN-γ response between these strains ( Figure 5B ) . The IL-2 SFC response to MTB H37Rv was significantly higher than to the H37RvΔRD-2 mutant ( median 228 SFC/106 PBMC , IQR 142–325 vs . 130 SFC/106 PBMC , IQR 53–268; p = 0 . 002 ) and complementation by Rv1986 alone substantially restored the SFC response ( 183 SFC/106 PBMC , IQR 86–285; p = 0 . 002 , when compared to H37RvΔRD-2 . Figure 5A ) . The CD4 deficiency caused by HIV infection is the greatest recognized predisposing factor to tuberculosis and conversely antiretroviral therapy ( cART ) reduces susceptibility by suppressing viral replication and allowing CD4 recovery [29] . We reasoned it would therefore be of interest to compare the IL-2 to Rv1986 and IFN-γ and IL-2 response to CFP-10 and ESAT-6 before and during the course of antiretroviral therapy . As the IFN-γ response to Rv1986 had not been prominent in HIV-1 uninfected persons this was not assayed . The ELISpot response of 19 HIV infected persons without evidence of active tuberculosis was therefore tracked longitudinally over the first 36 weeks of therapy . All patients experienced CD4 increases and suppression of HIV replication during cART . We could not sample all time points and patients for both cytokines due to limitation in the number of cells . Figure 6 shows results of patients whose IL-2 and IFN-γ response to CFP-10 and ESAT-6 was assayed at least twice and 9 patients in whom the corresponding IL-2 response to peptides p61–80 and p161–180 could be determined . Peptide responses were summed for analysis and compared to the values obtained from 20 HIV uninfected persons of similar background , age and sex ( i . e . those shown in Figure 2 ) . The IL-2 response to the peptides of Rv1986 was significantly lower in HIV infected persons prior to cART ( median 24 , IQR 11–43 ) than in HIV uninfected persons ( median 160 , IQR 114–256 , p = 0 . 009 , Figure 6A ) . A significant increase in response occurred during cART therapy such that the median at 36 weeks increased to 106 ( IQR 79–157 , p = 0 . 005 ) . By contrast the IFN-γ response to ESAT-6 and CFP-10 was not significantly lower in HIV infected persons prior to cART ( median 147 , IQR 50–965 ) than in HIV uninfected persons ( 232 , IQR 56–563 , p = 0 . 84 ) . Whilst the median response did increase during cART therapy , the overall trend was not significant p = 0 . 22 , Figure 6B ) . The IL-2 response to the peptides of ESAT-6 and CFP-10 was significantly lower in HIV infected persons prior to cART ( median 2 , IQR 0–31 ) than in HIV uninfected persons ( median 148 , IQR 44–323 , p = 0 . 02 , Figure 6C ) . No significant increase in response occurred during cART therapy . Taken together these findings indicate the decreased IL-2 to Rv1986 response prior to therapy correlates with increased susceptibility better than the IFN-γ response to CFP-10/ESAT-6; and that the partial but significant recovery of IL-2 to Rv1986 , but unchanged IFN-γ response to CFP-10/ESAT-6 also correlates with the recognized decrease in tuberculosis susceptibility that is conferred by cART . We have analyzed whole genome-based transcriptional profiles of M . tuberculosis subject to prolonged hypoxia to guide the discovery of potential antigens . Because the diagnostic potential of species-specific proteins is greatest we focused our initial consideration on two genes upregulated during hypoxia that are absent from all M . bovis strains including BCG by virtue of being RD11-encoded ( Rv2658c and Rv2659c ) [30] , [31] . We also investigated the RD2-encoded Rv1986 because it is absent from most commonly used BCG strains . When compared to the well-characterized immunodominant and species-specific molecules ESAT-6 and CFP-10 , RD11 proteins had inferior ability to restimulate IFN-γ from T cells of persons sensitized by either latent or active tuberculosis . However a striking finding was the immunodominance of Rv1986 for the IL-2 recall response , directed narrowly at two epitopic regions . The quantitative IL-2 response to Rv1986 was several times greater than the corresponding IFN-γ response to either ESAT-6 or CFP-10 ( Figure 1C ) . Our findings suggest Rv1986 to be a major target of long lived CD4+ central memory T cells and that the Rv1986 peptides are associated with a distinct pattern of cytokine production when compared to CFP-10 . There was significant recovery of IL-2 response to the peptides of Rv1986 than of IFN-γ response to ESAT-6 or CFP-10 during the course of cART in HIV infected persons . We also showed that deletion of Rv1986 from the genome of M . tuberculosis substantially decreases its ability to restimulate IL-2 secretion . These interesting findings are potentially important when considering vaccine-induced and natural immunity to tuberculosis and how immunodiagnosis may be improved . One hypothesis we were interested to test is whether , by virtue of upregulation during hypoxia , proteins encoded by such genes would be preferentially recognized by latently infected persons . With the exception of the weak IFN-γ response to Rv2659c pool 2 ( Figure 1A ) this proved not to be the case . Hypoxia does characterize tuberculous granulomas in vivo [18] but it is increasingly re-appreciated that both active and latent tuberculous lesions exhibit a dynamic spectrum of overlapping morphologies [32] , [33] , [34] , [35] , [36] , [37] , [38] and that hypoxic lesions likely occur in both clinical circumstances . A link between transient increases in transcript abundance during hypoxia and the immunogenicity of dosR regulated proteins has also been attempted and the term ‘latency antigen’ has been introduced [20] . A dominant antigenic target that is dosR regulated is Acr1 encoded by Rv2031c and under some assay conditions we , and subsequently others , have documented preferential T cell recognition of Acr1 by latently infected people [19] , [21] . Preferential recognition of Acr1 by latently infected persons was not observed in this study ( Figure 1 ) nor in our previous IFN-γ ELISpot analysis [22] , which is in fact consistent with expression of Acr1 throughout experimental infection [39] , [40] . Conversely a quantitatively higher IFN-γ response to the RD1 encoded CFP-10 and ESAT-6 antigens has sometimes been associated with active disease [41] , [42] , [43] , [44] , attributed to the secretion of these proteins by actively replicating bacilli . We did not however observe a higher response in active tuberculosis compared to latently infected persons . Differences in infection pressures between low and high incidence areas might feasibly contribute to these differences: the clinical environment in which we conducted this study suffers an extraordinarily high tuberculosis incidence of ∼1500/100 , 000 with much ongoing transmission [45] . It is also interesting to note that whilst the transcriptomic data showed a fold decrease in ESAT-6 and CFP-10 during hypoxia , the absolute abundance of these transcripts remained high ( Table 1 ) . Expression of ESAT-6 and CFP-10 under a variety of conditions is in agreement with other in vitro expression data [14] , [46] and adds to data suggesting these molecules may play a role in bacillary persistence as well as active infection [4] . The availability of expression profiles from latently infected human tissue rather than from axenic in vitro culture might provide a better starting point for antigen discovery . Although IFN-γ is essential to human defense against mycobacteria , it is increasingly recognized that assay of PBMC secretion of IFN-g is a poor correlate of protection in field studies of tuberculosis [47] . Greater attention to markers , such as IL-2 , that might better reflect immunological memory is now being paid and formed the basis for our investigation by ELISpot assay of this cytokine [28] , [41] , [48] , although IL-2 secretion itself is not established as a better correlate of protection than IFN-γ . Polyfunctional T cells that secret multiple cytokines are considered a potential correlate of protection in tuberculosis [49] , [50] although the finding that such cells are expanded in tuberculosis patients rather than healthy contacts has been interpreted by some to indicate a role in pathology rather than protection [51] . In this context our finding that Rv1986 was so dominant for the single positive IL-2 response yet elicited modest IFN-γ secretion was striking . The cytokine phenotype of antigen specific T cells is greatly influenced by co-stimulation and the cytokine milieu [52] , [53] . However it has also been suggested that the overall affinity of the TcR-peptide-MHC interaction may play a role as well [54] , [55] . Interestingly an epitope in Rv1986 with an anchor at position 167 ( corresponding with p161–180 ) is predicted for several DRB1*03 , *04 , *08 , *11 and *13 alleles [56]: a promiscuous binding ability that has been noted for other M . tuberculosis epitopes [55] and which might contribute to the almost universal response we observed to this peptide . Rv1986 is a putative membrane protein that is recognized by T cells from M . bovis infected cows [57] . Although the responses to other RD-2 encoded antigens has been previously evaluated in humans [58] , [59] , [60] , [61] , the human T cell response to Rv1986 has not been reported . Our finding that Rv1986 is a dominant target of IL-2 secreting memory T cells suggests that this recall response could contribute to protective immunity . Our findings also bring a novel twist to an old story: the partial and globally variable efficacy of BCG vaccine [62] , [63] . Henao-Tamayo and colleagues recently investigated the vaccine efficacy of BCG Pasteur concluding its ability to induce central memory T cells in the lung was poor perhaps contributing to its partial efficacy [64] . Although another recent study noted no experimental difference in short-term protective efficacy in Guinea Pigs between RD2- negative ( e . g . BCG Pasteur ) and RD2-positive ( e . g . BCG Japan ) strains [65] our discovery that a major target of the human IL-2 response is absent from the most commonly used strains is intriguing . Whilst the in vitro diagnostic potential of the two dominant peptides we have uncovered is considerable , the most important consequence of this work may be to re-evaluate by clinical trials whether BCG strains with and without RD2 vary in clinical efficacy . These techniques have been extensively described before [25] . Briefly , exponential phase cultures grown in rolling culture to an OD600 of 0 . 3 were diluted to a starting OD of 0 . 1 with warm media . This starting culture was transferred to a constantly stirred 1 liter flask , 500 mL of this starting culture per flask . Hypoxia was generated by introducing a constant flow of nitrogen with trace amounts of oxygen ( 0 . 2% O2 ) , leading to bacteriostasis . Samples were taken before hypoxia , at four hours , and after 1 , 4 , and 7 days of exposure to hypoxia . RNA was isolated from these samples using bead beating in the presence of Trizol , followed by chloroform extraction and precipitation of RNA . The RNA was further cleaned using an RNeasy kit purchased from Qiagen . Approximately 3 µg of purified RNA was converted to cDNA using Superscript III ( Invitrogen ) . Aminoallyl dUTP was included in the cDNA reaction , and subsequently conjugated to reactive Cy dye esters . The aerobically growing transcriptional profiles were directly compared to each subsequent hypoxic time point by cohybridization on the same microarray slide . The microarray slides and protocols were provided by the Pathogen Functional Resource Center at the J . C . Ventner Institute as part of their NIAID contract N01-AI-15447 . Slides were scanned with a GenePix 4000B purchased from Axon Technologies . Raw background subtracted intensities were normalized to SigA to provide an approximate measure of transcript abundance . The University of Cape Town research ethics committee approved this study ( REC 296/2007 ) . Written informed consent was provided by study participants . Patients with active or latent tuberculosis were recruited at the Ubuntu clinic at Khayelitsha site B , South Africa . All were of Xhosa ethnicity . Active tuberculosis ( ATB ) was defined by smear positivity for and/or culture of M . tuberculosis from one or more sputum specimens . Latent tuberculosis ( LTBI ) was defined by transverse TST reactivity of >15 mm in response to 2 TU PPD ( RT23 ) at 48–72 hours or an interferon-γ Enzyme linked immunospot ( ELISpot ) response to ESAT-6 or CFP-10 of >20 spot forming cells ( SFC ) /106 PBMC in the absence of clinical symptoms or radiographic abnormality and with a negative sputum smear and culture for M . tuberculosis . All subjects underwent voluntary counseling and testing for HIV-1 infection and positivity was an exclusion criterion . ATB patients were sampled prior to commencing antitubercular chemotherapy . Known immunosuppression for other reasons , age <18 years and pregnancy formed other exclusion criteria . Another group of HIV-1 infected adults who were starting antiretroviral therapy , followed up for 36 weeks were also included as previously described in detail [29] . Patients with ATB and/or HIV infection were treated according to South African national guidelines . The baseline characteristics of subjects enrolled to the study are shown in Table 2 . Peripheral blood mononuclear cells ( PBMC ) were separated over Ficoll . Cells were frozen and stored in liquid nitrogen until analyzed in batches . A total of 2 . 5×105 PBMC were added in 100 µl of RPMI/10%FCS ( R10 ) /well for ELISpot and in 200 µl of R10/well for cell culture . Antigenic stimuli were in the form of pools ( maximum 13 peptides in a pool ) of 20-mer peptides overlapping by 10 residues with each peptide used at a final concentration of 10 µg/ml . Peptides were purchased from Peptide Protein Research Ltd , Oxford , UK and from Pepscan Presto B . V , Netherlands . Peptides were HPLC purified and their mass verified by Mass spectrometry . Control stimuli for ELISpot included anti-CD3 mAb CD3-2 at 100 ng/ml final concentration and unstimulated wells . The interferon-γ ELISpot assay was performed as previously described with slight modifications [29] . Ninety-six well precoated ELISpot plates , mAb 1-D1K ( Pre-coated One-step , Mabtech; 3420-2ATP-10 ) were washed with sterile PBS , blocked with R10 for ≥30 min at room temperature . The blocking medium was removed and the PBMC were set up with respective antigenic stimuli . After incubation for16 h at 37°C with 5% CO2 , plates were washed with PBS , and 100 µl of secondary antibody , mAb 7-B6-1-ALP conjugate at 0 . 5 µg/ml final concentration in PBS containing 0 . 5% FCS was added . After 2 h of incubation at room temperature , 100 µl of filtered ready to use substrate solution ( BCIP/NBT-plus ) was added and developed until spots emerged , washed with tap water and allowed to dry . For the IL-2 ELISpot , 96- well polyvinylidene difluoride membrane based plates , type ELIIP ( MAIPSWU10; Milipore Corp ) , were activated by a brief treatment with 70% ethanol , coated overnight at 4°C with 15 µg/ml of mAb IL2-I ( Mabtech; 3440-2AW-Plus ) , and blocked with R10 for ≥30 min . The blocking medium was removed and the PBMC were set up with respective antigenic stimuli . After 16-h incubation at 37°C in 5% CO2 , the plates were washed , 100 µl of detection antibody ( IL-2-II-biotin ) at 1 µg/ml in PBS containing 0 . 5% FCS added and incubated at room temperature . After 2 hrs , 100 µl of Streptavidin-ALP 1∶1000 in PBS-0 . 5%FCS was added and incubated at room temperature . After 1 h , 100 µl of substrate solution ( BCIP/NBT-plus ) was added and developed until distinct spots emerged . Plates were washed with tap water and allowed to dry . Spot forming cells were enumerated by immunospot counter ( CTL , Cellular Technology Ltd ) and confirmed by microscope ( X4 ) . Results are quoted as cytokine spot forming cells ( SFC ) /106 PBMC . The ELISpot ( IFN-γ and IL-2 ) experiments using live M . tuberculosis strains H37Rv , H37RvΔRD2 and H37RvΔRD2::Rv1986 ( complemented by Rv1986 ) , were performed as described above and previously [66] with 200 , 000 PBMC/well cultured for 16–18 hrs in the presence of 200 , 000 live bacteria/well ( in duplicate wells ) . The MTB H37RvΔRD2 strain ( RD-2 mutant ) was prepared using homologous recombination and sucrose counter-selection as previously described [15] . This mutant was then electroporated with either the empty plasmid pMV306 or the same plasmid into which Rv1986 from H37Rv had been cloned . This gene was expressed under its native promoter . This resulted in the MTB H37RvΔRD2::pMV control ( H37RvΔRD2 ) and MTB H37RvΔRD2::Rv1986 ( complemented with Rv1986 ) strains , which were grown in 7H9 + ADC + 0 . 05% Tween 80 + Kanamycin ( 25 ug/ml ) and preserved as 25% glycerol stocks . 1 . 5–2×106 PBMC were incubated with the two Rv1986 peptides ( residues 61–80 and 161–180 ) at 10 µg/ml each ( i . e . 20 µg/ml peptide in total ) or a pool of 21 peptides from CFP-10 and ESAT-6 at 2 µg/ml each ( i . e . 42 µg/ml peptides in total ) at 37°C . Control stimuli included SEB as positive control at10 µg/ml and unstimulated cells as negative control . After 2 hrs , Brefeldin A at 5 µg/ml ( Sigma , St . Louis , MO ) was added to capture the newly formed cytokines in the Golgi apparatus . After 16 h incubation ( in total ) , the cells were washed with PBS ( 1X ) . For 8 color surface and intracellular staining the cells were first permeabilized , and fixed using Cytoperm/cytofix buffer ( BD ) for 20 min at 4°C , washed with BD Perm/wash and stained with antibody cocktail in BD perm/wash for 1 hr at 4°C . The antibodies used were as follows: CD3-Pacific Blue ( 1 µl/tube ) , CD4 QDot605 ( 0 . 5 µl/tube ) , CD8 Cy5 . 5PerCp ( 3 µl/tube ) , IFN-γ Alexa700 ( 1 µl/tube ) , IL-2 FITC ( 5 µl/tube ) , TNF Cy7PE ( 5 µl/tube ) , CD45RA- APC ( 3 µl/tube ) , CD27-PE ( 3 µl/tube ) , all of which were purchased from BD BioSciences . 106 cells were acquired on LSR II flow cytometer ( BD Bioscience ) . Cell doublets were excluded using forward scatter area vs . forward scatter height parameters . Unstained cells and single-stained mouse calibration beads were used to calculate compensations for every run . Data analysis was performed using FlowJo v 8 . 8 . 2 ( Tree Star ) , Pestle v 1 . 6 . 1 ( NIH ) and Spice v 5 . 05013 ( NIH ) . We defined T cell phenotypes based on the surface markers CD45RA and CD27: Central memory cells ( TCM ) as positive for CD27 and negative for CD45RA; effector memory ( TEM ) are negative for both CD27 and CD45RA and Terminally differentiated T cells ( Tdiff ) are negative for CD27 and positive for CD45RA . The results are expressed as the percentage of CD3+ CD4+ T cells . Bioplex , mixed-to-order panel ( premixed multiplex panel ) from Biorad was used for multiplex cytokine analysis . The assay was carried out according to the manufacturer's instructions . Briefly , the 96- well filter plate was pre- wet with 150 µl of Biorad assay buffer and the buffer removed by vacuum filtration . 50 µl of multiplex bead working solution was added to the wells and the buffer removed . 100 µl of Bioplex wash buffer was added to each well and washed twice and the buffer removed . 50 µl of standard and sample was added to the respective wells , the plate was sealed and then covered by aluminum foil and placed over a microplate shaker . The speed of the shaker was increased to 1100 RPM for 30 sec and then reduced to 300 RPM for 30 min , incubation at room temperature . After incubation , the plates were washed 3 times with Bioplex wash buffer . 25 µl of Bioplex detection antibody working solution was added , and incubated for 30 min as above on the microplate shaker at room temperature . The plates were washed 3 times with Bioplex wash buffer and 50 µl of streptavidin-PE was added , and incubated for 10 min , washed 3 times with Bioplex wash buffer . Beads were resuspended with 125 µl of Bioplex assay buffer , mixed over the microplate shaker at room temperature at 1100 rpm for 30 sec and read on the Bioplex suspension array system . The normality of data was assessed by the D'Agostino and Pearson omnibus test using Graphpad Prism 5 . 0 software ( www . graphpad . com ) . Parametric continuous variables were assessed by student's paired and unpaired t-tests , and non-parametric by Wilcoxon matched pairs , Kruskal Wallis test with Dunn's post test correction or Mann Whitney U tests . Contingency analysis was by Fisher's exact test of probability .
Mycobacterium tuberculosis ( the cause of tuberculosis ) can persist for many years in humans without causing disease but has the potential to reactivate . One of the conditions the bacterium must survive in these circumstances is hypoxia . In order to do so , the bacterium uses a characteristic set of genes that help alter its metabolism . It follows that the products of such genes may encode protein antigens that can be recognized by the immune response . We therefore analyzed gene response patterns of tuberculosis subject to prolonged hypoxia as a guide to the discovery of new antigens that might be useful as vaccines or diagnostic agents . Amongst the genes most strongly increased by low oxygen levels , one was identified ( known as Rv1986 ) that is missing from most strains of the tuberculosis vaccine Mycobacterium bovis BCG . When we analyzed human immune responses to this protein in tuberculosis infected people our experiments showed it was particularly well recognized by cells that produce a chemical messenger ( cytokine ) called interleukin-2 . Interleukin-2 is important for long-term immunological memory . The BCG vaccine is only partially effective and our experiments therefore suggest one of the reasons could be that an important immunological target is missing from many strains . Further evaluation of BCG strains in which Rv1986 is present or absent is therefore warranted in the hope that this might improve the efficacy of existing or new tuberculosis vaccines .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[ "immunology/antigen", "processing", "and", "recognition", "microbiology/immunity", "to", "infections", "immunology/immune", "response", "infectious", "diseases/bacterial", "infections", "immunology/immunity", "to", "infections" ]
2010
Hypoxia Induces an Immunodominant Target of Tuberculosis Specific T Cells Absent from Common BCG Vaccines
Hepadnavirus covalently closed circular ( ccc ) DNA is the bona fide viral transcription template , which plays a pivotal role in viral infection and persistence . Upon infection , the non-replicative cccDNA is converted from the incoming and de novo synthesized viral genomic relaxed circular ( rc ) DNA , presumably through employment of the host cell’s DNA repair mechanisms in the nucleus . The conversion of rcDNA into cccDNA requires preparation of the extremities at the nick/gap regions of rcDNA for strand ligation . After screening 107 cellular DNA repair genes , we herein report that the cellular DNA ligase ( LIG ) 1 and 3 play a critical role in cccDNA formation . Ligase inhibitors or functional knock down/out of LIG1/3 significantly reduced cccDNA production in an in vitro cccDNA formation assay , and in cccDNA-producing cells without direct effect on viral core DNA replication . In addition , transcomplementation of LIG1/3 in the corresponding knock-out or knock-down cells was able to restore cccDNA formation . Furthermore , LIG4 , a component in non-homologous end joining DNA repair apparatus , was found to be responsible for cccDNA formation from the viral double stranded linear ( dsl ) DNA , but not rcDNA . In conclusion , we demonstrate that hepadnaviruses utilize the whole spectrum of host DNA ligases for cccDNA formation , which sheds light on a coherent molecular pathway of cccDNA biosynthesis , as well as the development of novel antiviral strategies for treatment of hepatitis B . Hepadnavirus specifies a group of hepatotropic viruses that carry a single copy of the partially double stranded relaxed circular ( rc ) viral DNA genome in the enveloped virion particle [1] . Hepadnavirus infects mammalian and avian hosts with strict species-specific tropism , including human hepatitis B virus ( HBV ) and duck hepatitis B virus ( DHBV ) [2] . It is estimated that HBV has infected 2 billion people globally , resulting in more than 250 million chronically infected individuals who are under the risk of cirrhosis and hepatocellular carcinoma ( HCC ) [3 , 4] . Upon infection of an hepatocyte , the hepadnaviral rcDNA genome is delivered into the nucleus and converted into an episomal covalently closed circular ( ccc ) DNA , which exists as a minichromosome and serves as viral mRNA transcription template [5 , 6] . One mRNA species , termed pregenomic ( pg ) RNA , is packaged into the cytoplasmic nucleocapsid , where the viral polymerase reverse transcribes pgRNA into viral minus strand DNA , followed by asymmetric plus strand DNA synthesis to yield the major rcDNA genome or a minor double stranded linear ( dsl ) DNA form [7] . The mature nucleocapsid either acquires viral envelope proteins for virion egress , or recycles the viral DNA to the nucleus to replenish the cccDNA reservoir [8] . Therefore , cccDNA is an essential component of the hepadnavirus life cycle for establishing a persistent infection , and cccDNA elimination is an undisputed ultimate goal for a cure of hepatitis B [9] . However , the available drugs for treatment of chronic hepatitis B are rarely curative due to their failure to eliminate cccDNA [10] . Therefore there is an urgent unmet need to fully understand HBV cccDNA biology and develop novel effective treatments to directly target cccDNA formation and maintenance [11 , 12] . Unlike the episomal circular genomes of other DNA viruses , such as papillomaviruses and polyomaviruses [13 , 14] , HBV cccDNA does not undergo semiconservative replication , but is mainly converted from rcDNA [1] . The molecular mechanism by which rcDNA is converted into cccDNA remains obscure . Comparing the major differences between rcDNA and cccDNA ( Fig 1 ) , a series of well-orchestrated biological reactions are required to cope with the terminal molecular peculiarities of rcDNA during cccDNA formation , including: 1 ) completion of viral plus strand DNA synthesis; 2 ) removal of the 5’-capped RNA primer at the 5’ terminus of plus strand DNA; 3 ) removal of viral polymerase covalently attached to the 5’ end of minus strand DNA; 4 ) removal of one copy of the terminal redundancy on minus strand DNA; 5 ) ligation of both strands to generate the wildtype cccDNA [5 , 9] . Previous studies by us and others have identified a rcDNA species without the viral polymerase attachment on the minus strand DNA , namely deproteinized rcDNA ( DP-rcDNA ) or protein-free rcDNA ( PF-rcDNA ) , which is a putative functional precursor for cccDNA [15 , 16] . The molecular mechanism underlying rcDNA deproteinization is largely unknown . Previous studies have demonstrated that the host tyrosol-DNA phosphodiesterase 2 ( TDP2 ) is able to unlink the covalent bond between viral polymerase and rcDNA in vitro , but its role in cccDNA formation remains controversial [17 , 18] . We have further demonstrated that DP-rcDNA is produced in cytoplasmic viral capsid and transported into nucleus by cellular karyopherins [19] . Upon nuclear delivery of DP-rcDNA , it is hypothesized that the host functions , most likely the DNA repair machinery , recognize the gaps/nicks in rcDNA as DNA breaks ( damages ) , and repair it into the perfect circular cccDNA [5 , 9 , 20] . Such an assumption has only been theoretically conceivable and has not been experimentally confirmed until a recent study reported that host DNA polymerase κ ( POLK ) is involved in HBV cccDNA formation during de novo infection [21] . In addition , hepadnaviral dslDNA can also be converted into cccDNA through illegitimate self-circularization after preparing the termini for intramolecular ligation , however , dslDNA-derived cccDNA normally carries insertion-deletion mutations ( indels ) at the junction region , a typical phenotype of host non-homologous end joining ( NHEJ ) DNA repair activity [22–24] . In line with this , we have previously demonstrated that an NHEJ component Ku80 is responsible for cccDNA formation from DHBV dslDNA [25] , thus further confirming the involvement of DNA repair machinery in cccDNA formation . In order to systematically identify the host factors involved in cccDNA formation , we screened 107 human DNA repair genes for their effects on HBV cccDNA formation through shRNA knock-down in HepDES19 cells , and selected gene candidates ( screen hits ) for validation by functional inhibitions in multiple cell systems and assays . We report herein that the cellular DNA ligase ( LIG ) 1 and 3 are responsible for hepadnavirus cccDNA formation from rcDNA , and the conversion of dslDNA to cccDNA requires LIG4 . Such findings will shed light on the molecular mechanism of cccDNA biosynthesis and suggest novel antiviral targets for treatment of chronic hepatitis B . We screened a total of 107 cellular DNA repair genes for their effects on HBV cccDNA production by shRNA knock-down ( see Materials and methods for detailed screening procedures ) , and the primary screening result is summarized in S1 Fig . The screened genes were grouped based on their primarily associated DNA repair pathways . Knock-down of 8 genes showed more than 50% reduction of cccDNA compared to control knock-down , including LIG1 and LIG3 . The low hit rate of the shRNA screen may be due to the incomplete depletion of the targeted genes and/or functional redundancy of the cellular DNA repair genes/pathways . Surprisingly , a large number of gene knock-down resulted in cccDNA upregulation . Such unanticipated observation indicates that certain DNA repair genes may negatively regulate cccDNA formation through rcDNA sequestration or even degradation , which awaits further investigations . We prioritized DNA ligases for functional validation in cccDNA formation for two reasons . Firstly , although the redundant activities of host DNA repair factors may be involved in cccDNA formation , it is likely that the DNA ligation is an essential and final step to seal the nicks/gaps on rcDNA . Secondly , there are only three known DNA ligases in mammalian cells , including LIG1 , LIG3 , and LIG4 [26] , and we have previously ruled out the involvement of NHEJ repair apparatus , which contains LIG4 , in rcDNA to cccDNA conversion [25] . In addition to the above cell-based screening assay , considering that the conversion of rcDNA into cccDNA occurs in the cell nucleus where the DNA repair machinery functions , we established an in vitro cell-free cccDNA formation system with nuclear protein extract . With this assay , we have tested the ability of cccDNA formation from the purified DHBV virion rcDNA . As shown in Fig 2A , although cccDNA could not be directly detected by Southern blot , we were able to detect cccDNA signal by a sensitive PCR assay with primers targeting the sequences outside of the gap region in DHBV rcDNA ( S2A Fig ) , a similar principle has been used for the quantitative detection of HBV cccDNA by real-time PCR [27] . The assay specificity is sufficient enough to distinguish 0 . 3 pg of cccDNA from 30 pg of input rcDNA ( S2B and S2C Fig ) . Sequence analysis of the cccDNA amplicons demonstrated a perfect repair of rcDNA gaps in nuclear extract . It is worth noting that the observed cccDNA PCR signal from the above in vitro cccDNA formation assay might be derived from rcDNA-like template with one strand being repaired into a closed circular DNA . In line with this , a recent study has identified a HBV rcDNA species with a covalently closed minus strand but an open plus strand through digesting the Hirt DNA samples from cell cultures with 3’→5’ exonuclease I and III ( ExoI/III ) , which is possibly an intermediate during rcDNA to cccDNA conversion [28] . We thus attempted to search for such rcDNA-like species that might be generated during the in vitro cccDNA formation reaction by the similar approach . However , neither the minus-strand nor the plus-strand closed circular ssDNA was detected by Southern blot ( S3 Fig ) , indicating that the rcDNA species with covalently closed minus or plus strand in this in vitro cccDNA formation assay , if any , is below the detection limit of Southern blot . Further optimization is needed to improve the assay robustness . Nonetheless , the current in vitro cell-free cccDNA formation assay may still be able to partly recapitulate the nuclear events during rcDNA to cccDNA conversion , which provides a simple and convenient system to assist the study of DNA repair functions in hepadnavirus cccDNA formation . To assess the role of host DNA ligases in hepadnavirus cccDNA formation , LIG1/3 inhibitor L1 and L25 , and pan ligase inhibitor L189 that inhibit the interaction between human DNA ligases and nicked DNA [29] , were added into the in vitro cccDNA formation reaction . As shown in Fig 2B , all the ligase inhibitors tested completely blocked the rcDNA-to-cccDNA conversion in nuclear extract . Taking advantage of the fast kinetics and high productivity of DHBV cccDNA formation , we established a tetracycline ( tet ) -inducible DHBV stable cell line in the background of HepG2 cells , namely HepDG10 , to study hepadnavirus cccDNA formation in human cells . As shown in Fig 3A , DHBV pgRNA , core DNA and cccDNA were rapidly and robustly produced in HepDG10 cells upon withdrawal of tet , cccDNA could be detected by Southern blot as early as day 4 post induction and it accumulated to more significant levels after longer induction . The authenticity of the cccDNA band shown on Southern blot was confirmed by heat denaturation and EcoRI linearization ( S4 Fig ) . Considering that HBV cccDNA formation in cell cultures is extremely time-consuming [15] , the HepDG10 cell line thus provides a robust and convenient cell-based system for assessing human gene functions in hepadnavirus cccDNA formation through loss-of-function approaches ( e . g . chemical inhibitors , gene knock down/out ) , which the relative short assay window may avoid or reduce the potential cytotoxic effects and/or functional redundancy . Consistent with the results from the in vitro nuclear extract-based cccDNA formation assay ( Fig 2B ) , pan-ligase inhibitor L189 also inhibited DHBV cccDNA accumulation in HepDG10 cells without affecting core DNA replication ( Fig 3B ) , further indicating a role of ligase ( s ) in cccDNA formation . The incomplete inhibition of cccDNA formation by L189 might be due to partial inhibition of cellular ligases under the nontoxic concentrations tested in HepDG10 cells . To further assess the role of DNA ligases in cccDNA formation , the expression of LIG1 and LIG3 in HepDG10 cells was completely blocked through gene knock-out by CRISPR/Cas9 ( Fig 4A , top panel ) . The indel mutation sequencing result confirmed the disruption of LIG1 and LIG3 gene at genomic DNA level in the established HepDG10-LIG1 K . O . and HepDG10-LIG3 K . O . cells , respectively ( S5A , S5B , S6A and S6B Figs ) . Depletion of LIG1 or LIG3 did not affect the rcDNA production prominently but resulted in a significant reduction of cccDNA in HepDG10 cells ( Fig 4A ) , suggesting that both LIG1 and LIG3 are involved in cccDNA formation . Similar result was observed with another clone of HepDG10-LIG1 K . O . and HepDG10-LIG3 K . O . cells . The incomplete inhibition of cccDNA formation by knocking out either LIG1 or LIG3 might be due to a redundant function between these two DNA ligases . Consistent with this , Sanger sequencing of cccDNA from the control and LIG1/3 knock-out cells showed wild type DHBV sequence between DR1 and DR2 . We have also attempted to CRISPR-out LIG1 and LIG3 simultaneously in HepDG10 cells , but were unable to obtain clones with complete knock-out of both genes , which might be due to that at least one ligase gene is required for cell viability . However , in a single colony-derived cell line with LIG1/3 double knock-down , which the knock-down efficiency was confirmed by both Western blot ( Fig 4B , top panel ) and T7E1 indel assay ( S7 Fig ) , DHBV cccDNA formation was more profoundly inhibited without affecting core DNA replication ( Fig 4B ) , further confirming the requirement of LIG1/3 in cccDNA synthesis . In addition , the similar phenomena was observed in HepDG10 cells with LIG1/3 single or double knock-down by lentiviral shRNA ( S8 Fig ) . It is of note that the protein-free rcDNA on Hirt DNA Southern blot was also reduced under LIG1 or LIG3 knock-out ( Fig 4 , bottom panels ) . We reason that the concurrent decrease of protein-free rcDNA might be a consequence of cccDNA reduction because the Hirt DNA sample from HepDG10 cells contained a large quantity of nicked cccDNA , which has the indistinguishable electrophoretic mobility as true rcDNA but cannot be mild heat denatured into dslDNA form ( S9 Fig ) . The presence of nicked cccDNA in Hirt extraction has been described in previous studies [16 , 30] . To validate the role of LIG1/3 in HBV cccDNA formation , the individual ligase was knocked out in HBV stable cell line HepDES19 cells by CRISPR/Cas9 ( Figs 5A , S5A–S5C and S6A–S6C ) . Upon tet induction , the cytoplasmic rcDNA remained unchanged in LIG1 or LIG3 knock-out cells compared to control knock-out cells ( Fig 5B ) . cccDNA qPCR assay demonstrated a significant reduction of cccDNA in LIG1/3 knock-out HepDES19 cells ( Fig 5C ) . In order to avoid the contamination of rcDNA in cccDNA quantitation , we developed a method to eliminate rcDNA in Hirt DNA samples before qPCR , which involves a 85°C heat denaturation step that selectively denatures rcDNA into single stranded ( ss ) DNA , followed by Plasmid-safe ATP-dependent DNase ( PSAD ) treatment to remove non-cccDNA templates ( Materials and methods ) ( S10 Fig ) . During natural infection , hepadnavirus cccDNA is formed from both the invading rcDNA in virion and the newly synthesized rcDNA [1] . Because the DHBV or HBV stable cell line used in above studies only makes cccDNA through the rcDNA recycling pathway , we , thus , further assessed the role of LIG1/3 in HBV infection system . To do so , LIG1/3 expression was stably suppressed in HepG2-NTCP12 cells by lentiviral shRNA ( Fig 6A ) . The control and LIG1/3 knock-down HepG2-NTCP12 cells were infected with HBV particles in the presence of nucleoside analogue 3TC which is known to block de novo HBV DNA replication but not the initial cccDNA formation [21 , 31] , making the system suitable for studying the first round cccDNA formation from the input viruses . Upon infection , the levels of HBV cccDNA and core protein were markedly lower in HepG2-NTCP cells with LIG1 or LIG3 knock-down compared to the control knock-down cells ( Fig 6B , 6C and 6D ) , suggesting that LIG1 and LIG3 are also required for the first round HBV cccDNA formation during de novo infection . In order to further validate the role of DNA ligases in hepadnavirus cccDNA formation and rule out the potential off-target effects caused by CRISPR knock-out , we reconstituted LIG1 and LIG3 expression in their corresponding HepDG10 knock-out cells by transfecting plasmid expressing sgRNA-resistant LIG1 and LIG3 gene , respectively ( Fig 7A and 7B , top panels ) . Upon tet induction , DHBV cccDNA formation was successfully rescued in LIG1 and LIG3 knock-out cells by restoring the expression of each ligase ( Fig 7A and 7B , bottom panels ) , which further confirmed that both LIG1 and LIG3 play a critical and specific role in hepadnavirus cccDNA formation . Furthermore , LIG1 or LIG3 was ectopically expressed into their corresponding HepG2-NTCP12 shRNA knock-down cells through transfection , followed by HBV infection in the presence of 3TC . As shown in S11 Fig , restoration of LIG1 or LIG3 expression significantly enhanced HBV infection , as visualized by core immunostaining . With rcDNA serving as the major viral genome DNA form and cccDNA precursor , hepadnavirus replication produces a minor double stranded linear ( dsl ) DNA species , which can also be converted into cccDNA format [1] . We previously reported that Ku80 protein in cellular NHEJ DNA repair pathway is involved in DHBV cccDNA formation from such dslDNA form [25] . In this study , we set out to assess the role of LIG4 , the end effector of NHEJ pathway , in hepadnavirus cccDNA formation through gene knock-out approach . Because the classical CRSIPR/Cas9-based knock-out system requires NHEJ apparatus to introduce indel mutations at the DNA cleavage site , we made use of an alternative microhomology-mediated end-joining ( MMEJ ) DNA repair based CRISPR PITCh system to obtain the LIG4-null cell line [32] ( Materials and methods ) ( S12 Fig ) . As shown in Fig 8A , LIG4 expression was completely depleted in the established LIG4 knock-out HEK293T cells . Next , we transfected the control and LIG4 K . O . cells with either the wildtype DHBV-1S construct or a DSL-DHBV plasmid supporting dslDNA-only replication ( Fig 8B ) . The results demonstrated that the cccDNA formation in the context of wildtype DHBV replication was not affected by knocking out LIG4 ( Fig 8B , comparing lane 2 to lane 1 ) , while the cccDNA formation from dslDNA was completely abolished in the absence of LIG4 ( Fig 8B , lane 4 vs lane 3 ) . Furthermore , restoration of LIG4 expression was able to rescue cccDNA formation in DSL-DHBV transfected LIG4 K . O . cells ( Fig 8C ) . Collectively , the above data strongly supports a conclusion that LIG4 is specifically required for generating cccDNA from the hepadnaviral dslDNA genome . The establishment and persistence of hepadnavirus infection is dependent upon the viral cccDNA , which is a non-replicating episomal viral genome deposited in the nucleus of infected cell after conformational conversion from viral rcDNA [9] . Due to the limited gene-coding capacity of hepadnavirus genome , the virus needs to borrow host functions to complete its lifecycle [8] . The cellular DNA repair is a well-conserved surveillance and restoration system to detect and heal the damage in chromosomal DNA , by which maintains the stability and integrity of the host genome for replication and transcription [33 , 34] . It is plausible that hepadnaviruses hijack the cellular DNA repair apparatus for cccDNA formation by disguising the rcDNA as a “damaged” DNA [9 , 20] . The two gaps on rcDNA would be recognized as lesions for DNA repair by the host , and the DNA termini and their associated modifications are expected to undergo trimming , elongation , and ligation , during cccDNA formation ( Fig 1 ) . However , the host DNA repair pathway responsible for cccDNA formation remain largely unknown , and thus far , only a few host DNA repair enzymes have been reported to be involved in cccDNA formation , including the tyrosol-DNA phosphodiesterase 2 ( TDP2 ) [17] , polymerase κ ( POLK ) and λ ( POLL ) [21] . In this study , we screened 107 host DNA repair factors to assess their individual effect on HBV cccDNA formation by lentiviral shRNA knock-down , and identified and validated the host DNA LIG1 and LIG3 as key factors for hepadnavirus cccDNA formation by using a battery of in vitro and cell-based assays . Such work hence provides new insights into the mechanisms underlying hepadnavirus cccDNA formation in hepatocyte nucleus . As part of the cellular DNA replication and repair machineries , DNA ligases complete joining of DNA strands by catalyzing the phosphodiester bond formation . Specifically , LIG1 ligates the Okazaki fragments during chromosomal DNA synthesis , and it is involved in the ligation steps of homologous recombination repair ( HRR ) , long-patch base-excision repair ( BER ) and nucleotide excision repair ( NER ) ; LIG3 is responsible for sealing single strand DNA breaks during the process of short-patch BER and NER [26] . The involvement of DNA ligases in cccDNA formation indicated that the process of rcDNA termini generates ends that can be ligated and DNA ligases are the end-effectors for sealing the breaks of rcDNA . It is worth noting that previous studies have shown that LIG3 , but not LIG4 , is essential for nuclear DNA replication in the absence of LIG1 [35 , 36]; and LIG1 is a backup enzyme for LIG3 in BER and NER DNA repair pathways [37 , 38] . Such functional redundancy between LIG1 and LIG3 may explain the unaffected cell viability and the incomplete inhibition of cccDNA formation by knocking down/out LIG1 or LIG3 only ( Figs 4A , 5–7 and S8A ) or knocking down both ( Figs 4B and S8B ) . In previous studies , the redundant functions in cccDNA formation have also been observed between POLK and POLL [21] , and perhaps between TDP2 and an unknown TDP2-like protein ( s ) [17] . It is of note that the potential role of TDP2 in cccDNA formation remains controversial . While one study demonstrated that knock-down of TDP2 inhibited , or at least delayed , DHBV cccDNA formation [17]; another study suggested that TDP2 might even serve as a negative regulator of HBV cccDNA formation rather than a facilitator [18] , and a recent study showed that TDP2 chemical inhibitors did not inhibit HBV infection in cell cultures [39] . In our shRNA screen , POLK or TDP2 lentiviral shRNA did not significantly reduce cccDNA formation in HepDES19 cells ( S1 Fig ) , indicating that cellular functional redundancy for each enzyme might also exist in our experimental system . Further validation and mechanistic studies are required to reconcile these results . On the other hand , it is also possible that the first round cccDNA formation from virion DNA during infection and the intracellular cccDNA amplification pathway may have preference for different DNA repair enzymes , or there is hepatic cell line- or clone-specific requirement of host DNA repair factors/pathways for cccDNA formation . We had attempted to CRIPSR out both ligases in HepDG10 cells but only achieved partial double knock-down ( Fig 4B ) , suggesting that at least one of LIG1 and LIG3 is required by the cells , and perhaps by hepadnaviruses as well . However , our data does not completely rule out a possibility that a LIG1/3-independent ligation mechanism might be involved in cccDNA formation , such as DNA topoisomerase I ( TOP1 ) which has been suggested to play a role in rcDNA circularization through its DNA endonuclease and strand transferase activities [40] . Based on the previous and current data , it can be also inferred that hepadnaviruses have evolved to take advantage of the functional redundancy of host DNA repair machinery for a successful cccDNA formation . The mechanism underlying the different efficiency of cccDNA formation between HBV and DHBV remains largely unknown , but likely in a virus-specific but not host-specific manner [16 , 25 , 30] . Based on that , we created the cell-free and the human hepatoma cell-based DHBV system to facilitate the identification and validation of host and viral regulators of cccDNA formation ( Figs 2 and 3 ) . In addition to the possible determining factors for cccDNA formation in the steps of rcDNA maturation , deproteinization , nuclear importation and uncoating , whether the cellular DNA repair system differentially recognizes and repairs nuclear HBV and DHBV rcDNA into cccDNA remains obscure . In this study , we found that both viruses employ LIG1 and LIG3 for cccDNA formation ( Figs 4–7 , S8 and S11 ) , suggesting that the different repair process of HBV and DHBV rcDNA , if any , should be at the steps upstream of rcDNA end joining . Though LIG1 and LIG3 have overlapping functions , knocking out/down of LIG3 resulted in relatively lower level of cccDNA than LIG1 knock-out/down ( Figs 4–7 and S8 ) , which suggests that LIG3 may play a more important role in cccDNA formation . In line with this notion , the two separated nicks/gaps in rcDNA are reminiscent of single strand breaks , which are preferable substrates for LIG3 in BER- and NER-mediated single strand break repair ( Fig 1 ) . Moreover , during the primary screen , two other BER components , APEX1 and POLB [41] , emerged as candidates for positive regulator of cccDNA formation ( S1 Fig ) , further suggesting the potential involvement of short-patch BER in cccDNA formation , which awaits further systematic investigations . With the protein and RNA attachments at the 5’ end of minus- and plus-strand , respectively , hepadnavirus rcDNA is not a typical DNA break substrate for the major known repair pathways , and it is unknown whether the two gaps in rcDNA are repaired simultaneously or separately , including the final ligation step . A recent study revealed a nuclear rcDNA species with a covalently closed minus strand but an open plus strand , indicating that the nick on minus strand may be sealed first during cccDNA formation [28] . However , we did not observe an increased accumulation of protein-free rcDNA after blocking cccDNA synthesis in LIG1 or LIG3 knock-out cells ( Figs 4 , 6 , 7B and S8 ) . This phenomena may be due to a fact that the processed rcDNA ready for ligation is unstable . Previous studies have shown that the Hirt DNA samples from DHBV replicating cells contain high levels of nicked cccDNA which might be generated intracellularly or during the Hirt extraction [16 , 30] . We also found that the protein-free rcDNA in Hirt extraction from HepDG10 cells were largely nicked cccDNA ( S9 Fig ) , indicating that the observed reduction of protein-free rcDNA in LIG1/3 knock-out cells might be a consequence of cccDNA reduction ( Figs 4 and 7 ) . Nonetheless , further characterization of the nuclear rcDNA in LIG1/3 knock-out cells will provide further information for understanding the biological processes of rcDNA termini prior to the final ligation step during cccDNA formation . In parallel with the bona fide rcDNA-to-cccDNA formation during hepadnavirus infection , the viral dslDNA byproduct is also repaired into cccDNA with indel mutations at the joint region [16 , 23] . Although the dslDNA-derived cccDNA is generally defective of initiating a new round of viral DNA replication , it remains functional to express HBsAg and thus may play a role in viral pathogenesis . Based on the linear format of dslDNA and the indel mutations of its cccDNA derivative , it is hypothesized that dslDNA is a substrate for host error-prone NHEJ DNA repair system , and we have previously reported that another NHEJ component Ku80 is required for DHBV cccDNA formation from the dslDNA but not rcDNA [25] . LIG4 is the DNA ligase responsible for performing the last step of double strand DNA end joining in the NHEJ pathway [24] . In this study , we have demonstrated that LIG4 plays an essential role in cccDNA formation from DHBV dslDNA , and no functional redundancy was observed between LIG4 and other ligases ( Fig 8 ) . In addition , it has been reported that the chromosome DNA double strand breaks are targets for DHBV DNA integration [42] , which indicates that the NHEJ machinery , including LIG4 , is also responsible for the integration of hepadnavirus dslDNA into host genome . Altogether , our study revealed a critical role of cellular DNA ligases in hepadnavirus cccDNA biosynthesis . Another possible function of DNA ligases in hepadnavirus life cycle can be to maintain the integrity of cccDNA , provided the preexisting cccDNA undergoes DNA damage and the host cell is able to repair it . Based on our observations , the DNA ligase inhibitors , which are currently under development for anti-cancer therapy [43] , may be developed into host-targeting antiviral means to treat chronic hepatitis B by blocking cccDNA formation and/or repair . HepG2 and 293T cells were purchased from ATCC and cultured in DMEM/F12 medium ( Gibco ) supplemented with 10% fetal bovine serum , 100 U/ml penicillin and 100 μg/ml streptomycin . The tetracycline-inducible HBV ( Genbank accession number: U95551 ) stable cell line HepDE19 and HepDES19 were established previously [15] , and maintained in the same way as HepG2 , but with the addition of 1 μg/ml tetracycline ( tet ) and 400 μg/ml G418 . When required , the culture medium was switched to tet-free to initiate HBV replication in HepDE19 and HepDES19 cells . HBV infectious particles were collected from the supernatant of HepDE19 cells , and the infection of HepG2-NTCP12 cells and HBV core protein ( HBc ) immunofluorescence microscopy were conducted according to a previously published protocol [44] . DHBV virions were purified from the serum of virally infected ducks as previously described [19] . DNA ligase 1/3 inhibitors L1 ( 5- ( methylthio ) thiophene-2-carboxylic acid ) and L25 ( 2 , 3-dioxoindoline-7-carboxylic acid ) , and the pan ligase inhibitor L189 ( 6-Amino-2 , 3-dihydro-5-[ ( phenylmethylene ) amino]-2-4 ( 1H ) -pyrimidineone ) were purchased from Tocris Biosciences . Lamivudine ( 3TC ) was kindly provided by Dr . William Mason ( Fox Chase Center Center ) . DHBV total RNA in cell cultures was extracted by TRIzol ( Invitrogen ) and detected by Northern blot [25] . HBV and DHBV cytoplasmic core DNA and whole cell Hirt DNA were extracted and analyzed by Southern blot as previously described [15 , 45 , 46] . HBV total DNA and cccDNA qPCR were performed according to the literature with modifications [47–49] . Firstly , HBV total DNA in the total Hirt DNA sample , including protein-free rcDNA and cccDNA , was first quantified by qPCR with 0 . 8 μM of forward primer ( 5’-CCGTCTGTGCCTTCTCATCTG-3’ ( nt 1551–1571 ) ) , 0 . 8 μM of reverse primer ( 5’-AGTCCAAGAGTYCTCTTATGYAAGACCTT-3’ ( nt 1674–1646 ) ) , and 0 . 2 μM of TaqMan probe ( 5’-FAM- CCGTGTGCACTTCGCTTCACCTCTGC-TAMRA-3’ ( nt 1577–1602 ) ) . In the meantime , the relative cellular mitochondrial DNA ( COX3 gene ) level in each Hirt DNA samples was quantified by SYBR green qPCR with forward primer 5’-CCCTCTCGGCCCTCCTAATAACCT-3’ and reverse primer 5’-GCCTTCTCGTATAACATCGCGTCA-3’ . Next , to reduce the contamination of HBV rcDNA in the qPCR detection of cccDNA , the Hirt DNA sample was first heated at 85°C for 5 min to denature rcDNA into single-stranded DNA , followed by Plasmid-safe ATP-dependent DNase ( PSAD ) ( Epicentre ) treatment at 37°C for 16 h . The PSAD reaction was then stopped by heat inactivation at 70°C for 30 min , and the samples were further purified by DNA clean-up spin column ( Zymo Research ) . Real-time PCR amplification of 2 μl cleaned cccDNA sample was performed in a 20 μl reaction containing 0 . 9 μM forward primer ( 5’-ATGGAGACCACCGTGAACGCCC-3’ ( nt 1610–1631 ) ) , 0 . 9 μM reverse primer ( 5’-TCCCGATACAGAGCTGAGGCGG-3’ ( nt 2021–2000 ) ) , and 0 . 2 μM TaqMan probe ( 5’-FAM-TTCAAGCCTCCAAGCTGTGCCTTGGGTGGC-TAMRA-3’; nt 1865–1894 ) . The cccDNA qPCR primers were designed to target the HBV DNA sequences outside of the gap region in rcDNA and to avoid PCR amplification of the integrated HBV genome in HepDES19 cells . The FastStart Essential DNA Probes Master ( Roche ) and FastStart Universal SYBR Green Master ( Roche ) were used to assemble TaqMan and SYBR Green qPCR reactions , respectively . The qPCR was run by Roche LightCycler 96 under the following thermal cycling conditions: 10 min at 95°C , followed by 15 sec at 95°C and 1 min at 61°C for 50 cycles . The cccDNA qPCR data was normalized by the cellular mitochondrial DNA quantitation . The supercoiled cccDNA from DHBV transfected 293T cells were extracted by a previously developed alkaline lysis method with minor modifications [22] . Briefly , cells in a 6-well-plate were lysed in 200 μl lysis buffer containing 10 mM Tris-HCl ( pH7 . 4 ) , 1 mM EDTA , and 0 . 2% NP40 , at room temperature for 10 min . The lysate was mixed gently with 200 μl alkaline lysis buffer ( 0 . 1 M NaOH , 6% SDS ) and incubated at 37°C for 30 min , followed by adding 100 μl 3 M KAc ( pH5 . 0 ) and mixing gently . After incubating on ice for 10 min and centrifuging at 12 , 000 rpm for 5 min , supernatant was collected and extracted by phenol twice . DNA was precipitated by ethanol and subjected to Southern blot assay . The customized Mission lentiviral shRNA DNA repair gene family set was purchased from Sigma-Aldrich . The library contains 586 lentiviral shRNA targeting 140 DNA repair genes of all the known DNA repair pathways except for NHEJ , with an average of 3–5 different shRNA sequence against each target gene . The virus stocks were aliquoted in 96-well-plate with viral titer ranging from 1×107 to 3×107 g . e/ml . The TDP2 lentiviral shRNA was purchased from Santa Cruz Biotechnology and added into the library . To establish stable DNA repair gene knock-down cell lines . HepDES19 cells were infected with the pooled lentiviral shRNA targeting the same DNA repair gene or control lentiviral shRNA in the presence of tet , 2 days later , the cells were selected by puromycin ( 3 μg/ml ) for 1 week and the antibiotics-resistant cells were pooled and expanded into cell lines . The cells transduced by different lentiviral shRNA exhibited variable growth rate during puromycin selection but all were viable . A total of 107 DNA repair gene lenti-shRNA transduced cell lines were obtained . To assess the effect of RNAi on cccDNA production , the knock-down cells under confluent condition ( 1×106 cells per 35mm-dish ) were cultured in the absence of tet for 10 days , total Hirt DNA was extracted and subjected to Southern blot or HBV total DNA and cccDNA qPCR . It is known that the HBV Hirt DNA level is positively related to HBV DNA replication level , especially the cytoplasmic rcDNA level [15 , 16 , 19] . To assess the efficiency of cccDNA formation under gene knock-down , the relative cccDNA levels in DNA repair gene knock-down cells compared to control knock-down cells were normalized by total HBV Hirt DNA signals , which indicates a relative rcDNA-to-cccDNA conversion rate . Plasmid pTREHBVDE , the vector delivered the HBV transgene in HepDE19 cells , has been described previously [15] . DHBV-1S is a plasmid supporting DHBV ( Genbank Accession No . : K01834 ) DNA replication upon transfection into cell cultures [50] . Plasmid 1Sdsl-3 ( renamed to DSL-DHBV in this study ) was a derivative of DHBV-1S with an artificial point mutation of G2552C in viral genome that supports double stranded linear ( dsl ) DNA replication but fails to make rcDNA [25 , 51] . Plasmid pcLIG1-FLAG expressing human DNA ligase 1 ( LIG1 , GenBank Accession No . : NM_000234 ) with C-terminal FLAG-tag was purchased from Genescript ( clone ID: OHu14319 ) . To construct ligase 3 ( LIG3 ) expression plasmid pcLIG3 , the ORF of LIG3 was PCR amplified from MGC human LIG3 sequence-verified cDNA ( GE Healthcare Dharmacon , clone ID: 6092747 ) by forward primer 5’-5’-CGGGATCCATGTCTTTGGCTTTCAAGATCTTCTT-3’ ( BamHI site is underlined ) and reverse primer 5’-GCTCTAGACTAGCAGGGAGCTACCAGTCTCCGTTT-3’ ( XbaI site is underlined ) , and cloned into the BamHI/XbaI restricted pcDNA3 . 1 vector ( Invitrogen ) . Plasmid pcLIG4-FLAG expressing the C-terminal FLAG-tagged human Ligase 4 ( LIG4 ) was purchased from Genescript ( clone ID: OHu13291 ) . DHBV virion DNA which contain predominantly rcDNA and a minor portion of dslDNA were purified from serum derived virions and quantified by Southern blot using DHBV DNA marker as standard according to published literature [52] . DHBV cccDNA was extracted from Dstet5 cells and gel purified as previously described [15] . The nuclear extract was prepared from HepG2 cells and stored in aliquots following a published protocol [53] . To assemble the DNA repair reaction , the purified DHBV virion DNA was mixed with 4 μl nuclear extract in 200 μl reaction buffer containing 20 mM HEPES , 80 mM KCl , 10 mM MgCl2 , 1 mM ATP , 1 mM DTT , and 50 μM dNTPs , and incubated at 37°C for 30 min . To stop the reaction , 10 μl of 1% SDS , 20 μl of 0 . 5 M EDTA and 10 μl of 10 mg/ml pronase were added and incubated at 37°C for an additional 30 min . Next , the mixture was subjected to phenol and phenol:chloroform extraction , and viral DNA was precipitated down by ethanol and dissolved in 10 μl nuclease-free H2O . The obtained viral DNA was analyzed by Southern blot or cccDNA-specific PCR . The PCR reaction was assembled by mixing 0 . 5 μl DNA sample , 12 . 5 μl 2× PCR buffer ( Clontech ) , 1 μl of 20 μM forward primer ( 5’-GCCAAGATAATGATTAAACCACG-3’ ) , 1 μl of 20 μM reverse primer ( 5’-TCATACACATTGGCTAAGGCTC-3’ ) , 0 . 5 μl Terra polymerase ( Clontech ) , and 9 . 5 μl H2O . The DNA was amplified by 22 cycles of heat denaturation at 95°C for 30 sec , annealing at 55°C for 30 sec , and extension at 72°C for 30 sec . The PCR product was subjected to agarose gel electrophoresis and stained by ethidium bromide . To detect the possible closed minus strand or plus strand after in vitro cccDNA formation in nuclear extract , exonucleases Exo I and III ( ExoI/III ) were used to degrade DNA strands with a free 3’ end and preserve closed circular DNA in either single-stranded ( SS ) or double-stranded ( DS ) form as described previously [28] . Briefly , 5 ng of DHBV rcDNA were subjected to the in vitro cccDNA formation reaction as described above . After reaction , the recovered DNA were dissolved in 20 μl water and treated with 0 . 25 μl each of Exo I and Exo III at 37°C for 2 h in 1×NEB Cutsmart buffer . 5 ng of DHBV rcDNA without going through in vitro cccDNA formation reaction served as positive control for ExoI/III digestion . The digestion products were directly subjected to electrophoresis and Southern blotting for hybridization with p32-labeled DHBV minus- or plus-strand specific riboprobe . Plasmid pTRE-GFP-DHBV , which bidirectionally supports the tet-inducible expression of GFP and DHBV pgRNA , was constructed as follows . Firstly , a DNA fragment , which , in the 5’ to 3’ orientation , contained the partial sequence ( nt 12–430 ) of pBI vector ( Clontech , GenBank Accession No . : U89932 ) and the reverse complementary sequence of GFP ORF followed by SV40 polyadenylation signal , with unique PstI and KpnI restriction site at the 5’ and 3’ end , respectively , was chemically synthesized ( Genescript ) . Then the PstI/KpnI restricted fragment was cloned into the same endonuclease treated pTREHBVDE plasmid to generate pTRE-GFP-HBV . Next , another DNA fragment containing nt 425–468 sequence from pBI , a spacer sequence ( 5’-GCAGAGCTCGTTTGATC-3’ ) , and DHBV sequence ( nt 2524-3021/1 ) , with unique KpnI and EcoRI site at 5’ and 3’ end , respectively , was chemically synthesized ( Genescript ) , and the fragment was inserted into the KpnI/EcoRI sites of pTRE-GFP-HBV to generate pTRE-GFP-DHBV-EcoRI-HBV . The HindIII site in the backbone sequence downstream of the remaining HBV sequence in pTRE-GFP-DHBV-EcoRI-HBV was further mutated to SalI site by using QuikChange II Site-Directed Mutagenesis Kit ( Agilent Technologies ) to obtain pTRE-GFP-DHBV-EcoRI-HBV-SalI . One unit length of DHBV genome was amplified from DHBV-1S by PCR with forward primer ( 5’- CGGCTAGAATTCATGCTCATTTGAAAGCTT-3’ , nt 3011-3021/1-19 , DHBV EcoRI site is underlined ) and reverse primer ( 5’-AATTAAGTCGACAATTCTAGCCGTAATCGGATA-3’ , nt 3021–3001 , non-DHBV SalI site is underlined ) and digested by EcoRI and SalI and cloned into the same sites in pTRE-GFP-DHBV-EcoRI-HBV-SalI to generate the final product pTRE-GFP-DHBV . To establish tetracycline-inducible DHBV stable cell lines , HepG2 cells were cotransfected by pTRE-GFP-DHBV and pTet-off which expresses tet-responsive transcriptional activator ( tTA ) ( Clontech ) with 7:1 molar ratio . The transfected HepG2 cells were selected with 500 μg/ml G418 in the presence of 1 μg/ml tet . G418-resistant colonies were picked and expanded into cell lines . To determine DHBV positive cell lines , the candidate clones were cultured in 96-well-plate with tet-free medium for 6 days and subjected to fluorescence microscopy to select GFP-positive clones . Then , the GFP-positive cells were lysed in 1% NP40 and the cytoplasmic lysate was subjected to dot blotting as described previously [47] . The dot blot was hybridized by α-32P-UTP ( 800 Ci/mmol , Perkin Elmer ) labeled minus strand specific full-length DHBV riboprobe , and the obtained DHBV positive cell lines were further assessed for their tet-inducible DHBV core DNA replication and cccDNA production by Southern blot . A DHBV cccDNA highly producing cell line clone was named HepDG10 . The maintenance and induction of HepDG10 cells were performed in the same way as HepDES19 cells . Lentiviral particles expressing U6 promoter-driven shRNA for knocking down human LIG1 or LIG3 were purchased from Sigma-Aldrich . The shRNA coding sequences for knocking down LIG1 and LIG3 are listed in S1 Table . HepDG10 and HepG2-NTCP12 cells were transduced by above lentiviral LIG1 or LIG3 shRNA or control shRNA per manufacturer’s instruction . The transduced cells were selected with 3 μg/ml puromycin and the antibiotics-resistant cells were pooled and expanded into cell lines . The knock-down levels of LIG1 and LIG3 in knock-down cell lines were assessed by Western blot by using antibodies against LIG1 ( sc-271678 , clone C5 , Santa Cruz Biotechnology ) and LIG3 ( sc-390922 , clone E7 , Santa Cruz Biotechnology ) , respectively , and compared to control knock-down cells . β-actin served as loading control for Western blot by using anti-actin antibody ( MAB1501 , clone C4 , Millipore ) . To construct a lenti-vector expressing both LIG1 and LIG3 shRNA , a chemically synthesized DNA fragment containing the H1 promoter sequence and downstream LIG3 shRNA sequence ( S1 Table ) was cloned into the unique EcoRI site of lenti-shRNA plasmid DNA pLKO . 1-LIG1 ( TRCN0000048494 , Sigma ) , giving rise to the bicistronic lenti-shRNA plasmids with two orientations of the H1-shLIG3 cassette right downstream of the original U6-shLIG1 cassette . The head-to-tail and tail-to-tail dimer clones were named pLKO . 1-LIG1/3C1 and pLKO . 1-LIG1/3C2 , respectively , and pLKO . 1-LIG1/3C2 was used to prepare lentiviral shRNA particles with MISSION Lentiviral Packaging Mix ( Sigma ) . HepDG10 cells were transduced with lentiviral shLIG1/3 and the puromycin-resistant cells were pooled and expanded into stable cell line , namely HepDG10-shLIG1/3 . The double knock-down of LIG1 and LIG3 were determined by Western blot comparing to the aforementioned control knock-down cells . LIG1 and LIG3 knock-out cell lines were generated through CRISPR-mediated genome editing of LIG1 and LIG3 gene loci . The single guide ( sg ) RNAs targeting two different sites of human LIG1 and LIG3 gene were designed at http://www . e-crisp . org/E-CRISP and shown in S5 and S6 Figs . In addition to the general criteria for sgRNA design [54] , the sgRNAs were designed to target either the 5’ end of the ORF or the functional domain coding sequences of LIG1 or LIG3 . Furthermore , the designed sgRNA sequences do not possess any possible CRISPR sites in DHBV or HBV sequences . The synthetic sgRNA oligo pairs ( S2 Table ) were annealed and cloned into BbsI-digested lentiCRISPRv2 control vector ( Addgene # 52961 , gift from Dr . Feng Zhang ) . Lentivirus preparations were performed according to the protocols from Dr . Feng Zhang’s Lab ( genome-engineering . org ) . Briefly , each lentivector was co-transfected with packaging plasmids psPAX2 and pMD2 . G ( Addgene# 12260 and 12259 , respectively , gift from Dr . Didier Trono ) in molar ratio of 4:3:1 into 293T cells by Lipofectamine 2000 ( Invitrogene ) , and 48 h later , media was collected , centrifuged at 1 , 000 × g for 10 min , filtered through a 0 . 45um filter , and virus titers were determined by lentiviral titration kit . Lentiviral transduction of HepDG10 or HepDES19 and antibiotics-selection were performed as above described to generate control and LIG1/LIG3 stable knock-out cell lines , specifically HepDG10 LIG1 K . O . , HepDG10 LIG3 K . O . , HepDES19 LIG1 K . O . , and HepDES19 LIG3 K . O . cells . The LIG1 or LIG3 knock-out phenotype was confirmed by Western blot and indel sequencing assay . The corresponding coding sequences for epitopes of LIG1 and LIG3 antibodies have no overlap with the gene targeting sites of the designed sgRNAs . To knock out both LIG1 and LIG 3 in HepDG10 cells , the cells were transduced with lentiviruses encoding CRISPR/Cas9-LIG1-sgRNA1 and CRISPR/Cas9-LIG3-sgRNA1 together ( 1:1 ratio ) . The puromycin selection and clone screening was performed as described above . The indel mutations of sgRNA targeting sites in LIG1 and LIG3 gene loci was detected by T7E1 assay . For indel sequencing analysis of LIG1 and LIG3 genes , total genomic DNA from the control and LIG1 or LIG3 knock-out cells were extracted using DNeasy blood and tissue kit ( Qiagen ) according to the manufacturer’s protocol . The genomic sequence region covering the CRISPR target site was amplified by PCR using the indel detection primers ( S3 Table ) and cloned into T vector pMD19 ( Clontech ) for Sanger sequencing . The LIG1/3 DNA sequence from control and knock-out cells was aligned to determine the CRISPR-induced mutations . For T7E1-based indel assay , primers used to amplify DNA fragments containing LIG1-sgRNA1 and LIG3-sgRNA1 targeting region were listed in S3 Table , indel mutations were detected by Guide-it Mutation Detection Kit ( Clontech ) according to manufacturer’s manual . Because LIG4 is an essential component in NHEJ DNA repair pathway in eukaryotes [55 , 56] , the conventional NHEJ-based CRISPR/Cas9 knock-out system is not able to generate LIG4 knock-out cell lines . Therefore a newly developed microhomology-mediated end-joining ( MMEJ ) repair based CRISPR/Cas9 knock-in system [32] was used to establish LIG4 knock-out cells . Briefly , annealed sgRNA targeting the last exon of LIG4 gene ( S12 Fig and S4 Table ) was inserted into pX330A-1×2 ( Addgene# 58766 , gift from Dr . Takashi Yamamoto ) to obtain pX330A-1×2-LIG4-gRNA , and after Golden Gate assembly using BsaI ( New England Biolabs ) , the cassette of PITCh-gRNA from pX330S-2-PITCh ( Addgene# 63670 , gift from Dr . Takashi Yamamoto ) was inserted into pX330A-1×2-LIG4-gRNA to generate the All-in-One pCRISPR/Cas9-PITCh-LIG4-gRNA vector containing both LIG4 gRNA and PITCh-gRNA . Then , pCRIS-PITChv2-LIG4 was constructed based on pCRIS-PITChv2-FBL ( Addgene# 63672 , gift from Dr . Takashi Yamamoto ) by performing two separate PCR , one to amplify the vector backbone and one to amplify LIG4-specific microhomology arm containing EGFP-2A-Puro knock-in cassette . Generic 5′-reverse and 3′-forward primers were used for vector backbone amplification , and LIG4-specifc primers containing the desired microhomologies were used to amplify the insertion ( S4 Table ) . The above two purified PCR fragments were conjugated by using the In-Fusion HD cloning kit ( Clontech ) to generate plasmid pCRIS-PITChv2-LIG4 . To generate LIG4 knock-out cell line , 293T cells were transfected with pCRIS-PITChv2-LIG4 and the All-in-One plasmid pCRISPR/Cas9-PITCh-LIG4-gRNA with a molar ratio of 1:2 , followed by puromycin ( 1 μg/ml ) selection . The puromycin-resistant colonies were pooled together and subjected to fluorescence microscopy for determining EGFP-positive cells with successful knock-in , and then the knock-out efficiency of LIG4 was determined by Western blot using antibodies against LIG4 ( sc-28232 , clone H-300 , Santa Cruz Biotechnology ) . Control knock-in 293T cells were made by transfecting pCRIS-PITChv2-FBL and pX330A-FBL/PITCh ( Addgene# 63671 , gift from Dr . Takashi Yamamoto ) that target human fibrillarin ( FBL ) gene . The “NGG” protospacer adjacent motif ( PAM ) sequence of LIG1 sgRNA1 locates just ahead of the targeted exon ( S5 Fig ) , therefore plasmid pcLIG1-FLAG was directly used in the function rescue experiment by transfecting the HepDG10-LIG1 K . O . cells . To avoid the integrated lentiviral CRISPR/Cas9-LIG3 sgRNA system targets LIG3 ectopic expression plasmid , the LIG3 sgRNA1 corresponding PAM motif “CGG” in pcLIG3 was mutated to “CGA” by Q5 Site-Directed Mutagenesis Kit ( New England Biolabs ) with primers ( forward: 5’-AACTAGAGCGaGCCCGGGCCA-3’ , reverse: 5’- TCTCAAACATGCATTTAATGTGGTACCAC-3’ ) but without changing the amino acid sequence of LIG3 . The sgRNA-resistant pcLIG3 was used to transfect HepDG10-LIG3 K . O . cells in the function rescue experiment . Because the LIG4 knock-out 293T cells were made by transient transfection-mediated knock-out , pcLIG4-FLAG was used directly in rescue experiment by transfection .
Hepadnavirus cccDNA is the persistent form of viral genome , and in terms of human hepatitis B virus ( HBV ) , cccDNA is the basis for viral rebound after the cessation of therapy , as well as the elusiveness of a cure with current medications . Therefore , the elucidation of molecular mechanism of cccDNA formation will aid HBV research at both basic and medical levels . In this study , we screened a total of 107 cellular DNA repair genes and identified DNA ligase 1 and 3 as key factors for cccDNA formation from viral relaxed ( open ) circular DNA . In addition , we found that the cellular DNA ligase 4 is responsible for converting viral double-stranded linear DNA into cccDNA . Our study further confirmed the involvement of host DNA repair machinery in cccDNA formation , and may reveal new antiviral targets for treatment of hepatitis B in future .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "methods" ]
[ "molecular", "probe", "techniques", "enzymes", "enzymology", "plasmid", "construction", "dna", "replication", "dna", "construction", "dna", "molecular", "biology", "techniques", "gel", "electrophoresis", "ligases", "extraction", "techniques", "research", "and", "analysis", "methods", "electrophoretic", "techniques", "electrophoretic", "blotting", "artificial", "gene", "amplification", "and", "extension", "proteins", "molecular", "biology", "biochemistry", "southern", "blot", "dna", "extraction", "nucleic", "acids", "genetics", "biology", "and", "life", "sciences", "dna", "repair", "polymerase", "chain", "reaction" ]
2017
The role of host DNA ligases in hepadnavirus covalently closed circular DNA formation
TGF-β involvement in Chagas disease cardiomyopathy has been clearly demonstrated . The TGF-β signaling pathway is activated in the cardiac tissue of chronic phase patients and is associated with an increase in extracellular matrix protein expression . The aim of this study was to investigate the effect of GW788388 , a selective inhibitor of TβR1/ALK5 , on cardiac function in an experimental model of chronic Chagas’ heart disease . To this end , C57BL/6 mice were infected with Trypanosoma cruzi ( 102 parasites from the Colombian strain ) and treated orally with 3mg/kg GW788388 starting at 120 days post-infection ( dpi ) , when 100% of the infected mice show cardiac damage , and following three distinct treatment schedules: i ) single dose; ii ) one dose per week; or iii ) three doses per week during 30 days . The treatment with GW788388 improved several cardiac parameters: reduced the prolonged PR and QTc intervals , increased heart rate , and reversed sinus arrhythmia , and atrial and atrioventricular conduction disorders . At 180 dpi , 30 days after treatment interruption , the GW3x-treated group remained in a better cardiac functional condition . Further , GW788388 treatment reversed the loss of connexin-43 enriched intercellular plaques and reduced fibrosis of the cardiac tissue . Inhibition of the TGF-β signaling pathway reduced TGF-β/pSmad2/3 , increased MMP-9 and Sca-1 , reduced TIMP-1/TIMP-2/TIMP-4 , and partially restored GATA-6 and Tbox-5 transcription , supporting cardiac recovery . Moreover , GW788388 administration did not modify cardiac parasite load during the infection but reduced the migration of CD3+ cells to the heart tissue . Altogether , our data suggested that the single dose schedule was not as effective as the others and treatment three times per week during 30 days seems to be the most effective strategy . The therapeutic effects of GW788388 are promising and suggest a new possibility to treat cardiac fibrosis in the chronic phase of Chagas’ heart disease by TGF-β inhibitors . Chronic chagasic cardiomyopathy ( CCC ) is the most common form of non- ischemic cardiomyopathy and one of the main causes of complications and death in Latin America , where the disease is endemic . It is estimated that ~7 million people are infected by Trypanosoma cruzi worldwide [1] , but this is an underestimated screen as serology for Chagas disease is still not included in Public Health control programs in many countries . Chagas disease ( CD ) is caused by infection with T . cruzi parasites and presents an acute phase followed by a chronic phase , during which organ damage ( mainly cardiac ) can be observed in approximately one third of the patients [2] . CCC is a complex disease including host-parasite interactions contributing to an inflammatory and fibrotic scenario differing from other heart pathologies . As fibrosis is a major trait of CCC , specific anti-fibrotic therapies represent an alternative or complementary option to improve prognosis of this debilitating disease . Transforming growth factor ( TGF-β ) is a pleiotropic cytokine with strong pro-fibrotic properties that has been shown to actively contribute to cardiac damage in several fibrotic disorders [3] . Interestingly , patients with atrial fibrillation present an overexpression of TGF-β in atrial tissue [4] and atrial fibrillation accompanied by myocardial fibrosis predisposes to arrhythmia events [5] . TGF-β is secreted under a latent form by almost all types of cells and needs to be activated into its mature form by different molecules such as thrombospondin , integrins or matrix metalloproteases [6] . To develop its biological functions , mature active TGF-β must bind to its membrane receptors , known as TGF-β receptor-type I ( TβRI/ALK5 ) and -type II ( TβRII ) . Ligand binding stimulates the phosphorylation of intracellular proteins of the classical pathway , Smad2/3 , and some alternative pathways Erk , JNK , p38 , PI3K [7] . We and others have previously demonstrated the involvement of TGF-β in CD physiopathology [8–12] . Chagas disease patients presenting more severe forms of heart disease progression have higher levels of circulating TGF-β [8 , 11] . Single nucleotide polymorphisms in the TGF-β gene ( -509 C<T and +10 T<C ) were shown to be a risk factor for CD susceptibility , at least in the Latin American population [13 , 14] . Moreover , chronic chagasic patients with higher TGF-β levels present a worse clinical outcome after 10 years of follow up [15] . To understand the role of TGF-β in the physiopathology of CD , our group developed both in vitro and in vivo experimental models of the acute phase of the disease with important reproducible clinical features [16–19] . It was demonstrated that TGF-β favors T . cruzi cell invasion and the parasite intracellular cycle [16] . Data obtained in vitro with cardiomyocytes infected by T . cruzi and treated with anti-TGF-β compounds were confirmed in vivo , with an experimental model of acute phase infection [17–19] , in which we observed reduced parasitemia followed by reduced cardiac damage and extracellular matrix deposition [18 , 19] . One century after the initial description of CD , therapy has made little progress and is still based on two trypanossomicidal drugs: nifurtimox and benznidazole ( Bz ) . Recently , the BENEFIT randomized trail evaluated the efficacy of Bz on the clinical outcome of patients with CCC . This study showed a reduction of parasite load in serum but a lack of clinical effect on cardiac condition through a 5 year-follow-up [20] . The cardiac form of CD is characterized by progressive congestive heart failure with an inflammatory response , involving many cell populations such as CD4+ and CD8+ T cells , which triggers cardiac remodeling and myocardial fibrosis [21 , 22] , associated with sudden cardiac death [23] . Current options for the treatment of CCC are based on generic therapeutic strategies that do not differ from those in other cardiomyopathies: diuretics , beta-blockers , angiotensin-converting enzyme inhibitors , and spironolactone [24–28] . A pre-clinical assay using mouse models of CCC has been performed using anti-inflammatory agents , such as anti-TNFα compounds [29] . In the present study , we have used a model of chronic Chagas disease [30–32] induced over several months following the injection of a low inoculum of T . cruzi ( 102 parasites/mouse ) , to investigate whether GW788388 [33] , an oral compound that inhibits TGF-β receptor kinase activity , could reverse heart fibrosis and electrical conduction defects . This model reproduces much more closely the human pathological situation than the acute infection models used in previous studies [18 , 19] . All mice procedures were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the Brazilian National Council of Animal Experimentation ( http://www . cobea . org . br/ ) and the federal law 11 . 794 ( 8 October 2008 ) . Protocols used in this study were approved by the Institutional Committee for Animal Ethics of Fiocruz ( CEUA/Fiocruz , Licenses LW10/14 and LW42-11 ) . All efforts were made to minimize animal suffering . Four- to six weeks old female C57BL/6 ( H-2b ) mice were obtained from the animal facilities of the Oswaldo Cruz Foundation ( CECAL/Fiocruz , Rio de Janeiro , Brazil ) . Animals were housed for at least one week before parasite infection at the Cardoso Fontes Animal Facility/IOC under environmental factors and sanitation according to “Guide for the Care and Use of Laboratory Animals” . The Colombian strain of T . cruzi parasites was maintained by serial passage in mice every 35 days post-infection ( dpi ) and parasitemia was employed as a parameter to establish acute and chronic phases using 5 μL of blood obtained from the tail vein [29] . For all experimental procedures , C57BL/6 mice were infected by intraperitoneal injection of 100 blood trypomastigotes of the Colombian strain of T . cruzi [30] . The compound GW783388 ( GlaxoSmithkline , France ) or vehicle dilution buffer ( 4% DMSO , 96% [0 . 5% Hydroxypropylmethylcellulose ( HPMC ) , 5% Tween 20 , 20% HCl 1M in NaH2PO4 0 . 1M] ) was used in oral administration . Mice received GW788388 at 3 mg/kg from 120 dpi , when electrical abnormalities and fibronectin deposition in the heart tissue are detected [30 , 31] , to 150 dpi by gavage in three administration schemes ( 0 . 2 mL ) : single dose and one or three administrations per week for 30 days . The control group received vehicle buffer using the same scheme . Mice were divided into the following groups respecting the limit of 5 animals per cage: untreated non-infected ( NI ) , Non-infected and GW788388-treated ( NI+GW1x ) , infected and GW788388 untreated ( T . cruzi ) and infected and GW788388 treated , using 3 treatment schemes: single dose ( SD ) ; once ( GW1x ) and thrice ( GW3x ) a week . ECG recording and analysis were performed in all groups of infected and non-infected animals . Mice were intraperitoneally tranquilized with diazepam ( 20 mg/Kg ) , fixed in the supine position and the transducers were carefully placed subcutaneously according to chosen preferential derivation ( DII ) . Traces were recorded using a digital system ( Power Lab 2/20 ) connected to a bio-amplifier at 2 mV for 1 s ( PanLab Instruments , Spain ) . Filters were standardized between 0 . 1 and 100 Hz and traces were analyzed using the Scope software for Windows V3 . 6 . 10 ( PanLab Instruments , Barcelona , Spain ) . ECG parameters were recorded for at least 2 min and evaluated in the chronic phase at 120 , 150 and 180 dpi , using the following standard criteria: the heart rate , monitored by beats/minute ( bpm ) , and the variation at P wave and PR , QRS and correct QT intervals ( QTc ) , all measured in milliseconds ( ms ) . The ECG parameters were analyzed as previously described [29] . For analysis of cardiac function , ECHO recording and analysis were performed in all groups . Mice were anesthetized ( inhalation route ) with 1 . 5% isoflurane gas in 100% oxygen with flow 1L/minute , trichotomized in precordial region and examined with a Vevo 770 ultrasound apparatus ( Visual Sonics , Canada ) coupled to a 30 MHz transducer . Left ventricular ejection fractions ( LVEF ) were determined using Simpson’s method and left and right ventricular areas ( LV and RV ) were obtained in B-mode using a short axis view at the level of the papillary muscles . The estimation of TGF-β serum concentrations in samples of non-infected and infected animals at 120 and 150 dpi was performed using a TGF-β1 specific commercial enzyme linked immunosorbent assay ( ELISA ) kit ( Quantikine TGF- β1 ELISA , R&D Systems , USA ) according to the manufacturer’s instructions . Extraction of protein from frozen heart tissue was performed as previously described [12] . Proteins were analyzed by immunoblotting with specific primary antibodies against SMAD2/3 ( Cell Signaling– 8685 ) , pSMAD2/3 ( Cell Signaling– 3101 ) , Fibronectin ( Sigma–F3648 ) , Collagen type 1 ( Novotec , France , kindly provided by Dr . Daniella Areas Mendes-da-Cruz , IOC/Fiocruz ) , Timp-1 ( Sigma–SAB4502971 ) , Timp-2 ( Sigma-AB2965 ) , Timp-4 ( Sigma- T8312 ) . To confirm equal protein loading , the same membranes were stripped and reprobed with an antibody against GAPDH ( Ambion–AM300 ) . 40 μg of protein were loaded and separated on 12% SDS-PAGE with 0 . 1% gelatin incorporated as substrate . After running , gels were soaked in a sequence of baths ( 15 minutes in 2 . 5% Triton X-100 followed by 15 minutes in 2 . 5% Triton X-100 / 50 mM Tris-Cl pH 7 . 5 and 10 minutes , twice , in 50 mM Tris-Cl pH 7 . 5 ) , under constant shaking . Gels were incubated overnight at 37°C in a 50 mM Tris-Cl pH 7 . 5 / 10 mM CaCl2 solution and then stained with 0 . 5% Coomassie brilliant blue R-250 and scanned in a GS-800 scanner ( BioRad ) . The molecular masses of MMP-9 were estimated in Quantity One software ( BioRad ) by comparison with standards of PageRuler Plus Prestained Protein Ladder ( Thermo Scientific ) . In this study , RNA and DNA were extracted from the same heart tissue sample , using TRIzol . RNA and DNA were used to gene expression analysis and Parasite Load quantification , respectively . Following the TRIzol protocol , after the chloroform addition and separation in aqueous and organic phases , the aqueous phase was collected to extract RNA and DNA was extracted from the organic phase , following manufacture’s protocol . The parasite load estimation by qPCR was performed by absolute quantification , based on a standard curve produced from DNA samples extracted from 20 mg of heart tissue of a non-infected mice , spiked with 106 parasites . The standard curve was built from the serial dilution of DNA , ranging from 106 to 1 parasite equivalents . For parasite quantification , the qPCR reactions were carried out with 5 μL DNA; 10 μL FastStart Universal Probe Master Mix [2X] ( Roche ) ; 750nM cruzi1 ( 5′ASTCGGCTGATCG TTTTCGA3′ ) and cruzi2 ( 5′AATTCCTCCAAGCAGCGGATA3′ ) primers and 50nM cruzi3 probe ( 5′FAM-CACACACTGGACACCAA-NFQ-MGB3′ ) , specific for the satellite region of the nuclear DNA of T . cruzi . In parallel , cardiac tissue amount was estimated by the quantification of mouse GAPDH , using the Pre-developed TaqMan Assay Reagents Mouse GAPDH [20X] ( Applied Biosystems– 4352339E ) , at the final concentration of 1X , following manufacture’s protocol . The reactions were performed in ABI Prism 7500 Fast device ( Applied Biosystems ) . The PCR cycling conditions were: a first step at 95°C for 10 min , followed by 40 cycles at 95°C for 15 s and 58°C for 1 min . The parasite load was calculated by T . cruzi equivalents/mice heart tissue equivalents ratio and expressed as “Parasite load/cardiac GAPDH” . Extraction of total RNA from frozen heart tissue and the reverse transcription was carried out as previously described [34 , 12] . RT-qPCR was performed using TaqMan gene expression assays for MMP2 ( Mm00439498-m1 ) , MMP9 ( Mm00442991-m1 ) , GATA-4 ( Mm00484689-m1 ) , GATA-6 ( Mm00802636-m1 ) , Tbox-5 ( Mm00803518-m1 ) , Nkx2-5 ( Mm00657783 ) , Desmin ( Mm00802455-m1 ) , Titin ( Mm00658612-g1 ) , Troponin T ( Mm00449089-m1 ) and the endogenous housekeeping control genes glyceraldehyde 3-phosphate dehydrogenase GAPDH ( Mm99999915-g1 ) and β actin ( Mm00607939-s1 ) , which were purchased from Life Technologies ( USA ) . The reactions were performed and analyzed as previously described [34] . Fixed tissue was dehydrated and embedded in paraffin . Sections ( 3 μm ) were stained by Masson’s trichrome as previously described [19] , for fibronectin , pSMAD2/3 , connexin-43 , Sca-1+ and DAPI detection by immunofluorescence and for CD3+ cells by immunohistochemistry analysis . Sections were observed using a Nikon microscope coupled with image acquisition systems ( Nikon ) and the images were assessed for percentage area of collagen using CellProfiler image analysis software ( http://www . cellprofiler . org ) . Mice spleen specimens were obtained from the all groups . Spleen samples were processed for flow cytometric analysis within 1h from harvesting , using a simple and rapid procedure . The tissue was first disrupted with mechanic process into cell culture medium . The cell suspension was then washed and resuspended in phosphate-buffered saline ( PBS ) . Cell viability was assessed using Trypan Blue . For flow cytometric analysis , aliquots of cells were stained with saturating amounts of conjugated antibodies: CD4 ( Southern ) , CD8 ( BioLegend ) , CD44 ( BioLegend ) , CD62L ( BioLegend ) , CD49d ( eBioscience ) , CD11a ( Southern ) , CD45R ( BioLegend ) . Samples were run and analyzed on a FACSCalibur instrument ( BD Biosciences ) . Differences between infected and non-infected groups were considered statistically significant when $p< 0 . 05 , $ $p<0 . 01 , and $ $ $P< 0 . 001 and differences between infected mice GW788388 treated or not are indicated by *P< 0 . 05 , **P< 0 . 01 , and ***P< 0 . 001 , as determined by GraphPad Prism 4 . 0 software ( GraphPad Software Inc . , San Diego , CA , USA ) . All the analyses were performed using the non-parametric Mann–Whitney test . The chronic model of T . cruzi infection was set up as previously described [30 , 31] . This experimental model of CCC uses a different parasite strain ( Colombian , DTU-Tc I ) and a more resistant mouse strain ( C57BL/6 ) than in our previous experiments on acute infection [19] . C57BL/6 mice were infected by intraperitoneal injection of 100 blood trypomastigotes of the Colombian strain of T . cruzi . Infected mice were monitored by measuring the presence of circulating parasites during the peak of parasitemia at 42 dpi ( as previously shown in 30 ) . ECG parameters were evaluated in the chronic phase at 120 and 150 dpi , using the following standard criteria: ( i ) the heart rate was monitored by beats/minute ( bpm ) , and ( ii ) the variation at P wave and PR , and QTc intervals , all measured in milliseconds ( ms ) . ECG analysis demonstrated that at 120 dpi , all mice presented a significant decrease in heart rate , as measured by beats per minute ( bpm ) ( S1A Fig ) , associated with significant increase of P wave duration ( S1B Fig ) , PR ( S1C Fig ) and QT ( S1E Fig ) and QTc ( S1F Fig ) intervals and no difference in QRS interval ( S1D Fig ) , when compared with sex- and age-matched non-infected ( NI ) controls . Atrioventricular block type 1 ( AVB1 ) and type 2 ( AVB2 ) events occurred in 80% of the mice ( S1G Fig ) and arrhythmia was observed in 100% of the infected animals ( S1H Fig ) . Therefore , at 120 dpi the group of T . cruzi-infected mice show pivotal electrical abnormalities , as previously shown [30] . We tested the effects of the GW788388 compound using three different administration schemes starting at 120 dpi: i ) a single dose ( GW ) ; ii ) one dose per week ( GW1x ) or iii ) three doses ( GW3x ) per week during 30 days ( until 150 dpi ) . Heart parameters were measured at the end of the experiment at 150 dpi . All treatment schemes improved P wave duration and PR interval ( Table 1 ) whereas all but the single dose treatment decreased the prolonged QTc intervals ( Table 1 ) . In contrast , only the weekly treatments ( GW1x and GW3x ) were able to improve the heart rate ( Table 1 ) . In all further experiments , we only used two schemes for GW788388 treatment , once or thrice a week , for 30 days . When comparing the groups of mice before ( at 120 dpi ) and after the period of treatment ( at 150 dpi ) , we observed that non-treated mice presented a significant decrease of heart rate and an increase in P wave duration , PR and QTc intervals ( Fig 1A–1D ) . GW1x-treated mice presented a better cardiac rhythm , P wave duration , PR and QTc intervals ( Fig 1A–1D ) while GW3x-treated mice presented a significant decrease in P duration and QTc ( Fig 1A–1D ) as compare to non- treated mice . Importantly , we observed a reduction of AVB1 and AVB2 events after GW788388 treatment , as stated by ECG registers ( Fig 1E ) . Six out of 18 mice ( 33% ) treated with GW788388 once a week avoided sinus arrhythmia and 11 out of 18 mice ( 60% ) reversed AVB2 events . From the group of mice treated with GW788388 thrice a week , 19 out of 30 mice ( 60% ) avoided AVB1 events and 13 out of 30 mice ( 42% ) avoided AVB2 events ( Table 1 ) . Together , these data show that GW788388 treatment significantly improved heart function in infected mice . We next investigated the mechanism of action of this compound . We tested whether the positive effect on the heart rate could be mediated by rearrangement of gap junctions , more precisely connexin 43 ( Cx43 ) -enriched plaques known for their importance in heart electrical conduction [35] . During the chronic phase of T . cruzi infection , mice treated with GW788388 once or three times per week presented better-organized Cx43-enriched plaque distribution ( Fig 1F and 1G ) , possibly contributing to the improvement of electrical conduction ( Table 1 ) . Moreover , considering that Cx43 loss is established at 120 dpi [31] , our data support that GW788388 therapy reverses this pattern . As TGF-β has been described to influence T . cruzi cell invasion and the parasite intracellular cycle [17] , in order to assess the effect of the inhibition of TGF-β pathway on parasite load , we verified T . cruzi DNA quantity from the heart tissue of infected mice . As expected , we observed the presence of T . cruzi DNA in the heart of chronic infected mice . Interestingly , both treatments with GW1x and GW3x at 150 dpi did not modify the parasite load , suggesting that the inhibition of TGF-β pathway does not interfere with T . cruzi control on chronic infection ( Fig 2A ) . Recently , we have shown that T . cruzi-infected mice presented higher levels of circulating TGF-β during the acute phase of infection [12] . It is also known that TGF-β is involved in heart fibrosis of CD patients who present increased levels of serum TGF-β [8] . In the present mouse model of chronically Colombian-infected mice , levels of circulating TGF-β were increased at 120 dpi ( p<0 . 05 ) and 150 dpi ( 2-fold increase; p<0 . 01 ) . Although treatment with GW788388 once a week for four weeks did not impact TGF-β levels , GW788388 administration thrice a week from 120 dpi to 150 dpi significantly decreased TGF-β concentrations in serum ( p<0 . 001 ) ( Fig 2B ) . Then , we verified whether the canonical TGF-β signaling pathway was activated in cardiac tissue after T . cruzi infection . We investigated the phosphorylation pattern of Smad2/3 in heart extracts . We observed that chronic infection by T . cruzi increased expression of total Smad2/3 and its phosphorylation , in the heart of infected mice as compared to non-infected animals ( Fig 2C–2F ) . The treatment with GW788388 once or thrice a week decreased pSmad2/3 cardiac levels ( Fig 2C , 2D and 2F ) and the nuclear accumulation of pSmad2/3 , indicating that the canonical TGF-β signaling pathway was down-regulated under these conditions ( Fig 2C ) . Our group previously demonstrated the involvement of TGF-β in the development of cardiac fibrosis due to CD , both in experimental models of T . cruzi-infected mice during the acute phase [12 , 18 , 19] and in patients during the chronic phase [8] . Here , we investigated the expression of the extracellular matrix proteins fibronectin and collagen type I and found an accumulation of both proteins in response to chronic T . cruzi infection in the ventricular heart tissue , observed at 120 and 150 dpi ( Fig 3A–3D ) . Collagen expression was increased in the heart of infected mice as observed at 120 dpi and even more at 150 dpi . Interestingly , inhibition of TGF-β signaling by GW788388 administration using both schemes ( once and thrice a week ) , significantly decreased extracellular proteins expression after 30 days of treatment , demonstrating its capacity to reverse cardiac fibrosis ( Fig 3A–3D ) . Immunostaining for fibronectin ( Fig 3D ) and Masson´s trichrome staining for collagen deposition corroborated these data ( Fig 4 ) . We observed an increase in collagen deposition , visualized as light blue staining in the heart of the infected mice at 120 and 150 dpi as compared to non-infected animals ( Fig 4 ) . Moreover , GW788388 treatment clearly reduced collagen staining , suggesting cardiac tissue recovery . The process of fibrosis in response to a tissue damage could begin during the long-term injury stimulus , in which a loss of balance between the production and degradation of the ECM components is observed , leading to the gradual replacement of the functional tissue by a connective tissue [36] . On the other hand , decreased expression of extracellular matrix proteins could be associated to fibrosis reversion with replacement of functional tissue , indicating cardiac recovery . Then , we aimed to observe if GW788388 treatment could also ameliorate cardiac function and structure by echocardiogram ( ECHO ) . We investigated the percentage of the left ventricle ejection fraction ( LVEF ) by ECHO and observed that infected mice presented a reduced LVEF at 150 dpi ( from 62 , 5% to 51 , 3% ) , corroborating previous data [32] . Importantly , GW788388 treatment once a week significantly reversed heart pumping to normal values ( Fig 5A ) , reaching ~60% LVEF . We observed that other ECHO parameters were also altered , such as: i ) LV internal diameter ( LVID ) and ii ) LV stroke volume ( Table 1 ) . All parameters were significantly decreased at 150 dpi and GW treatment reversed to normal conditions ( Fig 5A and 5B , Table 1 ) . After GW788388 treatment cessation at 150 dpi , we followed up two different groups of T . cruzi infected mice non-treated and GW3x-treated for 30 days to verify whether therapy effects were sustained . At 180 dpi , we observed a low grade of LVEF in non-treated infected mice ( ~40% ) , while in the GW3x-treated group remained as the non-infected ( ~60% ) . Representative ECHO images from each group of mice are demonstrated in Fig 5B . n = 2–6 mice per group . To investigate the possible mechanisms involved in the reversal of heart electrical and functional abnormalities and fibrosis , we evaluated the involvement of the metaloproteinases MMP-2 and -9 activities in this process . At 150 dpi , T . cruzi infection significantly reduced MMP-9 mRNA expression and activity , ( Fig 6A , 6C and 6D ) . GW788388 treatment , in the two schemes ( once or thrice a week ) , significantly increased MMP-9 mRNA levels and its enzymatic activity ( Fig 6A , 6C and 6D ) . On the other hand , MMP-2 transcription and activity were not affected ( Fig 6B and 6E ) . As metaloproteinases are highly regulated and their transcript levels cannot be directly correlated to their activity , we investigated some of the most important MMPs regulators such as TIMP-1 , -2 and -4 . In order to understand the differences in MMPs expression and activities , we verified if one of the intrinsic regulators of MMPs , tissue inhibitor of matrix metaloproteinases -1 , -2 and -4 ( TIMP-1 , TIMP-2 and TIMP-4 ) presented altered protein expression in the heart , directly by Western blotting assays . We demonstrated that chronic T . cruzi infection induced the expression of TIMP-1 , TIMP-2 and TIMP-4 in heart tissue ( Fig 6D–6G ) . In GW788388-treated infected mice , both schemes significantly reduced the expression of TIMP-1 ( ~45% ) , TIMP-2 ( ~65% ) and TIMP-4 ( ~80% ) ( Fig 6D–6G ) . Next , we checked whether the effect of this therapy on reversion of heart fibrosis was associated with induction of heart recovery through recruitment or differentiation of cardiomyocytes . Thus , we investigated the mRNA levels of cardiac cell markers such as GATA-4 , GATA-6 , Nkx2-5 , Tbox-5 , troponin T , titin and desmin ( Fig 7A–7G ) . Almost all cardiac markers presented significantly reduced expression after T . cruzi infection , except for Nkx2-5 and desmin ( Fig 7A–7G ) . At 150 dpi , only GW3x -treatment significantly increased GATA-6 and Tbox-5 expression in the cardiac tissue of chronically T . cruzi-infected mice ( Fig 7B and 7D ) . In order to confirm the cardiac recovery process , we also analyzed the stem cell antigen-1 ( Sca-1+ ) , a marker for cardiac stem cell , directly in the heart tissue [37] . The presence of cells Sca-1+ was rare in the heart of non-infected mice and after 120 dpi . At 150 dpi , Sca-1+ cells were observed in the heart tissue , being more evident in infected mice treated with GW3x ( Fig 7H , white arrows ) , supporting that GW788388 therapy stimulated the arrival of stem cells with cardiac phenotype . At 180 dpi , 30 days after treatment interruption , the presence of Sca-1+ cells were still observed in the group of mice treated with GW3x from 120 to 150 dpi ( Fig 7H , white arrows ) . Altogether , the increased levels of GATA-6 and Tbox-5 mRNA and the presence of Sca-1+ cells suggested the emergence of cells with high potential of cardiac phenotype which could indicate cardiac recovery . During the chronic phase , inflammatory infiltrates and T . cruzi antigens are observed in the heart tissue , both processes could lead to cardiac damage and prominent fibrosis . The inflammatory infiltrate is composed mainly of T cells [29 , 32 , 34] . Thus , we also analyzed if the inhibition of TGF-β pathway could interfere on the inflammatory profile in our experimental model of chronic cardiomyopathy induced by T . cruzi infection . To this end , we assessed the presence of CD3+ cells in the heart and spleen tissues . Non-infected and GW-treated mice showed an increase in the number of CD3+ cells in the spleen , clearly demonstrating the potential of GW therapy on interfering on immune cells . Although this clear effect in avoiding the TGF-β control of T cells proliferation , GW treatment in the absence of T . cruzi infection did not altered the subset of memory T cells but decreased its activation profile , inhibiting the migration to the heart tissue as expected as no injury in the heart was observed . As already described [30 , 31] after T . cruzi infection , it was observed an increase in the frequency of CD3+ cells either in the spleen and in the heart ( Fig 8 ) . Interestingly , GW treatment three times a week decreases the frequency of CD3+ cells in the heart , which could indicate an effect in the migratory capacity of these cells from the spleen . T . cruzi infection promotes splenomegaly in this mice model [30] , in both acute and chronic phases , reproducing aspects of Chagas disease [31] . Here , we confirmed symptoms of spleen enlargement during the chronic infection and showed that GW treatment partially reversed this process ( Table 2 ) . Moreover , we investigated the CD4+ and CD8+ cell profile in the spleen of studied mice: naive , T central memory ( TCM ) and T effector memory ( TEM ) . We observed that T . cruzi chronic infection increased CD8+ cells from all CD8+ subtypes: CD44high+/CD62L- ( TEM ) , CD44high-/CD62L+ ( naïve ) and CD44high+/CD62L+ ( TCM ) . Regarding the CD4+ cells , no difference was observed , but analysis of the different CD4+ subtypes showed that CD44high+/CD62L- , CD44high-/CD62L+ and CD44high+/CD62L+ were increased ( Table 2 ) . The GW788388 treatment did not alter the increased expression of CD8+ cell profiles , but GW1x treatment showed a tended to decrease the number of these cells ( p = 0 . 053 ) . We also observed that GW1x treatment decreased the frequency of CD4+/CD44high-/CD62L+ , CD4+/CD44high+/CD62L+ , CD8+/ CD44high-/CD62L+ , CD8+/ CD44high+/CD62L- and CD8+/ CD44high+/CD62L+ , and GW3x only modulated the TCM CD4+/CD44high+/CD62L+ ( Table 2 ) . Among the activation molecules evaluated ( CD49d , CD11a and CD45R ) , both T lymphocytes , CD4+ and CD8+ , have increased activation profile with T . cruzi infection . Moreover , both schemes of GW treatments decreased this activation profile in at least one of the studied molecules ( Table 2 ) . This important pre-clinical proof of concept study places TGF-β regulator compounds as an alternative strategy for the treatment of heart fibrosis , once GW788388 treatment improved ECG and ECO profiles; modulated TGF-β levels and its intracellular proteins; reversed the loss of connexin-43 intercellular plaques; reduced fibrosis of the cardiac tissue; restored cardiac recovery markers and reduced the migration of CD3+ cells to the heart . Furthermore , at 180 dpi , 30 days after treatment interruption , the GW3x-treated group remained in a better cardiac functional condition . These data are promising and the inhibition of TGF-β pathway could be an important and relevant strategy of therapy for cardiac fibrosis during the chronic phase of Chagas’ heart disease .
TGF-β is a key molecule in many physiological processes as well as pathologies . We have previously described the role of TGF-β in Chagas disease , caused by the eukaryotic protozoan parasite Trypanosoma cruzi . Besides the high disease burden in many countries , one of the severe aspects of Chagas disease is the chronic heart condition developed by patients , for which there is no specific treatment . In search for a better treatment , we have investigated the potential of the TGF-β signaling blocker , GW788388 , on disease hallmarks in a mouse model of chronic Chagas’ heart disease . Oral administration of GW788388 produces a global reversion of cardiac damage , and , remarkably , reduces fibrosis , one of the most important death-associated features in chronic Chagas’ heart disease .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[ "medicine", "and", "health", "sciences", "cardiovascular", "anatomy", "tropical", "diseases", "fibrosis", "parasitic", "diseases", "parasitic", "protozoans", "animal", "models", "organisms", "developmental", "biology", "electrocardiography", "model", "organisms", "protozoans", "experimental", "organism", "systems", "bioassays", "and", "physiological", "analysis", "neglected", "tropical", "diseases", "research", "and", "analysis", "methods", "animal", "studies", "mouse", "models", "protozoan", "infections", "electrophysiological", "techniques", "trypanosoma", "cruzi", "signal", "transduction", "trypanosoma", "eukaryota", "chagas", "disease", "cardiac", "electrophysiology", "anatomy", "cell", "biology", "tgf-beta", "signaling", "cascade", "biology", "and", "life", "sciences", "cell", "signaling", "heart", "signaling", "cascades" ]
2019
TGF-β inhibitor therapy decreases fibrosis and stimulates cardiac improvement in a pre-clinical study of chronic Chagas’ heart disease
The metacercariae of the Clonorchis sinensis liver fluke excyst in the duodenum of mammalian hosts , and the newly excysted juveniles ( CsNEJs ) migrate along the bile duct via bile chemotaxis . Cholic acid is a major component of bile that induces this migration . We investigated the neuronal control of chemotactic behavior of CsNEJs toward cholic acid . The migration of CsNEJs was strongly inhibited at sub-micromolar concentration by dopamine D1 ( LE-300 and SKF-83566 ) , D2 ( spiramide , nemonapride , and sulpiride ) , and D3 ( GR-103691 and NGB-2904 ) receptor antagonists , as well as a dopamine reuptake inhibitor ( BTCP ) . Neuropeptides , FMRFamide , peptide YY , and neuropeptide Y were also potent inhibitors of chemotaxis . Meanwhile , serotonergic , glutamatergic , and cholinergic inhibitors did not affect chemotaxis , with the exception of fluoxetine and CNQX . Confocal immunofluorescence analysis indicated that dopaminergic and cholinergic neurons were colocalized in the somatic muscle tissues of adult C . sinensis . Our findings suggest that dopaminergic neurons and neuropeptides play a major role in the chemotactic migration of CsNEJs to bile , and their inhibitors or modulators could be utilized to prevent their migration from the bile duct . Clonorchis sinensis is the most common human liver fluke in East Asia , with more than 200 million people at risk of infection [1] . It primarily inhabits the bile duct , where it can induce serious pathological inflammatory changes , and chronic infection is associated with the development of cholangiocarcinoma [2] . Infections occur via the consumption of raw freshwater fish carrying C . sinensis metacercariae . The ingested metacercariae excyst in the duodenum , and the newly excysted juveniles ( hereafter termed CsNEJs ) quickly migrate to the intrahepatic bile duct in response to chemotactic cues in the bile [3 , 4] . In C . sinensis metacercariae , bile stimulates the expression of genes for energy generation to induce migration and those associated with maturation [5] . The presence of a bile acid transporter also indicates the role of bile for their survival and migration [6] . Indeed , cholic acid in bile has been found to be a primary attractant for CsNEJs to migrate into the intrahepatic bile duct [7] . The nervous system governing the cholic acid-induced chemotactic movement is , however , completely unknown . Helminthic locomotion , invasion and attachment are related to the neuromuscular system [8–12] . Neurotransmitters such as dopamine , serotonin , glutamate , acetylcholine , and neuropeptides have been observed to control the behavior . However , these processes are not well described in C . sinensis . In this study , we used a pharmacological approach to elucidate the neuronal control over the chemotaxis of CsNEJs to cholic acid . A New Zealand White rabbit ( 2 . 2 kg ) was purchased from Samtako Bio Korea Inc . ( Osan , South Korea ) . Animal was handled in an accredited Chung-Ang University animal facility in accordance with the AAALAC ( Association for Assessment and Accreditation of Laboratory Animal Care ) International Animal Care policies ( Accredited Unit , Korea FDA; Unit Number 36 ) . Approval for animal experiments was obtained from the Institutional Review Board of Chung-Ang University animal facility ( Approval Number CAU-2011-0053 ) . Topmouth gudgeon ( Pseudorasbora parva ) , the second intermediate host of C . sinensis , were purchased at a fish market in Shenyang , Liaoning Province , People’s Republic of China . Fish were ground and digested in artificial gastric fluid ( 0 . 5% pepsin , pH 2 . 0 , MP Biochemicals Co . , Solon , OH ) for 2 h at 37°C [13] . Solid matter was removed from digested content by filtration through a sieve of 212 μm mesh diameter . C . sinensis metacercariae were collected using sieves of 106 and 53 μm mesh diameter and washed with 0 . 85% saline , then collected under a dissecting microscope and stored in phosphate-buffered saline ( PBS ) at 4°C until use . The metacercariae were excysted in 0 . 005% trypsin ( Difco , Sparks , MD ) and used as experimental CsNEJs for downstream assays . A multi-trough bile chemotaxis assay panel was used per a previous study’s design [7] . Eight semi-cylindrical troughs 10 cm long , 1 cm wide , and 0 . 5 cm deep were carved into a polycarbonate block . Each trough was graduated 0 at the center , +1 to +5 cm on the left side , and −1 to −5 cm on the right side . In all chemotaxis assays , each trough was filled with 1 mL of 1× Locke’s buffer as a base solution [3] with various concentrations of the test compounds . Approximately twenty CsNEJs were placed at the center 0 point of each trough using a micropipette . After acclimating for 10 min , CsNEJs were attracted by dropping 4 μL of 50 mM cholic acid ( Sigma-Aldrich Co . , St . Louis , MO ) dissolved in dimethyl sulfoxide ( DMSO , Sigma-Aldrich ) at the +5 point . The same volume of DMSO only was dropped at the +5 point as a negative control . Behavior and migration distance of CsNEJs were observed using a dissecting microscope every 10 min for 60 min . All assays were performed inside a walk-in chamber maintained at 37°C and 80% humidity . To minimize temperature fluctuation , the chemotaxis panel was covered with a lid except when chemicals were applied or CsNEJs were observed . Effects of neuro-antagonists on the chemotaxis of CsNEJs were measured . Dopaminergic inhibitors used were LE-300 , SKF-83566 , sulpiride , remoxipride , nemonapride , spiramide ( AMI-193 ) , NGB-2904 , U-99194 , GR-103691 , and benzothiophenylcyclohexylpiperidine ( BTCP ) . Serotoninergic inhibitors were fluoxetine , spiroxatrine , ritanserin , and Y-25130 . Glutamatergic inhibitors were CNQX ( 6-cyano-7-nitroquinoxaline-2 , 3-dione ) , NBQX ( 6-nitro-2 , 3-dioxo-1 , 2 , 3 , 4-tetrahydrobenzo[f]quinoxaline-7-sulfonamide ) , MK-801 , and cyclothiazide . Cholinergic inhibitors were pirenzepine and benzoquinonium . Neuropeptides were Phe-Met-Arg-Phe amide ( FMRFamide ) , neuropeptide Y ( NPY ) , and peptide YY ( PYY ) . All test compounds ( Sigma-Aldrich ) were dissolved in 99 . 5% ethanol at stock concentration of 10 mM . Stock solution was either made on the day of assay or taken from aliquots stored at −20°C for no longer than 4 weeks prior to use . Each stock solution was serially diluted in 1× Locke’s solution . Mean chemotactic distance ( mm ) was calculated as the migration distance of all CsNEJs and dividing by their number . Percent chemotactic inhibition was calculated as a ratio of migration distance difference between test and positive control groups . All assays were performed in triplicate with different batches of CsNEJs and results presented as a mean ± standard error of the mean . Difference was statistically analyzed using Student’s t-test and considered statistically significant at P < 0 . 05 . Approximately 300 C . sinensis metacercariae were fed to a New Zealand White rabbit . Adult C . sinensis were recovered from the bile duct of the rabbit 4 months post-infection , then flat-fixed in 4% paraformaldehyde in 0 . 1 M PBS for 1 h , washed three times with AbD solution ( 0 . 1 M PBS , 0 . 1% Triton X-100 , 1% bovine serum albumin , and 0 . 1% NaN3 , at pH 7 . 4 ) for 10 min each , and incubated in AbD solution for 24 h at 4°C . The flukes were incubated in primary antibody solution containing goat anti-choline acetyltransferase polyclonal antibody ( 1:10 diluted , Millipore , Billerica , MA ) and mouse anti-tyrosine hydroxylase monoclonal antibody ( 1:250 diluted , Millipore ) for 5 d at 4°C . After washing in AbD for 24 h at 4°C , the flukes were incubated in donkey-anti-goat-Cy3 and donkey-anti-mouse-FITC secondary antibodies ( 1:500 diluted ) for 5 d at 4°C . After a final overnight wash in AbD , the immune-stained flukes were mounted in Gel/Mount ( Biomeda , Foster City , CA ) . The specimens were observed and photographed under a confocal microscope . Neuronal control of CsNEJs’ chemotaxis toward cholic acid was investigated . Various pharmacological agents acting on neuroreceptors such as dopamine , glutamate , serotonin , acetylcholine , and neuropeptide receptors were tested ( Table 1 ) . Of these , dopaminergic antagonists noticeably inhibited chemotactic migration of CsNEJs to 50 mM cholic acid , even at nanomolar concentrations ( Fig 1 ) . Inhibition showed some degree of concentration dependency . Dopamine D1 receptor antagonists LE-300 and SKF-83566 as low as 10 nM inhibited chemotaxis of CsNEJs toward 50 mM cholic acid by 35 . 5 ± 3 . 8% and 43 . 0 ± 4 . 0% , respectively ( Fig 1A ) . This inhibition increased with concentration of antagonists , and LE-300 and SKF-83566 at 100 μM inhibited chemotaxis as much as 92 . 0 ± 1 . 2% and 89 . 0 ± 5 . 8% . Dopamine D2 receptor antagonists spiramide , sulpiride , and nemonapride inhibited chemotaxis even more strongly ( Fig 1B ) , and remoxipride did so to a lesser degree . Spiramide , sulpiride , nemonapride , and remoxipride at 10 nM inhibited the chemotaxis by 82 . 7 ± 13 . 9% , 79 . 0 ± 17 . 6% , 76 . 0 ± 6 . 9% , and 8 . 3 ± 3 . 8% , respectively , and at 100 μM by 125 ± 14 . 9% , 87 . 3 ± 6 . 4% , 88 . 0 ± 2 . 3% , and 24 . 0 ± 0 . 6% . Peculiarly , CsNEJs moved in the opposite direction from cholic acid in the presence of 10–100 μM spiramide . Dopamine D3 receptor antagonists NGB-2904 and GR-103691 were also potent inhibitors of chemotaxis . U-99194 was only a weak inhibitor ( Fig 1C ) . At 10 nM , NGB-2904 and GR-103691 inhibited chemotaxis by 81 . 7 ± 7 . 9% and 79 . 3 ± 4 . 9% , and at 100 μM by 117 . 3 ± 2 . 3% and 104 . 0 ± 7 . 5% , respectively . Both NGB-2904 and GR-103691 caused CsNEJs to avoid cholic acid at high micromolar concentrations . A dopamine reuptake inhibitor BTCP moderately inhibited chemotaxis ( Fig 1D ) . At 10 nM , BTCP inhibited chemotaxis by 50 . 7 ± 5 . 0% , increasing to 83 . 0 ± 13 . 9% at 100 μM . Although dopaminergic inhibitors suppressed chemotaxis to cholic acid , they did not decrease motility or cause shrinkage of the CsNEJs . Serotonergic antagonists such as spiroxatrine ( 5-HT1 ) , ritanserin ( 5-HT2 ) , and Y-25130 ( 5-HT3 ) up to 100 μM did not influence the chemotactic migration of CsNEJs toward cholic acid ( Fig 2B–2D ) . However , fluoxetine , a selective serotonin reuptake inhibitor , reduced chemotaxis at concentrations higher than 1 μM ( Fig 2A , P < 0 . 05 ) . Fluoxetine alone at 100 μM caused a shrinkage of the worms . A glutamate AMPA ( α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid ) /kainate receptor antagonist CNQX inhibited chemotaxis only at a high concentration of 100 μM ( Fig 3A , P < 0 . 05 ) . However , the AMPA receptor antagonist NBQX , the NMDA ( N-methyl-d-aspartate ) receptor antagonist MK-801 , and the positive AMPA/kainate receptor modulator cyclothiazide did not inhibit chemotaxis ( Fig 3B–3D ) . Pirenzepine , a muscarinic receptor antagonist , and benzoquinonium , a nicotinic receptor antagonist , did not inhibit chemotaxis ( Fig 4 ) , and pirenzepine at 0 . 1 μM actually enhanced it ( P < 0 . 05 ) . Neuropeptides such as FMRFamide , peptide YY , and neuropeptide Y strongly inhibited chemotactic migration of CsNEJs at sub-micromolar to low micromolar concentrations ( Fig 5 ) , where FMRFamide at 10 nM inhibited chemotaxis by 85% , and peptide YY at 1 nM by 92% ( Fig 5A and 5B , P < 0 . 05 ) . The inhibition by these neuropeptides showed concentration dependency . On the other hand , neuropeptide Y did not affect chemotaxis up to 100 nM , but at over 1 μM , CsNEJs suddenly avoided CA , in an all-or-none manner ( Fig 5C ) . All three neuropeptides when applied alone had no effect on direction of migration . Confocal immunofluorescence microscopy indicated tissue distribution of dopaminergic and cholinergic neurons in adult C . sinensis . Tyrosine hydroxylase ( TH ) and choline acetyltransferase ( ChAT ) antibodies were used to detect dopaminergic and cholinergic neurons , respectively . Anti-TH green fluorescence and anti-ChAT red fluorescence indicated the colocalized presence of both neuronal types at low density throughout the body of adult C . sinensis , appearing more frequently in regions between oral and ventral suckers in the forebody , and between the testes in the hindbody . The neurons were aligned in the lateral margin of the intestine in the forebody ( Fig 6A–6F ) . The oral and ventral suckers were satellited with these neuronal cell bodies . The central nervous system of C . sinensis is composed of two cerebral ganglia whose anterior nerves contribute to the pharynx and to the oral sucker [14] . Serotonin has been found to stimulate the motility of adult C . sinensis . However , cholinergic agonists have an opposite effect , which is not reversed by traditional cholinergic antagonists , suggesting a disparate pharmacological profile of helminthic acetylcholine receptors from mammalian counterparts , or a non-receptor mediated action [15] . Nevertheless , the anthelmintic drug tribendimidine is a nicotinic acetylcholine receptor agonist effective to treat C . sinensis infection [16] . In the present study , however , the nicotinic antagonist benzoquinonium tended to inhibit chemotaxis of CsNEJs to cholic acid , whereas a muscarinic antagonist pirenzepine showed a tendency to enhance it . Dopamine neurons play a major role in the movement and muscle function of other flatworm and roundworm species [9 , 17–22] . The existence of a dopamine neuron , however , has not been previously established in C . sinensis . Our immunofluorescence findings confirm its existence in adult C . sinensis . The neuronal cell bodies were located in the somatic muscle region , suggesting dopaminergic control of its locomotory behavior . In accordance with this , the chemotactic migration of CsNEJs to cholic acid was suppressed by dopamine receptor antagonists even at low nanomolar concentration . Dopamine D2 and D3 receptor antagonists such as spiramide , sulpiride , nemonapride , NGB-2904 , and GR-103691 were particularly powerful inhibitors . These results suggest that dopamine neurons may control the chemotactic migration of CsNEJs to cholic acid . However , since remoxipride and U-99194 were only weakly effective , these results should be regarded with some caution . In addition , it is not certain whether the antagonistic effects on chemotaxis were due to an inhibition of muscle movement or a disturbance of chemosensation . Curiously , the dopamine reuptake inhibitor BTCP also moderately inhibited chemotaxis . Similarly , anomalous pharmacology has also been observed in the Schistosoma mansoni dopamine D2 receptor , where both the traditional agonists and antagonists inhibited the receptor [9] . In addition , atropine , a traditional muscarinic acetylcholine receptor antagonist , showed inverse agonist activity toward an acetylcholine receptor [23] . This discrepancy may arise from differences in the molecular structure of receptors between invertebrates and vertebrates . Serotonin stimulates motility in various flatworms , including adult C . sinensis [15] . Serotonin receptor is involved in the control of muscle motility at the neuromuscular junction of Fasciola hepatica and S . mansoni [24–26] . Fluoxetine and exogenous serotonin both produce a strongly hyperactive phenotype in S . mansoni schistosomula [27] . Serotonin also markedly stimulates F . hepatica activity , whereas fluoxetine oddly inhibits it [22] . We also observed that fluoxetine at a micromolar concentration inhibited chemotaxis , and that fluoxetine alone at 100 μM caused a shrinkage of CsNEJs , possibly indicating a toxic influence unrelated to its inhibition of serotonin transport . Serotonin antagonists inhibit serotonin-induced increases in motility of S . mansoni sporocyst [28] . However , in the present study , the serotonin antagonists spiroxatrine , ritanserin , and Y-25130 had no noticeable effects on the chemotaxis of CsNEJs . This also may be attributable to species differences in receptor structure in invertebrates . Glutamate-like immunoreactivity is widespread in the nervous system of F . hepatica [29] , where G protein-coupled glutamate receptors are predicted to exist [25] . Kainate binding sites were observed in adult S . mansoni , indicating an existence of ionotropic glutamate receptors [30] . However , MK-801 , a traditional NMDA antagonist in vertebrates , binds to nicotinic receptors , but not NMDA receptors in adult S . mansoni [31] . In the present study , only CNQX at 100 μM inhibited chemotaxis of CsNEJs , whereas NBQX , MK-801 , and cyclothiazide had no effect . These findings indicate that ionotropic glutamate receptors do not modulate the chemotactic movement of CsNEJs . Neuropeptides such as FMRFamide-like peptides and neuropeptide Fs are widely distributed in nervous systems in flatworms , including neurons serving the somatic musculature [11 , 32–34] . G protein-coupled receptors for neuropeptide F/Y binding are predicted to exist in F . hepatica [25] . FMRFamide , peptide YY , and neuropeptide Y , the neuropeptides tested here , all strongly inhibited the chemotaxis of CsNEJs to cholic acid . In S . mansoni , FMRFamide-like peptides elicit potent muscle contraction by enhancing Ca2+ influx through sarcolemmal voltage-gated Ca2+ channels [35] . The potent nature of muscle contraction by neuropeptides may lead to tetanic paralysis of CsNEJs , hampering their chemotactic movement . In summary , our results represent the first immunohistochemical demonstration of dopaminergic neurons in the somatic muscle layer of adult C . sinensis , and that dopamine receptor antagonists and neuropeptides are powerful inhibitors of CsNEJs chemotaxis to cholic acid . Since cholic acid is the most important chemoattractant to CsNEJs in bile fluids , this finding could be utilized for the development of drugs preventing C . sinensis migration to the bile duct [4 , 7] .
The liver fluke Clonorchis sinensis is a flatworm parasite dwelling in the bile duct , which can induce serious pathological inflammatory changes , and chronic infection is associated with bile duct cancer . In order to gain access to its habitat , C . sinensis larva follows chemical cues from the liver , a phenomenon called chemotaxis . Bile , including its component cholic acid , is essential for the juvenile fluke’s migration toward the intrahepatic bile duct from the host intestine . Here , we report that the chemotaxis to cholic acid is controlled through dopaminergic neurons and neuropeptides . This observation can be utilized to develop a preventive intervention against infection by C . sinensis .
[ "Abstract", "Introduction", "Materials", "and", "methods", "Results", "Discussion" ]
[ "cell", "motility", "neurochemistry", "chemical", "compounds", "dopaminergics", "neuroscience", "organic", "compounds", "hormones", "serotonin", "muscarinic", "acetylcholine", "receptors", "amines", "neurotransmitters", "neuropeptides", "catecholamines", "dopamine", "nicotinic", "acetylcholine", "receptors", "animal", "cells", "proteins", "neurochemicals", "chemistry", "transmembrane", "receptors", "chemotaxis", "peptide", "hormones", "acetylcholine", "receptors", "biochemistry", "signal", "transduction", "cellular", "neuroscience", "cell", "biology", "organic", "chemistry", "neurons", "biogenic", "amines", "biology", "and", "life", "sciences", "cellular", "types", "physical", "sciences" ]
2019
Chemotactic migration of newly excysted juvenile Clonorchis sinensis is suppressed by neuro-antagonists
Herbivores can gain indirect access to recalcitrant carbon present in plant cell walls through symbiotic associations with lignocellulolytic microbes . A paradigmatic example is the leaf-cutter ant ( Tribe: Attini ) , which uses fresh leaves to cultivate a fungus for food in specialized gardens . Using a combination of sugar composition analyses , metagenomics , and whole-genome sequencing , we reveal that the fungus garden microbiome of leaf-cutter ants is composed of a diverse community of bacteria with high plant biomass-degrading capacity . Comparison of this microbiome's predicted carbohydrate-degrading enzyme profile with other metagenomes shows closest similarity to the bovine rumen , indicating evolutionary convergence of plant biomass degrading potential between two important herbivorous animals . Genomic and physiological characterization of two dominant bacteria in the fungus garden microbiome provides evidence of their capacity to degrade cellulose . Given the recent interest in cellulosic biofuels , understanding how large-scale and rapid plant biomass degradation occurs in a highly evolved insect herbivore is of particular relevance for bioenergy . Plant cell walls contain the largest reservoirs of organic carbon on Earth [1] . This carbon is largely inaccessible to most organisms , occurring in the form of cellulose , hemicelluloses , and lignin . Certain bacteria and fungi are capable of deconstructing these recalcitrant plant polymers , and thus play a critical role in nutrient cycling in the biosphere . Lignocellulolytic microbes form symbiotic relationships with animals that feed on plant biomass , providing their hosts with access to nutrients in return for a constant supply of plant polymers . Recent microbiome studies have revealed how these communities mediate plant biomass deconstruction in animals , including detritivores [2] , ruminants [3] , and omnivores [4]–[6] . Here , we characterize the microbiome of an important Neotropical herbivore , the leaf-cutter ant Atta colombica . Leaf-cutter ants in the genus Atta are one of the most conspicuous and widespread insects in the New World tropics , forming massive colonies composed of millions of workers . Mature colonies forage hundreds of kilograms in leaves each year ( Figure 1A ) , substantially altering forest ecosystems and contributing to nutrient cycling [7] . Leaf-cutter ants do not feed directly on harvested leaves; rather , they use leaf fragments as substrate to cultivate a mutualistic fungus in specialized subterranean gardens ( Figure 1B and 1C ) . The fungus serves as the primary food source for the colony and in return is provided with substrate , protection from competitors , and dispersal through colony founding [7]–[9] . Despite the impact of these ants on tropical ecosystems , and the critical role leaves play in Atta colonies reaching immense sizes , our current understanding of plant biomass deconstruction within fungus gardens is limited . Our study presents the first functional metagenomic characterization of the microbiome of an insect herbivore . We reveal that the microbial community within the fungus gardens of leaf-cutter ants contains not only the fungal cultivar , but a diverse assembly of bacteria dominated by γ-proteobacteria in the family Enterobacteriaceae . We further show that these bacteria likely participate in the symbiotic degradation of plant biomass in the fungus garden , indicating that the fungal cultivar is not solely responsible for this process , as has been previously assumed . This suggests a model of plant biomass degradation in the fungus garden that includes both bacteria and the fungal cultivar , and we speculate that persistent cellulose-degrading bacterial symbionts like Klebsiella and Pantoea could work in concert with the fungal cultivar to deconstruct plant polymers . As an external digestive system , the fungus garden of leaf-cutter ants parallels the role of the gut in other plant biomass degrading systems like bovines and termites . The presence of a bacterial community dominated by Proteobacteria in leaf-cutter ant fungus gardens is similar to the gut microbiota reported for other insect herbivores , suggesting that bacteria in this phylum may be widespread in their association with herbivorous insects [28]–[30] . However , in contrast to other insect herbivores , the external nature of the leaf-cutter ant digestive system removes the restrictions imposed by the physical limitations of internal guts . This feature is likely responsible for these ants achieving massive colony sizes that harvest vast quantities of plant biomass to support their extensive agricultural operations . As a result , these herbivores have a considerable impact on their surrounding ecosystem by contributing significantly to the cycling of carbon and nutrients in the Neotropics . This study of the leaf-cutter ant fungus garden microbiome illustrates how a natural and highly-evolved microbial community deconstructs plant biomass , and may promote the technological goal of converting cellulosic plant biomass into renewable biofuels . A total of 25 fungus gardens from 5 healthy colonies ( 5 gardens each ) of the leaf-cutter ant Atta colombica were collected at the end of May and beginning of June , 2008 . These colonies are located along Pipeline Road in Soberanía National Park , Panama ( latitude 9° 7′ 0″ N , longitude 79° 42′ 0″ W ) and designated N9 , N11 , N12 , N13 , and N14 , respectively . Each fungus garden was vertically cross-sectioned into thirds with the top third designated as the “top” of the garden and the bottom third designated as the “bottom” of the garden . All material was frozen and transported back to the University of Wisconsin-Madison where it was stored at −20°C prior to processing . From all 5 colonies ( 3 gardens per colony ) , 5 independent samples from fungus garden tops and bottoms of each garden were collected for sugar composition analysis . Thus , a total of 75 fungus garden samples each from the top and bottom were used for this part of our study . This material was tested for crystalline cellulose and hemicellulose ( matrix polysaccharide ) content as follows . Cellulose content of fungus garden plant biomass was determined by first washing each sample using Updegraff reagent [31] , which removes matrix polysaccharides such as hemicelluloses , pectins and amorphous glucan . The remaining residue , containing only crystalline cellulose , was hydrolyzed using Saeman hydrolysis [32] . The resulting glucose monosaccharides were then quantified with an anthrone colourmetric assay as previously described [32] . For the composition of the matrix polysaccharide content , the following components were tested: arabinose , fucose , galactose , glucose , rhamnose , mannose and xylose . Quantification of these sugars were performed by treating finely ground materials with solvents to remove pigments , proteins , lipids , and DNA from the material as previous described [33] . The residue was de-starched with an amylase treatment , resulting in only cell wall material . This material was then treated with 2M trifluoroacetic acid solubilizing the matrix polysaccharides in form of their monosaccharides , and subsequently derivatized into their corresponding alditol-acetates , which were separated and quantified by GC-MS as previously described [34] . The same set of samples used for sugar composition analysis was also used for lignin content analysis , as previously described [35] . Briefly , all samples were dried to 60°C and ground using a 1-mm cyclone mill and analyzed for total non-lignin organic matter , lignin , and ash ( organic and inorganic ) content . Total carbohydrate content was assessed through a two-step acid hydrolysis with neutral sugars quantified using GC and uronic acids quantified using colorimetry . Klason lignin was quantified from the ash-free residue from the two-step acid hydrolysis . Ash content was quantified by combustion at 450°C for 18 h and the average µg/mg of material was calculated . Total DNA was extracted in preparation for either 16S rDNA sequencing or community metagenomic sequencing . For 16S rDNA sequencing , a total of 5 gardens each from 3 Atta colombica colonies ( N9 , N11 , and N12 ) were used . A total of 1 g ( wet weight ) of fungus garden material was sampled from the top layer of each garden corresponding to each colony , for a combined final weight of 5 g of fungus garden material . Total DNA from this sample was then extracted using a MoBio Power Soil DNA Extraction Kit ( MOBIO Laboratories , Carlsbad , CA , USA ) . The same procedures were performed for all fungus garden bottom layer samples for all 3 colonies . For community metagenomic sequencing , total community DNA was extracted from 5 whole fungus gardens each from all 5 Atta colombica colonies used in this study . A total of 1 g of fungus garden material was sampled from top , middle , and bottom layers from all fungus gardens and combined to produce a final sample weight of 75 g . This material was then enriched for bacteria using a modification of a previously-described protocol [36] . Briefly , total fungus garden material was buffered in 1X PBS ( 137 mM NaCl , 2 . 7 mM KCl , 10 mM Na2HPO4 , 2 mM KH2PO4 ) containing 0 . 1% Tween and then centrifuged for 5 minutes at 40×g . This resulted in a 3-layer mixture containing leaf-material at the top , fungal mass in the middle , and bacteria at the bottom . The top and middle layers were carefully removed , buffered with 1X PBS containing 0 . 1% Tween , and washed using the same centrifugation method an additional 3 times . The final mixture was then centrifuged for 30 minutes at 2800×g , re-suspended in 1X PBS containing 0 . 1% Tween and filtered through a 100 um filter . Total DNA from this resulting sample was then extracted using a Qiagen DNeasy Plant Maxi Kit ( Qiagen Sciences , Germantown , MD , USA ) . Extracted DNA from fungus gardens was PCR amplified ( 20 cycles ) using full-length universal bacterial ( 27F [5′-AGA GTT TGA TCC TGG CTC AG-3′] and 1391R [5′- GAC GGG CRG TGW GTR CA-3′] ) and archaeal ( 4aF [5′- TCC GGT TGA TCC TGC CRG-3′] and 1391R [5′- GAC GGG CRG TGW GTR CA-3′] ) primers and cloned into the pCR4-TOPO vector ( Invitrogen ) ( See http://my . jgi . doe . gov/general/protocols/SOP_16S18S_rRNA_PCR_Library_Creation . pdf ) . This was then sequenced using standard Sanger-based capillary sequencing and assembled as previously described [2] ( http://www . jgi . doe . gov/sequencing/protocols/prots_production . html ) . These same samples were then pyrosequenced by first PCR amplifying all samples with prokaryote-specific primers corresponding to the V6-V8 region ( 1492R [5′- TAC GCY TAC CTT GTT ACG ACT T - 3′] and 926F [5′- AAA CTY AAA KGA ATT GAC GG - 3′] fused to 5-base barcodes ( reverse primer only ) and 454-Titanium adapter sequences ) and then sequenced on a Roche 454 FLX GS Titanium pyrosequencer [14] . All 16S rDNA sequences generated in this study are deposited in GenBank with accessions HM545912–HM556124 and HM556125–HM559218 for near full-length 16S rDNA sequences and pyrotagged 16S rDNA sequences , respectively . Assembled full-length 16S contigs were first compared against the National Center for Biotechnological Information's ( NCBI ) non-redundant nucleotide ( nt ) and environmental nucleotide ( env_nt ) databases ( accessed: 05/01/2009 ) using BLAST [37] to verify that all sequences were bacterial . A small number of eukaryotic 18S sequences belonging to the fungus the ants cultivate , Leucoagaricus gongylophorus , which were likely amplified due to the cross-reactivity of the 16S primers , were removed . No sequences identified as archaeal were detected from our library generated using archaeal-specific primers , and only bacterial sequences were amplified . Sequences were prepared for alignment by orienting each sequence in the same direction using the computer program Orientation Checker [38] , putative chimeras were removed using Bellerophon [39] , and each set was de-replicated to remove exact duplicates . Finalized sets for each sample were then analyzed using the ARB [40] software environment as follows . All full-length 16S rDNA sequences were imported and then aligned using the ARB fast-aligner tool [40] against a user-constructed PT-Server ( constructed from the SILVA [41] 16S SSU rDNA preconfigured ARB reference database with 7 , 682 columns and 134 , 095 bacterial sequences; accessed: 01/15/2009 ) . The full alignment was manually curated using the ARBprimary editor ( ARB_EDIT4 ) in preparation for phylogenetic and community analysis . Once an acceptable alignment was obtained we created a PHYLIP [42] distance matrix in ARB using the filter-by-base-frequency method ( column filter; minimal similarity = 50%; gaps ignored if occurred in >50% of the samples; 1 , 320 valid columns ) . The PHYLIP distance matrix was exported to the MOTHUR software package v . 1 . 5 . 0 [43] for community analysis and OTU designation . Briefly , the distance matrix was read into MOTHUR and clustered using the furthest neighbor algorithm . From here , we performed rarefaction , rank-abundance , species abundance , and shared analyses . Representative sequences from each OTU at 97% were re-imported into ARB for phylogenetic analysis ( Figure S4 , S5 , S6 , S7 , and S8 ) . We used a Maximum Likelihood ( RAxML [44] ) method for all phylogenetic analyses ( normal hill-climbing search algorithm ) and the above-mentioned method for positional filtering . Closest taxonomic assignment of clones was performed using the Ribosomal Database Project ( RDP ) [45] by comparing sequences against the type strain database ( Table S5 ) . For pyrotagged short-read 16S rDNA sequences , all sequences were compared against the National Center for Biotechnological Information's ( NCBI ) non-redundant nucleotide ( nt ) and environmental nucleotide ( env_nt ) databases ( accessed: 05/01/2009 ) using BLASTN . Sequences were then classified as either bacterial , archaeal , or eukaryotic , and only those bacterial sequences ( 20 , 330 ) were retained for further analysis . These sequences were then processed through Orientation Checker , chimeras removed using the program Mallard [38] , and subsequently analyzed using MOTHUR in the following fashion . First the entire dataset was de-replicated to eliminate duplicate sequences . The remaining sequences were aligned in MOTHUR against the Greengenes [46] reference alignment ( core_set_aligned . imputed . fasta; 7 , 682 columns , accessed: 09/11/2009 ) using the Needleman alignment method with the following parameters: k-tuple size = 9; match = +1; mismatch penalty = −3; gap extension penalty = −1; gap opening penalty = −5 . Sequences were then screened to eliminate those shorter than 400 bp ( gaps included ) . Filtration eliminated 7 , 062 columns resulting in a total alignment size of 620 bp ( gaps included ) . The remaining dataset was again de-replicated to eliminate duplicate sequences and we constructed a furthest-neighbor distance matrix in MOTHUR using the twice de-replicated , filtered , alignment . All subsequent analyses ( rarefaction , rank-abundance , species abundance , and shared analyses ) were performed in MOTHUR using this distance matrix . A UniFrac [15] analysis was performed on all full-length 16S rDNA samples generated in this study , including 3 from the top and 3 from the bottom of fungus gardens . MOTHUR was used to generate phylip distance matrices and the computer program Clearcut [47] was then employed to construct neighbor-joining trees . UniFrac was then used to compare these samples as shown in Figure S3 . Whole community DNA was used to create a shotgun library which was then sequenced using a single pyrosequencing plate on a Roche 454 FLX GS Titanium sequencer . Assembly of the data was performed using the 454 de novo assembler software with default parameters . Total amounts of data generated and statistical coverage is presented in Table S2 . Raw sequence reads generated for this microbiome are deposited in NCBI's Short Read Archive under Study Accession SRP001011 . 1 , and assembled contigs and singletons have been deposited into DDBJ/EMBL/GenBank under the accession ADWX00000000 . The complete set of assembled contigs and singletons representing the fungus garden community metagenome was phylogenetically binned using the following approach . First , the metagenome was compared against NCBI's non-redundant nucleotide ( nt ) and environmental nucleotide ( env_nt ) databases ( accessed: 05/01/2009 ) using BLASTN ( e-value: 1e-05 ) and the top hit was retained . The designated phylogenetic classification of the top hit for each contig and singleton was then assessed and binned into one of the following 4 sets: Bacterial , Eukaryotic , Viral , or Unknown . We then performed in-depth phylogenetic binning of the bacterial portion of the fungus garden community metagenome using the current microbial genome collection ( http://www . ncbi . nlm . nih . gov/genomes/lproks . cgi , accessed: 05/15/2009 ) . We reasoned that using the current microbial genome collection is a likely a more accurate metric for classifying the bacterial set at the genus level because each genome in this collection is correctly annotated and the current iteration of this collection contains both phylogenetic breadth and depth for many represented genera . As a result , we performed two different phylogenetic bins using the current microbial genome collection . First , GeneMark [48] was used to predict open reading frames and their corresponding translated proteins of the bacterial portion of the fungus garden community metagenome using a generic bacterial gene model . This predicted proteome was then compared against a local database containing all proteomes in the current microbial genome collection ( http://www . ncbi . nlm . nih . gov/genomes/lproks . cgi , accessed: 05/15/2009 ) supplemented with the predicted proteomes of two bacterial strains ( Klebsiella variicola At-22 and Pantoea sp . At-9b , see below ) isolated from the fungus gardens of a related leaf-cutting ant species , Atta cephalotes . Comparison of the fungus garden proteome against our microbial reference database was done using BLASTP ( e-value: 1e-05 ) and the phylogenetic identity of the top hit was recorded . The total number of proteins was then tabulated to the genus level . Total nucleotide coding content for each predicted protein was then calculated to determine the total amount of nucleotide represented in each bin . Second , we performed phylogenetic binning on the bacterial portion of the fungus garden metagenome using the entire nucleotide content of the current microbial genome collection ( http://www . ncbi . nlm . nih . gov/genomes/lproks . cgi , accessed: 05/15/2009 ) , and again supplemented with the nucleotide content from the draft genome sequences of our two bacterial isolates from Atta cephalotes leaf-cutter ant fungus gardens . Using complete nucleotide content of the current microbial genome collection is advantageous because it includes both coding and intergenic regions , and provides a more robust measure of phylogenetic identity . We compared the entire bacterial portion of the fungus garden metagenome against this database using BLASTN ( e-value: 1e-05 ) and the phylogenetic identity of the top hit was recorded . The total number of contigs and singletons was then tabulated to the genus level and the corresponding nucleotide amounts were also calculated . Furthermore , we performed this same analysis using all high-quality reads from our fungus garden community metagenome . Finally , we employed the phylogenetic binning program PhymmBL [22] , which resulted in similar phylogenetic binning results as our comparison against the sequenced genome collection . We performed GC content analysis on the Bacterial , Eukaryotic , and Unclassified phylogenetic bins of the leaf-cutter ant fungus garden community metagenome . For the bacterial set , we divided the sequences according to the NCBI Taxonomic Groups Acidobacteria , Actinobacteria , α-proteobacteria , Bacteroidetes , β-proteobacteria , and γ-proteobacteria . We then calculated their GC content , and tabulated the total number of sequences within each group corresponding to each percentage as shown in Figure S11 . For Eukaryotic sequences , these were divided into fungal , metazoan , and plant classifications and GC content analysis was also performed as shown in Figure S12 . Furthermore , this same analysis was performed for the unclassified portion of the community metagenome and plotted alongside our Eukaryotic GC content analysis . The predicted proteome from the bacterial portion of the fungus garden community metagenome was annotated using the carbohydrate active enzyme ( CAZy ) database [23] as follows . A local database of all proteins corresponding to each CAZy family from the CAZy online database ( http://www . cazy . org/ , accessed: 06/01/2009 ) was constructed , and this was used to align the predicted proteome of the bacterial portion of the fungus garden community metagenome using BLASTP ( e-value of 1e-05 ) . This proteome was then annotated against the protein family ( Pfam [49] ) database ( ftp://ftp . ncbi . nih . gov/pub/mmdb/cdd/ , accessed: 05/01/2009 ) using RPSBLAST [50] ( e-value: 1e-05 ) . A CAZy to Pfam correlation list was then compiled based on the secondary annotations provided through the CAZy online database . Finally , only those proteins that had significant BLAST hits to a protein from our local CAZy database and its corresponding Pfam were retained and designated as a carbohydrate-associated enzyme . A similar process was performed using the eukaryotic portion of the fungus garden metagenome . However , because of the difficulty in accurately predicting proteins from this subset , due to the lack of good gene models , we compared the contigs and singletons in this subset to our local CAZy and Pfam databases using BLASTX ( e-value: 1e-05 ) . Only those hits with significant matches to a protein from our local CAZy database , and its corresponding Pfam were retained and designated as a carbohydrate-associated enzyme in this set . To determine the similarity of the fungus garden community metagenome with respect to other sequenced metagenomes , we performed a comparative analysis using protein domain and carbohydrate enzyme content as a comparative metric , as previously described [51] . In general , the predicted proteome from the bacterial portion of the fungus garden metagenome was annotated according to clusters of orthologous groups ( COGs [52] ) database ( ftp://ftp . ncbi . nih . gov/pub/mmdb/cdd/ , accessed: 05/01/2009 ) using RPSBLAST ( e-value: 1e-05 ) . The predicted proteomes from the following 13 metagenomes were also annotated in the same manner: bovine rumen [3] , chicken cecum [53] , fish gut and slime [54] , gutless worm [55] , human gut ( Gill ) [6] , human gut ( Kurokawa ) [56] , Minnesota soil [51] , lean mouse [5] , obese mouse [5] , termite hindgut [2] , wastewater sludge USA [57] , sastewater sludge OZ [57] , and whale fall [51] . The COG profiles from all of these metagenomes were divided according to their COG gene category designations and plotted as a proportion as shown in Figure S13 . Cluster analysis of COG profiles for these metagenomes were performed as follows . A matrix was generated with each row corresponding to a metagenome and each column corresponding to a COG ID . The proportion of each COG with respect to the total number of annotated COGs in that metagenome was calculated and populated in the appropriate cell of the matrix . Spearman's rank correlation was then applied to this matrix to generate a similarity matrix correlating each metagenome to each other based on the similarity of each metagenome's COG profile . A distance matrix was then calculated using the neighbor program from the computer suite Phylip [42] ( using the UPGMA method ) , and the resulting unrooted tree was visualized using the phylodendron tree drawing program ( http://iubio . bio . indiana . edu/treeapp/ , accessed 07/25/2009 ) . This same analysis was also performed using protein domains ( Pfam ) and no discernable difference in metagenome groupings was detected ( data not shown ) . A similar approach was used for clustering these metagenomes according to CAZy content . Each metagenome's predicted proteome was annotated using CAZy and correlated to its Pfam annotation as described above . Because each protein potentially encodes for domain that belong to multiple CAZy families ( i . e . a protein may contain both a GH and a CBM ) , we assigned multiple CAZy annotations to a particular protein . A carbohydrate enzyme matrix was then constructed with each row corresponding to a metagenome sample and each column corresponding to a CAZy family . Each cell in this matrix was then populated with the proportion of each CAZy family with respect to the total number of annotated CAZy families in each respective metagenome . Generation of an unrooted tree using this matrix was then constructed using the same procedure outlined for clustering based on the protein domain content metric . Pure isolates of Klebsiella variicola At-22 and Pantoea sp . At-9b were cultured from the fungus gardens of the leaf-cutter ant Atta cephalotes , as previously described [10] . Genomic DNA from these isolates were extracted , as previously described [10] . Draft genomes of Klebsiella variicola At-22 and Pantoea sp . At-9b were sequenced at the U . S . Department of Energy Joint Genome Institute ( JGI ) using a random shotgun approach through a combination of 454 standard and paired-end pyrosequencing ( 454 Life Sciences , a Roche Company ) and 36 bp read Illumina sequencing ( Illumina , Inc . ) . Sequencing using 454 was performed to an average depth of coverage of 30X for both Klebsiella and Pantoea . All general aspects of library construction and sequencing performed at the JGI can be found at http://www . jgi . doe . gov . A draft assembly for Klebsiella variicola At-22 was compiled based on 459 , 192 reads; for Pantoea sp . At9b , a draft assembly was constructed using 557 , 748 reads . The Phred/Phrap/Consed software package ( http://www . phrap . com ) was used for sequence assembly and quality assessment of both drafts [58]–[60] . After the shotgun stage , reads were assembled with parallel Phrap ( High Performance Software LLC ) . Automated annotation of these draft genomes were performed by the Computational Biology and Bioinformatics Group of the Biosciences Division of the U . S . Department of Energy Oak Ridge National Laboratory as described at http://genome . ornl . gov/ . The draft genome sequence and annotation for Klebsiella variicola At-22 and Pantoea sp . At-9b were deposited in GenBank under accession numbers CP001891 and ACYJ00000000 , respectively . The full set of reads used for the assembly of the fungus garden community metagenome was used to generate a recruitment plot against the draft genomes of Klebsiella variicola At-22 and Pantoea sp . At-9b , two isolates we cultured from the fungus garden of the leaf-cutter ant Atta cephalotes [10] , as previously described [27] . Briefly , the contigs from each draft genome were concatenated together in ascending size to produce a “pseudogenome” , and the reads from the fungus garden community metagenome were aligned against a database containing both pseudogenomes , and all genomes from the current microbial genome collection ( http://www . ncbi . nlm . nih . gov/genomes/lproks . cgi , accessed: 05/15/2009 ) using BLASTN . The top hit for each read was retained , and categorized to each genome . We then mapped reads corresponding to Klebsiella variicola At-22 and Pantoea sp . At-9b onto each organism's respective psuedogenome and further binned them according to their sequence identities as follows: 95%–100% , 90%–95% , 85%–90% , 80%–85% , and 70%–80% . Visualization of the mapped reads onto each respective draft genome was performed using the DNAPlotter software package [61] . A CAZy analysis was performed on the predicted proteomes of Klebsiella variicola At-22 and Pantoea sp . At-9b using the same approach as described for CAZy analysis of the leaf-cutter ant fungus garden community metagenome . Furthermore , both GH8 cellulases from each of these genomes were compared against the CAZyme of the fungus garden community metagenome at the nucleotide level using BLASTN ( e-value: 1e-05 ) . Bioassays were performed on pure cultures of Klebsiella variicola At-22 and Pantoea sp . At-9b to determine their capacity to degrade cellulose . These include carboxymethyl cellulose ( CMC ) assays and growth on microcrystalline cellulose . CMC assays were performed as previously described [62] . Briefly , pure cultures of both bacteria were inoculated onto yeast malt extract agar ( YMEA , 4 g yeast extract , 10 g Bacto Peptone , 4 g Dextrose , 15 g agar ) and grown at 25°C for 2 days . Single colonies were then spotted onto carboxymethyl cellulose plates ( 15 g agar , 5 g carboxymethyl cellulose [Calbiochem , La Jolla , CA] ) . Detection of cellulose degradation on CMC was performed using congo red , and the ability of each isolate's capacity for cellulose degradation was measured based on the zone of clearing present on the plate . Growth on microcrystalline cellulose was performed by inoculating 10 µl of pure culture into 150 µl of microcrystalline cellulose broth ( 1 L water and 5 g cellulose powder microcrystalline cellulose [MP Biomedicals , Solon , OH] ) and growth was measured using a DTX 880 Multimode Detector Plate Reader ( Beckman Coulter Inc . , Fullerton , CA ) at an absorbance of 595 for 2 days . Positive growth on microcrystalline cellulose was correlated to an increase in the measured absorbance over this period .
Leaf-cutter ants form massive subterranean colonies containing millions of workers that harvest hundreds of kilograms of leaves each year . They use these leaves to grow a mutualistic fungus that serves as the colony's primary food source . By farming fungus in specialized garden chambers , these dominant Neotropical herbivores facilitate rapid large-scale plant biomass conversion . Our understanding of this degradation process , and the responsible microbial community , is limited . In this study , we track the degradation of plant polymers in leaf-cutter ant fungus gardens and characterize the microbial community potentially mediating this process . We show that cellulose and hemicelluloses are degraded in the fungus gardens and that a previously unknown microbial community containing a diversity of bacteria is present . Metagenomic analysis of this community's genetic content revealed many genes predicted to encode enzymes capable of degrading plant cell walls . The ability of leaf-cutter ants to maintain an external microbial community with high plant biomass-degrading capacity likely represents a key step in the establishment of these ants as widespread , dominant insect herbivores in the Neotropics . This system is an important model for understanding how microbial communities degrade plant biomass in natural systems and has direct relevancy for bioenergy , given recent interest in cellulosic biofuels .
[ "Abstract", "Introduction", "Results/Discussion", "Materials", "and", "Methods" ]
[ "genetics", "and", "genomics/genomics", "genetics", "and", "genomics/microbial", "evolution", "and", "genomics", "evolutionary", "biology/microbial", "evolution", "and", "genomics", "genetics", "and", "genomics/comparative", "genomics", "evolutionary", "biology/evolutionary", "ecology", "computational", "biology/comparative", "sequence", "analysis", "microbiology/microbial", "evolution", "and", "genomics", "evolutionary", "biology/genomics", "biotechnology/environmental", "microbiology", "ecology/environmental", "microbiology", "computational", "biology/genomics", "evolutionary", "biology/bioinformatics", "computational", "biology/metagenomics", "genetics", "and", "genomics/bioinformatics" ]
2010
An Insect Herbivore Microbiome with High Plant Biomass-Degrading Capacity
The cohesin complex , which is essential for sister chromatid cohesion and chromosome segregation , also inhibits resolution of sister chromatid intertwinings ( SCIs ) by the topoisomerase Top2 . The cohesin-related Smc5/6 complex ( Smc5/6 ) instead accumulates on chromosomes after Top2 inactivation , known to lead to a buildup of unresolved SCIs . This suggests that cohesin can influence the chromosomal association of Smc5/6 via its role in SCI protection . Using high-resolution ChIP-sequencing , we show that the localization of budding yeast Smc5/6 to duplicated chromosomes indeed depends on sister chromatid cohesion in wild-type and top2-4 cells . Smc5/6 is found to be enriched at cohesin binding sites in the centromere-proximal regions in both cell types , but also along chromosome arms when replication has occurred under Top2-inhibiting conditions . Reactivation of Top2 after replication causes Smc5/6 to dissociate from chromosome arms , supporting the assumption that Smc5/6 associates with a Top2 substrate . It is also demonstrated that the amount of Smc5/6 on chromosomes positively correlates with the level of missegregation in top2-4 , and that Smc5/6 promotes segregation of short chromosomes in the mutant . Altogether , this shows that the chromosomal localization of Smc5/6 predicts the presence of the chromatid segregation-inhibiting entities which accumulate in top2-4 mutated cells . These are most likely SCIs , and our results thus indicate that , at least when Top2 is inhibited , Smc5/6 facilitates their resolution . In order to maintain chromosome stability , cells need to overcome topological problems caused by the structure of the DNA molecule . One example of such topological problem is DNA supercoiling induced by replication or transcription . Another is sister chromatid intertwinings ( SCIs ) , which is the wrapping of chromatids around each other ( Figure 1A and B ) . If not resolved by topoisomerases , supercoiling inhibits transcription and replication , and SCIs block chromosome segregation . While both type I and type II topoisomerases can resolve supercoils by making transient DNA breaks , the type II variant , called Top2 in the budding yeast Saccharomyces cerevisiae ( S . cerevisiae ) , is main responsible for the resolution of SCIs ( Figure 1C ) . [1]–[3] . If Top2 is rendered non-functional before anaphase , chromosome segregation with unresolved SCIs leads to DNA breakage and cell death [4]–[6] . In addition to presenting an obstacle for segregation , sister chromatid tethering by SCIs has been proposed to contribute to proper segregation by counteracting the force of the mitotic spindle , thereby facilitating chromosome alignment during metaphase [7] . The idea of such a positive function for SCIs was , however , challenged when the cohesin protein Scc1 ( also known as Mcd1 ) was shown to be essential for sister chromatid cohesion [8] , [9] . The cohesin complex , with a core consisting of Smc1 , Smc3 , Scc1 and Scc3 , is a so-called Structural Maintenance of Chromosomes ( SMC ) protein complex . In addition to the four core subunits , the Pds5 protein associates to the complex via Scc1 [10] , [11] . When either of the subunits is non-functional , sister chromatids are not held together , and chromosome alignment and segregation fail . Cohesin is loaded onto chromosomes before replication , and localizes to intergenic regions between genes that are transcribed in a convergent manner in S . cerevisiae [9] , [12] . Several observations indicate that transcription drives the translocation of cohesin to these regions after initial loading by the Scc2/4 complex at centromeres and other , so far mostly undefined , chromosome arm sites [12]–[14] . In addition to loading , cohesin has to be modified in order to establish cohesion . A key regulator of this process is the acetyltransferase Eco1 ( also called Ctf7 ) [15] , [16] , which acetylates Smc3 and thereby prevents the Pds5-associated protein Wpl1 ( also called Rad61 ) , to destabilize the cohesin-chromosome interaction [17]–[19] . In the absence of Wpl1 , Eco1 becomes largely dispensable for cohesion establishment [18] , [20] , [21] . At anaphase , chromosome segregation is made possible by removal of cohesin from chromosomes by separase , a protease which cleaves the Scc1 subunit [22] . In higher eukaryotes , the proteolytic cleavage is preceded by a specific dissociation of cohesin along chromosome arms in prophase , leaving cohesin only in the centromeric region until anaphase [23] , [24] . The identification of cohesin as the main constituent of chromatid cohesion provided an explanation of how sister chromatid cohesion could be maintained without the risk of chromosome breakage , which is inevitably linked to cohesion created by SCIs . A more recent study by Farcas et al . shows , however , that cohesin protects SCIs from resolution by Top2 on circular mini-chromosomes [25] , suggesting SCIs could contribute to cohesion . Intriguingly , the cohesin-related Smc5/6 complex ( Smc5/6 ) has also been connected to Top2 function [26] . Smc5/6 consists of Smc5 , Smc6 and six non-SMC proteins ( Nse1 , Mms21 , and Nse3-6 ) , and is best known for its function in DNA repair and recombination ( reviewed in [27] ) . The complex is recruited to DNA breaks in a process dependent on Mre11 , and central repair factor which accumulates early at the site of damage [28] . When Smc5/6 is non-functional , unresolved recombination intermediates accumulate between sister chromatids in the repetitive ribosomal DNA in unchallenged cells , and during S-phase repair of induced DNA damage [29]–[32] . Since DNA repair in the absence of proper Smc5/6 function is taken to a step that inactivates the DNA damage checkpoint , the unresolved DNA links will inhibit the subsequent segregation event . Also in meiosis , repair of DNA breaks without Smc5/6 leads to similar formation of unresolved recombination intermediates with following segregation failure [33]–[35] . In addition to this , Smc5/6 appears to have non-repair functions . In S . cerevisiae , Smc5/6 has been proposed to function in replication termination [36] , and removal of replication-induced supercoiling [26] . Moreover , Smc6 has been reported to allow full removal of cohesin at anaphase when Top2 function is partially compromised in the fission yeast Schizosaccharomyces pombe ( S . pombe ) [37] . Concerning the chromosomal association of Smc5/6 in the absence of DNA damage , it is independent of Mre11 , but requires the replication process as such , and increases after inactivation of the temperature-sensitive top2-4 allele in S . cerevisiae [26] , [28] . This opens for the possibility that the chromosomal association is triggered by the presence of SCIs , or another feature which accumulates in top2-4 . Regardless if the complex is recruited to SCIs or recombination structures in the absence of DNA damage , its chromosomal association should require that sister chromatids are in close proximity to one another . This predicts that the levels of Smc5/6 present on the replicated genome should decrease in the absence of cohesion , which leads to a separation of chromatids before anaphase . However , our earlier ChIP-on-chip analysis ( Chromatin immunoprecipitation ( ChIP ) , combined with analysis on microarrays ) of FLAG-tagged Smc6 indicated that the chromosome binding of Smc5/6 changed into more numerous , but narrower , binding sites in scc1-73 cells [28] . The finding that the chromosomal association of Smc5/6 was not reduced in the absence of cohesin argued against it being triggered by a structure which requires the proximity of sister chromatids . In contrast , the scc2-4 mutation , which inhibits cohesin loading , was shown to reduce the levels of chromosome-bound Smc6 . This , together with the aberrant binding pattern of Smc6 in scc1-73 cells , made it difficult to draw a definite conclusion on how cohesin influences the chromosomal association of Smc5/6 . Using ChIP-sequencing ( ChIP-seq , ChIP combined with DNA sequencing ) , together with ChIP-qPCR ( ChIP combined with quantitative PCR ) and in situ immunofluorescence , we now show that Smc5/6 chromosome binding is cohesin-dependent . The majority of the chromosome-bound Smc5/6 also co-localizes with cohesin in the vicinity of centromeres , and specifically accumulates along chromosome arms after Top2 inactivation . Evidence is provided that this accumulation is independent of recombination , DNA breaks and fork stalling . Our results also show that the amount of chromosome-bound Smc5/6 predicts the level of missegregation in top2-4 cells , and that the complex promotes the segregation of short chromosomes in the mutant . Altogether , the presented data suggests that Smc5/6 indicates the presence of SCIs in the duplicated genome , and that the complex promotes their resolution , at least when Top2 is inhibited . Triggered by the observations that Smc5/6 accumulates on chromosomes in top2-4 mutants [26] , and that cohesin is a protector of SCIs [25] , we revisited the chromosomal association of S . cerevisiae Smc5/6 using ChIP-seq . This method is more quantitative than ChIP-on-chip , and provides more clearly defined binding sites ( Figure 2A ) . The difference is likely caused by the requirement of additional amplification of the immunoprecipitated DNA in ChIP-on-chip , which increases the risk of false positive signals due to the preferential augmentation of certain DNA molecules . Moreover , 50 base pairs ( bp ) reads are mapped to a reference genome in ChIP-seq , while the amplified material is hybridized to 25 bases long oligonucleotides , each representing a specific genomic sequence , in ChIP-on-chip . The short length of the oligonucleotides , and the requirement for hybridization as such ( the efficiency of which varies from oligonucleotide to oligonucleotide ) , makes the ChIP-on-chip method less accurate . In contrast to previous results , the ChIP-seq analysis showed that the levels of Smc6 found on chromosomes were markedly reduced in the scc1-73 mutant after an S-phase at restrictive temperature ( Figure 2B ) . Western blot analysis confirmed that this reduction was not due to a general down-regulation of Smc6-FLAG protein levels in the mutant ( Figure S1A ) . At core centromeres the signal remained high , but at all other specific Smc6 binding sites it was abolished ( Figure 2B ) . The reduction of Smc6 was confirmed by ChIP-qPCR in the scc1-73 mutant ( Figures 2C and S2 ) . At arm loci , the amount of Smc6 was reduced to levels similar to those in G1-arrested wild-type cells , reflecting the background level before the complex has associated with chromosomes . The Smc6 signal around centromeres was also significantly reduced but remained at up to one third of the wild-type level ( Figure 2C ) . Thus , the ChIP-qPCR results show that the ChIP-seq data is quantitatively accurate , apart from at core centromeres where the signal is overestimated in ChIP-seq , when few other binding sites are present . Even though this indicates that Smc5/6 is largely absent from chromosomes in the cohesin mutant , it is possible that the reduction only reflects the spreading of the complex to an even distribution over the chromosomes . Such redistribution would make it undetectable by ChIP-seq and lead to a reduction in the ChIP-qPCR signal . To test if this was the case , immunofluorescence ( IF ) was utilized to detect the association of HA-tagged Smc6 on chromosome spreads . In scc1-73 cells , the fluorescence signal was reduced towards the levels detected in untagged cells ( Figure 2D ) . As for Smc6-Flag , the signal reduction was not due to lower levels of the Smc6-HA protein ( Figure S1B ) , showing that the chromosomal association , and not only positioning , of Smc5/6 requires a functional cohesin complex . The reduction of Smc6 binding in scc1-73 mutants indicates that Smc5/6 requires sister chromatid cohesion to associate with chromosomes . To test this further , Smc6 localization was analyzed in other cohesion-disrupting mutants . First , ChIP-seq and ChIP-qPCR analysis confirmed the earlier result that Smc5/6 requires the cohesin-loading protein Scc2 for chromosomal association ( Figure 3A and C ) [28] . The reduction of Smc6 binding in scc1-73 and scc2-4 mutants as measured by ChIP-qPCR was similar ( Figures 2C and 3C ) , and the reason for the difference previously seen by ChIP-on-chip remains unknown [28] . We also found that binding of Smc6 was prevented in the temperature sensitive pds5-101 mutant ( Figure 3A and C ) . ChIP-seq was also performed on Smc6-FLAG in eco1-1 cells , in which formation of sister chromatid cohesion is inhibited even though cohesin remains bound to the chromatids [15] , [16] . The reduction of Smc6 binding ( Figure 3A and C ) in this mutant therefore shows that the chromosomal association of Smc5/6 requires cohesion , and not merely the presence of cohesin on chromosomes . This was further supported by the observation that Smc6 chromosome binding in eco1-1 cells was increased by deletion of Wpl1 ( Rad61 ) ( Figure 3A and C ) , which restores cohesion [18] , [20] , [21] . On the other hand , the localization of cohesin remained unchanged in an smc6-56 mutant after an S-phase at restrictive temperature , showing that although cohesin controls Smc5/6 , the reverse is not true ( Figure 3B ) . Altogether , this shows that sister chromatids have to be held together for Smc5/6 to bind to the duplicated genome . To take full advantage of the higher resolution obtained by ChIP-seq as compared to ChIP-on-chip , we reinvestigated the chromosomal association of Smc5/6 during the cell cycle . This confirmed that the complex is mostly absent from chromosomes in G1 . Similarly to the Smc6 binding pattern in G2/M-arrested scc1-73 cells , ChIP-seq also revealed an association to the core centromeres in this cell cycle phase ( Figure 4A ) , but ChIP-qPCR analysis showed that the levels are low compared to the binding in G2/M-arrested cells ( Figure 2C ) . As shown before , Smc5/6 is detected at stalled forks in cells arrested in early S-phase by the addition of hydroxyurea ( HU ) [38] , which is a binding pattern that differs from the distribution found after completion of replication ( Figure 4B–D ) . This could indicate that the Smc5/6 is associated with the fork and follows fork progression , and to test this , Smc6 binding was analyzed in cells progressing through S-phase at 18°C . This condition is generally applied to slow down replication and to improve cell cycle synchronization . As in the HU-arrested cells , Smc6 displayed a different binding pattern as compared to in G2/M , but the signals were less well defined , likely due to a lower level of synchronization ( Figure 4B–D ) . Even though this left the question whether Smc5/6 follows the replication fork unanswered , it shows that the binding pattern detected in G2/M is not present in early S-phase . In addition to this , the following new features were revealed . First , Smc5/6 is absent from chromosomes in telophase cells , arrested through inactivation of the mitotic exit network kinase Cdc15 ( Figure 4E ) , well in line with the dependency on cohesin , which is removed from chromosomes at anaphase onset . Second , robust Smc5/6 binding sites are concentrated around the centromeres in G2/M-arrested cells , and all of these sites are found between convergently transcribed genes and co-localizes with cohesin ( Figures 4D and S3 ) . A third new feature of Smc6 chromosomal association was detected when comparing the level of association with the length of each chromosome . Earlier analysis showed that Smc5/6 enrichment per chromosome increased with its length [26] , [28] . Due to the new observation that strong Smc5/6 chromosome interaction sites clusters around centromeres , this analysis was repeated focusing on this region . Smc6 enrichment was calculated in a 100 kb region spanning the centromere ( Figure S4 ) , and when compared to chromosome length , a positive correlation was confirmed ( Figure 4F ) . In our earlier analysis , we suggested that this binding pattern reflects that SCIs can swivel off chromosome ends [26] . If so , enrichment on each side of the centromere should also correlate to the length of the corresponding chromosome arm . This is because kinetochores are re-attached to microtubules directly after their replication , which should confine SCIs to each individual arm [39] . However , the correlation between Smc6 enrichment in a 50 kb region on either side of the centromere and the length of corresponding chromosome arm is low , arguing against such an interpretation ( Figure 4G ) . On the other hand , the levels of Smc6 in the entire 100 kb region showed a stronger correlation with the distance to the closest telomere , i . e . the length of the shortest chromosome arm ( Figure 4H ) . This shows that the further away from a chromosome end a centromere is positioned , the more Smc5/6 will accumulate in its vicinity . Having determined that the chromosomal localization of Smc5/6 depends on cohesin and cohesion , the chromosomal binding pattern in top2-4 cells was determined using ChIP-seq and ChIP-qPCR . These analyses showed that Smc6 binding around centromeres in top2-4 was not significantly changed as compared to wild-type cells . However , along chromosome arms , Smc6 was strongly enriched at specific sites ( Figures 5A and S3 ) . Such an accumulation of Smc6 was also detected after depletion of Top2 by induced protein degradation , showing that it is was not specific effect of the top2-4 allele ( Figure S5 ) . In contrast to top2-4 cells , however , the Smc6 signal was increased at core centromere 9 after Top2 depletion , opening for a functional difference at these sites . The reason for this difference is unknown and here we focus on the increase along chromosome arms , which is common to both conditions . Similar to the binding sites in the pericentromeric region , the new binding sites were mainly found in intergenic regions between convergently oriented genes and co-localized with cohesin in top2-4 cells ( Figure 5A , B and E ) . The binding pattern of Scc1 , on the other hand , remained unaltered in top2-4 cells , showing that the change in Smc6 association does not reflect alterations in cohesin's chromosomal localization ( Figure 5B ) . Moreover , the levels of chromosome-bound Smc6 , as determined by ChIP-seq and IF , were reduced not only in scc1-73 cells , but also in top2-4 scc1-73 cells after an S-phase under restrictive conditions ( Figures 5C and 2D ) . This reduction was confirmed by ChIP-qPCR ( Figure 5D ) , and as in scc1-73 cells , it was not due to lower Smc6 protein levels in top2-4 scc1-73 cells ( Figure S1A and B ) . This suggests that the chromosomal binding of Smc5/6 in wild-type and top2-4 is due to the same underlying cohesin-dependent mechanism . However , even though the IF analysis showed that the level of chromosome-bound Smc6 was lower in top2-4 scc1-73 than in top2-4 cells , the signal was significantly stronger than in the scc1-73 single mutant ( Figure 2D ) . This , together with the ChIP results , indicates that some Smc5/6 remains on chromosomes in top2-4 scc1-73 , but distributes differently from cells with functional cohesin . Knowing that Top2 is needed for removal of transcription-induced supercoils [2] , [40] , the accumulation of Smc5/6 in top2-4 could be controlled by transcription alone . To investigate this , ChIP-seq and ChIP-on-chip analysis was performed after 1 hour of Top2 inactivation in G2/M-arrested cells , or after 30 minutes inactivation in a G1-arrest . The results revealed that without passage through S-phase at restrictive conditions , there was no alteration in Smc6 chromosomal interaction pattern as compared to the wild-type binding pattern ( Figure 6A and B ) . This shows that like in wild-type cells , the chromosomal positioning of Smc5/6 is set under replication in the top2-4 mutant . It is well established that Smc5/6 is recruited to DNA double-strand breaks ( DSBs ) and facilitates resolution of recombination intermediates [28] , [32] , [41] . To test whether Smc5/6 chromosome association in top2-4 cells was dependent on these structures , ChIP-seq and ChIP-qPCR analysis of Smc6 was performed on cells lacking RAD52 or MRE11 . Deletions of these genes inhibit recombination and Smc5/6 recruitment to DSBs , respectively [28] , [42] . The results showed that Smc6 still accumulates on chromosomes when Top2 is inhibited in these mutants , demonstrating that the complex binds chromosomes independently of DNA breaks and recombination in top2-4 cells ( Figure 7A and B ) . This was further supported by western blot analysis of the Rad53 kinase , which is part of the damage cell cycle checkpoint and becomes phosphorylated upon DNA damage [43] , [44] . This phosphorylation can be detected as a slower migrating form of Rad53 , and this was readily observed after replication inhibition through the addition HU to both wild-type and top2-4 cells ( Figure 7C ) . However , after passage through S-phase in the restrictive temperature without addition of HU , no phosphorylation was detected . This indicates that no DNA damage accumulates upon inhibition of Top2 during S-phase . Smc5/6 is also known to associate to stalled replication forks [38] , and Top2 has been shown to facilitate termination of replication [45] . It is therefore possible that Smc5/6 marks stalled forks that are still present in G2/M-arrested top2-4 cells . To test this , the chromosomal localization of the DNA polymerase epsilon subunit Dpb3 was analyzed . This showed that even though Dpb3 was detected on S-phase chromosomes , it was not found in G2/M-arrested top2-4 cells ( Figure 8A and B ) . Moreover , Smc6 did not accumulate on chromosomes in a helicase rrm3Δ mutant , known to elicit replication fork stalling ( Figure 8C ) [46] . Finally , to assay directly if replication or recombination intermediates accumulate at Smc5/6 binding sites , two-dimensional gel electrophoresis was performed at two loci displaying abundant Smc5/6 binding in top2-4 cells ( Figure 9A ) . At both loci , replication intermediates were detected in S-phase in wild-type and top2-4 cells , but not in G2/M-arrested cells , when Smc5/6 binding is most abundant ( Figure 9B ) . This shows that the accumulation of Smc5/6 at these loci in top2-4 mutant is not due to the presence of a DNA structure that can be detected by a standard two-dimensional gel electrophoresis assay . In addition , the UPB10-MRPL19 locus was investigated using two-dimensional gel electrophoresis on DNA prepared using a CTAB-extraction method [47] . This method preserves specific X-shaped structures , which have been suggested to be hemicatenated sister chromatids , and in early S-phase , these could be detected at a positive control locus , ARS305 , in wild-type and top2-4 cells ( Figure 9C ) . In the end of S-phase , no difference between wild-type and top2-4 cells could be seen at the Smc5/6 binding sites . This shows that it is not an increase in hemicatenane-like structures that causes Smc5/6 to accumulate after Top2 inhibition . Altogether , this indicates that the Smc5/6 binding pattern detected in top2-4 cells is independent of DNA breaks , recombination and replication fork stalling . So far , the data presented here show that Smc5/6 complex is recruited to a chromosome structure which requires sister chromatids that are held together by cohesin . It also accumulates at cohesin sites along chromosome arms after replication under Top2-inhibting conditions . The structure is not a recombination intermediate , a DNA break , nor a replication fork . Neither does it appear after inactivation of Top2 in G1- or G2/M-arrested cells . Altogether this points to that the chromosomal association of Smc5/6 indicates the presence of a recombination-independent structure , which forms during replication on cohesed sister chromatids , and normally is removed by Top2 . To test if Smc5/6 accumulation on chromosomes in top2-4 was sensitive to Top2 activity after the completion of DNA replication , the chromosomal association of Smc6 was investigated after restoration of Top2 function in G2/M-phase . Mutant top2-4 cells were first taken through an S-phase at restrictive temperature , and when arrested in G2/M , the temperature was decreased to permissive during 1 hour before sample preparation . Cell survival experiments suggest that SCIs are removed under these conditions [48] , and we confirmed this by showing that the temperature down-shift rescues the segregation of chromosome 5 , and removes the accumulation of SCIs on a reporter plasmid ( Figure 10A–C ) . Under these conditions , ChIP-seq and ChIP-qPCR showed that Smc6 dissociates from chromosomes in top2-4 cells to levels similar to those found in wild-type ( Figure 10B , D and E ) . If the temperature instead is maintained during the prolonged G2/M-arrest , Smc6 levels remained high . This indicates that the chromosomal association of Smc5/6 correlates with a segregation-inhibiting structure that can be removed by Top2 after the completion of replication , but persist during a prolonged G2/M-arrest when Top2 is non-functional . To investigate the correlation of the chromosomal association of Smc5/6 and missegregation in top2-4 further , we analyzed chromosome segregation after inactivation of Top2 in G2/M . Under these conditions Smc5/6 chromosomal association remains at wild-type levels ( Figure 6 ) , in contrast to the accumulation of Smc5/6 on chromosomes when Top2 is inactivated from G1 until G2/M ( Figure 5 ) . This allows segregation analysis under Top2-inhibiting conditions of chromosomes with either wild-type or increased levels of Smc5/6 . In preparation for this analysis , we first investigated chromosome segregation in wild-type and top2-4 cells released from a G1-arrest into restrictive conditions for the mutant . Using a system based on the association of fluorescently labeled tetracycline repressors with multiple repeats of tetracycline operators [9] , the centromere- and telomere-proximal regions of a short ( chromosome 1 ) , an intermediate ( chromosome 5 ) and a long chromosome ( chromosome 4 ) were observed ( see material and methods for details ) ( Figure 11 ) . All three chromosomes were marked 35 kb away from the centromere and within 100 kb from one of the telomeres . Note that on the short chromosome 1 , the centromere and telomere marker is one and the same . To get as detailed a picture of the segregation event as possible , sister chromatid separation was scored in relation to elongation of the mitotic spindle , and segregation was scored in relation to the separation ( Figure 11A ) . Chromatids were logged as separated as soon as two fully separated fluorescent dots were visible , and noted as segregated when these dots were partitioned into mother cell and bud . Both separation and segregation occurred simultaneously at the centromere of all three chromosomes in wild-type cells ( Figure 11B–D ) . Separation and segregation of the telomere-proximal regions of chromosomes 4 and 5 took place later , with the longer being most delayed ( Figure 11B–D ) . This result is expected since segregation starts at the centromeres due to their attachment to the mitotic spindle . In top2-4 , chromatid separation and segregation proceeded slower in the pericentromeric regions of all three chromosomes , and the delay was most pronounced on the longest chromosome 4 ( Figure 11B–D ) . Both separation and segregation of the telomere-proximal region of chromosomes 4 and 5 , but not 1 , were severely impaired . These length-dependent delays are in accordance with the observation that long linear chromosomes break more frequently than short ones in top2-4 cells , which was suggested to reflect the ability of SCIs to swivel off the ends of shorter chromosomes [6] . Having established this , segregation was scored after inactivation of Top2 in G2/M . Again , since chromosome-bound Smc5/6 is maintained at wild-type levels under these conditions , the segregation defects should be less severe than after Top2 inactivation in G1 , if the complex indicates the presence of the chromosome segregation-inhibiting structures in top2-4 cells . Moreover , a shorter chromosome is expected to segregate more efficiently than a longer one . We therefore analyzed segregation of telomeric markers on chromosome 4 ( long ) and 5 ( intermediate ) after a shift to restrictive conditions for the top2-4 allele in G2/M-arrested cells . This showed that in sharp contrast to the severe segregation defect of both chromosomes when Top2 is inactivated in G1 , the partitioning of the intermediate-size chromosome 5 now occurred at close to wild-type levels , while the telomeric marker on the long chromosome 4 still exhibited severely defective segregation ( Figure 12A and B ) . When the experiment was repeated , analyzing a region on chromosome 4 which was located at the same distance from the centromere as the telomeric marker on chromosome 5 ( approximately 350 kb from the centromere ) , an intermediate improvement of segregation was detected , as compared to when Top2 was inactivated in G1 ( Figure 12C ) . In yet another test of how well Smc5/6 chromosome association correlates with missegregation after Top2 inactivation , segregation was scored in an scc1-73 top2-4 double mutant after a G1-release into restrictive conditions . This analysis was also prompted by the observation that entanglements remains between circular mini-chromosomes in the double mutant [25] . Well in line with the ChIP and IF analyses , which indicate that some , but not all , Smc5/6 on chromosomes is retained but de-localized in scc1-73 top2-4 ( Figures 5A , C , D and 2D ) , these cells displayed an intermediate missegregation phenotype ( Figure 11E and F ) . While centromere-proximal regions of chromosomes 1 and 5 displayed premature separation similar to the scc1-73 single mutant , the telomere proximal site of chromosome 5 was inhibited as in top2-4 cells . If Smc5/6 accumulates in response to the accumulation of segregation-inhibiting structures in top2-4 mutants , it is expected to execute a function at these sites . To test this , we analyzed segregation of chromosome 1 . This chromosome segregates at near to wild-type levels in top2-4 cells despite the occurrence of new Smc6 binding sites along the arm ( Figure 11C ) . In an smc6-56 top2-4 double mutant however , there was a threefold increase in missegregation ( Figure 13A ) . This indicates that the additional Smc5/6 complexes recruited upon Top2 inhibition facilitate resolution of this chromosome . To test if this function is to promote removal of cohesin from mitotic chromosomes , levels of FLAG-tagged Scc1 was analyzed by western blot and ChIP-qPCR . This showed that both total levels and chromosome-associated Scc1 was equally reduced in telophase-arrested smc6-56 top2-4 and wild-type cells ( Figure 13B–E ) . This investigation was launched to understand why Smc5/6 accumulates on chromosomes under Top2-inhibiting conditions . Based on the current knowledge of both the complex and the topoisomerase this could either be due to the accumulation of SCIs or an increased level of sister chromatid recombination structures . Since Top2 impairment also delays replication termination , there is also a possibility that the accumulation of Smc5/6 is due to remaining forks in the mutant [45] . If recombination and/or SCI are the triggers , a central feature for Smc5/6 chromosome association should be a dependency on the proximity of sister chromatids . Using high-resolution ChIP-seq , ChIP-qPCR and IF in combination with a variety of mutations which disrupt sister chromatid cohesion , we show that this is the case in both wild-type and top2-4 cells ( Figures 2 and 5 ) . While it was already established that the cohesin loader Scc2 is needed for Smc5/6 chromosomal association [28] , the role of cohesin was more uncertain , making it possible that Scc2 directly loaded Smc5/6 on to chromosomes . However , the here presented results indicate that the absence of chromosome-bound Smc5/6 in scc2-4 cells is due to the lack of cohesion , and not to a direct role of Scc2 in Smc5/6 loading ( Figure 3 ) . On a more general level , the results also argue that phenotypes of mutations which disrupt cohesin function are caused by the combined loss of chromosome-bound cohesin and Smc5/6 . Mutations that change the localization of cohesin might also influence where Smc5/6 is found on chromosomes . Possibly , Smc5/6 contributes to some of the many functions assigned to cohesin ( reviewed in [49] ) . Importantly , however , while cohesin impairment leads to cohesion loss , inhibition of Smc5/6 only creates minor cohesion defects [50] , [51] , with replication delays and/or perturbations of chromosome structure and segregation being more common phenotypes . This suggests that the complex regulates a process and/or structure which is specific for tethered sister chromatid pairs . In addition to reveal that the chromosomal association of Smc5/6 in top2-4 cells is dependent on cohesion ( Figure 5 ) , this study shows that there are no signs of unfinished duplication in the mutant at the sites where Smc5/6 accumulates ( Figures 8 and 9 ) . This argues against the possibility that Smc5/6 binding is triggered by the presence of remaining replication forks . This is further supported by the persistence of chromosome-bound Smc6 during a prolonged G2/M-arrest ( Figure 10D and E ) , since termination of replication has been shown to be delayed but not prevented in mutants of Top2 [45] . Also , Smc6 does not accumulate on chromosomes in rrm3Δ cells ( Figure 8C ) , in which fork pausing is frequent . It is also unlikely that the trigger for Smc5/6 binding in top2-4 is a DNA break or a recombination structure since the damage checkpoint protein Rad53 remains un-phosphorylated , and the Top2-dependent increase in Smc6 binding is still present in top2-4 mre11Δ and top2-4 rad52 cells ( Figure 7 ) . Moreover , there are no signs of recombination intermediates detected by two-dimensional gel electrophoresis in the DNA regions bound by Smc5/6 in top2-4 mutants ( Figure 9 ) . This leaves SCIs as the most likely candidates as triggers for Smc5/6 binding , and the following results argue in favor for this assumption . First , as stated above , the buildup of Smc5/6 in top2-4 cells requires the proximity of chromatids . Second , the accumulation requires that the mutant pass through S-phase under restrictive conditions . After inactivation of Top2 in G1- or G2/M-arrested cells , the levels of Smc5/6 binding are unchanged , i . e . under conditions when Top2 inhibition is expected to perturb transcription only ( Figure 6 ) . Third , Smc5/6 dissociates from chromosomes when Top2 function is restored after replication , under conditions when Top2 resolves SCIs ( Figure 10 ) . Fourth , the level of Smc5/6 chromosome enrichment correlates to the degree of missegregation in top2-4 cells ( Figures 4 and 11 ) . Moreover , inactivation of Top2 in G2/M , which leaves the amount of Smc5/6 binding at wild-type levels , also leads to a lower degree of missegregation than after an S-phase without Top2 function ( Figure 12 ) . In addition to this , the observation that Smc5/6 is needed for segregation of short chromosomes in top2-4 cells ( Figure 13 ) , reveals yet another functional connection between Smc5/6 and SCIs . Results from our earlier analysis suggest that Smc5/6 facilitates formation of SCIs during replication , at least in top2-4 cells . This function was attributed a role of the complex in facilitating fork rotation , thereby decreasing the level of replication-induced supercoiling [26] . The here presented data suggests that Smc5/6 also is needed for Top2-independent resolution of SCIs when replication has been completed ( see below ) . Whether the replicative and post-replicative functions are functionally connected remains to be determined . In addition to providing evidence for Smc5/6 being controlled by the presence of SCIs on chromatids , the level of its chromosomal association indicates that it senses replication-induced superhelical tension . It is difficult to envisage another mechanism that would lead to a correlation between levels of chromosome-bound Smc5/6 and the length of the shortest chromosome arm ( Figure 4H ) . In a previous investigation , we proposed that the link between Smc5/6 binding and chromosome length reflected the ability of SCIs to swivel off chromosome ends [26] . But the relatively poor correlation between Smc6 enrichment and the length of each chromosome arm detected in this investigation ( Figure 4G ) argues against this , since SCI movements are expected to be confined between the microtubule-attached kinetochore and each telomere . We propose instead that the chromosomal association of Smc5/6 reflects the dissolution of replication-induced superhelical stress through rotation of the shortest arm . Such unidirectional dissolution should be possible since kinetochores become unattached from the mitotic spindle during their replication in early S-phase [39] , [52] , [53] . With increasing length of the shortest arm , the more difficult it will be to rotate , which will lead to higher levels of superhelical stress around the centromere . In addition to this , the chromosomal localization of Smc5/6 has to be promoted by a centromere specific-factor since superhelical tension is expected to reach high levels at centrally located , non-centromeric , regions of chromosomes as well . The specific maintenance of Smc5/6 close to the centromeres after Top2 reactivation in G2/M ( Figure 10D ) argues that this factor works by preventing Smc5/6 dissociation . Taken together , the presented results are consistent with a scenario where chromosome-bound Smc5/6 indicates the presence of SCIs in the duplicated genome . Based on the observations that cohesin protects SCIs from Top2-resolution , and that Smc5/6 facilitates their resolution , it is conceivable that SCIs are positioned at Smc5/6-containing cohesin sites . Even though this cannot be formally proven until SCIs are directly observed at these sites , we use the following sections to speculate based on this model and discuss what the distribution of the complex , taken into the context of chromosome segregation in wild-type and top2-4 cells , suggests about SCI dynamics in budding yeast ( summarized in Figure 14 ) . Smc5/6 distribution indicates that SCIs are preferentially found in the vicinity of centromeres in wild-type cells , and accumulate along chromosome arms when Top2 is inactivated during replication ( Figures 14 and S3 ) . During chromosome segregation in wild-type cells , the pericentromeric SCIs are removed by Top2 , which gain access to its substrates after proteolytic cleavage of cohesin . When Top2 is inactivated from G1 and onwards , SCIs accumulate also along chromosome arms and persist after cohesin cleavage in anaphase . The specific inhibition of segregation of intermediate and long chromosome arms under these conditions suggests that the pulling forces of the mitotic spindle drive SCIs from the centromere towards the ends of the chromosome . This will allow separation of all pericentromeric regions , and passive , Top2-independent , separation of short chromosome arms . If Top2 instead is rendered non-functional in G2/M , only centromere-proximal SCIs remain after cohesin removal , and this lower level of SCIs allows segregation of intermediate-sized chromosomes , and partial separation of central regions of a longer ones ( Figure 14 ) . Importantly , based on our observation that missegregation of the short chromosome 1 is increased in the smc6-56 top2-4 mutant as compared to both singles ( Figure 13A ) , Top2-independent SCI resolution appears to be facilitated by Smc5/6 function . Whether the complex achieves this by actively promoting SCI resolution via a separate mechanism and/or by preventing SCIs to be transformed into a structure which cannot be passively resolved over chromosome ends , remains to be established . However , in contrast to S . pombe , Smc5/6 does not appear to facilitate chromosome segregation in the absence of fully functional Top2 by promoting cohesin removal from mitotic chromosomes in S . cerevisiae ( Figure 13D and E ) . This difference might reflect that Top2 inhibition specifically perturbs cohesin removal which occurs independently of Scc1 cleavage [37] . Such a pathway has been reported to exist in fission , but not budding , yeast [54] . Regardless , taking the role of Smc5/6 in the resolution of late recombination intermediates into account , it is possible that recombination structures and SCIs have something in common which allows Smc5/6 to promote their resolution . In addition to the above , the premature chromatid separation of centromere-proximal regions in top2-4 scc1-73 ( Figures 11E and F ) , and the reduction in Smc5/6 chromosome association ( Figure 5 ) , suggest that cohesin does more to SCI dynamics than protecting them from Top2 resolution . If this was not the case , the segregation phenotypes of the double mutant should be identical to that of top2-4 cells , i . e . there should be a delay in segregation at all sites tested . A possible scenario is that cohesin is also needed to prevent SCI mobility along chromosome arms , leading to an even dispersal of SCIs in the top2-4 scc1-73 mutant . Moreover , in the lack of cohesin-imposed constraint , the pulling on the chromosomes by the mitotic spindle would be able to displace SCIs from the centromere-proximal region more readily than in a wild-type background . As a result , regions in the vicinity of centromeres would separate prematurely , while chromosome arm regions on longer chromosomes would remain entangled . On the shorter chromosomes , SCIs would also be passively resolved over chromosome ends more easily . In summary , this leads to a scenario where SCIs are resolved by Top2 decatenation and passive resolution in the scc1-73 mutant , and only by passive resolution in top2-4 scc1-73 cells . This is supported by the IF analysis which shows that there is more Smc6 left on chromosomes in top2-4 scc1-73 than in scc1-73 cells ( Figure 2D ) . Finally , in the light of the possibility that cohesin acts as a direct protector of SCIs we see two explanations for their preferential accumulation around centromeres in wild-type cells . One possibility is that SCI protection not only depends on cohesin , but also on a centromere-specific factor , as discussed above . The observation that reactivation of Top2 in G2/M allows removal of Smc5/6 from cohesin sites along chromosome arms , but not at centromeres ( Figure 10D ) , argues in favor for this . Another , not mutually exclusive , scenario is that SCIs only form when the topological tension reaches a certain threshold . In wild-type cells this would only occur in the vicinity of centromeres , while in top2-4 cells , in which replication-induced topological tension accumulates due to its function in supercoil relaxation , it would also happen at certain cohesin sites along chromosome arms . If so , chromatid entanglement after Top2 inhibition might not only be caused by lack of SCI resolution as the common view predicts , but also to an increase in SCI formation . In conclusion , this investigation reveals that cohesin and cohesion are required for the chromosomal association and localization of Smc5/6 . It also provides evidence that the chromosomal localization of Smc5/6 indicates the presence of SCIs , and that the complex is needed for their Top2-independent resolution . The localization of Smc5/6 to pericentromeric regions in G2/M-arrested cells thus opens for the possibility that SCI are maintained until anaphase , and therefore could contribute to chromatid cohesion , also on linear chromosomes . Taken together with the observation that the chromosomal localization of Smc5/6 is correlated to the length of the shortest chromosome arm , this leads to the unexpected prediction that replication-induced superhelical stress can influence chromosome segregation via the formation of SCIs . All strains are of W303 origin ( ade2-1 trp1-1 can1-100 leu2-3 , 112 his3-11 , 15 ura3-1 ) RAD5 with the modifications listed in Table S1 . Primer sequences used for site directed gene-modifications are available upon request . Strains used for live cell imaging: To integrate multiple copies of tetracycline operators at other sites than 35 kb away from centromere 5 , which is the location of the endogenous ura3-1 gene , ura3-1 was first replaced with the NAT gene , which confers resistance to nourseothricin . The ura3-1 gene was also cloned into the PFA6a-KanMX4 plasmid , which contains the kanamycin resistance gene ( KAN ) . Both ura3-1 and KAN were amplified by PCR using the primers listed in Table S2 . The resulting constructs were used in transformations , and correct integration at the chosen genomic sites was controlled by Southern blot . Finally , the TetO plasmid ( pWJ1378 ) containing multiple copies of tetracycline operons and URA3 , was integrated at the ura3-1 sites . Correct integration was again controlled by Southern blotting . If not stated otherwise , cultures were grown in YEP medium ( 1% yeast extract , 2% peptone , 40 µg/ml adenine ) supplemented with 2% glucose as carbon source , with the exception of the live cell imaging analysis , see below . For synchronization in G1 and a following release at restrictive temperature , 3 µg/ml α factor mating pheromone ( Innovagen ) was added every hour for 1 . 5 generation times . When a complete G1-arrest was achieved , cells were incubated at the restrictive temperature for 30 minutes , unless otherwise stated . For release into a synchronous S-phase , cells were filter-washed by three volumes of pre-heated YEP medium and subsequently resuspended in fresh medium . To achieve a subsequent arrest in the following G2/M , the release medium contained 15 µg/ml nocodazole ( Sigma ) . Chromatin immunoprecipitation was carried out as previously described [26] , [55] with the modification that cells were lysed using a 6870 Freezer/Mill ( SPEX , CertiPrep ) . Briefly , cells were crosslinked by 1% formaldehyde and then washed three times in ice-cold 1× TBS , before being lysed in the Freezer/Mill . Cell lysate was thawed on ice and suspended in lysis buffer . Chromatin was then sheared to a size 300–500 bp by sonication and IP reactions , with anti-FLAG antibody ( F1804 , Sigma ) conjugated to Dynabeads Protein A ( Invitrogen ) , were allowed to proceed over night . After washing and eluting the ChIP fraction from beads , crosslinks were reversed for input and ChIP fractions and DNA was purified . The DNA samples were then processed for sequencing ( see below ) , qPCR or hybridization to microarrays . qPCR was performed using SYBR green ( Applied Biosystems ) and primers listed in Table S2 on Applied Biosystem 7000 Real-Time PCR System according to the manufacturer's instructions . For ChIP-on-chip , hybridization of ChIP and input fractions to GeneChip S . cerevisiae Tiling 1 . 0R Array ( Affymetrix ) was performed as described [26] , [55] . BrdU-IP was performed as previously described [55] using monoclonal anti-BrdU antibody ( clone Bu 20a , Dako ) and Dynabeads Sheep Anti-Mouse IgG ( Invitrogen ) . DNA from ChIP and WCE fractions was further sheared to an average size of approximately 150 bp by Covaris ( Woburn , MA ) . Samples were then prepared for sequencing according to the manufacture's standard protocol ( Applied Biosystems SOLiD Library Preparation protocol ) and were sequenced on Applied Biosystems SOLiD platforms ( SOLiD3 , 4 and 5500 ) to generate single-end 50 bp reads . Sequenced reads of DNA-seq were aligned to the S . cerevisiae genome obtained from Saccharomyces Genome Database ( http://www . yeastgenome . org/ ) using Bowtie [56] , allowing three mismatches in the first 28 bases per read and filtering reads having more than 10 reportable alignments ( -n3 -m10 option ) . Each aligned read was extended to a predicted fragment length of 150 bp . Reads were summed in 10 bp bins along the chromosomes for ChIP and WCE , and further normalized and smoothed as previously described [57] , Nakato R . , et al , 2013 ) . For the number of total and mapped reads in each sample , see Table S3 . Sequence data are available at the Sequence Read Archive ( http://www . ncbi . nlm . nih . gov/sra ) with the accession number SRP018757 . To call peaks for Smc6 and Scc1 , we calculated the fold enrichment ( ChIP/WCE ) for each bin and identified bins which fulfilled following criteria: ( 1 ) fold enrichment was more than 2 . 0; ( 2 ) the maximum read intensity in ChIP bins was more than 1; and ( 3 ) fold enrichment of no tag sample was less than 1 . 8 . Chromosome arms ( Figure 5E ) were defined as the whole chromosomes excluding: 25 kb pericentromeric region spanning the centromere; subtelomeric regions ( 20 kb proximal to each telomere ) ; and long terminal repeats ( LTR ) . LTRs , defined by Saccharomyces Genome Database ( http://www . yeastgenome . org/ ) , were excluded from the upstream to the downstream open reading frame neighboring each LTR . The significance of Smc6 peaks clustering around pericentromeric regions ( Figure S3 ) was assessed with the binomial test by assuming that the Smc6 peaks distributed to the whole genome uniformly . The enrichment values of Smc6-FLAG for each chromosome ( Figure 4F–H ) were calculated by summing up the difference of fold enrichment between Smc6-FLAG and a no tag control experiment in 100 kb regions spanning the centromeres of each chromosome ( Figure S4 ) . Detailed information on the sequencing results is found in Table S3 . Mitotic spreads were prepared as described [58] with the exception that 5% Lipsol ( Dynalab ) was used as a detergent . Wild-type and mutated Smc6-3×HA-expressing cells were arrested in G2/M after a synchronous S-phase at 35° before preparation of spreads . Monoclonal rat-anti-HA ( Roche ) was used as the primary antibody followed by Cy3-conjugated goat-anti-rat ( Invitrogen ) to detect Smc6-3×HA on spreads . Each image was acquired under identical exposure conditions using a Leica microscope and 100× objective . Image analysis was carried out in Volocity ( Perkin Elmer ) . Signals from >50 chromosome spreads were quantified using the analysis tools provided by the Volocity software ( Perkin Elmer ) , and background staining in adjacent regions of the same size were subtracted . Box plots were made using standard statistical tools and represent all values measured between the maximum and the minimum . Statistical analysis to measure significance of differences between strains was done using a two-tailed T-test , with Welch's correction , which was used because the two populations compared had unequal variance . P-values greater than or equal to 0 . 05 were considered insignificant . If not stated otherwise , cells were grown at 23°C in synthetic medium lacking histidine and uracil supplemented with 2% glucose . For G1-release experiments , cells were first arrested by of alpha factor at a final concentration of 3 µg/ml , and moved to 35°C thirty minutes prior to release . 500 µl of cell suspension was then applied to Concanavalin A ( Sigma ) coated glass coverslips ( ∅ 12 mm ) , and were allowed to settle for 2 minutes . Medium was subsequently removed and 1 ml fresh medium without alpha factor was added . Cells were allowed to settle to the glass surface for another 40 minutes and were finally imaged through the following mitosis at 35°C . For G2-release experiments in Figure 10A , cells were first arrested in G1 as above , and after 30 minutes at 35°C , released into pre-warmed medium containing nocodazole at a final concentration of 15 µg/ml . Cells were then grown for one hour at 35°C and then either moved to 23°C or kept at 35°C for an additional hour prior to release from the G2/M-arrest . For experiments in Figure 6A , top2-4 cells were arrested in G1 , released and allowed to grow at 23°C for 90 minutes in medium containing nocodazole to reach a complete G2/M-arrest . The arrest was then maintained at 35°C for one hour prior to release . 500 µl of cell suspension was then put on Concanavalin A ( Sigma ) coated glass coverslips ( ∅ 12 mm ) and were allowed to settle for 2 minutes . Medium was then removed and 1 ml fresh 23°C or 35°C medium was added as appropriate . Cells were allowed to settle on the glass surface for another 5 minutes and then imaged through the following mitosis at either 23°C or 35°C . For both type of experiments , images consisted of a 7-layer Z-stack , with layers 0 . 8 µm apart . These were collected every 30 seconds in green ( GFP ) and red ( tdTomato ) channels , for a total of 70 minutes . Control experiments using wild-type and recombination-deficient rad52Δ cells showed that this setup left cell cycle progression unperturbed , and is therefore unlikely to introduce any significant DNA damage . The microscope used was Deltavision Spectris ( Applied Precision ) , and acquired images were analyzed using ImageJ ( version 1 . 44i ) . Automated tracking of spindle length was performed using CellProfiler version r10997 [59] . Briefly , images were segmented for nuclei based on tetR tdTomato fluorescence and each nucleus was tracked over time . Within each nucleus , the EGFP-tubulin structure was segmented and tracked over time . Spindle elongation was considered when the EGFP-tubulin structure exceeded 10 pixels in length , which is equal to 3 . 18 µm . Cells containing the plasmid pRS316-URA3 were collected and immediately fixed in ice-cold 70% ethanol . These cells were subsequently pelleted and incubated at 37°C for 30 minutes in 400 µl buffer containing 0 . 5 mg/ml zymolyase ( Seikagaku Biobusiness ) , 0 . 9 M sorbitol , 0 . 1 M EDTA ( pH 8 . 0 ) and 14 mM β-mercaptoethanol ( Sigma ) . After a second centrifugation , spheroblasts were resuspended in 400 µl of TE buffer and incubated at 65°C for 30 minutes with 90 µl of 270 mM EDTA ( pH 8 . 0 ) , 460 mM Tris-base and 2 . 3% SDS . Thereafter , 80 µl of 5 M potassium acetate was added , and samples were kept on ice during 60 minutes , subsequently centrifuged for 15 minutes at 13 000 rpm , and finally , the supernatant was collected into new tube . DNA was then precipitated using 1 ml of 100% ethanol , and resuspended in 500 µl of TE buffer . After treatment with 0 . 1 mg/ml RNaseI at 37°C for 30 minutes , the DNA was precipitated with 2-propanol , washed by 70% ethanol and resuspended in 50 µl of TE buffer . For nicking enzyme treatment , DNA was incubated with Nb . BtsI ( New England Biolabs ) for 2 hours at 37°C according to manufacturer's protocol . DNA samples were separated by electrophoresis in 0 . 8% agarose ( Lonza ) 0 . 5× TBE gel with 2 . 7 V/cm for 24 hours . Plasmids were detected by Southern blotting under standard conditions using radioactive probe that was generated by PCR using primer FW ( GTTCCAGTTTGGAACAAGAGTC ) , primer BW ( CATTAAGCGCGGCGGG ) and pRS316 as template . Genomic DNA isolation to study replication intermediates was performed according to [60] . Isolation of genomic DNA with CTAB extraction to preserve X-shape structures was performed according to [47] . Digestion was performed using PstI-HF ( New England Biolabs ) for the loci UBP10-MRPL19 and MPP10-YJR003C , and EcoRI and HindIII ( Roche ) for ARS305 locus . The DNA was then precipitated by the addition of 2 volumes ethanol containing 0 . 5 M potassium acetate and incubated at −80°C for 30 minutes . The precipitated DNA was spun down for 15 minutes at 13 000 rpm and washed with 70% ethanol , before being resuspended in loading buffer . The first dimension gel running was run in 0 . 35% agarose ( Melford , Molecular Biology Grade , MB1200 ) in 1× TBE at 1 V/cm , in room temperature for 24 hours . The second dimension gel running was run in 0 . 875% agarose ( same as above ) in 1× TBE with 0 . 3 µg/ml ethidium bromide at 5 V/cm , at 4°C for 8 hours , with buffer circulation from anode to cathode at 50 ml/min . Specific loci were detected by Southern blotting under standard conditions using radioactive probe that was generated by PCR using primer pairs GTTCGCCAGTCTCCGTTATT and CTGGGATACCCGAATGTGTATG for ARS305; ATGGTGAAGACATCGGCGAAGACA and AGTGGTAGAAGTGGTGGCTGAAGT for UBP10-MRPL19; GCTTCAGCGTATTGTAGCATTT and GCTCGTGGAACCTATCCTTATT for MPP10-YJR003C , with genomic DNA as template . To detect Rad53 , wild-type and top2-4 cells were G1-arrested at permissive temperature ( 23°C ) , incubated at restrictive temperature ( 35°C ) for 30 min , before being released into 0 , 2M HU or 15 µg/ml nocodazole at 35°C for 75 min . Cells were then collected and protein extracted using trichloroacetic acid ( TCA ) -precipitation . To detect Scc1-FLAG in telophase and G2/M-arrests , cells were G1-arrested as above before being released into media with or without 15 µg/ml nocodazole at 35°C for 2 hours . To detect Smc6-FLAG and -HA in various strains , cells were prepared as in Figure 2 . Cells were then collected and protein extracted a glass-bead disruption method [61]with the modifications that 1× PhosSTOP ( Roche ) was added to the lysis buffer and that after cell lysis , 2 µl of Benzonase nuclease ( Novagen 70664 ) and NaCl to 200 mM final concentration was added and incubated 30 min at 4°C to promote the release of chromatin-bound proteins . Bradford assay was then used to estimate protein concentration and 20 µg of protein was loaded for each sample . For Rad53 , Smc6-FLAG and Smc6-HA detection , membranes were cut after the blocking step and the lower part was incubated with anti-beta Actin antibody and the upper part of the membranes were incubated with anti-Rad53 , anti-FLAG and anti-HA , respectively . To detect Scc1-FLAG , the membranes were not cut . Instead , the membranes were incubated with anti-FLAG and anti-beta Actin antibody simultaneously . The following antibodies were used for detection: anti-Rad53 ( Abcam , ab104232 ) , anti-FLAG ( SIGMA , F1894 ) , anti-HA ( Roche , clone 3F10 ) and anti-beta Actin to detect Act1 ( Abcam , ab8224 ) .
When cells divide , sister chromatids have to be segregated away from each other for the daughter cells to obtain a correct set of chromosomes . Using yeast as model organism , we have analyzed the function of the cohesin and the Smc5/6 complexes , which are essential for chromosome segregation . Cohesin is known to hold sister chromatid together until segregation occurs , and our results show that cohesin also controls Smc5/6 , which is found to associate to linked chromatids specifically . In line with this , our analysis points to that the chromosomal localization of Smc5/6 is an indicator of the level of entanglement between sister chromatids . When Smc5/6 is non-functional , the resolution of these entanglements is shown to be inhibited , thereby preventing segregation of chromatids . Our results also indicate that DNA entanglements are maintained on chromosomes at specific sites until segregation . In summary , we uncover new functions for cohesin , in regulating when and where Smc5/6 binds to chromosomes , and for the Smc5/6 complex in facilitating the resolution of sister chromatid entanglements .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "mitosis", "biochemistry", "dna", "replication", "cell", "biology", "chromosome", "biology", "chromosome", "structure", "and", "function", "centromeres", "cell", "cycle", "and", "cell", "division", "genetics", "biology", "and", "life", "sciences", "dna", "cell", "processes", "chromatids", "chromosomes" ]
2014
The Chromosomal Association of the Smc5/6 Complex Depends on Cohesion and Predicts the Level of Sister Chromatid Entanglement
Trypanosoma brucei's mitochondrial genome , kinetoplast DNA ( kDNA ) , is a giant network of catenated DNA rings . The network consists of a few thousand 1 kb minicircles and several dozen 23 kb maxicircles . Here we report that TbPIF5 , one of T . brucei's six mitochondrial proteins related to Saccharomyces cerevisiae mitochondrial DNA helicase ScPIF1 , is involved in minicircle lagging strand synthesis . Like its yeast homolog , TbPIF5 is a 5′ to 3′ DNA helicase . Together with other enzymes thought to be involved in Okazaki fragment processing , TbPIF5 localizes in vivo to the antipodal sites flanking the kDNA . Minicircles in wild type cells replicate unidirectionally as theta-structures and are unusual in that Okazaki fragments are not joined until after the progeny minicircles have segregated . We now report that overexpression of TbPIF5 causes premature removal of RNA primers and joining of Okazaki fragments on theta structures . Further elongation of the lagging strand is blocked , but the leading strand is completed and the minicircle progeny , one with a truncated H strand ( ranging from 0 . 1 to 1 kb ) , are segregated . The minicircles with a truncated H strand electrophorese on an agarose gel as a smear . This replication defect is associated with kinetoplast shrinkage and eventual slowing of cell growth . We propose that TbPIF5 unwinds RNA primers after lagging strand synthesis , thus facilitating processing of Okazaki fragments . Trypanosomes and related parasites cause tropical diseases such as sleeping sickness and Chagas disease . As one of the earliest diverging eukaryotes that contain a mitochondrion [1] , this group of parasites is well known for unusual biological properties . For example , their mitochondrial genome , known as kinetoplast DNA ( kDNA ) , has an amazing and unprecedented structure , a giant DNA network residing in the cell's single mitochondrion [2] , [3] . The network is a planar structure composed of interlocked DNA rings including several thousand minicircles and a few dozen maxicircles . Within the mitochondrial matrix the kDNA network is condensed into a disk-shaped structure that is positioned near the flagellar basal body , which resides in the cytoplasm . The kDNA disk , called the kinetoplast , is actually connected to the basal body by a transmembrane filament system named the tripartite attachment complex ( TAC ) [4] . Like mitochondrial DNA in other organisms , maxicircles encode ribosomal RNAs and a handful of mitochondrial proteins such as subunits of respiratory complexes . Many maxicircle transcripts require editing before they can serve as functional mRNAs . Editing is an unusual RNA processing reaction involving addition or deletion of uridylate residues at specific internal sites of mRNAs ( reviewed in [5] , [6] ) . In some transcripts , editing occurs on a massive scale , with uridylates introduced by editing constituting more than half of the sequence of the resulting mRNA . Minicircles encode small guide RNAs that serve as templates for editing , thereby controlling its specificity . In this paragraph we will briefly discuss the kDNA replication mechanism in T . brucei , focusing on minicircles . The initial step in replication is the vectorial release of individual minicircles into the space , known as the kinetoflagellar zone ( KFZ ) , between the kDNA disk and the membrane near the flagellar basal body [7] . Here the free minicircles encounter proteins that assemble and propagate a replication fork , resulting in unidirectional replication as theta structures . The progeny minicircles are thought to segregate in the KFZ , and then migrate to the antipodal sites , two protein assemblies that flank the kDNA disk and are positioned about 180° apart [8] . At this time the monomeric minicircle replication products contain either a single continuously synthesized leading strand or they contain unligated Okazaki fragments [9] . Within the antipodal sites the Okazaki fragments are processed . Although the detailed processing mechanism is unknown it probably involves several enzymes that localize within the antipodal sites . These enzymes , which have been studied to varying degrees , include structure-specific endonuclease I [10] , [11] , DNA polymerase β [12] , and DNA ligase kβ [13] . These enzymes are thought to participate in removal of RNA primers and to fill and close the resulting gaps . The processed minicircles , containing either the newly synthesized leading strand or lagging strand and still containing at least one gap , are then attached to the network periphery by a topoisomerase II that is also situated in the antipodal sites [14] , [15] . Since two minicircles are attached for every one removed , the network grows in size . Only when the minicircle copy number has doubled are their remaining gaps repaired , most likely by DNA polymerase β-PAK [16] and DNA ligase kα [13] , two enzymes that reside within the kDNA disk . Then the network splits in two and its progeny , each identical to the parent , are pulled into the two daughter cells by their connection ( via TAC ) to the flagellar basal bodies [4] . Recently we discovered 8 proteins in T . brucei that are related to the Saccharomyces cerevisiae mitochondrial helicase ScPIF1 , and we named them TbPIF1-8 . Remarkably , six of these are localized at several different positions in the mitochondrion; of the other two , one is nuclear and the other appears to be in the cytoplasm [17] . We have so far studied only one of the mitochondrial proteins , TbPIF2 , and have found it to be a helicase that is essential for maxicircle replication [17] . Here we report that TbPIF5 ( Genbank accession No . : XP_847187; GeneDB accession No . : Tb927 . 8 . 3560 ) is a DNA helicase involved in minicircle Okazaki fragment processing , probably by unwinding the hybrid helices between RNA primers and the DNA template . We previously localized TbPIF5 to the antipodal sites by expressing an ectopic gene encoding a TbPIF5-GFP fusion protein [17] . To localize TbPIF5 encoded at its endogenous locus , we introduced a sequence encoding a myc epitope at the 3′ end of one endogenous allele of TbPIF5 gene . This protein would more likely be expressed at its normal level . Our immunofluorescence studies on this protein confirmed that TbPIF5 localizes within the antipodal sites ( Fig . 1 ) . Since almost all the cells in an asynchronous log phase culture had this localization , it is likely that this protein does not undergo significant change in its localization during the cell cycle . To determine whether TbPIF5 is actually a DNA helicase , we expressed it with a His-tag in E . coli and purified it by two steps of chromatography ( Fig . 2A ) . Recombinant TbPIF5 hydrolyzes ATP in the presence of Mg2+ and M13 ssDNA ( Fig . 2B ) , indicating that it has DNA-dependent ATPase activity . TbPIF5 also has helicase activity , releasing oligonucleotides that had been annealed to M13 single-stranded circles ( Fig . 2C ) . As expected , Mg2+ and ATP are required for this reaction ( Fig . 2D ) , and the optimal concentration for both was in the range of 0 . 5 or 1 mM ( Fig . 2D ) . To determine the polarity of helicase activity , we constructed substrates ( diagrammed in Fig . 2E ) with a short oligonucleotide ( either a or b; 5′ end-labeled with [32P]phosphate ) annealed to either the 5′ or 3′ terminus of oligonucleotide c . Under conditions in which we observed dissociation of oligonucleotide a from the duplex structure , we could not detect dissociation of oligonucleotide b . Therefore , as predicted from its homology to the yeast mitochondrial helicase , we conclude that TbPIF5 has a 5′ to 3′ helicase activity ( Fig . 2E ) . To study the function of TbPIF5 , we first tried RNAi using the pZJM vector [18] . Although ∼90% of the mRNA was degraded by 2 days after induction of RNAi ( Inset , Fig . S1A ) , there was no effect on cell growth ( Fig . S1A ) . Use of a stem-loop RNAi vector [18] gave the same result ( data not shown ) . We then tried to knock out both alleles of TbPIF5 by replacing each allele with a different drug marker . However , only one allele could be replaced as judged by Southern blot ( Fig . S1B ) . Because knockout of both alleles may be lethal , we introduced into the cell an ectopic TbPIF5 gene using the vector pLew79-MHTAP [19] . The ectopic gene expresses TbPIF5 only in the presence of tetracycline , and therefore it should allow deletion of the second genomic allele . For unknown reasons , this strategy was also unsuccessful using tetracycline concentrations ranging from 2–10 ng/ml ( data not shown ) , and thus we failed to knock out both genomic alleles . As discussed in the following paragraph , we found unexpectedly that a higher level of tetracycline , which causes overexpression of TbPIF5 , reduces the cell's growth rate . Using the ectopic expression system discussed in the previous paragraph ( except that both endogenous TbPIF5 alleles were still present ) , we found that 2 days of treatment with 1 µg/ml tetracycline caused more than a 15-fold increase ( judged by phosphorimaging ) in TbPIF5 mRNA ( see northern blot inset in Fig . 3A ) . Furthermore , this treatment reduced the cell's growth rate 4 days after tetracycline addition ( Fig . 3A ) , providing evidence that an elevated level of TbPIF5 is deleterious to the cell . TbPIF5 overexpression also caused shrinkage of kDNA networks as judged by DAPI staining of intact cells . Fig . 3B shows examples of fluorescence images of wild type cells and those that had undergone 6 days of overexpression . Fig . 3C shows kinetics of kDNA loss ( determined by visual inspection of fluorescence images like those in panel B ) following induction of overexpression . At day 6 , only ∼50% of the cells had normal-sized kDNA , 20% had small kDNA , and 30% had no detectable kDNA . We then used a different approach to evaluate minicircle and maxicircle abundance following induction of TbPIF5 overexpression . We digested total DNA with HindIII/XbaI , separated the fragments by agarose gel electrophoresis , and then probed a Southern blot for minicircles and maxicircles ( Fig . 3D ) . After 5 days of overexpression , minicircle abundance decreased by more than half , while there was only a mild effect on the level of maxicircles ( Fig . 3E ) . These results indicated that TbPIF5 overexpression selectively affects minicircles . We further examined the isolated kDNA networks by electron microscopy . The unit-sized network isolated from the uninduced cells has multiple maxicircle loops projecting from the periphery ( arrows in Fig . 4A ) . In the late stage of replication , maxicircle loops usually concentrate in the central region between the two segregating daughter networks ( arrows in Fig . 4B ) . After 6 days of TbPIF5 overexpression , some networks have become smaller in size ( Fig . 4C ) , and the structure of some networks is disorganized ( Fig . 4D ) . As usual , we observed multiple maxicircle loops extending from the edge of the networks in different stage of replication . However , they do not always concentrate in the central region of the double-sized network that is undergoing segregation ( see an example in Fig . 4E ) . To investigate whether minicircle loss caused by TbPIF5 overexpression is due to an effect on replication , we fractionated total DNA on an agarose gel in the presence of ethidium bromide and detected free minicircle replication intermediates by probing a Southern blot ( Fig . 5A ) . After 5 days of overexpression , both covalently-closed and gapped/nicked minicircles decreased by about half , consistent with the decrease in total minicircle abundance . One unexpected consequence of TbPIF5 overexpression was the appearance of a heterogeneous population of minicircle species migrating as a smear between covalently-closed and gapped/nicked minicircles . This smear , never before observed and which we call fraction H , is most prominent on days 1 to 4 . We next fractionated total DNA using neutral/alkaline 2-dimensional gel electrophoresis and analyzed free minicircle species by strand-specific hybridization ( Fig . 5B ) . In the first dimension , minicircle species were separated in TBE buffer containing ethidium bromide ( conditions identical to those used for the gel in Fig . 5A ) . In the second dimension , run in 30 mM NaOH , the double-stranded DNA was denatured . Using 5′-32P -labeled synthetic oligonucleotides , we separately probed for L- ( the leading strand ) and H-strands ( the lagging strand ) . The probes were complementary to sequences near the 5′ end of the L-strand and within the first Okazaki fragment on the H-strand . Interpretation of these gels was aided by comparison with our previous 2-D gels of minicircles from the closely-related parasite T . equiperdum [9] as well as from T . brucei [20] . As mentioned in the Introduction , these and other studies had shown that minicircles replicate unidirectionally via theta structures with the L-strand synthesized continuously and the H-strand discontinuously with ∼100 nucleotide Okazaki fragments . This mechanism is unusual in that Okazaki fragments are not joined until the θ-structures had segregated into monomeric products; joining is thought to occur within the antipodal sites [9] . In the control 2-D gel of wild type free minicircle intermediates ( Fig . 5B , upper panels ) , we show a fairly long exposure to reveal the unjoined Okazaki fragments ( OF ) derived from multiply-gapped circles ( MG ) and the diagonal of growing L-strands ranging in size up to ∼1 kb derived from θ-structures ( θ ) . Joining of most of the Okazaki fragments in a minicircle converts multiply-gapped minicircles to nicked or gapped minicircles ( N/G ) . Some minor minicircle species previously identified in wild type T . equiperdum and T . brucei such as the knotted minicircle ( K ) , linearized minicircle ( L ) , nicked dimer ( nD ) , and covalently-closed dimer ( ccD ) are not relevant to this paper and not discussed here [9] , [20] , [21] . Two-dimensional gels of minicircles from cells undergoing TbPIF5 overexpression for 1 day ( Fig . 5B , lower panels ) differed markedly from those from wild type cells ( Fig . 5B , upper panels ) . We found that fraction H has a ∼1 kb L-strand template and in the next paragraph we will present strong evidence that this strand is circular . These L-strands form a smear extending from CC to N/G ( H , left lower panel in Fig . 5B ) . The reason for smearing is that prior to denaturation they had been paired with H strands varying in size . The latter molecules form a diagonal , never observed previously , in the size range of 0 . 1 to near 1 kb ( H , right lower panel in Fig . 5B ) . Thus , fraction H likely consists of a circular L-strand paired with a family of growing H strands . Since the probe detects only the first Okazaki fragment to be synthesized , the H-strand fragments in the diagonal must include the first and form a family of ligated contiguous Okazaki fragments . Strand-specific hybridization also suggested a decrease in level of growing L-strands on θ-structures ( compare L-strand diagonal , designated θ , in left upper panel in Fig . 5B with corresponding area of left lower panel ) , although since different exposures were used it is not possible to make a firm conclusion on this point . To further characterize fraction H , we purified free minicircles by sucrose gradient centrifugation ( Fig . 5C ) and treated these molecules with T4 DNA polymerase ( plus all four dNTPs ) , T4 DNA ligase ( plus ATP ) , or both together ( Fig . 5D ) . DNA polymerase alone converts fraction H to the position of gapped/nicked minicircles , but DNA ligase alone barely affects the mobility of fraction H . However , both enzymes together convert a substantial portion of fraction H to covalently-closed minicircles . This experiment not only indicates that fraction H is a gapped molecule with ligated Okazaki fragments but also provides evidence that the L-strand of fraction H is a circle . If TbPIF5 is involved in primer removal , it is possible that its overexpression might reduce the number or length of primers on either free minicircles or those linked to the network . We previously reported that in T . brucei there are no ribonucleotides on the 5′ end of either the newly synthesized L-strand or the first Okazaki fragment on minicircles that were linked to the network [11] . However , we never had searched for primers on free minicircles . Using the strategy we developed previously [11] , we investigated whether primers were present before and after TbPIF5 overexpression ( Fig . 6 ) . We isolated kDNA networks and free minicircle intermediates ( from both uninduced and 1 day overexpression cells ) , digested them with TaqI , and fractionated the products on a denaturing 9% polyacrylamide gel . We then probed a Southern blot for the first Okazaki fragment . This fragment , containing ∼73 nucleotides but with a slightly heterogeneous 3′ end , had been converted by TaqI to a slightly smaller fragment ( 66 nucleotides ) with a homogeneous 3′ end ( Fig . 6A ) . This species , whether derived from free minicircles or network minicircles , was not altered by alkali treatment , indicating that there are no ribonucleotides on its 5′ end or anywhere else within the molecule . Using a similar strategy , we searched for ribonucleotides at the 5′ terminus of the continuously-synthesized L-strand . We cleaved the minicircles with HpyCH4V , which release a 69 nucleotide terminal L-strand fragment ( Fig . 6B ) . Again , there is no ribonucleotide attached at the 5′ end of the newly-synthesized L-strand . In a recent search for T . brucei mitochondrial DNA helicases , we found that the genome encodes 8 proteins related to ScPIF1 , a mitochondrial helicase of S . cerevisiae . Remarkably , 6 of the T . brucei PIF1-related gene products are mitochondrial [17] . Here we report the properties of one of these enzymes , TbPIF5 , which is localized in the antipodal sites ( Fig . 1 ) . As shown in Fig . 2 , we found that a recombinant protein had helicase activity , with a 5′ to 3′ polarity , similar to that of the yeast homolog [22] . We did not observe a phenotype following RNAi of TbPIF5 , even though ∼90% of the mRNA was depleted within 2 days ( Fig . S1 ) . We could knock out one , but not both alleles of TbPIF5 , raising the possibility that the gene is essential . Surprisingly , the genome of a related kinetoplastid , Leishmania major , encodes only 7 PIF1-like helicase genes , and the counterpart of TbPIF5 gene is apparently absent [17] . Although this fact might support an argument that TbPIF5 could be dispensable we cannot rule out the possibility that other PIFs may take over TbPIF5's functions in L . major . Like T . brucei , the T . cruzi genome contains 8 genes related to ScPIF1 . Although RNAi and single allele knockouts did not affect cell growth or kDNA size as determined by DAPI staining , we did observe a striking effect of TbPIF5 overexpression on the replication of minicircles . Not only was there a slowing of growth and loss of kDNA minicircles ( Fig . 3 ) , but there was an alteration in joining of Okazaki fragments ( Fig . 5 ) . Before we discuss these new data , we will review what is known about primer removal and other processing reactions of minicircle Okazaki fragments . We will also review Okazaki fragment joining in the nucleus of other eukaryotes . There is a fundamental difference between processing of trypanosome minicircle Okazaki fragments with that in other cells . In either prokaryotes or eukaryotes , Okazaki fragment primers are generally removed and fragments are ligated immediately after their synthesis [23] . In trypanosome mitochondria , on the other hand , minicircle Okazaki fragments are not joined until after the progeny minicircles have segregated . In T . brucei , theta-type replication apparently occurs in the KFZ , and then the segregated progeny are thought to migrate to the antipodal sites ( probably with one sister minicircle going to each antipodal site [2] ) . At this stage the progeny molecules with a newly-synthesized H-strand are designated multiply-gapped circles , and the gaps are positioned between the ∼100 nucleotide Okazaki fragments [9] , [24] . The presence in the antipodal sites of multiply-gapped minicircles ( with a 3′ OH terminus on each Okazaki fragment ) explains the intense in situ labeling of these sites by terminal deoxynucleotidyl transferase and a fluorescent dNTP [25] . The antipodal sites also contain enzymes that likely function in primer removal and gap repair . These include structure-specific endonuclease I ( SSE-1 , homologous to the 5′ exonuclease domain of bacterial DNA polymerase I ) [26] . RNAi of SSE-1 confirms its involvement in primer removal [11] . It is likely that following primer removal all but one of the gaps are repaired by DNA polymerase β and DNA ligase kβ , both of which are positioned in the antipodal sites [13] , [16] . Following repair of most but not all gaps , these minicircles , together with their sister minicircles ( also containing a single gap adjacent to or overlapping the L strand start site ) are reattached to the network periphery by a topoisomerase II that is also positioned in the antipodal sites [14] , [15] . Neither free minicircles nor network minicircles from procyclic T . brucei contain 5′ ribonucleotides derived from primers ( Fig . 6 and [11] ) , suggesting that in these cells primer removal is efficient . In contrast , we found one or two ribonucleotides on network minicircles ( both on the leading strand and at least on the first Okazaki fragment ) , in cells that had undergone RNAi knockdown of SSE-1 [11] . However , the newly-synthesized L-strands on network minicircles in T . equiperdum bloodstream forms have one or two 5′ ribonucleotides [27] and in the related parasite C . fasiculata has up to six [28] . No residual RNA primer was found associated with the minicircle H strand fragments in T . equiperdum [29] . Finally , we do not know for any of these parasites the initial length of the primer or all of the enzymes involved in their removal . C . fasciculata has a mitochondrial RNase H1 [30] and a comparable enzyme is found in T . brucei [31]; this enzyme may also contribute to primer removal . To understand processing of minicircle Okazaki fragments it is essential to consider the enzymology of this complex pathway in nuclei of other eukaryotes . Proteins involved in this process include flap endonuclease 1 ( FEN1 ) , RNase H , Dna2p , replication protein A ( RPA ) , DNA polymerase δ , and DNA ligase I [23] . RNase H removes the primer one nucleotide upstream of RNA-DNA junction [32] , and the remaining ribonucleotide is then cleaved by FEN1 [33] . S . cerevisiae also contains an RNase H-independent pathway in which DNA polymerase δ can strand-displace the RNA primer , forming a flap intermediate . Most flap intermediates are short and can be cleaved by FEN1 itself [34]–[37] . However , long flaps ( >30 bases ) may also be generated by DNA polymerase δ . The long flap is then coated by the single-strand binding protein RPA , which recruits Dna2p , a protein with both 5′ to 3′ helicase and nuclease activities . Dna2p cleaves the long flap into a shorter flap that is subsequently removed by FEN1 . Finally the resulting gap is repaired by polymerase δ and ligase I [38] , [39] . Recent studies in yeast have uncovered a role for PIF1 helicase in these reactions ( ScPIF1 is found in both the mitochondria and the nucleus ) [40]–[42] . The genetic interaction between PIF1 , DNA2 and a subunit of pol δ ( POL32 ) , together with the biochemical studies [43] , [44] , indicate that Pif1p may assist pol δ in generating the flap , which is processed subsequently by Dna2p [45] . The mechanism by which Pif1p functions in this process is still unclear . Here we found that TbPIF5 plays an important role in minicircle Okazaki fragment maturation . Our most significant finding was that overexpression of TbPIF5 causes accumulation of fraction H , which is a minicircle species that contains a growing lagging strand ( ranging from 0 . 1 kb to 1 kb ) on the 1 kb L-strand templates . We now propose a model explaining how TbPIF5 overexpression causes accumulation of fraction H ( Fig . 7B ) . As discussed above ( and diagramed in Fig . 7A ) , Okazaki fragment joining in wild type cells does not occur until after minicircle progeny have segregated and migrated to the antipodal sites . TbPIF5 ( alone or together with other proteins ) likely unwinds RNA primers , generating flaps that are subsequently degraded . The gaps are filled and repaired probably by DNA polymerase β and DNA ligase kβ . To prevent pre-maturation of Okazaki fragments , cells must tightly control the recruitment of some key enzymes such as TbPIF5 . For example , TbPIF5 may bind to the minicircle progeny only after their segregation and migration to the antipodal sites . It would not be surprising that overexpression of TbPIF5 perturbs the timing and location of Okazaki fragment processing . Excess TbPIF5 could bind to minicircle θ-structures , triggering premature removal of primers ( Fig . 7B ) and permitting joining of Okazaki fragments . If TbPIF5 also removes RNA primers that are not yet extended by a DNA polymerase , then further extension of the H-stand would be effectively blocked . L-strand synthesis would proceed to completion , allowing segregation of a sister with a full length newly-synthesized L-strand and another with a truncated H strand in which the Okazaki fragments had been joined . The latter molecules , with a heterogeneously-sized H-strand , form fraction H . Topoisomerase II might not recognize these molecules and therefore fail to reattach them to the network . Thus , fraction H gradually accumulates , presumably within the antipodal sites . This defect in minicircle attachment could explain the shrinking and eventual loss of kDNA that occurs following overexpression of TbPIF5 . Further studies are needed on this helicase and other proteins involved in primer removal to fully understand the mechanism of minicircle Okazaki fragment processing . Procyclic strain 29-13 ( from G . Cross , Rockefeller University ) was used for RNAi . Procyclic strain 927 was used for the localization experiment . Conditions for cell culture and transfection were described previously [18] , [46] . The first 500 bp of the TbPIF5 coding sequence were PCR-amplified using genomic DNA isolated from procyclic strain 427 and inserted into the pJZM and stem-loop vectors [18] . RNAi methodologies were described previously [18] . DNA and RNA purification , gel electrophoresis , Southern blotting , Northern blotting , and sucrose gradient sedimentation were performed as described previously [20] . The TbPIF5 knockout was conducted as described previously [47] . Electron microscopy of isolated kDNA networks was done as described [48] . Fragments of the 3′-end of TbPIF5 coding region ( 500 bp ) and its neighboring 3′ untranslated region ( 500 bp ) were PCR amplified using primers a–d: a , 5′GACCGGTACCCGTCTCACGCGCTTACCTATTG 3′; b , 5′ GCAGCTCGAGTTCTTCCACTTCCCCTTCATACTCCCC 3′; c , 5′ GCGGGGATCCCCGAGAGCGATGAGCGAAAAAG 3′; d , 5′ GCATCGGGGCGGCCGCACTCTCTCTCTCTCCATCTATGAATGC 3′ . PCR products were inserted into pMOTag33M [49] . After digestion with Acc65I and NotI , the DNA fragments were transfected into procyclic strain 927 . The coding sequence ( minus the first 49 amino acids which constitute a predicted mitochondrial targeting signal ) was amplified by PCR , cloned into pET28a ( Novagen ) , and transformed into the E . coli Rosetta™ ( DE3 ) pLysS strain ( Novagen ) . The cells were inoculated into 500 ml of LB medium ( containing 34 µg/ml chloramphenicol and 30 µg/ml kanamycin ) and grown at 37°C to an OD600 nm of 0 . 6 . After addition of 1 mM IPTG , the culture was incubated for another 3 h at 25°C . Cells were harvested by centrifugation ( 8000 g , 10 min ) and the cell pellet was resuspended in 20 ml buffer A ( 50 mM sodium phosphate , 300 mM NaCl , 10 mM imidazole , pH 8 . 0 ) . After lysis by sonication , the suspension was centrifuged ( 10000 g , 30 min ) and the supernatant was mixed gently with 2 ml Ni-NTA slurry ( Qiagen ) ( 1 h , 4°C ) . The Ni-NTA beads were then washed 4 times with 2 ml buffer B ( 50 mM sodium phosphate , 300 mM NaCl , 20 mM imidazole , pH 8 . 0 ) . Proteins were eluted 3 times with 0 . 5 ml buffer C ( 50 mM sodium phosphate , 300 mM NaCl , 250 mM imidazole , pH 8 . 0 ) . The eluates were dialyzed overnight at 4°C against buffer D ( 25 mM Tris-HCl , 300 mM NaCl , 1 mM DTT , pH 7 . 5 ) . The samples were loaded onto a 0 . 5 ml heparin-Sepharose FF ( Bioscience Healthcare ) column equilibrated with the same buffer . Recombinant protein was eluted at 0 . 8 M NaCl and dialyzed against buffer E ( 25 mM Tris-HCl , 100 mM NaCl , 1 mM DTT , pH 7 . 5 ) . Recombinant TbPIF5 is very unstable and it was freshly prepared for the activity assays . For ATPase assay , recombinant TbPIF5 ( 10 , 20 , and 50 ng ) was incubated ( 20 µl reaction , 10 min , 37°C ) with 8 . 25 nM [γ-32P] ATP ( 6000 Ci/mmol ) , 150 µM non-radioactive ATP , 50 mM Tris-HCl , pH 8 . 5 , 50 mM NaCl , 2 mM DTT , 2 mM MgCl2 , 0 . 25 mg/ml bovine serum albumin , and 50 ng M13mp18 ssDNA . Samples ( 1 µl ) were spotted onto a polyethyleneimine-cellulose plate ( J . T . Baker , USA ) and developed in 1 . 0 M formic acid/0 . 5 M LiCl followed by autoradiography . For helicase assays , the M13-based substrate was constructed as described [50] and the substrates for polarity assay were made as described [51] . Assays ( 20 µl each ) contained various amounts of TbPIF5 , 50 mM Tris-HCl , pH 8 . 5 , 50 mM NaCl , 2 mM DTT , 2 mM MgCl2 , 2 mM ATP , 0 . 25 mg/ml bovine serum albumin , and the substrate ( 15 fmol ) . Reactions were incubated at 37°C for 10 min and subjected to electrophoresis with a 12% polyacrylamide gel in 0 . 5×TBE ( 150 V , 1 h ) . The gel was dried and autoradiographed .
Trypanosoma brucei is a protozoan parasite that causes human sleeping sickness in sub-Saharan Africa . Trypanosomes are primitive eukaryotes and they have many unusual biological features . One prominent example is their mitochondrial genome , known as kinetoplast DNA or kDNA . kDNA , with a structure unique in nature , is a giant network of interlocked DNA rings known as minicircles and maxicircles . kDNA superficially resembles chain mail in medieval armor . The network structure dictates an extremely complex mechanism for replication , the process by which two progeny networks , each identical to their parent , are formed . These progeny networks then segregate into the daughter cells during cell division . One feature of this replication pathway , in which discontinuously synthesized strands of minicircles are joined together in a reaction involving an enzyme known as a helicase , is the subject of this paper . Since there is nothing resembling kDNA in human or animal cells , and since kDNA is required for viability of the parasite , enzymes involved in this pathway are promising targets for chemotherapy .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "molecular", "biology/dna", "replication", "microbiology/parasitology" ]
2009
TbPIF5 Is a Trypanosoma brucei Mitochondrial DNA Helicase Involved in Processing of Minicircle Okazaki Fragments
Although evolution is a multifactorial process , theory posits that the speed of molecular evolution should be directly determined by the rate at which spontaneous mutations appear . To what extent these two biochemical and population-scale processes are related in nature , however , is largely unknown . Viruses are an ideal system for addressing this question because their evolution is fast enough to be observed in real time , and experimentally-determined mutation rates are abundant . This article provides statistically supported evidence that the mutation rate determines molecular evolution across all types of viruses . Properties of the viral genome such as its size and chemical composition are identified as major determinants of these rates . Furthermore , a quantitative analysis reveals that , as expected , evolution rates increase linearly with mutation rates for slowly mutating viruses . However , this relationship plateaus for fast mutating viruses . A model is proposed in which deleterious mutations impose an evolutionary speed limit and set an extinction threshold in nature . The model is consistent with data from replication kinetics , selection strength and chemical mutagenesis studies . Mutations result from biochemical processes such as replication errors , editing , or nucleic acid damage , but their spread and fixation is a population-genetics process that takes place over much broader scales . Remarkably , the neutral theory of molecular evolution posits that the mutation rate should be the sole determinant of molecular evolution rates [1] , and adaptation theory also assigns a central role to mutation [2] , [3] . However , it is unclear to what extent this direct association between mutation and evolution holds true in nature because a variety of complex selective , ecological and demographical factors can potentially affect the evolutionary process at the molecular level [2] , [4] , [5] . Viruses offer an excellent system for addressing this question because their evolution is fast enough to be measured directly from isolates collected within timescales of years [6] and their mutation rates vary by several orders of magnitude [7] . The viral mutation rate has been shown to determine pathogenesis [8] , [9] , the risk of drug resistance [10] , vaccine efficacy [11] , [12] , the success of antiviral treatments [13]–[15] , or the likelihood of emergence of new diseases [16] , [17] . It is also known that most RNA viruses evolve extremely fast owing to their high mutation rates , which in turn are explained biochemically by the absence of proofreading or repair mechanisms [5] , [15] . However , our current knowledge of the evolutionary consequences of viral mutation is mainly qualitative or restricted to a small subset of viruses . By undertaking a systematic quantitative analysis , this work shows that the evolution rates of major viral groups in nature are consistent with mutation rate estimates obtained under controlled laboratory conditions . The size , polarity and number of genome strands are identified as major determinants of viral mutation and evolution . According to a purely neutral model , the evolution rate should increase linearly with the mutation rate , and this prediction is confirmed for viruses with relatively low mutation rates . However , evolution rates increase less than linearly as mutation rates become higher . A model in which the fitness load imposed by transient deleterious mutations retards molecular evolution is proposed , and the inferred parameters are tested using data from site-directed mutagenesis studies and other sources of evidence . This model predicts that further increases of the mutation rate would have a negative impact on viral evolution and suggests that RNA viruses replicate near an extinction threshold in nature . A recent compilation of experimentally-determined mutation rates yielded 37 standardized estimates for 23 viruses [7] . These rates range from 10−8 to 10−3 substitutions per nucleotide site per cell infection ( s/n/c ) and vary significantly among the major groups defined by the Baltimore classification of viruses ( Figure 1a; nested ANOVA: P = 0 . 002 ) . For evolution rates , 223 estimates corresponding to 84 different viruses were collected ( Text S1 ) , all of which were obtained using Bayesian analysis of dated sequences [18] and after validation of the molecular clock . This methodological consistency is critical to make reliable comparisons since evolution rates can vary strongly depending on the estimation procedure [19] , [20] . The collected evolution rates range from 10−6 to 10−2 substitutions per nucleotide site per year ( s/n/y ) and also vary significantly among Baltimore groups ( Figure 1b; nested ANOVA: P<0 . 001 ) . The fastest evolution corresponds to single-stranded ( ss ) RNA and reverse-transcribing ( RT ) viruses , followed by double stranded ( ds ) RNA and ssDNA viruses , whereas dsDNA viruses evolve more slowly on average ( Tukey's post-hoc test: P<0 . 05 ) . This confirms the well-known difference between RNA and DNA viruses [5] , [15] and , further , demonstrates that single-stranded viruses tend to evolve faster than double-stranded viruses regardless of whether their genetic material is RNA or DNA ( two-way nested ANOVA excluding RT viruses: P<0 . 001 ) . Among the seven viruses for which both mutation and evolution rates have been determined , these correlate positively ( Figure 2a; Pearson r = 0 . 813 , P = 0 . 026 ) . Furthermore , when averages are calculated for each Baltimore group using all available estimates , mutation and evolution rates show a strongly positive correlation ( Figure 2b; r = 0 . 946 , P = 0 . 004 ) . Consider first a purely neutral model in which the evolution rate K is proportional to the mutation rate μ [1] . Therefore , , or equivalently , , log-scales being here more appropriate for model fitting because the data range several orders of magnitude . Notice that the model specifically predicts that the linear relationship between log K and log μ has slope 1 . 0 . The value of a depends on the number of cell infection cycles ( generations ) per year ( g ) and on the fraction of effectively neutral mutations ( α ) such that . For viruses with relatively low mutation rates ( dsDNA , ssDNA and dsRNA viruses ) , this model fits the data accurately ( r2 = 0 . 995 , Figure 2b ) , yielding log10 a = 2 . 37±0 . 02 ( SEM ) . This is in full agreement with the neutral theory , although adaptive evolution may produce a similar pattern in some cases [2] . However , K increases less than linearly with μ for the fastest mutating viruses ( ssRNA an RT viruses ) and the overall fit of the above model is poor ( r2 = 0 . 432 ) . Because transient deleterious mutations are highly abundant in RNA virus populations [21] , the spread and fixation of mutations should be slowed down by the presence of deleterious mutations elsewhere in the genome . Specifically , the expected fraction of individuals not carrying these mutations is , where G is genome size and sH the harmonic mean of selection coefficients [3] . Taking this into account , the predicted evolution rate becomes , with and . This modification strongly improves the model ( Figure 2b; r2 = 0 . 995; partial F-test: P<0 . 001 ) , yielding log10 a = 2 . 387±0 . 027 and b = 3 . 744±0 . 172 . Therefore , short-lived deleterious mutations appear to play a key role is setting the rate of molecular evolution in RNA viruses . Although this effect concerns mainly neutral evolution , it should also be relevant to models of adaptation [3] . The inferred values for parameters a and b are consistent with independent sources of evidence . Site-directed mutagenesis studies in which the fitness effects of point mutations were determined for tobacco etch virus [22] , bacteriophage Qβ [23] and vesicular stomatitis virus [24] ( three ssRNA viruses ) gave α≈0 . 27 and sH values ranging from 0 . 172 to 0 . 338 ( Text S1 ) . Using allows us to obtain an estimate of b which ranges from 1 . 475 to 5 . 177 and includes the value b = 3 . 744 inferred above . However , the interval is relatively wide and thus does not provide a very stringent test of the model . Additionally , although the three viruses belong to different families and infect widely different hosts , some caution is granted because the estimated sH was based on three species only . Concerning parameter a , if we again assume α = 0 . 27 the estimated number of cell infection cycles per year averaged across viruses is . This is equivalent to one cell infection every 10 h , which is a realistic value for a variety of actively replicating eukaryotic viruses [25]–[28] . Therefore , the above model linking mutation and evolution rates is in broad agreement with empirical evidence from quantitative replication kinetics , selection strength and chemical mutagenesis studies . Previous experimental work has shown that slight elevations of the mutation rate ( on the order of threefold ) can lead to drastic fitness losses in a variety of ssRNA and RT viruses and often achieve mutagenesis-induced population extinction in the laboratory , suggesting a possible antiviral strategy [13]–[15] . However , evidence showing the relevance of these observations in natural populations has remained elusive . The above model predicts that the rate of evolution should be maximal when the genomic mutation rate is ( i . e . when ) , and then decays exponentially . The mean genomic mutation rates of ss ( + ) RNA viruses ( 0 . 663±0 . 417 ) , ss ( − ) RNA viruses ( 0 . 372±0 . 124 ) and RT viruses ( 0 . 445±0 . 116 ) are slightly higher but not significantly different from this value ( one-sample t-tests: P≥0 . 150 ) , implying that these viruses replicate close to the optimal mutation rate . However , this also means that further increases of the mutation rate would actually reduce the evolution rates of these viruses in nature and potentially endanger their survival . For instance , on average , a threefold increase in the mutation rate of ss ( + ) RNA viruses would produce a 48-fold evolutionary slowdown . Figure 3 shows the predicted relationship between mutation and evolution rates for hepatitis C virus , poliovirus 1 , influenza A virus , and human immunodeficiency virus 1 , four well-studied human viruses . Drake's rule establishes that the genomic mutation rate is approximately constant across DNA microorganisms ( including viruses ) and equal to 0 . 003 substitutions per generation [29] , [30] . Since μ≈0 . 003/G , it is possible to use G as an inverse correlate of μ to further test the association between mutation and evolution rates . If mutation rates determine evolution , DNA viruses with small genomes should tend to evolve faster than those with large genomes . Confirming this prediction , the evolution rates of 19 different DNA viruses correlate negatively with their genome sizes ( Figure 4a; partial r = −0 . 707 , P = 0 . 001 ) and , importantly , this correlation remains significant after accounting for the fact that ssDNA viruses usually have smaller genomes than dsDNA viruses ( partial r = −0 . 551 , P = 0 . 022 ) . A linear regression of the form gives m = −0 . 906±0 . 216 and p = −0 . 349±0 . 890 . The estimate of m does not deviate significantly from 1 . 0 ( t-test: P = 0 . 667 ) , further supporting the linear relationship between K and μ shown above for slowly-mutating viruses . An apparent outlier is the human papillomavirus ( HPV ) 16 , which evolves faster than expected from its genome size . However , this rate was obtained from sequences sampled only three years apart , and there is a known tendency for evolution rate estimates to become inflated in the short-term [20] . Indeed , the HPV-16 estimate and those for varicella zoster and human adenovirus C are considered unreliable [31] . Small dsDNA viruses ( HPV-16 and two polyomaviruses ) are also problematic because their mutation rates have not been determined and there is little consensus about their evolution rates [31] . However , supporting the robustness of the results , the above correlations remain unaffected or even improve after removing these five viruses ( r = −0 . 791 , P = 0 . 001 and r = −0 . 629 , P = 0 . 029 , respectively ) . Concerning RNA viruses , recent work suggests that there may also be an inverse relationship between mutation rates and genome sizes , although less evident than for DNA viruses [7] probably because their narrower genome size range makes it more difficult to demonstrate this association . Interestingly , genome sizes and evolution rates also correlate negatively among RNA viruses [32] , yet this correlation is much weaker than for DNA viruses ( Figure 4b; partial correlation excluding RT viruses: r = −0 . 267 , P = 0 . 038 ) . Several factors can determine viral evolution in addition to the mutation rate . For instance , viruses undergoing fewer replication cycles per time unit should evolve more slowly , and this appears to be the case of water- or vector-borne viruses which spend longer periods of inactivity than directly transmitted viruses [32] , [33] . In latently integrated retroviruses , viral replication is carried out by the host machinery , thus reducing dramatically the mutation rate compared to actively replicating viruses . This can explain why a very low rate of evolution was inferred for foamy virus based on a well-supported host-virus co-speciation pattern [34] . Similar results have been obtained for vertically transmitted human T-cell leukemia viruses [35] and for papilloma viruses coevolving with felids [36] . On the other hand , positive selection associated with recent host jumps or immune pressure can accelerate evolution , and a similar effect occurs in viruses experiencing strong transmission bottlenecks because this reduces the effective population size and relaxes selection against deleterious mutations [17] . Taking these factors into account and given that estimation errors are usually large , it is not surprising that mutation and evolution rates show considerable scatter , and that their relationship becomes evident only after averaging large and comparable datasets . Another necessary caveat is that time-structured sequence data spanning years or decades often contain short-lived polymorphisms . Among other factors , this explains why evolution rates inferred in this manner are generally higher than those based on long-term calibration points such as co-speciation events [20] , and warns against comparing rates obtained by such different methods . Although the dataset used here was based on dated samples only and was methodologically consistent , sampling timespans were inevitably variable , but this was accounted for in the statistical analysis . It was not possible to use studies based on co-speciation events for testing the model because estimates obtained in this way are not reliable for most viral types . Finally , a potential methodological pitfall is that some viral species and families have been more extensively studied than others , thus introducing sampling bias . A more robust analysis consists of giving the same weight to all species and families , independent of the number of estimates available for each . This alternative averaging method gave very similar results ( see Methods for details ) . The classical notion that RNA viruses are the fastest mutating and evolving entities in nature has been revised , after several recent reports showing that the evolution rates of ssDNA viruses are similarly high [5] . This is compatible with the finding that mutation and evolution rates are generally higher in single-stranded viruses than in double-stranded viruses ( RNA or DNA ) , a possible explanation being the greater instability of single-stranded nucleic acids [37] . On the other hand , mutation and evolution rates appear to vary smoothly across viruses and , therefore , defining discrete categories may not be a helpful approach . From a broader perspective , despite the extreme diversity of viral types , the above model provides a simple and general framework for how mutation rates determine viral evolution . This generality is achieved after incorporating the impact of deleterious mutations on evolution , which should be particularly significant in ssRNA and RT viruses . The proposed model is consistent with several independent sources of evidence but , to further consolidate it , additional empirical studies will be needed . For instance , selection strength data have been obtained experimentally only for a handful of viruses . Also , the mutation rates in eukaryotic ssDNA and dsRNA viruses are largely unknown , and a similar uncertainty exists for evolution rates in bacteriophages and small dsDNA viruses . The present work provides well-defined predictions for addressing these issues in the future . Mutation rates were taken directly from a recent meta-analysis [7] . Evolution rates were selected from the literature according to the inclusion criteria indicated in the text . For methodological consistency , estimates based on long-term virus-host co-speciation events were not used . Although co-speciation is well supported for some DNA viruses , it is not for most RNA viruses [38] . Since evolution rates are known to be time-dependent [20] , inclusion of co-speciation data for DNA viruses would inflate differences between DNA and RNA viruses . Similarly , within and among host evolution rates differ systematically [39] and thus the former were not used . Estimates from the same study but corresponding to different datasets ( genes or groups of sequences ) were treated as independent observations , whereas those obtained from the same dataset using different methods were averaged before analysis or the best-fit value was used if available . Because raw rates ranged several orders of magnitude log-transformed data were used . Data normality was satisfied within each Baltimore category after this transformation for both mutation and evolution rates ( Kolmogorov-Smirnov tests: P≥0 . 258 ) . In ANOVA and correlation tests , the sampling timespan was used as a covariate to account for evolution rate time-dependency . Mean rates for each Baltimore category were calculated directly from individual observations . Although the Baltimore classification distinguishes between RT-DNA and RT-RNA viruses , these two groups were pooled because there were data for few species , but similar results were obtained using the full set of Baltimore groups ( not shown ) . For model fitting , log-scale means were also used except for the μG term appearing in equation for which arithmetic means were used . For the ss ( + ) RNA group this mean ( 1 . 185±0 . 478 ) was strongly affected by a few extremely high estimates and thus , a previously-defined subset of more reliable estimates [7] was used instead . However , using μG = 1 . 185 , the model with deleterious mutations also provided a significant improvement over the purely neutral model ( partial F-test: P = 0 . 017 ) , yielding log10 a = 2 . 287±0 . 105 , b = 2 . 116±0 . 449 , and r2 = 0 . 914 . To account for the fact that some viral species or families were more represented than others , the analysis was repeated after calculating unweighted averages hierarchically first for each species , then for each family , then for each Baltimore group . In addition to reducing bias , this method accounts for phylogenetic relatedness ( on the other hand , it increases error because of the fewer estimates available for some groups ) . The correlation between mutation and evolution rates was maintained ( r = 0 . 923 , P = 0 . 009 ) and the model with deleterious mutations provided again the best fit ( r2 = 0 . 961; partial F-test: P<0 . 001 ) . To further test the robustness of the results , the analysis was also redone using medians instead of means . The correlation between mutation and evolution rates was r = 0 . 956 ( P = 0 . 003 ) and , again , the model with deleterious mutations explained the data better than the purely neutral one ( r2 = 0 . 926; partial F-test: P = 0 . 012 ) . In these two alternative analyses , the average μG for ss ( + ) RNA viruses was also calculated using the subset of more reliable estimates .
Viruses are an excellent system for addressing the evolutionary implications of mutation because their mutation rates vary by orders of magnitude , and their evolution takes place within the time frame of human observation . Theory posits a direct relationship between these two processes , but this has rarely been tested empirically . This work shows that evolution rates in nature correlate with experimentally-determined mutation rates for the major viral groups , and identifies key genome properties determining these rates . Current theory allows us to predict evolution rates accurately for slowly-mutating viruses but fails for the fastest mutating viruses . To solve this limitation , a model in which deleterious mutations play a key evolutionary role is proposed .
[ "Abstract", "Introduction", "Results/Discussion", "Methods" ]
[ "microbial", "mutation", "virology", "population", "genetics", "biology", "evolutionary", "biology", "microbiology", "evolutionary", "genetics" ]
2012
From Molecular Genetics to Phylodynamics: Evolutionary Relevance of Mutation Rates Across Viruses
To gain insights into complex biological processes , genome-scale data ( e . g . , RNA-Seq ) are often overlaid on biochemical networks . However , many networks do not have a one-to-one relationship between genes and network edges , due to the existence of isozymes and protein complexes . Therefore , decisions must be made on how to overlay data onto networks . For example , for metabolic networks , these decisions include ( 1 ) how to integrate gene expression levels using gene-protein-reaction rules , ( 2 ) the approach used for selection of thresholds on expression data to consider the associated gene as “active” , and ( 3 ) the order in which these steps are imposed . However , the influence of these decisions has not been systematically tested . We compared 20 decision combinations using a transcriptomic dataset across 32 tissues and showed that definition of which reaction may be considered as active ( i . e . , reactions of the genome-scale metabolic network with a non-zero expression level after overlaying the data ) is mainly influenced by thresholding approach used . To determine the most appropriate decisions , we evaluated how these decisions impact the acquisition of tissue-specific active reaction lists that recapitulate organ-system tissue groups . These results will provide guidelines to improve data analyses with biochemical networks and facilitate the construction of context-specific metabolic models . Most biological systems can be structured as networks , from cell signaling pathways to cell metabolism . These networks are invaluable for describing and understanding complex biological processes . For example , metabolic network reconstructions can illuminate the molecular basis of phenotypes exhibited by an organism , when used as a platform for analyzing data measuring gene expression , protein expression , enzymatic activity , or metabolite concentrations . For these analyses , the data are overlaid on the biological networks using Boolean rules that describe the relationship between the measured molecules ( e . g . , mRNAs , metabolites ) and the network edges and nodes . These logical rules capture how the molecules influence each other’s activity ( i . e . , activation , inhibition , or cooperation ) , and allow users to quantify each network edge or define the status of each network component as either “on” or “off” . Therefore , a biological process can be described in a given context by adding or removing nodes and/or edges based on genome-scale data . Genome-scale metabolic networks utilize this Boolean formulation connecting genes to reaction , and therefore have been used extensively as platforms for analyzing mRNA expression data to elucidate how changes in gene expression impacts cell phenotypes [1–8] . These studies have spanned diverse applications from identification of disease mechanisms [9 , 10] to identification of drug targets [11 , 12] , and the evaluation of cell responses to drugs [13] . Despite the success of the many studies integrating omics data with biochemical networks , there are several challenges in the integration of omics data with networks that are infrequently discussed . These challenges impact the accuracy of context-specific networks , and include experimental and inherent biological noise , differences among experimental platforms , detection bias , and the unclear relationship between gene expression and reaction flux [14] . Furthermore , algorithmic assumptions influence the quality and functionality of resulting models and the physiological accuracy of their predictions [15–19] . While previous work has discussed the impact of various algorithms on obtaining physiologically accurate metabolic networks , the influence of the initial steps of data integration with biological networks has not been clearly evaluated and discussed in the literature . Thus , no universal rules have been established on how to integrate transcriptomic data , referred to here as “preprocessing” , leading often to inappropriate decisions for model development . These preprocessing steps include ( 1 ) how to account for network elements ( e . g . , reactions ) that do not have a one-to-one relationship with genes and reactions ( e . g . , isozymes , complexes , and promiscuous enzymes ) , referred to here as gene mapping , and ( 2 ) how to define which genes are expressed or not , referred to here as thresholding , and ( 3 ) the order of gene mapping and thresholding in data integration . Here we evaluate the influence of the transcriptomic preprocessing steps and their consequences on the biological meaning captured by the data . Specifically , we do this by evaluating 20 different combinations of preprocessing steps , using transcriptomic data from 32 tissues . By evaluating the resulting 640 tissue-specific active reaction lists , we identify which decisions have the largest impact on list content , and which decisions best capture the similarities seen within tissues from the same organ-systems . This study aims to help researcher make proper decision on omics data integration by stress-testing existing methods and proposing improvements on their implementation . This results in guidelines for overlaying transcriptomic data in metabolic networks and the lessons learned should be applicable to the analysis of transcriptomic data in all sorts of biological networks used for systems biology analysis . Biochemical and other network types provide valuable platforms for analyzing and interpreting data . In these networks , links between nodes often represent enzyme-catalyzed reactions , and as such there is often not a one-to-one relationship between the genes and reactions . This relationship is represented using logical rules , referred as Gene-Protein-Reaction rules ( GPRs , Fig A in S1 Text ) . When overlaying mRNA abundances on biochemical networks , GPRs are used to define which genes are the main determinants of the enzyme activity catalyzing a reaction . We refer here to this step as gene mapping . The most common assumption for multimeric enzyme complexes is that the gene with the minimum expression governs the activity . For isoenzymes , the activity may either depend on the total expression of all isoenzyme genes [20] or the isoenzyme gene with highest expression [21] ( Fig 1A ) . Furthermore , the absolute mRNA abundance is often considered to represent a gene’s potential activity by using a thresholding approach . That is , if the gene is expressed at a level above a threshold , it is often considered to be active . This threshold definition has been implemented in many different ways in the literature , from the use of only a single threshold to more complex rules involving multiple thresholds . For example , one unique threshold value can be applied to all genes ( i . e . , the global thresholding approach , [22 , 23] ) while others have applied different thresholds to each gene ( i . e . , the local thresholding approach , [24 , 25] ) ( Fig 1B ) . When using one single threshold in a global context ( i . e . , global T1 ) , the genes presenting an expression above this value are considered as active ( i . e . , ON ) while the others are inactive ( i . e . , OFF ) ( Fig 1C ) . However , when multiple samples are available , one can compute a gene-specific threshold based on the distribution of the expression levels observed for this gene over all the samples ( e . g . , a local rule that sets a threshold equal to the mean expression level across all samples ) . This gene-specific thresholding approach can be implemented in combination with a defined global threshold for genes presenting low expression values among all the samples ( e . g . , below the usual detection level associated to the measurement method ) to prevent their inclusion with the active genes for some samples ( i . e . , local T1 ) . Therefore , the genes whose expression is below the value defined by this global lower bound will always be considered as inactive ( i . e . , OFF ) , while other genes will fall under the local rule for gene-specific threshold definition ( i . e . , MAYBE ON ) ( Fig 1C ) . Another similar extreme case can be encountered when the gene expression level is high in all the samples . Therefore , we propose to also analyze the influence of using one lower and one upper threshold values defined based on the distribution of expression level off all the genes in all the samples ( i . e . local T2 ) . Doing so , the local rule for gene-specific threshold definition is actually applied only to the genes whose expression is between the range of values defined by the lower and upper bounds ( i . e . , MAYBE ON ) , ensuring that genes presenting low expression values among all the samples are always considered as inactive ( i . e . , OFF ) while the ones with very high expression values among all the samples are always considered as active ( i . e . , ON ) ( Fig 1C ) . Preprocessing of transcriptomic data for their integration into biochemical networks relies mainly on these two decisions: gene mapping and thresholding , but these can be implemented in different orders , with either gene mapping or thresholding occurring first ( Fig 1D ) . Therefore , multiple combinations of these decisions could be made when overlaying data onto biochemical networks , and these decisions may influence the data integration and the subsequent biological interpretation ( see Table 1 , Fig 1 , and a detailed explanation about the decisions presented in Methods section ) . Here , we integrated transcriptomic data from 32 different tissues in the Human Protein Atlas [25] with the Human genome scale model Recon 2 . 2 [26] using 20 different combinations of the 3 main preprocessing decisions ( Table 1 , Fig 1 ) . This resulted in 640 different tissue-specific profiles of “expression” values for all gene-associated reactions in Recon 2 . 2 . To specifically evaluate the immediate impact of the preprocessing decisions on the resulting networks ( i . e . , list of reactions of a genome-scale model ( GEM ) with a non-zero expression level after overlaying the data ) , we focused our analysis on the content of the networks themselves ( i . e . , the definition of active biochemical pathways therein ) and the biological interpretation of these networks . Decisions regarding gene mapping , thresholding ( i . e . , approach and number of states ) , and order of steps affect the definition of active reaction sets . Specifically , the sets of active reactions ( i . e . , reactions of the GEM with a non-zero expression level after overlaying the data ) varied considerably in size from 358 reactions to 3286 reactions across all tissues , depending on preprocessing decisions and tissue type ( Fig 2A ) . To assess the impact of each decision , we conducted a principal component analysis ( PCA ) of the reaction sets considered as active , depending on the preprocessing decisions ( i . e . , a PCA on the matrix of all active reactions vs . all combinations of decisions and tissues; see Methods for details ) . The first principal component explains >35% of the overall variance in active reaction content ( Fig 2B ) . The thresholding related parameters ( global/local and T1/T2 ) provide the most significant contribution to the variation in the first principal component ( 38 . 5% ) , with the differences between the global and local approaches having the greatest impact ( Fig 2C , 2F and Fig B in S1 Text ) . The effect of thresholding impacted the networks more than the differences across tissues , which only explained 15 . 8% of the variation in the first principal component . Tissue specific effects did not dominate until the second principal component , where it explained 73% of the variation in the component . The order of the preprocessing steps only provides a small contribution to the explained variation in the first principal component ( Fig 2C and 2D ) . Meanwhile , the type of gene mapping has the least influence on active reaction sets ( Fig 2C and 2E ) . These results indicate that the identification of active reactions is most heavily affected by the thresholding approach ( as defined in Fig 1B ) , followed by the state definition used for thresholding ( as defined in Fig 1C ) and the order of preprocessing steps ( as defined in Fig 1D ) while the gene mapping method does not seem to have an influence . We assessed the similarities of tissues belonging to the same organ-system , based on the knowledge of the set of active reactions . We assumed that organ-system groups are formed by tissues working collaboratively to achieve a specific function ( e . g . , the gastrointestinal system turns food into energy ) . Therefore , we hypothesized that similarities of tissues within an organ system may lead to a more similar set of active metabolic reactions within the system , in comparison to other systems , as suggested by previous transcriptomic analyses [27 , 28] . To this end , we calculated Euclidean distances between pairs of tissues belonging to the same organ-system ( Fig 3 , Fig C in S1 Text , see Methods for more details ) . Our results highlight the influence of preprocessing decisions on the significance of tissue grouping at the reaction level . Moreover , we observed that some decisions improved the significance of tissue grouping: Order 2 works generally better than Order 1 . Local T2 also is better than GlobalT1 and LocalT1 . However , there was not a clearly superior approach for gene mapping in our analysis ( Fig 4 , Fig D in S1 Text ) . Some organ classification systems will group dissimilar organs together into a single organ-system , and we wondered if our analysis would still suggest the removal of such tissues from the organ-systems based on metabolic differences . For example , our previous analysis was done without associating the placenta to the Female reproductive organ-system group . However , the Human Protein Atlas groups it into the Female reproductive organ group ( S1 Table ) . The placenta is functionally and histologically different from the other tissues of this group , being derived from both maternal and fetal tissue . This biological difference was successfully captured when we compared the tissue similarity analysis with and without the placenta in the Female reproductive organ-system group ( Fig E in S1 Text ) . Our results above showed that tissues belonging to the same organ system grouped together . This suggests that active reaction sets of tissues contain biological meaning that facilitates grouping tissues belonging to the same organ system . To this end , we identified pathways known to be active in a tissue based on literature . We found 154 pathway-tissue pairs ( i . e . a given pathway is known to take place in a given tissue; S2 Table; Fig F1 in S1 Text ) and used it to evaluate different thresholding methods ( See Methods , Fig F in S1 Text ) . This resource suggested that pathways could be classified into three categories based on observed ubiquity ( i . e . , how many tissues wherein the pathways were active ) : tissue-specific ( low ubiquity , pathway is known to occur in 1–2 tissues ) ; group-specific ( medium ubiquity , pathway is known to occur in 3–10 tissues belonging to the same organ-system group ) ; and ubiquitous ( pathway is known to occur in nearly all tissues ) ( Fig F2 in S1 Text ) . The coverage of each of the pathways ( as defined in Recon 2 . 2 ) in each of the active reaction sets was evaluated using two metrics: ( i ) ubiquity , and ( ii ) false negative rate . We evaluate here the false negative rate ( i . e . , when a pathway known to be present in a tissue is not enriched in this tissue ) as opposed to false positive rate since pathways may also be present in non-canonical tissues ( i . e . tissues in which the pathway may be poorly studied ) . We first quantified across tissues the ubiquity of each pathway for each thresholding method , resulting in a predicted ubiquity matrix . Pathways grouped together into 5 clusters ( Fig G4 in S1 Text ) . We found that 2 of our 3 ubiquitous pathways matched the pathway cluster P5 , 2 out of our 6 group-specific pathways were found in clusters P2 & P3 , and 8 of the 20 tissue-specific pathways matched to cluster P4 ( Fig G4 in S1 Text ) . The coverage of pathway types in different pathway clusters suggests that different thresholding methods can be distinguished in their ability to enrich certain types of pathways based on where they are localized across tissues ( tissue-specific , group-specific , or ubiquitous ) . Therefore , we tested the ability of thresholding methods to accurately predict presence of a pathway in the right tissue ( hypergeometric test , enriched if p < 0 . 05 ) . For this , we compared the false negative rates for each of the thresholding methods ( global50 , global75 , local25 , local25-90 , and local25-75 ) in the 5 clusters . We found that global75 generated the most false negatives among pathways in P2 , global50 and local25 generated highest false negatives in P3 , and local25 and local25-90 generated the most false negatives in P5 ( Figs . H and I in S1 Text ) . Interestingly , difference in the number of false negative pathways in P4 were never significant between the different methods . However , nearly 40% of tissue-specific pathways existed in cluster P4 , likely because nearly all methods perform equally in making accurate predictions about tissue-specificity of pathways ( i . e . , pathways with very low ubiquity ) . The results , here , indicate that nearly all methods perform equally well in predicting highly tissue-specific pathways such as heme synthesis ( Fig 5A ) and bile acid synthesis ( Fig 5B ) . However , they perform differently when considering group-specific and/or ubiquitous pathways such that global50 and global75 captured fewer accurate tissues for androgen/estrogen metabolism ( Fig 5B ) and xenobiotic metabolism; and local25 and local25-90 did not capture ubiquity of glycolysis/gluconeogenesis and citric acid cycle ( see Methods; Fig 5C ) . Further , we also found that local25-75 never produced significantly ( Fig I in S1 Text ) less false negatives in pathway clusters P2 , P3 , and P5 ( Fig H in S1 Text ) . Therefore , these results together suggest that local25-75 presents most accurate list of active reactions . Several methods have been developed to integrate transcriptomic data in GEMs , thus enabling the comprehensive study of metabolism for different cell types , tissue types , patients , or environmental conditions [8 , 12 , 22 , 23 , 29 , 30] . However , while these , and many other studies rely on preprocessing decisions to integrate the transcriptomic data in biochemical networks , each study makes different decisions without reporting the reason for their approach . Indeed , no rigorous comparison of the impacts of such decisions has been previously reported clearly in the literature . Here , we highlighted how different preprocessing decisions might influence information extracted from tissue specific gene expression data . We evaluated the influence of each preprocessing decision quantitatively by studying the active reaction sets and qualitatively by evaluating tissue grouping at an organ-system level . Our analysis suggested that thresholding related decisions have the strongest influence over the set of active pathways , and more specifically the thresholding approach ( i . e . , global or local; Fig 1C ) . This can be explained by the considerable influence of the decision on thresholding on the number of genes selected as expressed ( Fig J in S1 Text ) . We note that threshold value choice for global thresholding was previously found to be the dominant factor influencing cell type-specific model content when context specific extraction methods were benchmarked [18] . When using global thresholds , the number of the genes selected to be active significantly decreases with increasing threshold value . However , the use of local thresholding leads to a smaller variation in the number of genes predicted to be active ( Fig K in S1 Text ) . Furthermore , for similar state and value attribution ( e . g . , T1 25th percentile ) , the use of the global thresholding approach leads to the selection of a larger number of genes predicted to be active in all tissues than the local approach ( Fig J in S1 Text ) . Therefore , using a global threshold leads to fewer differences between tissues and a higher correlation of active reaction sets across tissues ( Fig L in S1 Text ) , thus losing improved tissue specificity of the networks seen with the local thresholding approaches ( Fig 4 ) . This may have an important impact on analyses of tissue specific metabolism . Furthermore , the use of global thresholding is likely to lead to many false-negative reactions ( i . e . , reactions predicted to be inactive but are active ) , such as housekeeping genes that might be lowly expressed since they make essential vitamins , prosthetic groups , and micronutrients that are needed in low concentrations . Interestingly , the use of the T2 state definition seems to be less dependent on threshold values attributed than the T1 state definition when using a local approach ( Fig K in S1 Text ) . Therefore , the use of a T2 state definition in combination of a local approach seems to successfully overcome the arbitrary aspect of threshold value selection and its influence on data preprocessing . The order of preprocessing steps only moderately influences the definition of active reactions sets ( Fig 1C ) . This decision implies two different interpretations of the influence of the RNA transcript levels on the determination of the enzyme abundance and activity associated to a given reaction . Indeed , the Order 1 suggests that the measured expression levels determine the enzyme abundance available for a reaction while its associated activity will be defined depending on the gene chosen as the main determinant of the reaction behavior . On the other hand , the Order 2 relies on a comparison of the activities of each gene associated with enzymes that might catalyze a reaction without directly accounting for the absolute transcript abundance . Our analyses suggest that Order 2 provides more significant grouping for the Gastrointestinal and Lymphoreticular systems and does not considerably influence the grouping of the Female reproductive system . Advances in fluxomic measurement techniques will be invaluable to further investigate this preprocessing decision . Indeed , this would allow the analysis of the correlation between the RNA transcript levels and gene activity ( expression data transformed using thresholding ) of all the genes contributing to the definition of a reaction activity . Furthermore , this correlation analysis will further help with biological interpretation of this preprocessing decision and further refine guidelines for gene mapping decisions . In our analysis , both gene mapping methods handle the AND relationships within a GPR rule in the same way but they differ in the treatment of OR relationships by either considering the maximum expression value ( GM1 ) or a sum of expression values ( GM2 ) . Therefore , GM1 assumes that a reaction activity is determined by only one enzyme while GM2 accounts for the activity of all potential isoenzymes for a reaction . Surprisingly , while most of the reactions in Recon 2 . 2 are associated with at least two isoenzymes ( Fig M1 in S1 Text ) , the distributions of these reaction activities do not significantly change between the gene mapping approaches ( Fig N in S1 Text ) . Indeed , even if there is a significant difference in the number of genes mapped to the model depending on the techniques used: an average of 58 . 3% of the genes present in the model and available in the HPA dataset are mapped to the model reactions using GM1 while 89 . 5% are mapped using GM2 . The expression value of genes that are unmapped using GM1 but mapped with GM2 is often below the 50th percentile of the overall transcriptomic data available ( Fig O in S1 Text ) and therefore seems to not significantly influence the distribution of the reaction activities obtained . This is why the decisions relating to the gene mapping method do not influence the set of active reactions in the case of the transcriptomic dataset used in this study . However , it may not be the case for all transcriptomic datasets , especially if more metabolic genes are associated to high gene expression values . In this context , the development of more biologically meaningful gene mapping methods might be the key to capture differences between cell-types or tissues . Current gene mapping methods consider all enzymes as specialists ( i . e . , one enzyme is associated to one reaction ) . However , numerous enzymes are actually “generalists” as they exhibit promiscuity [31 , 32] ( Fig M4 in S1 Text ) . This functional promiscuity of an enzyme may be manifested in the form of competition between reactions catalyzed by this enzyme , and therefore influence the catalytic activity of an enzyme . In this context , future work may benefit from exploring strategies to handle enzyme promiscuity [33] . In conclusion , decisions must be made on how to best handle and incorporate transcriptomic data into biochemical networks . This benchmarking study emphasizes for the first time the importance of carefully evaluating these decisions and associated parameters . Our analysis highlights that the choice of thresholding approach influences the active reaction sets the most , even more than tissue-specific effects . Meanwhile , gene mapping decisions had the lowest influence . We showed that some decisions better capture the functional tissue similarity across different organ systems . Overall , our analysis showed that transcriptomic data preprocessing decisions influence the ability to capture meaningful information about tissues . However , current preprocessing techniques present important limitations and decisions associated to this process should be made very carefully . Indeed , numerous steps and decisions involved in the estimation of enzyme abundance and activity from transcriptomic data rely on biological assumptions that have not yet been leveraged . With the increasing availability and affordability of omic measurement techniques , studies filling the gap between mRNA expression and enzymatic activity will be of crucial importance . In this context , we hope that the guidelines provided by this study will help researchers develop more robust and biologically meaningful preprocessing techniques , leading to more accurate models , and deeper insights into the tissue-specific behavior of an animal . We used the Human Protein Atlas transcriptomic dataset ( HPA ) which includes RNA-Seq data of 20344 genes across 32 different human tissues [25] . Out of 20344 genes , 1663 can be mapped to the metabolic genes present in Recon 2 . 2 ( 99 . 4% of coverage ) [26] . S3 Table presents the 10 genes of Recon 2 . 2 that are not associated with expression values in the HPA dataset and Fig P in S1 Text presents the distribution of gene expression values in the HPA dataset . Recon 2 . 2 [26] includes 1673 genes , 5324 metabolites and 7785 reactions . 3061 reactions do not have GPR associations . The remaining 4724 reactions are associated to 1797 different enzymes and about 20% of these reactions can be catalyzed by multiple isoenzymes . Almost 21% of the enzymes are formed by enzyme complexes ( up to 46 subunits—reaction: NADH2_u10m ) and about 54% of the enzymes are promiscuous enzymes ( Fig M in S1 Text ) . In metabolic networks , the relationship between genes and reactions is represented using logical rules , referred as Gene-Protein-Reaction rules ( GPRs ) . These rules describe the association between the genes responsible for the expression of protein subunits forming the enzyme that catalyzes a reaction ( AND for enzyme complexes; OR for isoenzymes ) . This relationship linking enzymes to reactions may have different types of GPR patterns . Some relationships are simple , with one gene encoding one enzyme that catalyzes one reaction . However , many are more complicated , in which one enzyme catalyzes multiple reactions ( promiscuous ) , multiple proteins form an enzyme complex that catalyzes one reaction ( multimeric ) , multiple enzymes catalyze one reaction ( isoenzymatic ) , or multiple enzymes could catalyze multiple reactions ( isoenzymatic promiscuous ) [32] ( Fig A in S1 Text ) . Gene mapping methods ( GMMs ) require combined use of the GPR rule and gene expression data to determine the enzyme activity associated to a reaction . In this regard , two methods have been used prominently in the field: Thresholding Approaches: Thresholding approaches describe the scheme of threshold imposition on the gene expression value for a gene and/or reaction to be considered as “active” . The definition of thresholding criteria requires one to decide on how to partition the gene expression or reaction activity . In this regard , the ON/OFF state definition is often used in the literature . This type of state definition requires only one value to qualify if a gene/reaction is active . For example , when using one single threshold in a global context ( i . e . , hereafter referred as global T1 ) , the genes presenting an expression above this value are considered as active ( i . e . , ON ) while the others are inactive ( i . e . , OFF ) . However , this type of gene expression partition in a local context ( e . g . , expression threshold of a gene defined by its mean expression across all samples ) presents limitations when facing genes with very low or very high expression values for all the samples . Indeed , when a gene presents always very low expression values , the use of the mean as threshold will lead to the consideration of its expression in some samples . Contrarily , some genes may be associated with very high expression values in all the samples . Doing so , while this gene should be considered as active , the current state partition will lead to considering this gene as non-expressed in all the samples presenting an expression value below the mean . To overcome this problem , we propose to implement the gene-specific thresholding approach in combination with the definition of a global threshold definition for genes presenting low expression values among all the samples ( e . g . , below the usual detection level associated to the measurement method ) to prevent their definition as an active set for some samples ( i . e . , local T1 ) . Therefore , the genes whose expression is below the value defined by this global lower bound will always be considered as inactive ( i . e . , OFF ) , while other genes will fall under the local rule for gene-specific threshold definition ( i . e . , MAYBE ON ) . The T1 state definition of local thresholding approach can be defined as follows “the expression threshold for a gene is determined by the mean of expression values observed for that gene among all the tissues BUT the threshold must be higher or equal to a lower percentile bound globally defined” . Another similar extreme case can be encountered for genes with high expression values in all samples . To this end , we propose to introduce an upper and a lower bound can be introduced to define the expression values for which a gene should always be considered as expressed or non-expressed . This will ensure that genes with very low expression values across all the samples will never be considered as active ( i . e . , OFF ) and genes with very high expression across samples are always considered as active ( i . e . , ON ) . Doing so , the local rule for gene-specific threshold definition is applied only to the genes whose expression is in between the range of values defined by the lower and upper bounds ( i . e . , MAYBE ON ) . The definition of the local threshold with a T2 state definition can be expressed as follows: “the expression threshold for a gene is determined by the mean of expression values observed for that gene among all the tissues BUT the threshold: ( i ) must be higher or equal to a lower percentile bound globally defined and ( ii ) must be lower or equal to an upper percentile bound globally defined . ” Threshold values: The threshold values depend on the approach ( i . e . , local or global ) and on the number of states ( i . e . , T1 or T2 ) used for thresholding . The global approach can only be associated with T1 state definition as it requires the assignment of only one threshold value . On the other hand , the local thresholding approach can be used in combination with either a T1 or a T2 state definition , as mentioned above . In the context of this study , we have chosen to compare the following combination of threshold value attribution: The preprocessing for transcriptomic data could be done in two possible ways as described below: For Order 1 , thresholding is imposed on these “reaction expressions” and no longer on the gene expression . This leads to the necessity to adapt the local threshold definition in the case of a preprocessing combination using Order 1 with GM2 gene mapping . Indeed , as the GM2 approach map multiple genes to a reaction , the activity of this reaction can no longer be defined by using the gene expression distribution . Therefore , the activity threshold for a reaction is determined by the sum of mean expression values observed for the genes mapped to this reactions ) among all the tissues , but the mean expression value of each gene mapped to the reaction must be higher or equal to a lower percentile bound globally defined . Furthermore , it must be lower or equal to upper percentile bound globally defined . A binary matrix is constructed in which each row represents one of the 20 preprocessing approaches for each tissue ( i . e . , total of 640 rows ) and each column represents a reaction of the GEM ( i . e . , 7785 columns ) : a reaction being active ( 1 ) or not active ( 0 ) in the GEM . The PCA analysis was conducted on this matrix after the removal of reactions being active for all or no preprocessing combinations and having each row centered to have zero mean . The variance explained by the different factors ( each preprocessing decision and the tissue origin ) within each of the principal components is calculated as follows . Within one factor , the maximum Pearson correlation coefficient ( R ) of the component scores and categories is calculated across all possible orderings of the categories . Reported is the R2 scaled to percentages . The set of active reactions have been used to compute the Euclidean distance between each tissue . We associated each tissue to an organ system using the classification proposed in the Human Protein Atlas ( S1 Table ) and computed the average Euclidean distance between tissues belonging to the same organ system . Note that , we only considered organ systems presenting more than two tissues within the same group ( i . e . , Female Reproductive , Lymphatic and Gastrointestinal ) . To compute the significance of our results , we generated the mean Euclidean distance for 10000 randomly selected group with the same number of tissues and computed the exact p-value ( i . e . , proportion of random distance lower than the observed distance ) associated to each organ system . Pathway-tissue pairs were manually searched using the tissue and pathway names . The search results directed to published articles were then manually selected . The articles that specifically studied the presence of enzymes in the tissues ( extracted post-mortem ) or culture of human cells were chosen as evidence . We also considered review articles that discussed pathways , their ubiquity among human cell types , or role of a pathway in the tissue ( S2 Table ) . To evaluate the biological differences in core reaction lists obtained from different thresholding methods , we tested if , for a thresholding method , a given pathway was enriched ( hypergeometric p-value < 0 . 05 ) in a given tissue based on the number of reactions associated to the pathway ( K ) , the total number of reactions selected ( N ) , the number of reactions selected that associated to the pathway in that tissue ( x ) , and the total number of reactions associated to a pathway . Hypergeometric p-value was calculated using the MATLAB function , hygecdf . For a given thresholding method , our analysis resulted in enrichment matrix ( a binary matrix ) , E . The value of Eij was assigned as 1 if ith pathway was enriched in jth tissue , otherwise the value was assigned as 0 ( Fig G3 in S1 Text ) . The enrichment of a pathway in the active reaction set of a tissue was interpreted as a strong statistical support that the pathway is predicted to be active in the tissue by the thresholding method . We compared this matrix to the known pathway-tissue pairs ( Fig G5 in S1 Text ) for analysis of false negatives in each thresholding method ( Fig G6 in S1 Text ) . This matrix was then converted into a vector by summing along the tissue dimension; thus , indicating the number of tissues where a given pathway was enriched . This vector was made for each thresholding method; thus , we ended up with a matrix describing number of tissues a pathway is predicted to occur when a thresholding method is used . Here , we call this matrix predicted ubiquity matrix , UP . The value of UPik indicates the number of tissues ith pathway was enriched in by kth thresholding method . The predicted ubiquity matrix was then clustered to identify similarity in ubiquity predictions of thresholding methods among different pathways ( Fig G4 in S1 Text ) . We used complete linkage and Euclidean distance metric for hierarchical clustering . The pathways were then divided into 5 clusters . The Human Protein Atlas transcriptomic dataset ( HPA ) were acquired from supplementary material of [25] . Recon 2 . 2 model was downloaded from https://github . com/u003f/recon2/tree/master/models . The MATLAB code for applying the different preprocessing combinations is available in COBRA toolbox v3 . 0 , in the papers section [34] .
Over the past two decades , systems biology approaches have permeated all fields of biology . Indeed , biological networks are commonly used for the analysis of omics datasets and to test hypotheses through computational simulations . Many of these types of studies require one to overlay omics data to a biological network . In many biochemical networks , such as metabolic networks , there is not a one-to-one relationship between genes and network edges ( e . g . , metabolic reactions ) , due to the existence of isozymes and protein complexes . Therefore , decisions must be made on how to overlay data onto networks . For example , these decisions include ( 1 ) how to integrate gene expression levels using the Boolean relationships between genes , proteins , enzymes , and reactions , ( 2 ) the selection of thresholds on expression data to consider the associated gene as “active” or “inactive” , and ( 3 ) the order in which these steps are imposed . While these decisions have been made in many published studies , there has been no systematic evaluation of the impact of these decisions on the biological accuracy of the resulting networks . To this end , we benchmarked the different existing decisions made to integrate the data into the network . The results provide guidelines to improve data analyses with biochemical networks that should translate to many other network types .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[ "genetic", "networks", "protein", "metabolism", "engineering", "and", "technology", "enzymes", "metabolic", "networks", "enzymology", "network", "analysis", "genome", "analysis", "molecular", "biology", "techniques", "research", "and", "analysis", "methods", "computer", "and", "information", "sciences", "genomics", "gene", "mapping", "proteins", "gene", "expression", "molecular", "biology", "biochemistry", "gene", "identification", "and", "analysis", "software", "engineering", "genetics", "biology", "and", "life", "sciences", "transcriptome", "analysis", "preprocessing", "computational", "biology", "metabolism" ]
2019
Assessing key decisions for transcriptomic data integration in biochemical networks
Measureable rates of genome evolution are well documented in human pathogens but are less well understood in bacterial pathogens in the wild , particularly during and after host switches . Mycoplasma gallisepticum ( MG ) is a pathogenic bacterium that has evolved predominantly in poultry and recently jumped to wild house finches ( Carpodacus mexicanus ) , a common North American songbird . For the first time we characterize the genome and measure rates of genome evolution in House Finch isolates of MG , as well as in poultry outgroups . Using whole-genome sequences of 12 House Finch isolates across a 13-year serial sample and an additional four newly sequenced poultry strains , we estimate a nucleotide diversity in House Finch isolates of only ∼2% of ancestral poultry strains and a nucleotide substitution rate of 0 . 8−1 . 2×10−5 per site per year both in poultry and in House Finches , an exceptionally fast rate rivaling some of the highest estimates reported thus far for bacteria . We also found high diversity and complete turnover of CRISPR arrays in poultry MG strains prior to the switch to the House Finch host , but after the invasion of House Finches there is progressive loss of CRISPR repeat diversity , and recruitment of novel CRISPR repeats ceases . Recent ( 2007 ) House Finch MG strains retain only ∼50% of the CRISPR repertoire founding ( 1994–95 ) strains and have lost the CRISPR–associated genes required for CRISPR function . Our results suggest that genome evolution in bacterial pathogens of wild birds can be extremely rapid and in this case is accompanied by apparent functional loss of CRISPRs . Populations of animals are under constant threat from bacterial pathogens , which can be particularly destructive following a switch to a new host or the evolution of novel virulence mechanisms . Understanding the rate and process of evolutionary change in pathogens is thus important to assessing the risks of pandemics and developing means to predict and avoid such catastrophic events . In 1994 , a strain of Mycoplasma gallisepticum ( MG ) was identified as the causative agent of an emerging epizootic in House Finches , a wild songbird inhabiting Eastern North America [1] . This bacterial pathogen frequently causes disease in commercial chicken and turkey flocks , but it had never been reported in House Finches or any songbird , leading to the suggestion that the epidemic began when MG expanded its host range from poultry to this phylogenetically distant songbird . MG prevalence reached 60% in some areas , and killed an estimated 225 million finches in the first three years after detection [2] . The early detection of the epizootic allowed research and citizen-science teams to track its rapid spread throughout eastern North America in exceptional detail , making it one of the best documented wildlife pathogen outbreaks [3]–[7] . Although previous genome-wide studies have clarified rates of measurable evolution in viral pathogens [8] , [9] and in bacterial populations evolving under laboratory conditions or as human pathogens [10]–[18] , less is known about rates of genetic change in bacterial pathogens of non-mammalian vertebrates , particularly on short evolutionary time scales . Genome-wide and gene-specific estimates of point substitution in bacterial lineages measured over centuries [19] to millions of years [20] suggest maximum substitution rates on the order of 10−7 to 10−9 per site per year . Although recent work suggests the rate may be even faster for several bacterial species [12] , [14] , [19] , the number of studies documenting whole-genome changes in bacteria during host switches is still small , particularly for wildlife pathogens [21] , [22] . As part of ongoing surveillance , field isolates of MG obtained from infected finches were sampled at multiple time points from the start of the epidemic in 1994 to 2007 , providing a genetic time series beginning immediately after the host switch , as well as an opportunity to directly measure the tempo and mode of evolution in a natural bacterial population whose genome is as yet uncharacterized . To characterize patterns of genomic change during its host switch between distantly related avian species , we sequenced whole genomes of 12 House Finch MG isolates from this 13-year time series , with four samples each from the beginning ( 1994–1996 ) , middle ( 2001 ) and recent ( 2007 ) periods ( Table S1 ) . In addition , to identify putative source strains as well to determine if differences between the House Finch MG strains and the ∼1 Mb published reference Rlow strain from chicken [23] were ancestral or derived , we sequenced four additional strains from chicken and turkey based on phylogenetic analysis of a smaller multistrain data set ( Figure S1 ) . Our sequence , SNP filtering and between-platform cross-validation protocols yielded a high quality 756 , 552 bp alignment encompassing 612 genes ( Tables S2 , S3 , S4 , Text S1 , Figure S2 ) , and allowed us to monitor point substitutions , genomic indels , IS element insertions , and other changes across the entire genome ( Figure 1 ) , including the entire array of clustered regularly interspaced short palindromic repeats ( CRISPR ) of all 17 strains ( finch and poultry isolates ) . All House Finch MG samples were collected in the southeastern U . S . ( Table S1 ) , with an emphasis on the well studied population in Alabama [24] , [25] . The population structure of Eastern House Finches before the epizootic was virtually panmictic [26] , suggesting that there is likely to be little geographic structuring of MG in the east , a hypothesis that could be tested with additional data . The 12 House Finch strains from the three time periods spanned the known temporal and phylogenetic diversity of this lineage , and included strains that have been used to study host response to pathogen infection in House Finches [27] . To determine genetic diversity and phylogenetic identity of putative source populations of the House Finch MG strains , and to aid in sampling chicken and turkey strains for sequencing , we first analyzed a previously published data set [28] . Phylogenetic analysis of 1 , 363 bp obtained from four genomic regions for a large sample ( n = 82 ) of MG strains suggests that turkeys rather than chickens were the source of House Finch MG and that the MG lineage colonizing House Finches first passed multiple times among chickens and turkeys ( Figure S2 ) . Although this analysis suggests frequent host switches between chickens and turkeys , which diverged 28–40 MYA [29] , [30] , it also suggests a single switch to the House Finch , a songbird species diverged from chickens by ∼80 MYA [31] . The whole genome alignment contained strong signals of a founder event as a result of colonization of House Finches . The total nucleotide diversity ( π ) in the House Finch strains for the four-gene region was only 3 . 1% of the diversity in circulating poultry strains prior to the epizootic , and only 2 . 3% of the poultry diversity when considering the entire House Finch MG genome [28] ( Figure 2 and Table S5 ) . In agreement with the four-gene analysis , our whole genome sequencing showed that the four sequenced poultry isolates were much more genetically diverse than the 12 House Finch isolates , possessing a total of 13 , 175 SNPs as compared to only 412 SNPs among the House Finch isolates ( Table S2 ) . The House Finch MG diversity corresponds to π = 0 . 00014 , or roughly 1 SNP every 1 , 800 bp . Consistent with purifying selection acting over the longer time period encompassing the divergence of House Finch and poultry MG strains ( as opposed to acting after the host-switch among House Finch strains alone ) , there was a stronger bias against non-synonymous substitutions among the more diverged poultry strains than among the recently diverged House Finch MG strains ( Table S6 ) . Across the entire genome , only 147 ( 35% ) of the SNPs among the House Finch isolates were phylogenetically informative; the majority ( 265 or 64% ) appeared as singletons . To further quantify House Finch MG demography , we used a statistical model , the Bayesian skyline plot implemented with BEAST , that utilizes information on dates of sampling to estimate changes in genetic diversity through time [32] , [33] ( Text S2 ) . The analysis is broadly consistent with field observations suggesting a mid-1990s origin followed by rapid population expansion , though it estimates that the House Finch MG lineages coalesced roughly in 1988 , several years prior to the observation of sick birds in the field ( estimated MRCA of the House Finch MG strains is 19 . 2 years prior to 2007 [95% HPD 16 . 9 – 21 . 7]; Figure 2d ) . Discrepancies between coalescence times and observed outbreaks in host populations have been observed for other pathogens , and could possibly be due to selective or demographic effects , or in our case low sample size [12] . Phylogenetic analysis suggests substantial turnover in the standing SNP variation between sampling intervals , with strong clustering of the 2007 strains , which are distinguished from other House Finch strains by 85 diagnostic SNPs ( Figure 3 ) . We found that one of the sequenced turkey strains , TK_2001 , was highly similar in sequence to the House Finch strains and shares a number of genomic deletions and transposon insertions as well as duplications and losses of CRISPR spacers ( see below ) with the House Finch MG strains . This turkey strain may represent a poultry lineage close to the source lineage for House Finch MG ( Figure 3 ) . In addition to SNPs in House Finch MG we found five large genomic deletions that occurred by 2007 and amounted to ∼42 , 245 bp and encompassing 34 genes relative to the chicken Rlow strain ( Figure 1 and Figure 3 , Table S7 ) . Three of these deletions are phylogenetically informative among the 17 MG strains ( Table S7 ) , but their conflicting phylogenetic distribution underscores the presence of recombination ( see next section ) . Two deletions totaling 9 , 275 bp were shared among all strains except the reference . In addition , we detected six novel IS element insertions in the House Finch MG lineage ( Text S3 , Table S8 ) and three of the genomic deletions were likely mediated by illegitimate recombination between flanking IS elements ( Table S7 ) . In addition to the 34 genes deleted as part of genomic deletions , we found evidence for pseudogenization of 19 genes relative to the chicken MG reference ( Text S3 , Table S9 ) . Two genes appear to have been disrupted by transposon insertions and 17 genes were pseudogenized by frameshift or nonsense mutations ( Table S9 ) . The substantial gene losses we detected , a total of 52 genes ( ∼8 . 6% ) fixed in the House Finch MG lineage , presumably as a result of the bottleneck during host switch . By contrast , we failed to find a single novel gene in House Finch MG that was not also found in the poultry MG strains ( Text S5 ) . Comparative analysis with other Mycoplasma genomes showed that 15% of these lost genes also lacked a homologue in the other genomes surveyed whereas 13% had a homologue in every genome ( Table S9 ) . Despite the small amount of genetic variation segregating among our House Finch Mycoplasma samples ( only 412 SNPs ) , it is not possible to construct a phylogenetic tree for these strains that is free of homoplasies . Although the four 2007 strains and all 2001 strains except AL_2001_17 clearly formed well defined clades based on 85 and 28 SNPs , respectively , establishing the phylogenetic relationships for the other 5 House Finch MG strains exclusively via SNPs was not possible ( Text S6 , Figure 3 ) . Although a total of 16 SNPs were phylogenetically informative for the placement of these five strains , the largest cluster of SNPs that were phylogenetically consistent was seven , and overall , 13 different trees were supported by at least 3 SNPs each . Similarly , substantial homoplasy was found among the four newly sequenced poultry strains and the Rlow reference . Although 6 , 152 SNPs were parsimony informative for these five strains , the unrooted tree with the best support was in conflict with 4 , 619 ( 75% ) of these SNPs . These patterns are expected if sites are being shuffled by recombination or horizontal gene transfer ( HGT ) among isolates , and analysis of the entire data set found strong support for this ( Text S4 , Figures S3 , S4 , S5 ) . Using the pairwise homoplasy index test [34] revealed a statistically significant signal of recombination ( p<10−9 ) . This signal comes predominantly from the four newly sequenced poultry strains because there is not enough genetic variation to make this test significant when only the House Finch strains are considered . However if we apply to the House Finch MG strains the homoplasy test by Maynard-Smith and Smith [35] , which is found to perform well in situations of low nucleotide diversity [36] , we again obtain a significant signal for recombination ( p<10−6 ) . We conclude that , despite a significant signal for recombination in both the poultry and House Finch strains , the House Finch MG cluster as a whole is a distinct and easily identifiable phylogenetic lineage with a long branch separating it from the poultry strains ( Figure 3 ) . Coalescent analysis [32] of the 12 House Finch isolates sampled at different dates suggested an extraordinary point substitution rate of 1 . 02×10−5 substitutions per site per year ( 95% HPD 7 . 95×10−6 to 1 . 23× 10−5 ( Text S2 ) , consistent with earlier suggestions that Mycoplasma may be among the fastest evolving bacteria [37] . This rate of point substitution is not restricted to House Finch MG strains but was also found in the poultry strains when analyzed separately ( Text S2 ) , suggesting that rapid evolution was characteristic of MG prior to the House Finch epizootic . We estimated a similar substitution rate when considering only the four-gene multistrain alignment use to identify poultry strains for sequencing ( Text S2 ) . We verified that our estimate of substitution rate is robust to different protocols for SNP identification , statistical models and data sets ( Figure 4; Text S7 ) . Altogether we estimated the substitution rate within a coalescent framework on 34 combinations of SNP calling and model assumptions and found consistent estimates throughout ( Text S1 , Figure 4 , Figure S6 ) . In addition , we achieved a similar estimate using a Poisson regression approach as well as a root-to-tip regression ( Text S7 and Figure 4 ) . In addition to a high estimated substitution rate in MG , we found a mutation in the gene-encoding UvrB that could elevate this rate yet further . UvrB is an essential part of the nucleotide excision repair system , which has been posited to be the most important pathway for maintaining genomic integrity in Mycoplasma [38] . The mutation truncates the UvrB protein by three amino acids ( Table S10 ) and raises the possibility of the origin of a mutator strain in House Finch MG [39] as the C-terminal of this protein is essential for its function [40] . Consistent with this idea , we found 14 instances of adjacent SNPs among the 12 House Finch isolates , a notable excess in an alignment with only 412 variable sites ( Table S11 ) . Moreover , 12 of these 14 are CC→TT double substitutions , which are normally repaired by the UVR system ( Table S10 ) . For 13 of the 14 doublets , both sites are inferred to have mutated on the same branch of the tree , suggesting single mutational events , and the proportion of doublet mutations involving the same base was drastically higher ( 92 . 8% ) in lineages with the UvrB mutation as compared to those without ( p<0 . 0001; Table S10 ) . Nonetheless , these doublet mutations are not required to achieve the high rate of substitution that we measured . They account for less than 7% of the segregating variation and removal of these doublet sites does not affect the high estimated substitution rate . The UvrB mutation is found in all of our House Finch MG strains as well as the turkey strain TK_2001 , but not in the ancestral chicken strains or the reference chicken strain . Thus , the mutation appears to have arisen on the lineage leading to the House Finch . In some bacterial systems , CRISPRs have a well-recognized function in bacterial immunity and defense against phage , although they may possess additional functions , such as gene regulation [41]–[44] . We extensively catalogued CRISPR repeats in the House Finch and ancestral poultry strains ( Figure 5 , Text S8 , Table S12 ) . In so doing we observed drastic changes in the CRISPR system between House Finch and poultry strains ( Figure 5 ) [45]–[48] . The House Finch MG strains from 1994–96 contain up to 50 unique spacers , none of which is shared with the four divergent poultry genomes , which each contained a unique set of 36 to 147 spacer regions consistent with a high rate of turnover for a population actively acquiring new spacer sequences . We found that less than 1% of the 302 unique spacer sequences had similarity to any sequences in the House Finch MG genomes and that none of the remaining spacers had any similarity to sequences in Genbank , indicating an external source for these sequences ( Text S8 ) . Surprisingly , no novel spacer elements are present in any of the House Finch MG samples or TK_2001 , indicating that the CRISPR array ceased recruiting additional spacers around the time of host switch into the House Finch . In fact , over the 13-year period of the epizootic , the number of unique spacers present in the CRISPR array of the samples decreased to 28 ( Figure 5 ) . Further evidence for degradation of the CRISPR locus following the host switch is the complete loss of the four CRISPR-associated ( i . e . “CAS” ) genes in all of the 2007 isolates , a loss that likely renders the CRISPR system in House Finch MG non-functional [45] . We conducted whole-genome sequencing on a unique 13-year serial sample of Mycoplasma strains circulating in wild House Finches to characterize genomic changes accompanying a host shift from poultry in the mid-1990s as well as to obtain a very high substitution rate for this avian pathogen . Previous estimates using serial samples and/or the known timing of events presumably tied to the divergence of bacterial strains have generally found much lower rates . An estimate of 2 . 0×10−6 was obtained for Staphyloccous aureus [12] , 1 . 1×10−7 for Buchnera [19] , 7 . 42×10−7 in Yersinia pestis and 1 . 4×10−6 in Heliobacter pylori [14] . Disentangling the effects of recombination and point substitution can be challenging and some previously published substitution rates are likely to be upper bounds rather than point estimates [12] . Our estimate appears to be among the highest reported for a bacterium , and is consistent with other reports of exceptionally high substitution rates in mycoplasmas [37] . Estimates of substitution rates can be influenced by the interval over which sequences are sampled , with estimates taken from short time intervals often exceeding those taken on biogeographic or geological time scales [49] . However the small number of SNPs that we detected segregating in House Finch MG populations suggest negligible effects of multiple hits on our estimate , and our use of a coalescent model suggests that effects of ancestral polymorphism on substitution rate estimates should be adequately accounted for [32] , [50] . Additionally , our estimates of substitution rate were robust to many potential complicating factors , including SNP calling protocol and whether poultry or House Finches were used as the host for sampled sequences . Given the history and genetic isolation of the House Finch MG strains , the influence of recombination or lateral gene transfer on our estimate of substitution rate is likely also minimized ( Text S7 ) . The CRISPR dynamics we observed in House Finch MG differ from that seen in other pathogen and bacterial populations . A recent study of Y . pestis CRISPR arrays from 131 strains [51] indicated a slower pace of CRISPR evolution than observed in MG and pattern of evolution in which acquisition of novel sequences does not play a prominent role . This study found that in Y . pestis the first part of the CRISPR arrays were conserved and that over 76% of all spacer sequences derived from within the Y . pestis genome . Similarly , a recent study of E . coli and Salmonella genomes found that strains within 0 . 02% divergence typically have identical CRISPR loci [52] and that spacer sequences were often matched to elements of the E . coli genome . Additionally , some spacer sequences were shared between strains within a species exhibiting over 1% sequence divergence . These observations and an estimated substitution rate on the order of 10−10 per site per year suggested that E . coli strains that had diverged for 1 , 000 years sometimes shared identical CRISPR loci , suggesting patterns of evolution different from that expected for a rapidly changing adaptive immune system primed to combat phages , a conclusion that was supported by later work [53] . By contrast to the pattern seen in these γ-proteobacteria , none of the House Finch MG strains in this study have the same CRISPR locus despite differing at only 0 . 01–0 . 02% of sites and likely having last shared a common ancestor less than 20 years ago . Our serial sampling suggests that the loss of spacer sequences and the CRISPR system itself can take place on very short time scales in Mycoplasma . Unlike the patterns seen in E . coli , Y . pestis , and Salmonella , the poultry MG strains in our study did not share any spacer sequences , even though they differed by ∼1% . These strains had very large CRISPR arrays and 99% of all spacer sequences did not match any known sequence in their genome or in the databases . Therefore the MG CRISPR loci studied here differ from the those observed in some γ-proteobacteria , a group for which CRISPR dynamics can appear functionally unrelated to ecology or immunity [53]–[55] . Instead , our finding of rapid evolution and degradation of the CRISPR loci more closely resembles patterns found in other bacterial groups , particularly those in which CRISPR is involved in phage defense [56] . CRISPRs are found in only 40% of sequenced bacteria investigated thus far , and often have major roles in bacterial immunity in several lineages investigated in detail [45] . We were surprised to find a gradual degradation and ultimate apparent functional loss of the CRISPR system in House Finch MG after the host switch and a shift in CRISPR dynamics appears to be a major correlate of host switch in this system . One possible explanation for this pattern is that MG experienced release from its ancestral phage parasite community ( or other mobile genetic elements such as plasmids ) following introduction into the House Finch . Loss of traits upon removal of the agent of selection is a common evolutionary response , as are population expansions of animals and plants when introduced into novel habitats unaccompanied by their parasites [57] . Despite the large amount of ecological research focusing on this host-pathogen system [3]–[7] , at present nothing is known about phages that infect MG or their role in its evolutionary dynamics . Therefore the hypothesis of parasite release as a driver of CRISPR loss is purely speculative . We know of no phage known to infect the Pneumoniae phylogenetic group of mycoplasmas and the few phages known to infect Mycoplasma have proven difficult to characterize [58] . We might expect Mycoplasma bacteriophages to be host-specific given that they seem to be unusual in their ability to bind to a bacterium with no cell wall and a diverse assortment of surface proteins [58] . However , we are not aware of even basic data on the degree to which Mycoplasma might be susceptible to the many bacteriophages that they presumably encounter in their environment . Although phage represent one possible source for these novel ∼30 bp sequences , another possible explanation for the source of the spacer sequences is that they derive from plasmids . Although unprecedented ( we know of no examples of a naturally occurring plasmid in the Pneumoniae mycoplasmas ) , such a scenario could raise the possibility of easier genetic manipulations in MG where development of such tools has been challenging [59] . Of the many other possibilities that could explain the observed degradation of the CRISPR loci , we can at least rule out self-interference as an explanation in derived MG strains , given that there is only a single CRISPR cluster in House Finch MG [54] . Measurement of costs , possible advantages and consequences of CRISPR loss , as well as functional and evolutionary assays and surveys of phage diversity will help determine if the rapid and deadly spread of Mycoplasma following their expansion into the House Finch was facilitated by a lack of phage predation , a short-term advantage of CRISPR degradation or some other , possibly neutral , mechanism . Although our sequence data is suggestive , explicit functional studies will also be required to demonstrate CRISPR functionality or lack thereof in poultry and House Finch MG and its role , if any , in phage defense . Genome evolution of MG during its host-switch from poultry to House Finches adds to a growing list of host-switches that are successful in the complete absence of novel genes [21] , [60] , [61] and bacterial lineages exhibiting high rates of point substitution [14] . Mycoplasmas are some of the fastest evolving organisms on earth [62] having lost many of the repair mechanisms present in other bacteria [38] and this high mutation rate could help introduce deleterious mutations and contribute to the substantial level of pseudogenization that was observed in this study . The high basal substitution rate in MG may well be elevated yet further by UvrB mutation that we detected , a mutation that could have consequences for the long term genomic integrity of this MG lineage , particularly if it remains genetically distinct from and unable to exchange genes with the poultry MG lineages with a functional UvrB . Alternatively , given the short ( 3 amino acid ) truncation of this gene in the House Finch strains , another explanation for the greatly increased number of doublet mutations in the lineage carrying the UvrB truncation is that selection has not had enough time to remove them as it has for poultry strains without this mutation . Although mutator strains are known to have a selective advantage in rapidly evolving laboratory and natural populations [39] , [63] , additional functional and experimental work will be required to determine the selective and functional effect of the mutation we have detected in UvrB , and over what time scales such selective effects might persist . For this and other endeavors , serial sampling of additional bacterial populations in nature will further clarify the rate at which genomes are remolded during host switches in the wild . DNA sequence data for 4 gene fragments collected from 74 strains in Ferguson et . al . [28] , was combined with data from 8 strains newly sequenced in this study to yield a Large Sample Multiple Sequence Alignment ( LS-MSA ) 1 , 363 bp in length ( Figure S2 ) . We estimated nucleotide diversity and the standard deviation of this estimate within and among subgroups of these sequences using DNAsp version 4 . 10 . 9 [64] ( Table S5 ) . In estimating diversity of MG strains sampled from chickens and turkeys , we restricted analysis to those strains sampled during 1994–1996 for comparison with our earliest House Finch strains sampled in a similar time interval . Twelve strains of MG isolated from House Finches in the Southeastern US were sequenced with the Roche 454 Gene Sequencer . The average coverage level was 9 . 4X ( Table S1 ) . Additionally , four MG strains isolated from poultry hosts and selected based on their positions in the multistrain phylogenetic tree were sequenced with the Illumina sequencing platform to an average coverage of ∼410 X ( Tables S2 , S3 , S4 , Text S1 , Figure S2 ) . Using a coalescent model and a Bayesian framework as implement in BEAST v1 . 52 [32] we estimated the mutation rate and times to common ancestry from a 13-taxon alignment composed of the reference MG genome and all of the House Finch MG strains whose genomes were sequenced in this study ( Text S2 ) . We also ensured that the conclusions from this inference were not sensitive to the SNP calling procedures or the choice of substitution models ( Text S2 , S7 , Figure S6 ) . In order to compare the mutation rate between the poultry and House Finch MG populations , these quantities were similarly estimated from the 82 taxon LS-MSA after removing nine laboratory strains from the alignment that likely experienced different population dynamics than the wild strains and had unknown sampling dates . A Poisson regression model was also used to estimate substitution rates by counting mutations along a single lineage assumed to span the dates of sampling for each strain ( Text S7 ) . We catalogued IS elements using BLAST and the ISFinder database [65 , Text S4] . We tested for evidence of genetic recombination between MG strains using the genome sequences from our 4 poultry and 2 House Finch strains using the pairwise homoplasy index test [34] as implement in splitstree4 [66] , and the homoplasy test by Maynard-Smith and Smith [35] . Further evidence for the presence of recombination and the number of nonrecombining blocks was provided by other methods ( Text S6 , Figures S3 , S4 , S5 ) .
Documenting the evolutionary changes occurring in pathogens when they switch hosts is important for understanding mechanisms of adaptation and rates of evolution . We took advantage of a novel host–pathogen system involving a bacterial pathogen ( Mycoplasma gallisepticum , or MG ) and a songbird host , the House Finch , to study genome-wide changes during a host-shift . Around 1994 , biologists noticed that House Finches were contracting conjunctivitis and MG from poultry was discovered to be the cause . The resulting epizootic was one of the best documented for a wildlife species , partly as a result of thousands of citizen science observers . We sequenced the genomes of 12 House Finch MG strains sampled throughout the epizootic , from 1994–2007 , as well as four additional putatively ancestral poultry MG strains . Using this serial sample , we estimate a remarkably high rate of substitution , consistent with past implications that mycoplasmas are among the fastest evolving bacteria . We also find that an array of likely phage-derived sequences known as CRISPRs has degraded and ceased to recruit new repeats in the House Finch MG strains , as compared to the poultry strains in which it is diverse and rapidly evolving . This suggests that phage dynamics might be important in the dynamics of MG infection .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "zoology", "biology", "microbiology", "evolutionary", "biology", "population", "biology" ]
2012
Ultrafast Evolution and Loss of CRISPRs Following a Host Shift in a Novel Wildlife Pathogen, Mycoplasma gallisepticum
The gene regulation mechanism along the life cycle of the genus Schistosoma is complex . Small non-coding RNAs ( sncRNAs ) are essential post transcriptional gene regulation elements that affect gene expression and mRNA stability . Preliminary studies indicated that sncRNAs in schistosomal parasites are generated through different pathways , which are developmentally regulated . However , the data of sncRNAs of schistosomal parasites are still fragmental and a complete expression profile of sncRNAs during the parasite development requires a deep investigation . We employed high-throughput genome-wide transcriptome analytic techniques to explore the dynamic expression of microRNAs ( miRNAs ) and endogenous siRNAs ( endo-siRNAs ) of Schistosoma japonicum covering the free-living cercarial stage and all stages in the definitive host . This led us to analyze over 70 million clean reads represented both high and low abundance of the small RNA population . Patterns of differential expression of miRNAs and endo-siRNAs were observed . MiRNAs was twice more than endo-siRNAs in cercariae , but gradually decreased along with the development of the parasite . Both small RNA types were presented in equal aboudance in lung-stage schistosomula , while endo-siRNAs accumulated to 6 times more than miRNAs in adult female worms and hepatic eggs . Further , miRNAs were found mainly derived from genes located in the intergenic regions , while endo-siRNAs were mainly generated from transposable elements ( TEs ) . The expression pattern of TE-siRNAs , as well as the pseudogene-derived siRNAs clustered in mRNAs of cytoskeletal proteins , stress proteins , enzymes related to energy metabolism also revealed distinction throughout different developmental stages . Natural antisense transcripts ( NATs ) -related siRNAs accounted for minor proportion of the endo-siRNAs which were dominantly expressed in cercariae . Our results represented a comprehensive expression profile of sncRNAs in various developmental stages of S . japonicum with high accuracy and coverage . The data would facilitate a deep understanding of the parasite biology and potential discovery of novel targets for the design of anti-parasite drugs . Schistosomiasis is a chronic debilitating disease that afflicts more than 200 million individuals in the tropics and sub-tropics regions [1] . The agents of this disease , parasitic flatworms of the genus Schistosoma , have a complex developmental life cycle characterized by a distinct parasitic phase in mammalian and molluscan hosts and a free-living phase in freshwater . There are at least seven discrete developmental stages of the parasite within the definitive ( lung-stage schistosomula , juvenile , adult male and female worms , and eggs ) and intermediate ( sporocysts ) hosts as well as the aquatic , free-swimming miracidia and cercariae , with dramatically morphological changes [2] . And they are among the few platyhelminth parasites to adopt a dioecious lifestyle and possess heteromorphic sex chromosomes , which are arrayed in 7 pairs of autosomal chromosomes and one pair of sexual chromosomes ( Z , W ) , homozygous ( ZZ ) for male and heterozygous ( ZW ) for female [3] , [4] . Previous investigations on Schistosoma japonicum had revealed that a complex gene regulation pattern was deployed by this genus of parasites [5] and its haploid genome , which is about 270 Mb in size , has been recently decoded as a valuable entity for identification of small regulatory RNAs of this parasite [6] . Small non-coding RNA transcripts , approximately 18–30 nucleotides in length , are critical regulators in silencing of target genes in fungi , plants , and metazoans [7]–[9] . Three major categories of sncRNAs , siRNAs , miRNAs , and Piwi-interacting RNAs ( piRNAs ) have been well established and extensively studied [10] . SncRNAs exert their regulatory functions in chromatin architecture modelling , post-transcriptional repression and mRNA destabilization , mobile genetic elements suppression , and virus defence , usually through guiding the RNA-induced silencing complex ( RISC ) to their target genes [7] , [11]–[13] . In Drosophila , sncRNAs are generated through Dicer-dependent or independent pathways [14] . Dicer-1 generates miRNAs whereas Dicer-2 creates endo-siRNAs . Recently , it was found that the Argonaute protein family , which include the ubiquitous AGO ( AGO1 and AGO2 ) and the germline-specific Piwi ( AGO3 ) were devoted to different small RNA-mediated regulatory pathways [15] . AGO1 functions primarily in the miRNA-dependent pathway that silences messenger RNA , whereas AGO2 has been involved in RNAi-mediated silencing directed by exogenous and endogenous siRNAs . Further study in Drosophila somatic cells revealed that there were two classes of endo-siRNAs , one was generated from TEs and involved in retrotransposon repression; the other was produced in a Dicer-2-dependent manner from distinctive genomic loci , through refolding of RNA transcripts . The function of the second class of endo-siRNAs was likely to regulate mRNA stability in somatic cells [14] . Recently , several groups have endeavored to identify and characterize sncRNAs of schistosome with conventional cloning method and the deep-sequencing technique , mainly focused on juvenile and mixed adult worms , the two relatively closed developmental stages of the parasite [16]–[21] . A repertoire of miRNA transcripts unique to S . japonicum or those conserved to other metazoan lineages was identified . Differential expression of certain miRNAs was observed between the two developmental stages of S . japonicum ( hepatic schistosomula and adult worms ) and S . mansoni ( 7d schistosomula and adult worms ) , suggesting that miRNAs play a distinct role in modulating development , maturation , and reproduction of the parasite [17]–[19] , [21] . Importantly , miRNA genes within one cluster could be differentially expressed , which emphasized that individual transcript might be developmentally regulated by distinct mechanisms [17] , [19] . Meanwhile , a set of endo-siRNAs derived mainly from transposable elements ( TEs ) and the natural antisense transcripts ( NATs ) of S . japonicum has also been defined [17] , [19] . Interestingly , the distinct length and 3′ end heterogeneity of endo-siRNAs derived from both TEs and NATs were also associated with the developmental differentiation of the parasite [17] . Though the knowledge regarding sncRNA biology within the juvenile and mixed adult worms of S . japonicum is expanding , it is indispensable to systematically profile the repertoire of sncRNAs in other stages , especially the cercariae , which is the only infectious stage to penetrate its mammalian hosts; the lung-stage schistosomula , that is viewed as the most susceptible stage for intervention [22] , [23]; the tissue trapped eggs , which is the critical mediator for severe pathology in schistosomiasis , and the difference between the two sexes of adult worms . In this study , the expression profile of sncRNAs in the four critical developmental stages of S . japonicum was systematically investigated . The data , for the first time , provide a broader view of small non-coding RNAs in the parasite . The freshly released cercariae of S . japonicum were harvested from parasite-infected Oncomelania hupensis purchased from Jiangxi Institute of Parasitic Diseases , Nanchang , China . The lung-stage schistosomula ( 3 days post infection ) were isolated from lung tissues of infected Kunming strain mice as previously described [24] . Adult worms were obtained by hepatic-portal perfusion of New Zealand White rabbits or BALB/c mice 7-weeks post infection . Male and female worms were manually separated with the aid of a light microscope . Liver tissues deposited with schistosome eggs were obtained from BALB/c mice at day 30 and 45 post infection , respectively . All procedures performed on animals within this study were conducted following animal husbandry guidelines of the Chinese Academy of Medical Sciences and with permission from the Experimental Animal Committee . All animal work have been conducted according to Chinese and international guidelines . Total RNAs of S . japonicum at different developmental stages ( cercariae , lung-stage schistosomula , adult male and female worms perfused from infected rabbits ) and liver total RNAs of BALB/c mice 30d and 45d post infection were extracted using Trizol reagent ( Invitrogen , CA , USA ) . RNA quantification and quality were evaluated by Nanodrop ND-1000 spectrophotometer ( Nanodrop Technologies , Wilmington , DE ) and Agilent 2100 Bioanalyzer ( Agilent Technologies , Palo Alto , CA ) . Construction of small RNA libraries was carried out as described previously . Briefly , RNAs between 15–30 nucleotides ( nt ) were excised from a 15% TBE urea polyacrylamide gel electrophoresis ( PAGE ) . RNA samples were purified and ligated to Illumina's proprietary 5′ and 3′ adaptors , and further converted into single-stranded cDNA with Superscript II reverse transcriptase ( Invitrogen , CA , USA ) and Illumina's small RNA RT-Primer . The cDNA was amplified with high fidelity Phusion DNA polymerase ( Finnzymes Oy , Finland ) in 18 PCR cycles using Illumina's small RNA primer set . The purified PCR products were sequenced by an Illumina Genome Analyzer at the BGI ( Beijing Genomics Institute , Shenzhen , China ) . Raw datasets produced by deep sequencing from the libraries ( cercariae , lung-stage schistosomula , adult male and female worms , and infected liver tissues ) were pooled . Clean reads were obtained after removing of the low quality reads , adaptor null reads , insert null reads , 5′ adaptor contaminants , and reads with ployA tail . Adapter sequences were then trimmed from both ends of clean reads . All identical sequences were counted and merged as unique sequences , herein referred to as sequence tags . The unique reads along with associated read counts were mapped to the S . japonicum genome sequences ( http://lifecenter . sgst . cn/schistosoma/cn/schdownload . do ) using the program SOAP [25] . As for the liver libraries , the unique reads were mapped to the genome of mouse ( http://hgdownload . cse . ucsc . edu/downloads . html#mouse ) with SOAP , and those perfectly matched ones were removed prior to mapping to the S . japonicum genome . Briefly , the perfectly matched reads were first BLAST-searched against the 78 known mature miRNAs of S . japonicum deposited in Sanger miRBase [26] , [27] ( Release 15 ) using the program Patscan [28] . The remains were then BLAST-searched against Metazoa other than S . japonicum miRNAs allowing two mismatches to identify homologs of known Metazoa miRNAs . These homologs , as well as non-conserved reads ( with rRNA , tRNA , snoRNA and high repetitive reads being filtered out [29] ) were considered as potential miRNAs . To avoid repeated prediction and reduce the calculation redundancy , we then searched against the genome of S . japonicum and combined candidate unique reads located in close proximity in the reference genome with less than 150 bp and we called the joint genomic fragment as a block . For each block , 150 nt upstream and 150 nt downstream sequence were extracted for secondary structure analysis . We used software Einverted of Emboss [30] to find the inverted repeats ( step loops or hairpin structure ) , with the parameters threshold = 30 , match score = 3 , mismatch score = 3 , gap penalty = 6 , and maximum repeat length = 240 as described [31] . Each inverted repeat was extended 10 nt on each side , the secondary structure of the inverted repeat was folded using RNAfold [32] and evaluated by mirCheck using default parameters [31] . MiRNA candidates passed mirCheck were Blast-searched against Metazoa miRNAs except those of S . japonicum using the program Patscan again and labeled with conserved and non-conserved ( novel ) miRNAs , respectively . The novel unique reads that sequenced less than 2 times were removed . Finally , miRNA precursors that passed MirCheck were inspected manually in order to remove the false prediction . We employed IDEG6 to identify miRNAs showing statistically significant difference in relative abundance ( as reflected by TPM value ) between any two small RNA libraries [33] . The general Chi2×2 test was applied to determine whether one particular miRNA was significantly differentially expressed between any two samples ( P value < = 0 . 01 ) . Hierarchical clustering of the known S . japonicum miRNAs was constructed based on the transformed data of log10 of TPM value . The transposable elements in the S . japonicum genome were predicted by using REPET ( http://urgi . versailles . inra . fr/index . php/urgi/Tools/REPET ) . TE-derived siRNAs were identified as previously described [17] . Figures were constructed to reflect the relative abundance of sense and antisense of TE-derived siRNAs during the parasite development . Briefly , the expression value of each base on TE was the sum of the expression of siRNAs that mapped to the position . After a proper bin ( 20–50 bases ) was selected based on the length of TE sequences , the average expression value was calculated for each bin , and the expression level for four stages was marked by different colors . The natural antisense transcripts of S . japonicum were annotated and NAT-derived siRNAs were confirmed as described [17] . The small RNAs that failed to pass mirCheck were aligned to S . japonicum predicted mRNA sequences of SGST ( http://lifecenter . sgst . cn/schistosoma/cn/schdownload . do ) using the program SOAP , and perfectly matched reads were retained . Then a Perl script was wrote to scan the predicted mRNAs , if the region continuous covered by small RNAs is longer than 100 bp , the region was deemed as a “siRNA-Cluster” . Stem-loop qRT-PCR was performed to quantitate the sex-biased expressed miRNAs [20] , [34] . Stem-loop RT primers were designed to reverse-transcribe target miRNAs into cDNAs using total RNAs isolated from male and female adult worms , respectively ( from the same smaples used for Solexa sequecing ) . The 20 µl reaction system contained 1 µg of total RNA , 50 nM of each individual stem-loop RT primer , 1×RT buffer , 0 . 5 mM dNTPs ( Takara ) , 0 . 01 M DTT ( Invitrogen ) , 0 . 25 µl Superscript III reverse transcriptase ( 200 U/µl , Invitrogen , CA , USA ) , and 0 . 1 µl RNaseOUT inhibitor ( 40 U/µl , Invitrogen ) . cDNA was synthesized by incubation for 30 min at 16°C , 30 min at 42°C , 15 min at 70°C . Real-time quantification was carried out using an Applied Biosystems StepOne Plus system . PCR reactions were set up by combining 0 . 5 µM miRNA-specific forward primer , 0 . 5 µM common reverse primer , 2 µl of RT product ( 1∶1 dilution ) , 10 µl of Power SYBR Green PCR Master Mix ( ABI , CA , USA ) , and adjusted to a final volume of 20 µl with DEPC-treated water in triplicates . For endogenous control , 1 µg of male or female total RNA was converted to cDNA with oligo ( dT ) . The forward primer: 5′-CCTTCATGGTAGACAACGAAGCT-3′ and reverse primer: 5′-TGTAGGTTGGACGCTCTATGTCC-3′ , were used to amplify the α-tubulin gene as an endogenous control . The reaction conditions were as follows: 95°C for 5 min , followed by 40 cycles of 95°C for 5 sec and 60°C for 30 sec . The quantification of each miRNA relative to α-tubulin mRNA was calculated using the equation: N = 2−ΔCt , ΔCt = CtmiRNA - Ctα-tubulin [35] . All primers used are listed in Table S1 . 5′-DIG-labeled miRCURY LNA probes were ordered from Exiqon ( Vedbaek , Denmark ) ( Http://www . exiqon . com ) . Northern blot analysis was performed mianly by a method described in a previous study [36] . Total RNAs were isolated from male adult female adult worms perfused from BALB/c mice 7-weeks post infection . 10 µg total RNA of each smaple was resolved by 15% denaturing ( 7 M urea ) PAGE and were blotted by capillary transfer to neutral Nylon Membranes ( Hybond-NX , GE ) with 20×SSC . RNAs were further cross-linked to the membrane by EDC ( 1-ethyl-3- ( 3-dimethylaminopropyl ) carbodiimide ) method [37] . Blots were pre-hybridized by incubation with DIG Easy Granule ( Roche ) at 37°C for 3 h . And hybridization were carried out in the same buffer containing 1 nM DIG-labeled LNA probe at temperature recommended by manufacturer ( RNA Tm - 30°C ) overnight . Blots were washed twice in a low stringently buffer ( 2×SSC , 0 . 1% w/v SDS ) , and four times in a high stringently buffer ( 0 . 1×SSC , 0 . 1% w/v SDS ) , for 30 min each , both at hybridization temperature . The membrane was rinsed in washing buffer , and incubated in blocking solution at room temperature for at least 2 h ( DIG washing buffer and blocking solution Set , Roche ) . Subsequently , blots were incubated with a 10 , 000-fold dilution of anti-DIG-AP ( Roche ) in blocking solution at room temperature for 30 min , washed 5 times for 15 min each in washing buffer . After rinsing in detection buffer for 5 min , the blots were detected using CDP-star chemiluminescent substrate for alkaline phosphatase ( Roche ) . Blots were stripped by boiling for 1 min at 100°C in 10 mM Tris-HCl , pH 8 . 0 , 5 mM EDTA , and 0 . 1% SDS and probed up to three times . Six small RNA libraries were generated by high-throughput RNA sequencing ( see Materials and Methods and Table S2 ) . Four libraries , SjC , SjL , SjM , and SjF , were constructed from sequences that were directly derived from the cercariae , lung schistosomula , adult male , and female worms , respectively . The two remaining libraries , SjE30 and SjE45 , were egg libraries derived from the hepatic tissues of BALB/c mice 30 and 45 days post-infection , respectively . The reads that aligned to the mouse genome were filtered before they were mapped to the genome of S . japonicum . In total , 65 , 630 , 916 clean reads were obtained from libraries SjC , SjL , SjM , and SjF , which were merged into 6 , 989 , 949 unique tags , thus resulted in an average redundancy as high as 89 . 3 ( Redundancy = 100- ( Total Unique Tags/Total Clean Reads ×100 ) ) . Among these unique tags , 1 , 593 , 604 ( 22 . 80% ) could be aligned to the genome of S . japonicum ( Table S3 ) . The match rate was varied among different libraries , from the lowest of 20 . 46% ( SjM ) to the highest of 31 . 95% ( SjF ) , this phenomenon may related with the change of ratio of different small RNAs during development and/or between sexes . The low match ratio to the genome may be caused by either genome variation of different parasite isolates or due to less sequence information of the intergenic regions where most of the miRNAs were generated . The phenomenon was also observed in similar studies by others , and several explanations have been offered [18] . Regarding the egg libraries , 15 , 774 and 20 , 800 unique tags from libraries SjE30 and SjE45 , respectively , mapped to the S . japonicum genome ( Table S4 ) . These datasets contain roughly an order of magnitude more sequence than previous similar studies . The short ncRNA transcripts were categorized according to features related to primary and secondary structure ( Figure 1 and Table S5 ) . The majority of the ncRNA transcripts were miRNAs and TE-derived endo-siRNAs , accounting for 26 . 75% and 44 . 77% , respectively , of the total sncRNA pool ( Figure 1A ) . Only 2 . 21% of the miRNAs identified in our libraries were novel , indicating that most miRNAs have been recovered from the genome . Long terminal retrotransposons ( LTR ) and un-annotated transposons were predominant in the set of endo-siRNAs . Interestingly , the sets of miRNAs and endo-siRNAs displayed stage- or sex-related variation in expression ( Figure 1B and C ) . The percentage of miRNA was approximately double than that of the TE-derived endo-siRNA set in the cercariae library; the amount of miRNAs and endo-siRNAs was almost equal in lung-stage schistosomula , while endo-siRNAs were dominant in the adult worms and eggs , especially in female worms and early deposited eggs ( 6 times more than that of miRNAs ) . A class of endo-siRNAs derived from unclassified transposons was dominantly expressed in the male and female parasite compared to other stages ( Figure 1B ) . The clear pattern of preferential expression of the genes encoding the two classes of small RNAs suggests that they play stage-specific regulatory functions . Before invasion into a mammalian host , the parasite is likely to mainly utilize miRNA pathways to regulate gene expression , while endo-siRNA mediated regulation is suppressed . The high percentage of TE-derived endo-siRNAs in females and early deposited eggs suggests that siRNAs are more functional at these developmental stages . Earlier studies in D . melanogaster and mouse oocytes demonstrated that endo-siRNAs were critical elements for maintaining genomic stability through suppression of TE activity [38]–[40] . S . japonicum possesses a faster reproductive rate than flies or mice , and thousands of eggs are released by one female adult worm each day . Efficient suppression of TE activity is likely a prerequisite for continuity of parasite development and transmission , a possible explanation for why TE-derived endo-siRNAs were dominantly found in late-stage parasites . When the sequences of the small RNAs containing classical miRNA structure were aligned to the Sanger miRBase ( Release 15 ) , 77 sequences homologous to known or well-characterized miRNAs were identified . We found 74 , 71 , 69 , 70 , 18 , and 25 such sequences in libraries SjC , SjL , SjM , SjF , SjE30 , and SjE45 , respectively . Only one miRNA , the previously reported sja-miR-8-5p [19] , was not detected in this study ( Table S6 ) . Among the set of 77 known miRNAs , approximately 20 miRNAs were conserved homologues of sequences from the planarian Schmidtea mediterranea , the genus most closely related to Schistosoma , in previous investigations [17] , [19] , [20] , [41]–[43] , indicated that phylum Platyhelminthes contains common miRNAs that carry out similar biological function . The maximum read number of a single miRNA was 1 , 044 , 358 ( library SjC , sja-miR-1; Table S7 ) , illustrating the sequencing depth of our investigation . The range of read numbers was from the single-digits to millions , highlighting the sequencing capacity of next-generation sequencing technology and suggesting that expression variation of these miRNAs does indeed exist . This observation most likely reflects functional differentiation among the miRNAs . Apart from the known miRNAs , 193 hairpins containing 45 conserved mature miRNAs derived from 19 families were predicted in our sequence libraries . These miRNAs along with their expression level ( reflected by transcripts per million , TPM ) during development were shown in Table S8 . Additionally , we identified 199 novel miRNAs with various expression levels and stage specificities ( Table S9 ) . In contrast with the common or evolutionarily conserved miRNAs , most novel miRNAs identified in this study possessed low read numbers , with the exceptions of sequences sja-novel-23-5p and sja-novel-48-3p , which was mainly expressed in female adult worms and cercariae , respectively . Previous investigations of miRNA biogenesis revealed that miRNA genes are located either in intergenic regions [44] that are controlled by their own miRNA promoters and regulatory units [45] , or in introns , non-protein coding genes , or exons , and thus they are likely to be regulated in concert with host genes [46] . In an earlier study , we found that many S . japonicum miRNA genes were clustered together , and that genes within the same cluster may be regulated independently [17] . In the present study , we mapped all identified miRNA sequences to the S . japonicum genome and found that miRNAs were generated from 5′ or 3′ UTRs , intragenic , and intergenic regions in the genome; however , a majority of sequences ( 87 . 2% of the total miRNAs identified ) were transcripts derived from genes located in intergenic loci ( Figure 2 ) . Thus , compared to Caenorhabditis elegans , S . japonicum has evolved more sophisticated control mechanisms for regulation of miRNA expression , possibly explaining the complicated nature of the transcription profiles of individual miRNAs in various developmental stages of the parasite . Although the relative expression level of a particular miRNA has been proposed to be represented by the number of sequence reads , other investigations have argued that neither read counts nor northern blot signal accurately reflect actual abundance or expression level [19] , [47] , [48] . Here , the expression levels of each unique tag in cercariae , lung-stage schistosomula , separated adult worms and eggs libraries were normalized to TPM as previously described [18] , [49] , [50] . Thus , the read abundance should basically reflect the expression level of the tags in the parasites . The scale of the relative miRNA abundance during the various developmental stages appears in Figure 3 . Of 77 known miRNAs , 28 miRNAs exhibited high expression levels in one or more developmental stages . The expression levels of the novel miRNAs identified in this study were generally low ( Table S9 ) . However , four miRNAs were with relatively higher expression level in one particular stage , as sja-Novel-23-3p and sja-Novel-23-5p were dominantly expressed in the female parasite , while sja-Novel-48-3p and sja-Novel-74-3p were substantially expressed in cercarial stage . Like C . elegans , schistosomal parasites need to complete a series of biological and physiological activities , including protease secretion , tail detachment , glycocalyx shedding , and tegument transformation before developing to the schistosomula stage [51] , [52] . A particular gene repertoire of S . mansoni parasites was previously shown to be up-regulated during the transition from schistosomula to adult worms [22] . Here , we observed that the expression of a set of miRNAs including sja-bantam , sja-miR-1 , sja-miR-124-3p , sja-miR-2a-3p , sja-miR-3492 , and sja-miR-36-3p was substantially down-regulated in lung-stage schistosomula compared to cercariae ( Table S6 ) , suggesting that the target mRNAs of these miRNAs may encode proteins fulfilling important functions at this stage . We further explored the differential expression of miRNA genes between male and female adult worms . The expression of a set of miRNAs , sja-miR-7-5p , sja-miR-61 , sja-miR-219-5p , sja-miR-125a , sja-miR-125b , sja-miR-124-3p and sja-miR-1 were dominant in male worms , while sja-bantam , sja-miR-71b-5p , sja-miR-3479-5p , and sja-Novel-23-5p were predominantly found in the female parasites ( Table S6 and S9 ) . The expression of these sex-biased miRNAs was validated by stem-loop RT-PCR ( Figure 4A ) . The expression level of sja-miR-1 was relatively high in male adult worms ( 1 . 098±0 . 228 ) and female adult worms ( 0 . 358±0 . 021 ) when compared to other miRNAs , and was not shown in Figure 4A . The correlation between the TPM values and qPCR was investigated by a method described in a previous study ( R = 0 . 882 , Spearman's Rho , p<0 . 0001 , n = 11 ) [53] . However , among individual miRNAs , the qPCR results did not exactly reflect the TPM values of the maximally expressed miRNAs , probably due to the existence of extensive miRNA isomiRs , or asymmetrical amplification during library construction . We further validated the expression differences of ten sex-biased miRNAs by northern blot analysis using the total RNA extracted from adult male and female worms isolated from BALB/c mice 7-weeks post infection ( Figure 4B ) . The differential expression pattern of these miRNAs except sja-miR-71b-5p between male and female worms was quite consistent with the TPM values of high-throughput sequencing and the qRT-PCR results . The phenomenon was also observed in a recent study which noted that several miRNAs were expressed at similar levels in protoscoleces of G1 and G7 genotypes Echinococcus granulosus , which parasitized in different hosts [54] . Thus , these data indicated that host factors may have little impact on the expression profile or level of sncRNAs . Although the function of these miRNAs remains to be elucidated , the significant differential expression between male and female adult worms indicated that they may participate in regulation of sexual differentiation and maintenance , pairing and reproduction of the parasite . Moreover , miRNAs may cooperate with other small RNAs ( such as endo-siRNAs ) and transcription factors to form a comprehensive network to regulate growth , development , differentiation , and reproduction for adaptation to a variety of environments [19] . Further studies on these miRNAs may contribute to better understanding of the developmental mechanism of sexual dimorphism in this parasite . Recent observations of endo-siRNAs in D . melanogaster , mice , and schistosome have added more complexity to our knowledge of small RNA-mediated regulatory pathways [14] , [17] , [38]–[40] , [55]–[58] . So far , endo-siRNAs have been found to be mainly derived from TEs , complementary annealed NATs , and the long “fold-back” transcripts known as hairpin RNAs [59] . We previously found that the TE-derived siRNAs in S . japonicum were more predominant than other endo-siRNAs , including NAT-derived siRNAs [17] . Here , we systematically analyzed the expression levels of sense and antisense endo-siRNAs that derived from various TEs in cercariae , lung-stages schistosomula , male and female adult worms ( Table S10–14 ) . The read numbers of endo-siRNAs in egg libraries were much lower than the other libraries , leading us to exclude the egg libraries from further analysis . We observed that LINE , TIR , and LTR transposon classes were the main sources of endo-siRNAs , while the endo-siRNAs derived from other TEs were much less abundant ( Figure 5 ) . Further , endo-siRNAs mapped to the top ( sense siRNA ) and bottom ( antisense siRNA ) strands of LTR and non-LTR TEs . The expression patterns of LTR-derived sense and antisense siRNAs were relatively symmetrical , though there were obvious stage and sex specificities in expression loci ( Figure 5A , B , and C ) . Reads mapped to the S . japonicum LTR retrotransposon SjCHGCS11 [6] were annotated as SACI-7_2p in our analysis ( Figure 5A ) . Both sense and antisense siRNAs were concentrated in the coding region of reverse transcriptase in a manner similar to that observed in D . melanogaster somatic cells [38] . Sjpido , SjR1 , and SjR2 are three classes of non-LTR retrotransposons that make up 5% of the S . japonicum genome; siRNAs generated from these elements mainly mapped to certain sequence regions ( Figure 5B ) , contrary to our observations of LTR retrotransposons . The expression levels of siRNAs derived from SjR1 were much higher in cercariae than in male and female adult worms , indicating that these siRNAs are more functional in the earlier developmental stage ( Figure 5B ) . Sj-alpha-1 derived siRNAs were predominantly generated from the antisense strand , while Sj-alpha-2 derived siRNAs were generated from the sense strand; however , both types of siRNAs had low expression levels ( Figure 5C ) . In cercariae and lung-stage schistosomula , the TIR ( Sj_Blaster_Recon_7337_MAP_14 annotated as SmTRC1_1p in the genome ) derived siRNAs were highly expressed , while the MITE ( Sj_Blaster_Grouper_1934_MAP_4 ) derived siRNAs were mainly expressed in the adult worms , and predominantly corresponded to the antisense strand ( Figure 5D ) . Thus , the TE-derived endo-siRNAs of S . japonicum were more diverse than those found in D . melanogaster . Although the origin of the antisense siRNAs is not known ( cis- or trans-transcription ) , their abundance suggests that they are stable and likely participate in regulatory pathways . Previous studies of mouse oocytes revealed that antisense transcripts from pseudogenes formed double-strand RNAs with their functional counterparts , the sources of the endo-siRNAs , and the sense siRNAs were predominant in the endo-siRNA [55] . It has been proposed that the “passenger strand” of an siRNA is unstable due to the thermodynamic asymmetry of the two strands [60] . However , this hypothesis cannot explain our identification of many reads corresponding to the antisense siRNAs; in some cases , only the antisense strands were identified . Further dissection of the function of the two endo-siRNA classes would be an essential step toward understanding the network of gene regulation during the parasite development . NAT-derived siRNAs are a second source of endo-siRNAs; these endo-siRNAs are further classified as cis-NAT- or trans-NAT-derived endo-RNAs [56] , [61] , [62] . Cis-NAT-derived endo-siRNAs are generated from transcripts from the same gene locu , while trans-NAT-derived endo-siRNAs come from NAT transcripts of distinct loci . We detected potential NAT pairs by aligning the predicted mRNA sequences to each other . Only one cis-NAT pair and 1772 trans-NAT pairs were identified in silico using data from SGST . Our sequencing results were remarkably similar to the in silico prediction; one cis-NAT pair and 225 trans-NAT pairs were detected ( Table S15 ) , indicating that bi-directional transcription was much less prevalent in schistosomal parasites and transcripts from duplicated genes are more common . Thus , trans-NAT-derived endo-siRNAs are likely the main sources of NAT-derived siRNA in S . japonicum , a scenario that differs from other organisms [63] . However , we cannot rule out the possibility that most of the cis-NAT pairs may be undetectable given the lack of information about the non-protein-coding regions of the S . japonicum genome . The identification of long non-coding RNAs in S . japonicum is still underway , and may provide an important source of NAT-derived siRNAs [64] . A previous study of D . melanogaster somatic cells demonstrated that endo-siRNAs mapped to protein-coding mRNAs rather than to transcripts of transposons that regulate mRNA expression [38] . Here , we also mapped the endo-siRNAs to the predicted mRNA sequences of S . japonicum , and found that nearly half of the siRNA-related transcripts clustered within predicted mRNAs . These mRNAs mainly encoded proteins from four categories: 1 ) proteins similar to pol polyprotein and endonuclease-reverse transcriptase; 2 ) cytoskeletal proteins such as myosin , actin , and tropomyosin; 3 ) enzymes or transporters such as COX1 , COX2B , superoxide dismutase 1 , glyceraldehyde 3-phosphate dehydrogenase , lactate dehydrogenase A , ATP-dependent RNA helicase , cation-transporting ATPase , H+-transporting ATPase , and cathepsin B and L; 4 ) stress proteins including heat shock protein , cold shock protein , and stress-induced phosphoprotein 1 ( Table S16 ) . We were unable to distinguish whether siRNAs clustered in pol polyprotein and endonuclease-reverse transcriptase transcripts were derived from retrotransposons or NATs . We speculated that some of the siRNAs in the other three categories were derived from trans-NATs formed by transcripts of pseudogenes and their parental genes , as suggested recently [65]; for example , the pseudogenes of hsp70 and cathepsin B exist in schistosome genomes [66] , [67] . Furthermore , the pseudogenes of actin , COX , GAPDH , FABP , and histone are common in mammalian genomes . Pseudogene-derived endo-siRNAs were previously detected in mouse oocytes , with two transcripts , Hsp90ab1 ( heat shock protein 90 kDa alpha , class B member 1 ) and Dynll1 ( dynein , light chain ) , possessing features similar to our findings [55] . Thus , unlike the silencing of selfish genetic elements by TE-related siRNAs , trans-NAT-derived endo-siRNAs mainly regulate the expression of mRNAs coding for a diverse set of proteins . Our current study generated comprehensive profiles of endogenous small RNAs ( miRNAs and endo-siRNAs ) during the four crucial developmental stages of S . japonicum . Reverse expression patterns of miRNAs and endo-siRNAs during the parasite development and differentiation process were observed . Two classes of endo-siRNAs derived from TEs and trans-NATs were identified , and the LTR retrotransposon derived siRNAs were more abundant than siRNAs from non-LTR TEs . There are likely two layers of regulatory function employed by the parasite; the antisense siRNAs directly affect the stability of mRNA transcripts , while the sense siRNAs may function indirectly by affecting the amount of antisense transcripts . Thus , the small RNA-mediated network in schistosomal parasites is more complex than networks reported in other organisms .
Schistosomiasis , a debilitating disease , caused by agents of the genus Schistosoma afflicts more than 200 million people worldwide . Schistosomes could serve as an interesting model to explore gene regulation due to its evolutional position , complex life cycle and sexual dimorphism . We previously indicated that sncRNA profile in the parasite S . japonicum was developmentally regulated in hepatic and adult stages . In this study , we systematically investigated mircoRNA ( miRNA ) and endogenous siRNA ( endo-siRNA ) profile in this parasite in more detailed developmental stages ( cercariae , lung-stage schistosomula , separated adult worms , and liver tissue-trapped eggs ) using high-throughput RNA sequencing technology . We observed that the ratio of miRNAs to endo-siRNAs was dynamically changed throughout different developmental stages of the parasite . MiRNAs were expressed dominantly in cercariae , while endo-siRNAs accumulated in adult female worms and hepatic eggs . We demonstrated that miRNAs were mostly derived from intergenic regions whereas siRNAs were mostly derived from transposable elements . We also annotated miRNAs and siRNAs with stage- and gender- biased expression . Our findings would facilitate to understand the gene regulation mechanism of this parasite and discover novel targets for anti-parasite drugs .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results", "and", "Discussion" ]
[ "medicine", "infectious", "diseases", "schistosomiasis", "parasitic", "diseases", "helminth", "infection" ]
2011
Profiles of Small Non-Coding RNAs in Schistosoma japonicum during Development
The spiking activity of principal cells in mammalian hippocampus encodes an internalized neuronal representation of the ambient space—a cognitive map . Once learned , such a map enables the animal to navigate a given environment for a long period . However , the neuronal substrate that produces this map is transient: the synaptic connections in the hippocampus and in the downstream neuronal networks never cease to form and to deteriorate at a rapid rate . How can the brain maintain a robust , reliable representation of space using a network that constantly changes its architecture ? We address this question using a computational framework that allows evaluating the effect produced by the decaying connections between simulated hippocampal neurons on the properties of the cognitive map . Using novel Algebraic Topology techniques , we demonstrate that emergence of stable cognitive maps produced by networks with transient architectures is a generic phenomenon . The model also points out that deterioration of the cognitive map caused by weakening or lost connections between neurons may be compensated by simulating the neuronal activity . Lastly , the model explicates the importance of the complementary learning systems for processing spatial information at different levels of spatiotemporal granularity . How does the decay of the connections affect the net structure of the flickering complex F τ ? As shown on Fig 5A , the numbers of links N2 ( t ) and of the triple connections N3 ( t ) rapidly grow at the onset of the navigation and begin to saturate in about ts ≈ 4 minutes ( i . e . , by the time when a typical link had time to make an appearance ) , reaching their respective asymptotic values in ta ≈ 7 minutes . To put the size of the resulting flickering complex into a perspective , note that the number of simplexes in a decaying complex F τ < ∞ can never exceed the number of simplexes that would have existed in absence of decay , i . e . , in the “perennial” coactivity complex , F ∞ ≡ T . Thus , the size of the complex at a moment t , F τ ( t ) , can be characterized by the proportion of simplexes that happened to be actualized at that moment . As illustrated on Fig 5A , these numbers fluctuate around 60% for the second order simplexes and around 40% for the third order simplexes , with the relative variances ΔN2/N2 ≈ 12% and ΔN3/N3 ≈ 17% respectively . In other words , the perennial coactivity complex F ∞ ( t ) loses about a half of its size due to the flickering of the simplexes , and fluctuates within about 15% margins from the mean . To quantify the changes in the complexes’ structure as a function of time , we evaluated the number of two- and three-vertex simplexes that are present at a given moment of time ti , but are missing at a later moment tj , normalized by the size of F τ ( t i ) , i . e . , d i j ( k ) = N k ( F τ ( t i ) \ F τ ( t j ) ) / N k ( F τ ( t i ) ) , k = 2 , 3 . As shown on Fig 5B , these numbers , which we refer to , respectively , as the second and third asymmetric distances between F τ ( t i ) and F τ ( t j ) , rapidly grow as a function of temporal separation |ti − tj| . In fact , after approximately the effective decay time τ e ( 2 ) , the difference between F τ ( t i ) and F τ ( t j ) becomes comparable to the sizes of either F τ ( t i ) or F τ ( t j ) , which implies that the pool of simplexes in the simplicial complex is replenished at the effective decay timescale . However , the shape of the coactivity complex changes slowly: Fig 5C demonstrates that nearly 100% of the connections are shared at two consecutive moments , i . e . , the changes in flickering complex from one moment of time to the next are marginal . Over longer periods , the flickering complex can change significantly . For example , the proportion of simplexes that are present at t* = 9 minutes , when F τ is particularly inflated , and at other moments , varies around N 2 ( F τ ( t * ) ∩ F τ ( t ) ) ≈ 82 % for second and N 3 ( F τ ( t * ) ∩ F τ ( t ) ) ≈ 64 % for the third order simplexes ( Fig 5C ) . Despite the rapid recycling of the individual simplexes , the large-scale topological characteristics of the flickering complex remain relatively stable . As demonstrated on Fig 5D , after the initial stabilization period of about two minutes ( which biologically may be interpreted as the initial learning period ) , F τ contains only one zero-dimensional and a single one-dimensional topological loop—as the simulated environment E . Some topological fluctuations appear around t* ≈ 9 minutes , as indicated by an outburst of short-lived spurious loops , most of which last less than a minute . After this period , the first two Betti numbers of F τ retain their physical values b 0 ( F τ ) = b 1 ( F τ ) = 1 ( Fig 5E ) . Since Zigzag homology theory allows tracing individual loops in F τ continuously across time , these persistent topological loops can be viewed as ongoing representations of the simply connected environment E and of the physical hole in it . Thus , the coactivity complex F τ preserves , for the most time , not only its approximate size , but also its topological shape—despite transience at the “microscale” , i . e . , at the individual simplex level . Physiologically , these results indicate that the large-scale topological information significantly outlasts the network’s connections: although in the discussed case about a half of the functional links rewire within a τ e ( 2 ) -period , the topology of the cognitive map encoded by the cell assembly network remain stable . In other words , a transient cell assembly network can encode stable topological characteristics of the ambient space , despite transience of the connections . We investigated the topological stability for a set of proper decay times τ ranging from one to five minutes . As one would expect , the number of simplexes in the flickering complex increases with growing τ: in the studied map , the number of links raises from about 40% of the maximal value at τ = 75 secs to just under 60% for τ = 200 secs , whereas the number of the third order connections raises from 60% to about 80% ( Fig 6A ) . The distributions of the effective lifetimes for the short-lived ( fluttering ) connections retain their exponential shapes ( see Suppl . Materials ) with the means that are approximately proportional to the proper decay times , τ e ( 2 ) ≈ 2 τ and τ e ( 3 ) ≈ τ ( Fig 6B and 6C ) . The contribution of the surviving simplexes also steadily grows with τ ( see Suppl . Materials ) ; as a result , the net average lifetimes , computed for the entire population of simplexes , grow faster Δ t ς 2 ≈ 3 τ and Δ t ς 3 ≈ 2 τ . As τ increases , the Betti numbers rapidly reduce to their physical values , b 0 ( F τ ) = b 1 ( F τ ) = 1: the lower is the connection decay rate , the smaller are the topological fluctuations generated in the flickering complex ( Fig 6D and Suppl . Materials ) . This is a natural result: the longer the simplexes survive , the closer the topological shape of F τ is to the topological shape of the environment E . Physiologically , it implies that the lower is the cell assembly decay rate , the more stable is the cognitive map’s topological structure . As shown on Fig 6D , a stabilization of topological barcode is achieved around τ ∼ 2 minutes . This value can also be naturally interpreted: for such τ , the rat moving at the mean speed of about 25 cm/sec has time to visit most of the environment and reactivate connections in all parts of F τ before they may decay , which allows the induced coactivity complex to contract the spurious topological loops , to assume and to retain the correct topological shape . Note however , that this is only a qualitative argument since the expected lifetimes of over 63% of links is smaller than τ and the lifetimes of 15% of them live longer than 2τ . To test how these results are affected by the spread of the link lifetimes , we investigated the case in which the lifetimes of all the links are fixed , i . e . , the decay probability is defined by the function p ( t ) = { 1 if t = τ 0 if t ≠ τ , ( 2 ) while keeping the other parameters of the model unchanged . The results shown on Fig 7A demonstrate that due to the rejuvenation effects , the range of the effective lifetimes widens and becomes qualitatively similar to the histograms induced by the decay distribution ( 1 ) . As before , there appear two distinct populations of links: the short-lived links whose lifetimes concentrate around the singular proper lifetime τ , and the “survivor” links , whose lifetimes approach Ttot . However , the topological structure of the “fixed-lifetime” coactivity complex F τ * differs dramatically from that of the decaying complex F τ . As shown on Fig 7C , F τ * contains a large number of short-lived , spurious topological loops even for the values of τ that reliably produce physical Betti numbers in the case of the exponentially distributed lifetimes . For example , at τ = 100 secs , the zeroth Betti number of F τ * hovers at the average value of 〈b0〉 ≈ 40 , reaching at times b0 ∼ 100 , with nearly unchanged b1 = 1 , which indicates that F τ * is split into a few dozens of disconnected , contractible islets . As the proper decay time increases , the population of survivor links grows and the disconnected pieces of F τ * begin to pull together: at τ = 200 secs , the Betti numbers b k ( F τ * ) retain their physical values most of the time , yielding occasional splashes of topological fluctuations ( Fig 7C and 7D ) . These differences between the topological properties of F τ and F τ * indicate that the tail of the exponential distribution ( 1 ) , i . e . , the statistical presence of long-lasting connections is crucial for producing the correct topology of the flickering complex . Physiologically , this implies that the statistical spread of the connections’ lifetimes plays important role: without it , the dynamical cell assembly network fails to represent the topology of the environment reliably . These observations led us to another question: might the topology of the flickering complex be controlled by the shape of the lifetimes’ distribution and the sheer number of links present at a given moment , rather than the specific timing of the links’ appearance and disappearance ? To test this hypothesis , we computed the number N2 ( t ) of links in the decaying coactivity graph G τ ( t ) for τ = 100 sec at every discrete moment of time t ( see Methods ) , and randomly selected the same number of links from the maximal available pool , i . e . , from the graph G ∞ ( t ) that would have formed by that moment without links’ decay ( Fig 8A ) . The collections of links randomly selected at consecutive moments of time can be viewed as instances of a random connectivity graph G r ( t ) , i . e . , as a graph whose links can randomly appear and disappear , in contrast with the decaying links of G τ ( t ) ( compare Figs 8B and 3B ) . As it turns out , the random and the decaying graphs G r ( t ) and G τ ( t ) , as well as their respective clique complexes F r ( t ) and F τ ( t ) exhibit a number of similarities . First , the histogram of the net lifetimes of the links in G r ( t ) shown on Fig 8C is bimodal , with an exponential component characterized by the mean 〈T2〉 = 124 sec , and a component representing a population of surviving connections , similar to the histograms shown on Fig 4G and 4H . Second , the Betti numbers of the random coactivity complex F r converge to the Betti numbers of the environment in about 3 minutes—about as quickly as the Betti numbers of its decaying counterpart F τ ( Fig 8D ) . However , in contrast with the decaying flickering complex F τ , the random flickering complex F r keeps producing occasional one-dimensional loops over the entire navigational period at a low rate ( about 3% of the time , see Suppl . Materials ) . Thus , according to the model , the topological properties of the map encoded by a network with randomly formed and pruned connections are similar to the properties of a map produced by a network with decaying connections , as long as the net probability of the links’ existence are same . In either case , rapidly rewiring connections do not preclude the appearance of a stable topological map , which once again demonstrates that the latter is a generic phenomenon . The turnover of memories ( encoding new memories , integrating them into the existing frameworks , recycling old memories , consolidating the results , etc . ) is based on adapting the synaptic connections in the hippocampal network [47] . In particular , these processes require a balanced contribution of both “learning” and “forgetting” components , i . e . , of forming and pruning connections [11 , 12] . The imbalances and pathological alterations in the corresponding synaptic mechanisms are observed in many neurodegenerative conditions , e . g . , in the Alzheimer’s disease , which is known to affect spatial cognition [48] . However , interpreting the physiological meaning of these alterations is a challenging task , in particular because certain changes in neuronal activity may not be direct consequences of neurodegenerative pathologies . For example , it is believed that neuronal ensembles may increase the spiking rates of the active neurons in order to compensate for the reduced synaptic efficacies [49–55] . Such considerations motivate deep brain stimulation and other treatments that help to improve cognitive performance in animal models of Alzheimer’s diseases and in Alzheimer’s patients , by increasing the electrophysiological activity of hippocampal cells [56 , 57] . Previous studies , carried out for the models of perennial cell assembly networks [58] , provided a certain theoretical justification for these approaches . It was demonstrated that a place cell ensemble that fails to produce a reliable topological map of the environment due to an insufficient number of active neurons might be forced to produce a correct map by increasing the active place cells’ firing rates . Similarly , the reduction in the firing rates or poor spatial selectivity of spiking may sometimes be compensated by increasing the number of active cells and so forth . Since the current model allows modeling networks with transient connections , we wondered whether it might indicate a theoretical possibility to compensate for the reduced cell assemblies’ lifetimes by altering the place cell spiking parameters . To that end , we varied the mean firing rate f and the number of cells N in the simulated place cell ensemble and studied the topological properties of the resulting coactivity complex as a function of the links’ proper half-life , τ . The results shown on Fig 9 demonstrate that indeed , increasing neuronal activity helps to suppress topological fluctuations in the flickering coactivity complex for a wide range of the connections’ decay times . Moreover , these changes also increase the proportion of trials in which the place cell ensemble captures the correct signature of the environment ( see Suppl . Materials ) . Physiologically , these results indicate that recruiting additional active cells and/or boosting place cell firing rates helps to overcome the effect of overly rapid deterioration of the network’s connections , i . e . , increasing neuronal activity stabilizes the topological map . In particular , a higher responsiveness of the Betti numbers of the flickering coactivity complex to an increase of the mean firing rate ( Fig 9C and 9E ) as compared to the number of active place cells ( Fig 9A ) suggests that targeting the active neurons’ spiking may provide a better strategy for designing clinical stimulation methods . The formation and disbanding of dynamical place cell assemblies at the short- and intermediate-memory timescales enable rapid processing of the incoming information in the hippocampal network . Although many details of the underlying physiological mechanisms remain unknown , the schematic approach discussed above provides an instrument for exploring how the information provided by the individual cell assemblies may combine into a large-scale spatial memory map and how this process depends on the physiological parameters of neuronal activity . In particular , the model demonstrates that a network with transient connections can successfully capture the topological characteristics of the environment . Previously , we investigated this effect using an alternative model of transient cell assemblies , in which the connections were constructed by identifying the pool of cells that spike within a certain “coactivity window , ” ϖ , and building the coactivity graph G ϖ from the most frequently cofiring pairs of neurons [58] . The accumulation of topological information within each ϖ-period , was then described using persistent homology theory techniques . The results indicate that if ϖ extends over 4-6 minutes or more , the topological fluctuations in the flickering complex are suppressed and the topological shape of F ϖ becomes equivalent to the shape of the environment . In the current model , enabled by a much more powerful Zigzag persistent homology theory [34–36] , we employ an alternative approach , in which the links of the coactivity graph appear instantly following pairwise place cell coactivity events . Thus , in contrast with the model discussed in [58] , the current model involves no selection of the “winning” coactivity links , which one might hold responsible for stabilizing the shapes of the flickering coactivity complexes . Nevertheless , this model demonstrates the same effect: the large-scale topological shapes of resulting coactivity complexes stabilize , given that the connections decay sufficiently slowly and have sufficiently broadly distributed lifetimes . The connections’ lifetimes required to achieve such stabilization in the “latency free” model are longer than in the input integration model ( τ ≈ 100 sec vs . τ ϖ ≈ 10 sec ) , which indicates that physiological networks may integrate spiking information over a certain extended period ϖ and optimize the network’s architecture over this information . However , the fact that stable topological maps can emerge in all these different types of transient networks ( including randomly flickering networks ) suggests that this is a generic effect that fundamentally may be responsible for the appearance of stable cognitive representations of the environment in the physiological neuronal networks with transient connections . In other words , the emergence of stable topological maps may represent a common “umbrella” phenomenon that can be implemented via different physiological mechanisms . In all cases , the model reveals three principal timescales of spatial information processing . First , the ongoing information about local spatial connectivity is rapidly processed at the working memory timescale , which physiologically corresponds to rapid forming and disbanding of the dynamical place cell assemblies in the hippocampal network . The large-scale characteristics of space , as described by the instantaneous Betti numbers , unfold at the intermediate memory timescale . At the long-term memory timescale the topological fluctuations average out , yielding stable , qualitative information about the environment . While the former may take place in the hippocampus , the latter two might require involvement of the cortical networks . Thus , the model reaffirms functional importance of the complementary learning systems for processing spatial information at different timescales and at different levels of spatial granularity [47 , 59 , 60] . We consider a group of place cells c0 , … , ck coactive , if they produce spikes within two consecutive θ-periods [26 , 43] . As a result , the time interval [0 , Ttot] splits into 1/4 sec long time bins that define the discrete time steps t1 , … , tn . We use simplexes and simplicial complexes to represent combinatorially the topology of the neural activity . An abstract simplex of dimensionality n is a set containing n + 1 elements . A subset of a simplex is called its face . A simplicial complex is a collection of simplexes closed under the face relation: if a simplex belongs to a simplicial complex , then so do all of its faces ( Fig 11 ) . In the constructions studied in this paper , our simplicial complexes consist of coactive place cells . If all cells {c0 , … , ck} are coactive within a given time window , then so is any subset of them , meaning coactive simplexes form a complex . In fact , because coactivity is defined for a pair of cells , our simplexes are precisely the cliques in the coactivity graph . A simplex {c0 , … , ck} is present if and only if all of its cells are pairwise coactive . In flickering clique complexes , certain pairwise connections may decay over time , while others appear as time progresses . The effect on the simplicial complex is that some simplexes are removed from the complex , while others are added to it . So we get a sequence of “flickering complexes , ” Xi , connected by alternating inclusions: X 1 ⊆ X 2 ⊆ X 3 ⊇ X 4 ⊆ X 5 ⊇ … A k-dimensional chain is a set of k-dimensional simplexes ( Fig 11 ) that can be combined with suitable coefficients . If the coefficients form an algebraic field , then the chains form a vector space . Here we use the simplest algebraic field Z 2 , which consists of two Boolean values 0 and 1 . A boundary of the simplex is the sum of its one-dimension-lower faces: ∂ k { c 0 , … , c k } = ∑ i = 0 k { c 0 , … , c i - 1 , c i + 1 , … , c k } . The map extends linearly to the entire simplicial complex , X , mapping its k-dimensional chains to its ( k − 1 ) -dimensional chains . The kernel of this map , i . e . , all the chains without a boundary , is the set of cycles of the complex , denoted by Zk ( X ) = ker ∂k . The image of ∂k+1 consists of the k-dimensional chains that are boundaries of some ( k + 1 ) -dimensional chains , denoted by Bk ( X ) = im ∂k+1 . Cycles count “k-dimensional holes” in the complex . But not all such holes are independent of each other . We consider two cycles equivalent , or homologous , if they differ by a boundary . Algebraically , one can verify that boundaries themselves have no boundaries , ∂k ∘ ∂k+1 = 0 . In other words , boundaries are cycles . This allows us to take a quotient , Hk ( X ) = Zk ( X ) /Bk ( X ) , called the k-dimensional homology vector space . By definition , it considers two cycles equivalent , if their difference is a boundary of some ( k + 1 ) -dimensional chain . The dimension of this vector space , called the k-th Betti number , βk ( X ) = dim Hk ( X ) , counts the number of independent holes in the topological space . Given the sequence of flickering complexes above , we compute homology of each one . Inclusions between complexes induce maps between the homology vector spaces: the homology class of a cycle in the smaller complex maps to the homology class of the same cycle in the larger complex . Accordingly , we get a sequence of homology vector spaces , connected by linear maps: H k ( X 1 ) → H k ( X 2 ) → H k ( X 3 ) ← H k ( X 4 ) → H k ( X 5 ) ← … This sequence , called zigzag persistent homology , generalizes ordinary persistent homology [36] , where all the maps between homology groups point in the same direction . It is this generalization to the alternating maps that allows us to handle the flickering complexes . On the surface , zigzag persistent homology tracks how the Betti numbers of the flickering complexes change . But the maps that connect homology vector spaces provide extra information . It is possible to select a basis for each vector space in this sequence , so that the bases for adjacent vector spaces are compatible [34] . Specifically , we can select a collection of elements { z i j } j for each vector space Hk ( Xi ) , such that the non-zero elements form a basis for the homology vector space Hk ( Xi ) —in other words , they represent a set of independent holes in Xi . Furthermore , such collections are compatible in the sense that adjacent basis elements map into each other: if we have a map f: Hk ( Xi ) → Hk ( Xi±1 ) , then f ( z i j ) = z i ± 1 j , if z i j ≠ 0 . The experiments in this paper use the algorithm of Carlsson et al . [35] to compute such compatible bases . It follows that the sequence of homology vector spaces can be decomposed into a barcode , where each bar represents the part of the sequence , where a particular basis element is non-zero . The bars capture when independent holes appear in the flickering complex , how long they persist , and when they eventually disappear . The authors will provide the software used for these computations upon request .
The reliability of our memories is nothing short of remarkable . Synaptic connections between neurons appear and disappear at a rapid rate , and the resulting networks constantly change their architecture due to various forms of neural plasticity . How can the brain develop a reliable representation of the world , learn and retain memories despite , or perhaps due to , such complex dynamics ? Below we address these questions by modeling mechanisms of spatial learning in the hippocampal network , using novel algebraic topology methods . We demonstrate that although the functional units of the hippocampal network—the place cell assemblies—are unstable structures that may appear and disappear , the spatial memory map produced by a sufficiently large population of such assemblies robustly captures the topological structure of the environment .
[ "Abstract", "Results", "Discussion", "Methods" ]
[ "learning", "cell", "physiology", "action", "potentials", "medicine", "and", "health", "sciences", "neural", "networks", "membrane", "potential", "social", "sciences", "electrophysiology", "neuroscience", "learning", "and", "memory", "cognitive", "psychology", "mathematics", "algebra", "computer", "and", "information", "sciences", "animal", "cells", "vector", "spaces", "psychology", "cellular", "neuroscience", "linear", "algebra", "cell", "biology", "physiology", "neurons", "algebraic", "topology", "topology", "biology", "and", "life", "sciences", "physical", "sciences", "cellular", "types", "cognitive", "science", "neurophysiology" ]
2018
Robust spatial memory maps encoded by networks with transient connections
Noninvasive imaging and tractography methods have yielded information on broad communication networks but lack resolution to delineate intralaminar cortical and subcortical pathways in humans . An important unanswered question is whether we can use the wealth of precise information on pathways from monkeys to understand connections in humans . We addressed this question within a theoretical framework of systematic cortical variation and used identical high-resolution methods to compare the architecture of cortical gray matter and the white matter beneath , which gives rise to short- and long-distance pathways in humans and rhesus monkeys . We used the prefrontal cortex as a model system because of its key role in attention , emotions , and executive function , which are processes often affected in brain diseases . We found striking parallels and consistent trends in the gray and white matter architecture in humans and monkeys and between the architecture and actual connections mapped with neural tracers in rhesus monkeys and , by extension , in humans . Using the novel architectonic portrait as a base , we found significant changes in pathways between nearby prefrontal and distant areas in autism . Our findings reveal that a theoretical framework allows study of normal neural communication in humans at high resolution and specific disruptions in diverse psychiatric and neurodegenerative diseases . The functional specialization of the human cerebral cortex is critically dependent on the structural organization and connectivity of its cortical areas [1–5] . In recent years , evidence to support this structure-function relationship has focused on noninvasive methods through functional imaging and tractographic studies in humans . These approaches have led to analysis of large datasets and findings on broad brain communication networks [6–8] . Notwithstanding the introduction of approaches with increasing resolution to probe structure-function relationships in recent years , little is known about the efficacy of image-based tracing methods to capture the wealth of existing connections in humans [9–12] . This drawback is particularly acute in depicting even strong pathways to and from small subcortical nuclei or distinct cortical layers ( e . g . , [13–19] ) while avoiding emergence of false pathways [20] . By contrast , a wealth of high-resolution information on connections has been amassed in nonhuman primates using invasive neural tracing methods , which have contributed to theories about the organization of connections ( reviewed in [21 , 22] ) . Moreover , numerous studies of the cyto- , myelo- , and receptor-architecture have systematically quantified neocortical laminar patterns and correlated them with gene expression or activity patterns ( e . g . , [5 , 23–26]; reviewed in [27] ) . One key principle that has emerged is that the architectonic differences in the cortex of mammalian species are not random but systematic . Moreover , connections critically depend on the systematic variation in cortical structure ( reviewed in [21] ) . Can we use the rich information from monkeys to understand connections in humans and then examine their disruption in disease ? We addressed this issue using the prefrontal cortex ( PFC ) as a model system in monkeys and humans , because this region is affected disproportionally in psychiatric and neurological diseases . The overall organization of the PFC and associated white matter bundles appear to be largely preserved in primate evolution , rendering nonhuman primates an invaluable animal model for the study of connections in humans [1 , 17 , 19 , 28–33] . Our goal was to compare first the fundamental architecture of distinct prefrontal areas in humans and monkeys and then their axons below the cortex , which make up the highway system for connections . The methods to address this issue in humans and monkeys were identical . We then used detailed connection data from monkeys to determine whether pathways studied through axon features at high resolution are correlated with actual connections , studied with neural tracers . We used the PFC as a model system because it has a large array of areas with distinct laminar structure and functions [34 , 35] , and there are extensive quantitative data on its connections in nonhuman primates ( reviewed in [28 , 36–40] ) . Within the framework of systematic variation in the cortex , we studied 3 functionally and structurally distinct PFC regions: anterior cingulate cortex ( ACC ) , posterior orbitofrontal cortex ( OFC ) , and lateral prefrontal cortex ( LPFC ) , which are broadly associated with attention , emotions , and executive function [41] . We provide evidence for a strong similarity in the architecture of the cortex and white matter axons that participate in short- and long-distance communication in monkeys and humans and a high correlation between axon features and actual connections in monkeys and , by extension , in humans . Against the template of the high correlation of axon features and connections , further analyses revealed a changed trajectory of axons in autism spectrum disorder ( ASD ) , suggesting disruption in both nearby and distant neural communication . Lastly , guided by the rules linking connections with the systematic structural variation of the cortex , we provide a map of connections that can be used in future studies to test hypotheses regarding networks likely affected in ASD . Fig 1 shows the overall experimental design , and Tables 1–3 and S1 and S2 Tables include information on the humans and monkeys used in the present study . We first quantified and compared key cytoarchitectonic features of the PFC in both species , including the laminar density of neurons . The investigated ACC areas 25 and 32 and OFC areas 13 and orbital proisocortex ( OPro ) are dysgranular , an architectonic term applied to areas that have overall poor laminar definition and inconspicuous layer 4 , as shown in Fig 2 . We refer to these areas collectively as limbic , to describe operationally areas that are either dysgranular or agranular ( lacking layer 4 ) . LPFC areas 46 and 8 are eulaminate , a term used for areas that have 6layers , including a clearly visible layer 4 ( Fig 2 ) . These terms apply to all areas of the cerebral cortex , including primary motor area 4 , which has been incorrectly described as “agranular” in the literature ( for discussion , see [42 , 43] ) . We delineated areas based on the detailed descriptions and maps by von Economo , Koskinas , and Sanides [35 , 44 , 45] for the human cerebral cortex and Barbas and Pandya [38] for the rhesus macaque PFC . For this study , we avoided regions near the borders of areas , as described in recent detailed studies on the architecture of PFC regions [46–49] , and relied on multiple salient features used in classic and modern architectonic studies to determine what constitutes an architectonic area , despite focal architectonic variations seen particularly across large areas . A good example of this approach is the characterization and mapping of area 25 in the rhesus macaque , which was based on cyto- and myeloarchitecture , as well as the distribution of neurofilament markers ( SMI-32 ) and calcium-binding proteins [50] . The position of area 25 in the posterior , ventromedial ( subgenual gyrus ) , and orbital surface of the PFC is comparable in rhesus monkeys and humans ( [51] , mainly areas FL and FH in the human cortical atlas by von Economo and Koskinas [44] ) . There is a gradual increase in the laminar elaboration of area 25 along the posterior to anterior and medial to lateral axes such that its most posterior and medial portions are agranular , whereas the anterior and lateral segments are dysgranular [35 , 44 , 50–52] . The salient and most distinguishing feature of area 25 across its entire extent is the increased thickness and neuronal density of deep layers 5 and 6 and the very low density of myelinated axons in both primate species [35 , 50] . Here we focused on the dysgranular portion of area 25 for analysis . Gradual changes in the laminar architecture of the cortex have also been described in the OFC , where layer 4 is sparser in area OPro than in the rostrally situated area 13 , which is also dysgranular [44 , 53–56] , and in the LPFC , where eulaminate area 8 is progressively more granular than the anteriorly situated area 46 , which is also eulaminate [29 , 30 , 38 , 44 , 57–59] . Quantitative comparison of cytoarchitectural features showed that despite differences between humans and monkeys , the differences between limbic and eulaminate PFC areas followed similar trends in the 2 species . Human PFC areas had overall lower neuron density compared to rhesus macaques ( Figs 2 and 3 ) , but the relationship among areas was the same ( Fig 3 ) . Most neurons within a unit volume of human PFC ( relative density ) were found in layer 3 , and neurons were most densely packed in layers 2 and 4 , whereas in rhesus macaques neuron distribution and relative density were more balanced among layers , with the highest packing density seen in layer 4 , as in the human ( Fig 4 ) . Pyramidal projection neurons in layers 3 and 5 were overall larger in humans . Limbic and eulaminate PFC cortices with their distinct cytoarchitecture had differences in the density of neurons that followed similar trends across prefrontal areas in monkeys and humans ( Figs 3 and 4 ) . As shown in Fig 4 , ACC cortices , and especially area 25 , had relatively more neurons in deep layers 5 and 6 , whereas LPFC had more neurons in the superficial layers . Fig 4 also shows that the packing density ( neurons/layer volume in mm3 ) followed the same trend across areas in monkeys and humans . LPFC areas stand out by having 2 peaks in the density of neurons centered on layers 2 and 4 , consistent with qualitative classic studies and earlier quantitative studies [34 , 38 , 45] . The OFC had an intermediate density of neurons , about equally distributed in the infragranular and supragranular layers . The relative size of pyramidal projection neurons in layers 3 and 5 was also different among PFC areas . A characteristic feature of ACC was the presence of relatively large pyramidal neurons in layer 5 , whereas in LPFC the largest neurons were found at the bottom of layer 3 ( Fig 2 ) . Once again , OFC had a more balanced appearance with slightly larger neurons in infragranular layer 5 . We then studied myelin in the gray matter , which coats a large number of axons in the primate cortex . First used in classic studies , myelin is an extremely useful feature for delineating architectonic areas , including the often misunderstood primary motor cortex ( M1 ) in primates , which is heavily myelinated , as are other eulaminate areas [43] . In addition , myelin may be used to detect differences across cortical regions in the living human brain by imaging [60] . We also studied the density of oligodendrocytes , which myelinate axons . The least myelinated areas were within the ACC , and the most myelinated areas were within LPFC , with OFC areas showing a pattern between the 2 extremes ( Figs 5 and 6 ) . Quantitative analysis showed that myelin density was highly and positively correlated with the laminar density of oligodendroglia in both species ( Fig 6; R2 = 0 . 56 ) , consistent with the role of oligodendroglia in myelinating axons . There were no significant differences in the myeloarchitecture of PFC in monkeys and humans , so the following description applies to both . Optical density measurements of the gray matter level index showed that myelin was relatively high in layer 1 and the transition from layer 5 to layer 6 and to the white matter in all areas examined ( Fig 6 ) . ACC area 25 had overall the lowest levels of myelin staining among the areas examined . Human PFC areas had overall higher oligodendrocyte density compared to rhesus macaques ( Figs 3 and 6 ) . The above analyses revealed that humans and rhesus monkeys show common trends in the cytoarchitecture and myeloarchitecture of the gray matter across prefrontal areas . We then addressed whether this pattern extends to the architecture of axons beneath the areas studied in the 2 species . The significance of this analysis is based on the fact that axons make up pathways that connect cortical areas . We thus systematically examined the density and thickness of individual myelinated axons in ultrathin sections of the white matter beneath ACC , OFC , and LPFC at very high resolution with an electron microscope ( EM ) . Myelinated axons made up about 50% of the white matter beneath the areas studied ( range: 35%–60% ) . The remaining space was taken up primarily by unmyelinated axons and glia , especially oligodendrocytes . As shown in Fig 7 , there was a striking positive correlation between the density of neurons in the overlying gray matter and the density of myelinated axons in the white matter . Significantly , in both humans and monkeys , we found a gradually increasing trend from ACC to OFC to LPFC in the density of myelinated axons in the white matter and neurons in the gray matter ( R2 = 0 . 8 , monkey; R2 = 0 . 96 , human ) . We next probed additional axon features in the white matter , in view of the fact that axons vary in thickness , a feature that affects axon dynamics , including the well-known differences in conduction velocity of neural impulses [61 , 62] . Axon diameters below prefrontal areas ranged between 0 . 1–7 μm , in line with previous studies in human and macaque cortex [63–66] . The thickness of axons overlapped extensively but extended further within the large end of the spectrum in the human PFC ( 0 . 1–7 μm ) than in rhesus macaques ( 0 . 1–5 μm; Fig 8A–8D ) . The finding of thicker axons in humans is consistent with the greater distances that axons must travel in the larger human brain , as well as the size of neurons in the 2 species . As shown in Fig 8E , there was a positive linear correlation between the thickness of axons and the thickness of their myelin sheath . This finding is consistent with the classic relationship of the inner to outer diameter of axons , known as the g-ratio , which is an indicator of the efficiency of conduction velocity and neurotransmission . In all areas , the g-ratio increased significantly with axon size . In both species , the average g-ratio of axons was near the optimal average value ( approximately 0 . 6 ) . We then used cluster analysis to segregate axons into 4 groups by thickness ( Fig 8C and 8D ) , based on their outer diameters , which included the myelin sheath . Cluster analysis was conducted separately for monkeys and humans because of the differences in the thickness of axons between the 2 species . The cluster analysis separated axons in macaque monkeys into thin ( 0 . 1–0 . 49 μm ) , medium ( 0 . 5–0 . 86 μ m ) , thick ( 0 . 87–1 . 4 μm ) , and extra-large ( larger than 1 . 4 μm ) axons . In humans , cluster analysis similarly grouped axons into thin ( 0 . 1–0 . 83 μm ) , medium ( 0 . 84–1 . 51 μm ) , thick ( 1 . 52–2 . 65 μm ) , and extra-large ( larger than 2 . 65 μm ) . In both humans and monkeys , thin and medium axons were the most numerous . Thick or extra-large axons constituted fewer than 30% of all axons beneath all PFC areas ( Fig 9 ) . The relative position of axons within the white matter and overall diameter are also indicators of their termination in nearby or distant brain areas . Thus , thin axons in the outer white matter , near cortical layer 6 , link nearby areas . This is consistent with 2 structural principles . First , as axons approximate their destination to innervate the cortex , they split into thinner terminal axons , akin to the thickness of the human forearm in comparison with the fingers . Second , fibers that link nearby areas follow the shortest route to their cortical destination within the superficial white matter , consistent with the principle of economy of wiring [67 , 68] . By extension , thicker axons dive through the deep white matter as they travel to farther destinations [61 , 69] . We used all structural parameters of axons to create unique fingerprints of the white matter in humans and monkeys for the 3 prefrontal regions , and the results are seen in Fig 9 . The superficial white matter ( SWM ) , which extended about 2 mm below layer 6 , had relatively more thin axons compared to the deep white matter ( DWM ) , which had relatively thicker axons in all areas and in both species . As seen in the fingerprint diagrams , the similarities between the 2 species outnumber the differences , which were restricted to the outer diameter of axons in all areas , as well as myelin thickness across all areas except for LPFC ( Fig 9; asterisks show statistically significant differences between humans and monkeys ) . Another species difference was in the SWM below ACC , which had a higher proportion of thin axons in rhesus monkeys than in humans ( Fig 9A , lower asterisk ) . In both species , the ACC and , to a lesser extent , the OFC had relatively more medium and thin axons compared to LPFC , especially in the superficial parts of the white matter , which contain mainly short- and medium-range pathways ( Fig 9 ) . In contrast , below LPFC there were relatively more thick to extra-large axons , especially in the DWM , which contains long-distance pathways ( Fig 9 ) . It should be noted that the SWM contains some thick axons as well , since pathways destined to travel over long distances must pass through the SWM in order to enter the DWM . The SWM also contains U-shaped fibers that connect neighboring gyri [28 , 70] , though these appear to be sparser than previously thought [71] . To assess interspecies and interareal similarities and differences , we used hierarchical cluster analysis and subsequent multidimensional scaling of 34 features derived from brightfield analysis and 12 feature dimensions from EM analysis ( see S1 Data for data features used ) . This analysis made it possible to project high-dimensional data into a 2-dimensional space . The distance between data points reflects the similarity/dissimilarity of areas . Analyses revealed a clear separation of ACC , OFC , and LPFC areas in each species , based on cellular and axon features of the gray and white cortical matter . The separation of regions was strikingly similar in the 2 species ( Fig 10A and 10B ) . The stress was low , indicating that the low-dimensional representation accurately captured the relationships between areas . The nonmetric multidimensional scaling ( NMDS ) analysis shows that for both species , there is a gradient from dysgranular to eulaminate cortices . Trends for the monkey and human regions were parallel but separated , indicating that the feature set also reflects species-level differences . We also examined whether feature sets of each region between species were correlated . Because the estimated variables in the brightfield and EM feature sets of each region included values that varied considerably in scale ( e . g . , neuron density in tens or hundreds of thousands and myelin optical density from 0 to 255 ) , we log-transformed the data and then regressed the monkey feature set onto the human feature set for each area . The R-squared values were high , indicating high correlation between the monkey and human feature sets for each PFC region for architectonic gray matter data obtained using brightfield microscopy ( Fig 10C–10E ) and white matter data obtained with EM ( Fig 10F–10K ) . The above analysis established that rhesus monkeys and humans show similar trends in the architecture of axons below distinct prefrontal regions . We then investigated whether axon features can be used to infer connectivity . This was accomplished by comparing the high-resolution EM data on the relative ratio of thin and thick white matter axons in humans and monkeys with connectivity data for ACC , OFC , and LPFC areas in monkeys . Monosynaptic connections were studied at high resolution using tract tracing in monkeys . We used available data on projection neurons throughout the cortex that are directed to ACC , OFC , and LPFC areas in rhesus monkeys to compare directly with axon features . We expressed labeled projection neurons as the relative ratio of short/medium-range versus long-range cortical connections ( N = 11 animals , 12 tracer injections; 6 , female ) , based on the distances of interconnected areas . Short/medium-range cortical connections of ACC , OFC , and LPFC were restricted within the frontal lobe; long-range axons linked prefrontal areas with temporal , parietal , and occipital cortices; the results are shown in Fig 11 . All areas had overall more local ( within the frontal lobe ) than distant connections . However , the relative proportion of short-range versus long-range connections for each region was significantly different such that LPFC made more long-range connections than ACC ( Fig 11C and 11E ) . As in the other analyses ( above ) , OFC showed intermediate features with values between the other 2 PFC regions ( not shown ) . These relative ratios closely resembled the EM data that included all axons in the white matter below ACC , OFC , and LPFC , placing thin axons in short/medium-range pathways and thick axons in long-range pathways ( Fig 11B and 11C ) . Our findings of systematic trends in the architecture of the cortex confirm and extend classical architectonic studies and , importantly , provide novel findings on the architecture of the white matter beneath the cortex . An important principle is that the architectonic differences in the cortex are systematic and connections are predicated on the relationship of the laminar structure ( cortical type ) of the linked areas , according to the structural model for connections ( reviewed in [21] ) . Moreover , the principle of the relationship of cortical type and connections transcends the model of cortical connectivity based on the distance between areas [4 , 72] , as illustrated by the fact that some distant areas of similar cortical type are strongly interconnected [50 , 73–75] . We thus expressed the relative number of labeled projection neurons from the above analysis as the relative ratio of connections of areas with similar or different structural types . We grouped cortical areas into 3 major , well-established structural types that are widely and consistently used in the literature: agranular , dysgranular , and eulaminate ( for reviews , see [76 , 77] ) . Results showed that the limbic ACC and posterior OFC areas were primarily connected with other dysgranular cortices , whereas LPFC areas were primarily connected with other eulaminate areas ( Fig 11D ) . The relative ratios of connections grouped by structural type closely resembled the EM data that included all axons in the white matter below PFC areas , placing thinner axons in pathways that connect mostly neighboring areas of similar type ( Fig 11B and 11D ) ; this finding is consistent with the principle of systematic variation of areas across the cortical mantle ( reviewed in [21] ) . The relationship of thin and thick axons in the white matter of human PFC showed the same trend as in monkeys , with overwhelming predominance of thin compared to thick axons . As was the case in rhesus macaques , human ACC had relatively more thin axons compared to LPFC ( Fig 11A ) , while OFC showed intermediate levels . We then used the rich tract-tracing connectivity database in rhesus macaques to quantitatively map the strength of ACC , OFC , and LPFC connections in relation to the distance between linked areas , in order to identify short- and long-distance connections , and classify them also by the structural type of interconnected cortices ( Fig 11E–11H ) . This analysis revealed that areas within all 3 PFC regions were primarily connected with nearby cortices of similar type . In addition , LPFC , in particular , had relatively more distant connections with areas of similar type , compared to ACC . The findings in ACC were of particular interest in view of our previous data , which showed that individuals with ASD had significantly more thin axons in the SWM and fewer thick axons in the DWM below ACC [64] . We used the expanded dataset of neurotypical-control subjects in this study ( NCTR = 6 , Table 3 shows subjects used in EM study of white matter ) , which contained axons from the portion of area 32 anterior to the corpus callosum that precisely matched the ACC white matter region studied earlier [64] , and compared axon features in the brain of adults with ASD ( NASD = 5 , Table 3 ) . For this analysis we used the outer diameter of axons , which takes into account the myelin sheath , and parcellated the population of thin and thick axons based on the cluster analysis from the expanded control dataset . In agreement with our previous report , the brains from adults with ASD had significantly more thin axons and fewer thick axons in the white matter below ACC area 32 ( Fig 12A ) . The addition of control subjects in the expanded dataset increased the power of the analysis and showed that the differences between control and ASD subjects were pronounced . Moreover , the changes in white matter axons in ASD could be reliably detected within the entire population of axons and throughout the entire white matter below ACC ( Fig 12A and 12B ) . Further , the extended dataset of control subjects additionally revealed , for the first time ( to our knowledge ) , similar changes in the SWM below dorsal LPFC area 46 . Thus , area 46 also had significantly more thin axons and fewer thick axons in adults with ASD compared to neurotypical controls ( Fig 12C and 12D ) . It should be noted that some of the additional control subjects we used in this study were older ( 58–67 years ) than other control and ASD subjects ( 30–50 years , Table 1 ) . This is important because structural changes in the white matter with aging have been reported—in particular , decreases in axon density and the thickness of the myelin sheath in major pathways of the monkey ( reviewed in [78 , 79] ) and human cortex ( reviewed in [80–83] ) . To assess potential effects of age on estimates of axon size and density , we examined the correlation of all estimated variables between subjects within control and ASD groups , using multivariate analysis of covariance ( MANCOVA ) with age as a covariate . Age had no effect on the results . Estimated variables from older control subjects were well within the range of values from younger control subjects ( see S1 Data ) . The organization of the PFC appears to be largely preserved in primate evolution , rendering nonhuman primates an invaluable animal-model to study connections in humans [1 , 17 , 19 , 28–33 , 84] . The frontal lobe , in particular , has expanded significantly in humans compared to nonhuman primates [1 , 31 , 85–89] , accompanied by more and larger neurons , more synapses , and greater complexity in networks [84 , 90–95] . Indeed , a notable architectonic difference in PFC between the 2 primate species was the lower neuron density found in the human PFC , leaving more space available for neuronal processes and synapses [45] . Our analyses suggest that the human PFC and , in particular , its limbic components are endowed with high plasticity as well as vulnerability in neurological and psychiatric diseases [96–98] . The human PFC white matter had overall thicker axons than in monkeys , consistent with the longer distances that pathways must traverse in a larger brain , as also observed in other comparative studies [1 , 65 , 89] . The overall enlargement of axons and the thickness of the myelin sheath in human PFC is consistent with the increased density of oligodendroglia [32 , 99 , 100] and helps explain the increase in the glia/neuron ratio in human cortex compared to nonhuman primates [90] . These findings thus reveal that in addition to gray matter architecture [5 , 24 , 27 , 34 , 53 , 60 , 64] , we can use white matter features to construct unique fingerprints of cortical areas and facilitate their distinction in primates . The relationships among the 3 PFC regions in humans and monkeys followed remarkably similar patterns . Thus , overall density of neurons , oligodendrocytes , and intracortical myelin increased progressively from the limbic ACC and OFC to eulaminate LPFC areas [34 , 35 , 45 , 64 , 90 , 96 , 101] . The gradual increase in laminar elaboration and density of neurons from limbic to eulaminate PFC areas was marked by a parallel increase in the density of myelinated axons in the white matter [32] . In both species , the ACC and the OFC , 2 limbic regions , had relatively more thin and fewer thick axons compared to the LPFC . Parallel analysis of the rhesus macaque connectome showed that ACC and OFC also had denser short/medium-range pathways compared to LPFC , which also had relatively dense long-distance connections . Myelination can accelerate impulse conduction and regulate the timing of communication among local or distant connections [62] and the synchronization of functionally related areas [65 , 102–104] . Our findings are thus consistent with a role of limbic ACC areas in networks that process signals at relatively slow speeds , such as pain [105 , 106] . On the other hand , LPFC areas maintain strong connections with select distant sensory and other association areas implicated in cognitive functions . Our findings thus showed a similar architecture in the gray and white matter of human and rhesus monkey ACC , OFC , and LPFC . Moreover , differences in the architecture of the gray and white matter varied systematically and in the same direction in both species . In addition , the gray and white matter architecture accurately reflected the connectivity in macaque monkeys and , by inference , in humans . These findings are rooted in the general principle of systematic variation in the cortex . This principle is based on strong evidence that architectonic variations in the cortex are not random but systematic [45] . Further , corticocortical connections mirror the systematic architectonic variation , as formulated in the “structural model for connections” [107 , 108] . According to this model , connections between areas are biased toward a “feedforward” or a “feedback” pattern depending on the ( dis ) similarities of the architecture between linked areas . The bigger the architectonic differences , the bigger the bias . For example , dysgranular limbic areas , like those of the ACC and the posterior OFC , which have less delineated layers and lower cell density , send mainly feedback projections to eulaminate areas , like those of the LPFC , which have 6 well-defined layers and higher neuron density . These projections originate mainly from the deep layers of ACC or OFC and terminate mostly in the superficial layers of the LPFC [38 , 108 , 109] . Projections in the opposite direction , from LPFC to ACC , are feedforward: they originate mostly from the superficial layers and terminate in the middle/deep layers of ACC [38 , 108] . Connections between areas with similar laminar structure are columnar: they originate in most layers and terminate in all layers , as seen in the connections linking dysgranular ACC and OFC areas [56] . Over the past 20 years , studies have consistently supported this model for ipsilateral and callosal connections among diverse cortices in nonhuman primates ( e . g . , [57 , 73 , 75 , 107 , 108 , 110–112] ) and other species [113 , 114] . The relative distribution of neurons in different layers in areas that differ in laminar structure ( cortical type ) is consistent with their predominant projection pattern . For example , the limbic ACC areas have relatively higher density of neurons in the deep layers , their main output layers , and neurons in layer 5 of ACC are bigger than in other layers . In contrast , eulaminate LPFC areas have denser layers 2 and 3 , and their biggest neurons are found in layer 3 , their predominant output layer to other cortices [107 , 108] . The present findings further showed that the principle of systematic variation is also embedded in the fine structure and density of myelinated axons in the white matter . The white matter below PFC thus contained mostly thin myelinated axons and few thick axons , consistent with their connections being primarily with neighboring frontal areas [1 , 64 , 65 , 115] , in line with previous studies [65 , 116] . This finding is also consistent with the observation that nearby areas tend to be strongly interconnected , forming “rich club” hubs that attract and disperse communication paths [67 , 117] . Further , our results revealed that the relative ratio of thin to thick axons in the white matter below primate PFC ( and by extension connections ) is associated with a more fundamental principle than proximity—namely , cortical type . This principle most parsimoniously explains why both nearby as well as distant areas can be strongly connected [50 , 57 , 73–75 , 108 , 118] . Consequently , areas that are robustly interconnected , whether they are neighbors or not , are more likely to have comparable overall laminar structure . By probing the organization of the components of the white matter , our findings now also differentiate between short/medium-range from long-distance pathways . Short/medium-range pathways connect nearby areas , as verified in the connections in monkeys and supported by the prevalence of thin myelinated axons in the SWM . On the other hand , long-range pathways that must travel in the DWM contain thicker axons , as shown by the preponderance of thick axons below LPFC , in accord with its strong connections with distant parietal , temporal , and occipital areas . These findings are in line with recent imaging studies that use MRI and diffusion-weighted tractography methods to map the human and monkey cortex based on the relationship of whole-brain white matter connectivity with macroscopic structural features , like T1- and T2-based intracortical myelin content , or the correlated activity of areas during rest , unimodal , and transmodal functions [3 , 5 , 24 , 119–121] . The regularity and similarity in the structure of the gray and white matter in the PFC of primates suggests that architectonic information can be used to predict the strength and laminar pattern of connections in humans . While invasive procedures for the detailed study of connections are precluded in humans , the architecture can be studied in postmortem brain tissue [63 , 64] , as in this study . Based on the similarities and functional data for ACC , OFC , and LPFC areas [122–124] , we predict that these cortical regions will have a similar pattern of connections in humans as in rhesus macaques [39 , 56 , 107 , 109] , which awaits further test in future studies . Our findings have significant implications for the status of the architecture and , consequently , for connections in neurological and psychiatric disorders and suggest that parallel gray and white matter changes may reflect common pathological mechanisms , in line with other imaging and neuropathological studies [63 , 125–128] . The PFC , in particular , is consistently affected in ASD , along with the processes of attention , social interactions , emotions , and executive control [36 , 97 , 129–132] . Prefrontal pathways are structurally and functionally disorganized in autism , exhibiting local overconnectivity and long-distance disconnection [63 , 64 , 133–141] . Our findings of more thin than thick axons in the PFC of individuals with ASD support the hypothesis that PFC is “talking to itself” in autism [132] . The disruption appears to be particularly extensive below ACC , with an exuberance of thin axons ( typically found in short-range pathways ) and a decrease in thick axons ( typically found in long-range pathways ) . Our analyses of the architectonic profile of the primate PFC and connectome delineate the preponderant pathways linking PFC with other areas ( Fig 12E ) . In the case of dysgranular ACC areas , these networks include strong columnar connections with nearby medial PFC and OFC , as well as relatively weaker long-range connections with similarly dysgranular rhinal , temporal , posterior cingulate , and prostriate visual cortices [14 , 50 , 56 , 73–75 , 111 , 142 , 143] . In the case of eulaminate LPFC , these networks include strong columnar connections with nearby PFC , as well as moderate/strong long-range connections with eulaminate lateral parietal , superior temporal , and occipital cortices [57 , 73 , 144] . On the other hand , in connections between structurally dissimilar cortices , our results highlight the potential significance of robust short-range connections between dysgranular ACC and eulaminate LPFC areas [38 , 145 , 146] . In rhesus macaques , for example , excitatory axons from ACC send feedback pathways that mostly target the superficial layers of LPFC . When these pathways form synapses with inhibitory neurons in LPFC , they innervate preferentially inhibitory neurons that have modulatory effects on pyramidal neurons [109] and likely increase the signal-to-noise ratio and facilitate focused attention on a task [147] . The exuberance of thin axons in the white matter below ACC and the SWM below LPFC in ASD could underlie behavioral challenges typical of autism , like excessive focus and inability to shift attention when necessary . Our findings support a potential disruption of mechanisms that rely on a fine balance of excitation and inhibition in the PFC and are consistent with atypical ACC and LPFC activation in ASD [135 , 148 , 149] , including desynchronized and reduced activity during working memory tasks ( [137 , 150]; reviewed in [151] ) . In conclusion , our findings provide the basis to relate data from invasive neuroanatomical tract-tracing studies in a nonhuman primate model , the rhesus monkey , to neuroanatomical , histopathological , or noninvasive imaging studies in humans . This analysis is a prerequisite for understanding brain function and disruption in neuropathology and for identification of potential regions for intervention in disease . Our approach establishes a framework of structural principles based on detailed high-resolution quantitative findings that complement a large body of data on the cortical architecture in human [35 , 45 , 64] and nonhuman primates [34 , 38 , 107 , 108] and connections of prefrontal areas [14 , 57 , 75 , 107 , 108 , 110] . Application of these principles makes it possible to study and model functionally relevant cortical circuits at an exquisite level of refinement in humans , as exemplified by the study of prefrontal networks , and their consistent disruption in disorders like ASD . We used prefrontal postmortem brain tissue from neurotypical adult humans ( N = 8 , female: 3 ) and rhesus monkeys ( Macaca mulatta; N = 25 , female: 12; Tables 1–3 ) . We quantitatively studied the cyto- and myeloarchitecture of the gray and white matter of ACC ( areas 25 and 32 ) , OFC ( areas 13 and OPro ) , and LPFC ( areas 46 and 8 ) , at high resolution , using brightfield and electron microscopy ( Fig 1 ) . We then quantitatively characterized cellular and axonal structures and densities to distinguish PFC areas within and across species . We used these data to compare and correlate structural features of short- and long-range pathways between areas of similar or different structure with connectivity data from quantitative tract-tracing studies in rhesus macaques . Finally , using this framework that links the structure of cortical areas with their connections , we compared PFC white matter organization and connectivity between neurotypical adults ( N = 6 , female: 3 ) and individuals with ASD ( N = 5 , female: 1 ) , to study short- and long-range ACC and LPFC pathways with structurally similar or different cortices . Table 3 shows the human subjects and rhesus monkeys used in each type of experiment and analysis . Data of individual human subjects and rhesus monkeys are included in the Supporting Information ( S1 Data ) . We used coronal PFC sections from formalin-fixed postmortem brain tissue of neurotypical adults and individuals with ASD . Brain tissue was obtained from the Harvard Brain Tissue Resource Center through the Autism Tissue Program and Anatomy Gifts Registry . The study was approved by the Institutional Review Board of Boston University . Human subjects were matched as closely as possible based on tissue availability . Brain tissue from all 8 neurotypical ( control ) subjects was processed and used for quantitative light microscopic analysis . Brain tissue from 6 of these subjects ( 3 female subjects ) was processed and used for quantitative electron microscopic analysis ( Table 3 ) . Two control subjects ( HBJ and HBK ) were not used for EM because of suboptimal tissue quality . The diagnosis of autism was based on the Autism Diagnostic Interview–Revised ( ADI-R ) . Some subjects from individuals with ASD were diagnosed with seizure disorder ( subject AN 08792 ) , depression ( subject AN 18892 ) , and schizophrenia ( subject AN 06746 ) . Results from the analysis of the features of axons in these and the female subjects ( HAW , B-5353 , B-6004 , and AN-07770 ) did not differ from other subjects within each group , in this and other studies that used tissue from the same subjects [64 , 152 , 153] . Clinical characteristics , including ADI-R scores , and other data of human subjects and the experiments they were used for are summarized in Tables 1 and 3 and S1 Table . Coronal PFC blocks were matched ( Fig 1A ) based on human brain atlases [44 , 154] and additional cytoarchitectonic studies of human PFC [35 , 45 , 59] . We postfixed tissue in 2% paraformaldehyde and 2 . 5% glutaraldehyde , in 0 . 1 M phosphate buffer ( PB , pH: 7 . 4 ) for 2–4 days at 4 °C . To preserve the ultrastructure until processing , tissue blocks were cryoprotected in 25% sucrose solution and then immersed in antifreeze solution ( 30% ethylene glycol , 30% glycerol , 40% 0 . 05 M PB , pH: 7 . 4 with 0 . 05% azide ) and stored at −20 °C . Tissue blocks were then rinsed in 0 . 1 M PB and cut coronally in 50-μm-thick sections on a vibratome ( Pelco , series 1000 ) or frozen in −70 °C isopentane and cut in a cryostat ( CM 1500 , Leica ) in the coronal plane at 20–50 μm in 10 series of free-floating sections . Sections used for histological stains were mounted on chrome-alum gelatin–coated slides . We used archival postmortem rhesus monkey ( M . mulatta ) brain tissue for architectonic and pathway analyses ( Tables 2 and 3 and S2 Table ) . We used 12 tracer injections and data plots from 11 rhesus monkeys to study the cortical connections of ACC ( N = 4 ) , OFC ( N = 4 ) , and LPFC ( N = 4 ) through tract tracing . Tracer injection sites and quantification of projection neurons have been described in detail in previous studies and will be briefly described here ( [14 , 56 , 57 , 73 , 75 , 108–112 , 143 , 155–161]; see S2 Table for a detailed list of relevant quantitative studies ) . The available rhesus monkeys were young adult animals , obtained from the New England Primate Research Center . Procedures were designed to minimize animal suffering and reduce the number of animals used . Detailed protocols of the procedures were approved by the Institutional Animal Care and Use Committee at Harvard Medical School and Boston University School of Medicine in accordance with NIH guidelines ( DHEW Publication no . [NIH] 80–22 , revised 1996 , Office of Science and Health Reports , DRR/NIH , Bethesda , Maryland , United States ) . Table 3 summarizes the data of rhesus monkeys and their specific use here . Imaging for injection of tracers , surgery , perfusion , and tissue preparation were described previously [13 , 14 , 16] . The retrograde tracers injected in the ACC , the OFC , and the LPFC included biotinylated dextran amine ( BDA , 10% solution , Invitrogen , Carlsbad , California , US ) , Fast Blue ( FB , 1% solution; Sigma , St . Louis , Missouri , US ) , Diamidino Yellow ( DY , 3% solution; Sigma ) , and horseradish peroxidase conjugated to wheat germ agglutinin ( HRP-WGA , 8% solution , Sigma ) . After removal from the skull , all brains were photographed , cryoprotected in a series of sucrose solutions ( 10%–30% in 0 . 01 M PBS ) , and frozen in −70 °C isopentane ( Fisher Scientific , Pittsburg , Pennsylvania , US ) for rapid and uniform freezing . Brains were cut in the coronal plane on a freezing microtome at 40 or 50 μm to produce 10 matched series . In animals with injection of fluorescent tracers , 1 series was mounted on glass slides , coverslipped , and used to map labeled neurons . Series of sections of monkey and human PFC were mounted on gelatin-coated slides ( Gelatin Type A , G8-500 , Fisher Scientific , Fair Lawn , New Jersey , US ) and stained for Nissl using thionin blue ( Thionin powder , T-409 , Fisher Chemicals ) to view neurons and glia and examine the cytoarchitecture of each area , as described [162] . Briefly , sections were dried , defatted in a 1:1 solution of chloroform ( C298-1 , Fisher Scientific ) and 100% ethanol ( Pharmco-AAPER , Brookfield , Connecticut , US ) for 1 to 3 hours , rehydrated through a series of graded alcohols and dH2O , stained with 0 . 05% thionin ( pH 4 . 5 ) for 15 minutes , differentiated through graded alcohols , cleared with xylenes ( UN1307 , Fischer Scientific ) , and coverslipped with Entellan ( Merck , Whitehouse , New Jersey , US ) . Other series of sections mounted on gelatin-coated slides were stained using the Gallyas silver technique to label intracortical myelin [163 , 164] . We postfixed human and monkey tissue blocks or sections in 2%–4% paraformaldehyde and 2 . 5%–3% glutaraldehyde , in 0 . 1 M PB , pH 7 . 4 , for 2 days at 4 °C . To preserve the ultrastructure until processing , tissue blocks and sections were immersed in antifreeze solution ( 30% ethylene glycol , 30% glycerol , 40% 0 . 05 m PB , pH 7 . 4 , with 0 . 05% azide ) and stored at −20 °C . Blocks were rinsed in 0 . 1 M PB and cut coronally in 50-μm-thick sections on a vibratome ( series 1000 , Pelco ) or on a freezing microtome . Sections were rinsed in 0 . 1 M PB and postfixed in a variable wattage microwave oven ( Biowave , Pelco ) with 6% glutaraldehyde at 150 W . Small regions of sections containing the superficial or deep parts of the white matter below the PFC were cut under a dissecting microscope . We confirmed the presence of ACC , OFC , and LPFC cortical regions of interest in adjacent sections that were stained with Nissl . White matter regions of interest were postfixed in 1% osmium tetroxide with 1 . 5% potassium ferrocyanide in PB , washed in buffer ( PB ) and water , and dehydrated in an ascending series of alcohols . While in 70% alcohol , they were stained with 1% uranyl acetate for 30 minutes . Tissue sections were then cleared in propylene oxide and embedded in araldite or LX112 at 60 °C . Serial ultrathin sections ( 50 nm ) were cut in the horizontal plane with a diamond knife ( Diatome , Fort Washington , Pennsylvania , US ) using an ultramicrotome ( Ultracut; Leica , Wein , Austria ) and collected on single slot grids to view with a transmission EM ( 100CX; Jeol , Peabody , Massachusetts , US ) or a scanning EM ( Zeiss Gemini 300 with STEM detector and Atlas 5 software modules ) , as described [64 , 162] . Myelinated axons were easily identified at the EM by the darkly stained electron-dense myelin sheath [165] . We subdivided white matter beneath ACC , OFC and LPFC in 2 regions , by determining axon alignment in serial sections under the microscope , at gradually increasing distances from the gray-white matter border , as described [63 , 64] . The outer , gyral , or superficial part ( SWM ) , which mainly consists of axons that participate in short-range pathways , was immediately adjacent to layer 6 of the overlying cortical areas . The SWM , which was segmented based on the predominant radial alignment of axons , had a thickness of 0 . 5–2 . 5 mm . However , some axons in the SWM either had variable trajectories or followed the curvature of the overlying gray matter , resembling the likely trajectory of U-shaped fibers that connect neighboring gyri [28 , 70] . The inner or deep part of the white matter ( DWM ) , which mainly consists of axons participating in long-distance pathways , was segmented based on the predominance of axons that run mainly sagittally to the cerebral surface . To estimate the sample size , we took into account the number of human subjects and rhesus monkeys and the volume fraction of areas sampled so that the number of individual cells and axons examined produced estimates with a small coefficient of error ( <10% ) , as described [16 , 64 , 96 , 163] . Pilot studies with exhaustive sampling , progressive means analysis , and the formula of West et al . , [166] , as well as a posteriori power analysis , using data from our previous studies of cell and axon densities , or tract tracing of pathways , took into consideration all known and estimated variables , including age , sex , postmortem interval ( PMI ) , and other diagnoses . These considerations showed that the sampling ratios used exceeded the samples needed to detect differences with a greater than 90% probability and with an estimated large effect size in the population ( 0 . 80 ) . We estimated the overall and laminar density of neurons and oligodendrocytes in representative columns along the depth of a straight portion of the gyral part of anterior cingulate areas 25 and 32 , orbital areas 13 and OPro , and lateral area 46 ( Fig 1 ) , based on previous maps for rhesus macaques [38] and humans [44 , 59 , 154] . We examined OFC area 13 in human PFC and the structurally similar OFC area OPro in rhesus macaques . We used the unbiased stereological method of the optical fractionator [167 , 168] with the aid of commercial software ( StereoInvestigator; Microbrightfield , Williston , Vermont , US ) , as described [64 , 96 , 163] . We used a minimum of 3 sections from 1 series of coronal sections from each human subject and rhesus monkey and drew contours of layers in each column . We counted Nissl-stained neurons or oligodendrocytes at 1000× , using systematic random sampling . We identified neurons and oligodendrocytes based on their characteristic features , following a detailed cytology algorithm , as we have described [[162]; Fig 13] . Briefly , we first split labeled cell profiles into 2 broad groups . One group included cells with darkly stained nuclei ( microglia and oligodendrocytes ) , and the other group included cells with a lighter nuclear stain ( neurons , astrocytes , and endothelial cells ) . The level of stain tends to be correlated with the size of the nucleus , that is , darkly stained nuclei tend to be smaller than lightly stained nuclei . Once a cell was allocated into 1 of these 2 broad groups , we followed the detailed neurocytology algorithm [162] to distinguish microglia and oligodendrocytes in the darkly stained nucleus group and neurons , astrocytes , and endothelial cells in the other group , using key cytological features . These included the presence or absence of cytoplasm around the nucleus ( present in neurons; absent in glial cell types and endothelial cells ) , the distribution of heterochromatin grains , and the staining of euchromatin in the nucleus . Rounded and darkly stained nuclei with 2–4 granules of heterochromatin , often with a perinuclear halo and/or a pinkish crescent of cytoplasm , are typical of oligodendrocytes; in the human , some oligodendrocytes have clear nuclei with lightly stained euchromatin . Elongated , comma-shaped or polylobular darkly stained nuclei with numerous small granules forming a grid across the nucleus , often with greenish inclusions next to the nucleus ( in the unstained cytoplasm ) , are typical of microglia . Cells with lightly stained nuclei and unstained cytoplasm , with a rim of peripheral heterochromatin under the nuclear envelope and several heterochromatin granules attached to this rim or in the heterochromatin net , were classified as astrocytes . Some astrocytes have yellow inclusions and pinkish threads in the perinuclear cytoplasm . Homogeneous staining of the euchromatin helps make the distinction between astrocytes and endothelial cells . Lightly stained nuclei and stained cytoplasm are the features of neurons . Nuclei with an “empty” appearance and small granules of heterochromatin around a distinct nucleolus are typical of large pyramids as well as of large nonpyramidal neurons , like fusiform von Economo neurons . Lightly stained nuclei with a nucleolus partially or totally surrounded by irregular clumps of heterochromatin and 1–2 additional heterochromatin granules in the nucleoplasm are typical of small neurons . Nuclear folding , present in neurons , additionally helped distinguish small neurons from astrocytes , especially in humans . The counting frame ( disector ) size for cell counts was 50–60 μm . To ensure an unbiased estimate of the number of cells , we first measured the thickness of each section and set a guard zone at the bottom and top of each section to correct for objects plucked during sectioning ( minimum 2 μm in 10–15-μm sections after tissue shrinkage ) ; the disector thickness was thus smaller than the thickness of the section [166–168] . The height of the counting frame was 5 μm , and grid spacing was 100–300 μm . Cells were counted if their nuclei fell within the counting frame or touched the 2 inclusion lines , but not the 2 exclusion lines [168] . These parameters yielded a sampling fraction with a coefficient of error of <10% per contour , as recommended [167 , 168] . The use of uniform random sampling ensured that every part of each area examined had the same chance of being included in the sample . We computed cell density by dividing the estimated number of counted cells with the estimated volume of each contour to assess the relative density ( cells/volume of all layers in mm3 ) and the packing density ( cells/volume of each layer in mm3 ) in each area and each human subject and rhesus monkey . The relative laminar density of neurons is an indicator of the average density of neurons within a cortical gray matter column and shows which layers have more neurons . On the other hand , the packing laminar density of neurons , which was estimated as the average density of neurons in each layer divided by the volume ( mm3 ) of that layer , shows which layers have more densely packed populations of neurons within a cortical gray matter column of set volume . We estimated the density of axons and the thickness of axons and myelin in the white matter at the EM in humans and rhesus monkeys . We sampled a volume of approximately 1 cm3 below each prefrontal cortical area , using a systematic random sampling fraction of 1:1 , 000 that yielded more than 2 , 000 axons , per human subject and rhesus monkey , per area . We divided the white matter ( as described above ) into a superficial part ( close to the gray matter ) and a deep part . We captured high-resolution images of areas of interest that were imported in ImageJ and calibrated . We estimated the overall density of axons at low magnification ( 2 , 000×–3 , 300× ) by dividing the surface area of axon profiles by the total surface area of the sampled region . We estimated the inner and outer diameter as well as the thickness of the surrounding myelin sheath at high magnification ( 10 , 000× ) , as described [64] . In our analysis , we included all axon profiles: those that were perpendicular to the cutting plane and appeared cylindrical , as well as elongated profiles . To ensure consistency , we measured the diameter perpendicular to the center of the maximum diameter of the axon profile . We quantified the laminar myelin content from images of representative columns from each area in the human and monkey PFC captured at the light microscope ( Olympus BX 51 ) , under brightfield , with a CCD camera ( Olympus DP70 ) , connected to a workstation running imaging software ( DP Controller , Olympus ) . We captured images using a UPlanFl 10× /0 . 30 lens with the same light exposure to minimize background variability . We obtained optical density measurements from 5–15 images of cortical gray matter columns taken from at least 3 sections per area in each human subject or rhesus monkey . We imported images into ImageJ , converted them into gray scale , and inverted them . We measured levels of background staining in gray matter regions with no myelinated axons within each image and subtracted background pixel values from each image to eliminate staining inconsistencies , due to experimental variability . These gray matter regions with few or no myelinated axons were typically in the superficial , supragranular layers of the cortex ( layers 1–3 ) in each image of a cortical column . As a result , pixels with strong staining had consistently higher optical density values than the background . We measured the mean gray level density for each normalized image and obtained overall mean gray level values for each area . We also estimated the gray level density along the depth of background-corrected columns , encompassing cortical thickness from the pial surface to the white matter . To normalize gray level density profiles to a standard depth for all cortical regions , we divided each profile into 20 bins and averaged density values across images for each bin . We analyzed an extensive connection dataset that includes quantitative information on the existence or absence and numbers of labeled projection neurons in the cortex of macaque monkeys after injections of retrograde neural tracers in the ACC , OFC , and LPFC . These data were recompiled from databases used in published studies from our laboratory [e . g . , [107 , 108 , 157]; see S2 Table for a detailed list of publications with these data] . Labeled projection neurons were plotted throughout the entire cortex from a series of sections from each animal ( separated by 400–1 , 000 μm ) , using a commercially available semiautomated system ( Neurolucida-Stereoinvestigator; Microbrightfield ) . For earlier studies , we used a custom-built system that included an analog/digital-microscope/computer interface ( Nikon , Optiphot/Austin/i486 PC running DOS ) with plotting software developed in our laboratory [156] . We estimated relative frequencies of connections that were present in each animal by dividing the number of labeled projection neurons in an area by the total number of labeled projection neurons across the entire cortex . Analysis was limited to ipsilateral connections . This procedure normalized connectivity data and minimized variability due to experimental differences across rhesus monkeys ( e . g . , size of tracer injection or tracer transport ) . For the analysis of short- and long-range ipsilateral corticocortical connections of ACC , OFC , and LPFC , we classified ratios of projection neurons based on their spatial adjacency ( distance; S3 Table ) . Connections with areas in the frontal lobe , including prefrontal and premotor areas , were considered short-range connections . All other connections with projection neurons in the temporal , parietal , and occipital lobes were considered to be long-range connections . This classification of connections is largely in line with previous studies that have described or analyzed these and other connections and datasets in primates and other species [4 , 28 , 37 , 72 , 73 , 118] . We also classified cortical areas that innervated ACC , OFC , and LPFC , based on their structural type ( S3 Table ) . This classification is based first on observations that cortical areas tend to connect mainly with other cortical areas of similar structure [74 , 107 , 169] , and second , it is based on the Structural Model of Connections , which states that the laminar structural similarity of cortical areas is related to the laminar origin and termination pattern of connections between them , as well as their relative strength ( e . g . , [21 , 73 , 107 , 108 , 113 , 114 , 118] ) . Classical and recent studies have determined the structural ( dis ) similarity of areas based on qualitative terms or quantitative measurements of several features , including the number and definition of individual cortical layers , assessed in Nissl- or myelin-stained sections , neuronal density , or the density of various receptors [21 , 23 , 27 , 34 , 38 , 44 , 45 , 170] . Laminar definition and the absence or the presence , as well as the thickness , of layer 4 have been useful for distinguishing cytoarchitectonic features . Neurons in cortical layer 4 are small , granular , and densely packed and can be further distinguished by cellular and neurochemical criteria , like the lack of labeling with an antibody for a neurofilament protein ( SMI-32 ) , which labels neurons in layers 3 and 5 [42 , 43 , 162] . Pyramidal neurons in adjacent layers 3 and 5 are also considerably larger and not as densely packed . Using these criteria , we defined 3 major cortical types that have been consistently and widely used in the literature and assigned cortical areas in frontal , temporal , parietal , and occipital lobes in one of these categories ( S3 Table ) . The first structural type of cortex includes agranular areas , identified as those that lack layer 4 altogether . The second type includes dysgranular areas , which have a poorly developed layer 4 . Finally , the third type includes eulaminate cortical areas , which describes the rest of the areas with 6 well-defined layers and laminar elaboration . Areas classified based on these 3 structural types also mostly differ quantitatively in the laminar density of neurons . The density of neurons in association cortices gradually increases from agranular to dysgranular to eulaminate areas with small variations , and agranular and dysgranular ( limbic ) areas have a lower density of neurons in upper layers 2 and 3 compared to deep layers 5 and 6 . We gathered data blind to condition and cortical region in the part of the study involving human postmortem tissue . Random codes for human subjects and images were broken after completion of each part of the study . In most instances , data collection was performed by at least 2 investigators . We employed 1-way ANOVA for overall comparison of cell or optical densities and estimates across species , areas , and laminar groups . For analyses that showed significant differences ( p < 0 . 05 ) , we performed post hoc pair comparisons ( Bonferroni method ) . Data were tabulated in Excel ( Office 365 , Microsoft ) , and analyses were performed using Statistica ( StatSoft , Tulsa , Oklahoma , US; RRID: SCR_014213 ) . For the EM analysis of axon densities and sizes , we obtained samples from widely spaced ultrathin sections ( 1 every 10 ) and fields of view through systematic random sampling to minimize the likelihood of sampling axons from the same parent branch . This sampling scheme and the fact that most axons branch very close to or after they enter the gray matter minimized the likelihood of counting segments of the same axon more than once . We evaluated data through scatter and frequency distribution plots and K-means cluster analysis with parameters set to maximize initial between-cluster distances . Data distributions for continuous variables were not significantly different from normal as determined by the Kolmogorov–Smirnov test and thus allowed the use of parametric statistics . We initially used x2 and Kolmogorov–Smirnov tests to examine axon size distributions and multiple linear regression analysis to examine correlations . To demonstrate global similarities/differences among prefrontal areas , we performed NMDS , which allows visualization of high-dimensional data into a low 2-dimensional space that approximates pairwise distances between data points . We used all estimated variables at the light and electron microscopic level . These included 34 architectural features , like the relative and packing laminar density of neurons , oligodendrocytes , and intracortical myelin , as well as 12 ultrastructural features of the white matter , including relative axon density ( all myelinated axons ) , outer diameter , myelin thickness , g-ratio , and relative proportion of thin , medium , thick , and extra-large axons . The relative proximity among items in an NMDS diagram represents their relative similarity . For each of the 3 brain regions in the 2 species , the values were averaged to produce a feature vector comprised of 34 features derived from brightfield analysis ( neuron and oligodendroglia density all layers , relative and packing laminar density of neurons and oligodendroglia , and relative laminar density of myelin ) . This feature set was z-scored and used to create a distance matrix , which served as the basis for nonmetric multidimensional scaling . This technique allows high-dimensional data to be assigned to new lower-dimensional coordinates that preserve distances between data points . We thus projected the 34-dimensional feature data into a 2-dimensional space . We performed a similar analysis for the EM measurements in 6 white matter regions ( SWM and DWM below ACC , OFC , and LPFC ) and 12 feature dimensions ( axon density , outer diameter , myelin thickness , g-ratio , ratio of thin axons , ratio of medium axons , ratio of thick axons , ratio of XLarge axons , myelin thickness of thin axons , myelin thickness of medium axons , myelin thickness of thick axons , and myelin thickness of XLarge axons ) . We additionally performed discriminant analysis to identify experimental measures that minimize the overlap and clearly separate the distributions of individual data points belonging to different cortical areas across species . Moreover , we performed hierarchical cluster analysis ( HCA ) to group areas based on ( dis ) similarities in their parameter profiles . In this test , the relative similarity of areas is expressed as the distance between two branching points in a cluster tree diagram . HCA and NMDS analyses employed squared area ( dis ) similarity matrices derived from the normalized area profiles by Pearson′s correlation . In addition , we examined whether the feature set of each region between species was correlated . As the feature set included values that vary considerably in scale , we log-transformed the data and then regressed the monkey feature set onto the human feature set . Finally , we examined the potential effects of sex , PMI , age at death , and other diagnoses ( i . e . , seizures ) on all estimates for axon size as well as axon and cell density , using correlation analysis . In addition , we compared all estimated variables between and within species using MANCOVA with sex , PMI , age at death , and other diagnoses as covariates . These analyses did not yield significant effects . Captured digital images that were used for analyses were not modified . Images used in the figures were imported into Adobe Illustrator CC software ( Adobe Systems , San José , California , US ) to assemble in panels . Minor adjustments of overall brightness and contrast were made , but the images were not retouched .
Can the wealth of information from animal studies on the structure and connections of the cerebral cortex—the brain’s outer rim—be translated to understand neural communication in humans and disruption in brain diseases ? To address this question , we examined the prefrontal cortex , which is associated with attention , emotions , and executive control—functions that are disrupted in psychiatric and neurologic diseases . We compared the architecture of the human and rhesus monkey cortex , using identical methods in both species to maximize the accuracy of comparisons . High-resolution microscopy revealed features of gray matter regions , made up of many cells , as well as of the white matter beneath , which contains axons that form connections between brain regions . We found that the architecture of the gray and white matter in humans and monkeys varies systematically ( and in parallel ) and reflects connections assessed by tracers . Using the template established with control human brains , we found significant differences in short- and long-distance pathways in the brains of individuals with autism . The framework established here helps predict patterns in the architecture and connections of areas across mammalian species and sets the stage to integrate functional imaging data from control subjects to compare with pathological states in humans .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "methods" ]
[ "medicine", "and", "health", "sciences", "nervous", "system", "prefrontal", "cortex", "pervasive", "developmental", "disorders", "brain", "vertebrates", "social", "sciences", "neuroscience", "animals", "mammals", "developmental", "psychology", "animal", "models", "primates", "experimental", "organism", "systems", "nerve", "fibers", "autism", "spectrum", "disorder", "old", "world", "monkeys", "research", "and", "analysis", "methods", "rhesus", "monkeys", "monkeys", "animal", "cells", "axons", "macaque", "cellular", "neuroscience", "psychology", "eukaryota", "cell", "biology", "anatomy", "central", "nervous", "system", "neurons", "biology", "and", "life", "sciences", "cellular", "types", "amniotes", "organisms" ]
2018
Parallel trends in cortical gray and white matter architecture and connections in primates allow fine study of pathways in humans and reveal network disruptions in autism
All vertebrate brains develop following a common Bauplan defined by anteroposterior ( AP ) and dorsoventral ( DV ) subdivisions , characterized by largely conserved differential expression of gene markers . However , it is still unclear how this Bauplan originated during evolution . We studied the relative expression of 48 genes with key roles in vertebrate neural patterning in a representative amphioxus embryonic stage . Unlike nonchordates , amphioxus develops its central nervous system ( CNS ) from a neural plate that is homologous to that of vertebrates , allowing direct topological comparisons . The resulting genoarchitectonic model revealed that the amphioxus incipient neural tube is unexpectedly complex , consisting of several AP and DV molecular partitions . Strikingly , comparison with vertebrates indicates that the vertebrate thalamus , pretectum , and midbrain domains jointly correspond to a single amphioxus region , which we termed Di-Mesencephalic primordium ( DiMes ) . This suggests that these domains have a common developmental and evolutionary origin , as supported by functional experiments manipulating secondary organizers in zebrafish and mice . The vertebrate brain is arguably the most complex structure in nature . All vertebrates show a highly conserved construction plan , or Bauplan , of their central nervous system ( CNS ) , which involves several major anatomical and genetic partitions and their subsequent subdivisions [1] . Understanding how this Bauplan has originated during evolution has been a matter of intense research and debate , but there is still no satisfactory answer . Do homologues to major vertebrate brain partitions exist in invertebrate species ? Have new vertebrate partitions originated by subdivision and specialization of preexisting structures ? Did positional genetic patterning mechanisms predate the origin of recognizable neuroanatomical regions , or did both originate concomitantly ? These and related questions have been investigated mainly from an evolutionary developmental ( Evo-Devo ) perspective , since early developing brains have not yet undergone complex morphogenetic deformations and are thus more amenable to evolutionary comparisons between distantly related species . In the case of vertebrates , the CNS arises very early in embryonic development via neural induction . The neuroectodermal plate represents the earliest CNS primordium , which then folds into a closed tube during neurulation . Already at neural plate stages , the CNS becomes regionalized molecularly into large anteroposterior ( AP ) regions . According to the prosomeric model [2–4] , this Bauplan includes the secondary prosencephalon and diencephalon proper ( forebrain ) , midbrain , hindbrain , and spinal cord ( Fig 1A and 1B ) . These primary regions are further partitioned into smaller transverse AP units , identified as brain segments or neuromeres ( Fig 1B ) . Two lineal neuroepithelial signal sources known as secondary organizers are crucial for this process: the zona limitans intrathalamica ( ZLI or mid-diencephalic organizer ) and the isthmic organizer ( IsO , located in the Midbrain–Hindbrain Boundary , MHB; Fig 1B ) . These organizers are characterized by the release of diffused morphogen signals ( SHH and FGF8/WNT1 , respectively ) and are involved in AP regionalization and differential specification of the diencephalic , mesencephalic , and some rostral rhombencephalic neuromeres [5–13] . Furthermore , along the neural tube , each neuromere is composed of four continuous dorsoventral ( DV ) domains: roof , alar , basal , and floor plate regions ( Fig 1A–1D ) . Importantly , the prospective DV pattern is already observed at neural plate stages , corresponding to its mediolateral dimension ( Fig 1A , 1C and 1D ) : the future floor corresponds to the neural plate midline , whereas the future roof lies at the border of the neural plate . The rostral end of the neural plate is thus morphologically singular because the floor does not reach the anterior border of the plate but ends rostrally at the prospective mamillary hypothalamic region , in coincidence with the underlying rostral tip of the notochord [14]; therefore , the roof , alar , and basal plates concentrically cross the midline at the terminal wall ( future acroterminal domain ) , curving around the rostral end of the floor plate ( Fig 1A , [2 , 4 , 15] ) . An important breakthrough in the study of comparative neuroanatomy and the evolutionary origin of CNSs has been the observation that each established AP and DV anatomical partition in a given species is characterized by the differential expression of specific gene markers early in development in a combinatorial code that we refer to as genoarchitecture [14] . These molecular codes create clear-cut molecular boundaries between the neuromeres , and often correspond with visible external bulges due to the differential proliferation of the progenitors because of their distinct genoarchitectonic profiles [3 , 15] . Strikingly , the number of neuromeric units and their associated genoarchitecture is highly conserved in all vertebrate groups , including the basal-branching agnathans [16–34] . This implies that a fundamentally conserved anatomical CNS Bauplan and its corresponding genetic blueprint have existed at least since the last common ancestor of vertebrates . Therefore , a major approach to understanding the origins of this Bauplan has been to investigate the expression of orthologs of key gene markers in chordate and nonchordate invertebrate species . Remarkably , a subset of these markers show fixed relative AP positions , suggesting that some of the regional genoarchitectonic codes of vertebrates were established prior to the origin of the vertebrate brain Bauplan . For example , the transverse genetic boundaries defined by the abutting expression of Fezf/Irx and Otx/Gbx—which in vertebrates correspond to the anatomical positions in which the ZLI and IsO secondary organizers will develop , respectively—are observed in the CNSs of species as diverged as amphioxus and fruit flies [35–37]; although , these sites lack the expression of the morphogens responsible for the organizer activity in vertebrates [38–41] . Moreover , some markers expressed in the annelid Platynereis durmeilii show remarkable topologic similarity with the mediolateral and AP molecular pattern in vertebrates [42–44] . In one of the most striking cases of genetic patterning conservation observed between vertebrates and invertebrates , the diffuse epidermal nervous system of hemichordates displays multiple vertebrate-like AP genetic codes , including a ZLI-like domain with equivalent relative expression of hh , six3 , fng , otx , and wnt8 orthologs and an IsO-like region coexpressing fgf8/17/18 and wnt1 , suggesting conservation of the underlying genetic programs despite the fact that they are patterning divergent structures in the two lineages [45–48] . Altogether , these studies thus suggest that multiple defining genetic programs that pattern the vertebrate brain predate its evolutionary emergence . However , the major limitation of these nonchordate model systems to investigate the origin of the vertebrate brain Bauplan is the lack of an unambiguous anatomical and topological reference system . Even under the assumption that the nervous systems of these invertebrate phyla are truly homologous to the vertebrate CNS , each one has its own set of clade-specific characters and thus correspond to different variational modalities of CNSs [49] . This impedes direct topological comparisons , leaving similarities of gene expression patterns as the only support for any hypothesized homology assignment . For this reason , the cephalochordate amphioxus has traditionally been the most studied invertebrate species for comparative analyses with vertebrates . Unlike nonchordates , amphioxus develops its tubular CNS from a neural plate in the same way that vertebrates do , thus allowing direct topological comparisons of prospective brain regions . Furthermore , unlike tunicates , cephalochordates have undergone slow evolutionary rates , both genomically and morphologically [50 , 51] . Multiple studies on this organism have shown , for instance , that the Otx/Gbx and Fez/Irx genetic boundaries [36 , 37 , 52] as well as part of its neural Hox AP patterning [53–57] are conserved with vertebrates . Similarly , orthologs of many other key vertebrate genes have been implicated in neural function and development in amphioxus ( see S1 Table for a list of previously described gene expression patterns in amphioxus with relevance to CNS development that have been used in this study ) . These reports , together with multiple comprehensive and integrating reviews [41 , 58–64] , have provided important insights on the presence of molecularly-defined partitions in the developing amphioxus CNS . Nonetheless , these studies have been performed by different research groups , using different amphioxus species , and usually focused on the expression of a single gene at multiple embryonic stages . This has made the systematic integration and accurate combinatorial analyses of these expression patterns a complex task . To address these difficulties , we mapped here 48 genes with well-known roles in vertebrate CNS patterning on a single amphioxus developmental stage , the 7-somite mid-neurula , in which a wide spectrum of orthologs of vertebrate neural gene markers is expressed . With these data , we propose an integrative model of the molecular regionalization of the amphioxus developing CNS that is consistent and comparable with the prosomeric model of the vertebrate CNS Bauplan . Our results show that , at the mid-neurula stage , the amphioxus CNS primordium has an unexpectedly complex genoarchitecture , with three major molecularly distinguishable AP divisions ( and some secondary subdivisions ) and a set of standard DV zones . Strikingly , direct topological comparison between the molecular models of the two lineages , as well as extensive novel and previously reported functional data , suggest that the vertebrate territory comprising the diencephalic neuromeric units corresponding to thalamus and pretectum ( prosomeres p2 , p1 ) , but not the prethalamus ( p3 ) , share with the midbrain a common ontogenetic and evolutionary origin , and , altogether , are homologous to a nonregionalized Pax4/6-positive domain in amphioxus , which we termed Diencephalo-Mesencephalic primordium ( DiMes ) . Whether resulting from an increase in complexity in vertebrates or , alternatively , a simplification in amphioxus compared to the last common ancestor of chordates , these results suggest that the differences in AP Bauplan complexity between the two lineages are likely linked to the secondary organizers of vertebrates ( ZLI and IsO ) , which are absent in amphioxus . Experimental abrogation and manipulation of these organizers in vertebrate species generate phenotypic defects that are consistent with this hypothesis . AP and DV subdivisions in developing chordate neural tubes are defined according to axial references . Conventionally , such references are provided in vertebrates by the axial mesoderm ( the notochord ) , the floor plate , roof plate , and alar–basal boundary within the lateral walls of the neural tube , all of which are topologically parallel to each other ( Fig 1 ) . Amphioxus has a notochord , which extends singularly beyond the forebrain [65] , and a floor plate [66–70] . As previously reported for the Floridan amphioxus Branchiostoma floridae [71 , 72] , we observed in the European amphioxus B . lanceolatum that the gene FoxA2-1 is a selective marker of the notochord ( Fig 2A–2A′′′ , 2E and 2F ) , while Nkx2 . 1 seems to be a general floor plate marker at the 7-somite neurula stage ( Fig 2B–2B′′′ ) . As in vertebrates , in which its expression in the floor plate is transient [73] , Nkx2 . 1 expression is highly dynamic during amphioxus CNS development ( S1 Fig ) . Nkx2 . 1 is observed along the entire presumptive floor plate at early- and mid-larval stages , but it subsequently becomes restricted rostralwards . Since the whole neural tube of amphioxus sits on top of the notochord , it should be , in theory , regarded as topographically epichordal . Thus , as the floor plate is induced vertically by the notochord [74–76] , we a priori expected the amphioxus floor plate to extend all along the acroterminal neural midline ( up to the neuropore ) , in contrast to the vertebrate floor plate , which stops at the mamillary pouch of the hypothalamus , coinciding with the approximate position of the rostral tip of the notochord ( Fig 1A and 1B; [2 , 4] ) . Instead , we observed that the floor plate , defined by Nkx2 . 1 expression , does not reach the anterior neural border , but it ends in a slightly expanded median patch that recalls the mamillary hypothalamic ending observed in vertebrates ( Fig 2B and 2B′ insets , K; see also [66] ) . Interestingly , Hedgehog ( Hh ) , which is a well-established floor plate marker in vertebrates [74 , 77 , 78] , and Nkx6 are also expressed in the amphioxus floor plate , but their anterior limit of expression is more caudal than that of Nkx2 . 1 ( Figs 2C–2C′′′ , 2K , 2L and 7D–7D′′; a similar expression for Hh has been reported in B . floridae [69] ) . Goosecoid ( Gsc ) is also expressed in the floor plate ( in contrast to previous reports [79] ) in a variable and patchy pattern that might reflect cyclic dynamic changes ( Fig 2D–2D′′′ ) . These markers differentiate two major floor plate AP regions: ( i ) a rostral-most median floor domain characterized by only Nkx2 . 1 expression , which corresponds to the floor plate of the forebrain region that we refer to as the amphioxus hypothalamo-prethalamic primordium ( HyPTh; see below and Fig 2K and 2L ) ; and ( ii ) the rest of the floor plate , defined by Hh , Gsc , Nkx6 , and Nkx2 . 1 expression . We next examined the genoarchitecture of the amphioxus axial mesoderm to assess the existence of a putative prechordal plate homolog . According to the updated prosomeric model [2 , 4] , the latter tissue lies topologically rostral to the neural primordium and the notochord ( Fig 1B ) . As mentioned above , FoxA2-1 labels the whole amphioxus prospective notochord ( Fig 2A–2A′′′ ) . On the other hand , the expression of both Hh and IrxB in the axial notochordal tissue does not reach the rostral tip of the FoxA2-1–positive domain , stopping beneath the rostral end of the Nkx2 . 1–positive HyPTh floor plate ( Fig 2L , 2E–2J; it should be noted , however , that IrxB expression seems to reach the anterior tip of the notochord in B . floridae [80] ) . Moreover , in the amphioxus axial mesoderm , Six3/6 expression was observed exclusively in the rostral tip of the FoxA2-1–positive domain , beyond the Hh/IrxB–positive part of the notochord ( Fig 2K , 2L and 2J; see also [81] ) . Interestingly , this Six3/6 expression is maintained at later stages , when the notochord is fully formed [81] , indicating that its rostral tip has a distinct molecular signature compared to the rest of the notochord . Remarkably , in vertebrates , Six3 is expressed in the prechordal plate but not in the notochord at any level [24]; therefore , the rostral notochordal tip of amphioxus might represent a possible prechordal plate homologue , previously unrecognized due to its histologic similarity to the notochord proper ( see Discussion ) . Finally , we found that Gbx expression appears restricted to a more caudal sector of the notochord , whose rostral border is posterior to the caudal boundary of the HyPTh neural domain ( Figs 2K , 2L , 3F and 3L inset ) . Previous studies [37] and other observations described below suggest that the Gbx-expressing domain of the notochord and overlying neural tissue begins at the rostral end of the major region we term Rhombencephalo-Spinal primordium ( RhSp; Fig 2L ) . In summary , we observed that the amphioxus axial mesoderm is subdivided molecularly into various regions , which have direct correspondence with major subdivisions in the overlaying neural plate . Previous gene expression studies provided evidence for the presence of longitudinal zones positioned parallel to the floor plate , implying DV patterning in the amphioxus CNS [69 , 82] . We thus investigated the extent of DV regionalization and its related boundaries by systematically searching for gene expression patterns with specific DV domains . We found that most of the examined patterns could be classified into three groups ( Figs 3–9 ) : ( i ) peripheral genes , with expression restricted to the periphery of the neural plate ( future topologically dorsal or alar zone; Six3/6 , Lhx2/9b , Zic , Msx , Pax2/5/8 , Pax3/7 , Nova ) ; ( ii ) internal genes , with expression domains respecting the former peripheral longitudinal zone ( Pou3f , Sim , FoxD , Meis , Lef , Lhx1/5 , Hox3 , Hox6 , FoxB ) ; and ( iii ) pan-DV genes , expressed across both aforementioned domains ( Otx , Gbx , Fezf , Irx , Pax4/6 , Six3/6 , Nkx2 . 2 , Meis , Rx , Hox1 , Wnt3 , Wnt7 , Nova , Ebf ) . A few markers were ascribed to two of these categories since they have DV expression subdomains that differ depending on the AP partition in which they are expressed ( see below and Fig 9 ) . Altogether , these patterns suggest the existence of continuous basal and alar plate zones that extend longitudinally throughout the amphioxus neural tube primordium . As in vertebrates , the right and left moieties meet frontally around the rostral end of the floor plate ( Fig 10A ) , as clearly exemplified by the alar expression of Lhx2/9b ( Fig 6C′ and 6C′′ ) . In all vertebrates , dynamic antagonistic expression of Otx ( rostral ) and Gbx ( caudal ) in the neural plate eventually reaches an equilibrium at the caudal end of the midbrain , defining the MHB ( Fig 3A and 3D; [5 , 83–88] ) . Clonal labeling studies performed in frogs at the 64 blastomere stage showed that this is the earliest detectable brain transverse boundary [89] . A comparable boundary is also present in amphioxus , aligned between the first and second somites [37] , which we further corroborated at the 7-somite neurula stage in B . lanceolatum ( Fig 3B–3F ) . Accordingly , it was suggested that the first intersomitic limit of amphioxus roughly marks the genetic homolog of the MHB of vertebrates [37 , 52 , 53 , 58 , 90] . Thus , it can be postulated that these early expression domains in both lineages define the boundary between a rostral Otx-positive “archencephalic prototagma” ( ARCH; Fig 3J and 3M ) and a caudal Gbx-positive “deuteroencephalic prototagma” ( DEU; Fig 3J and 3M ) . In addition to Gbx , several other amphioxus genes show specific expression within DEU at this stage , abutting rostrally the ARCH–DEU boundary ( Wnt3 , Wnt7 , FoxB , Pax2/5/8 , and Msx; Figs 6G–6H′′ and 7E–7G′′ ) . We previously showed that the ARCH domain can be subdivided anteroposteriorly based on Fezf and Irx expression [36] . In both vertebrates and amphioxus , Fezf genes are expressed in the rostral-most part of the CNS at early neural tube stages , creating an anterior subdomain within the Otx-positive territory ( Fig 3G–3I ) and thus leaving a gap between the caudal end of their expression and the start of that of Gbx in DEU ( Fig 3K and 3L ) . On the other hand , Irx genes are expressed within this gap , abutting rostrally with Fezf and extending posteriorly into the Gbx-positive DEU tagma ( Fig 3K–3O ) . Studies in Xenopus , zebrafish , and mice , comparing Fezf and Irx expression patterns with fate mapping data , have shown that the transverse Fezf-Irx interface marks the prethalamo–thalamic boundary where the ZLI will develop [21 , 91–94] . Based on the expression patterns observed in the 7-somite amphioxus neurula , we accordingly defined a rostral Fezf-positive HyPTh and a caudal Irx-positive DiMes intercalated between the Fezf-positive and Gbx-positive domains ( Fig 3K–3O ) . Remarkably , several genes , including Pax4/6 , Six3/6 , Pou3f , and Sim , are expressed specifically or most strongly within the DiMes ( Fig 4 , see also Fig 8A–8C′′ ) , supporting the distinct identity of this region . Moreover , other genes in addition to Fezf appear restricted to HyPTh ( e . g . , Rx throughout it , Fig 5B and 5B′′ , and FoxD in its basal plate subdomain , Fig 5C and 5C′′; see also Figs 8D–8E′′ and 9 ) or have distinct expression subdomains within HyPTh ( e . g . , Nova , S2C and S2′′ Fig ) . On the other hand , other markers , such as Ebf , are expressed caudally to the Fezf/Irx limit , similarly to the three Irx genes ( IrxA-C ) ( Figs 3N , 3O , 4G , 4G′ , 6A and 6B′′ ) . Triple fluorescent in situ hybridization and confocal 3-D reconstruction show that the Fezf-positive HyPTh , the Pax4/6-positive DiMes and the Gbx-positive RhSp domains abut sharply one another . Interestingly , the intermediate domain , DiMes , is very small , consisting only of two rows of cells along the AP dimension ( Fig 8A–8B′′ ) . Analogous fluorescent in situ hybridization comparison of Fezf , Six3/6 , and Gbx patterns shows that Six3/6 is also strongly expressed in the DiMes compartment ( Fig 8C–8C′′ ) . The Fezf and Gbx markers are expressed with similar mutual relationships also at the 4/5-somite ( early neurula ) stage ( S3A–S3B′ Fig ) , leaving an expression gap where weak Six3/6 signal can already be detected ( S3E–S3F′ Fig ) . Therefore , both the ARCH/DEU limit and the HyPTh and DiMes subdivisions within ARCH are established very early in amphioxus CNS development . We next sought to identify further AP molecular partitions within the HyPTh and DiMes forebrain domains of the 7-somite neurula . Unlike the DiMes , for which we could not identify any molecular subdivision , eight examined markers showed restricted expression domains within HyPTh , sometimes limited to either alar or basal regions . Their combined pattern is consistent with the existence of three AP subdivisions within the HyPTh , which we termed Rostral HyPTh , Intermediate HyPTh and Caudal HyPTh ( Rostral-HyPTh , Interm-HyPTh , Caudal-HyPTh; Fig 10A; see Discussion for possible homology relationships with partitions in the vertebrate forebrain ) . For instance , the expression of six rostral markers ( Nkx2 . 2 , Nova , Meis , Pou3f , Lef , Lhx2/9b ) appears across Rostral-HyPTh and Interm-HyPTh , but seems to respect a transverse double row of cells that lie anterior to the Irx-expressing DiMes; topologically , this caudal negative domain of HyPTh ( Caudal-HyPTh ) would correspond in vertebrates to the primordium of the prethalamic region ( Fig 10A ) . This partition can be visualized as a gap of negative labeling , e . g . , by double in situ hybridization for Meis and IrxC or for Nkx2 . 2 and IrxB ( Figs 5G–5G′ and 8F–8F′′ , respectively ) . While Nkx2 . 2 and Nova signals are present at both alar and basal levels of Rostral-HyPTh and Interm-HyPTh ( but not in the corresponding part of the floor plate; Fig 5D–5D′′ and S3C–S3C′′ Fig; see schematic details of expression in Fig 9 ) , Pou3f , Meis , and Lef expression is restricted to the local basal region ( also respecting the floor plate; Figs 4F–4F′′ , 5F–5F′′ and 5H–5H′′ , respectively ) , and Lhx2/9b expression appears selectively in the peripheral alar region ( Fig 6C–6C′′ ) . Six3/6 is the only studied marker restricted to the Rostral-HyPTh , specifically in the alar plate ( Figs 2J , 4D–4D′′ and 8C–8C′′; similar expression of Six3/6 in the anterior-most part of the neural plate was also reported in B . floridae [81] ) . On the other hand , Zic , which is a well-known marker of the alar and roof plates in the CNS of vertebrates [95] , is expressed throughout the presumptive alar plate of the HyPTh , but its expression is significantly stronger at the Caudal-HyPTh , showing decreasing signal towards the rostral alar parts of the HyPTh complex ( S2G–S2G′′ Fig ) . In contrast to the major HyPTh and DiMes partitions , the three HyPTh molecular subdivisions are not fully established at the 4/5-somite stage ( S3 Fig ) . Numerous amphioxus genes have been previously reported to show iterative expression domains within the DEU , suggesting the existence of characteristic subdivisions within this partition ( [96–99] , and see S1 Table ) . Consistent with these studies , we identified several molecular AP partitions within the rostral-most subdomain of DEU , referred to here as the RhSp , which roughly ends caudal to the fifth somite at the 7-somite neurula stage ( Figs 9C and 10A ) . Gbx , Wnt3 , Wnt7 , and FoxB appear selectively expressed throughout the RhSp; all of them abut rostrally the DiMes/RhSp boundary , and their expression domains end at different caudal levels , either coinciding with the end of somite five or extending further caudalwards ( Figs 3E , 3F , 6G–6G′′ , 6H–6H′′ and 7E–7E′′ ) . Gbx , Wnt3 , and Wnt7 occupy both alar and basal regions ( but not the floor plate ) , as previously described [37 , 100 , 101] , while FoxB is restricted to basal areas ( Fig 7E; see also [102] ) . Hox1 , Hox3 , and Hox6 genes are also expressed along alar and basal parts of the RhSp , with rostral expression borders that correspond with the intersomitic limits S3/S4 , S4/S5 , and S5/S6 , respectively ( Fig 7A–7C′′; see also [53–57] ) . As mentioned above , these molecular partitions are complemented by patterns of iterated spots with negative intervals , which can be aligned with the center ( Lhx1/5 and Pou3f , Figs 6E–6E′′ and 4F–4F′′ ) or posterior half of the somites ( Nova , S2C–S2C′′ Fig ) or the inter-somitic boundaries ( Pax3/7 , S2E–S2E′′ Fig ) . In the case of Pax2/5/8 , patches are less well defined , particularly caudally , where they become nearly continuous ( Fig 7F–7F′′ ) . Finally , some genes show isolated spots of expression located at different positions within the RhSp AP subdivisions: Msx ( Fig 7G–7G′′ ) , Meis ( Fig 5F–5F′′ ) , Zic ( S2G–S2G′′ Fig ) , Nkx6 ( Fig 7D–7D′′ ) , and Otp ( Fig 6F–6F′′ ) , sometimes correlating with the prospective position of the future pigmented photoreceptor spot . A major implication of our comparison of the overall CNS genoarchitecture between amphioxus and vertebrates is that the small amphioxus Pax4/6-positive DiMes corresponds topologically to the large vertebrate region comprising the thalamus , pretectum , and midbrain ( Fig 10A–10B′′ , see Discussion ) . Patterning of this territory in vertebrates occurs under the dual control of the secondary brain organizers ( ZLI and IsO , see Introduction ) , which induce the molecular subdivision and differential growth of an initially Pax6-positive primordium . Among other effects , these organizers inhibit the expression of Pax6 at the two ends of the territory so that Pax6 signal becomes restricted to the caudal pretectum ( p1 ) and the epithalamus ( dorsal-most part of p2 ) ( Fig 11C′ ) . Accordingly , we hypothesized that the small size and the lack of internal regionalization of the DiMes , particularly with respect to Pax4/6 expression , may be , at least in part , related to the absence of ZLI-like and IsO-like effects in amphioxus . To gather support for this hypothesis , we turned first to loss-of-function transgenic mouse lines in which either the ZLI or the IsO are absent . Double Fezf1-/-Fezf2-/- mutants [103] lack the ZLI organizer and largely lose the molecular identity of the alar thalamic field , displaying expanded expression of the pretectal markers Pax6+ and Ebf1+ ( Fig 11B and 11B′–11B′′′′ ) ; the midbrain was not altered in these mice . We also studied En1cre/+; Fgf8flox/flox mice ( see Materials and Methods ) in which the IsO is deleted across the MHB [104] . The resulting phenotype showed a reduction of the AP dimension of the pretecto-mesencephalic region down to one-third of its normal size and an abnormal caudal expansion of PAX6 immunoreaction ( as well as of the posterior commissure ) , suggesting a lack of differential specification of the midbrain versus the pretectum ( Fig 11C and 11C′–11C′′′′ ) ; in this case , the thalamus seemed normal . Next , we tried to eliminate both the ZLI and the IsO together in zebrafish , using quadruple morpholino ( 4MO ) treatment against otx1a , otx2 , eng2a , and eng2b . Although effects in neural progenitors of other areas cannot be ruled out , double morpholinos against otx1a and otx2 were successfully used previously to specifically abolish the ZLI [105] , while double morpholino treatment against eng2a and eng2b caused the loss of the IsO [106] , and expression of pax6a throughout the midbrain remnant [107] . Strikingly , the normal pax6a-negative gaps corresponding to the alar plate of the midbrain and diencephalic thalamus were abolished or severely reduced in nearly all ( 84% ) 4MO specimens tested ( Fig 11F; n = 75 , p = 3 . 84 × 10−31 , one-sided Fisher Exact test ) , often resulting in a continuous expression of pax6a between the rostral conserved part of the forebrain and the hindbrain ( Fig 11G , 11H , 11K and 11L and sagittal sections in insets in Fig 11H and 11L ) . Supporting the effective suppression of the two organizers in this experiment , we observed disappearance of the dorsal ZLI spike expression of shha and of the MHB-related transverse band of wnt1 expression ( Fig 11D , 11E , 11H , 11I , 11L and 11M ) . In addition , all 4MO embryos showed a significant reduction of the zebrafish DiMes-like remnant at 28 h post fertilization ( hpf ) compared to the controls ( Fig 11G′ and 11K′ ) . Consistent with previous results [37 , 59 , 64] , our data show that the incipient amphioxus neural tube is molecularly divided anteroposteriorly into a rostral archencephalic ( ARCH ) and a caudal deuterencephalic ( DEU ) portions from very early stages , similarly to vertebrates ( Fig 10A and 10B ) . Traditionally , three main AP divisions are defined in the vertebrate ARCH ( Fig 1 ) : the secondary prosencephalon ( encompassing hypothalamus plus telencephalon ) , the diencephalon , and the midbrain . On the contrary , the ARCH of amphioxus shows only two main divisions , which we termed DiMes and HyPTh ( Fig 10B ) . DiMes is a small caudal region consisting of two rows of cells that occupies the topological position corresponding in vertebrates to the midbrain and the two diencephalic segments that lie caudal to the ZLI organizer; no internal subdivisions were detected within DiMes . In contrast , the HyPTh encompasses three molecularly distinct segments: a relatively large , bipartite , putative hypothalamus-homolog region ( where neither telencephalic nor optic vesicles are present [113] ) plus a caudal region that occupies the topological position corresponding to the vertebrate prethalamus . In the case of the DEU , its rostral portion , referred to here as RhSp primordium , may represent a field-homolog of the vertebrate hindbrain , and shows a number of gene expression patterns that configure periodic segment-like territories ( Fig 9C ) . Notably , these major AP partitions of the developing CNS are mirrored by molecularly defined subdivisions in the underlying axial mesoderm . Indeed , we provide evidence that the distinct molecular entity at the rostral tip of the amphioxus notochord may be homologous to the vertebrate prechordal plate , being thus essentially different from the notochord proper that underlies the brain floor plate ( Fig 2 ) . This potentially prechordal region lies topologically rostral to the HyPTh ( not under it ) , as occurs with the prechordal plate in vertebrates [2 , 4] , and is characterized by the absence of Hh and Nkx6 and the specific expression of Six3/6 , which is also characteristic of the vertebrate prechordal plate [24] . Therefore , it is possible that this special notochord-looking region—which also shows unusual proliferation and rostralward growth [70 , 114]—may correspond to a variant prechordal plate homolog and/or plays partly equivalent signaling functions to this structure in amphioxus , despite the absence of some key vertebrate prechordal markers ( Gsc , noggin , and chordin [79 , 115] ) . Finally , regarding DV patterning , multiple markers provide extensive evidence for continuous molecularly distinct floor , basal , and alar zones throughout the length of the incipient neural tube ( Figs 9 and 10 ) . Although we did not find selective markers for the roof plate , it is possible that these may exist at later stages , upon neural tube closure . Consistent with the idea that the alar plate and roof plate are not differentially specified in amphioxus at these stages , orthologs of several vertebrate neural plate border makers ( e . g . , Pax3/7 , Msx , and Zic ) were found to be expressed broadly in the alar plate ( Fig 9; observed also in B . floridae [116] ) . According to the updated prosomeric model [4] , the nontelencephalic part of the vertebrate secondary prosencephalon can be subdivided into two main neuromeres: terminal ( THy , hp2 ) and peduncular ( PHy , hp1 ) hypothalamic prosomeres . In addition , THy includes a specialized rostral-most median part extending dorsoventrally , the acroterminal area [2 , 4 , 117] ( Fig 10C′′ ) . This molecularly distinct domain produces a number of specialized formations along the DV axis , including the alar preoptic lamina terminalis , the optic chiasma , the eye vesicles , the basal median eminence , and the neurohypophysis . In amphioxus , HyPTh represents a relatively large , molecularly distinct forebrain region lying rostral to the DiMes . This domain has specific expression of Fezf throughout ( Fig 9 ) , which is also absent caudal to the ZLI limit in vertebrates [93 , 103 , 118] . Our analysis suggests that there are three molecularly distinct AP subdivisions within the amphioxus HyPTh , which we termed Rostral-HyPTh , Interm-HyPTh , and Caudal-HyPTh . By direct topological ascription , these might correspond , respectively , to the transverse THy ( including a rostromedian acroterminal region ) and PHy hypothalamic segments and a prethalamus-like segment next to the DiMes . Six3/6 was the only studied marker that selectively labeled Rostral-HyPTh . Remarkably , in mice , Six3 is expressed extensively dorsoventrally across the alar and basal zones of THy ( including the acroterminal area ) , whereas Six6 signal is restricted to a ventral suprachiasmatic part of the THy acroterminal alar plate , but none of them are expressed at PHy [2 , 117] . These data support a genetic equivalence between the Rostral-HyPTh and THy , in addition to their topological correspondence . Moreover , amphioxus develops in its acroterminal region ( orange domain in Fig 10C and 10C′ ) a median primordial eye patch and , ventral to it , a median group of “infundibular cells” [90] , which are located above the most anterior floor plate cells ( gray cells in Fig 10C ) and might represent a homologue of the vertebrate neurohypophysis . As mentioned above , in vertebrates , both the eyes and the neurohypophysis develop from the acroterminal area [2] , further supporting the homology of vertebrate and amphioxus acroterminal domains and thus of Rostral-HyPTh and THy ( Fig 10C–10C′′ ) . In the case of the Caudal-HyPTh primordium , its topological position , lying directly rostral to the Fezf-Irx boundary , provides grounds to suggest field homology with the vertebrate prethalamus . Importantly , previous studies indicate that Fezf genes are essential to specify the prethalamic domain in vertebrates; however , unlike regions within the vertebrate DiMes counterpart ( see below ) , this specification is independent of the ZLI organizing activity and occurs prior to its formation [93 , 103] and is thus compatible with the amphioxus scenario at the examined stage . Nonetheless , it should be noted that , although more weakly expressed , the presence of Rx expression in Caudal-HyPTh ( absent in the prethalamus of vertebrates [118] ) , suggests the alternative possibility that this partition may represent a primordium homolog to both the peduncular hypothalamus and prethalamic region . One of the most striking implications of our results is that the small , Pax4/6-positive DiMes of amphioxus corresponds topologically to the region comprising the vertebrate thalamus , pretectum , and midbrain ( Fig 10 ) . While this area is not subdivided in amphioxus and consists only of two cell rows at the neurula stage , the equivalent vertebrate region shows three major partitions and extensive cell proliferation . These partitions in vertebrates originate during development as a consequence of the action of the secondary brain organizers on a Pax6-positive primordium . In particular , Shh signaling from the ZLI is crucial for the specification of the thalamus [6–8 , 119] , and Fgf8 and Wnt1 expression from the IsO are necessary for proper midbrain specification and differential caudal growth [5 , 10 , 12 , 13 , 120–123] . Moreover , due to the action of these organizers , the expression of Pax6 in this primordium is mainly restricted to the pretectum and the epithalamus and becomes absent in the ventricular zone of the thalamus and midbrain ( Fig 11 ) [124 , 125] . Therefore , altogether , these data suggest that the vertebrate thalamus , pretectum , and midbrain share a common origin , both ontogenetically ( from an early and transient Pax6-positive area found between the prospective ZLI and IsO levels ) and phylogenetically ( homologous to the amphioxus DiMes region ) . This hypothesis has two major implications for our understanding of the vertebrate brain Bauplan and its evolutionary origins . First , it implies that two of the diencephalic prosomeres—pretectum ( p1 ) and thalamus ( p2 ) —are more evolutionarily related to the midbrain than they are to the third diencephalic prosomere—the prethalamus ( p3 ) —which would , in turn , be more related with the secondary prosencephalon ( see previous section ) . That is , the diencephalon proper would be neither an evolutionarily nor an ontogenetically primordial subdivision of the vertebrate brain . This striking implication is further supported by the differential responses of these regions to experimental manipulation of the organizers and their associated signaling molecules . Chicken-quail heterotopic grafts of the ZLI , as well as focalized ectopic expression of SHH using beads in chicken embryos , show that only pretectum and midbrain , but not the prethalamus , are competent to be re-patterned to a thalamic fate [6 , 7 , 119] . Similarly , quail-chick , rat-chick , or mouse-chick heterotopic grafts of the IsO generate an ectopic midbrain in pretectal and thalamic regions , but never in the prethalamus and secondary prosencephalon [126–129] . That is , thalamus , pretectum , and midbrain have similar developmental potentials that are not shared by the prethalamus . In fact , our hypothesis provides a plausible ontogenetic explanation that has long been missing for these intriguing observations , underscoring its explanatory power . A second major related implication of our hypothesis is that the vertebrate thalamus , pretectum , and midbrain jointly share altogether a common ancestor with the amphioxus DiMes . Since neither Hh nor Fgf8 and Wnt1 , the key morphogens involved in ZLI and IsO activity , respectively , are expressed at the corresponding topological positions in amphioxus [40 , 69 , 130–132] , it is plausible to speculate that vertebrate thalamus , pretectum , and midbrain partitions may have emerged evolutionarily from an ancestral Pax4/6-positive DiMes-like region concomitantly to the evolution of the ZLI and IsO brain organizers as orthogonal signaling centers . Alternatively , the undivided , small amphioxus DiMes may represent an evolutionary simplification upon the loss of the organizers [47 , 48] , if they were already patterning the neural plate-derived CNS of the last common ancestor of chordates . Irrespectively , a major prediction of both evolutionary hypotheses is that suppression of the organizers during vertebrate development should result in a ( relatively ) homogeneous , smaller , undivided , and fully Pax6-positive region lying between recognizable prethalamus and hindbrain , as we observed in mouse and zebrafish embryos with suppressed ZLI and/or IsO ( Fig 11 , and see also [103 , 107 , 133–136] ) . Although the converse experiment—the induction of ectopic organizers in amphioxus—is still not technically possible , future methodological developments could allow assessing if and how the DiMes may respond to these morphogens . Finally , an independent line of evidence supporting the functional homology between the amphioxus DiMes and the corresponding vertebrate regions comes from the retinal projections in the two lineages . In vertebrates , primary eye projections target mainly the midbrain ( optic tectum/superior colliculus ) , while secondary eye projections target mainly the pretectum and thalamus and , to a lesser extent , prethalamus and hypothalamus [137] . In amphioxus , projections from the single frontal eye have recently been mapped to a Pax4/6-positive region in the four gill slit larval stage [138] , which likely corresponds to a DiMes derivative based on its topological position and Pax4/6 expression . Our comprehensive genoarchitectonic model of the developing amphioxus CNS at mid-neurula stage sheds new light onto the origins of the vertebrate brain . First , it shows that the basic blueprint of the vertebrate brain Bauplan was already present in the last common ancestor of chordates . The major AP and DV partitions identified in amphioxus have direct topological correspondence with vertebrate counterparts , even though these may be further elaborated in vertebrates . Such is the case of the eye vesicles and the telencephalon developing as alar expansions of a HyPTh-like region or the growth and regionalization of a DiMes-like region into thalamus , pretectum , and mesencephalon . Secondly , it highlights the importance of the evolution of secondary organizers in the gain or loss of brain partitions . Thirdly , it allowed us to propose novel homologies between amphioxus and vertebrate structures , such as the acroterminal hypothalamic area and the prechordal plate . Finally , it casts doubts on the relevance of the classic separation between forebrain and midbrain in vertebrates from an evolutionary and developmental perspective , suggesting that a redefinition of the main AP regions into which the vertebrate brain is classically divided ( forebrain , midbrain , and hindbrain ) could provide a better conceptual framework to understand the origins of the vertebrate brain . All animal work in this study has been conducted following the Spanish and European legislation . Adult fish were only used to obtain eggs through natural mating ( ethical committee approval number: 635/2014 ) . All mouse experiments were performed according to protocols approved by the Universidad Miguel Hernandez OEP committee ( UMH . IN . EP . 01 . 13 ) and Conselleria Generalitat Valenciana ( 2014/VSC/PEA/00055 ) . Chicken experiments were performed according to protocols approved by the ethical committee from the University of Murcia ( 137/2015 ) . For all the previously annotated genes in the B . floridae genome , primer pairs were designed to span the full-length coding sequence when possible . A liquid cDNA library from different developmental stages of the European amphioxus ( B . lanceolatum ) was screened by PCR using B . floridae specific primers . For previously unannotated genes , we performed tBLASTN searches in the B . floridae JGI v1 . 0 genome , using the aminoacidic sequences of the vertebrate orthologs . The corresponding genomic sequences were retrieved and a gene model was predicted by GeneWise2 and GeneScan , as previously described [139] . Cloned B . lanceolatum mRNAs used for in situ hybridization are available in S2 Table . Ripe adult amphioxus specimens were collected in Argelès-sur-mer , France . Spawning was induced as previously described [140] in a dry laboratory in Barcelona , Spain . After in vitro fertilization , embryos were cultured at 18 ºC for 15 h or 21 h ( 4/5 somite and 7 somite stages , respectively ) and fixed with 4% PFA in MOPS buffer overnight at 4°C . Chromogenic whole-mount in situ hybridization was performed as previously described [36] using Nitrobluetetrazolium/bromochloroindolyl phosphate ( NBT/BCIP ) or BMP purple ( Roche ) as chromogenic substrate for the final alkaline phosphatase . Following whole-mount in situ hybridization , selected embryos were embedded in a 0 . 1 M PBS solution with 15% gelatine and 20% sucrose , frozen in isopentane , and sectioned with a cryostat at 12–14 μm-thick . Double-fluorescent in situ hybridizations were performed essentially as nonfluorescent in situ hybridizations , as described in [141] with two extra steps of incubation in 5% NAC and ( 50 mM DTT , 1% NP40 , 0 . 5% SDS ) in PBS1X before the hybridization step . Dinitrophenol ( DNP ) -labeled antisense riboprobes were synthesized using DNP-11-UTP labeling reagent ( PerkinElmer ) , and DIG-labeled antisense riboprobes were synthesized using DIG RNA labeling mix ( Roche ) . Labeled riboprobes were detected using anti-DNP-POD ( Perkin Elmer ) and anti-DIG-POD ( Roche ) antibodies , and green and red fluorescent signals amplified with TSA -Plus -Fluorescein and Tetrarhodamine systems ( Perkin Elmer ) , respectively . Images were acquired using a Leica TCS-SPII confocal microscope or a Zeiss Axiophot . Confocal datasets were deconvolved with Huygens Professional version 16 . 05 ( Scientific Volume Imaging , The Netherlands , http://svi . nl ) , analyzed , and assembled with ImageJ; for panels B and B′ in Fig 8 , images were further processed with Imaris ( 7 . 2 . 3 , Bitplane AG , software available at http://bitplane . com ) . Breeding zebrafish ( Danio rerio ) were maintained at 28°C on a 14 h light/10 h dark cycle as described in [142] . To disrupt the ZLI and IsO secondary organizers together , we performed a quadruple transient knockdown using four morpholino-antisense oligomers ( MOs ) that had been previously described to abolish each of the organizers individually: otx1a and otx2 MO’s for the ZLI [105] , and eng2a and eng2b for the IsO [106] . As injection controls , we used a combination of the two nontargeting MOs that were used in the original articles ( a morpholino-sense oligomer against twhh ( Cont1 ) [105] and a standard control MO ( Cont2 ) [106] ) . The combination of experiment or control MOs was injected at the one-cell stage into the yolk at the following concentrations ( based on the original sources ) : otx1a ( 0 . 25 mM ) , otx2 ( 0 . 25 mM ) , eng2a ( 0 . 5 mM ) , eng2b ( 0 . 5 mM ) , Cont1 ( 0 . 5 mM ) , Cont2 ( 1 mM ) . Each embryo was injected with 1 . 5 nl of the MO mix ( injection of 1 . 0 nl produced similar , yet milder , phenotypes , whereas injection of 2 . 0 nl resulted in full mortality ) . Four independent experiments were performed ( in different days ) , injecting approximately 100 eggs per condition and experiment . Injected embryos were fixed in 4% PFA overnight at 4 ºC and used for whole-mount in situ hybridization as previously described [143] . A subset of stained embryos was cryosectioned , and both sections and whole embryos were mounted in 80% glycerol-PBS and photographed in a Zeiss Axiophot microscope . The full list of probe sequences is available in S3 Table . The Fgf8 conditional mutant was generated by the Gail R . Martin laboratory [120] , and the transgenic mouse line expressing cre under the En1 promoter was generated in the Dr . Wolfgang Wurst laboratory [144] . Mutant embryos were generated by crossing double heterozygous males ( En1cre/+; Fgf8flox/+ ) with homozygous Fgf8flox/flox conditional females . Immunohistochemistry ( PAX-6 ) and in situ hybridization ( Otx2 ) in paraffin sections were performed as previously described [145] . The primary PAX-6 rabbit polyclonal IgG antibody was diluted in PBTG ( 1:500; PRB-278P/Covance ) . The Otx2 probe was synthesized as in [86] . All the procedures involving extraction of brain samples and further tissue processing were done as previously described [146] . Fertilized chicken ( Gallus gallus domesticus ) eggs were bought from a national farm ( Granja Santa Isabel; Córdoba , Spain ) and incubated at 38 ºC and 65% controlled humidity in a forced draft incubator until the Hamburger–Hamilton stage five ( HH5 ) [147] . Embryos were fixed by immersion in 4% paraformaldehyde in 0 . 1M phosphate buffered saline ( PBS , pH 7 . 4 ) during 16 h at 4 ºC . Whole-mount in situ hybridization was done as previously described [146] using probes for Otx2 and Gbx2 reported in [24] . Fezf2 probe was cloned using the following primers: F , GCTACAAACCCTTCGTCTGC and R , GCTCAGGGTCACTTGCTACC .
According to textbooks , vertebrate brains develop from a neural tube that rapidly becomes regionalized into the forebrain ( which includes the secondary prosencephalon and diencephalon ) , midbrain , and hindbrain . These regions are then further subdivided; in particular , the diencephalon gives rise to the prethalamus , thalamus , and pretectum . However , embryological manipulations of brain signaling centers showed that the prethalamus behaves very differently than the thalamus and pretectum , which largely share their developmental potential with the midbrain . Therefore , this classic partition scheme might not be fully consistent from a developmental perspective . To better understand the origin and evolution of the regionalization of the vertebrate brain , we built a comprehensive molecular model of the incipient neural tube of amphioxus , an invertebrate chordate that shares multiple features with its vertebrate relatives . This model shows that the amphioxus nervous system is unexpectedly complex , sharing its basic blueprint with that of vertebrates . However , a single undivided region in amphioxus , which we termed Di-Mesencephalic primordium ( DiMes ) , unambiguously corresponds to the region encompassing the thalamus , pretectum , and midbrain in vertebrates , indicating that these regions are also more closely related evolutionarily . Therefore , the diencephalon as a neuroanatomical compartment as well as the classic separation between forebrain and midbrain in vertebrates seem inconsistent from both an evolutionary and developmental perspective .
[ "Abstract", "Introduction", "Results", "Discussion", "Concluding", "remarks", "Materials", "and", "methods" ]
[ "medicine", "and", "health", "sciences", "in", "situ", "hybridization", "molecular", "probe", "techniques", "nervous", "system", "brain", "vertebrates", "animals", "developmental", "biology", "molecular", "biology", "techniques", "amphioxus", "research", "and", "analysis", "methods", "embryology", "probe", "hybridization", "gene", "expression", "thalamus", "brainstem", "molecular", "biology", "anatomy", "central", "nervous", "system", "midbrain", "genetics", "biology", "and", "life", "sciences", "notochords", "organisms" ]
2017
Molecular regionalization of the developing amphioxus neural tube challenges major partitions of the vertebrate brain
Spike timing dependent plasticity ( STDP ) is believed to play an important role in shaping the structure of neural circuits . Here we show that STDP generates effective interactions between synapses of different neurons , which were neglected in previous theoretical treatments , and can be described as a sum over contributions from structural motifs . These interactions can have a pivotal influence on the connectivity patterns that emerge under the influence of STDP . In particular , we consider two highly ordered forms of structure: wide synfire chains , in which groups of neurons project to each other sequentially , and self connected assemblies . We show that high order synaptic interactions can enable the formation of both structures , depending on the form of the STDP function and the time course of synaptic currents . Furthermore , within a certain regime of biophysical parameters , emergence of the ordered connectivity occurs robustly and autonomously in a stochastic network of spiking neurons , without a need to expose the neural network to structured inputs during learning . Synfire chains , consisting of distinct groups of neurons that project to each other in a sequential order ( Fig 1A ) , were originally proposed as a model for sequence generation in the mammalian cortex [20] . Subsequently , compelling evidence has pointed to the possibility that this architecture underlies the synchronous neural activity observed in the songbird premotor nucleus HVC [21 , 22] , with ∼100 excitatory neurons in each layer . Theoretical reasoning indicates that this architecture can indeed produce stable propagation of synchronous spiking activity if the groups are sufficiently large [23 , 24] . It was shown theoretically that STDP , combined with heterosynaptic competition , can lead spontaneously to the formation of thin chains [25] , in which single neurons project to each other sequentially ( see also [26] ) . Formation of wide synfire chains , in which many neurons participate in each group along the sequence , proved to be more difficult , and was successfully demonstrated when structured inputs were fed into the network [25] . Thus it remains unclear whether structured inputs are required for the formation of wide chains in networks of spiking neurons . Here we demonstrate that with appropriate choice of the biophysical parameters , STDP dynamics can lead to robust formation of wide synfire chains , without the need to provide any structured inputs to the network . The grouping of neurons occurs spontaneously , assisted by high order synaptic interactions that arise from the STDP dynamics . As a second example for the role of high order synaptic interactions in plasticity , we consider whether STDP can promote the formation of distinct , self connected groups of cells ( Fig 1B ) . Theoretical works demonstrated that such structures can lead to multiple persistent states in the neural dynamics [27–31] . Another motivation for consideration of these structures arises from anatomical studies of local connectivity: for example , connections among excitatory neurons in the rat visual and somatosensory cortices tend to be clustered [1 , 32] . In the context of learning , it has been demonstrated theoretically that STDP , combined with additional plasticity mechanisms and structured inputs , can lead to formation of self connected assemblies [27 , 33 , 34] . However , in similarity to the formation of synfire chains by STDP , it remains unclear whether self connected assemblies can emerge in an initially unstructured network without the inclusion of correlated inputs that are fed into the network during learning . Here we show that high order synaptic interactions enable the spontaneous formation of self connected assemblies , without the inclusion of such inputs . In studying the effects of STDP , it is necessary to consider models of neural activity that explicitly involve the timing of action potentials ( as opposed to simpler rate models ) . Due to the difficulty in evaluating spike correlation functions in most models of spiking neurons , analytical treatments of STDP have made certain approximations for the spiking statistics: typically , pre and post synaptic spike trains were treated as if they follow inhomogeneous Poisson statistics [12–14 , 16] . Therefore , we explicitly consider a recurrent network of neurons which follow linear Poisson ( LP ) dynamics ( Methods ) . The activity in the network is stochastic , and the probability of each neuron to emit an action potential is proportional to a weighted sum of the previous activity in the network and a constant external input ( see Fig 2A and Methods ) . Networks of LP neurons have been shown to approximate well the correlation in spike timing of neurons with more elaborate leaky integrate-and-fire dynamics , operating in an asynchronous regime [35] . The availability of an exact expression for spike correlations in such networks ( see below ) allows us to develop a precise theory for the weight dynamics driven by STDP . We consider synaptic efficacies in a recurrent neural network with an arbitrary structure ( Fig 2A ) . The efficacies undergo long term potentiation or depression in response to each pair of spikes , depending on the time interval between the firing of the pre and post synaptic neurons ( Fig 2B ) . In the case of slow learning rate , the rate of change in the synaptic efficacies can be expressed in terms of the product between the time dependent pair correlation and the STDP function ( see Methods ) : Δ i j STDP ≡ ∫ d τ F τ C i j τ . ( 1 ) Here , Δ i j STDP is the drift in the synaptic efficacy from neuron j to neuron i , defined as the average change in the synaptic efficacy per unit time . The correlation function Cij ( τ ) represents the probability density that neurons j and i emit a pair of spikes temporally separated by a duration τ ( Eq 9 ) , and F ( τ ) is the STDP function that describes how potentiation ( or depression ) depends on the time interval τ ( Fig 2C ) . Using an analytic expression for the spike correlation function [36] , we obtain an exact expression for STDP in recurrent networks with arbitrary connectivity ( Methods ) : Δ STDP = f 0 r r T + 1 2 π ∫ - ∞ ∞ d ω F ˜ - ω I - a ˜ ω W - 1 D I - a ˜ - ω W T - 1 . ( 2 ) Here , W is the connectivity matrix , a ( t ) is the time course of synaptic currents , a ˜ ( ω ) is its Fourier transform , ri is the average firing rate of neuron i ( Eq 13 ) , Dij = δij rj , and f0 is the area under the STDP function ( Eq 16 ) . The derivation of Eq 2 does not involve any assumptions on the specific form of the synaptic currents , STDP function , or the network architecture . S1 Fig demonstrates that this expression provides an accurate description of the average learning dynamics in networks of LP neurons , over time scales which are relevant to plasticity in the brain . Eq 2 expresses how plasticity in one synapse depends on the connectivity of the full network . Additional insight on this expression , which may seem elaborate , is obtained by noticing that the spike correlation functions in the network can be written as a power series , obtained as an expansion in the strength of the synaptic efficacies [37] . This allows us to reformulate Eq 2 as follows ( Methods ) : Δ i j STDP = f 0 r i r j + ∑ α β f α , β · ∑ k r k W α i k W β j k , ( 3 ) where the coefficients fα , β are defined below . Each one of the terms in Eq 3 has a relatively simple , intuitive interpretation that we discuss next . The first term in Eq 3 , f0 ri rj , represents the contribution to STDP arising from the mean firing rates of the pre and post synaptic neurons , while ignoring any correlations in the timing of their spikes . Accordingly , this term is simply proportional to the firing rates ri and rj ( Fig 3A ) . Such a term is often postulated in phenomenological models of synaptic plasticity [38] , and its emergence from STDP dynamics has been described , e . g . , in [11 , 12 , 16 , 17] . The probability of neurons i and j to emit a spike is transiently modulated whenever a spike is emitted anywhere within the network . In the sum on the right hand side of Eq 3 , each term quantifies how a spike in a neuron k modulates the probability of neurons i and j to emit pairs of spikes at various latencies—and through this modulation , how the spikes of neuron k influence the drift in the synaptic efficacy Wij . This contribution to the drift is written as a sum over structural motifs , which share a common organization shown schematically in Fig 3B . In each structural motif , the source neuron k projects to the post synaptic neuron i via a path of α synapses , and to the pre synaptic neuron j via a path of β synapses . The synaptic drift driven by the motif is proportional to all the synaptic weights along the two paths , and to the firing rate of the source neuron . In addition , the drift is proportional to a motif coefficient fα , β . This coefficient depends on the number of synapses in the two paths , the time course of the synaptic currents , and the detailed form of the STDP learning function . We discuss this dependence in detail later ( see also Methods ) . The first order contributions in the above sum are those in which {α , β} = {1 , 0} ( Fig 3C ) , or {α , β} = {0 , 1} ( Fig 3D ) : Δ i j STDP = f 0 r i r j + f 1 , 0 r j W i j + f 0 , 1 r i W j i + … . ( 4 ) These terms are local: they depend only on the direct synapses that link neurons i and j , and on the firing rate of these two neurons ( Fig 3C and 3D ) . Previous works [17 , 18] derived these contributions to STDP using heuristic arguments that focused on the pre and post synaptic neurons , and studied their consequences when embedded in a recurrent network . Under an asymmetric STDP function , the first order terms induce a competition between a synapse Wij and the opposite synapse Wji [17 , 18] , whereas a symmetric STDP function tends to promote the development of a symmetric weight matrix . Here , these local plasticity rules are obtained as the first order terms in a systematic expansion , which includes also higher order terms . Next , we demonstrate that high order motifs can promote the formation of large-scale structures in the synaptic connectivity . We focus on two types of structures: synfire chains ( Fig 4A ) , and clusters of self connected assemblies ( Fig 4B ) . In both structures , synapses of different neurons are highly correlated . The purpose of this section is to illustrate by specific examples that high order motifs , beyond the first order , can lead to emergence of these structural correlations . A more systematic study is presented in later sections . We consider networks that consist of recurrently connected excitatory neurons , and inhibitory interneurons with fast synapses ( see Methods ) . The inhibitory input to each neuron depends on the total activity of the excitatory network , and is adjusted such that the inhibitory weights balance the excitatory ones [14 , 15] . The main role of inhibition in the network is to suppress runaway excitation in the neural dynamics . In addition , all neurons receive constant , identical inputs ( Methods ) . The plasticity mechanisms acting on the excitatory neurons are summarized in Fig 4E . The excitatory connections are modifiable through STDP and are bounded between zero and a positive bound , denoted by wmax . In addition to STDP , the excitatory synapses undergo heterosynaptic competition that limits the total synaptic input and output of each neuron: the sum of the incoming excitatory weights to each neuron , and the sum of outgoing excitatory weights from each neuron are bounded to lie below a positive hard bound , denoted by Wmax . The competition , combined with STDP and with the hard bound on each synaptic weight , can lead to a steady state connectivity pattern in which each neuron receives input from a certain number of pre-synaptic partners , and projects to a certain number of post-synaptic partners [25] . These numbers are tuned by the ratio between Wmax and wmax [25] . To test the influence of individual motifs on the STDP dynamics , we perform simulations in which we include only a few of the terms in Eq 3 , starting from initial weights that were drawn independently from a uniform random distribution ( Methods ) . Instead of obtaining the exact expressions for the coefficients fα , β ( Eq 15 ) , we artificially tune their values and observe the consequences on the structures that emerge . The motif {2 , 1} ( Fig 3G ) , when acting between excitatory neurons , encourages neurons that receive input from the same pre synaptic neuron to project into the same post synaptic neuron ( Fig 4A ) . Consequently , this motif can induce correlations between the synaptic connections formed by neurons that belong to the same layer of a synfire chain . This is a key feature which differentiates wide synfire chain structures from other connectivity patterns in which each neuron has a prescribed number of presynaptic and postsynaptic partners . This observation raises a hypothesis , that the motif {2 , 1} can promote formation of wide synfire chains , by favoring these structures over other connectivity patterns which are compatible with the constraints set by the heterosynaptic competition . To test this hypothesis , we perform simulations that include only contributions from the motifs {α , β} = {1 , 0} , {2 , 0} , {2 , 1} , and the opposing terms in which α and β are exchanged . Additionally , we set fα , β = −fβ , α , as expected if the STDP function is antisymmetric . An example is shown in Fig 4C . When including only the first and second order motifs , a simulation of the plasticity dynamics leads to a structure in which each row and column contains a small number of active synapses , without any reciprocal connections ( left ) . However , the synaptic weights are not organized in a synfire chain structure . When including the third order motifs {2 , 1} and {1 , 2} , the synaptic efficacies self organize into a perfect synfire chain ( right ) , despite the absence of any correlations in the external inputs to the network . Results from a wider set of simulations , in which we systematically vary the strength of motifs , are presented in the section Self organization into synfire chains . As a second example , we examine the influence of the motif {1 , 1} on the formation of self connected cell assemblies . The motif {1 , 1} ( Fig 3E ) , when acting between excitatory neurons , enhances reciprocal connections between neurons that receive common input ( note that the contribution from this motif vanishes if the STDP function is antisymmetric , but for a symmetric STDP function this motif can significantly contribute to the plasticity dynamics ) . This raises the hypothesis that formation of self connected assemblies can be promoted by the contribution of the motif {1 , 1} . To check this hypothesis , we conduct plasticity simulations that contain only contributions from first and second order motifs: {0 , 1} , {1 , 0} , and {1 , 1} in Eq 3 . In addition , we impose the relation f1 , 0 = f0 , 1 , as expected if the STDP function is symmetric ( see Eq 15 in Methods ) . In the example shown in Fig 4D , the first order motifs , together with the synaptic competition , lead to a symmetric connectivity matrix in which the number of active synapses is limited in each row and in each column ( left ) . With the inclusion of the {1 , 1} motif ( right ) , strong correlations emerge between synapses of different neurons , and fully connected cell assemblies emerge . So far , we illustrated how high order motifs can promote the formation of global structures by artificially tuning the contribution of specific motifs to plasticity ( through the coefficients fα , β ) . In the actual STDP dynamics , the coefficients fα , β cannot be controlled independently . Instead , these coefficients are determined by the temporal structure of the STDP function and the synaptic currents . Each motif induces temporal correlations of spikes between the pre and post synaptic neurons with a characteristic time course . The time course of correlations depends on the structure of the motif ( characterized by α and β ) , and on the time course of the synaptic currents . Therefore , the synaptic current , together with the STDP function , affects how strongly each motif influences the synaptic drift , as quantified by the magnitude of the corresponding motif coefficient fα , β ( Eq 15 ) . The influence of the synaptic currents on the motif coefficients is illustrated in detail in Methods . As a specific example for how the time course of synaptic currents can affect the motif coefficients , we consider in Fig 5 the influence of a delay in the onset of the post synaptic current ( abbreviated below as the synaptic latency ) . Synaptic latencies depend on diverse physiological properties , such as the length and conductance velocity in the axon [39] and location on the dendrite [40] . Here we model the synaptic latency as a temporal shift in the synaptic current ( Fig 5 ) . In the example shown in Fig 5 the motif coefficients f1 , 0 and f2 , 0 decrease with an increase of the synaptic latency . The coefficient f2 , 1 is influenced by the synaptic latency as well , but for synaptic latencies ranging from 0 to 10 ms this dependence is extremely weak . Therefore , an increase in the synaptic latency reduces the contribution of the motifs {1 , 0} and {2 , 0} to STDP relative to the motif {2 , 1} . The reason for these trends is explained qualitatively in Methods . Higher order motifs exhibit a similar behavior , depending on the difference between α and β ( S5 Fig ) . The possibility to tune the relative contribution of motifs through the interplay of post synaptic currents and the STDP function , suggests that global structures could be spontaneously generated via STDP with appropriate choice of these biophysical parameters . Furthermore , this observation provides a principled way to search for parameters that enable emergence of specific structures . We next focus on the emergence of synfire chains under the influence of STDP and heterosynaptic competition . For simplicity , we consider mainly the case where the STDP function is antisymmetric . First , we consider in detail the interplay between the third order motif {2 , 1} , which facilitates the formation of synfire chains , and the first order motifs {1 , 0} and {0 , 1} , whose contribution to STDP dynamics was the focus of previous theoretical works [17 , 18] . Intrinsic plasticity mechanisms other than STDP can act to self-regulate the efficacy of each synapse , in similarity to the effect of the first order motif {1 , 0} ( see below ) . Therefore , it is interesting to consider f1 , 0 and f0 , 1 as separate parameters even if the STDP function is antisymmetric . In Fig 6 we examine the phase space spanned by f1 , 0 , f0 , 1 , and f2 , 1 , while assuming that f1 , 2 = −f2 , 1 due to the antisymmetric form of the STDP function . To avoid decay of all weights to zero when f1 , 0 is strongly negative , we include in the dynamics also a term that drives growth of each weight at a fixed rate ( see Methods , Eq 19 ) . For simplicity , we first consider a situation in which other motifs do not contribute to the dynamics . Fig 6A shows results from simulations in which we varied f0 , 1 and f1 , 0 while fixing f2 , 1 . To quantify whether synfire chains emerge robustly we constructed a score that quantifies similarity between the steady state excitatory connectivity and a perfect synfire chain structure ( abbreviated below as the chain score ) . This chain score ranges from 0 to 1 , where 1 corresponds to a perfect match ( Methods ) . The figure shows the chain score , averaged over multiple random choices of the initial weights . Over a wide range of values of f1 , 0 and f0 , 1 precise synfire chain structures are obtained robustly . We note two characteristics of this parameter regime: first , the coefficient f0 , 1 is negative . Thus , each synapse inhibits its reciprocal synapse . Second , f1 , 0 lies within a range of values which is fairly insensitive to f0 , 1 when |f0 , 1| is sufficiently large . Similarly , in Fig 6B we fix f0 , 1 , while varying f1 , 0 and f2 , 1 . As expected , synfire chains emerge only when f2 , 1 is sufficiently large . Furthermore , with increase of f2 , 1 the range of values of f1 , 0 that permits formation of synfire chains becomes wider . Thus , the high order motif {2 , 1} plays a pivotal role in the spontaneous emergence of wide synfire chains . Results from additional simulations , in which we include additional low order motifs are shown in S2 Fig . Under typical conditions relevant to the full plasticity dynamics ( discussed below ) , the second order motifs {2 , 0} and {1 , 1} have a detrimental effect on synfire chain formation ( a contribution from the motif {1 , 1} may be present if the STDP function is not antisymmetric ) . Next , we address the emergence of synfire chains under the full STDP dynamics . Finally , we check whether self connected assemblies can spontaneously emerge under the full STDP dynamics . The motif {1 , 1} promotes formation of such structures ( Fig 4B and 4D ) . Therefore , we choose biophysical parameters that increase the relative contribution of this motif . First , we choose an STDP function with a Mexican hat structure ( Fig 9A ) , which increases synaptic efficacies between neurons that spike at similar times , regardless of the temporal order of the spikes . Second , we note that the contribution of the {1 , 1} motif is independent of synaptic latency , because both the pre and post synaptic neurons i , j accrue the same latency relative to the source neuron k . On the other hand , coefficients of other low order motifs do depend on the synaptic latency ( Fig 9B ) . Based on the above reasoning , we expect that self connected assemblies will emerge , under the influence of STDP and synaptic competition , for finite synaptic latencies , in which contributions from motifs other than {1 , 1} are suppressed . Fig 9C shows results from a stochastic simulation of a Poisson network , starting from initial random connectivity , with a synaptic delay of ∼5 ms . Here , a precise structure of self connected assemblies emerges robustly . Other structures were observed for alternative choices of the synaptic latency . We focused on the influence of biophysical parameters on the contribution of motifs up to third order . However , Eqs 2 and 3 include also contributions from higher order motifs . This raises a question , why an analysis up to third order allowed us to predict the emergence of global structures . A partial answer to this question is that for small synaptic weights , the contribution of high order motifs decays with the number of participating synapses . One important factor contributing to the decay of motif coefficients is that when |α − β| is large the spike correlation function is shifted outside the range of the STDP window . In addition , spike correlation functions widen with increase of the number of participating synapses , thus decreasing their overlap with the STDP function . Another , more formal argument is based on the derivation of Eq 3: as long as the synaptic weights are sufficiently weak , such that all eigenvalues of the connectivity matrix are smaller than unity , the expansion in Eq 14 converges for all ω , and therefore the sum in Eq 3 must converge as well ( this is also the condition for stability of the linear neural dynamics ) . This implies that the combined contribution from all motifs of order n must decay as a function of n . Even though contributions of high order motifs must eventually decay , our simulations of the full STDP dynamics were performed in a regime where motifs beyond the third order do influence the plasticity . To a large extent , the effect of higher order motifs can be predicted by the contribution of second and third order motifs ( S5 Fig ) . For example , all high order motifs that satisfy α − β = 1 are expected to assist the formation of synfire chains structures , based on the same intuition that was demonstrated in Fig 4A . All these motifs also share a similar time course since they involve a delay of one synapse between the activity of the pre and post synaptic neurons . Similarly , all the motifs with α = β are expected to contribute to formation of self connected assemblies . In similarity to the motif {1 , 1} , and in contrast to the other motifs , their contribution is not influenced by the synaptic latency . The similar dependence on the synaptic latency in motifs with the same value of α − β is illustrated in S5 Fig . Fiete et al . [25] demonstrated that narrow synfire chains , in which single units project sequentially to each other , can emerge spontaneously under the combined influence of STDP and heterosynaptic competition . However , wide synfire chains did not emerge unless correlated inputs were fed into the network , even though the STDP simulations implicitly included motifs of all orders . Our results suggest why it was difficult to obtain wide synfire chains robustly in this work: first , Fiete et al . did not include in their model self depression , which can suppress the contribution of the first order motif , bringing the plasticity dynamics to an appropriate regime ( see Fig 6A ) . In particular , it is likely that the choice of parameters was such , that the relative contribution of the third order motif was not sufficiently large . Interestingly , wide ( but sparsely connected ) synfire chains were spontaneously produced in another recent work [44] , which considered a simplified model of neural and synaptic dynamics , operating in discrete time bins . By applying our framework to this model , it is straightforward to see that only a small subset of the possible structural motifs contributed to plasticity , due to the simplified and discrete dynamics . In addition , the synaptic plasticity rules included an effective form of self-depression . Thus , the spontaneous formation of synfire chains in [44] is consistent with the predictions of our work . Considerable theoretical attention has been devoted to the influence of STDP on the steady state distribution of synaptic weights [12 , 13 , 45–49] . This interest is partially motivated by the observation in specific brain areas of unimodal distributions of synaptic efficacies , often following approximately a log-normal distribution [1 , 50] . Due to our interest in formation of synfire chains and self-connected assemblies , we focused on situations in which the steady-state weight distribution is bimodal . However , under certain choices of parameters which do not lead to the formation of ordered structures , we observe unimodal weight distributions ( see , for example , the black area in Fig 6A which is characterized by strong synaptic self-inhibition [48] ) . It will be of interest in future studies , to ask whether it is possible to obtain highly ordered structures , in which the non-vanishing weights follow broad distributions , perhaps under a softer implementation of the synaptic competition . It will also be interesting in future studies to consider situations in which connections exist between a subset of neuron pairs: for example , the structural connectivity may be sparse . The analytical framework that we developed can be directly applied to networks with an arbitrary adjacency matrix . In this case , only efficacies of structurally existing synapses should be updated based on Eqs 2 and 3 ( note that in Eq 3 , only those motifs that are realized in the structural connectivity graph can contribute to the sum , since the synaptic efficacies associated with non-existing connections vanish ) . Moreover , the formalism can be easily generalized to consider synapses with heterogeneous biophysical properties . Nucleus HVC plays a key role in timing the vocal output of songbirds [21 , 22] . This nucleus is a compelling candidate for a brain area that can organize autonomously to produce structured dynamics , since auditory deprived songbirds generate a song with a stereotypical temporal course [51 , 52] . However , in almost all theoretical works that addressed how local plasticity rules give rise to temporal sequences of neural activity , it was necessary to provide some form of structured input into the network in order to robustly produce the sequential neural activity [25 , 53–55] . Similarly , structured inputs were required in order to robustly produce self connected assemblies , which give rise to another useful form of neural dynamics , characterized by multiple stable states [27–31] . It is therefore significant that synaptic structures which support structured neural dynamics can emerge in a neural circuit without any preexisting order in the synaptic organization , and without any exposure to external stimuli . Appropriate choices of the biophysical parameters , which enable this type of autonomous organization , became apparent by applying the theoretical formalism and reasoning developed in this work . Finally , we briefly mention another area of future work , in which the formalism developed here may find a useful application: we expect that the analysis of synaptic dynamics in terms of contributions from structural motifs , will be valuable for assessing the role of STDP in shaping the high order statistics of cortical connectivity , as experimental data on these statistics become increasingly available [1 , 2 , 32] . The time dependent activity of neuron i is a stochastic realization of an inhomogeneous Poisson process , with expectation value λ i t = ∑ k = 1 N W i k ∫ - ∞ t d t ′ a t - t ′ S k t ′ + b i , ( 5 ) where N is the number of neurons , W is the connectivity matrix , a ( t ) is the synaptic current , bi is a constant external input , and S k ( t ) = ∑ μ δ ( t - t k μ ) is the spike train of the neuron k ( where t k μ are spike times of the neuron ) . We assume that the neurons do not excite themselves , meaning that ∀i Wii = 0 . Assuming that all spike pairs contribute to STDP , the change in the synaptic efficacies due to STDP can be expressed as follows: W ˙ i j STDP t = W ˙ i j + t + W ˙ i j - t , ( 6 ) where W ˙ i j + t = S i t ∫ - ∞ t S j t ′ F t - t ′ d t ′ ( 7 ) is the change in synaptic efficacy arising from spikes in the post synaptic neuron i at time t , and W ˙ i j - t = S j t ∫ - ∞ t S i t ′ F t ′ - t d t ′ ( 8 ) is the change following a spike in the pre synaptic neuron j at time t . In both terms , the integration is over all previous spikes of the presynaptic neuron ( Eq 7 ) or the postsynaptic neuron ( Eq 8 ) . We define the correlation function of spikes in each pair of cells as follows , C i j τ ≡ S i t + τ S j t , ( 9 ) where 〈 ⋅ 〉 denotes averaging over different realizations of the Poisson dynamics for a given connectivity . For constant external input , and under the assumption of slow learning , the correlation function is stationary , and does not depend on t . We denote the rate of change in the synaptic efficacy , averaged over realizations of the Poisson dynamics as Δ i j STDP ≡ W ˙ i j STDP , ( 10 ) and refer to it in short as the synaptic drift . Using the correlation function we can express the synaptic drift as follows , Δ i j STDP = ∫ - ∞ ∞ C i j τ F τ d τ . ( 11 ) For linear dynamics the correlation function can be written exactly . In the frequency domain [36] , C ˜ ω = 2 π δ ω r r T + I - a ˜ ω W - 1 D I - a ˜ - ω W T - 1 , ( 12 ) where ri = 〈λi〉 , the diagonal matrix Dij = δij rj , I denotes the unit matrix , and we use the convention in which the Fourier transform of a function g ( t ) is defined as g ˜ ( ω ) = ∫ - ∞ ∞ e - i ω t g ( t ) d t . The average firing rate can be easily obtained from Eq 5: r = I - a ˜ 0 W - 1 b . ( 13 ) By substituting Eq 12 in Eq 11 , we obtain Eq 2 . Note that the diagonal terms should be ignored , since we assume that there is no synapse from a neuron to itself . We consider N excitatory neurons , with modifiable recurrent connections . In addition to the STDP , the excitatory synapses undergo heterosynaptic competition , and possibly constant self depression and growth at a constant rate . The full plasticity dynamics of these synapses can be summarized by the following expression: W ˙ i j ex t = η W ˙ i j STDP t - ψ Δ i in - ψ Δ j out - μ W i j ex + γ . ( 19 ) The terms Δ i in , Δ j out represent the heterosynaptic competition [25]: competition over the input to neuron i , Δ i in = ∑ k W i k ex - W max · Θ ∑ k W i k ex - W max , ( 20 ) and competition over the outputs from neuron j: Δ j out = ∑ k W k j ex - W max · Θ ∑ k W k j ex - W max . ( 21 ) Here , Θ ( x ) is the Heaviside step function: Θ x = 0 x < 0 1 x ≧ 0 When ψ is sufficiently large , the competition guarantees that the sum over each row and column of Wex does not exceed Wmax . Finally , the term μ W i j ex represents self depression ( constant weakening of the synapse in proportion to the synaptic efficacy ) , and the term γ represents a constant growth of each weight at a fixed rate . In addition to these rules , the excitatory synapses are restricted to the range [0 , wmax] . To classify groups of neurons that share similar connectivity , we performed k-means classification on a set of N vectors , where the i-th vector includes all the excitatory input and output synaptic efficacies of neuron i: { W i k ex , W i k T ex } k = 1 … N , and using a squared Euclidean distance . We then reordered the neurons ( and the connectivity matrix ) based on the groups identified by the k-means clustering . When searching for synfire chain structure , we chose the order of groups as follows: We randomly chose one group and set it as the first group . We then looked for a remaining group that receives the largest total synaptic input from the first group , and set it as the next group . This process was repeated to include all groups . Next , we compared the ordered connectivity matrix to an “ideal” binary connectivity matrix that represents complete feed forward connectivity between the groups , or complete clustering into self connected groups . The chain score is defined as: 1 − x , where x is the normalized square distance between the ordered matrix , scaled by the largest element , and a matrix representing ideal feed forward connectivity ( or a perfect arrangement of self connected clusters ) . Finally we maximized the similarity score over a range of values of k .
Plasticity between neural connections plays a key role in our ability to process and store information . One of the fundamental questions on plasticity , is the extent to which local processes , affecting individual synapses , are responsible for large scale structures of neural connectivity . Here we focus on two types of structures: synfire chains and self connected assemblies . These structures are often proposed as forms of neural connectivity that can support brain functions such as memory and generation of motor activity . We show that an important plasticity mechanism , spike timing dependent plasticity , can lead to autonomous emergence of these large scale structures in the brain: in contrast to previous theoretical proposals , we show that the emergence can occur autonomously even if instructive signals are not fed into the neural network while its form is shaped by synaptic plasticity .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[ "medicine", "and", "health", "sciences", "action", "potentials", "neural", "networks", "nervous", "system", "membrane", "potential", "electrophysiology", "neuroscience", "synaptic", "plasticity", "network", "analysis", "neuronal", "plasticity", "computer", "and", "information", "sciences", "developmental", "neuroscience", "network", "motifs", "animal", "cells", "biophysics", "physics", "cellular", "neuroscience", "cell", "biology", "anatomy", "synapses", "physiology", "neurons", "biology", "and", "life", "sciences", "cellular", "types", "physical", "sciences", "neurophysiology" ]
2016
Shaping Neural Circuits by High Order Synaptic Interactions
Transcription factor ( TF ) regulation is often post-translational . TF modifications such as reversible phosphorylation and missense mutations , which can act independent of TF expression level , are overlooked by differential expression analysis . Using bovine Piedmontese myostatin mutants as proof-of-concept , we propose a new algorithm that correctly identifies the gene containing the causal mutation from microarray data alone . The myostatin mutation releases the brakes on Piedmontese muscle growth by translating a dysfunctional protein . Compared to a less muscular non-mutant breed we find that myostatin is not differentially expressed at any of ten developmental time points . Despite this challenge , the algorithm identifies the myostatin ‘smoking gun’ through a coordinated , simultaneous , weighted integration of three sources of microarray information: transcript abundance , differential expression , and differential wiring . By asking the novel question “which regulator is cumulatively most differentially wired to the abundant most differentially expressed genes ? ” it yields the correct answer , “myostatin” . Our new approach identifies causal regulatory changes by globally contrasting co-expression network dynamics . The entirely data-driven ‘weighting’ procedure emphasises regulatory movement relative to the phenotypically relevant part of the network . In contrast to other published methods that compare co-expression networks , significance testing is not used to eliminate connections . Evolution , normal development , immune responses and aberrant processes such as diseases and cancer all involve at least some rewiring of regulatory circuits [1]–[3] . Indeed it is the subtle ( and sometimes not so subtle ) differences in circuit wiring that makes each individual unique . The key nodes in regulatory circuits are frequently transcription factors ( TF ) [4] . Thus , there is a great deal of interest in developing methods for decoding TF changes . Regulator-target interactions can be assessed by ChIP-on-chip but this requires large amounts of homogenous starting material and TF-specific reagents . Furthermore , the recruitment of a TF to a promoter does not necessarily correlate with transcriptional status , so biological interpretation can be complex [5] . Likely sites of key regulatory mutations can be revealed by Whole Genome Scans ( WGS ) but this approach requires large numbers of individuals and very dense SNP panels . Even so , the exact causal gene may remain ambiguous if there are several genes near the marker . In any case , little insight is gained into the underlying regulatory mechanisms . In order to gain further insights into the regulatory apparatus , computational approaches are continuously being proposed . To date , they all operate by integrating information from multiple levels of biological organisation particularly eQTL , protein-protein interaction and TF binding site data [6]–[9] . Identifying regulatory change solely through contrasts in gene expression data has been elusive because TF tend to be stably expressed at baseline levels [10] close to the sensitivity of standard high-throughput expression profiling platforms . Further , TF activation is often regulated post-translationally and thereby can act somewhat independently of expression level . Biologically important common TF activation processes ( localisation to the nucleus , phosphorylation , ligand binding , formation of transcriptionally ‘open’ euchromatin , and presence of cofactors , all in addition to mutations in the protein coding region of the regulator ) are poorly detected by conventional differential expression ( DE ) analysis . We hypothesised that a system-wide network approach might have utility , on the grounds that while a differentially-regulated TF might not be DE between two systems , its new position in the network of the perturbed system might allow detection of the ‘smoking gun . ’ To allow reliable evaluation of such a hypothesis a well-defined experimental model system is required . Piedmontese cattle are double-muscled because they possess a genomic DNA mutation in the myostatin ( GDF8 ) mRNA transcript [11] . The resulting dysfunctional myostatin protein is a transcriptional regulator that releases the brakes on muscle growth reflecting the importance of TGF-β signalling pathways in the determination of final muscle mass and fibre composition [12] , [13] . A preliminary analysis of the expression of myostatin in Piedmontese×Hereford versus Wagyu×Hereford animals found that DE of myostatin was not detectable using cDNA-based expression microarrays [14] , [15] . Thus we have a system in which we know the identity of the gene containing the causal mutation , myostatin ( MSTN ) , but we cannot identify it by DE of the mRNA in muscle samples . By contrasting the muscle transcriptomes of the Piedmontese and Wagyu crosses across 10 developmental time points , our aim was to establish the question to which myostatin is the answer . In other words , what question do we need to ask of the gene expression data for it to reveal the identity of the transcriptional regulator containing the causal mutation ? We found that 11 , 057 genes gave valid expression signal: noise data across the 10 developmental time points for the 2 crosses ( Table S1 ) . Of these 11 , 057 genes 920 were deemed to be gene expression regulators ( Table S2 ) . The experimental design ( Figure 1 ) allowed us to assess DE between the crosses and we visualised the data on an MA plot ( Figure 2 ) , identifying 85 DE genes using conservative statistical criteria . The most DE genes included slow twitch muscle structural genes ( e . g . MYL2 ) , which were higher in the Wagyu crosses ( W×H ) than in the Piedmontese crosses ( P×H ) and immune genes ( e . g . HLA-DQA2 ) , which were higher in the P×H than in W×H . The most DE transcriptional regulator was CSRP3 which was higher in W×H than in P×H . Consistent with previously published data using a cDNA-based microarray [14] myostatin was not DE between the crosses . Next , we examined the difference in the specific behaviour or co-expression of targeted pairs of genes between the two crosses , by subtracting the correlation coefficient in Wagyu from that in Piedmontese . This approach has a very recent precedent [16] . However , two important modifications presented in the current co-expression work include an absence of significance analysis and the decision to limit the computation to a targeted subset of genes ( i . e . , transcriptional regulators versus DE genes ) . This targeting better emphasises the transcriptional regulation of the change in two systems . This quantification of a gene's differential connectivity is more sensitive than the majority of published approaches where only the total number of ‘significant’ connections is contrasted . Instead , like [16] we exploit the identity of the connectors and the differential magnitude of each connection , even in circumstances where the correlation is weak in either ( or both ) of the networks . As this principal forms the basis of the rest of our analysis and appears to capture the regulatory rewiring that takes place in myostatin mutant muscle , it will be referred to from this point on as differential wiring ( DW ) . In circumstances where we do analyse changes in total numbers of ‘significant’ connections , we elected to use the term differential hubbing ( DH ) on the grounds that the total number of connections determines the extent to which a gene can be considered a hub . The PCIT algorithm was used to establish significance in these cases [17] . Table 1 contains definitions for the main terms used in our new analysis , and identifies those aspects which are completely novel ( PIF and RIF ) from those which have been published in some form ( DE , DH and DW ) . The most DE gene in our dataset is MYL2 , and myostatin is the third most DW regulator to it , with a value of 1 . 103 . The derivation of DW is illustrated for the myostatin-MYL2 connection in Figure 3 . It is built on the differences in the myostatin-MYL2 co-expression patterns across development in the Piedmontese cross minus the Wagyu cross . A positive DW is generated where the expression of the target ( e . g . MYL2 ) is positively correlated with the regulator ( e . g . myostatin ) in P×H , while in W×H the expression is either less positive or negative . The converse applies for negative DW . In an attempt to assess the importance of each DE gene to the change in phenotype , we propose a new metric: the “phenotypic impact factor ( PIF ) . ” PIF is a mathematical abstraction designed to ‘weight’ for the contribution the various DE genes make to the difference in the molecular anatomy of the two systems , based purely on their numerical properties . The values were generated by combining the amount of DE between the crosses , coupled with the average abundance calculated for both crosses at all time points for each of the 85 DE genes . Abundant transcripts that were highly DE scored highly , whereas scarce transcripts that were only slightly DE scored poorly . The high phenotypic impact genes enriched for slow twitch muscle structural genes ( MYL2 , MYL3 , TNNT1 , MYH7 , ACTN2 and MYOZ2 ) correctly highlighting the observed phenotype change between the breed crosses , namely the gross muscle fibre transition . The coherence of the output is very consistent with an expectation based on the observed gross anatomical fibre change [18] . We formalised this observation using the GOrilla tool [19] comparing the GO terms enriched by high DE to those enriched by high PIF , computed for all 11 , 057 genes . Not surprisingly , the extremes of both lists strongly enrich for muscle structural components because the transcriptome data was derived from muscle tissue . However , GOrilla assigned a p-value for ‘contractile fibre part’ – the top match within the ‘cellular component’ context - that was 7 orders of magnitude , or 10 , 000 , 000 times more significant for extreme PIF than for extreme DE ( p = 3 . 14E-21 versus p = 2 . 01E-14 ) . We thus conclude that PIF performs well at enriching those genes which appear to contribute strongly to the difference in phenotype between the two states , although a full justification of this conclusion requires further experimental evidence . On the other hand , the PIF metric is not particularly well suited to regulators , although they were included in the analysis . Regulators are often stably expressed at close to baseline levels making detection of isolated changes in expression level challenging and possibly misleading . To account for this , we ascribed “regulatory impact factors” ( RIFs ) to each of the 920 regulators based on their cumulative , simultaneous , DW to the DE genes , accounting for the PIF of the DE genes . This metric was intended as a mathematical abstraction to represent the relative importance of the regulators in driving the phenotypically relevant part of the network described above , based on differences in their correlations . Those regulators that were highly DW to many of the high PIF genes received strong scores , whereas those that were DW to a few , low PIF genes scored poorly . Figure 4 illustrates the extent to which myostatin is highly DW to the high PIF genes , with Piedmontese and Wagyu co-expressions plotted on the two axes . The 85 red circles correspond to the 85 myostatin-DE gene co-expression values . Circle size corresponds to the PIF of the DE gene co-expressed with myostatin at that particular co-expression intersection . The perpendicular distance from the diagonal line corresponds to the amount of differential wiring . For myostatin , this distance tends to be greatest for the high PIF genes ( largest circles ) . The five largest circles are MYL2 , CSRP3 , MYH6 , CA3 and MYL3 . It is important to note that Figure 4 also reveals that most of the mass ( i . e . , most of the correlation coefficients ) are either close to −1 or close to +1 . This bimodal distribution in the correlation coefficients affecting DE genes has already been documented [20] and contrasts with the expected uni-modal distribution that would be obtained across all genes where most of the mass is centred around zero . We explored 2 alternative methods to compute RIF scores ( Eq4 and Eq5 Materials and Methods ) . Myostatin had the fourth most positive RIF using Eq4 and the second highest using Eq5 ( Table 2 ) . Overall the RIF values calculated using the two equations had a correlation efficient of ∼0 . 7 . In the absence of evidence favouring one approach over the other we decided to follow the original thread of defining the question to which myostatin was the answer . When we calculated the mean of the two different RIF values , myostatin received the highest score out of the 920 regulators with a RIF of 3 . 49 ( Figure 5 and Table 2 ) . Two muscle transcription factors MEF2C and MYOD1 also appeared in the top ten , although the former was ranked much lower by Eq4 . In addition , SUV39H2 ( a histone methyltransferase that cooperates with SMADs [21] , components of the TGF-ß pathway though which myostatin is proposed to act , lay in third place ( Table 2 ) . No major muscle TF , or components of the TGF-ß pathway , were in the top ten negative RIF genes ( Table 3 ) . The remainder of the top 10 positive and negative RIF regulators are annotated in Tables 2 and 3 , and can be compared and contrasted to the top 10 positive and negative DE regulators in Tables 4 and 5 . To highlight which cluster of DE genes are being ‘perturbed’ by which cluster of regulators , the DW values for the 920 regulators ( in rows ) and the 85 DE genes ( in columns ) ( Table S3 ) can be assembled into a ‘perturbation matrix’ which we visualised using PermutMatrix software [22] . This novel representation of gene expression data ( derived from the more traditional configuration with genes in rows and samples in columns ) allows for the separation of DE genes from regulators and , after hierarchical clustering , reflects the way in which the regulators are co-differentially wired with each other ( i . e . , where the differential wiring behaves in a co-ordinated manner ) . In Figure 6 a small section of the perturbation matrix is shown . Of particular note was a tight cluster of 18 DE genes comprising 5 genes encoding high PIF slow twitch structural proteins ( MYL3 , TNNT1 , MYH7 , ACTN2 and MYOZ2 ) and also featuring SMPX and 2 DE regulators ( ANKRD1 and CSRP3 ) . MYL2 , another gene encoding a slow twitch structural protein , did not feature in this DE module , but clustered on its own . The regulatory axis contained several high impact regulatory ‘hot spots . ’ One of these included myostatin and MYOD1 at its heart , and also included CSRP1 , USF1 , POU5F1 , NR3C2 , SBNO1 and PITX2 . The very tight clustering of myostatin and MYOD1 reflects closely coordinated patterns of DW between the two crosses across the 85 DE genes . These biologically-sensible clusters imply that co-differential wiring can be used as an explicit criterion to form an edge in a regulatory perturbation network . We used a hard 0 . 9 threshold to establish network edges between those regulators that were highly co-differentially wired to the 85 DE genes . We visualised the deduced network in Cytoscape [23] . This approach led to an enormous cohesive module of low impact regulators ( those regulators that apparently do not contribute to the change in phenotype ) , plus a number of smaller high impact modules ( those regulators that apparently do contribute to the change in phenotype ) . A notable high impact module comprised 3 transcriptional regulators: MSTN , MYOD1 and IFRD1 . The derivation of the high co-differential wiring between the crosses for myostatin and MYOD1 is illustrated in more detail with specific respect to the slow twitch module genes in Figure 7 . In contrast to myostatin and MYOD1 , which are highly positively co-differentially wired to each other , the other member of the module , IFRD1 , is highly negatively co-differentially wired to them . The greatest DW values for all three transcriptional regulators tend to be associated with the high PIF muscle structural genes at the far right of the x axis ( ANKRD1 , MYOZ2 , TNNT1 , MYH6 , SMPX , CSRP3 and ACTN2 ) . However , positive correlation of DW of regulators does not necessarily imply positive correlation , or indeed any significant correlation , of expression of the regulators themselves and vice versa . In other words , neither the clustered regulators on the y axis of the perturbation matrix nor the clustered DE genes on the x axis are actually significantly co-expressed with each other in any combination , based on a PCIT analysis ( unpublished data ) . Furthermore , Myostatin , MyoD1 and IFRD1 are not significantly co-expressed with any of the other 11 , 057 genes in the system , let alone the subset in the matrix . The same applies to ACTN2 , MYH6 , CSRP3 , ANKRD1 , MYL3 and MYOZ2 ( unpublished data ) . Rather , it is the coordinated manner in which two genes differ in their behaviour in the two systems that drives co-differential wiring . We tested the distributional and numerical properties of RIF1 and RIF2 ( Eq4 and Eq5 ) on a simulated data to assess the extent to which our real output could be ascribed to chance . The simulated data comprised 5 , 000 genes surveyed across 10 experimental conditions ( in line with the 10 time points ) in two treatments ( in line with the two breed crosses ) . In accordance with the real data , expression values were simulated from a normal distribution with a mean of 8 . 6 and a standard deviation of 2 . 8 and truncated at 4 and 16 . Also , for each gene , its expression profile across the two treatments was simulated to have a correlation of 0 . 95 . Simulations were performed under the null hypothesis of no differential expression between treatments , no correlation between genes across conditions , and no regulator-target relationships . Therefore , in these settings any observed association could be attributed to chance alone . For the computations of RIF1 and RIF2 , a random 920 genes were selected and treated as potential regulators and their regulatory impact factor computed against the 85 genes showing the most extreme measure of differential expression across the two conditions . Based on this approach a simulated version of Figure 4 was constructed ( data not shown ) which , unlike the observed Figure 4 from our real data , bore most of its mass in its centre ( indicative of a bell-shaped distribution of correlation coefficients ) . Both distributions were found to be statistically different as indicated by the Kolmogorov-Smirnov two-sample test ( P<0 . 0001 ) . We used the PCIT algorithm [17] to establish the number of significant connections for each regulator in the P×H and W×H datasets , to determine how well this conventional approach performed in comparison to RIF . In line with previous authors we discovered that the DH axis ( i . e . , the change in the number of significant connections between the two breeds ) enriches at its extremes for transcriptional regulators ( Figure 8 ) . The extreme 1% DH ( i . e . , 110 genes out of the 11 , 057 available ) contains 15 transcriptional regulators rather than the 9 expected by chance alone ( hypergeometric p-value = 0 . 0192 ) . This enrichment is not true for the DE axis , which contains 9 transcriptional regulators . However , DH failed to capture myostatin in its extremes , which suggests its usefulness as a metric for the identification of transcriptional regulators of relevance may not be broadly applicable ( Figure 8 ) . We also ran the ‘signed’ hubbing analysis of [24] on our data and plotted the output ( Figure S1 ) . As with the PCIT DH approach , myostatin was not enriched at the extremes of the DiffK DH axis . This means it failed to identify the regulator containing the known causal mutation as being differentially behaved in the two muscle systems . In the introduction we posed a computational challenge: identify the question in P×H versus W×H muscle development to which myostatin is the answer . The subsequent analysis suggests the following: “Which transcriptional regulator is cumulatively most differentially wired to the abundant most differentially expressed genes ? ” This question is clearly very different to the conventional “which transcriptional regulator is the most differentially expressed ? ” and unsurprisingly the latter gives quite different answers , including the notable failure to identify myostatin out of the 920 candidates . This result suggests that traditional microarray approaches generating lists of DE regulators may be committing type III statistical error , the error committed when giving the right answers to the wrong questions [25] , [26] . Regulators may indeed be correctly identified as DE , but this does not mean that they are differentially activated . The converse is also true . For example , TF activity can be regulated in many ways , localisation to the nucleus , chemical modification , change in accessibility of DNA binding sites and availability of cofactors that by and large are independent of TF expression level . It is therefore inevitable that these common forms of regulatory change will be overlooked by DE analysis . The positive identification of myostatin as the major regulatory perturbation in this specific set of experimental contrasts is noteworthy , despite the stated aims of the approach . The Piedmontese causative mutation exists at the first level of organisation ( genomic DNA ) , and manifests its effect at the third ( protein ) and higher levels ( phenotype ) . Despite this , we can identify it using only data at the second level of biological organisation – the transcriptome . In addition , all animals were Hereford hybrids so 50% of the protein translated by the P×H animals was as functional as the myostatin protein translated by the W×H; in line with this , the increase in muscle mass was correspondingly subtle ( ∼9% ) ( unpublished data ) . The new algorithm works , in effect , by firstly establishing a Phenotypic Impact Factor ( PIF ) for each of the DE genes . Thus , genes that are both highly abundant and highly DE between the crosses derive a correspondingly high PIF , or discrimination factor . Taken together , this weighting provides an abstract molecular description of the phenotype perturbation specific to the treatments under consideration . In the P×H versus W×H comparison , the genes with the highest PIF ( i . e . , those that are abundant and highly DE ) tend to be slow twitch muscle structural genes ( MYL2 , MYL3 , TNNT1 , MYH7 , ACTN2 and MYOZ2 ) . This correctly reflects the most pervasive phenotypic change in Piedmontese myostatin mutants ( along with the increase in muscle mass ) namely the gross fibre type transition . We therefore conclude that DE , in the context of transcript abundance , is a powerful measure of phenotypic / anatomical change ( but not necessarily , as we have already argued , regulatory change ) . RIF is based on the cumulative , simultaneous , differential wiring ( DW ) of each regulator to the DE genes , ‘weighted’ for the PIF of each DE gene . Satisfactorily , the regulator awarded the highest RIF by this approach is myostatin , the gene that bears the known causal mutation ( SNP ) in Piedmontese genomic DNA [11] . This positive result suggests that our concept and method of assigning RIF represents a promising approach to the identification of causal mutations , and additionally those regulatory ‘hot spots’ resulting from non-genetic perturbations in other systems . The highest impact regulators are documented in Table 2 , and include known muscle master regulators like MyoD1 . A caveat: some known muscle master regulators ( e . g . Myf5 ) do not perform strongly in our analysis , i . e . , they are ascribed relatively low RIF's . This suggests not that these regulators are unimportant to bovine muscle development , but rather that they play only a small role in the rewiring that directs these two muscle phenotypes down different developmental paths . During the conceptual development of the algorithm we tried several permutations . The best performer , as described above and in the results section , incorporates the average abundance and differential expression of the DE genes ( which tend not to be transcriptional regulators ) , and the cumulative DW of the regulators to those weighted DE genes . Surprisingly , inclusion of either the average abundance or DE of the regulators themselves actually impairs the ability of the algorithm to identify myostatin ( data not shown ) . While we assessed several versions of the algorithm , there is no evidence that the data has been over-fitted because ( 1 ) the model is relatively simple compared to the data it analyses , ( 2 ) like in any other expression experiment , only the normalized gene expression levels for each gene in each of the samples ( or experimental conditions ) are needed , ( 3 ) it is built on sound mathematical principles ( mixed-ANOVA models and model-based clustering ) , and ( 4 ) those mathematical principles mesh well with our biological understanding of the behaviour of both structural proteins ( where DE and abundance are always important ) and transcriptional regulators ( where DE and abundance are not necessarily important , but transcriptional connectivity is important ) in a range of living systems . The two versions of the algorithm provided ( Eq4 and Eq5 ) are alternatives in the sense that they are built on the same set of concepts . However , at this stage , it is not clear whether one can be considered superior to the other . Consequently , we have derived our impact factor discussion from the combined , averaged output of both equations ( Table S2 and Tables 2 and 3 ) . Equation 5 has the appealing intrinsic mathematical feature that it allows for auto-regulation , a biological feature thought to be true for myostatin itself [11] . Our observations imply caution when assessing isolated DE lists of TF . That TF can behave differently in two systems without being strongly DE , has been discussed before [24] , [27] and is graphed for these data in Figures 2 , 5 , and 8 . It is interesting that the top 10 candidates generated by the new algorithm and the top 10 DE regulators do not overlap , although HNRNPD and MYOD1 lie just outside the top ten most DE regulators; the most DE regulator ( CSRP3 ) was assigned only a modest RIF . CSRP3 has been reported to be a potential structural component of the sarcomere [28] , but also as a soluble component [29] and as a TF involved in the transduction of mechanical stress signals from the structural proteins to the nucleus [30] , [31] . ANKRD1 , another major DE regulator , may have a similar role [32] , [33] . The possible structural roles of these regulators may place them in an intermediate category between structural protein and regulator , which complicates the decision to include them in a transcriptional regulator list and may also have implications for interpretation of the output . Assigning impact factors to the regulators ( based on the behaviour of its co-expression with respect to the phenotypically most relevant part of the network ) forms step 1 in a 2-step process , and it yields biologically valid results . The next step is to computationally wire up the high impact regulators into coherent transcriptional modules , whose coordinated behaviour drives the phenotype change . We attempted to do this by establishing relationships between regulators who were ‘similarly’ or co-ordinately differentially-wired between the two crosses . To our knowledge this is the first time co-differential wiring has been used for reverse-engineering regulatory circuitry . The resultant output captures the phenotypic and regulatory differences between the two crosses and so we view it as a ‘perturbation matrix . ’ The building and clustering of the perturbation matrix satisfactorily resolves both axes into biologically sensible modules . For example the DE axis generates a very tight module of high phenotypic impact slow twitch muscle fibres ( ACTN2 , MYH7 , TNNT1 , MYL3 and MYOZ2 ) . Equally , the regulator axis resolves a high impact regulatory module comprising myostatin and MYOD1 , among others . Myostatin is embedded in the middle of this high impact module . We interpret these clusters of regulatory disturbance as representing ‘hot spots’ of circuit rewiring that account for the major phenotypic changes between the crosses . The exceptionally tight coupling of myostatin and MYOD1 on the y axis is the product of a near perfect matching of co-differential wiring across all 85 DE genes ( p = 0 . 917 ) . Figure 7 illustrates the co-differential wiring of these two regulators against the slow twitch module DE genes . Unfortunately , our success in correctly determining the rewiring of myostatin in ( i . e . , identifying the other regulators through whom it communicates its effect on muscle mass and muscle fibre composition ) is harder to evaluate than the impact factor data . This is because the regulatory events that transduce myostatin's influence on Piedmontese muscle mass and fibre composition have not been well established . To establish the validity of the co-differential wiring approach we examined the biological identity of those genes with the highest co-DW coefficients . The distinction belongs to PTTG1 and TOP2A ( 0 . 994 ) which are involved in the same highly fundamental biological process , that of chromatid separation during DNA replication . With specific regard to the myostatin and MyoD1 clustering , the high co-DW congruence makes a clear prediction that the myostatin SNP in Piedmontese exerts its effect on skeletal muscle via circuit rewiring with MyoD1 . MyoD1 has not only been shown to drive the expression of a set of genes necessary for fast muscle differentiation [34] , but also to be specifically regulated by myostatin in mice [35]; thus , our prediction appears robust . The separation of abscissa clustering predicts that MYL2 is under a different regulatory program to the other slow twitch muscle structural genes in the system . When we next used the co-DW patterns to generate edges in a network , myostatin was linked to 2 other high impact regulators , MYOD1 and IFRD1 . It is highly noteworthy that IFRD1 , which is required for myoblast differentiation , forms a known , experimentally-verified regulatory circuit with MYOD1 [36] adding further support to our co-differential wiring method . Taken together , these results are very appealing because they indicate a single method that not only correctly clusters regulators who behave the same in the two systems ( PTTG1 and TOP2A ) but also those who behave differently in the two systems ( Myostatin , MyoD1 and IFRD1 ) . The additional myostatin module connections , i . e . , between myostatin and IFRD1 ( −0 . 903 ) and between MYOD1 and IFRD1 ( −0 . 925 ) ( Figure 7 ) are not represented in Figure 6 because Permut Matrix does not recognise inverted patterns ( i . e . , the signs on the edges are negative instead of positive ) . In contrast to MYOD1 , the high RIF MEF2C is not part of this cluster . The utility of this algorithm clearly relies on appropriate data selection . Presumably , the microarray data must be assayed on the right tissue and at biologically important times . However , the dataset that we analysed was not designed to address the specific question of identifying the gene containing the causal mutation , rather it was designed to study the impact of nutrition restriction of the mother on the subsequent performance of the calves [37] , [38] . A limitation of our method is that the regulators must be identified at the start of the analysis . However , other than the initial identification of the regulators , all the downstream information such as PIF , RIF and the topology of the co-differentially wired network is entirely data-driven , i . e . , computed directly from the normalised microarray expression values . Finally , during the development of the algorithm we initially attempted to determine regulatory changes via a simpler version of connectivity , i . e . , describing changes in the number of connections of each regulator , what we have termed DH ( and what is sometimes referred to as differential connectivity in the literature ) . DH approaches have previously proved useful in identifying genes that appear to play key regulatory roles in evolution [39] , cancer [40] and the development of sexual dimorphism [41] . However , the procedure is limited because it requires the application of a significance analysis to isolate the significant from the non-significant connections . Which significance analysis to use is a subject of ongoing debate with weighted networks appearing to hold the most promise [17] , [24] , [42] , [43] . While it was true that high DH ( coupled with low DE ) proved diagnostic of regulators in general ( sensu [24] , [27] ) , it performed poorly as a discriminatory metric with specific regard to myostatin ( Figure 8 ) . A possible reason why can be illustrated by the following hypothetical example . Consider a regulator with 100 positive associations in one system , and 100 negative associations in another , and two entirely different sets of target genes . In its most basic form , a DH analysis would suggest this regulator is not differentially hubbed ( as 100−100 = 0 ) , clearly a false negative . Thus , a DH analysis may suffer from ( 1 ) ignoring the identity of the connected genes , ( 2 ) ignoring the sign on the edges , and ( 3 ) ignoring the phenotypic impact of each connected gene . To further compare RIF to published network approaches , we ran the DiffK hubbing analysis of [24] , which is a sophisticated ‘signed’ differential hubbing algorithm . This positioned myostatin 147th out of the 920 regulators on the DiffK axis , i . e . , it failed to identify myostatin as behaving differently in the two muscle systems ( Figure S1 ) . This result suggests that hubbing analyses , in their various forms , are unable to identify the causal mutation in this particular data . Our definition of RIF does not require computation of the number of connections of a given regulator in each of the two networks . Therefore , algorithms for network re-construction ( weighted or otherwise ) are of no relevance . Instead , the difference between the connection weight of a given gene with each of the DE genes , accounting for PIF , appears to be sufficient . In other words , RIF has a set of refinements which make it highly sensitive . These refinements include recognising the specific identity of target genes , recognising the possible importance of ‘weak’ edges that would be deemed non-significant by other methods and recognising the phenotypic importance of the target genes . This principle is well illustrated by the DW of myostatin to MYL2 . The co-expression relationship significantly changes from +0 . 761 in the P×H system to −0 . 342 in the W×H system . The −0 . 342 Myostatin-MYL2 ‘edge’ in the Wagyu network would be unequivocally discarded by all statistical methods as being insignificant ( whether by ARACNE , PCIT or some other approach ) whilst the +0 . 761 Myostatin-MYL2 ‘edge’ in the Piedmontese would be borderline insignificant depending on the exact analysis used . Therefore , comparisons between these arrangements ( which underpin the success of our present analysis ) cannot be sensitively quantified by DH . Further , the fact that MYL2 is highly abundant and highly DE ( and therefore of great phenotypic importance ) would be overlooked by DH , unless the PIF metric was applied . It is a telling observation that myostatin is neither DE nor DH ( Figure 8 ) , yet is cumulatively the highest RIF regulator on the array by some margin ( Figure 5 ) . We have argued that the algorithms success is built on controlling type III error , i . e . , it gives the right answer because it asks the right question . The approach should be generalisable to other ‘omics data because its mathematical approaches mesh well with the known biology of regulatory and non-regulatory molecules . Unlike other causal mutation finding computational approaches [6]–[9] , RIF requires data at only one level of organisation ( the transcriptome ) . Having said this , the future availability of more complete TF binding data and other resources will enable the determination of a more complete path from causal mutation to phenotype . By extracting richer regulatory information RIF may help establish novel regulatory perturbations . These include a better understanding of the network topologies that underpin evolutionary novelty and the mis-wiring events that lead to aberrant development such as cancer . Use of animals and the procedures performed in this study was approved by the New South Wales North Coast Animal Care and Ethics Committee ( Approval No . G2000/05 ) . Hereford cows were artificially inseminated or mated to one of 5 different Wagyu sires or one of 6 different Piedmontese sires . All Piedmontese sires were homozygous for the MSTN ( GDF8 ) missense mutation in exon 3 and none of the Wagyu sires carried the mutation . We sequenced the myostatin transcript from cDNA and found it to be heterozygous for the SNP mutation in all Piedmontese samples with approximately equal peak heights for both alleles . Muscle tissue from these animals has been contrasted previously across both pre- [14] and post- [15] natal development using a custom cDNA array derived from adult muscle and adipose tissue libraries . Further details relating to experimental design can be found therein . Total RNA was prepared as previously described [14] . We used a bovine oligonucleotide microarray , developed in 2006 by ViaLactia Bioscience in collaboration with Agilent , containing 21 , 475 unique 60-mer probes , representing approximately 19 , 500 distinct bovine genes . Four microarrays are present on each Agilent chip . Issues considered in the experimental design included the availability of biological replicates as well as the quality of the extracted mRNA . The experimental layout was designed to allow a focus on the cross comparison , but to also permit a developmental aspect to be carried out ( Figure 1 ) . Two clear components were included: gestation and post-natal . For the gestation component of the experiment , a total of 12 dual-channel hybridizations were performed including three biological replicates for each cross and at the 4 time points ( 60 , 135 , 195 and 280 days ) . For the post-natal component , 36 hybridizations were performed including the same four biological replicates for each cross surveyed at six ages from 3 to 30 months old . Alternate dye channel was applied to allow accounting for systematic effects due to dye bias . Microarrays were hybridized at the SRC Microarray Facility of the Institute for Molecular Biosciences in Brisbane , Australia ( http://microarray . imb . uq . edu . au/ ) . We used a number of approaches to establish a reasonably definitive list of genes encoding proteins that directly or indirectly modify gene expression , including chromatin remodelers . We made use of a comprehensive list of TF previously published in humans [44] and identified the homologs on the Agilent bovine array . This list was augmented by examining files available at ftp://ftp . ncbi . nih . gov/gene/DATA/ which were obtained and searched by accession number to identify gene ontology information for each sequence . We also took advantage of a range of online databases with information on TF binding motifs to further corroborate the list . While we discriminated between modifiers of gene expression , such as TF , non-transcription factor regulators ( e . g . myostatin ) , chromatin remodelers ( e . g . HDAC2 ) and signalling molecules ( e . g . FRZB ) in the list , the phrase transcriptional regulator covers all together . Gene expression intensity signals were subjected to a series of data acquisition criteria based on signal to noise ratio and mean to median correlation as detailed previously [45] . In brief , we employed the following two editing criteria for data acquisition: First , we required that the signal to noise ratio ( computed by dividing the background corrected intensity by the standard deviation of the background pixels ) be greater than unity; Second , we required that the correlation between the mean and the median signal intensities ( computed by dividing the smaller of the mean or median by the larger ) to be greater than 0 . 85 . Tran et al . [46] suggested that a correlation of 0 . 85 or higher not only retains more data than other methods , but retained data are more accurate than traditional thresholds or common spot flagging algorithms . However , these criteria were applied separately for the red and for the green intensity channels so that a different number of observations for each channel were obtained . These resulted in a total of 2 , 083 , 641 gene expression intensity readings ( 1 , 027 , 379 red and 1 , 056 , 262 green ) on 11 , 057 genes that were background corrected and base-2 log transformed . The arithmetic mean and standard deviation ( in brackets ) for the red and green intensities were 8 . 67 ( 3 . 16 ) and 8 . 14 ( 2 . 82 ) , respectively . Data normalization was carried out using a linear mixed ANOVA model as described in [47] and differentially expressed ( DE ) genes identified by model-based clustering via mixtures of distributions on the normalized expression of each gene at each cross and time point as detailed in [47] , [48] . In brief , the following linear mixed-effect model was fitted to the data: ( 1 ) where Y ijkvmn represents the n-th background-adjusted , normalized base-2 log-intensity signal from the m-th gene at the v-th experimental variety treatment ( breed cross and time point ) from the i-th chip , j-th array ( i . e . , there are four microarrays per chip ) and k-th dye channel; μ is the overall mean; C represents a comparison fixed group effects defined as those intensity signals from the same chip , array and dye channel; G represent the random gene effects with 11 , 057 levels; AG , DG , and VG are the random interaction effects of array by gene , dye by gene , and variety by gene , respectively . Finally , ε is the random error term . In what follows , it is understood that the v-th variety treatment incorporates both the main class treatment of cross ( e . g . P×H versus W×H ) as well as the sub-class level ( e . g . the 10 time points ) . That is: v = 1 , 2 , … , 10 for the Piedmontese cross at the 10 time points; and v = 11 , 12 , … , 20 for the Wagyu cross also at the same 10 time points . For the random effects in model ( 1 ) , standard stochastic assumptions are:where iid denotes independently and identically distributed and N denotes the normal distribution . Variance components are between genes ( σ2 g ) , between genes within array ( σ2 ag ) , between genes within dye ( σ2 dg ) , between genes within treatment ( σ2 vg ) and within genes ( σ2 e ) . Variance components were estimated using restricted ( to zero error contrasts ) maximum likelihood ( REML; see [49] for detailed formulae ) . To determine which genes are DE between the two crosses , the following t-statistic was computed for each gene in g: ( 2 ) This definition of DE is likely to be conservative as it is based on overall variation in expression across all time points . However , it has the advantage of dealing with irregular time intervals compared with dynamic clustering methods based on autoregressive models [50] , where the time points have to be evenly spaced . Finally , the DE measurement contrasts in ( 2 ) were processed by fitting a two-component normal mixture model and posterior probabilities of belonging to the non-null component were used to identify DE genes with an estimated experiment-wise false discovery rate of <1% as described by [51] . We introduce the term differential wiring ( DW ) which , defined for every pair of genes , is computed from the difference between the co-expression correlation observed between these two genes in the Piedmontese network minus the co-expression correlation between the same pair of genes in the Wagyu network . In algebraical terms , DW is computed as follows: ( 3 ) where f and i indicate the f-th TF and the i-th DE gene , respectively; For every regulator in our dataset , we introduce a new term , namely Regulatory Impact Factor ( RIF ) which simultaneously combines the DW between the TF and each of the DE genes , weighted for the PIF of the DE genes , i . e . , their expression averaged across the two crosses ( denoted as A i , for the i-th DE genes ) and their measure of differential expression given in Equation ( 2 ) . In algebraical terms , the RIF associated with the f-th TF is computed as follows: ( 4 ) where nde denotes the number of DE genes . An alternate definition of RIF f , providing similar rankings to ( 6 ) is given by: ( 5 ) where and indicate the expression of the i-th DE gene in Piedmontese and Wagyu , respectively; and and indicate the square of the co-expression correlation between the f-th TF and the i-th DE gene in the Piedmontese and Wagyu networks , respectively . This alternate definition of RIF has the additional appealing features of expressing the product of the average and the differential expression as the difference of the squared expression in each cross ( i . e . , a computational simplification ) , as well as the squared correlations ( i . e . , coefficient of determination ) between the f-th TF and the i-th DE gene indicating the strength of one variable ( the TF ) explaining variation in a second variable ( the DE gene ) . It also allows for the existence of self-regulation which more realistically reflects biology ( i . e . , note that for DW fi = 0 for f = i; a situation where a TF is also DE , impacting on the computation of RIF as per Equation 4 ) . RIF scores were normalized to a mean of zero and a standard deviation of one . PIF is implicit in the Equation 4 representation of RIF and is defined as the product of the average and the differential expression of a gene , computed as follows: ( 6 ) Differential hubbing was calculated in two ways . Firstly , by subtracting the number of significant connections a gene has in Wagyu from the number of significant connections it has in Piedmontese where significance was established using the PCIT algorithm [17] . Secondly , we also computed a ‘signed’ DH using the network strategy detailed in [24] .
Evolution , development , and cancer are governed by regulatory circuits where the central nodes are transcription factors . Consequently , there is great interest in methods that can identify the causal mutation/perturbation responsible for any circuit rewiring . The most widely available high-throughput technology , the microarray , assays the transcriptome . However , many regulatory perturbations are post-transcriptional . This means that they are overlooked by traditional differential gene expression analysis . We hypothesised that by viewing biological systems as networks one could identify causal mutations and perturbations by examining those regulators whose position in the network changes the most . Using muscular myostatin mutant cattle as a proof-of-concept , we propose an analysis that succeeds based solely on microarray expression data from just 27 animals . Our analysis differs from competing network approaches in that we do not use significance testing to eliminate connections . All connections are contrasted , no matter how weak . Further , the identity of target genes is maintained throughout the analysis . Finally , the analysis is ‘weighted’ such that movement relative to the phenotypically most relevant part of the network is emphasised . By identifying the question to which myostatin is the answer , we present a comparison of network connectivity that is potentially generalisable .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "physiology/muscle", "and", "connective", "tissue", "developmental", "biology/developmental", "molecular", "mechanisms", "computational", "biology/systems", "biology" ]
2009
A Differential Wiring Analysis of Expression Data Correctly Identifies the Gene Containing the Causal Mutation
Neurons display a wide range of intrinsic firing patterns . A particularly relevant pattern for neuronal signaling and synaptic plasticity is burst firing , the generation of clusters of action potentials with short interspike intervals . Besides ion-channel composition , dendritic morphology appears to be an important factor modulating firing pattern . However , the underlying mechanisms are poorly understood , and the impact of morphology on burst firing remains insufficiently known . Dendritic morphology is not fixed but can undergo significant changes in many pathological conditions . Using computational models of neocortical pyramidal cells , we here show that not only the total length of the apical dendrite but also the topological structure of its branching pattern markedly influences inter- and intraburst spike intervals and even determines whether or not a cell exhibits burst firing . We found that there is only a range of dendritic sizes that supports burst firing , and that this range is modulated by dendritic topology . Either reducing or enlarging the dendritic tree , or merely modifying its topological structure without changing total dendritic length , can transform a cell's firing pattern from bursting to tonic firing . Interestingly , the results are largely independent of whether the cells are stimulated by current injection at the soma or by synapses distributed over the dendritic tree . By means of a novel measure called mean electrotonic path length , we show that the influence of dendritic morphology on burst firing is attributable to the effect both dendritic size and dendritic topology have , not on somatic input conductance , but on the average spatial extent of the dendritic tree and the spatiotemporal dynamics of the dendritic membrane potential . Our results suggest that alterations in size or topology of pyramidal cell morphology , such as observed in Alzheimer's disease , mental retardation , epilepsy , and chronic stress , could change neuronal burst firing and thus ultimately affect information processing and cognition . Neurons exhibit a wide range of intrinsic firing patterns with respect to both spike frequency and spike pattern [1]–[3] . A distinct type of firing pattern that is critically involved in neuronal signaling and synaptic plasticity is burst firing , the generation of clusters of spikes with short interspike intervals [4] . Bursts can improve the signal-to-noise ratio of neuronal responses [5] and may convey specific stimulus-related information [6] . Bursts of spikes can be more effective than single spikes in inducing synaptic long-term potentiation ( LTP ) [7] , [8] , or can even determine whether LTP or LTD ( long-term depression ) occurs [9] . In synapses with short-term facilitation , bursts can be transmitted more reliably than isolated spikes [10] . Electrophysiology , in combination with computational modeling , has elucidated the ionic mechanisms underlying intrinsic neuronal burst firing . Two main classes of mechanisms have been distinguished [4] . In so-called dendrite-independent mechanisms—responsible for bursting in thalamic relay neurons [11] , for example—the fast , spike-generating conductances and the slow , burst-controlling conductances are co-localized in the soma . Conversely , in dendrite-dependent mechanisms—involved in pyramidal cell burst firing—these conductances are distributed across the soma and dendrites , with the interaction between somatic and dendritic conductances playing an essential role in burst generation . Dendritic voltage-gated Na+ and K+ channels , which promote propagation of action potentials from the soma into the dendrites , cause the dendrites to be depolarized when , at the end of a somatic spike , the soma is hyperpolarized , leading to a return current from dendrites to soma . The return current gives rise to a depolarizing afterpotential at the soma , which , if strong enough , produces another somatic spike [12] , [13] . This whole process was described by Wang [13] as ‘ping-pong’ interaction between soma and dendrites . Although ion channels play a pivotal role in burst firing , dendritic morphology also appears to be an important factor . In many cell types , including neocortical and hippocampal pyramidal cells [14]–[17] , neuronal firing patterns and the occurrence of bursts are correlated with dendritic morphology . Results from modeling studies also suggest a relationship between dendritic morphology and firing pattern [18]–[21] . However , these studies are mainly correlative [21] , focus on morphologically very distinct cell classes [18] , use only the physiologically less appropriate stimulation protocol of somatic current injection , and do not investigate the impact of topological structure of dendritic arborizations . Consequently , the effects of dendritic size and dendritic topology on burst firing , and the underlying mechanisms , remain poorly known . Considering that dendritic morphology can undergo significant changes in many pathological conditions , such as Alzheimer [22] , [23] , mental retardation [24] , [25] , epilepsy [26] , and chronic stress [27]–[29] , it is important to examine the implications of altered morphology for neuronal firing . Using computational models of neocortical pyramidal neurons , we here explore in a systematic and rigorous way the impact of a cell's dendritic morphology on the ping-pong mechanism of burst firing , under either somatic current injection or synaptic stimulation of the apical dendritic tree . Importantly , we thereby distinguish between the effects of size and topology of the apical dendrite . Furthermore , we identify the underlying mechanism by which morphology affects burst firing in the model . We use a morphologically and biophysically realistic model of a bursting layer 5 pyramidal cell from cat visual cortex ( Fig . 1A ) that is based on [18] ( which in turn builds upon [30] ) . The model is implemented in NEURON [31] and captures the general features of bursting in pyramidal cells , particularly the interaction between soma and dendrites in burst generation [12] , [13] . In the soma , the voltage-dependent currents and associated maximal conductances ( in pS µm−2 ) are as follows: a fast sodium current , ; a slow voltage-dependent non-inactivating potassium current , ; a fast non-inactivating potassium current , ; a slow calcium-activated potassium current , ; and a high voltage-activated calcium current , . In the axon hillock , and In the apical dendrite , the conductances are as in the soma , except that . In both the soma and the dendrites , the membrane capacitance µF cm−2 , the axial resistance 80 Ω cm , and . Internal calcium concentration is computed using entry via the high-voltage activated calcium current and removal by a first order pump , where the baseline calcium concentration is 0 . 1 µM and the time constant of calcium removal is 200 ms [18] , [32] . The reversal potentials ( in mV ) are , , , and . All currents are calculated using conventional Hodgkin-Huxley-style kinetics . For the specific rate functions for each current , we refer to [18] . The pyramidal cell is activated by either somatic or dendritic stimulation . For somatic stimulation , the cell is continuously stimulated with a fixed current injection of 0 . 2 nA at the soma . For dendritic stimulation , the cell is stimulated by synapses that are regularly distributed across the apical dendrite , with a density of 1 synapse per 20 µm2 . For this synaptic density , the total input current , based on the current transfer at a single synapse , is approximately the same as with somatic stimulation . The excitatory synaptic input is mediated by AMPA receptors . The time course of conductance changes follows an alpha function , where , with the peak time ms , the peak conductance , nS , and the reversal potential mV [33] . Each synapse is randomly activated , whereby the time intervals between the activations of a synapse are drawn from a negative exponential distribution , where is the mean of the distribution . Over the time period of a complete simulation , this results in a Poisson distribution of synaptic activation times . For each synapse , the mean activation frequency is set to 1 Hz , and each synapse is activated independently of the other synapses . The firing patterns were recorded from the soma . Each simulation lasted 10000 ms , of which the first 1000 ms were discarded in the analysis in order to remove possible transient firing patterns . In studying the influence of pyramidal cell morphology on burst firing , we distinguish between dendritic size and dendritic topology . The size of a dendritic tree is the total length of all its dendritic segments . The segment between the soma and the first branch point is called the root segment ( see Fig . 2B ) . Dendritic segments between two branch points are intermediate segments , and segments between a branch point and a terminal tip are terminal segments . The topology of a dendritic tree is the way in which the dendritic segments are connected to each other . For example , a tree with a given number of terminal segments can be connected in a fully asymmetrical or a fully symmetrical way ( see Fig . 2B , dendritic trees 1 and 23 , respectively ) . To investigate how the dendritic size of the pyramidal cell influences burst firing , we varied the total length of the cell's apical dendrite according to two methods . In the first method , we successively pruned terminal segments from the apical dendritic tree . Starting with the full pyramidal cell morphology , in each round of pruning we randomly removed a number of terminal segments from the apical dendritic tree . Each terminal segment had a chance of 0 . 3 to be removed . From the reduced dendritic tree , we again randomly cut terminal segments , and so on , until in principle the whole apical dendrite was eliminated . This whole procedure was repeated 20 times . The density of synapses was kept constant during pruning , so with dendritic stimulation pruning also changed the total input to the cell . With somatic simulation , the total input to the cell did not change when the apical dendrite was pruned . In the second method , we kept the dendritic arborization intact and changed the size of the apical dendrite by multiplying the lengths of all its segments by the same factor . Thus in this way the entire apical dendritic tree was compressed or expanded . For dendritic stimulation , we kept the total synaptic input to the cell constant by adapting the density of the synapses . So , both with somatic and dendritic stimulation , the total input to the cell did not change when the size of the apical dendrite was modified . To examine the impact of the cell's dendritic branching structure on burst firing , we varied the topology of the apical dendritic tree by swapping branches within the tree . The apical dendritic trees that were generated in this way all have exactly the same total dendritic length and other metrical properties such as total dendritic surface area and differ only in their topological structure . The total input to the cell , both with somatic and dendritic stimulation , did not change when the topological structure was altered . To facilitate a systematic analysis of the role of dendritic size and dendritic topology in shaping burst firing , we also use a set of morphologically simplified neurons . The neurophysiological complexity of these neurons is similar to that of the full pyramidal cell model . For a systematic study , one must use trees with a relatively small number of terminal segments , because otherwise the number of topologically different trees becomes so large that simulating all of them becomes impossible . For a tree with only 12 terminal segments , for example , there already exist as many as 451 different tree topologies [34] . Here , we use a set of 23 neurons consisting of all the topologically different trees with 8 terminal segments ( Fig . 2 ) . The trees may also be thought of as representing the backbones of potentially much larger dendritic arborizations . All segments in the tree ( intermediate and terminal segments; see Fig . 2B ) have the same length , so that the different tree topologies do not differ in total dendritic length . In almost all types of neurons , including neocortical pyramidal cells , the diameters of dendritic segments decrease at each branch point , with terminal segments having the smallest diameter [35] , [36] . In the trees we use , the diameter of a parent segment , , is related to the diameters of its daughter segments , and , as , where the branch power e is equal to 1 . 5 ( Rall's power law ) [37] . For pyramidal cells , values of the branch power were found to range between 1 . 5 and 2 [35] , [36] . Since terminal segment diameters show only a narrow range of values [38] , all terminal segments were given the same diameter ( 0 . 7 µm ) [35] , [38] , while the diameters of intermediate segments were calculated using Rall's power law . This implies that asymmetrical topologies will have a higher total dendritic surface area than symmetrical topologies . We therefore also considered the case in which all segments in the tree have the same diameter ( 3 µm ) , so that the different tree topologies do not differ in dendritic surface area . All neurons also have a soma compartment , with a diameter and length of 14 µm [39] . The ion channel types and densities are based on Mainen and Sejnowski's [18] reduced model . In the soma , there is a fast sodium current , , and a fast non-inactivating potassium current , ( maximal conductances , in pS µm−2 ) . The dendrites contain a fast sodium current , ; a slow voltage-dependent non-inactivating potassium current , ; a slow calcium-activated potassium current , ; a high voltage-activated calcium current , ; and a leak current , . Internal calcium concentration is computed using entry via the high-voltage activated calcium current and removal by a first order pump , where the baseline calcium concentration is 0 . 1 µM and the time constant of calcium removal is 200 ms [18] , [32] . For both the soma and the dendrites , the membrane capacitance µF cm−2 and the axial resistance 80 Ω cm . The reversal potentials ( in mV ) are , , , and . All currents are calculated using conventional Hodgkin-Huxley-style kinetics . For the specific rate functions for each current , we refer to [18] . As in the pyramidal cell model with full morphological complexity , the neurons are activated by either somatic or dendritic stimulation . All the tree topologies receive the same input . For somatic stimulation , the neurons are continuously stimulated with a fixed current injection of 0 . 03 nA ( 0 . 1 nA for the non-Rall neurons , in which segment diameter is equal throughout the dendritic tree ) . For dendritic stimulation , the cells are stimulated by 600 synapses , with on each terminal or intermediate segment ( in total 15 segments for a tree with 8 terminal segments ) 40 uniformly distributed synapses . With this number of synapses , the total input current , based on the current transfer at a single synapse , is approximately the same as with somatic stimulation . The synaptic input is mediated by AMPA receptors , with the same parameters as in the full pyramidal cell model . Also as in the full pyramidal cell model , each synapse is randomly activated according to a Poisson process , with a mean activation frequency of 1 Hz . The simulations were performed in NEURON [31] . The firing patterns were recorded from the axosomatic compartment . Each simulation lasted 10000 ms , of which the first 1000 ms were discarded in the analysis in order to remove possible transient firing patterns . To examine how the size of the dendritic tree influences firing pattern , we changed the total dendritic length of a given tree topology by multiplying the lengths of all its segments by the same factor . For dendritic stimulation , the number of synapses on the tree was thereby kept constant . So , both with somatic and dendritic stimulation , the total input to the cell did not change when the size of the dendritic tree was modified . For presenting the firing patterns from the different tree topologies , we ordered the trees according to the degree of symmetry of their branching structure . To do this , we used a variant of the ranking scheme proposed by Harding [34] , [40] . A binary tree can be described by denoting at each node , i . e . , a branch point or the root , the sizes of the subtrees ( in number of terminal segments ) it carries . For example , a tree of size 1 is simply denoted as 1 . A tree of size 2 is denoted as 2 ( 1 , 1 ) . For trees of size 3 ( see Fig . 2B ) , there are two possibilities , 3 ( 1 , 2 ( 1 , 1 ) ) and 3 ( 2 ( 1 , 1 ) , 1 ) , but they do not represent topologically different trees , because the only difference is the order of the two subtrees . For trees with four terminal segments , there are two topologically different trees , 4 ( 2 ( 1 , 1 ) , 2 ( 1 , 1 ) ) and 4 ( 3 ( 2 ( 1 , 1 ) , 1 ) , 1 ) . Note that the last tree can also be written in several other ways , e . g . , 4 ( 3 ( 1 , 2 ( 1 , 1 ) ) , 1 ) . To obtain a unique notation , we applied the following two rules . First , if the subtrees at a particular node have a different size , the largest subtree is put left of the comma . So we write 3 ( 2 ( 1 , 1 ) , 1 ) instead of 3 ( 1 , 2 ( 1 , 1 ) ) . Second , to order ( sub ) trees of equal size , we consider to be larger the tree that has the highest number at the first figure in which the tree descriptions differ . So of the following two trees , 4 ( 2 ( 1 , 1 ) , 2 ( 1 , 1 ) ) and 4 ( 3 ( 2 , 1 ) , 1 ) , the second one is considered to be larger ( since 3>2 ) . Thus , in a description of an 8-terminal tree of which these two trees are the subtrees , the second subtree is put first , i . e . , 8 ( 4 ( 3 ( 2 , 1 ) , 1 ) , 4 ( 2 ( 1 , 1 ) , 2 ( 1 , 1 ) ) ) . Once all tree topologies were written in a unique form , they were ordered according to their size ( in the extended sense , as described above ) , whereby the largest one was put first in the list . Since two trees can now be ordered simply by looking at the first figure in which their descriptions differ , this ordering is called a reverse lexicographical ordering . Applied to trees of the same size , it puts trees in order of symmetry , with the most asymmetrical tree first and the most symmetrical tree last . Fig . 2A shows the ordering of the 23 topologically different trees of 8 terminal segments . Note that the ordering is only used for presentation purposes and does in no way affect the results . Bursting is defined as the occurrence of two or more successive spikes with short interspike intervals followed by a relatively long interspike interval . To quantify bursting , we used the burst measure developed in [41] . This measure is based solely on spike times and detects the correlated occurrence of one or more short ( intraburst ) interspike intervals followed by a long ( interburst ) interspike interval . It quantifies the extent of bursting in the whole spike train; it does not try to identify individual bursts , as some other approaches do [42] . The burst measure is based on the following notion ( see also Text S1 ) . If a spike train consists of spikes with independent successive interspike intervals , the variance of the sum of two successive interspike intervals [ , where is the time of the ith spike in the spike train] is twice the variance of the single interspike intervals [] . If bursting occurs , successive intervals are no longer independent , and this relation is violated . Thus , the difference between the two variances is a measure for bursting . If we divide this difference by the squared average interspike interval , we obtain a normalized burst measure ( called B ) that is sensitive only to the relative sizes of the interspike intervals and not to the average interval size: ( 1 ) where means , and stands for taking the expectation or average value of the interspike intervals between two successive spikes in the spike train . We used eqn ( 1 ) to quantify the extent of bursting in a spike train , thus taking into account all interspike intervals and to calculate the average interspike intervals and their variances . If a spike train consists of spikes with independent successive interspike intervals ( i . e . , no bursting ) , , and . Although B is a continuous measure assessing the degree of bursting and not classifying spike trains as either bursting or non-bursting , we will for practical purposes consider spike trains with a value of B below 0 . 15 as non-bursting ( see also Fig . S8 ) . In Text S1 , we derive that , for a periodic spike train with two-spike bursts , the expected value of B is ( 2 ) where is the average interspike interval between two consecutive bursts and is the average interspike interval within a burst . If , there is no bursting , and . The higher the ratio of inter- to intraburst interspike intervals , the stronger the bursting and the higher the value of . In the limiting case if goes to infinity , B goes to 1 . The burst measure in eqn ( 1 ) is a general measure for detecting bursting in a spike train , and in Text S2 we show that it is equally valid for spike trains containing bursts that consist of more than two spikes . The input conductance of a pyramidal cell with a given dendritic morphology was determined by applying a static , subthreshold current injection at the soma . The ratio of the magnitude of the injected current to the resulting change in membrane potential at the soma is defined as the input conductance of the cell [43] . The input conductance is the reverse of the input resistance . To quantify the electrotonic extent of a dendritic tree , we introduce a new measure called mean electrotonic path length ( MEP ) . For a given terminal segment ( see Fig . 2B ) , the electrotonic path length is the length ( normalized to the electrotonic length constant ) of the path from the tip of the segment to the soma . This electrotonic path length is determined for each terminal segment , and the sum of all electrotonic path lengths is divided by the total number of terminal segments to obtain the MEP of the dendritic tree . More precisely , to obtain the MEP of a dendritic tree , we first normalize the length of each terminal , intermediate or root segment i ( see Fig . 2B ) with respect to its electrotonic length constants , yielding a dimensionless electrotonic length [44] , in which is defined as [43] ( 3 ) where is the radius of dendritic segment i , and and are constants denoting the specific membrane resistance and the intracellular resistivity , respectively . The MEP of a dendritic tree with terminal segments is then given by ( 4 ) where is the sum of the electrotonic lengths of all the dendritic segments in the path from the tip of terminal segment to the soma . The analysis and model code for this paper including a tool for NEURON parameter scanning is available from ModelDB at http://senselab . med . yale . edu/modeldb via accession number 114359 . To facilitate a better understanding of our findings obtained with the pyramidal cell model and to analyse more precisely the role of dendritic morphology in shaping burst firing , we also investigated a set of 23 morphologically simplified neurons consisting of all the topologically different trees with 8 terminal segments . Because the cells have relatively few terminal segments , the impact of dendritic topology on burst firing can be studied in a systematic way . Given the crucial role of bursts of action potentials in synaptic plasticity and neuronal signaling , it is important to determine what factors influence their generation . Using a standard compartmental model of a reconstructed pyramidal cell [18] , we have investigated how size and topology of dendritic morphology affect intrinsic neuronal burst firing . We have shown that either shortening or lengthening the apical dendrite tree beyond a certain range can transform a bursting pyramidal cell into a tonically firing cell . Remarkably , altering only the topology of the dendritic tree , whereby the total length of the tree remains unchanged , can likewise shift the firing pattern from bursting to non-bursting or vice versa . Moreover , both dendritic size and dendritic topology not only influence whether a cell is bursting or not , but also affect the number of spikes per burst and the interspike intervals between and within bursts . The influence of dendritic morphology on burst firing is attributable to the effect dendritic length and dendritic topology have , not on input conductance , but on the spatial extent of the dendritic tree , as measured by the mean electrotonic path length between soma and distal dendrites . For the spatiotemporal dynamics of dendritic membrane potential to generate burst firing , this electrotonic distance should be neither too small nor too large . Because the degree of symmetry of the dendritic tree also determines mean electrotonic path length , with asymmetrical trees having larger mean path lengths than symmetrical trees , dendritic topology as well as dendritic size affects the occurrence of burst firing . In Mainen and Sejnowski's [18] two-compartment model for explaining the role of dendritic morphology in shaping firing pattern , the spatial dimension of morphology was completely reduced away . Although the model is able to reproduce a wide range of firing patterns , it does not capture the essential influence of dendritic morphology on burst firing , in which , as we have shown here , the spatial extent of the dendritic tree and the resulting spatiotemporal dynamics of the dendritic membrane potential are crucially involved . The effect of dendritic size and topology on burst firing and the correlation of burst firing region with mean electrotonic path length are robust to changes in model properties , including morphology , strength of input stimulus , ion channel densities , and keeping the number of ion channels constant as morphology is changed . The specific range of dendritic sizes that supports burst firing , as well as the impact of dendritic topology , does not strongly depend on the strength of the input stimulus , especially with somatic stimulation ( Figs . S1 and S2 ) . More importantly , the overall way in which dendritic morphology influences burst firing is independent of stimulus strength . Likewise , the impact of dendritic morphology is qualitatively insensitive to the value of the branch power used in the morphologically simplified cells: even in dendritic trees in which the segment diameters are uniform we observe the same effect of dendritic length and topology ( Fig . 8 ) . In changing dendritic size or topology , we held the density of ion channels constant ( i . e . , the conductances were fixed ) , which implies that the total number of ion channels also changed when dendritic morphology was varied . Keeping the conductances fixed seems biologically the most appropriate choice , since removing membrane to shrink the dendritic tree ( as well as adding membrane to enlarge it ) will include the membrane's ion channels and is therefore not expected to affect ion channel density . But even if we hold the number of ion channels constant , by adjusting the values of the conductances as the surface area of the dendritic tree is changed when dendritic topology or total length is varied , we obtain surprisingly similar results ( Fig . S3 ) . Although the precise values of mean electrotonic path length that support burst firing are slightly different , the overall effect of dendritic size and topology on burst firing and the correlation of burst firing region with mean electrotonic path length remain the same . Since recent studies have shown that the same firing patterns can be produced by different combinations of conductances [47] , [48] and even by different combinations of conductances and morphological properties [49] , it is important to ensure that our results are not specific for the particular choice of conductance values in the Mainen and Sejnowski model [18] . Provided the model supports the ping-pong mechanisms of burst firing , our main findings are indeed robust to considerable changes in ion channel densities , both under somatic and under dendritic stimulation . Although the range of tree sizes that supports burst firing may be different for different ion channel densities , Figs . S4 , S5 , S6 , S7 show that the general impact of dendritic size and topology on burst firing , as well as the correlation of burst firing with mean electrotonic path length , is maintained for a wide range of dendritic ion channel densities . Interestingly , the value of the mean electrotonic path length where burst firing commences , going from small to large trees , is not affected by ion channel density , as opposed to the value of the mean electrotonic path length where burst firing stops . This suggests that when the dendritic tree is reduced in size so that the cell no longer exhibits burst firing , compensatory changes in ion channel conductances [49] may not be able to bring back the cell to a bursting mode . In contrast , when the dendritic tree is enlarged beyond the range where the cell bursts , compensatory changes in ion channel conductances ( e . g . , increased dendritic ) may be able to restore burst firing . Since it has experimentally been shown that removal of the apical dendrite abolishes bursting in layer 5 pyramidal cells [50] and pathological conditions often affect the apical dendrite , but not the basal dendrites [29] , we focused in this study on the morphology of the apical dendritic tree . The simulations with the morphologically simplified cells , which do not have basal dendrites , show that basal dendrites are not essential for burst firing . In the pyramidal cell and the simplified cells , burst firing is similarly affected by dendritic morphology , which again emphasizes the robustness of our findings . Compared with other modeling studies investigating the relationship between dendritic morphology and firing pattern [18]–[21] , [51] , our study is unique in that it focuses on burst firing , adopts a systematic approach , investigates morphological changes within a cell type , considers not only somatic stimulation but also physiologically more appropriate dendritic stimulation , and especially examines the impact of topological structure of dendritic arborizations . Moreover , our study is not just correlative but provides insight into the mechanisms underlying the influence of morphology on firing pattern . We stimulated the cells either by a current injection at the soma , as is done in most experimental and modeling studies [18] , or by synapses distributed over the dendritic tree , which is physiologically more relevant . Importantly , we found that the influence of dendritic morphology on burst firing is essentially the same under both stimulation regimes . Our study confirms a suggestion by Krichmar et al . [21] that dendritic branching structure might have a direct influence on neuronal firing activity . In simulation studies , they found that although dendritic size could account for much of the differences in neuronal firing behavior between CA3 pyramidal cells , it did not provide a complete explanation for the observed electrophysiological variability . Our results are in accord with empirical observations suggesting that pyramidal cells should have reached a minimal size to be capable of burst firing . In weakly electric fish , the tendency of pyramidal cells to fire bursts is positively correlated with the size of the cell's apical dendritic tree [52] . In rat prefrontal cortex [16] and visual cortex [14] , [53] , the classes of pyramidal cells that exhibit burst firing have a greater total dendritic length than the other classes . In addition , the developmental time course of bursting shows similarities with that of dendritic morphology . In rat sensorimotor cortex , the proportion of bursting pyramidal cells progressively increases from postnatal day 7 onwards , while at the same time the dendritic arborizations become more complex [54] . In pyramidal cells from rat prefrontal cortex , the total lengths of apical and basal dendrites increase dramatically between postnatal days 3 and 21 , with neurons capable of burst firing appearing only from postnatal day 18 onwards [55] , [56] . Direct experimental testing of the influence of dendritic morphology on burst firing could be done by physically manipulating the shape or size of the dendritic tree , e . g . , by using techniques developed by Bekkers and Häusser [50] . They showed that dendrotomy of the apical dendrite indeed abolished bursting in layer 5 pyramidal cells . Dendritic morphology can undergo significant alterations in many pathological conditions , including chronic stress [27]–[29] , [57] , epilepsy [26] , hypoxic ischemia [58] , Alzheimer [22] , [23] , and disorders associated with mental retardation [24] , [25] . Functional consequences of these morphological changes are usually interpreted in terms of loss or formation of synaptic connections as a result of a diminished or expanded postsynaptic surface area . Our modeling results indicate that alterations in dendritic morphology can directly modify neuronal firing , irrespective of changes in total synaptic input . Chronic stress , as well as daily administration of corticosterone , induces extensive regression of pyramidal apical dendrites in hippocampus [27] , [57] , [59] and prefrontal cortex [28] , [29] . As a result of a decrease in the number and length of terminal branches , the total apical dendritic length can reduce by as much as 32% [29] , while basal dendrites are not affected . Similarly large alterations have been observed in response to mild , short-term stress [60] . Our results predict that stress and the accompanying reduction in apical dendritic length could turn a bursting neuron into a non-bursting one . Indeed , Okuhara and Beck [61] found that two weeks of high corticosterone treatment caused a decrease in the relative number of intrinsically bursting CA3 pyramidal cells . Since burst firing of CA3 pyramidal cells is critically involved in LTP [62] , this could have profound functional consequences for hippocampal information processing [63] . With regard to epilepsy , a significant decrease in total dendritic length and number of branches has been found in pyramidal cells following neocortical kindling [26] . In line with our results , Valentine et al . [64] reported that activity of single cells recorded from the primary auditory cortex of kindled cats showed a reduction in the amount of burst firing and a decrease in the number of spikes per burst . In Alzheimer's disease , various aberrations in dendritic morphology have been observed— including a reduction in total dendritic length and number of dendritic branches [22] , [23] and alterations in the pattern of dendritic arborization [65]—which may contribute to the abnormal neurophysiological properties of Alzheimer pyramidal cells [66] . The anomalies in morphology could influence the cells' ability to produce burst , and , because of the role of burst firing in LTP and LTD [8] , [9] , ultimately affect cognition . In disorders related with mental retardation , the observed alterations in dendritic length and pattern of dendritic branching [24] , e . g . , changes in the degree of symmetry of the apical dendrite [67] , may likewise be hypothesized to contribute to impaired cognition . Importantly , our results indicate that dendritic sprouting—which too has been observed in Alzheimer [68] , [69]—may also be able to change neuronal burst firing . Since firing patterns characteristic of different classes of neurons may in part be determined by total dendritic length , we expect on the basis of our results that a neuron may try to keep its dendritic size within a restricted range in order to maintain functional performance . Indeed , Samsonovich and Ascoli [70] have shown that total dendritic size appears to be under intrinsic homeostatic control . Statistically analyzing a large collection of pyramidal cells from hippocampus and prefrontal cortex , they found that , for a given morphological class and anatomical location , fluctuations in dendritic size in one part of a cell tend to be counterbalanced by changes in other parts of the same cell , so that the total dendritic size of each cell is conserved . We predict that dendritic topology may similarly be protected from large variations . In fact , there could be a trade-off between dendritic size and dendritic topology . In a set of bursting pyramidal cells , we expect that apical dendritic trees with a lower degree of symmetry are shorter in terms of total dendritic length or have thicker dendrites to reduce electrotonic length than those with a higher degree of symmetry . Although changes in dendrite morphology of pyramidal cells have been observed in response to environmental enrichment [71] and learning [72] , recent long-term in vivo imaging studies have demonstrated remarkable stability of dendrites in adult animals [73] , [74] . Like the homeostatic control of dendritic size , this stability may point to the functional relevance of dendritic topology . An intriguing possibility is that firing pattern and dendritic morphology could mutually tune each other during development , as a result of a reciprocal influence between dendritic growth and neuronal activity . Dendritic morphology affects firing pattern , and neuronal activity in turn is known to modulate dendritic growth and branching [75] , with , interestingly , firing frequency and firing pattern having distinct effects on outgrowth [76] . As our study underscores , differences in neuronal firing properties may not necessarily reflect differences in ion channel composition . In some cases , variability in dendritic morphology may even have a relatively bigger effect on firing pattern than variability in membrane conductances [49] , [77] . Our results show that alterations in either the size or the topology of dendritic arborizations , as have been observed in many pathological conditions , could have a marked impact on pyramidal cell burst firing and , because of the critical role of bursting in neuronal signaling and synaptic plasticity , ultimately affect cognition .
Neurons possess highly branched extensions , called dendrites , which form characteristic tree-like structures . The morphology of these dendritic arborizations can undergo significant changes in many pathological conditions . It is still poorly known , however , how alterations in dendritic morphology affect neuronal activity . Using computational models of pyramidal cells , we study the influence of dendritic tree size and branching structure on burst firing . Burst firing is the generation of two or more action potentials in close succession , a form of neuronal activity that is critically involved in neuronal signaling and synaptic plasticity . We found that there is only a range of dendritic tree sizes that supports burst firing , and that this range is modulated by the branching structure of the tree . We show that shortening as well as lengthening the dendritic tree , or even just modifying the pattern in which the branches in the tree are connected , can shift the cell's firing pattern from bursting to tonic firing , as a consequence of changes in the spatiotemporal dynamics of the dendritic membrane potential . Our results suggest that alterations in pyramidal cell morphology could , via their effect on burst firing , ultimately affect cognition .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[ "neuroscience/theoretical", "neuroscience", "computational", "biology/computational", "neuroscience" ]
2010
Impact of Dendritic Size and Dendritic Topology on Burst Firing in Pyramidal Cells
The goal of influenza-like illness ( ILI ) surveillance is to determine the timing , location and magnitude of outbreaks by monitoring the frequency and progression of clinical case incidence . Advances in computational and information technology have allowed for automated collection of higher volumes of electronic data and more timely analyses than previously possible . Novel surveillance systems , including those based on internet search query data like Google Flu Trends ( GFT ) , are being used as surrogates for clinically-based reporting of influenza-like-illness ( ILI ) . We investigated the reliability of GFT during the last decade ( 2003 to 2013 ) , and compared weekly public health surveillance with search query data to characterize the timing and intensity of seasonal and pandemic influenza at the national ( United States ) , regional ( Mid-Atlantic ) and local ( New York City ) levels . We identified substantial flaws in the original and updated GFT models at all three geographic scales , including completely missing the first wave of the 2009 influenza A/H1N1 pandemic , and greatly overestimating the intensity of the A/H3N2 epidemic during the 2012/2013 season . These results were obtained for both the original ( 2008 ) and the updated ( 2009 ) GFT algorithms . The performance of both models was problematic , perhaps because of changes in internet search behavior and differences in the seasonality , geographical heterogeneity and age-distribution of the epidemics between the periods of GFT model-fitting and prospective use . We conclude that GFT data may not provide reliable surveillance for seasonal or pandemic influenza and should be interpreted with caution until the algorithm can be improved and evaluated . Current internet search query data are no substitute for timely local clinical and laboratory surveillance , or national surveillance based on local data collection . New generation surveillance systems such as GFT should incorporate the use of near-real time electronic health data and computational methods for continued model-fitting and ongoing evaluation and improvement . Influenza remains a paradox for public health: While influenza epidemics are expected seasonally in temperate climates , their exact timing and severity remain largely unpredictable , making them a challenge to ongoing preparedness , surveillance and response efforts [1] . Surveillance efforts for influenza seek to determine the timing and impact of disease through characterizing information on reported illnesses , hospitalizations , deaths , and circulating influenza viruses [2] . Since establishment of the first computerized disease surveillance network nearly three decades ago [3]–[5] , the use of information and communications technology for public health disease monitoring has progressed and expanded . During the last decade , the use of electronic syndromic surveillance systems have allowed for automated , detailed , high volume data collection and analysis in near-real time [6]–[9] . In parallel , novel approaches based on influenza-related internet search queries have been reported to yield faster detection and estimation of the intensity of influenza epidemics [10]–[16] . The public health utility of such systems for prospective monitoring and forecasting of influenza activity , however , remains unclear [17]–[21] , particularly as occurred during the 2009 pandemic and the 2012/2013 epidemic season [22]–[24] . In November 2008 , Google began prospectively monitoring search engine records using a proprietary computational search term query model called Google Flu Trends ( GFT ) to estimate national , regional and state level ILI activity in the United States ( US ) [12] . The goal of GFT was to achieve early detection and accurate estimation of epidemic influenza intensity [13] . The original GFT model was built by fitting linear regression models to weekly counts for each of the 50 million most common search queries , from the billions of individual searches submitted in the US between 2003 and 2007 [13] . An automated query selection process identified the exact text searches that yielded the highest correlations with national and regional influenza-like-illnesses ( ILI ) surveillance in the US during the period of model fitting; the top scoring 45 search terms constituted the original GFT ILI search definition . The GFT search algorithm was revised in the autumn of 2009 , following the emergence and rapid spread of the pandemic A/H1N1pdm09 influenza virus in the US , which had gone wholly undetected by the GFT system . The updated GFT model used surveillance data from the first 20 weeks of the pandemic and a qualitative decision process with less restrictive criteria for additional ILI-related search terms to be included [14] . By September 2009 the historical GFT model was replaced with retrospective estimates from the revised algorithm . Currently , the updated GFT model provides real-time estimates of influenza intensity at three geographic scales in the US: national , state and select local cities , as well as estimates for many countries worldwide [16] . The original and updated GFT models have both shown high retrospective correlation with national and regional ILI disease surveillance data [13] , [14]; however , the prospective accuracy of this surveillance tool remains unclear , even though GFT estimates are used in forecasting models for influenza incidence [15] , [20] , [21] . We present a comparative analysis of traditional public health ILI surveillance data and GFT estimates for ten influenza seasons to assess the retrospective and prospective performances of GFT to capture season-to-season epidemic timing and magnitude . We compared weekly ILI and GFT data from June 1 , 2003 through March 30 , 2013 , a period of ten influenza seasons which included a range of mild and moderately severe seasonal influenza epidemics as well as the emergence of the first influenza pandemic in over forty years . The surveillance systems were assessed at three geographical levels: entire US , Mid-Atlantic region ( New Jersey , New York and Pennsylvania ) and New York City . All public health surveillance data used in the study came from systems operating prospectively on a daily or weekly basis throughout the study period [2] , [25]–[27] . Nationwide and regional ILI surveillance data were compiled from the US Centers for Disease Control and Prevention ( CDC ) sentinel ILI-Net surveillance system , which includes sources ranging from small physician practices to large electronic syndromic surveillance networks [2] . The CDC ILI-Net system is publically available each week , typically on Friday for the previous week ending Saturday during the respiratory season ( October to May ) , with a recognized reporting lag of 1–2 weeks [2] , [13] . Local ILI data came from the New York City Department of Health and Mental Hygiene ( DOHMH ) emergency department ( ED ) syndromic surveillance system , which is collected and analyzed daily , with a reporting lag of about one day [25]–[27] . In each system , all weekly public health surveillance ILI proportions were calculated as total ILI visits divided by all visits each week . Internet search query data came from the original [13] and updated GFT models [14] , using weekly estimates available online [16] from both the periods of retrospective model-fitting ( 4 seasons for the original model and 6 seasons for the updated model ) and prospective operation for both models ( 1 season and 4 seasons , respectively; Table 1 ) . Finalized weekly GFT estimates were publically available each Sunday for the previous week , with a reporting lag of about one day . The original and updated GFT models used scaled measures of ILI-related searches to be directly comparable to the weighted ILI proportions from the CDC ILI-Net system [2] , [13] , [14] , [16] ( Figure 1 ) . For additional details on data sources , see Supporting Information . All observed ILI weekly proportions were analyzed with a traditional Serfling regression approach to establish weekly expected baselines and estimate the “excess” ILI proportions attributable to influenza and identify epidemic periods ( [28]–[33]; Supporting Information ) . The GFT system presents ILI search query estimates as a qualitative measure of influenza activity on a scale ranging from “minimal” to “intense” each week [16]; neither GFT model provided quantitative measure for detection or estimation of impact [13] , [14] . For all public health surveillance and GFT estimates we assessed two epidemiological criteria to characterize influenza outbreaks: epidemic timing and intensity . Timing was based on estimates of epidemic onset and peak week for each season and ILI surveillance system . The onset each season was defined as the first of consecutive weeks exceeding the surveillance threshold ( upper limit of the 95% confidence interval of the Serfling baseline ) . The peak week was identified as the week with the greatest proportion of ILI visits each season or epidemic ( Table 2 ) . For each data source and season we assessed epidemic intensity by determining the proportion of excess ILI for peak weeks and by summing the weekly excess ILI proportions for each epidemic period as a measure of cumulative ILI intensity for each season and epidemic . All Serfling regression confidence intervals represented the upper and lower 95% limit , calculated as the predicted non-epidemic baseline ±1 . 96 standard deviations [28]–[33] . We calculated the ratio of excess GFT divided by excess ILI at each geographic level for each epidemic ( Table 3 ) , with a constant ratio indicating consistent influenza monitoring by GFT for the period . To further evaluate the week-to-week accuracy and timing of GFT and potential asynchrony with traditional ILI surveillance , we calculated Pearson correlations in the national , regional and local datasets , following the original methods used in the development [13] and evaluation of GFT [14] . Original and updated GFT model estimates were assessed for the periods of retrospective model-fitting and prospective monitoring ( Table 2 ) , and for specific epidemic seasons ( Table 4 ) . We measured cross-correlations at negative and positive lags for each influenza season to identify the corresponding lead or lag with the highest correlation values between GFT and traditional ILI systems , indicating the degree of shift in the timing of the GFT trends compared to ILI surveillance . While correlations are useful to assess GFT [14] , they only provide a measure of relative correspondence between ILI and internet search systems , and do not provide an indication of the nature of the relationship between the trend estimates or the observed lags . As a complementary measure , we compared the regression slope of public health ILI data with GFT estimates during retrospective model-fitting and prospective periods , and for specific seasons . For further details , see Supporting Information . Historical estimates from the original GFT model were based on the model-fitting period from September 28 , 2003 to March 17 , 2007; the system was evaluated during March 18 , 2007 to May 11 , 2008 , and has run prospectively since then . The week-to-week GFT estimates during the model-fitting period were highly correlated with ILI surveillance data at the national ( R2 = 0 . 91 ) , regional ( Mid-Atlantic , R2 = 0 . 79 ) and state/local level ( New York , R2 = 0 . 89; Table 1 ) . Similarly , GFT estimates were highly correlated with CDC ILI surveillance at the national and regional levels during the validation period [13] , and remained high through the period of prospective use prior to the emergence of the 2009 A/H1N1 pandemic , from May 12 , 2008 to March 28 , 2009 ( R2≥0 . 75; Table 4 ) . Seasonal and epidemic onset and peak weeks varied considerably during the period ( Table 2 ) . Estimation of excess ILI visits and GFT search query fractions were also well correlated on a week to week basis during this period ( Supporting Tables; Figure 2 ) . In late-April 2009 , detection of novel A/H1N1 influenza in an outbreak in Queens , New York , was immediately followed by a spike in ILI surveillance data across much of the nation during the week ending May 2 , 2009 [2] . Mid-Atlantic States and New York City experienced a substantial spring pandemic wave ( Figure 1B , C ) , unlike many other regions of the US [2] . Despite recognized pandemic activity , the national GFT estimates were below baseline ILI levels for May–August 2009 , indicating no excess impact ( red line , shaded 2009 period , Figure 1A ) . The correlations between the surveillance ILI and GFT estimates , however , were very high during this period at the US level for observed ( R2 = 0 . 91 ) as well as estimated excess values ( R2 = 0 . 81; Figure 2A ) . At the Mid-Atlantic level , correlations were lower for observed ( R2 = 0 . 51 ) , but still high for estimated excess values ( R2 = 0 . 80 ) , while the slope of the linear relationship between the two surveillance systems was near zero ( slope = 0 . 11 ) , indicating that there was little or no excess ILI estimated by GFT ( Figure 2B ) . The discrepancy at the Mid-Atlantic level was exacerbated for New York City , where the pandemic impact was greater than any other epidemic that decade , while the original GFT estimates remained near expected baseline levels for the entire period ( R2 = 0 . 78 ) . Accordingly , the slope of the GFT regression against ILI was near zero ( slope = 0 . 05 ) , indicating that GFT data did not accurately measure the intensity of the pandemic ( Figure 2C ) . Taken together , the original GFT model missed the spring 2009 pandemic wave at all levels ( Figure 1 ) , providing incidence estimates 30–40 fold lower than those based on ILI surveillance ( Table 3 ) . The original and updated GFT estimates appeared very similar during the pre-pandemic period 2003–2009 , but diverged considerably by May 2009 ( red and blue lines , Figure 1 ) . Like the original GFT model , the updated GFT estimates during the model-fitting period were highly correlated with CDC ILI surveillance at the national and regional levels ( R2≥0 . 77 , Table 1 ) . In contrast for New York City , the updated GFT estimates were less well correlated with local ILI syndromic surveillance data during this period ( R2 = 0 . 51 , Table 1 ) . Of particular interest is the retrospective characterization of the 2009 pandemic by the updated GFT algorithm , which tracked the spring wave very well at the national level , but underestimated the magnitude at the regional level by nearly 30% , and at the New York City level by 70% ( Figure 1; Table 3 ) . In September 2009 , the updated GFT algorithm began running prospectively , providing estimates that tracked CDC ILI surveillance data well for the remainder of 2009 , a period in which most pandemic A/H1N1 infections occurred . Updated GFT estimates were highly correlated with ILI surveillance at the national ( R2 = 0 . 98 ) , and regional ( R2 = 0 . 92 ) levels ( Figure 1A–B; Table 4 ) . Mid-Atlantic ILI surveillance , however , demonstrated two peaks , consistent with different timing of pandemic waves in states within the region ( Figure 1B ) . For New York City , the updated GFT estimates and ILI surveillance were less well correlated when measured directly ( R2 = 0 . 51 ) , though highly correlated when lagged by three weeks ( R2 = 0 . 89 ) , showing the updated GFT model estimates for the fall 2009 pandemic wave to increase and peak 3 weeks earlier than ILI surveillance ( Figure 1C; Table 4 ) . Overall , GFT underestimated the cumulative ILI incidence of the main pandemic period , May–December 2009 , by 52% for New York City ( 25% for the broader region ) , with non-overlapping confidence intervals between the GFT and ILI surveillance systems ( Table 3 ) . Correlations between the updated GFT model and ILI data during the first two years of prospective post-pandemic surveillance were high at the national level during the 2010/2011 ( R2 = 0 . 95 ) and 2011/2012 ( R2 = 0 . 88 ) seasons ( Table 4 ) . At the regional level , there was high correlation in 2010/2011 ( R2 = 0 . 83 ) with a slight underestimation of incidence , and low correlation in 2011/2012 ( R2 = 0 . 37 ) with a slight overestimation of ILI incidence ( Figure 1B ) . At the New York City level , updated GFT estimates for 2010/2011 were reasonably well correlated with observed ILI ( R2 = 0 . 74 ) , though with ILI surveillance increasing and peaking earlier ( Figure 1C ) , and showing an improved lagged correlation ( R2 = 0 . 80 , lagged 1 week; Table 4 ) . For the relatively early and moderately severe 2012/2013 epidemic season , observed GFT estimates greatly overestimated the initial onset week and magnitude of the outbreak at all three geographical levels ( Figure 1; Table 2 ) . The correlations between the updated GFT model estimates and ILI surveillance , however , were very high at all levels ( R2≥0 . 86 , Table 4 ) . GFT model estimates of epidemic intensity were far greater than ILI surveillance data at the national ( 268% ) , regional ( 208% ) and local ( 296% ) levels ( Table 3 ) . Accordingly , the slopes of the weekly regression of ILI surveillance against GFT estimates during 2012/2013 ( United States , slope = 1 . 91; Mid-Atlantic , slope = 2 . 29; New York City , slope = 2 . 63 ) were far greater than those for other epidemic and pandemic seasons ( Figure 3 ) , and substantially different from a slope of 1 ( p<0 . 05 ) . Following Google's development of GFT in 2008 , and the considerable excitement generated by the original publication and release of the online tool [12] , [13] , [16] , concerns were raised regarding the tenuous relationship between internet searches and the presentation of illness to clinical or emergency medical providers [17] . We used clinical ILI surveillance data at local , regional and national scales as a proposed “ground truth” to test the ability of GFT to perform as a timely and accurate surveillance system in the US . We identified substantial errors in GFT estimates of influenza timing and intensity in the face of pandemic and seasonal outbreaks , including prospectively missing the early wave of the 2009 pandemic and overestimating the impact of the 2012/2013 epidemic . Although we are not the first to report issues in GFT estimates for seasonal and pandemic influenza [22] , our study is the first to carefully quantify the performance of this innovative system over a full decade of influenza activity and across three geographical scales . The 2009 A/H1N1 pandemic is a particularly important case study to test the performance of GFT , with its unusual signature pandemic features of out-of-season activity in the spring of 2009 , atypical ( young ) age pattern of cases , recurring waves and substantial geographic heterogeneity [34]–[38] . Immediately following the spread of the pandemic virus in the US , public health officials and electronic surveillance networks found that local and state level surveillance data did not correspond with estimates provided by the original GFT model , particularly in some urban areas and harder hit regions of the Northeastern and Midwestern US [18] , [39] . Clearly , the original GFT algorithm was not able to track sentinel ILI patterns that deviated from the influenza seasons that occurred during the model-fitting period . Even after the GFT algorithm was revised in September 2009 , we have shown that the retrospective estimates for the spring 2009 pandemic wave were still not in agreement with regional and local surveillance . Further , the updated GFT model that has been used prospectively failed to accurately capture the autumn 2009 pandemic wave in New York City , presenting the outbreak three weeks before it actually occurred . This assessment echoes earlier concerns regarding the timeliness and accuracy of internet search data for public health monitoring at the local level [17] and during the early wave of the 2009 pandemic [18] . To have missed the early wave of the 2009 pandemic is a serious shortcoming of a surveillance system – as these are times when accurate data are most critically needed for pandemic preparedness and response purposes . Although the GFT system provided relatively accurate estimates during post-pandemic years which were characterized by mild influenza activity , it overestimated the 2012/2013 epidemic by 2–3 fold relative to traditional ILI surveillance systems , across national , regional and local geographical levels in the US ( see also [22] ) . While the intensity of the 2012/2013 influenza season was roughly comparable to the 2003 A/H3N2-Fujian epidemic as measured by traditional surveillance and assessed by CDC as “moderately severe” [2] , the 2012/2013 season was scored by the GFT tool as by far the most severe epidemic in over a decade . A limitation of our study is its focus on US systems . Many international syndromic , physician consultation , laboratory and internet survey surveillance systems provide rapid , detailed and accurate influenza-related surveillance [3]–[5] , [40]–[48] . These systems allowed for development of GFT search query algorithms which were trained to mimic the specific regional influenza-related patterns [16] . While international GFT search query estimates are publically available earlier than many government run surveillance systems , it is important to note that public health data typically undergo monitoring for data quality and investigation prior to public release . It is also important to note that GFT has been set up where robust surveillance systems already exist , providing ILI search query data for populations that are already under surveillance . An additional limitation of our study is the imperfect nature of our assumed “ground truth” surveillance . Our study sought to assess the ability of GFT to estimate physician consultation and syndromic ILI surveillance patterns , not necessarily the true incidence of influenza infection and illness . We recognize that physician sentinel and syndromic data can be biased , particularly during periods of heightened public health concern . This has been well described in a study of online survey data and health-seeking behavior during the two waves of the 2009 pandemic in England [48] . This recognized bias highlights the need for multiple sources of surveillance information in the community . In a previous evaluation of GFT , the authors and engineers at Google and the US CDC concluded that their original GFT model had “performed well prior to and during” the 2009 pandemic , when assessed as simple correlations at national and regional levels [14] . Regarding this measure of performance , however , we found the use of simple correlation to be inadequate , as values greater than 0 . 90 often occurred during periods when critical metrics such as peak magnitude and cumulative ILI revealed that the GFT models were actually greatly under- or over-estimating influenza activity . Our study demonstrates that simple correlation measures can mischaracterize the performances of a novel surveillance system , and instead we recommend the use of additional and alternative metrics based on estimates of onset and peak timing and cumulative intensity of influenza epidemics . Because the search algorithm and resulting query terms that were used to define the original and updated GFT models remain undisclosed , [13] , [14] , it is difficult to identify the reasons for the suboptimal performance of the system and make recommendations for improvement . Concerns were raised early-on that the data-mining nature of GFT might over-fit the historical data and introduce bias in prospective use [17] . After the original GFT model missed the spring 2009 pandemic wave – an outbreak with different timing and characteristics than the outbreaks present in the retrospective model-fitting period – the GFT algorithm was modified , potentially addressing the possible over-fitting issue . The revised GFT model , however , appeared to be susceptible to bias in the opposite direction , possibly due to changes in health information searching and care seeking behavior driven by the media . Further , important epidemiologic information such as patient age , location , illness complaint or clinical presentation remain un-available in GFT ( an adult person could be performing a search on behalf of a sick minor in another state ) . In contrast , public health information systems are less prone to such biases , as they collect demographic and geographic data as well as additional health outcomes , which can be used to investigate atypical signals . Ultimately , public health actions are taken locally . As such , the accuracy and timeliness of local disease surveillance systems are critical; as is the utility of the information in supporting decisions . The additional detail in local syndromic ILI surveillance data , and its direct link to individuals seeking care , facilitates public health action . Computerized surveillance , such as the New York City syndromic chief complaint ED system , can accurately capture the impact of influenza activity [25] , [26] . In the present study , we have shown that these systems are more accurate than , yet equally timely as the GFT tool , which indicates the need for further research and support for computerized local disease surveillance systems . We believe there is a place for internet search query monitoring in disease surveillance , and for continued research and development in this area [13]–[21] , [49]–[58] . For now , in the US CDC's national and regional ILI surveillance data remain the “ground truth” source of influenza activity at national and regional levels , but timeliness , detail and coverage remain issues . Thus , we believe there is a broader need for electronic clinically-based disease surveillance at the local level , similar to the ED system in place in New York City [25]–[27] , and for collaborative and distributed networks connecting these systems for research and practice [39] , [58]–[60] . Careful evaluation of the strengths and limitations of GFT and other innovative surveillance tools should be expanded to encompass a range of developed and developing country settings , following the approach proposed here , in order to improve local , regional and global outbreak surveillance methods and inform public health responses . The way forward using high volume search query data such as GFT may be through integration of near-real time electronic public health surveillance data , improved computational methods and disease modeling – creating systems that are more transparent and collaborative , as well as more rigorous and accurate , so as to ultimately make them of greater utility for public health decision making .
In November 2008 , Google Flu Trends was launched as an open tool for influenza surveillance in the United States . Engineered as a system for early detection and daily monitoring of the intensity of seasonal influenza epidemics , Google Flu Trends uses internet search data and a proprietary algorithm to provide a surrogate measure of influenza-like illness in the population . During its first season of operation , the novel A/H1N1-pdm influenza virus emerged , heterogeneously causing sporadic outbreaks in the spring and summer of 2009 across many parts of the United States . During the autumn 2009 pandemic wave , Google updated their model with a new algorithm and case definition; the updated model has run prospectively since . Our study asks whether Google Flu Trends provides accurate detection and monitoring of influenza at the national , regional and local geographic scales . Reliable local surveillance is important to reduce uncertainty and improve situational awareness during seasonal epidemics and pandemics . We found substantial flaws with the original and updated Google Flu Trends models , including missing the emergence of the 2009 pandemic and overestimating the 2012/2013 influenza season epidemic . Our work supports the development of local near-real time computerized syndromic surveillance systems , and collaborative regional , national and international networks .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[]
2013
Reassessing Google Flu Trends Data for Detection of Seasonal and Pandemic Influenza: A Comparative Epidemiological Study at Three Geographic Scales
The HIV-specific cytotoxic T lymphocyte ( CTL ) response is a critical component in controlling viral replication in vivo , but ultimately fails in its ability to eradicate the virus . Our intent in these studies is to develop ways to enhance and restore the HIV-specific CTL response to allow long-term viral suppression or viral clearance . In our approach , we sought to genetically manipulate human hematopoietic stem cells ( HSCs ) such that they differentiate into mature CTL that will kill HIV infected cells . To perform this , we molecularly cloned an HIV-specific T cell receptor ( TCR ) from CD8+ T cells that specifically targets an epitope of the HIV-1 Gag protein . This TCR was then used to genetically transduce HSCs . These HSCs were then introduced into a humanized mouse containing human fetal liver , fetal thymus , and hematopoietic progenitor cells , and were allowed to differentiate into mature human CD8+ CTL . We found human , HIV-specific CTL in multiple tissues in the mouse . Thus , genetic modification of human HSCs with a cloned TCR allows proper differentiation of the cells to occur in vivo , and these cells migrate to multiple anatomic sites , mimicking what is seen in humans . To determine if the presence of the transgenic , HIV-specific TCR has an effect on suppressing HIV replication , we infected with HIV-1 mice expressing the transgenic HIV-specific TCR and , separately , mice expressing a non-specific control TCR . We observed significant suppression of HIV replication in multiple organs in the mice expressing the HIV-specific TCR as compared to control , indicating that the presence of genetically modified HIV-specific CTL can form a functional antiviral response in vivo . These results strongly suggest that stem cell based gene therapy may be a feasible approach in the treatment of chronic viral infections and provide a foundation towards the development of this type of strategy . Human hematopoietic stem cells ( HSCs ) , through development in the thymus , are capable of producing progeny T cells that generally display one of a vast repertoire of T cell receptors ( TCRs ) . In the case of many non-persistent viral infections , T cells bearing TCRs specific to viral antigens mediate a potent antiviral response that results in the clearance of the virus from the body . Even in the presence of most persistent viral infections , a potent T cell response is mounted; however it often fails to clear the virus from the body . A critical component of the T cell antiviral response is the CD8+ cytotoxic T lymphocyte ( CTL ) , whose primary function is to recognize viral antigens ( in the context of human leukocyte antigen class I ( HLA I ) ) and kill virus-infected cells . In HIV infection , the potent antiviral CTL response is critical for establishing relative control of viral replication during the acute and chronic infection stages of the disease [1]–[6] . However , unlike what is observed in most non-persistent viral infections , the CTL response fails to clear HIV from the body . The magnitude , breadth , functional quality , and kinetics of the antiviral CTL response all are critical in controlling ongoing viral replication; however , the reasons for the failure to rid the body of virus are not completely understood [7] , [8] . Ongoing viral replication and viral evolution in the infected host is one important , although highly confounding , factor in the persistence of HIV in chronic infection [4] , [5] . Even under effective antiretroviral therapy ( ART ) , the virus is not cleared from the body and the level of HIV specific CTLs declines , likely due to lower levels of antigen to stimulate the persistence/generation of these cells [9] , [10] . Due to the importance of T cell responses in controlling and eliminating viral infection there exists a great need to explore ways to enhance antiviral T cell immune responses . Recently , much of attention in HIV research has focused on ways to enhance or correct the defects in HIV-specific CTL responses . Gene therapy-based approaches that augment immunity towards viral antigens represent unique , yet largely unexplored , strategies towards the treatment of HIV disease . We have previously examined the feasibility of a stem cell-based gene therapy approach to enhance cell-mediated immunity towards chronic HIV infection . In these studies , we demonstrated that human HSCs genetically modified with genes encoding a human HIV-specific TCR can produce mature , fully functional T cells in human thymus implants in severe combined immunodeficient ( SCID ) mice . The resulting genetically directed CD8+ T cells are capable of killing HIV antigen-expressing cells ex vivo [11] . Further , we showed that the appropriate restricting human leukocyte antigen ( HLA ) class I molecule is required for proper development of transgenic TCR-containing CTLs . In all , our earlier studies demonstrated that TCR-modified human HSCs can be directed to develop into mature CTLs in a human thymus environment in the context of the proper HLA type . However , as the SCID-hu mouse model demonstrated poor peripheral reconstitution and function of human immune cells , these studies did not address the ability of these cells to suppress HIV replication in vivo . In the present studies , we examined the ability of genetically modified T cells derived from HSC transduced with a single HIV-specific TCR to suppress viral replication in vivo . We utilized a modified version of a newly established humanized mouse model , the non-obese diabetic ( NOD ) -SCID , common gamma chain knockout ( γc−/− ) , humanized bone marrow , fetal liver and thymus ( the NSG-BLT ) mouse model , which allows the generation of peripheral human immune responses , and serves as an effective model for HIV infection and pathogenesis [12]–[14] ( see Figure 1A ) . These humanized mice display multilineage human hematopoiesis and systemic engraftment of peripheral organs with human blood cell types including T lineage cells , B lineage cells , myeloid lineage cells , NK cells , as well as cells from other lineages [12] ( and see Figure 1B ) . We modified human hematopoietic stem cells in this model with molecularly cloned genes corresponding to a TCR specific to the HIV-1 Gag 77–85 SLYNTVATL ( SL9 ) epitope to allow the production of mature HIV-specific CTLs in multiple organs of these reconstituted mice . We determined that human T cells that expressed the HIV-specific TCR were capable of suppressing HIV replication in vivo and preventing or slowing viral damage to the engrafted human immune system . These studies establish a system to examine “genetic vaccination” approaches that target chronic viral infection and to more closely examine mechanisms of human antiviral immunity in vivo . We have previously demonstrated that human hematopoietic stem cells can be genetically modified by delivering a gene for an HIV-specific TCR , and develop into mature T cells in an HLA-restricted fashion in the human thymus of SCID-hu mice [11] . These newly produced , SL9 gag antigen-specific , naive T cells were determined to be capable of producing IFN-γ in response to peptide stimulation and were found to be lytic to SL9 peptide loaded target cells . However , it was not known whether these genetically modified HIV-specific CTLs could traffic to relevant organs in the mice and whether they were capable of killing HIV infected cells in vivo . To address this question , we established an improved model , based on the NSG-BLT model previously shown to allow HIV replication [15] , [16] , as a surrogate system to assess the antiviral efficacy of engineered , HIV-specific T cells in vivo . NSG mice were implanted with human fetal liver-derived CD34+ HSCs that had been modified with a lentiviral vector containing the genes for a TCR targeting the HIV Gag SL9 epitope , or as a control , with HSCs modified with a lentiviral vector containing a non-HIV-specific TCR with unknown specificity . In addition , these mice received implantation of human fetal Thymus and Liver under the kidney capsule to facilitate human T cell development . Hence , we term this the NSG-CTL model ( Figure 1A ) . As genetic manipulation of HSCs is required in this model , we initially determined the effects on this type of lentiviral transduction on multilineage hematopoietic potential of HSCs in the humanized mice . Phenotypic markers of human hematopoiesis were examined by flow cytometry in mice within 6 weeks following implantation of human tissues . One hundred percent of the mice receiving human tissue had human cells in the peripheral blood , including myeloid , natural killer ( NK ) , T cell , and B cell lineages ( Figure 1B ) . In these mice , the average percentage of human CD45+ cells in the peripheral blood was 53% of the total cells ( with a standard deviation of 29% and a range of 19%–80% , n = 12 ) . We more closely examined the bone marrow in these mice for the presence of human cell engraftment , particularly human HSC engraftment . We found a significant population of human CD34+ HSCs in the bone marrow ( Figure 1C ) . The majority of these cells coexpressed the CD45 molecule , which is indicative of cells with lymphopoietic potential [17] . In addition , there were significant populations of both CD3 expressing T cells and CD19 expressing B cells in the bone marrow of these mice . This indicated that multilineage human hematopoiesis occurs in these mice and provides evidence that , in addition to T cells , other components of the human immune system are present . These data demonstrated that our modification of the NSG-BLT humanized mouse utilizing genetically modified human hematopoietic stem cells does not negatively affect human hematopoiesis . We then examined the animals for the presence of cells expressing the transgenic , HIV specific TCR by MHC tetramer staining . We found CD3+ T cells expressing the transgenic TCR in all organs assessed , including the bone marrow , thymus , spleen , liver , and peripheral blood of the mice receiving transduced human hematopoietic stem cells ( Figure 1D ) . Thus , we have observed long-term , multilineage human immune reconstitution and the development of mature T cells that express the transgenic , HIV-specific TCR in multiple organs in the NSG-CTL mouse . To assess if peripheral cells resultant from human hematopoietic stem cells that expressed the recombinant SL9-specific TCR were capable of suppressing HIV replication in vivo , NSG-CTL mice containing the HIV specific TCR or a control TCR were infected with HIV-1NL-HSA-HA . HIV-1NL-HSA-HA is an engineered variant of HIV-1NL4-3 that contains the murine heat stable antigen ( HSA ) reporter gene modified to contain an Influenza hemagglutinin ( HA ) antibody epitope , which is cloned into the open reading frame of the vpr gene to allow detection of HIV-infected cells by cell surface detection of HA expression using flow cytometry [18] . Peripheral blood was assessed for the level of productively infected cells two and six weeks post infection . Within 2 weeks post infection , we observed a reduced level of productively infected cells in mice containing the HIV-specific TCR versus mice containing the control TCR ( Figure 2A ) . In addition , there was less initial CD4 depletion in mice containing the HIV-specific TCR versus mice containing the control TCR . Within six weeks post infection , while there was an overall increase in virus-expressing cells from the earlier time point , we observed a marked reduction in productively infected cells in mice containing the HIV specific TCR versus the control TCR , indicating suppression of viral replication over time ( Figure 2B ) . At this time point , mice containing cells expressing the HIV-specific TCR had a greater preservation of CD4+ T cells and higher CD4 to CD8 T cell ratios when compared to mice expressing the control TCR . Amongst all mice in the experiment , there was no statistically significant difference 2 weeks following infection with either CD4 cell count or with the percentage of cells expressing HIV , however there was a trend towards better preservation in CD4+ cell numbers as well as lower levels of virus-expressing cells in mice containing the HIV-specific TCR ( Figure 3 ) . However , by 6 weeks post-infection , there was a statistically significant difference in CD4 cell numbers and levels of infected cells between mice with cells expressing the HIV-specific TCR and mice with cells expressing the control TCR . Thus , genetic modification of HSCs with a single HIV-specific TCR produces peripheral T cells capable of suppressing cellular HIV expression and CD4 depletion in vivo . We next sought to determine if cells modified with an HIV-specific TCR could suppress virus levels in peripheral blood plasma . However , quantitating plasma viremia in the mouse model is difficult due to the amount of plasma obtained per blood draw ( typically ∼50 microliters ) , the limit of detection obtainable with this amount of blood , and the high cost associated with commercial assays . Therefore to measure viremia in this system , we developed a novel quantitative PCR-based technique for HIV in mouse plasma . Based on the recently elucidated secondary structure of the HIV genome [19] , primers were designed to specifically target relatively “open” regions of the RNA genome that contain minimal secondary structure to attempt to allow increased sensitivity to detect viral RNA . Utilizing this technique , which has a reliable sensitivity of 5 copies of HIV RNA per sample , we determined that the viral load 2 weeks and 6 weeks post infection was significantly lowered in mice receiving the HIV-specific TCR versus mice receiving cells transduced with the control TCR ( Figure 4A ) . This suggested systemic suppression of HIV replication in vivo . Surprisingly , analysis of the viral RNA for mutations in the SL9 epitope did not reveal the presence of any mutations in this epitope in the majority quasispecies , which was identical in comparison to the sequence of the input virus and the virus of infected mice containing the non-specific TCR control ( Figure 4B ) . This suggested that in this period of time , viral escape to the selective pressure of the SL9 specific TCR had not occurred in the blood of these mice , possibly due to limited viral replication in this model . Thus , there was significant suppression of viral replication in vivo in mice expressing the HIV-specific TCR versus the control TCR and this suppression did not result in significant viral escape within 6 weeks following infection . As illustrated in Figure 1 , T cells expressing transgenic HIV-specific TCRs were found in multiple organs in mice receiving genetically modified HSCs . Based on this , we next addressed suppression of HIV in multiple organs in the lymphoid compartment in mice containing cells expressing the HIV-specific TCR . NSG-CTL mice that had received HSCs transduced with the HIV SL9-specific TCR or , separately , the non-specific control TCR were infected with HIV-1NL-HSA-HA . Sets of infected animals were then assessed 2 weeks and 6 weeks post infection for the quantity of HIV proviral DNA sequences in human cells in the spleen , bone marrow , and human thymus implant ( Figure 5 ) . We observed significant suppression of HIV replication in human cells in these organs as early as 2 weeks post infection ( in the bone marrow ) in mice receiving HSC containing the HIV-specific TCR . 6 weeks post-infection , HIV levels were significantly lower in the spleen , bone marrow , and human thymus implant in animals receiving the HIV-specific TCR as compared to mice receiving the control TCR . In addition , analysis for proviral DNA in human cells in the pooled peripheral blood cells ( n = 3 mice per treatment group ) , revealed a similar trend , with 37 copies and 529 copies of HIV per 10 , 000 human cells at 2 weeks and 6 weeks post infection respectively , in mice containing the HIV-specific TCR , and 356 copies and 792 copies of HIV per 10 , 000 human cells at 2 weeks and 6 weeks , respectively , in mice containing the control TCR . Thus , these data indicate that there is significant suppression of HIV in multiple lymphoid tissues in animals receiving HSCs genetically modified to produce cells that specifically target HIV infected cells . We assessed the antiviral effector function of CTLs expressing HIV-specific transgenic TCRs in mice receiving genetically modified HSCs . In an additional series of experiments , mice containing the SL-9 specific TCR were infected with HIV or left uninfected and cells from the peripheral blood were assessed for phenotypic changes that would suggest differentiation . HIV infection resulted in phenotypic differentiation of HIV specific cells , as determined by SL9 MHC tetramer staining , into cells possessing an effector phenotype [20] , [21] ( CD8+SL9Tetramer+CD45RA-CCR7- ) ( Figure 6A ) . This was similar to the phenotypic changes we observed in previous studies following ex vivo peptide stimulation of SL9-specific , TCR transgenic thymocytes [11] and in vivo responses to the MART-1 tumor antigen by MART-specific CD8 cells [22] . This increased loss of CD45RA and CCR7 expression that we observed in HIV- specific TCR-expressing cells in infected mice versus uninfected mice is indicative of antigen-specific induction of cellular differentiation . We then more closely analyzed the differences we observed viral suppression by and expansion of HIV-specific CTLs in vivo in infected mice . We found a significant correlation between the highest levels of reconstitution of HIV-specific TCR-expressing cells prior to infection and more effective suppression of viral loads in the serum six weeks following infection ( Figure 6B ) . Interestingly , we noted that at six weeks following infection , mice that had greater levels of HIV-specific TCR-expressing cells in the peripheral blood had higher viral loads at this time point ( Figure 6C ) . In addition , we saw significant antigen-driven expansion of HIV-specific TCR-expressing CTLs in infected animals compared to controls , with the greatest levels of expansion seen in animals with the lowest initial ( week -2 ) transgenic TCR reconstitution ( Figure 6D ) . Taken together , these results suggest that greater initial reconstitution of transgenic HIV-specific cells is more effective at controlling early viral replication . Furthermore these data suggest that the higher resultant viral loads in animals with initially low human immune reconstitution drive greater antigen-specific cell expansion over time . Thus , CTLs expressing the HIV-specific TCR undergo antigen-driven phenotypic differentiation and expansion in this model , which correlates with control of viral replication . The CTL response has a pivotal role in controlling HIV replication in infected individuals . While HIV generates a potent natural immune response during the acute stage of infection , this response does not result in the control of viral replication or clearance of the virus from the body [4]–[6] . There are critical defects in the CTL response that result during chronic viral infection . These defects include the inadequate generation of a functional response due to low antigen-specific precursor frequency , expression of functional inhibitory molecules such as programmed death-1 ( PD-1 ) and T-cell immunoglobulin domain and mucin domain 3 ( TIM-3 ) , and Cytotoxic T-Lymphocyte Antigen 4 ( CTLA-4 ) , and activation of suppressor cell activity [23]–[26] . In addition , HIV can directly or indirectly perturb viral antigen presentation , immunoregulatory cytokine production , T cell differentiation and effector/memory generation , and can infect CTLs themselves [27]–[33] . The maintenance of a potent antiviral CTL response is critical in all stages of infection and there are strong associations between the preservation of CTL responses specific for more conserved HIV epitopes , greater control of viral replication , and slower disease progression [5] , [6] . In the present study , we demonstrate the feasibility of engineering human hematopoietic stem cells to become peripheral T cells capable of targeting HIV replication in vivo . Our previous studies provided evidence that the genetic modification of human hematopoietic stem cells with a lentiviral vector containing an antigen-specific TCR ( specific to the SL9 Gag epitope ) allowed the development of functional human T cells in human thymus implants in SCID-hu mice [11] . While this study demonstrated that transgenic TCR-containing T cells are capable of developing in the human thymus , the ability of these cells to target and kill HIV infected cells in vivo was not known . In the present study , we use an improved chimeric mouse model exhibiting a high degree of human immune cell reconstitution to significantly extend these observations to demonstrate that mature T cells expressing an antigen-specific human TCR are capable of developing and migrating to peripheral organs in vivo . In contrast to the SCID-hu Thy/Liv model , which is an excellent model for studies examining human thymopoiesis but limited in examining peripheral immune responses [34] , we utilized a variation of the humanized BLT mouse model utilizing the NSG strain that allows multilineage hematopoiesis and human cell repopulation in peripheral organs [35] , [36] . The generation of natural immune responses to HIV in these systems appears to be relatively limited , particularly the ability of these mice to elicit HIV specific human T cell responses which is likely due to incomplete human immune cell reconstitution , particularly antigen presenting cell reconstitution , to the levels seen in humans [12] , [36] , [37] . In addition , lower antigen-specific cell precursor frequency and the lack of or lower levels of human-specific cellular support immune components ( such as costimulatory or immunoregulatory molecules , adhesion molecules , and cytokines ) likely contribute to the lower levels of antiviral immune responses generated in humanized mice . The incomplete and varied immune reconstitution in the current humanized mouse systems results in differences in immune responses and kinetics of viral pathogenesis compared to natural HIV infection in humans . The reasons for this are unclear and vary between the different types of humanized mouse models; however , there are many similarities and parallels between HIV infection in humanized mice and humans which makes these surrogate models very powerful in their ability to allow the close examination of many aspects of HIV infection , transmission , pathogenesis , immunity , and therapeutic intervention [36] . While natural antiviral T cell immune responses are limited in current humanized mouse models , our studies suggest that the genetic “programming” of HSCs to produce T cells specific for HIV can overcome this limitation in this system and produce measurable T cell responses that have a significant antiviral effect in vivo . Further , we found it startling that the use of a single HIV-specific TCR can result in significant HIV suppression while natural suppressive antiviral CTL responses are polyclonal . These observations can provide the platform for future studies that allow the closer examination of the generation of human antiviral immune responses and the identification of factors involved in the persistence and potential eradication of HIV infection . Previous attempts utilizing a gene therapy approach towards enhancing antigen specific cellular immune responses have focused on “redirecting” mature T cells towards viral or cellular antigens [38]–[46] . In these cases , genes for HIV-specific T cell receptors ( TCRs ) or chimeric antigen specific receptors were utilized to modify mature T cells to specifically target virus infected cells or malignancies . In some cases of the latter , tumor regression has occurred in treated individuals [45]–[47] , which suggests that the genetic modification of T cells towards a specific antigen is feasible in vivo in humans and alludes to the potential for the further development of these strategies to target other diseases . However , the modification of mature T cells has several limitations , including the possibility of endogenous TCR miss-pairing with the newly introduced TCR , the development of intrinsic functional defects and/or the alteration of cellular effector/memory maturation pathways in the cells following heavy ex-vivo manipulation [47] , and the maintenance of long-lived fully functional cells . A stem cell-based approach where HSCs are modified with an antigen specific receptor , however , may abrogate these complications by allowing the long term , continual natural development of mature T cells that bear the transgenic antigen-specific molecule . Normal development of these cells in the bone marrow and selection in the thymus would reduce the possibility of producing cells that are autoreactive through TCR miss-pairing and functionally altered through ex vivo manipulation , major drawbacks of mature T cell modification . We have recently shown that genetic modification of human HSC with a TCR specific for human melanoma allows the generation of melanoma-specific human T cells capable of clearing tumors in BLT mice [48] . Our current studies extend this type of approach to demonstrate the in vivo efficacy of TCR-modified stem cells to generate antigen-specific T cells that target a rapidly replicating viral infection in vivo . Our results document the ability of the resulting HIV-specific CTLs to dramatically reduce viral replication and consequent CD4 cell loss in a relevant model of HIV pathogenesis . Recent stem cell-based attempts at protecting cells from direct infection by HIV through the modification of HSCs with antiviral genes or genes that knock down viral coreceptors [16] , [49] , [50] require high percentages of HSCs to be genetically modified to be protected from infection . Our results suggest that the approach of genetically vaccinating cells to target HIV infection would require much lower levels of genetic modification . Modification of human HSCs with a transduced TCR results in significantly increased naïve , antigen specific precursors . This level of transduction is sufficient to result in decreased viral replication and increased immune protection . Correspondingly in humans , uninfected HLA-A*0201+ individuals have an estimated natural SL9 epitope-specific , naïve CTL precursor frequency of approximately one in 3 . 3×106 cells in the peripheral blood , which is similar to the precursor frequency of naïve cells specific to a variety of other viral antigens [51] . In our studies , the TCR-transduced population typically accounted for 0 . 75–5 . 5% of the CD8+ T cell population in a given organ in the mouse following their differentiation from HSCs ( the illustration in Figure 1D represents a single mouse from a single experiment ) . The frequency of transgenic cell reconstitution did not correlate with transduction efficiencies of the vector into the stem cells , rather it appears to be due to individual engraftment rates of CD34+ cells into each mouse . However , even at low frequencies of transgenic TCR expressing cells , this represents a significant increase in the naïve cell precursor frequency for cells specific to the SL9 Gag epitope , as mice harboring the control non-specific TCR and untransduced mice had undetectable levels or very low levels of natural SL9-specific CTLs as determined by MHC tetramer staining . Utilizing TCR gene transductions to yield increases of HIV-specific precursor frequency to conserved antigenic epitopes could potentially reconstitute innate defects in the ability of peripheral T cells to clear infected cells . While the human thymus involutes over time , thus producing fewer T cells in adults than in children , it does retain some activity throughout life [52] , [53] . A recent study involving introduction of an antiviral gene into the autologous HSC of HIV infected adults illustrated that naïve T cells bearing the transgene were detected in the peripheral blood of these subjects , indicating that genetically engineered T cells can develop from HSC in adult HIV infected subjects [54] . Through this type of therapeutic intervention , our results suggest the feasibility of supplying newly developed , naïve antigen-reactive cells , that could allow the overall T cell response to overcome limits in the magnitude of the response that inhibit effective viral clearance . This type of gene therapy-based approach could further diversify the breadth of the responses by naïve , antigen specific cells by utilizing TCRs specific to other epitopes of HIV . The use TCR gene transduction as a therapeutic approach would have to be tailored to the HLA type of the individual receiving treatment in order to produce cells that survive T cell selection processes . Immune epitope escape from the transduced TCR , which did not occur in the time frame of our experiments , is likely to occur in vivo in a clinical setting . One potential caveat of the humanized mouse model is the lower level of human immune cell reconstitution than is seen in humans; which significantly , yet incompletely , recapitulated the human immune system in the mouse . While HIV replication rates and viral loads persist detectably over weeks , they do not achieve the levels observed in natural infection in humans . This lower level of viral replication is one potential reason that viral escape mutants to the SL9-specific TCR may be slower to develop . The potential for viral immune escape necessitates the use of multiple TCRs in a therapeutic setting targeted to the antigen or antigens of interest . Careful selection of multiple TCRs targeted to relatively conserved antigenic epitopes within defined HLA types could reduce the possibility of viral epitope evolution and immune escape , perhaps driving the evolution of the virus into a less fit state [55] . The evidence that immune escape and viral evolution against many specific epitopes occurs relatively slowly suggests that an engineered immune response and the immune pressure created by these antigen-specific cells may be therapeutically beneficial by lowing viral replication , decreasing levels of infected cells , and impairing the fitness state of the virus [55] , [56] . In sum , our results demonstrate the feasibility of a therapeutic approach that involves the modification of human HSCs by delivering genes for antigen-specific TCR to produce peripheral , naïve , antigen-specific T cells that are capable of reducing HIV replication in vivo . These studies provide a foundation and a model system that would allow the closer examination of human antiviral T cell responses and the development of therapeutic strategies that target chronic viral infection . Peripheral blood mononulear cells was obtained at the University of California , Los Angeles in accordance with UCLA Institutional Review Board ( IRB ) approved protocols under written informed consent using an IRB-approved written consent form by the UCLA Center for AIDS Research Virology Laboratory and Dr . Yang and was distributed for this study without personal identifying information . Human fetal tissue was purchased from Advanced Biosciences Resources or from StemExpress and was obtained without identifying information and did not require IRB approval for its use . Animal research described in this manuscript was performed under the written approval of the UCLA Animal Research Committee ( ARC ) in accordance to all federal , state , and local guidelines . Specifically , these studies were carried out under strict accordance to the guidelines in The Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and the accreditation and guidelines of the Association for the Assessment and Accreditation of Laboratory Animal Care ( AALAC ) International under UCLA ARC Protocol Number 2010-038-02B . All surgeries were performed under ketamine/xylazine and isofluorane anesthesia and all efforts were made to minimize animal pain and discomfort . The following antibodies were used in flow cytometry: CD3 , CD4 , CD11c , CD8 , CD45 , CD45RA , CD34 , HLA-DR ( Coulter ) , CD19 , CD56 , CCR7 , HSA , and IgG controls ( eBioscience ) , hemagglutinin ( HA ) sequence YPYDVPDYA ( Roche ) , and HLA-A*02 ( Serotech ) . HLA-A*0201 tetramer containing the HIV Gag SL9 SLYNTVATL ( SL9 ) peptide was purchased from Coulter . Cell surface marker expression was analyzed utilizing antibodies conjugated to either fluorescein isothiocyanate ( FITC ) , Peridinin Chlorophyll Protein ( PerCP ) -Cy5 . 5 , phycoerythrin ( PE ) , electron coupled dye ( ECD ) , PE-Cy5 , PE-Cy7 , allophycocyanin ( APC ) , APC-Alexa750 , APC-eFluor780 , Alexa700 , eFluor405 , Pacific Blue , or Pacific Orange in appropriate combinations . Cells were acquired on a LSR II flow cytometer ( BD Biosciences ) using FACSDiva software . FlowJo software was used for analysis . Lentiviral production from the plasmid containing the HIV SL9 specific TCR ( pCCL . PPT . hPGK . 1 . 9 . IRES . eGFP ) or a control TCR with an unknown specificity ( pCCL . PPT . hPGK . α4 . IRES . eGFP ) was produced using the Invitrogen ViraPower Lentiviral Expression System using the pCMV . ΔR8 . 2 . Δvpr packaging plasmid and the pCMV-VSV-G envelope protein plasmid as previously described [11] . NSG mice were initially purchased from Jackson Laboratories and bred at the UCLA Division of Laboratory Animal Medicine . To construct NSG-CTL mice , fresh human HLA-A*0201+ fetal liver and thymus pairs from the same donor were obtained from Advanced Biosciences Resources or from StemExpress . Fetal liver was then homogenized and CD34+ cells were isolated as described [11] . Briefly , fetal liver is diced into small ( ∼3 mm ) pieces , homogenized and digested with collagenase type IV ( 1 mg/ml ) , hyaluronidase ( 1 mg/ml ) , DNase I ( 2 U/ml ) ( Sigma ) . CD34+ cell were purified using magnetic activated cell sorting ( Miltenyi ) . The negative fraction of cells , which contains fetal liver stromal cells ( CD34− cells ) is saved . CD34+ cells were then genetically transduced following resuspension in Yssel's medium containing 2% human serum albumin and placed in a tissue culture plate coated with 20 µg/ml retronectin ( Takara Bio , Inc . ) along with lentiviral vector at a multiplicity of infection of 5 overnight at 37° . Fetal liver stromal cells and the matched fetal thymus , cut into small pieces ( 2 mm ) , were cultured at 37° overnight in RPMI-1640 containing 10% fetal calf serum ( FCS ) and 0 . 44 mg/ml Piptazo . The next day , tissue and cells were washed with PBS and a fraction of the transduced CD34+ cells were then viably frozen . The remaining CD34+ cells were combined with fetal liver stromal cells in cold Matrigel ( BD Biosciences ) in a 1∶9 ratio ( CD34+ cells:fetal liver stromal cells , typically 500 , 000 transduced CD34+ cells:4 , 500 , 00 fetal liver stromal cells ) and combined with a 2 mm fetal thymus piece in a trocar and placed under the kidney capsule of NSG mice . Transduction efficiency was determined following culturing in IMDM containing 20%FCS , 50 ng/ml of IL-3 , IL-6 , and SCF for 3 days , and subsequent assessment of GFP fluorescence by flow cytometry . Transduction efficiency of CD34+ cells occurred at a mean rate of 12 . 7% ( Standard deviation = 12 . 6% , range 1 . 63%–38 . 5% , n = 14 ) . Three weeks following implantation , mice were irradiated with 3 Gy using a cobalt-60 source to clear a niche for human CD34+ cell engraftment of the bone marrow . The frozen CD34+ cells were then thawed and then injected intraveneously into the mice . Mice were then checked for human cell engraftment 6–10 weeks post-injection . Multiple experiments were performed with a minimum of 3 mice per experimental group to yield statistical significance . Each experiment utilized humanized mice that were made from human tissue from same donor and the donor tissue was unique experiment to experiment . A virus variant of HIV-1NL4-3-HSA-HA containing the mouse heat stable antigen ( HSA ) which has been modified to contain the influenza virus hemaggluttin YPYDVPDYA sequence ( HA ) cloned into the vpr open reading frame , and which also contains the SL9 Gag epitope , has been previously described [18] . Virus was grown in CEMx174 from plasmid-derived virus stock . Viral infectivity was determined by limiting dilution titration on CEMx174 cells . Mice were infected by intraperitoneal injection 10–12 weeks following CD34+ cell injection with 50–100 ng of previously frozen virus stock . Mouse blood peripheral blood was drawn by retro-orbital bleeding into glass capillary tubes coated with 330 mM EDTA ( Gibco ) , and 3% sterile human serum albumin ( Baxter Healthcare ) . Viral RNA was extracted from plasma with the High Pure Viral RNA Kit ( Roche ) . The kit is designed to extract 200 µl of plasma . Since there is generally less plasma than this , the volume was estimated by weight and brought up to 200 µl with phosphate buffered saline ( Gibco ) . DNA standards and the template for in vitro transcribed RNA for quantitative PCR was derived from pNL101 linearized with EcoRI , checked with electrophoresis , and quantitated by spectrophotometry ( A260 ) . A section of the gag gene in pNL101 was amplified with the primers NG1CF , position 366–398 ( 5′-GGAGAATTAGATAAATGGGAAAAAATTCGGTTA-3′ ) and NG1CR position 679–648 ( 5′-GCCTTTTTCTTACTTTTGTTTTGCTCTTCCTC-3′ ) , and cloned into pCR4TOPO . The product containing the cloned gag was then digested with SpeI and BsrGI and gel purified . This fragment was then translated to RNA with T7 RNA polymerase ( Promega Riboprobe Transcription Kit ) and quantitated by spectrophotometry ( A260 ) . This RNA was serially diluted in The RNA Storage Buffer ( Ambion ) with 0 . 4 U/µl Rnasin and 5 ng/µl Lambda DNA/HindIII ( as carrier ) , to make a stock of 500 , 000 copies/µl . Before each RT run , a fresh vial of RNA was serially diluted in the RNA Storage Buffer ( Ambion ) to make standards of 100 , 000 to 10 copies . Quantitative reverse transcriptase-PCR ( RT-PCR ) was performed using the following primers/probe specific for gag sequences: NG1F ( position 453–480 ) 28 bp ( 5′-GAGCTAGAACGATTCGCAGTTAATCCTG-3′ ) , NG1R ( position 570–534 ) 37 bp ( 5′-ATAATGATCTAAGTTCTTCTGATCCTGTCTGAAGGGA-3 ) , NG1Z probe ( position 482–520 ) 39 bp ( FAM-5′ -CCTTTTAGAGACATCAGAAGGCTGTAGACAAATACTGGG-3-BHQ ) . The final reaction concentration consists of 2 . 5 µM NG1F , 7 . 5 µM NG1R , and 2 . 5 µM NG1Z . These primers were based on sequences identified to be in relatively “open” regions of HIV RNA not impeded by secondary structure interference as determined by [19] . Reverse transcription was performed using the SuperScript III kit ( Invitrogen ) . The annealing step consisted of 5 µl of template RNA plus 3 µl of a mixture consisting of 1 . 5 parts 20 µM NG1R , 0 . 5 parts Rnasin plus ( Promega ) , 16 parts 5× RT buffer , and 12 parts water . The resulting 8 µl was heated to 70° for 2 minutes , then at 60° for 5 minutes , then cooled to room temperature . The RT step was run by adding 2 µl of a mixture of 8 parts water , 4 parts 5× buffer , 5 parts DTT , 2 parts 25 mM dNTPs ( Invitrogen ) , and 1 part SuperScript III . This was heated to 55° for 30 minutes , 85° for 5 minutes , then cooled to room temperature . For quantitative DNA PCR following the reverse transcription step , 15 µl of the PCR mix consisting of 38 . 5 parts water , 44 parts 25 mM MgCl2 , 50 parts NG-FRZ oligos , 5 parts 500 mM Tris buffer pH 8 . 3 , 8 . 5 parts 1 M KCl , 2 . 5 parts 25 mM dNTPs , and 1 . 25 parts Platinum Taq was added to all wells that underwent the reverse transcription reaction and mixed . Real-time , quantitative PCR was performed with 5 minutes activation at 95° , and followed by 45 cycles of 95° for 15 seconds and 60° for one minute on a BioRad CFX96 thermocycler . An additional set of DNA standards , serially diluted from 2×105 copies to 20 copies , of linearized pNL101 was run in parallel to control for the efficiency of the RT step . Results from samples were interpolated within the quantitation derived from the RNA standards . Statistical support was provided through the UCLA Center for AIDS Research ( CFAR ) Biostatistical Core . Experiments were analyzed utilizing the Student's t test , the Spearman rank correlation test ( SRCT ) , or the Wilcoxon Rank Sum Test ( WRST ) ( when n = 3 ) , as indicated . Evolution/mutation of the dominant version of the introduced Gag-SL9 epitope sequences from the plasma of mice infected with HIV-1NL4-3-HSA-HA was determined by bulk sequencing of the segment of the Gag coding region . Plasma viral RNA was isolated as described above and cDNA was synthesized utilizing the Superscript cDNA synthesis kit ( Applied Biosystems ) . Alternatively , proviral DNA from lymphocytes on infected mice was isolated as described above . Utilizing these DNAs , the region of the HIV-1 Gag flanking the SL9 epitope ( a . a . 77–85 ) was PCR amplified using the 737-Forward primer ( 5′-GCGACTGGTGAGTACGCC-3′ ) and the 1255-Reverse primer ( 5′-ACCCATGCATTTAAAGTTC-3′ ) and purified by Gel-purification ( Qiagen Inc . , USA ) . This purified bulk PCR product was then directly used for dye-terminator sequencing with both 737-Forward and 1255-Reverese primers in parallel . The data was then analyzed by the ABI-3130 genetic analyzer ( Applied Biosystems , USA ) .
There is a desperate need for the development of new therapeutic strategies to eradicate HIV infection . HIV actively subverts the potent natural immune responses against it , particularly cellular cytotoxic T lymphocyte ( CTL ) responses . The development of a therapy that allows long-lived immune self-containment of HIV and restoration of these CTL responses by the host would be ideal . Through genetic manipulation of human blood-forming stem cells , we introduced a molecule– an HIV-targeting T cell receptor ( TCR ) –that allowed the generation of functional HIV-specific CTLs following differentiation within human tissues in a humanized mouse model . To assess if these newly developed , HIV-specific CTLs can allow active suppression of HIV replication , we infected these mice with HIV . We found that the development of genetically modified , HIV-specific CTLs in these mice results in the presence of a functional antiviral CTL response in vivo that significantly lowers viral replication following HIV infection . These results have strong implications for the use of this technology to engineer the human immune response to combat viral infections and suggest that genetic engineering via HSCs may allow tailoring of the immune response to target and eradicate HIV .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "immune", "cells", "viral", "transmission", "and", "infection", "immunology", "microbiology", "immunodeficiency", "viruses", "adaptive", "immunity", "immunotherapy", "animal", "models", "of", "infection", "viral", "clearance", "t", "cells", "biology", "immunity", "virology", "antivirals" ]
2012
In Vivo Suppression of HIV by Antigen Specific T Cells Derived from Engineered Hematopoietic Stem Cells
Patients with Chagas disease have migrated to cities , where obesity , hypertension and other cardiac risk factors are common . The study included adult patients evaluated by the cardiology service in a public hospital in Santa Cruz , Bolivia . Data included risk factors for T . cruzi infection , medical history , physical examination , electrocardiogram , echocardiogram , and contact 9 months after initial data collection to ascertain mortality . Serology and PCR for Trypanosoma cruzi were performed . Of 394 participants , 251 ( 64% ) had confirmed T . cruzi infection by serology . Among seropositive participants , 109 ( 43% ) had positive results by conventional PCR; of these , 89 ( 82% ) also had positive results by real time PCR . There was a high prevalence of hypertension ( 64% ) and overweight ( body mass index [BMI] >25; 67% ) , with no difference by T . cruzi infection status . Nearly 60% of symptomatic congestive heart failure was attributed to Chagas cardiomyopathy; mortality was also higher for seropositive than seronegative patients ( p = 0 . 05 ) . In multivariable models , longer residence in an endemic province , residence in a rural area and poor housing conditions were associated with T . cruzi infection . Male sex , increasing age and poor housing were independent predictors of Chagas cardiomyopathy severity . Males and participants with BMI ≤25 had significantly higher likelihood of positive PCR results compared to females or overweight participants . Chagas cardiomyopathy remains an important cause of congestive heart failure in this hospital population , and should be evaluated in the context of the epidemiological transition that has increased risk of obesity , hypertension and chronic cardiovascular disease . Chagas disease , caused by the parasite Trypanosoma cruzi , affects an estimated 8 million people in the Americas [1] . Transmission occurs via inoculation of T . cruzi-infected feces of the triatomine vector through the bite wound or mucosal surfaces . Infection may also be acquired congenitally , through blood transfusion , organ transplant or through consumption of contaminated food or drink . The acute phase is associated with high level parasitemia , and mild , non-specific symptoms in the majority of individuals [2] . After 8–12 weeks , infected individuals pass into the chronic phase , in which parasites are no longer detectable by peripheral blood microscopy and diagnosis relies on demonstration of anti-T . cruzi antibodies . Infection is life-long in the absence of successful treatment . Over a period of decades , 20–30% of infected individuals develop specific patterns of end-organ damage . The most common form , chronic Chagas cardiomyopathy , is characterized by conduction system abnormalities , brady- and tachyarrhythmias , dilated cardiomyopathy , apical aneurysm , and thrombus formation in the aneurysm or enlarged left ventricle [3] . Patients with Chagas heart disease have a high rate of mortality from ventricular arrhythmias , pulmonary or cerebral emboli , and intractable congestive heart failure [3] . Historically , T . cruzi transmission occurred predominantly in rural areas of Latin America where poor housing conditions promoted vector infestation . Since 1991 , Chagas disease control programs have made striking progress in decreasing vector- and blood-borne T . cruzi transmission , leading to dramatic declines in infection prevalence among children [4] , [5] . However , millions of T . cruzi-infected adults remain , and massive population movements over the past 3 decades have brought many of these individuals to cities across Latin America . Urban populations are in transition from an epidemiology of predominantly infectious diseases to patterns similar to industrialized countries where obesity , hypertension , diabetes and atherosclerosis are the leading causes of illness and death [6] , [7] . Adults infected with T . cruzi as children form a transitional generation , experiencing the simultaneous impact of past infectious exposures and current cardiovascular risk factors . Bolivia has the highest prevalence of T . cruzi infection in the world , estimated at 6% of the national population , and reaching 30–40% in surveys of pregnant women , blood donors or endemic community members [1] , [8] , [9] , [10] . The major objective of this study was to assess T . cruzi cardiac morbidity and its coincidence with common cardiovascular risk factors and disease among patients attending a large urban public hospital . In addition , we explored risk factors for T . cruzi infection and disease severity , and clinical and epidemiological associations with positive results by T . cruzi PCR . The protocol was approved by the institutional review boards of the study hospital , Asociación Benéfica PRISMA , and the Centers for Disease Control and Prevention . The study was conducted in the Hospital Universitario Japonés in Santa Cruz , Bolivia from August 25 to November 13 , 2008 . The hospital is one of two public hospitals and serves approximately 60% of the city's uninsured population . Although the city of Santa Cruz does not have vector-borne T . cruzi transmission , infection prevalence is high because many residents migrated from rural areas with intense transmission . All patients attending the cardiology clinic , admitted to the cardiology inpatient service or seen by the cardiology service in inpatient consultations were eligible , with the exception of those younger than 18 years , pregnant , unable to provide informed consent or unable to undergo cardiac evaluation . After written informed consent , a structured questionnaire was administered by a study nurse . Data included demographics and factors potentially associated with T . cruzi infection risk ( e . g . , type of housing , location of and infestation in residences throughout life , family history of sudden death and Chagas disease , history of transfusion ) . Each patient underwent a focused history and physical examination , electrocardiogram ( EKG ) and echocardiogram . The study physician and two cardiologists who performed the echocardiograms and interpreted the EKGs were blinded to patient infection status . Data on existing structural heart disease or risk factors for structural heart disease were collected from medical records , EKGs and echocardiograms . A 12-ml blood specimen was obtained from each participant , centrifuged and separated into serum and clot . In August 2009 ( 9 months after the completion of the cross-sectional study ) , telephone contact was attempted for all participants or their family members to ascertain whether the patients were alive . The date of death was sought for those no longer alive at that time . Serum specimens were screened for antibodies to T . cruzi using two commercial enzyme-linked immunosorbent assays ( ELISA ) , one based on whole epimastigote lysate ( Chagatek , bioMérieux , Buenos Aires , Argentina ) and the other on recombinant antigens ( Chagatest , Wiener laboratory , Rosario , Argentina ) . Results were interpreted following the manufacturers' instructions . For discordant results , an immunofluorescent antibody test ( IFA ) was performed using a titer of 1∶32 as the positive cut-off . A patient was considered to have confirmed T . cruzi infection if he or she had positive results by at least two serologic tests [11] . Polymerase chain reaction ( PCR ) was performed using 500 µl specimens of clot , based on an earlier analysis showing higher sensitivity in this specimen compared to buffy coat or whole blood [12] . DNA was extracted following a standard phenol-chloroform extraction protocol [13] . PCR amplifications were performed using the 121/122 primer set ( 5-AAATAA- TGTACGGGKGAGATGCATGA-3 and 5-GTTCGATTGGGGTTGGTGTAATATA-3 ) targeting the kinetoplast minicircle , using published methods and conditions [12] , [14] . A positive result was based on the appearance of the characteristic 330-bp product [12] . Quantitative real time PCR followed published methods [15] . The primer set Cruzi 1 ( 5′–ASTCGGCTGATCGTTTTCGA–3′ ) and Cruzi 2 ( 5′–AATTCCTCCAAGCAGCGGATA–3′ ) was used to amplify a 166 base-pair DNA fragment . The probe Cruzi 3 ( 5′–CACACACTGGACACCAA–3′ ) was labeled with 5′FAM ( 6 – carboxyfluorescein ) and 3′MGB ( minor groove binder ) . TaqMan Human RNase-P detection reagent ( Applied Biosystems ) was included as an internal control . A result was considered valid only when the internal control was efficiently amplified . A non-template negative control was included in each run . A positive result was defined by the threshold cycle ( Ct ) , the first cycle where fluorescence was detected above baseline . The C ( t ) was determined by the respective standard curve for the specimen batch , and was always between 37 and 38 cycles . A clot specimen was inoculated with 1×106 T . cruzi Y strain trypomastigotes , extracted and diluted successively to determine the minimum quantity detectable; the limit was found to be 1 parasite/ml [8] , [15] . The American College of Cardiology and American Heart Association ( ACC/AHA ) heart failure classification system was applied for all patients with available echocardiograms [16] . For T . cruzi-infected patients , we also assigned a severity classification based on a slight modification of published methods:[17] Stages were not assigned for infected subjects whose clinical and echocardiogram findings indicated an etiology other than Chagas cardiomyopathy . All other T . cruzi-infected patients with echocardiogram data were assigned a Chagas cardiomyopathy stage , regardless of the presence or absence of comorbidities . Hypertension , diabetes , coronary disease were defined based on patient history and/or physician documentation in the medical record chart . Body mass index ( BMI ) was calculated ( weight in kilograms divided by the square of the height in meters ) . Cross-sectional analyses were performed in SAS version 9 . 2 ( SAS Institute ) for three outcome measures: ( 1 ) T . cruzi infection among all subjects; and ( 2 ) positive T . cruzi PCR results and ( 3 ) Chagas cardiomyopathy severity stage among T . cruzi-infected subjects only . Categorical variables were compared by the exact Pearson's chi-square test; continuous variables were analyzed using the Wilcoxon rank-sum test . Variables significant at α = 0 . 05 were considered for multivariable logistic regression models , and appropriate 2-way interactions were tested . Of 549 patients seen by the cardiology service during the study period , 394 ( 72% ) enrolled in the study . Three excluded patients were pregnant , 8 were younger than 18 years , 44 declined participation , and 100 left the hospital prior to contact with the research staff . One patient failed to meet the serological criteria for confirmed infection ( positive results by Wiener ELISA , negative by Chagatek ELISA and IFA ) but had positive results by T . cruzi PCR; this patient was excluded from the epidemiological analyses . Of the 393 patients included in the analysis , 65% were female . The mean age was 52 . 0 years ( range 20–86 ) for females and 51 . 4 years ( 18–87 ) for males . The prevalence of confirmed T . cruzi infection was 63 . 9% ( 251/393 ) . Overall , 64% of participants had a history of hypertension and 67% were overweight or obese ( defined as body mass index [BMI] >25 ) . Other markers or risk factors for structural heart disease were less common . Of the 342 patients with ACC/AHA staging data , 155 ( 45% ) had stage C or D , indicating symptomatic congestive heart failure; of these , 91 ( 58 . 7% ) had Chagas cardiomyopathy , whereas 23 ( 14 . 8% ) had T . cruzi infection but another probable etiology for their cardiac insufficiency . Chagas cardiomyopathy severity stage was assigned for 191 T . cruzi-infected patients: 27 ( 14 . 1% ) stage A ( equivalent to indeterminate ) , 97 ( 50 . 8% ) stage B , 51 ( 26 . 7% ) stage C and 16 ( 8 . 4% ) stage D . Chagas severity stage was not determined for the 38 T . cruzi-infected patients whose heart disease was judged to be from another etiology . This group included 2 patients with isolated left ventricular hypertrophy ( LVH ) , 2 with left atrial dilatation associated with severe mitral stenosis or mitral valve prolapse , 26 with diastolic dysfunction and/or LVH due to hypertensive disease , 4 with segmental hypokinesis and systolic dysfunction in setting of an acute coronary syndrome , and 1 each with a pericardial tumor and congenital heart disease . Twenty-two infected patients were lacking echocardiogram data and could not be assigned a Chagas severity stage for that reason . Of the 393 patients in the analysis , survival status 9 months after completion of the cross-sectional study was determined for 325 ( 83% ) ; 25 ( 7 . 7% ) patients died during that time , but dates of death were only available for 19 , precluding a formal survival analysis . Patients with T . cruzi infection were more likely to have died than those without infection ( 21/210 ( 10% ) versus 4/115 ( 3% ) ; p = 0 . 05 ) . However , the strongest predictor of death was ACC/AHA heart failure stage: 19 ( 15% ) of 125 patients with stage C or D died , compared to 0 ( 0% ) of 59 with stage A and 1 ( 1% ) of 102 with stage B ( p<0 . 001 ) . Of the 19 patients with ACC/AHA stage C or D , 15 ( 79% ) had Chagas disease; the one deceased patient with ACC/AHA stage B also had Chagas disease . Five of the deceased patients , all with T . cruzi infection , had not had echocardiograms and therefore were missing ACC/AHA stage data . Males and females were equally likely to have T . cruzi infection ( Table 1 ) . Infected patients were older than uninfected patients ( mean 53 . 7 vs . 48 . 3 years , p = 0 . 001 ) . T . cruzi infection prevalence increased with age to a peak at approximately 58 for males and 69 for females , and then declined , demonstrating a significant quadratic relationship . Age was therefore included as a quadratic function in models with T . cruzi infection as the outcome . The prevalence of hypertension , coronary artery disease , diabetes and BMI >25 did not differ among T . cruzi-infected versus uninfected patients ( Table 1 ) . There were no significant differences in findings for overweight ( BMI 25–29 . 9 ) compared to obese patients ( BMI >30 ) . We therefore used a combined category that included overweight and obesity . In both unadjusted and adjusted analyses , T . cruzi-infected patients had significantly higher prevalence of arrhythmias and severe congestive heart failure as measured by ACC/AHA classification , and were more likely to have been recruited from the inpatient service compared to uninfected patients . In the model adjusted for age and sex , infected patients were less likely to report coronary artery disease; this difference was borderline significant . There were no differences by infection status in the frequency of valvular heart disease , congenital heart disease , stroke , or transitory ischemic attacks , or of reported symptoms such as dizziness or lightheadedness , syncope , palpitations , typical or atypical chest pain , exertional dyspnea , dysphagia , or constipation ( data not shown , but available upon request ) . Infected and uninfected patients were equally likely to have an implanted pacemaker ( 14/251 versus 5/142; p = 0 . 47 ) . In models adjusted for age and sex , T . cruzi-infected patients were significantly more likely than uninfected patients to have a right bundle branch block ( RBBB ) , left ventricular dilatation , ‘pure’ left atrial dilatation ( defined as left atrial end diastolic diameter >40 mm not explained by diastolic dysfunction or LVH ) , low ejection fraction , diffuse wall motion abnormalities , apical aneurysm or intracavitary thrombi , and less likely to have LVH ( Table 2 ) . In a multivariable logistic regression model , rural residence , poor housing conditions and increasing duration of residence in an endemic province were associated with increased risk of T . cruzi infection , with a significant interaction between the latter two variables ( Table 3 ) . For each additional decade living in an endemic province , people who lived in poor housing experienced a 3 . 2-fold increase risk of T . cruzi infection compared a 1 . 7-fold increase for those who lived in moderate or good housing . In univariate analyses , increasing age ( odds ratio [OR] per 10-year increase in age 1 . 40; 95% confidence interval [CI] 1 . 09–1 . 81 ) , male sex ( OR 5 . 87; 95% CI 3 . 04–11 . 31 ) , having lived in poor housing ( OR 2 . 46; 95% CI 1 . 29–4 . 71 ) and decades living in an endemic province ( OR 1 . 24; 95% CI 0 . 99–1 . 54 ) were associated with elevated risk of severe disease , defined as Chagas cardiomyopathy stage C or D . Having BMI >25 was associated with a trend to decreased risk of severe disease ( OR 0 . 56; 95% CI 0 . 29–1 . 10 ) . In a multivariable model , male sex , increasing age and poor housing remained strong predictors of severe disease ( Table 4 ) . In the adjusted models , no significant association was detected for other variables ( including BMI >25 , hypertension , coronary artery disease , or diabetes , positive results by PCR , prior treatment for T . cruzi infection , and cumulative time living in an infested house or rural area ) . Of 251 patients with confirmed positive results by serology , 109 ( 43% ) had positive results by conventional PCR , and of these , 89 ( 81 . 7% ) also had positive results by real time PCR . Among specimens with positive results by real time PCR , the median number of parasite copies/mL was 79 . 4 ( range 1 . 1 – 313 , 330 ) . One patient ( excluded from the epidemiological analyses ) had positive results by 1 of 3 serological tests , but positive results by conventional and real time PCR . If this patient is considered uninfected based on the serological results , both conventional and real time PCR had calculated specificities of 99 . 3% ( 142/143 ) . Specimens with positive results by PCR had higher mean Chagatek and Wiener ELISA absorbance values ( defined as optical density minus ELISA plate cut-off value ) than those with negative PCR results ( median Chagatek absorbance 1 . 34 vs . 1 . 23 , p = 0 . 006; Wiener 2 . 28 vs . 2 . 21 , p<0 . 001 ) . However , the distribution of absorbance values for specimens positive and negative by PCR had a high degree of overlap . There was no difference in ELISA absorbance values by Chagas severity stage ( P = 0 . 73 ) . Epidemiological associations with positive results by conventional PCR and real time PCR were similar . Among seropositive participants , males were significantly more likely than females to have positive results by conventional PCR ( OR 2 . 16; 95% CI 1 . 28–3 . 65 ) or by real time PCR ( OR 1 . 830; 95% CI 1 . 03–3 . 26 ) . In additional univariate analyses , positive conventional PCR results were less frequent among patients with BMI >25 ( OR 0 . 40; 95% CI 0 . 22–0 . 71 ) , a history of hypertension ( OR 0 . 54; 95% CI 0 . 31–0 . 92 ) or signs of left atrial dilatation on echocardiogram ( OR 0 . 50; 95% CI 0 . 29–0 . 85 ) . There was no association between severity of heart failure or Chagas cardiomyopathy and positive results by conventional or real time PCR . In a multivariable logistic regression model adjusted for age and disease severity , male sex remained associated with higher odds and BMI >25 with lower odds of positive results by conventional PCR ( Table 5 ) . Parasite load as measured by real time PCR was higher among males than females ( mean copies/mL 4292 vs 1286 , P = 0 . 0032 ) , but there was no difference by BMI category ( P = 0 . 94 ) . A profound epidemiological and nutritional transition has been underway in Latin America for the last several decades [7] , [18] , [19] . In some countries , such as Chile and Uruguay , a pattern of non-communicable disease epidemiology similar to the United States or Europe is now predominant [18] . In other countries , such as Guatemala and Bolivia , the rural population still lives in pre-transitional conditions with limited access to adequate housing , diet , clean water and sanitation [20] . However , when the rural poor migrate to cities , they often experience an abrupt transition to more sedentary lives and calorie-dense diets , and may add obesity , diabetes and hypertension to pre-transition conditions , including Chagas disease [21] . Our data offer a graphic illustration of the challenge facing health care systems in T . cruzi-endemic areas . We found that 59% of congestive heart failure cases were attributable to Chagas cardiomyopathy , and 79% of deaths in patients with advanced congestive heart failure occurred in those with Chagas disease . These figures are much higher than the 6–22% of congestive heart failure cases and 8% of congestive heart failure deaths attributed to Chagas disease in facility-based data from Brazil and Argentina reviewed in a recent publication [22] , and illustrate the very high disease burden still caused by T . cruzi in Bolivia today . Nevertheless , a majority of both T . cruzi-infected and uninfected patients in Santa Cruz had additional cardiovascular risk factors such as hypertension , obesity and diabetes . Although the increasing burden of chronic non-infectious disease in developing countries is well recognized [22] , there are no published data regarding potential interactions of these conditions with infectious diseases such as T . cruzi . Our data represent a first snapshot of their co-occurrence and highlight issues that merit further investigation , including more comprehensive assessment of cardiovascular risk factors and co-morbidities , population-based disease burden assessments , and longitudinal follow-up to evaluate prognosis and disease progression over time . Our study suffers from the inherent limitations of facility-based data: patterns may reflect differences in health care seeking behavior rather than biological differences , and the magnitude of the morbidity burden does not reflect what would be found in a population-based study because patients with more severe disease are more likely to seek care than those with milder disease . Consistent with the published literature [23] , [24] , [25] , we found male sex to be associated with a substantially increased risk of severe Chagas heart disease . As expected from the chronic progressive nature of the cardiomyopathy , severity also increased steadily with age . The fall-off in seroprevalence among older adults in endemic communities has long been attributed to excess mortality among infected individuals as their heart disease worsened after age 45 [26] , [27] . Survival is markedly shortened among Chagas cardiomyopathy patients once they have heart failure or other indicators of advanced cardiomyopathy [25] , [28] , [29] . We also observed similar age- and sex-related patterns . However , because our data are facility-based , we cannot rule out the possibility that they reflect differences in health care access . Morbidity and mortality from Chagas disease tend to occur at an older age now than previously [30] , and this may be reflected in the older age of peak seroprevalence in our data compared to earlier studies . In part , this represents a cohort effect: as vector control has decreased exposure and infection incidence among younger people , Chagas disease is becoming a disease of older adults [30] . In addition , better clinical management and health care access may have improved survival , especially for those living in urban areas [31] . Many experts also believe that intense exposure to infected vectors and repeated T . cruzi reinfection contributed to accelerated progression of cardiomyopathy and high rates of severe disease in the decades before the Southern Cone Initiative began in 1991 , and that vector control has led to an amelioration of Chagas cardiomyopathy since then [32] , [33] , [34] . Although this hypothesis is difficult to test in human populations , animal models provide supporting evidence that repeated infection worsens the long-term prognosis of Chagas heart disease [35] . We observed that patients who had lived in a house with earthen floor and walls were more likely to have severe Chagas cardiomyopathy; this finding may reflect poor housing conditions acting as a proxy for more intense vector exposure and reinfection risk . Our clinical findings are consistent with many published studies , demonstrating associations with right bundle branch block , left ventricular dilatation , low ejection fraction , apical aneurysms , and intracavitary thrombi [3] . The risk factors for infection identified in our data were not surprising , and as shown in many previous studies , included markers for or consequences of poverty , such as lower educational level , rural upbringing and having lived in houses with adobe walls and earthen floors . The intensification of risk seen when poor quality housing was combined with longer residence in an endemic province underscores the importance of maintaining both vector control and housing improvement programs in rural Bolivia , where infestation has not yet been eliminated [36] . We found no consistent association between parasite burden as indicated by positive PCR results and disease severity; this negative finding may result from the limitations of the cross-sectional design or the molecular methods , or may reflect the pathophysiology of the disease . In the last 10 years , a consensus has emerged that parasite persistence underlies the development and progression of Chagas cardiomyopathy [34] , [37] . In a study of left ventricular heart muscle tissue from necropsies of patients who died with different stages of chronic Chagas cardiomyopathy , a positive association among inflammation , fibrosis and parasite DNA was found [38] . Other authors endorse the importance of parasite persistence in tissue the pathogenesis of chronic Chagas' cardiomyopathy [39] , [40] Nevertheless , demonstration of a quantitative link between parasite burden and cardiac pathology or disease severity has been elusive [34] . This failure may be due in part to the limitations of existing detection methods . PCR assays , while more sensitive that culture or xenodiagnosis , are imperfect [41] . Sampling biases may also exist: when multiple specimens are examined , whether by xenodiagnosis , culture or PCR , many individuals have detectable parasitemia on some occasions but not others and parasite levels in peripheral blood may not correlate well with parasite loads in target tissues such as the heart [41] , [42] . In a life-long infection such as Chagas disease , the question of timing is also essential . In theory , the impact of parasite load on pathology could occur early but with enduring effects . In an earlier study of 56 infected individuals , positive PCR results at baseline was predictive of increased progression risk over a several-year period of longitudinal follow-up [23] . In contrast , other studies have failed to show any association between the presence or level of parasitemia and cardiomyopathy progression [24] , [41] , [42] . We found intriguing epidemiological associations with T . cruzi PCR results . Males were significantly more likely to have positive PCR results than females , and obesity was associated with a lower prevalence of positive results . Animal models suggest that male gonadal hormones decrease immune control of the parasite [43] . Orchiectomy of male mice enhances clearance of T . cruzi infection and reduces the amastigote burden in heart tissue , whereas parasitemia levels increase when castrated mice receive supplemental testosterone . In female mice , the parasitemia level rises after oophorectomy , and can be suppressed by estrogen and/or progesterone supplementation [44] . To our knowledge , the association of obesity with lower likelihood of positive results by T . cruzi PCR has not previously been reported in a clinical study . However , mouse models of chronic T . cruzi infection reveal large numbers of parasites in adipocytes . Indeed , more parasites are seen per adipocyte than in the tissues such as heart where parasites are classically thought to reside during chronic infection , suggesting that adipose tissue may represent a long-term reservoir and potentially a parasite sink [45] . Further investigation is needed to confirm the validity of these epidemiological observations and their physiological significance . One of our study participants had discordant serology and positive PCR results; we excluded her from the epidemiological analyses , because her results did not meet the standard criteria for confirmed chronic T . cruzi infection , which require at least 2 positive serological tests [11] . Nevertheless , we believe these results reflected a true infection . Positive PCR results in specimens from seronegative individuals have been reported in a number of studies , reflecting the fact that none of the diagnostic options for chronic T . cruzi infection has perfect sensitivity and specificity , and any given set of test results should be interpreted with care [41] , [46] , [47] , [48] . As control programs decrease T . cruzi incidence , attention is shifting to the large number of adults infected before these programs began . At the same time , T . cruzi-endemic countries are experiencing population shifts , urbanization and the epidemiological transition . Better understanding of the role of cardiac comorbidities and metabolic factors in the course and evolution of chronic Chagas cardiomyopathy will be essential to optimizing management of these patients in the future .
Latin America is undergoing a transition from disease patterns characteristic of developing countries with high rates of infectious disease and premature deaths to a pattern more like industrialized countries , in which chronic conditions such as obesity , hypertension and diabetes are more common . Many rural residents with Chagas disease have now migrated to cities , taken on new habits and may suffer from both types of disease . We studied heart disease among 394 adults seen by cardiologists in a public hospital in the city of Santa Cruz , Bolivia; 64% were infected with T . cruzi , the parasite that causes Chagas disease . Both T . cruzi infected and uninfected patients had a high rate of hypertension ( 64% ) and overweight ( 67% ) , with no difference by infection status . Nearly 60% of symptomatic congestive heart failure was due to Chagas disease; mortality was also higher for infected than uninfected patients . Males and older patients had more severe Chagas heart disease . Chagas heart disease remains an important cause of congestive heart failure in this hospital population , but often occurs in patients who also have obesity , hypertension and/or other cardiac risk factors .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[ "cardiovascular", "disorders" ]
2010
Chagas Cardiomyopathy in the Context of the Chronic Disease Transition
Pathogenic bacteria may modify their surface to evade the host innate immune response . Yersinia enterocolitica modulates its lipopolysaccharide ( LPS ) lipid A structure , and the key regulatory signal is temperature . At 21°C , lipid A is hexa-acylated and may be modified with aminoarabinose or palmitate . At 37°C , Y . enterocolitica expresses a tetra-acylated lipid A consistent with the 3′-O-deacylation of the molecule . In this work , by combining genetic and mass spectrometric analysis , we establish that Y . enterocolitica encodes a lipid A deacylase , LpxR , responsible for the lipid A structure observed at 37°C . Western blot analyses indicate that LpxR exhibits latency at 21°C , deacylation of lipid A is not observed despite the expression of LpxR in the membrane . Aminoarabinose-modified lipid A is involved in the latency . 3-D modelling , docking and site-directed mutagenesis experiments showed that LpxR D31 reduces the active site cavity volume so that aminoarabinose containing Kdo2-lipid A cannot be accommodated and , therefore , not deacylated . Our data revealed that the expression of lpxR is negatively controlled by RovA and PhoPQ which are necessary for the lipid A modification with aminoarabinose . Next , we investigated the role of lipid A structural plasticity conferred by LpxR on the expression/function of Y . enterocolitica virulence factors . We present evidence that motility and invasion of eukaryotic cells were reduced in the lpxR mutant grown at 21°C . Mechanistically , our data revealed that the expressions of flhDC and rovA , regulators controlling the flagellar regulon and invasin respectively , were down-regulated in the mutant . In contrast , the levels of the virulence plasmid ( pYV ) -encoded virulence factors Yops and YadA were not affected in the lpxR mutant . Finally , we establish that the low inflammatory response associated to Y . enterocolitica infections is the sum of the anti-inflammatory action exerted by pYV-encoded YopP and the reduced activation of the LPS receptor by a LpxR-dependent deacylated LPS . Lipopolysaccharide ( LPS ) is one of the major surface components of Gram-negative bacteria . The molecular structure of LPS is rather unique: an amphiphile with a hydrophobic region , the so-called lipid A , adjacent to a densely negatively charged polysaccharide . In Escherichia coli K-12 , the lipid A is a β ( 1′-6 ) -linked disaccharide of glucosamine phosphorylated at the 1 and 4′ positions with positions 2 , 3 , 2′ , and 3′acylated with R-3-hydroxymyristoyl groups , the so-called lipid IVA . The 2′and 3′R-3-hydroxymyristoyl groups are further acylated with laureate ( C12 ) and myristate ( C14 ) , respectively , by the action of the so-called late acyltransferases LpxL ( HtrB ) and LpxM ( MsbB ) , respectively [1] . When E . coli is grown at 12°C , LpxP , the cold-temperature-specific late acyltransferase , acts instead of LpxL adding palmitoleate ( C16∶1 ) [1] . Although the enzymes required to synthesize the lipid A are conserved throughout all Gram-negative bacteria there is heterogeneity on lipid A structure among Gram-negative bacteria compared to the E . coli K-12 . This is due to differences in the type and length of fatty acids , in the presence of decorations such as aminoarabinose or phosphoethanolamine and even in the removal of groups such as phosphates or fatty acids from lipid A [2] . LPS plays a crucial role during recognition of microbial infection by the host immune system . In fact , the lipid A moiety is a ligand of the Toll-like receptor 4 ( TLR4 ) /myeloid differentiation factor 2 complex [3] . The stimulation of this receptor complex triggers the activation of signalling cascades resulting in the induction of antimicrobial genes and release of cytokines , thereby initiating inflammatory and immune defence responses . Perusal of the literature demonstrates that changes in the number of acyl chains and in the phosphorylation status of the headgroup greatly affect the biological activity of lipid A . It is not surprising that some pathogens modulate their lipid A structure to alter their detection by the host; being these regulated changes important virulence traits ( for a review see [4] ) . Furthermore , given the importance of the LPS structure to the homeostasis of the outer membrane , it is possible that the aforementioned changes may also affect the physiology of the outer membrane as was recently demonstrated for Salmonella [5] . The genus Yersinia includes three human pathogens: Y . pestis , Y . pseudotuberculosis and Y . enterocolitica . The latter can cause food-borne infections in animals and humans ( yersiniosis ) , with symptoms such as enteritis and mesenteric lymphadenitis [6] . Y . enterocolitica is endowed with a repertoire of virulence factors that help bacteria to colonize the intestinal tract and to resist host defence mechanisms [7] , [8] . Temperature regulates most , if not all , virulence factors of yersiniae [7] , [8] . Recent studies have shown that temperature also regulates the structure of yersiniae lipid A [9]–[14] . Thus the number and type of the lipid A fatty acids and the substitutions of the 1- and 4′-positions in the glucosamine disaccharide can vary . Rebeil and co-workers [12] elegantly demonstrated that a shift in temperature induces a change in the number and type of acyl groups on the lipid A of the three Yersinia species . At 21°C , lipid As are mainly hexa-acylated whereas at 37°C they are tetra-acylated [12] . The temperature-dependent regulation of the lipid A acyltransferases underlines the shift in lipid A acylation both in Y . pestis and in Y . enterocolitica [12] , [14] . Pathogenic yersiniae also express hepta-acylated lipid A due to the addition of C10 , in Y . pestis and Y . pseudotuberculosis , or C16 ( palmitate ) , in Y . enterocolitica [12] , [14] , [15] . PagP is the acyltransferase responsible for the addition of palmitate to the lipid A in Y . enterocolitica [15] . Other lipid A species are consistent with the substitution of the phosphate at the 4′ end of the glucosamine disaccharide with aminoarabinose [15] . The aminoarabinose content is temperature-regulated in Y . pestis and in Y . enterocolitica [12] , [15] , [16]; being higher in bacteria gown at 21°C than at 37°C . Similar to other Gram-negative bacteria , the products of ugd and pmrHFIJKLM ( arnBCADTEF ) ( hereafter pmrF operon ) are required for the synthesis and addition of aminoarabinose to lipid A in Y . enterocolitica [15] . Finally , we and others [9]–[14] , [17] have reported a unique tetra-acyl lipid A species ( m/z 1388 ) found only in Y . enterocolitica grown at 37°C . Evidence support the notion that this species lacks the ester-linked R-3-hydroxymyristoyl group further acylated with laureate ( C12 ) [12] , [14] , [17] . Indeed , mass spectrometry analysis did confirm that the nonreducing glucosamine of the lipid A is substituted with only one ( amide-linked ) R-3-hydroxymyristoyl group further acylated with myristate ( C14 ) [17] . Altogether , these findings strongly suggest that the tetra-acyl lipid A species ( m/z 1388 ) may be caused by a deacylase removing the 3′-acyloxyacyl residue of the lipid A . The work described in this article gives experimental support to this hypothesis and explores the impact of the lipid A structure on Y . enterocolitica virulence traits . Further confirming previous findings [14] , [15] , lipid A isolated from Y . enterocolitica 8081 serotype O:8 ( hereafter YeO8; Table 1 ) grown at 37°C appeared to be identical to those reported by Rebeil et al . and Oertelt et al . [12] , [17] . The main species were a 3′-O-deacylated form ( m/z 1388 ) containing two glucosamines , two phosphates , three 3-OH-C14 , and one C14; and a hexa-acylated form ( m/z 1797 ) ( Figure 1A ) . In bacteria grown at 21°C , a minor species ( m/z 1414 ) was detected and may represent a 3′-O-deacylated form containing three 3-OH-C14 and one C16∶1 [12] . S . enterica serovar typhimurium and Helicobacter pylori also express 3′-O-deacylated lipid A species [18] , [19] . A membrane located hydrolase , named LpxR , removes the 3′-acyloxyacyl residue of lipid A in both organisms [18] , [19] . In silico analysis of the YeO8 genome ( accession number AM286415; [20] ) revealed that this pathogen may encode an LpxR orthologue ( locus tag YE3039 ) . The predicted YeO8 LpxR ( YeLpxR ) has 73% and 20% amino acid identities to S . enterica and H . pylori LpxR proteins , respectively . Furthermore , YeLpxR has 100% amino acid identity to Y . enterocolitca Y11 serotype O:3 ( locus tag Y11_05741; accession number FR729477 ) and Y105 serotype O:9 ( locus tag YE105_C2442; accession number CP002246 ) LpxR homologs . Analysis of the available Y . pestis and Y . pseudotuberculosis genomes revealed that they do not encode any gene similar to lpxR . YeO8 lpxR was mutated to determine whether this gene is indeed responsible for removing the 3′-acyloxyacyl residue of lipid A . MALDI-TOF mass spectrometry studies showed that , at 37°C , the lpxR mutant ( YeO8-ΔlpxRKm ) produced a lipid A which lacked the unique tetra-acyl lipid A species ( m/z 1388 ) found in YeO8 and only contained the hexa-acylated species ( m/z 1797; four 3-OH-C14 , one C12 and one C14 ) ( Figure 1B ) . At 21°C , lipid A isolated from YeO8-ΔlpxRKm was similar to that of YeO8 although without the minor species m/z 1414 ( Figure 1B ) . Complementation of the mutant with pTMLpxR restored the presence of the tetra-acyl species ( Figure 1C ) . In summary , our results confirmed the predicted function of Y . enterocolitica O:8 lpxR homolog as the lipid A 3′-O-deacylase . The LpxR-dependent lipid A deacylation was more evident on bacteria grown at 37°C than at 21°C , hence suggesting that the expression and/or function of the deacylase might be temperature-regulated , being higher at 37°C than at 21°C . To monitor transcription of lpxR quantitatively , a transcriptional fusion was constructed in which a promoterless lucFF gene was under the control of the lpxR promoter region ( see Material and Methods ) ; thereafter lpxR::lucFF was introduced into YeO8 and the luciferase activity was determined . The expression of the fusion was higher at 21°C than at 37°C ( Figure 2A ) . Real time ( RT ) quantitative PCR ( RT-qPCR ) experiments showed that lpxR mRNA levels were also higher at 21°C than at 37°C ( Figure 2B ) . To assess LpxR levels , the C-terminus of the protein was tagged with a FLAG epitope and the construct was cloned into the medium-copy plasmid pTM100 to obtain pTMLpxRFLAG ( see Materials and Methods ) . This plasmid restored the presence of the tetra-acyl species ( m/z 1414 and m/z 1388 ) in the lipid A of YeO8-ΔlpxRKm ( data not shown ) . Western blot analysis of purified membranes from YeO8-ΔlpxRKm containing pTMLpxRFLAG showed that LpxR levels were higher in membranes from bacteria grown at 21°C than at 37°C ( Figure 2C ) . Altogether , it can be concluded that the expression of lpxR is indeed temperature-regulated but , in contrast to our initial hypothesis , its expression is higher at 21°C than at 37°C . The apparent contradiction between the mass spectrometry analysis , more deacylation at 37°C , and the Western blot data , higher levels of LpxR at 21°C than at 37°C , led us to explore whether low temperature may affect the function of the enzyme . Since E . coli has been used as surrogate host to characterize Salmonella LpxR ( StLpxR ) function [18] , we mobilized pTMLpxR into E . coli MG1655 to analyze lipid A species by mass spectrometry in bacteria grown at 21°C and 37°C . Results shown in figure 3 demonstrate that LpxR did deacylate the E . coli lipid A from bacteria grown either at 21 or 37°C as detected by the presence of species m/z 1360 ( Figure 3C–D ) . This species was found previously in E . coli expressing StLpxR [18] . Of note , the species m/z 1414 , which is consistent with the deacylation of the species m/z 1850 containing palmitoleate ( C16∶1 ) instead of laureate ( C12 ) , was observed only in E . coli grown at 21°C . LpxP is the cold-temperature-specific late acyltransferase responsible for the addition of palmitoleate [1] . Altogether , our results indicate that the reduced LpxR-dependent deacylation found in YeO8 grown at 21°C cannot be attributed to a general lack of function of the enzyme at this temperature . We sought to determine why LpxR activity was not observed in YeO8 grown at 21°C despite the detection of the enzyme in the membrane . Among other possibilities , we speculated that specific features of YeO8 lipid A found only at 21°C might be responsible for the reduced LpxR activity . Furthermore , these features should be absent in E . coli grown at 21°C since LpxR-dependent activity was observed here . A conspicuous difference between YeO8 and E . coli lipid As is the presence of aminoarabinose and palmitate ( m/z 1954 and 2063 , respectively ) only in the former [14] , [15] . Therefore , we explored whether any of these modifications could account for the reduced LpxR activity . In YeO8 , similarly to other Gram-negative pathogens , the products of the pmrF operon are required for the synthesis and addition of aminoarabinose to lipid A whereas the acyltransferase PagP is required for the addition of palmitate to lipid A [15] . The lipid A from the pagP mutant , YeO8-ΔpagPGB , grown at 21°C resembled that of the wild-type strain , except that the species containing palmitate ( m/z 2063 ) was not detected ( Figure 4A ) . In contrast , the tetra-acylated species ( m/z 1414 ) was clearly observed in the lipid A from YeO8-ΔpmrF grown at 21°C ( Figure 4C ) . This was dependent on LpxR activity since the peak was absent in the double mutant YeO8-ΔpmrF-ΔlpxRKm ( Figure 4E ) . LpxR-dependent deacylation of lipid A ( m/z 1388 ) observed in bacteria grown at 37°C was not affected in either pmrF or pagP single mutants ( Figure 4B , D ) . Control experiments revealed that lpxR expression was not affected in YeO8-ΔpmrF since the expression of the lpxR::lucFF fusion was not significantly different between YeO8 and the pmrF mutant either grown at 21°C or at 37°C ( Figure 4G ) . On the whole , these results are consistent with the notion that the reduced LpxR activity observed in YeO8 at 21°C is associated with the lipid A modification with aminoarabinose . Our findings might suggest that aminoarabinose-containing LPS may directly inactivate the lipid A deacylase activity of YeLpxR . Alternatively , modification of lipid A with aminoarabinose could inhibit the physical interaction of LPS with YeLpxR . To explore this , the 3-D structure of YeLpxR was modeled ( Figure 5A ) . The amino acids 1–296 ( following the putative signal sequence ) could be modeled based on the crystal structure of StLpxR ( PDB code 3FID; [21] ) and the sequence alignment between StLpxR and YeLpxR ( Figure S1 ) . The fold of the resulting model is likely to be of good quality , since YeLpxR has such a high sequence identity to StLpxR ( 75% ) . Additionally , the important StLpxR amino acids identified by Rutten and co-workers [21] are conserved in YeLpxR . Six amino acids differ between the YeLpxR and the StLpxR active sites ( Figure S1 ) . Major differences are D31 and Q35 in YeLpxR , of which D31 is closer to the active site ( Figure 5B ) . The corresponding amino acids are much smaller in StLpxR , glycine and an alanine , respectively , which cause StLpxR to have a bigger cavity . StLpxR has a protruding cavity close to K67 , which cannot be found in YeLpxR ( Figure 6A ) . The difference is induced by D31 in YeLpxR , which occupies more space than G31 in StLpxR . As a consequence , the conserved K67 adopts a different conformation in the YeLpxR model . Due to D31 , the cavity in YeLpxR is divided into two parts with a narrow connection , and this amino acid also prevents YeLpxR from forming an inward protruding cavity similar to the one found near G31 in StLpxR ( Figure 6A ) . Docking of a modified Kdo2-lipid A molecule ( see Materials and methods ) to the model of YeLpxR showed that the phosphate group , which attaches aminoarabinose to Kdo2-lipid A , binds into the cavity in the vicinity of K67 and D31 ( Figure 6B ) . Docking of the same molecule to the crystal structure of StLpxR yielded a result where the phosphate group was located in the protruding cavity close to K67 ( Figure 6C ) . As expected , docking of the modified Kdo2-lipid A molecule with aminoarabinose to the YeLpxR model did not give any valuable result . On the other hand , when the same molecule was docked to the StLpxR crystal structure , aminoarabinose was bound close to G31 . It occupies the space corresponding to the narrow connection of the two larger cavities in YeLpxR ( Figure 6D ) As a result from the modeling and docking studies , we suggest that Kdo2-lipid A with aminoarabinose cannot fit into the active site of YeLpxR due to D31 , hence leading to the inability of YeLpxR to deacylate Kdo2-lipid A with aminoarabinose . To confirm our predictions , we constructed LpxR mutants by site-directed mutagenesis ( see Material and Methods ) . In addition to the amino acids corresponding to the active site amino acids in StLpxR , we wanted to study the effect of the D31G mutation for YeLpxR as the modelling and docking studies suggested that D31 has an important role in the YeLpxR specificity for the Kdo2-lipid A species . The constructs were introduced into E . coli MG1655 and the lipid A from the transformants grown at 37°C was analyzed by MALDI-TOF mass spectrometry . Most of the constructs containing LpxR mutants did trigger the deacylation of E . coli lipid A , detected by the presence of species m/z 1360 , ( Table 2 ) . In contrast , constructs containing LpxR mutants , LpxR ( N9A ) , LpxR ( D10A ) , LpxR ( S34A ) , and LpxR ( H122A ) did not deacylate E . coli lipid A . These results were expected since Rutten and co-workers have reported that these residues are located in the StLpxR active site and all of them are conserved in LpxR homologues [21] . Next , only those constructs triggering deacylation of E . coli lipid A were introduced into YeO8 . When the YeO8 strains were grown at 37°C , all LpxR mutants restored the presence of the tetra-acyl species ( m/z 1388 ) in the lipid A of YeO8-ΔlpxRKm ( Table 2 ) . Additionally , the mass spectrometry analysis revealed that LpxR ( D31G ) mutant did trigger the deacylation of lipid A in bacteria grown at 21°C as it was detected the presence of lipid A species m/z 1414 and m/z 1545 ( Figure 7B ) . The latter is consistent with the deacylation of the lipid A species modified with aminoarabinose ( m/z 1954 ) . In summary , our results further confirmed the amino acids important for the catalytic activity of YeLpxR . Moreover , our results confirmed the molecular modelling predictions , thereby demonstrating that the presence of D31 in the active site pocket of YeLpxR causes steric hindrance for the binding and deacylation of lipid A species modified with aminoarabinose . In YeO8 the expression of the loci responsible for the lipid A modification with aminoarabinose , ugd and pmrF operon , is temperature regulated , being higher at 21°C than at 37°C [15] . Mechanistically , this is so because the expression of the positive regulators phoPQ and pmrAB , which control the expression of ugd and the pmrF operon , is also higher at 21°C than at 37°C [15] . In turn , the temperature-dependent regulation of phoPQ and pmrAB is explained by H-NS-dependent negative regulation alleviated by RovA , another major regulator of Yersinia [22] , [23] , at 21°C [15] . Moreover , there is cross-talk between the regulators in such way that PhoPQ and PmrAB regulate positively the expression of rovA and the effect of PhoPQ is more important [15] . The inverse correlation between the substitution of the lipid A with aminoarabinose and lipid A deacylation , prompted us to evaluate whether phoPQ and pmrAB might negatively regulate lpxR . Results shown in figure 8 revealed that the expression of lpxR::lucFF was significantly up-regulated in the phoPQ and pmrAB mutants at 21°C and 37°C ( Figure 8A ) . However , the expression of lpxR reached wild-type levels in the double phoPQ-pmrAB mutant regardless the bacteria growth temperature ( Figure 8A ) . RT-qPCR experiments showed that the levels of lpxR mRNA were higher in the phoPQ and pmrAB mutants than in the wild type and double phoPQ-pmrAB mutants , which were not significantly different ( Figure S2 ) . Recently , we have shown that rovA expression is downregulated in the phoPQ and pmrAB single mutants , being the lowest in the phoPQ mutant , whereas in the phoPQ-pmrAB double mutant rovA expression is not significantly different to that in the wild type [15] . Therefore , the fact that lpxR expression follows the opposite trend in these mutants led us to analyze whether rovA negatively regulates the expression of lpxR . Indeed , luciferase activity was higher in the rovA mutant than in the wild type and the levels were not significantly different that those observed in the phoPQ mutant when bacteria were grown either at 21°C or 37°C ( Figure 8A ) . Similar results were obtained when the lpxR mRNA levels were analyzed by RT-qPCR ( Figure S2 ) . The increased lpxR expression observed in rovA and phoPQ single mutants at 21°C was no longer found in the double mutant rovA-phoPQ ( Figure 8A and Figure S2 ) . When bacteria were grown at 37°C , lpxR expression in the rovA-phoPQ mutant was significantly lower than those observed in the rovA and phoPQ single mutants ( p<0 . 05 for each comparison versus rovA-phoPQ mutant ) although still higher than that in the wild type ( Figure 8A and Figure S2 ) . Of note , the expression of lpxR was no longer temperature regulated in the rovA-phoPQ mutant ( Figure 8B ) . The fact that the expression of lpxR::lucFF in the triple mutant rovA-phoPQ-pmrAB at 21°C was less than in the wild-type strain may support the notion that , in the absence of the negative regulator RovA , PmrAB and/or a PmrAB-modulated regulator positively regulates lpxR . At 37°C , lpxR expression in the triple mutant was not significantly different than those found in the double mutant phoPQ-pmrAB and the wild type ( Figure 8A and Figure S2 ) . Collectively , our data revealed that the expression of lpxR is negatively controlled by the same regulators that activate the loci necessary for the substitution of the phosphate at the 4′ end of the glucosamine disaccharide with aminoarabinose . In a previous study , we observed the down regulation of YeO8 virulence factors in mutants lacking the lipid A late acyltransferases LpxM , LpxL or LpxP [14] . These results raised the possibility that lipid A acylation may act as a regulatory signal by acting on a transduction pathway ( s ) [14] . In this context , we sought to determine the impact of LpxR to the expression/function of YeO8 virulence factors . Virulence genes can be regulated as part of the flagellar regulon , indicating that this regulon contributes to Y . enterocolitica pathogenesis [24] . YeO8 is motile when grown at 21°C but not at 37°C [25] and previously we showed that LpxM and LpxP mutants are less motile than the wild type [14] . We examined the influence of LpxR on the flagellar regulon . We quantified the migration of the wild type and YeO8-ΔlpxRKm in motility medium ( 1% tryptone-0 . 3% agar plates ) . Figure 9 shows that YeO8-ΔlpxRKm was less motile than the wild type . Yersinia motility is related to the levels of flagellins which , in turn , are regulated by the expression of flhDC , the flagellum master regulatory operon [25] , [26] . We hypothesized that the expression of flhDC could be lower in the lpxR mutant than in the wild type . To address this , the flhDC::lucFF transcriptional fusion [26] was introduced into the chromosome of the strains and the luciferase activity was determined . At 21°C , luminescence was lower in the lpxR mutant than in the wild type ( Figure 9B ) . Complementation of the lpxR mutant with pTMYeLpxR restored flhDC::lucFF expression to wild-type levels ( Figure 9B ) . Notably , the catalytic inactive LpxR mutants LpxR ( N9A ) and LpxR ( S34A ) , encoded by pTMLpxR ( N9A ) and pTMLpxR ( S34A ) respectively , also complemented the lpxR mutant ( Figure 9B ) . Western blot analysis of purified membranes from YeO8-ΔlpxRKm containing pTMLpxR ( N9A ) FLAG or pTMLpxR ( S34A ) FLAG showed that the mutant proteins were expressed ( Figure S3 ) . When the strains were grown at 37°C , YeO8 and YeO8-ΔlpxRKm produced the same luminescence ( Figure 9B ) . One virulence gene that is regulated as part of the flagellar regulon is yplA and hence its expression is regulated by flhDC [24] , [27] , [28] . Considering that flhDC expression was downregulated in the lpxR mutant , we speculated that yplA expression could be affected in this mutant . The transcriptional fusion yplA::lacZYA [29] was introduced into the chromosome of the wild type and the lpxR mutant and their β-galactosidase activities were measured . Indeed , the β-galactosidase activity was lower in YeO8-ΔlpxRKm than in the wild type ( Figure 9C ) . Plasmids pTMYeLpxR , pTMLpxR ( N9A ) and pTMLpxR ( S34A ) complemented the phenotype ( Figure 9C ) . In summary , these results indicate that the flagellar regulon is downregulated in the lpxR mutant with a concomitant decrease in motility and downregulation of yplA expression . Inv is an outer membrane protein of Y . enterocolitica responsible for invasion of the host [30] , [31] . Since YeO8 lipid A mutations affect inv expression [14] , we asked whether inv expression is altered in the lpxR mutant . An inv::phoA translational fusion [32] was introduced into the genome of YeO8 and YeO8-ΔlpxRKm and inv expression was monitored as alkaline phosphatase ( AP ) activity ( Figure 10A ) . AP activity was significantly lower in the lpxR mutant than in the wild type . Plasmids pTMYeLpxR , pTMLpxR ( N9A ) and pTMLpxR ( S34A ) restored AP activity to wild-type levels . These differences in inv expression prompted us to study the ability of YeO8-ΔlpxRKm to invade HeLa cells by using a gentamicin protection assay . The amount of intracellular bacteria was 55% lower when cells were infected with the lpxR mutant than with the wild type ( Figure 10B ) . RovA is required for inv expression in Y . enterocolitica [33] . Therefore , among other possibilities , the low inv expression found in the lpxR mutant could be caused by downregulation of rovA expression . To address this , the rovA::lucFF transcriptional fusion [14] was introduced into the genome of the wild type and the lpxR mutant and the luminescence was determined . Results shown in figure 10C demonstrate that rovA expression was dowregulated in YeO8-ΔlpxRKm . This phenotype was complemented with plasmids pTMYeLpxR , pTMLpxR ( N9A ) and pTMLpxR ( S34A ) . Together , our data show that the down-regulation of inv expression found in the lpxR mutant is most likely caused by downregulation of rovA expression , the positive transcriptional regulator of inv . Y . enterocolitica harbours a plasmid ( pYV ) -encoded type III secretion system which is required for virulence . A set of virulence factors , called Yops , are secreted by this system and enable Y . enterocolitica to multiply extracellularly in lymphoid tissues [34]–[36] . In several pathogens , LPS polysaccharide status affects the expression of the type III secretion systems [37]–[39] . Therefore , we asked whether the production of the Yersinia pYV-encoded type III secretion system is altered in the lpxR mutant . At 37°C and under low calcium concentrations , this system secretes the Yops to the culture supernatant [40] . Analysis of Yop secretion revealed that the wild type and the lpxR mutant secreted similar levels of Yops ( Figure 11A ) . We sought to determine whether the translocation of Yops to the cytosol of eukaryotic cells is affected in the lpxR mutant . Detection of cytoskeleton disturbances upon infection of epithelial cells is one of the most sensitive assays to establish Yop translocation [41] . The injection of YopE into the cytosol of A549 cells by wild-type bacteria induced disruption and condensation of the actin microfilament structure of the cells whereas this was not the case when cells were infected with YeO8-ΔyopE mutant ( Figure 11B ) . YopE translocation to A549 cells was not affected in the lpxR mutant background ( Figure 11C ) . As expected , A549 cells infected with YeO8-ΔlpxRKm displayed similar cytoskeleton disturbances than those cells infected with the wild type ( Figure 11B ) . yadA is another pYV-encoded virulence gene whose expression is only induced at 37°C [42] . YadA is an outer membrane protein mediating bacterial adhesion , bacterial binding to proteins of the extracellular matrix and complement resistance ( for a review see [43] ) . Analysis of YadA expression by SDS-PAGE demonstrated that YeO8-ΔlpxRKm and YeO8 produced the same amount of the protein ( Figure 11D ) . To assess YadA functionality , we asked whether the YadA-dependent binding to collagen is altered in the lpxR mutant . To this end , we analyzed the binding of YadA-expressing whole bacteria to collagen type I by immunofluorescence ( see Material and Methods ) . In contrast to the negative control , a pYV-cured strain ( YeO8c ) , YeO8 and YeO8-ΔlpxRKm bound to collagen without differences between them ( Figure 11E–F ) . Taken together , these results suggest that the production and function of the pYV-encoded virulence factors Yops and YadA are not altered in the lpxR mutant . Cationic antimicrobial peptides ( CAMPs ) belong to the arsenal of weapons of the innate immune system against infections . In the case of Gram-negative bacteria , CAMPs interact with the lipid A moiety of the LPS [44]–[47] and lipid A modification is one of the strategies employed by Gram-negative bacteria to counteract the action of CAMPs . We and others have used polymyxin B as a model CAMP since it also binds to lipid A . Furthermore , resistance to this peptide reflects well the resistance to other mammalian peptides and correlates with virulence [48]–[51] . Therefore we evaluated the resistance of the lpxR mutant to polymyxin B . Results shown in figure 12A demonstrate that the mutant was as resistant as the wild type to the peptide when grown either at 21°C or at 37°C . Of note both strains were more susceptible to polymyxin B when grown at 37°C than at 21°C ( Figure 12A ) . The mammalian immune system recognizes and responds to E . coli LPS via the TLR4 complex , resulting in the synthesis and secretion of pro-inflammatory cytokines that recruit immune cells to the site of infection . The ability of LPSs to evoke inflammatory responses and the potency of them are directly related to the structure of the molecule . It has been reported that underacylated LPSs are less inflammatory than hexa-acylated ones , being the E coli lipid A ( m/z 1797 ) the prototype of hexa-acylated LPSs [52] . Therefore , the dramatic changes in lipid A acylation displayed by the lpxR mutant at 37°C led us to evaluate the immunostimulatory properties of YeO8 and YeO8-ΔlpxRKm . As cellular read-out , we determined TNFα levels secreted by macrophages infected either with the wild type or the lpxR mutant grown at 21°C and 37°C . YeO8 and YeO8-ΔlpxR induced similar levels of TNFα although the levels induced by bacteria grown at 37°C were significantly lower than those triggered by bacteria grown at 21°C ( p<0 . 05 for comparison of TNFα levels between temperatures for a given strain ) ( Figure 12B ) . This was dependent on the well known anti-inflammatory action of the pYV-encoded YopP [53] , [54] , since a yopP mutant grown at 37°C induced similar levels of TNFα than those induced by wild-type bacteria grown at 21°C ( Figure 12B ) . Therefore we sought to determine whether YopP could be counteracting the inflammatory response induced by YeO8-ΔlpxRKm . Indeed , YeO8-ΔyopP-ΔlpxRKm induced the highest levels of TNFα ( Figure 12B ) . Further sustaining this notion , the TNFα levels induced by the lpxR mutant cured of the pYV virulence plasmid grown at 37°C were significantly higher than those induced by the virulence plasmid negative wild-type strain but not different than the YeO8-ΔyopP-ΔlpxRKm-triggered TNFα levels ( Figure 12B ) . Of note , the TNFα levels induced by the virulence plasmid negative wild-type strain grown at 37°C were significantly lower than those triggered by bacteria grown at 21°C hence further highlighting the importance of lipid A acylation on the immunostimulatory properties of YeO8 . Pathogenic yersiniae show a temperature-dependent variation in lipid A acylation [9]–[14] . At 21°C , Y . enterocolitica synthesizes hexa-acylated lipid A containing four 3-OH-C14 , one C12 and either one C16∶1 or one C14 . At 37°C , Y . enterocolitica lipid A presents a tetra-acylated species ( m/z 1388 ) and a hexa-acylated one containing four 3-OH-C14 , one C12 and C14 . In a previous work , we identified and characterized the acyltransfreases , lpxM , lpxL and lpxP , responsible for the addition of C12 , C14 and C16∶1 , respectively , to lipid A [14] . Moreover , we demonstrated that the expressions of these enzymes are temperature regulated [14] . However , the unique tetra-acyl lipid A found in the wild type grown at 37°C ( m/z 1388 ) remained to be explained at the molecular level . We and others have established that this species is consistent with 3′-O-deacylation of lipid A [12] , [14] , [17] . In this work by combining biochemistry , genetics and molecular modelling we present evidence that LpxR is the lipid A 3′-O-deacylase of Y . enterocolitica . YeLpxR is one of the closest homologues to StLpxR . Despite the presence of StLpxR in the Salmonella outer membrane , the bacterium does not produce 3′-O-deacylated lipid A species under any growth conditions tested to date [18] . This has been termed as enzyme latency and similar findings have been reported for the Salmonella lipid A 3-O-deacylase PagL and E . coli PagP [55] , [56] . Our data revealed that YeLpxR is also latent in the membrane of YeO8 grown at 21°C . However , this is not a general feature of lipid A deacylases since H . pylori LpxR is constitutively active [19] . Several explanations could underlie YeLpxR latency at 21°C . Firstly , we explored whether low temperature may affect the function of the enzyme . The fact that YeLpxR did deacylate E . coli lipid A when grown at 21°C does not support that low temperatures grossly inhibit the enzyme activity . Nevertheless , we do not by any means completely rule out that temperature may affect YeLpxR activity , and thorough biochemical analyses are warranted to rigorously define the functional parameters of YeLpxR activity . This will be the subject of future studies . We next hypothesized that specific features of YeO8 lipid A , which do not exist in the E . coli lipid A , may be responsible for YeLpxR latency . The first conspicuous difference is the type of secondary fatty attached to the lipid IVA . In E . coli the late acyltransferases LpxL and LpxM add laureate ( C12 ) and myristate ( C14 ) respectively [1] whereas in YeO8 these enzymes transfer myristate ( C14 ) and laureate ( C12 ) respectively [14] . However , this cannot account for the reduced LpxR activity since the enzyme did deacylate E . coli lipid A . The presence of palmitoleate in YeO8 lipid A at 21°C but not at 37°C cannot be the reason since YeLpxR deacylated E . coli lipid A containing palmitoleate , found in E . coli grown at 21°C . Instead , our results revealed that the lipid A substitution with aminoarabinose is associated with YeLpxR latency since LpxR-dependent lipid A deacylation was clearly observed in the pmrF mutant grown at 21°C . Notably , the lack of aminoarabinose also releases Salmonella PagL from latency [56] , hence suggesting a key role for the lipid A modification with aminoarabinose in LPS remodelling . The molecular modelling and docking experiments further highlighted the importance of lipid A substitution with aminoarabinose for YeLpxR function . D31 in YeLpxR forces the conserved K67 to adopt a different conformation compared to StLpxR . According to the docking results , the resulting loss of cavity space in the vicinity of K67 in YeLpxR , causes the phosphate at the 4′ end of Kdo2-lipidA to bind somewhat differently to YeLpxR than to StLpxR . In the latter , the phosphate binds in the cavity near K67 , while in YeLpxR it is forced to bind more outwards from the enzyme . The docking of Kdo2-lipidA with aminoarabinose to StLpxR showed that aminoarabinose occupies the cavity space , which corresponds to a narrow connection between two larger cavities in YeLpxR . The large reduction in cavity volume at this particular site causes this space to be too small for the accommodation of aminoarabinose . Hence , D31 seems to cause steric hindrance for the binding of aminoarabinose-containing Kdo2-lipidA to YeLpxR . Therefore , we predicted that D31 could have an important role for the YeLpxR substrate specificity . Indeed , the site-directed mutagenesis experiments validated that the presence of D31 in the active site pocket of YeLpxR causes a steric hindrance for the binding and deacylation of lipid A species modified with aminoarabinose . Nevertheless , at present we do not rule out that other residues of YeLpxR also contribute to its latency . In this regard , Salmonella PagL is released from latency when specific amino acid residues located at extracellular loops of the enzyme are mutated and it has been postulated that these residues are involved in the recognition of aminoarabinose-modified lipid A [56]–[58] . Studies are going to explore whether residues located at extracellular loops of LpxR also contribute to enzyme latency . The inverse correlation between the aminoarabinose content in the LPS and the LpxR-dependent lipid A deacylation prompted us to evaluate whether the same regulatory network governing the expression of the pmrF operon and ugd could regulate lpxR . Recently , we have shown that the global regulators RovA , PhoPQ , and PmrAB positively control the expression of the loci necessary for aminoarabinose biosynthesis at 21°C [15] . Furthermore , there is a cross-talk between these regulators since the expressions of phoPQ and pmrAB are downregulated in the rovA mutant whereas rovA expression is downregulated in phoPQ and pmrAB single mutants [15] . Our findings support the notion that RovA and PhoPQ are negative regulators of lpxR since its expression was higher in phoPQ and rovA single mutant backgrounds than in the wild type . In turn , the two-component system PmrAB and/or a PmrAB-regulated system may act as a positive regulator because lpxR expression was similar in the wild-type and rovA-phoPQ backgrounds . One striking finding of our study is that motility and invasion of eukaryotic cells were reduced in the lpxR mutant grown at 21°C . Mechanistically , our data revealed that the expressions of flhDC and rovA , the key regulators controlling the flagellar regulon and invasin respectively [22] , [25] , [33] , were down-regulated in the lpxR mutant . Although we have reported that lipid A acylation status affects motility and invasion [14] , the phenotypes were found in mutants lacking the late-acyltransferases and hence displaying major changes in the lipid A structure at 21°C [14] . This is in contrast to the lpxR mutant grown at 21°C , where the LpxR-dependent deacylation was hardly observed . The fact that YeLpxR is in latent stage at this growth temperature may suggest that , in the lpxR mutant background , the absence of the enzyme in the outer membrane , not the lipid A deacylation , acts as the regulatory signal underlying the reduced expressions of flhDC and rovA . Given experimental support to this hypothesis , the catalytically inactive mutants LpxR ( N9A ) and LpxR ( S34A ) restored the expressions of flhDC , ylpA , inv and rovA to wild-type levels . These results are in good agreement with the notion that membrane-intrinsinc β-barrel proteins , such as LpxR , may launch transmembrane signal transduction pathways upon sensing outer membrane perturbations [59] , in our case , the absence of the protein itself . Therefore , it can be speculated that those systems sensing extracytoplasmatic stresses could underlie the regulatory connection between the absence of LpxR and the expression of Y . enterocolitica virulence factors . Giving indirect support to our speculation , it has been reported that lipid A deacylation induces σE-dependent responses in E . coli [60] , the Cpx system senses changes in LPS O-polysaccharide [61] . Experiments are underway to test whether the activation status of the Cpx and/or σE systems is altered in the lpxR mutant background and whether any of these systems is responsible for the reduced expression of flhDC and rovA found in the mutant . The LPS contains a molecular pattern recognized by the innate immune system thereby arousing several host defence responses . On one hand , CAMPs target this LPS pattern to bind to the bacterial surface , which is necessary for their microbicidal action . On the other hand , recognition of the LPS by the LPS receptor complex triggers the activation of host defence responses , chiefly the production of inflammatory markers . Not surprisingly , the modification of the LPS pattern is a virulence strategy of several pathogens to evade the innate immune system , and Y . enterocolitica is not an exception . Recently , we have demonstrated that the temperature-dependent lipid A modifications with aminoarabinose and palmitate help Y . enterocolitica to avoid the bactericidal action of CAMPs [15] . In this context , it was not totally unexpected to find out that the lpxR mutant was as susceptible as the wild type to polymyxin B , a model CAMP , since the mass spectrometry analysis indicated that the aforementioned lipid A modifications were not affected in the lpxR mutant background . Concerning the activation of inflammatory responses , several studies highlight the critical role of pYV-encoded Yops , chiefly YopP , to prevent the activation of inflammatory responses in a variety of cells , including macrophages . Nevertheless , Rebeil and co-workers [12] conclusively demonstrated that purified LPS from Y . enterocolitca grown at 37°C is less inflammatory than that purified from bacteria grown at 21°C . This is in agreement with the concept that underacylated LPSs are less inflammatory than hexa-acylated ones [52] . Therefore , it was plausible to speculate that the LpxR-dependent deacylation of LPS at 37°C was responsible for the reduced stimulatory potential of the LPS described by Rebeil and co-workers . To confirm this speculation we chose to challenge macrophages with alive bacteria instead of using purified LPS since there might be differences between the cellular recognition of purified LPS and the LPS expressed in the complex lipid environment of the bacterial outer membrane . To our initial surprise , we observed that the lpxR mutant elicited similar inflammatory response than the wild type when both strains were grown at 37°C . The fact that these responses were significantly lower than those elicited by bacteria grown at 21°C suggested that pYV-encoded factors were attenuating the inflammatory response . Therefore , we hypothesized that the arsenal of Yops injected to the cell were efficiently counteracting the activation of inflammatory responses evoked by the lpxR mutant LPS . In fact , our data demonstrated that the production and function of the pYV-encoded virulence factors were not affected in the lpxR mutant . Giving support to our hypothesis , the inflammatory response elicited by the lpxR mutant cured of the pYV virulence plasmid grown at 37°C was significantly higher than that induced by the virulence plasmid negative wild-type strain . Moreover , our findings suggest that , among all Yops , YopP plays a major role in counteracting the inflammation elicited by the lpxR mutant since the TNFα levels induced by the lpxR mutant cured of the pYV virulence plasmid grown at 37°C were not different than those triggered by YeO8-ΔyopP-ΔlpxR . On the whole , our results and those reported by Rebeil and co-workers [12] are consistent with a model in which the characteristic low inflammatory response associated to Y . enterocolitica infections might be the sum of the anti-inflammatory action exerted by YopP and the reduced activation of the LPS receptor complex due to the expression of a LpxR-dependent deacylated LPS . In this scenario , the latency of LpxR may facilitate a quick bacterial response upon entering the host to reduce the initial recognition of the pathogen by the LPS receptor complex . This will allow the pathogen to activate other host countermeasures , among others the pYV-encoded type III secretion system , which is a time consuming process . Bacterial strains and plasmids used in this study are listed in Table 1 . Unless otherwise indicated , Yersinia strains were grown in lysogeny broth ( LB ) medium at either 21°C or 37°C . When appropriate , antibiotics were added to the growth medium at the following concentrations: ampicillin ( Amp ) , 100 µg/ml for Y . enterocolitica and 50 µg/ml for E . coli; kanamycin ( Km ) , 100 µg/ml in agar plates for Y . enterocolitica , 50 µg/ml in agar plates for E . coli , and 20 µg/ml in broth; chloramphenicol ( Cm ) , 20 µg/ml; trimethoprim ( Tp ) , 100 µg/ml; tetracycline ( Tet ) 12 . 5 µg/ml; and streptomycin ( Str ) , 100 µg/ml . In silico analysis led to the identification of Y . enterocolitica 8081 homologue of lpxR ( YE3039 ) , yopP ( YEP0083 ) and yopE ( YEP0053 ) [accession number AM286415; [20]] . To obtain the lpxR , yopP , and yopE mutants two sets of primers ( Table S1 ) were used for each gene to amplify two different fragments from each gene , LpxRUP and LpxRDOWN , YopPUP and YopPDOWN , YopEUP and YopEDOWN , respectively . Both fragments were BamHI-digested , purified , ligated , amplified as a single PCR fragment using a mixture of GoTaq Flexi polymerase ( 2 . 5 units/reaction; Promega ) and Vent polymerse ( 2 . 5 units/reaction; New England Biolabs ) , gel purified and cloned into pGEMT-Easy ( Promega ) to obtain pGEMTΔlpxR , pGEMTΔyopP , and pGEMTΔyopE respectively . A kanamycin resistance cassette flanked by FRT recombination sites was obtained as a BamHI fragment from pGEMTFRTKm and it was cloned into BamHI-digested pGEMTΔlpxR and pGEMTΔyopP to generate pGEMTΔlpxRKm and pGEMTΔyopPKm respectively . ΔlpxR::Km , and ΔyopP::Km alleles were amplified using Vent polymerase ( New England Biolabs ) and cloned into SmaI-digested pKNG101 to obtain pKNGΔlpxRKm and pKNGΔyopPKm , respectively . ΔyopE allele was obtained by PvuII-digestion of pGEMTΔyopE , gel purified and cloned into SmaI-digested pKNG101 to obtain pKNGΔyopE . pKNG101 is a suicide vector that carries the defective pir-negative origin of replication of R6K , the RK2 origin of transfer , and an Str resistance marker [62] . It also carries the sacBR genes that mediate sucrose sensitivity as a positive selection marker for the excision of the vector after double crossover [62] . Plasmids were introduced into E . coli CC118-λpir from which they were mobilized into Y . enterocolitica 8081 by triparental conjugation using the helper strain E . coli HB101/pRK2013 . Bacteria were diluted and aliquots spread on Yersinia selective agar medium plates ( Oxoid ) supplemented with Str . Bacteria from 5 individual colonies were pooled and allowed to grow in LB without any antibiotic overnight at RT . Bacterial cultures were serially diluted and aliquots spread in LB without NaCl containing 10% sucrose and plates were incubated at RT . The recombinants that survived 10% sucrose were checked for their antibiotic resistance . The appropriate replacement of the wild-type alleles by the mutant ones was confirmed by PCR and Southern blot ( data not shown ) . In the case of YeO8-ΔlpxRKm and YeO8-ΔyopPKm mutants , the kanamycin cassette was excised by Flp-mediated recombination [63] using plasmid pFLP2Tp . This plasmid is a derivative from pFLP2 constructed by cloning a trimethoprim resistance cassette , obtained by SmaI digestion of p34S-Tp [64] , into ScaI-digested pFLP2 . The generated mutants were named YeO8-ΔlpxR and YeO8-ΔyopP , respectively . YeO8-ΔyopP-ΔlpxRKm and YeO8-ΔpmrF-ΔlpxRKm double mutants were obtained mobilizing the pKNGΔlpxRKm plasmid into YeO8-ΔyopP and YeO8-ΔpmrF , respectively . The replacement of the wild-type alleles by the mutant ones was done as described above and confirmed by PCR ( data not shown ) . To cure the pYV plasmid from YeO8-ΔlpxRKm , bacteria were grown at 37°C in Congo Red Magnesium oxalate agar plates [65] . Colony size and lack of uptake of Congo Red were used to detect loss of the virulence plasmid . This was further confirmed by testing the YadA-dependent autoagglutination ability [66] . A 443 bp DNA fragment containing the promoter region of lpxR was amplified by PCR using Vent polymerase ( see Table S1 for primers used ) , EcoRI digested , gel purified and cloned into EcoRI-SmaI digested pGPL01Tp suicide vector [15] . This vector contains a promoterless firefly luciferase gene ( lucFF ) and a R6K origin of replication . A plasmid in which lucFF was under the control of the lpxR promoter was identified by restriction digestion analysis and named pGPL01TpYelpxR . This plasmid was introduced into E . coli DH5α-λpir from which it was mobilized into Y . enterocolitica by triparental conjugation using the helper strain E . coli HB101/pRK2013 . Strains in which the suicide vectors were integrated into the genome by homologous recombination were selected . This was confirmed by PCR ( data not shown ) . To complement the lpxR mutant , a DNA fragment of 1 . 5 kb was PCR-amplified using TaKaRa polymerase ( see Table S1 for primers used ) gel purified , and cloned into pGEMT-Easy ( Promega ) to obtain pGEMTComlpxR . A fragment , containing the putative promoter and coding region of the deacylase , was obtained by PvuII digestion of pGEMTComlpxR , gel purified and cloned into the ScaI site of the medium copy plasmid pTM100 [40] to obtain pTMLpxR . For the construction of plasmid pTMLpxRFLAG , the lpxR coding region with its own promoter and a FLAG epitope sequence right before the stop codon was PCR amplified using Vent polymerase , primers LpxRtagging and LpxrFLAG ( Table S1 ) and genomic DNA as template . The fragment was phosphorylated , gel purified and cloned into ScaI-digested pTM100 . pTMLpxR and pTMLpxRFLAG were introduced into E . coli DH5α-λpir and then mobilized into Y . enterocolitica strains by triparental conjugation using the helper strain E . coli HB101/pRK2013 . Lipid As were extracted using an ammonium hydroxide/isobutyric acid method and subjected to negative ion matrix-assisted laser desorption ionization time-of-flight ( MALDI-TOF ) mass spectrometry analysis [14] , [67] . Analyses were performed on a Bruker Autoflex II MALDI-TOF mass spectrometer ( Bruker Daltonics , Incorporated ) in negative reflective mode with delayed extraction . Each spectrum was an average of 300 shots . The ion-accelerating voltage was set at 20 kV . Dihydroxybenzoic acid ( Sigma Chemical Co . , St . Louis , MO ) was used as a matrix . Further calibration for lipid A analysis was performed externally using lipid A extracted from E . coli strain MG1655 grown in LB at 37°C . Interpretation of the negative-ion spectra is based on earlier studies showing that ions with masses higher than 1000 gave signals proportional to the corresponding lipid A species present in the preparation [9] , [12] , [17] , [68] . Important theoretical masses for the interpretation of peaks found in this study are: lipid IVA , 1405; C12 , 182 , C14 , 210; C16∶1 , 236 . 2; aminoarabinose ( AraNH ) , 131 . 1; C16 , 239 . Site-directed mutagenesis of the lpxR gene was performed by PCR [69] . Plasmid pTMLpxR , obtained with a minipreparation kit ( Macherey-Nagel ) , was used as template and the desired mutations were introduced by the primer pairs described in Table S1 . Amplifications were carried out in 50 µl reaction mixture using Vent DNA polymerase ( New England BioLabs . ) . The PCR was started with initial 70 sec incubation at 95°C and then steps ( 95°C 50 sec , 60°C 75 sec and 72°C 6 min ) were repeated 20 times followed by a 10 min extension time at 72°C . The obtained PCR products were gel purified , phosphorylated with T4 polynucleotide kinase , ligated , and digested with DpnI to break down any remaining template plasmid . The ligated PCR-product was transformed into E . coli C600 . Plasmid DNA was isolated from transformants and the lpxR gene was completely sequenced to confirm the generated mutations and to ensure that no other changes were introduced . The name of each mutant construct includes the wild-type residue ( single-letter amino acid designation ) followed by the codon number and mutant residue ( typically alanine ) . For the construction of plasmids pTMLpxR ( N9A ) FLAG and pTMLpxR ( S34A ) FLAG , the lpxR alleles encoded into pTMLpxR ( N9A ) and pTMLpxR ( S34A ) were PCR amplified using Vent polymerase , and primers LpxRtagging and LpxrFLAG ( Table S1 ) . The fragments were phosphorylated , gel purified and cloned into ScaI-digested pTM100 [40] . Plasmids were introduced into E . coli DH5α-λpir and then mobilized into Y . enterocolitica strains by triparental conjugation using the helper strain E . coli HB101/pRK2013 . Overnight 5-ml cultures of Y . enterocolitica strains were diluted 1∶21 into 100 ml of LB in a 250-ml flask . Cultures were incubated with aeration at 21°C or 37°C until OD600 0 . 8 . Bacteria were recovered by centrifugation ( 6500×g; 10 min , RT ) and they were resuspended in 2 ml of 10 mM Tris/HCl ( pH 7 . 4 ) -5 mM MgSO4 containing 2% Triton X-100 ( v/v ) . Cells were broken by sonication ( Branson digital sonifier; microtip 1/8″ diameter , amplitude 10% ) for 15×1 min cycles , each cycle comprised 1 min sonication step separated by 1 min intervals . Unbroken cells were eliminated by centrifugation ( 2000×g , 20 min ) , and cell envelopes were recovered by ultracentrifugation ( Beckman 70 . 1 Ti rotor; 45 000×g; 1 h , 4°C ) . The cell envelopes were resuspended in 500 µl of distilled water . The protein concentration was determined using the BCA Protein Assay Kit ( Thermo Scientifc ) . 80 µg of proteins were separated on 4–12% SDS-PAGE , and semi-dry electrotransferred onto a nitrocellulose membrane using as transfer buffer SDS-PAGE-urea lysis buffer [a freshly prepared 1∶1 mix of 1× SDS running buffer ( 12 mM Tris , 96 mM glycine , 0 . 1% SDS] and urea lysis buffer ( 10 mM Na2HPO4 , 1% β-mercaptoethanol , 1%SDS , 6 M urea ) ] [70] . Membrane was blocked with 4% skim milk in PBS . Membranes were stained using anti-Flag antibody ( 1∶2000; Sigma ) following the instructions of the supplier . A homology model of YeLpxR was constructed based on the crystal structure of StLpxR ( PDB code 3FID; [21] . The YeLpxR sequence was used as bait to search Protein Data Bank with the Basic Local Alignment Search Tool ( BLAST ) at NCBI ( http://blast . ncbi . nlm . nih . gov/ ) . A pairwise sequence alignment was made using the program MALIGN [71] in the BODIL modeling environment [72] , and a picture of the alignment was created using ESPript [73] . The essential water molecule and the zinc ion in the StLpxR crystal structure were also included in the YeLpxR model . A set of ten models was created with the program MODELLER [74] , from which the model with the lowest value of the MODELLER objective function was analyzed and compared to the crystal structure of StLpxR by superimposing with the program VERTAA ( Johnson & Lehtonen , 2004 ) in BODIL . Different rotamers for D10 and D31 were searched with the program Jackal ( http://wiki . c2b2 . columbia . edu/honiglab_public/index . php/Software:Jackal ) . D10 was changed to the same rotamer as in the crystal structure of StLpxR , while the rotamer used for D31 was the one with the lowest energy according to Jackal . SURFNET [75] was used to detect surface cavities , while PyMOL ( Version 1 . 4 , Schrödinger , LLC ) was used for preparing pictures . For the SURFNET calculations , the minimum radius for gap spheres was set to 1 . 5 Å and the maximum radius was 4 . 0 Å . For the docking studies , a Kdo2-lipid A , both with and without aminoarabinose , was modified from the coordinates for the LPS molecule in the crystal structure of FhuA [76] . The fatty acyl chains were removed from the Kdo2-lipid A molecule in order to reduce the number of rotatable bonds and make the docking more reliable . Aminoarabinose was added to the modified Kdo2-lipid A molecule with SYBYL ( Version 8 . 0 , Tripos Associates , Inc . , St Louis , MO , USA ) , and the structure was minimized with the conjugate gradient method and Tripos force field . The modified Kdo2-lipid A , both with and without aminoarabinose , was docked to the YeLpxR model and the StLpxR crystal structure ( PDB code 3FID ) with GOLD via Discovery Studio ( CSC IT Center for Science Ltd , Espoo , Finland ) , with default docking parameters and the receptor cavity defined to D10 , Q16 , T/S34 , K67 , and Y130 . The reporter strains were grown at 21°C or at 37°C on an orbital incubator shaker ( 180 r . p . m . ) until OD540 1 . 6 . The cultures were harvested ( 2500×g , 20 min , 24°C ) and resuspended to an OD540 of 1 . 0 in PBS . A 100 µl aliquot of the bacterial suspension was mixed with 100 µl of luciferase assay reagent ( 1 mM D-luciferin [Synchem] in 100 mM citrate buffer pH 5 ) . Luminescence was immediately measured with a Luminometer LB9507 ( Berthold ) and expressed as relative light units ( RLU ) . All measurements were carried out in quintuplicate on at least three separate occasions . Phenotypic assays for swimming motility were initiated by stabbing 2 µl of an overnight culture at the centre of agar plates containing 0 . 3% agar and 1% tryptone [25] , [26] . Plates were analyzed after 24 h of incubation at RT and the diameters of the halos migrated by the strain from the inoculation point were compared . Experiments were run in quadruplicate in three independent occasions . To measure flhDC expression , plasmid pRSFlhDC08 [26] encoding the transcriptional fusion flhDC::lucFF was integrated into the genomes of the strains by homologous recombination . This was confirmed by Southern blot ( data not shown ) . Luminescence was determined as previously described . β-galactosidase activity was determined as previously described with bacteria grown in 1% tryptone at RT [77] . Alkaline phosphatase activity was determined in permeabilized cells and the results are expressed in enzyme units per OD600 as previously described [78] . Experiments were run in duplicate in three independent occasions . Bacteria were grown at 21°C or at 37°C in 5 ml of LB medium on an orbital incubator shaker ( 180 r . p . m . ) until an OD600 of 0 . 3 . 0 . 5 ml of ice-cold solution EtOH/phenol [19∶1 v/v ( pH 4 . 3 ) ] were added to the culture and the mixture was incubated on ice for 30 min to prevent RNA degradation . Total RNA was extracted using a commercial NucleoSpin RNA II kit as recommended by the manufacturer ( Macherey-Nagel ) . cDNA was obtained by retrotranscription of 2 µg of total RNA using a commercial M-MLV Reverse Transcriptase ( Sigma ) , and random primers mixture ( SABiosciences , Quiagen ) . 50 ng of cDNA were used as a template in a 25-µl reaction . RT-PCR analyses were performed with a Smart Cycler real-time PCR instrument ( Cepheid , Sunnyvale , CA ) and using a KapaSYBR Fast qPCR Kit as recommended by the manufacturer ( Cultek ) . The thermocycling protocol was as follows; 95°C for 3 min for hot-start polymerase activation , followed by 45 cycles of 95°C for 15 s , and 60°C for 30 s . SYBR green dye fluorescence was measured at 521 nm . cDNAs were obtained from three independent extractions of mRNA and each one amplified by RT-qPCR in two independent occasions . Relative quantities of lpxR mRNAs were obtained using the comparative threshold cycle ( ΔΔCT ) method by normalizing to rpoB and tonB genes ( Table S1 ) . Overnight cultures of Y . enterocolitica strains were diluted 1∶50 into 25 ml of TSB supplemented with 20 mM MgCl2 and 20 mM sodium oxalate in a 100-ml flask . Cultures were incubated with aeration at 21°C for 2 . 5 h , and then transferred at 37°C for 3 h . The optical density at 540 nm of the culture was measured and the bacterial cells were collected by centrifugation at 1500×g for 30 min . Ammonium sulphate ( final concentration 47 . 5% w/v ) was used to precipitate proteins from 20 ml of the supernatant . After overnight incubation at 4°C , proteins were collected by centrifugation ( 3000×g , 30 min , 4°C ) and washed twice with 1 . 5 ml of water . Dried protein pellets were resuspended in 50 to 80 µl of sample buffer and normalized according to the cell count . Samples were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( SDS-PAGE ) using 12% polyacrylamide gels and proteins visualized by Coomassie brilliant blue staining . Control experiments revealed that the secretion of Yops was not affected in yopE and yopP mutants except that each mutant did not produce either YopE or YopP , respectively ( data not shown ) . Bacteria were grown overnight in 2 ml RPMI 1640 medium lacking phenol red at 37°C without shaking . The OD540 of the culture was measured and CFUs were determined by plating serial dilutions . Bacteria from 1-ml aliquot were recovered by centrifugation ( 16 000×g , 10 min , 4°C ) and resuspended in 200 µl of SDS-sample buffer . Samples were incubated for 4 h at 37°C and kept frozen at −20°C . Samples were analyzed by SDS-PAGE using 10% polyacrylamide gels and proteins visualized by Coomassie brilliant blue staining . Samples were normalized according to the cell count and they were not boiled before loading the gel . Overnight cultures of Y . enterocolitica strains grown at 37°C were diluted 1∶10 into 5 ml of LB and grown with aeration at 37°C for 2 . 5 h . bacteria were pelleted , washed once with PBS and resuspended to an OD540 of 0 . 3 in PBS . 12 mm circular coverslips in 24-well tissue culture plates were coated overnight at 4°C with 10 µg/ml human collagen type I ( Sigma ) in PBS ( final volume 100 µl ) . Coverslips were washed three times with TBS and later they were blocked for 1 h at 4°C with 2% BSA in TBS . Finally , coverslips were washed three times and were incubated at 37°C with 100 µl of the bacterial suspension . After 1 h incubation , the coverslips were washed three times with PBS and then bacteria fixed with 3 . 7% paraformaldehyde ( PFA ) in PBS pH 7 . 4 for 20 min at room temperature . PFA fixed cells were incubated with PBS containing 0 . 1% saponin , 10% horse serum and Hoechst 33342 ( 1∶25000 ) for 30 min in a wet dark chamber . Finally , coverslips were washed twice in 0 . 1% saponin in PBS , once in PBS and once in H2O , mounted on Aqua Poly/Mount ( Polysciences ) and analysed with a Leica CTR6000 fluorescence microscope . Bacteria were counted in images from three randomly selected fields of view obtained at a magnification of ×100 taken with a Leica DFC350FX camera . Wild-type adhesion was set to 100% . Carcinomic human alveolar basal epithelial cells ( A549 , ATTC CCL-185 ) were maintained in RPMI 1640 tissue culture medium supplemented with 1% HEPES , 10% heat inactivated foetal calf serum ( FCS ) and antibiotics ( penicillin and streptomycin ) in 25 cm2 tissue culture flasks at 37°C in a humidified 5% CO2 atmosphere as previously described [79] . For infections , A549 cells were seeded on 12 mm circular coverslips in 24-well tissue culture plates to 70% confluence . Cells were serum starved 16 h before infection . Overnight cultures of Y . enterocolitica strains grown at 21°C were diluted 1∶10 into 5 ml of LB and grown with aeration at 21°C for 1 . 5 h and then 1 h at 37°C . Bacteria were pelleted , washed once with PBS and resuspended to an OD600 = 1 ( approximately 109 CFU/ml ) in PBS . Cells were infected with this suspension to get a multiplicity of infection of 25∶1 . After 1 h incubation , the coverslips were washed three times with PBS and then cells fixed with 3 . 7% PFA in PBS pH 7 . 4 for 20 min at room temperature . PFA fixed cells were incubated with PBS containing 0 . 1% saponin , 10% horse serum , Hoechst 33342 ( 1∶2500 ) , and OregonGreen 514-phalloidin ( 1∶100 ) ( Invitrogen ) for 30 min in a wet dark chamber . Finally , coverslips were washed twice in 0 . 1% saponin in PBS , once in PBS and once in H2O , mounted on Aqua Poly/Mount ( Polysciences ) and analysed with a Leica CTR6000 fluorescence microscope . Images were taken with a Leica DFC350FX camera . YopE translocation into A549 cells was done as previously described [38] . Briefly , A549 cells were seeded in 12-well tissue culture plates to 80% confluence . Cells were serum starved 16 h before infection . Overnight cultures of Y . enterocolitica strains grown at 21°C were diluted 1∶10 into 5 ml of LB and grown with aeration at 21°C for 1 . 5 h and then 1 h at 37°C . Bacteria were pelleted , washed once with PBS and resuspended to an OD600 = 1 ( approximately 109 CFU/ml ) in PBS . Cells were infected with this suspension to get a multiplicity of infection of 25∶1 . To synchronize infection , plates were centrifuged at 200×g during 5 min . After 1 h infection , cells were washed twice with PBS and resuspended in 400 µl of PBS with the help of a rubber policeman . Cell suspensions were transferred to a 1 . 5 ml microcentrifuge tube and cells pelleted ( 16 000×g; 12 sec ) . Supernatant was carefully removed and cells were resuspended in 100 µl of 1% digitonin ( w/v ) in PBS supplemented with a cocktail of protease inhibitors ( Halt protease inhibitor single-use cocktail EDTA-free; Thermo ) . After 2-min incubation at RT , samples were centrifuged ( 16 000×g; 10 min , 4°C ) . 80 µl of the supernatant , containing cytosolic proteins , were collected to whom 20 µl of 5× SDS sample buffer were added . The pellet , containing intact bacteria and cell membranes , was resuspended in 100 µl 1× SDS sample buffer . Aliquots corresponding to approximately 6×104 infected A549 cells were analysed by SDS-polyacrylamide gel electrophoresis and Western blotting using rabbit polyclonal antiserum raised against YopE ( 1∶2000 dilution ) . Strains were grown aerobically for 16 h at RT , pelleted and resuspended to an OD540 of 0 . 3 in PBS . Bacteria suspensions were added to subconfluent HeLa cells at a multiplicity of infection of ∼25∶1 . After a 30 min infection , monolayers were washed twice with PBS and then incubated for an additional 90 min in medium containing gentamicin ( 100 µg/ml ) to kill extracellular bacteria . This treatment was long enough to kill all extracellular bacteria . After this period , cells were washed three times with PBS and lysed with 0 . 5% saponin in PBS and bacteria were plated . Experiments were carried out in triplicate on three independent occasions . Invasion is expressed as CFUs per monolayer . Bacteria were grown either at 21°C or 37°C in 5 ml LB in a 15-ml Falcon tube with shaking ( 180 rpm ) , and harvested ( 2500×g , 20 min , 24°C ) in the exponential growth phase ( OD540 0 . 8 ) . Bacteria were washed once with PBS and a suspension containing approximately 1×105 CFU/ml was prepared in 10 mM PBS ( pH 6 . 5 ) , 1% Tryptone Soya Broth ( TSB; Oxoid ) , and 100 mM NaCl . Aliquots ( 5 µl ) of this suspension were mixed in 1 . 5 ml microcentrifuge tubes with various concentrations of polymyxin B ( Sigma ) . In all cases the final volume was 30 µl . After 1 h incubation at the bacterial growth temperature , the contents of the tubes were plated on LB agar . Colony counts were determined and results were expressed as percentages of the colony count of bacteria not exposed to antibacterial agents . All experiments were done with duplicate samples on at least four independent occasions . Murine macrophages RAW264 . 7 ( ATCC , TIB71 ) were grown on DMEM tissue culture medium supplemented with 10% heat-inactivated foetal calf serum ( FCS ) and Hepes 10 mM at 37°C in an humidified 5% CO2 atmosphere . For bacterial infection , cells were seeded in 24-well tissue culture plates 15 h before the experiment at a density of 7×105 cells per well . Overnight cultures of Y . enterocolitica strains grown at 21°C were diluted 1∶10 into 5 ml of LB and grown with aeration at 37°C or 21°C for 3 h . Bacteria were pelleted , washed once with PBS and resuspended to an OD600 = 1 ( approximately 109 CFU/ml ) in PBS . Cells were infected with this suspension to get a multiplicity of infection of 25∶1 . To synchronize infection , plates were centrifuged at 200×g during 5 min . After a 30 min infection , cells were washed twice with PBS and then incubated for an additional 180 min in medium containing gentamicin ( 100 µg/ml ) . Supernatants were removed from the wells , cell debris removed by centrifugation , and samples were frozen at −80°C . TNFα present in supernatants of culture cells was determined by ELISA ( Bender MedSystems ) with a sensitivity <4 pg/ml . The results were analyzed by the one-sample t test using GraphPad Prism software ( GraphPad Software Inc . ) . Results are given as means ± SD . A P value of <0 . 05 was considered to be statistically significant .
Lipopolysaccharide ( LPS ) is one of the major surface components of Gram-negative bacteria . The LPS contains a molecular pattern recognized by the innate immune system . Not surprisingly , the modification of the LPS pattern is a virulence strategy of several pathogens to evade the innate immune system . Yersinia enterocolitica causes food-borne infections in animals and humans ( yersiniosis ) . Temperature regulates most , if not all , virulence factors of yersiniae including the structure of the LPS lipid A . At 21°C , lipid A is mainly hexa-acylated and may be modified with aminoarabinose or palmitate . In contrast , at 37°C , Y . enterocolitica expresses a unique tetra-acylated lipid A . In this work , we establish that Y . enterocolitica encodes a lipid A deacylase , LpxR , responsible for the lipid A structure expressed by the pathogen at 37°C , the host temperature . Our findings also revealed that the low inflammatory response associated to Y . enterocolitica infections is the sum of the anti-inflammatory action exerted by a Yersinia protein translocated into the cytosol of macrophages and the reduced activation of the LPS receptor complex due to the expression of a LpxR-dependent deacylated LPS .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "bacteriology", "gram", "negative", "biology", "microbiology", "host-pathogen", "interaction", "bacterial", "biochemistry", "bacterial", "pathogens" ]
2012
Deciphering the Acylation Pattern of Yersinia enterocolitica Lipid A
Little is known about the origin and long-term evolutionary mode of retroviruses . Retroviruses can integrate into their hosts' genomes , providing a molecular fossil record for studying their deep history . Here we report the discovery of an endogenous foamy virus-like element , which we designate ‘coelacanth endogenous foamy-like virus’ ( CoeEFV ) , within the genome of the coelacanth ( Latimeria chalumnae ) . Phylogenetic analyses place CoeEFV basal to all known foamy viruses , strongly suggesting an ancient ocean origin of this major retroviral lineage , which had previously been known to infect only land mammals . The discovery of CoeEFV reveals the presence of foamy-like viruses in species outside the Mammalia . We show that foamy-like viruses have likely codiverged with their vertebrate hosts for more than 407 million years and underwent an evolutionary transition from water to land with their vertebrate hosts . These findings suggest an ancient marine origin of retroviruses and have important implications in understanding foamy virus biology . Foamy viruses are complex retroviruses thought exclusively to infect mammalian species , including cats , cows , horses , and non-human primates [1] . Although human-specific foamy viruses have not been found , humans can be naturally infected by foamy viruses of non-human primate origin [2]–[4] . Comparing the phylogenies of simian foamy viruses ( SFVs ) and Old World primates suggests they co-speciated with each other for more than 30 million years [5] . Retroviruses can invade their hosts' genomes in the form of endogenous retroviral elements ( ERVs ) , providing ‘molecular fossils’ for studying the deep history of retroviruses and the long-term arms races between retroviruses and their hosts [6] , [7] . Although ERVs are common components of vertebrate genomes ( for example , ERVs constitute around 8% of the human genome ) [8] , germline invasion by foamy virus seems to be very rare [9] , [10] . To date , endogenous foamy virus-like elements have been discovered only within the genomes of sloths ( SloEFV ) [9] and the aye-aye ( PSFVaye ) [10] . The discovery of SloEFV extended the co-evolutionary history between foamy viruses and their mammal hosts at least to the origin of placental mammals [9] . However , the ultimate origin of foamy virus and other retroviruses remains elusive . The continual increase in eukaryotic genome-scale sequence data is facilitating the discovery of additional ERVs , providing important insights into the origin and long-term evolution of this important lineage of viruses . In this study , we report the discovery and analysis of an endogenous foamy virus-like element in the genome of the coelacanth ( Latimeria chalumnae ) , which we designate ‘coelacanth endogenous foamy-like virus’ ( CoeEFV ) . The discovery CoeEFV offers unique insights into the origin and evolution of foamy viruses and the retroviruses as a whole . We screened all available animal whole genome shotgun ( WGS ) sequences using the tBLASTn algorithm using the protein sequences of representative foamy viruses ( Table S1 ) and identified several foamy virus-like insertions ( Table S2 and Fig . S1 ) within the genome of L . chalumnae , one of only two surviving species of an ancient Devonian lineage of lobe-finned fishes that branched off near the root of all tetrapods [11]–[15] . There are numerous in-frame stop codons and frame-shift mutations present in these CoeEFV elements , suggesting that the CoeEFV elements might be functionally defective . Although more than 230 vertebrate genome scale sequences are currently available , endogenous foamy virus elements have been only found in the aye-aye , sloths , and coelacanth , indicating that germline invasion of foamy virus is a rare process [9] , [10] . We extracted all contigs containing significant matches and reconstructed a consensus CoeEFV genomic sequence ( Fig . S2 ) . The resulting consensus genome shows recognizable and typical foamy virus characteristics ( Fig . 1 ) . Its genome has long terminal repeat ( LTR ) sequences at both 5′ and 3′ ends and encodes the three main open reading frames ( ORFs ) , gag , pol , and env , in positions similar to those of exogenous foamy viruses ( Fig . 1 ) . Two additional putative ORFs were found at positions similar to known foamy virus accessory genes but exhibit no significant similarity ( Fig . 1 ) . Notably , we found that the Env protein is conserved among foamy viruses and the coelacanth virus-like element ( Fig . 2 ) . A Conserved Domain search [16] identified a conserved foamy virus envelope protein domain ( pfam03408 ) spanning most ( 887 of 1016 residues ) of the CoeEFV Env protein , with an E-value of 1 . 3×10−69 ( Fig . 2 ) . The CoeEFV Env protein shares no detectable similarity with other ( non-foamy virus ) retroviral Env proteins or with retroviral elements within available genomic sequences of other fishes , such as the zebrafish ( Danio rerio ) . Hence , it provides decisive evidence that CoeEFV originated from a foamy-like virus . To exclude the possibility that these CoeEFV elements result from laboratory contamination , we obtained a tissue sample of L . chalumnae and succeeded in amplifying CoeEFV insertions within the genome of L . chalumnae via PCR with degenerate primers designed for conserved regions of foamy virus pol and env genes . To establish the position of CoeEFV on the retrovirus phylogeny , conserved regions of the Pol protein sequences of CoeEFV and various representative endogenous and exogenous retroviruses were used to reconstruct a phylogenetic tree with a Bayesian approach . The phylogenetic tree shows that CoeEFV groups with the foamy viruses with strong support ( posterior probability = 1 . 00; Figs . 3 and S3 ) , confirming that CoeEFV is indeed an endogenous form of a close relative of extant foamy viruses . The discovery of CoeEFV establishes that a distinct lineage of exogenous foamy-like viruses existed ( and may still exist ) in species outside the Mammalia . Endogenous retroviruses are likely to undergo a gradual accumulation of neutral mutations with host genome replication after endogenization [17] . To date the invasion of CoeEFV into coelacanth genome , we identified two sets of sequences , each of which arose by segmental duplication because each set of sequences shares nearly identical flanking regions ( Fig . S4 ) . The two sets contain five and two sequences , respectively . Because the divergence time of the two extant coelacanth species ( L . chalumnae and L . menadoensis ) is uncertain [11] , it is impossible to obtain a reliable neutral evolutionary rate of coelacanth species . Nevertheless , even using the mammalian neutral evolutionary rate [18] as a proxy for the coelacanth rate , the invasion dates were conservatively estimated at 19 . 3 ( 95% highest posterior density [HPD]: 15 . 3–23 . 6 ) million years ago for the dataset of five sequences . For the dataset containing two sequences , the divergence between the pair is estimated to be 4 . 1% and the invasion time is estimated to be approximately 9 . 3 million years ago . Because the CoeEFV invasion almost certainly occurred earlier than the duplication events within the host genome and because the evolutionary rate of coelacanth species is thought to be lower than other vertebrate species [19] , [20] , the time of CoeEFV integration might much more than 19 million years . Additional phylogenetic evidence ( see below ) suggests that its exogenous progenitors likely infected coelacanths for hundreds of millions of years prior to the event that fossilized CoeEFV within its host's genome . To further evaluate the relationship of foamy viruses , we reconstructed phylogenetic trees based on the conserved region of Pol proteins of foamy viruses and Class III retroviruses , the conserved region of foamy virus Pol and Env protein concatenated alignment , and the conserved region of foamy virus Env protein alignment , respectively . The three phylogenies have the same topology in terms of foamy viruses ( Figs . 4 , S5 , and S6 ) . CoeEFV was positioned basal to the known foamy viruses ( Fig . 4 ) , suggesting a remarkably ancient ocean origin of foamy-like viruses: the most parsimonious explanation of this phylogenetic pattern is that foamy viruses infecting land mammals originated ultimately from a prehistoric virus circulating in lobe-finned fishes . The branching order of the three foamy virus phylogenies ( Fig . 4 , S5 , and S6 ) is completely congruent with the known relationships of their hosts , and each node on the three virus trees is supported by a posterior probability of 1 . 0 ( except the node leading to equine , bovine , and feline foamy viruses on the Env phylogeny , which is supported by a posterior probability of 0 . 94; Fig . S6 ) . The common ancestor of coelacanths and tetrapods must have existed prior to the earliest known coelacanth fossil , which is 407–409 million years old [21] . The completely congruent virus topology , therefore , strongly indicates that an ancestral foamy-like virus infected this ancient animal . Crucially , the foamy viral branch lengths of the three phylogenies are highly significantly correlated with host divergence times ( R2 = 0 . 7115 , p = 1 . 10×10−5 , Fig . 5; R2 = 0 . 7024 , p = 1 . 41×10−5 , Fig . S5; and R2 = 0 . 7429 , p = 4 . 26×10−6 , Fig . S6 ) , a pattern that can reasonably be expected only if the viruses and hosts codiverged . It is worth emphasizing that we used a consensus sequence to represent CoeEFV in these analyses , so its branch length should correspond roughly to that of the exogenous virus that integrated >19 million years ago , rather than within-host mutations since that time . There are two alternative explanations for these phylogenetic patterns . One is that the exogenous progenitor of CoeEFV is not truly the sister taxon to the mammalian foamy viruses , but a more distant relative . The robust posterior probability ( 1 . 00 ) placing them in the same clade and the absence of evidence for viruses or virus-like elements from other species disrupting this clade argue against this view , as does the significant similarity between the Env proteins of CoeEFV and the foamy viruses ( Fig . 2 ) . Moreover , its branch length would be difficult to explain under such a scenario . If the coelacanth foamy-like virus lineage and the mammalian foamy virus lineage did not share a most recent common ancestor in their ancestral host , why is CoeEFV neither more nor less divergent from the mammalian foamy viruses than one might expect if they did ? The other alternative to the hypothesis that these viruses have co-diverged over more than 407 million years is that they somehow moved , in more recent times , from terrestrial hosts to sarcopterygian hosts that inhabited the deep sea , and that the similarity of the coelacanth virus to the mammalian viruses is due to cross-species ( in fact cross-class ) transmission , rather than shared history . However , as illustrated by the significant correlation between host divergence times and viral distances ( Figs . 5 , S5 , and S6 ) , the long branches leading to CoeEFV and the clade of mammal foamy viruses suggest the virus had already circulated in vertebrates for an extremely long time before the origin of mammal foamy virus . Given that there is strong evidence that placental mammals were already being infected with foamy viruses by about 100 million years ago [9] , the distinctness of the coelacanth virus suggests that it would have to have crossed from some other unidentified host , one whose foamy-like virus was already hundreds of millions of years divergent from the mammalian viruses . This seems highly unlikely . Although cross-species transmission of SFVs has been observed [2]–[5] , [22] , foamy viruses seem to mainly follow a pattern of co-diversification with their hosts [5] , [9] . If one accepts that the endogenous foamy viruses within the genomes sloths indicate more than 100 million years of host-virus co-divergence , it seems plausible that CoeEFV extends that timeline by an additional 300 million years . Moreover , the habitat isolation of the coelacanth and terrestrial vertebrates would have provided limited opportunities for direct transfer of foamy viruses to coelacanths . Taken together , these lines of evidence strongly suggest that foamy viruses and their vertebrate hosts have codiverged for more than 407 million years , and that foamy viruses underwent a remarkable evolutionary transition from water to land simultaneously with the conquest of land by their vertebrate hosts . Our analyses provide compelling evidence for the existence of retroviruses going back at least to the Early Devonian . This is the oldest estimate , to our knowledge , for any group of viruses , significantly older than the previous estimates for hepadnaviruses ( 19 million years ) [23] and large dsDNA viruses of insects ( 310 million years ) [24] . Although highly cytopathic in tissue culture , foamy viruses do not seem to cause any recognizable disease in their natural hosts [1] , [25] , [26] . Such long-term virus-host coevolution may help explain the low pathogenicity of foamy viruses . The fact that the Env is well conserved between CoeEFV and foamy viruses is consistent with the fact that these viruses are asymptomatic and mainly co-evolve with their hosts in a relatively conflict-free relationship . It is easy to imagine that previously overlooked examples of such a non-pathogenic virus may yet be found in hosts that fill in some of the gaps in the phylogeny , namely amphibians , reptiles , and birds . It will be of interest to screen these hosts , but also various fish species , for evidence of exogenous and/or endogenous foamy-like viruses . Dating analyses provide the clearest evidence for when and where retroviruses originated . There is strong evidence that foamy viruses shared a common , exogenous retroviral ancestor more than 400 million years ago ( since Env was present in both terrestrial and marine lineages ) . The discovery of endogenous lentiviruses demonstrates that lentiviruses , a distinct retroviral lineage that includes HIV , are also millions of years old [27]–[30] . Foamy viruses and lentiviruses share a distantly related ancestor ( Figs . 3 , S3 ) and the foamy virus clade alone almost certainly accounts for more than 407 million years of retroviral evolution . It follows that the origin of at least some retroviruses is older than 407 million years ago . As with the coelacanth lineage in the foamy virus clade , we found that retroviruses of fishes occupy the most basal positions within both the Class I and Class III retroviral clades ( walleye dermal sarcoma virus ( WSDV ) and snakehead retrovirus ( SnRV ) , respectively , blue asterisks ) , ( Figs . 3 , S3 ) . This pattern provides additional evidence of a marine origin and long-term coevolution of these major retroviral lineages . However , to be specific , the phylogenetic reconstruction in Fig . 3 reflects the history of only of the Pol protein , not a comprehensive history of retroviral genomic evolution . Nevertheless , our analyses support a very ancient marine origin of retroviruses . All available animal whole genome shotgun ( WGS ) sequences from GenBank were screened for endogenous foamy viruses using the tBLASTn algorithm and the protein sequences of representative exogenous and endogenous foamy viruses ( Table S1 ) . Sequences highly similar to foamy virus proteins discovered within the coelacanth WGS were aligned to generate a CoeEFV consensus genome . Conserved domains were identified using CD-Search service [16] . Ethanol preserved Latimeria chalumnae tissue sample was obtained from Ambrose Monell Cryo Collection ( AMCC ) at the American Museum of Natural History , New York . Genomic DNA was extracted using the DNeasy tissue kit ( QIAGEN , MD ) following the manufacturer's instructions . Amplification of ∼680 bp gag gene and ∼650 bp env gene fragments was performed with the degenerate primer pairs , FVpol-F ( 5′-AACAGTGYCTYGACCMAACC-3′ ) and FVpol-R ( 5′-TAGTGAGCGCTGCTTTGAGA-3′ ) , FVenv-F ( 5′-CTGGGGATGACAAYCAGAGT-3′ ) and FVenv-R ( 5′-CCACTCRGGAGAGAGGCAAC-3′ ) . PCR was performed in 25 µl of final volume reactions with 0 . 1 µl Platinum Taq HiFi enzyme ( Invitrogen , CA ) , 1 µl primer mix ( 10 µM each ) , 0 . 5 µl of 10 mM dNTP mixture , 1 µl of 50 mM MgSO4 , 2 . 5 µl of 10× PCR buffer , and 1 µl of template DNA . The PCR reactions were cycled under the following conditions: initial denaturation at 94°C for 2 minutes , 45 cycles of ( 94°C for 15 seconds , 60°C for 60 seconds , and 72°C for 30 seconds ) , and final elongation at 72°C for 5 minutes . The PCR products were purified using QIAquick spin columns ( QIAGEN , MD ) . Purified PCR products were cloned into the pGEM-T Easy vector ( Promega , WI ) . Cloned products were sequenced by the University of Arizona Genetics Core with an Applied Biosystems 3730XL DNA Analyzer . The sequences have been deposited in GenBank ( Accession Nos . JX006240-JX006251 ) . All protein sequences were aligned using Clustal Omega [31] . Gblocks 0 . 91b was used to eliminate ambiguous regions and extract conserved regions from the alignments [32] . To determine the phylogenetic relationship between CoeEFV and other retrovirus , we reconstructed a phylogeny based on the conserved region of Pol proteins of CoeEFV and various representative exogenous and endogenous retroviruses ( Table S1; Dataset S1 ) . To further evaluate the relationship and divergence of foamy viruses , the conserved region of the foamy viruses and Class III endogenous retroviruses Pol protein ( Dataset S2 ) , the conserved region of foamy virus Pol and Env protein concatenated alignment ( Dataset S3 ) , and the conserved region of foamy virus Env protein alignment ( Dataset S4 ) were used to infer phylogenetic trees . We were unable to discern positional homology for the first 143 residues of the Pol protein with reasonable certainty . These regions were excluded from all subsequent analyses . All the phylogenetic analyses were performed with MrBayes 3 . 1 . 2 [33] using 1 , 000 , 000 generations in four chains , sampling posterior trees every 100 generations . The rtREV amino acid substitution model [34] was used . The first 25% of the posterior trees were discarded . MCMC convergence was indicated by an effective sample size >300 as calculated in the program Tracer v1 . 5 . For the phylogenetic tree based on the foamy viruses and Class III endogenous retroviruses Pol protein , Class III endogenous retroviruses were used to root the foamy viral phylogeny ( Fig . 4 ) . Because there is no obvious outgroup for foamy virus Env protein , we rooted the phylogenetic trees inferred from foamy virus Pol and Env concatenated alignment and Env alignment using midpoint method ( Figs . S5 and S6 ) . Because the topologies of the host and virus trees were identical for the foamy viruses ( Figs . 4 , S5 , and S6 ) , we were able to plot host branch length ( in millions of years ) versus virus branch length ( in expected amino acid substitutions per site ) for every branch ( both internal and external ) . The vertebrate host divergence times are based on references [21] , [35] , and [36] . The nucleotide sequences were aligned using MUSCLE [37] . To estimate the age of the CoeEFV invasion , we identified two sets of sequences , which contain five sequences ( contig270160 , contig184752 , contig185880 , contig245863 , and contig236769 ) ( Dataset S5 ) and two sequences ( contig243355 and contig219087 ) ( Dataset S6 ) . Sharing the same flanking region , each set of sequences arose from segmental duplication . I ) For the dataset of five sequences: the best-fitting model of nucleotide substitution was determined using jModelTest [38] . The typical mammal neutral evolutionary rate ( 2 . 2×10−9 substitutions per site per year , standard deviation = 0 . 1×10−9 ) was used as the rate prior [18] . The HKY substitution model was used . BEAST v1 . 6 . 1 ( http://beast . bio . ed . ac . uk ) was employed for Bayesian MCMC analysis with a strict clock model [39] and Yule model of speciation . MCMC chains were run for 100 million steps twice to achieve adequate mixing for all parameters ( effective sample size >200 ) . Tracer v1 . 5 was used to summarize and analyze the resulting posterior sample . II ) For the dataset of two sequences: we calibrated the genetic distance between the pair based on the Kimura two-parameter model , in which transitions and transversions are treated separately .
The deep history of retroviruses is still obscure . Retroviruses can leave integrated copies within their hosts' genomes , providing a fossil record for studying their long-term evolution . Endogenous forms of foamy viruses , complex retroviruses known to infect only mammalian species , appear to be extremely rare , so far found only in sloths and the aye-aye . Here , we report the discovery of endogenous foamy virus-like insertions within the genome of a so-called ‘living fossil’ , the coelacanth ( Latimeria chalumnae ) . We provide evidence suggesting that foamy viruses and their hosts share a coevolutionary history of more than 407 million years , and that foamy viruses accompanied their vertebrate hosts on the evolutionary transition from water to land . These findings indicate that the retroviruses originated in the primeval ocean millions of years ago .
[ "Abstract", "Introduction", "Results/Discussion", "Materials", "and", "Methods" ]
[ "genome", "evolution", "biology", "genomics", "evolutionary", "biology", "genetics", "and", "genomics" ]
2012
An Endogenous Foamy-like Viral Element in the Coelacanth Genome
In evolution strategies aimed at isolating molecules with new functions , screening for the desired phenotype is generally performed in vitro or in bacteria . When the final goal of the strategy is the modification of the human cell , the mutants selected with these preliminary screenings may fail to confer the desired phenotype , due to the complex networks that regulate gene expression in higher eukaryotes . We developed a system where , by mimicking successive infection cycles with HIV-1 derived vectors containing the gene target of the evolution in their genome , libraries of gene mutants are generated in the human cell , where they can be directly screened . As a proof of concept we created a library of mutants of the human deoxycytidine kinase ( dCK ) gene , involved in the activation of nucleoside analogues used in cancer treatment , with the aim of isolating a variant sensitizing cancer cells to the chemotherapy compound Gemcitabine , to be used in gene therapy for anti-cancer approaches or as a poorly immunogenic negative selection marker for cell transplantation approaches . We describe the isolation of a dCK mutant , G12 , inducing a 300-fold sensitization to Gemcitabine in cells originally resistant to the prodrug ( Messa 10K ) , an effect 60 times stronger than the one induced by the wt enzyme . The phenotype is observed in different tumour cell lines irrespective of the insertion site of the transgene and is due to a change in specificity of the mutated kinase in favour of the nucleoside analogue . The mutations characterizing G12 are distant from the active site of the enzyme and are unpredictable on a rational basis , fully validating the pragmatic approach followed . Besides the potential interest of the G12 dCK variant for therapeutic purposes , the methodology developed is of interest for a large panel of applications in biotechnology and basic research . Broadening the repertoire of natural molecules and generating variants that confer new phenotypes to human cells are appealing perspectives for the development of biomedical applications , and for understanding fundamental cellular processes . To this end , in classical procedures , libraries of mutants are generated in vitro by degenerated PCR or DNA shuffling , and then screened on biochemical bases or with genetic tests in bacteria [1] . However , when the modification of human cells is sought , the mutants identified in these preliminary screenings often do not confer the desired phenotype due to differences in protein folding , post-translational modifications , and to the complex epistatic network that regulates the expression of the phenotype in the cells of higher eukaryotes . Alternatively , the library can be cloned in eukaryotic expression vectors for the screening step , albeit with the drawback of a considerable loss of complexity . The generation and screening of libraries of mutants directly in human cells would constitute an ideal solution to circumvent these problems . Nature provides organisms that are perfectly exploitable for this purpose: retroviruses . Indeed , after entry into the target cell , the viral polymerase ( reverse transcriptase , RT ) converts the viral genomic RNA , through an error-prone process that generates genetic diversity , into double-stranded DNA , which is then permanently integrated in the genome of the cell . The human immunodeficiency virus ( HIV-1 ) is the retrovirus with the strongest mutation rate , and constitutes therefore the ideal candidate for developing such approaches . During the process of Reverse Transcription , point mutations are introduced in the proviral DNA with a rate of 1–3 . 4×10−5 nt/cycle [2]–[4] and genetic diversity is further amplified by recombination [5] , [6] . We report here the development of a new methodology ( called “Retrovolution” ) aimed at generating and screening libraries of cellular genes directly in human cells . In Retrovolution the error-prone replication machinery of HIV-1 is diverted to drive the evolution of cellular genes: by performing successive infection cycles of cell cultures with HIV-1-derived viral vectors containing the sequence target of the evolution inserted in their RNA , libraries of gene mutants are generated in human cells , where they can be directly screened for the desired phenotype . A system for the evolution of non-viral sequences based on the use of engineered HIV-1 had been previously developed for the optimization of the Tetracycline-regulated expression system [7] . In that system , the isolation of an improved phenotype was strictly linked to the efficiency of replication of the viral particle , whereas in Retrovolution selection is not linked to viral replication and the use of replication-defective , non-cytopathic vectors allows the isolation and expansion of cells with the desired phenotype , which is not possible when replication-competent HIV particles are used . While relatively simple proteins or sequences , as a modified tag or reporter gene , have traditionally constituted the target for setting up evolution procedures [8] , [9] , to evaluate the power of the Retrovolution system we aimed at generating and isolating variants of a protein modifying a complex cellular phenotype , which could have a potential therapeutic interest . Namely , we wished to mutagenize the human deoxycytidine kinase ( hdCK , NM_000788 . 2 mRNA NCBI accession number ) gene with the goal of isolating an enzyme variant able to increase the sensitivity of cells to a widely used category of anticancer compounds , deoxycytidine ( dC ) analogues . The hdCK , physiologically involved in the phosphorylation of deoxycytidine , deoxyadenosine and deoxyguanosine in the salvage pathway for the deoxyribonucleotide biosynthesis , is a pivotal enzyme in the activation of deoxycytidine analogues used in chemotherapy ( AraC and Gemcitabine , used in the treatment of leukemias and solid tumours ) . These compounds , administered as nucleosides , must be transformed by cellular enzymes in their triphosphate form to become active and induce apoptosis in the cell by interfering with DNA replication and repair [10] . The first phosphorylation step , catalyzed by the human deoxycytidine kinase , determines the overall efficiency of the process [11] . Accordingly , there is a direct correlation between the level of dCK expression and the sensitivity to deoxycytidine analogues in different types of tumor [12] , [13] . The appearance of resistance forms among tumor cells , often due to a loss of dCK activity , and the toxicity of high concentrations of nucleoside analogues for non-tumor cells , constitute the major limits of the treatments with these compounds [14] , [15] . The sensitization of cancer cells to low doses of drugs by insertion of a gene coding for a drug-metabolizing enzyme constitutes an appealing approach for gene therapy applications [16]–[18] . A dCK variant that could preferentially phosphorylate nucleoside analogues with respect to the natural substrate deoxycytidine , and therefore induce cell suicide upon exposure to low concentrations of prodrug , would constitute a good candidate for these “suicide gene” therapies in cancer treatment , but also in transplantation medicine . In this case , indeed , it would provide a human ( and therefore poorly immunogenic ) gene to use as negative selection marker to insert in transplanted cells to counteract their eventual uncontrolled proliferation in vivo . So far , attempts at constructing a dCK mutant with improved ability in phosphorylating deoxycytidine analogues have been based on the information obtained from the 3D structure of the protein [19] , [20]; these dCK variants based on the rational design displayed an overall increased catalytic activity , resulting in a highly improved affinity for the natural substrate and a less improved affinity for the prodrug , which is not predictive of the induction of a sensitization phenotype in vivo . By evolving the dCK gene with Retrovolution procedure we could directly screen for dCK variants inducing cell death in presence of doses of the deoxycytidine analogue Gemcitabine lower than the ones needed to kill cells bearing a wt dCK , and this allowed us to isolate a mutant consistently sensitizing cancer cells to the prodrug . The Retrovolution system , outlined in Figure 1 , is based on the insertion of the gene target of the evolution process in the genomic RNA ( gRNA ) of conditional replication-defective , HIV-1 derived vectors pseudotyped with the Vesicular Stomatitis Virus envelope ( VSV-G ) ( Figure 1A ) . Repeated cycles of transduction of producer cells with these vectors , as indicated in Figure 1B , mimic successive infection cycles by HIV-1 , during which mutations will be inserted in the gRNA and re-shuffled by recombination [21] , [22] , an event favored when transductions are performed at a high Multiplicity Of Infection ( MOI ) . During the procedure , HIV-1 Reverse Transcriptase will introduce mutations all along the viral gRNA and , while those falling in functional regions of the vector will be selected ( positively or negatively ) , the others , including those falling in the target gene , will accumulate in an unbiased fashion . This process generates a library of mutants of the target gene directly inserted into the vector tasked with delivering the transgene to the human cell , and requires minimal intervention from the experimenter . During the evolution procedure , cells that did not receive the viral vector and therefore do not contain an integrated proviral DNA are removed by puromycin treatment ( the puromycin resistance gene being inserted as a selectable marker in the gRNA of the vectors , see Figure 1A ) . As an alternative , GFP can be used as a selectable marker instead of puroR and successfully transduced cells can be sorted by FACS at every round: the use of GFP would fasten the evolution procedure ( a complete puromycin selection takes three to four days ) , but would require a much more important manual intervention . The average number of positions that will be mutated in the target gene will depend on the number of “infection cycles” performed , on the size of the target gene itself , and on the mutation rate of the reverse transcription process , as outlined in Figure 2A . Once reached the ideal complexity , the library of mutants will be screened by transducing target cells ( the human cells to which the intended phenotype is to be conferred ) at a low MOI ( ≪1 ) to minimize the possibility that one cell will receive more than one vector , followed by clonal analysis ( Figure 1B ) . As the vectors are pseudotyped with the VSV envelope , which can infect virtually any kind of mammalian cells , the library can be screened in all human cell types . A library of dCK variants was generated by using the dCK coding sequence ( dCK-cs ) as the target gene ( Figure 1A ) of the Retrovolution procedure . By sequencing part of the library after 9 cycles of transduction we could calculate an ongoing mutation rate of 1×10−4 nt/infectious cycle . Based on this , the library was screened upon reaching generation 16 , at which a complexity of 1 . 5 mutated positions per dCK-cs ( 783 bp ) was expected ( Figure 2A ) . A further sequencing of a subset of the library at this generation confirmed the presence of 1 . 4 mutated positions per dCK gene ( Table 1 ) . In the dCK variants sequenced all kinds of mutations were present with a bias toward G>A transitions ( 62 . 3% ) , consistent with the mutational pattern of HIV-1 RT [23] . While the sequencing of the dCK gene at different generations of the Retrovolution system confirmed that mutations were inserted randomly along the gene ( Figure 2B ) , the sequencing of the viral backbone of the corresponding clones revealed a consensus of 12 mutations in the U3 region of each vector ( Figure 2C ) . These consensus mutations appeared in the promoter region of HIV-1 LTR since the F8 generation and were fixed in the population ever since . Moreover , their nature diverged from the typical HIV-1 mutation pattern since the percentage of G>A transitions was consistently lower ( 33% ) . No mutations were instead found in the other cis-acting regions essential for genome packaging , reverse transcription , integration and nuclear export . Overall , this strongly suggests that the viral sequences underwent selection for an optimized transcription of the genomic RNA from the U3 promoter in the specific context of the producer cells , while preserving the functionality of the vector , and allowing the production of a library of randomly mutated variants of the target gene . To screen the library for the presence of dCK variants that confer an increased sensitivity to low concentrations of the deoxycytidine analogue Gemcitabine , viral vectors from generation 16 were used to transduce HEK-293T cells at a low MOI ( MOI = 0 . 03 ) to ensure that each cell did not contain more than one variant of the transgene , and single clones were isolated by limiting dilution and puromycin selection . We then measured the viability of 76 individual clones in the presence of increasing concentrations of Gemcitabine ( 10 , 35 and 70 nM ) and calculated the cell death rate as the number of alive cells at concentration X of Gemcitabine divided by the number of cells at 0 Gemcitabine . The experiment was repeated three times using , as controls , HEK-293T cells either untransduced or transduced with a wt-dCK containing vector . The two higher Gemcitabine concentrations tested resulted to be too strong for an effective screening since all samples , including controls , displayed more than 80% of cell death ( data not shown ) . At 10 nM Gemcitabine , instead , 6 of the 76 clones tested yielded , in at least one of the three independent experiments , a cell death rate higher than both untransduced HEK-293T cells and cells bearing the wt dCK transgene ( Figure 3A ) . These six clones were selected for further analysis . The 6 clones isolated in the preliminary screening were further characterized for the ability of their transgenes to induce sensitization to Gemcitabine in the Messa10K cell line , uterine sarcoma cells that express an inactive dCK and are , therefore , highly resistant to this drug [13] . These cells provide an ideal background for testing the ability of the library of dCK mutants to increase sensitivity to Gemcitabine , since the phenotype generated by the introduction of a single copy of a dCK variant will not be masked by the activity of the wt dCK protein present in the cells . For each HEK-293T clone selected , viral particles were rescued by transfection with the plasmids encoding HIV-1 Gag-Pol and VSV Env , and used to transduce Messa10K cells at a MOI<1 . Since the effect induced by a transgene on the cell phenotype can be strongly influenced by the position of integration of the proviral DNA within the target cell genome , transduction of the Messa10K cells was performed following a procedure , outlined in Figure S1 , that generates , for each transgene , a population of cells containing the same transgenic sequence inserted at different genomic locations ( “polyclonal population” ) . For each dCK variant analysed we established 4 to 9 independent Messa 10K populations with this procedure , that allows to average the possible effects of the insertion site and to highlight the phenotype induced by the transgenic sequence itself . From the analysis of cell death ratios of Messa 10K populations in presence of increasing concentrations of Gemcitabine appeared that one of the 6 transgenes identified on HEK-293T cells , G12 ( E171K , E247K , L249M ) , strongly sensitized cells to the drug . In the 9 G12-Messa10K polyclonal populations tested , indeed , 50% of the cells died at 75 nM Gemcitabine , while at the same concentration only 10% of Messa 10K containing a wt dCK were killed ( Figure 3B ) . The analysis of a broader range of Gemcitabine concentrations revealed that in G12-Messa 10K cells the Gemcitabine IC50 was reduced by 60-fold compared to cells transduced with wt-dCK , and by 300-fold relative to untreated cells [13] ( Figure 3C ) . As shown by Western Blot ( Figure 4A ) , sensitization of Messa10K cells by G12 is not the consequence , as could result from mutations in the EF1-alpha promoter ( Figure 1A ) , of a higher level of protein expression compared to wt-dCK . Nevertheless , sequencing of the entire G12 gRNA revealed the presence of mutations within as well as outside the dCK-cs and these mutations could influence the phenotype observed . To rule out this possibility we inserted the G12 dCK-cs in a wild-type gRNA backbone plasmid ( generating “G12/wt-backbone” gRNA ) , and tested the deriving viral vectors on Messa10K cells . Also in this case a sensitization to the nucleoside analogue ( Figure 4B ) was detected , indicating that the Gemcitabine sensitization phenotype is rather due to properties of the dCK variant encoded by G12 than to mutations arisen in other portions of the genomic RNA . Messa10K cells constitute a model of Gemcitabine-resistant cells lacking an active endogenous dCK , as frequently emerge during chemotherapy [15] . To investigate the effect of the insertion of G12 in tumour cells that express an endogenous functional dCK protein , we assessed the phenotype conferred by the insertion of the mutated dCK to the colon carcinoma cell line HT29 , and to the pancreatic cancer cell line BxPC3 . In HT29 , sensitivity to Gemcitabine is strictly linked to the levels of dCK activity , whereas in BxPC3 the sensitivity to the prodrug depends on a different mechanism , the expression levels of the cytidine deaminase CDA [24] , [25] , an enzyme that inactivates the fully phosphorylated Gemcitabine by deamination . Therefore , while in the former cells the insertion of an improved variant of dCK should lead to a significant sensitization to the drug , in the latter ones the effect should be reduced . We generated two independent G12 polyclonal populations for each cell line and tested them . Consistently with our hypothesis , we observed a considerably stronger effect of G12 in HT29 than in BxPC3 ( Figure 4C and 4D , respectively ) , with a decrease in the Gemcitabine IC50 with respect to cells containing a wt-dCK from 1 . 6 µM to 80 nM in HT29 and only of less than two folds ( from 40 to 22 nM ) in BxPC3 . These results support the idea that the Gemcitabine sensitization phenotype is due to an improved ability of the G12 mutated kinase to activate the prodrug in cell culture and underscores that the insertion of a single copy of the transgene can increase the sensitivity of cells to Gemcitabine even in the presence of a wt dCK activity . To further support the existence of a link between the effect of the mutant in cell culture and its enzymatic activity , we expressed in E . coli the G12 mutant and the wt type dCK proteins , purified them and characterized their efficiency of phosphorylation of Gemcitabine and of dC in vitro . While Gemcitabine was phosphorylated with comparable efficiencies by the wt and G12 dCKs ( left panel of Figure 5A ) , the natural substrate dC was phosphorylated at almost undetectable levels by G12 , with a dramatic drop with respect to the efficiency of phosphorylation observed with the wt dCK enzyme ( right panel of Figure 5A ) . The G12 dCK mutant therefore displays an altered substrate specificity , constituting a kinase that specifically phosphorylates Gemcitabine , even in the presence of the natural substrate of wt dCK . These observations underscore that , in vivo , an increased sensitivity to the prodrug can be efficiently achieved through a decreased ability of the kinase to phosphorylate its natural substrate , rather than through an improved efficiency of phosphorylation of the prodrug itself . So far , efforts at improving the efficiency of phosphorylation of nucleoside analogues by dCK have relied on the rational design of mutants either of the active site region [19] or of Ser74 , the phosphorylation of which has been described to modulate the dCK enzymatic activity in vivo [20] , [26] . Although the resulting enzymes displayed an increased overall catalytic activity in vitro , the relative efficiency of phosphorylation ( expressed as the ratio Kcat/KmGemcitabine/Kcat/KmdC ) of the drug with respect to the natural substrate was decreased compared to the wt kinase , a trend opposed to that observed for the G12 mutant we describe ( Figure 5B ) . Consistently with the view that these mutants should not impact sensitivity of cells to Gemcitabine treatment , when we constructed a vector carrying the sequence of the triple mutant A100V/R104A/D133A [19] and used it to transduce Messa10K cells , no altered sensitivity to Gemcitabine was observed ( Figure 5C ) . We describe here the development of a new system of mutagenesis of cellular genes aimed at identifying genetic variants of interest for applications in basic and applied research . A crucial feature of the system is the possibility of performing , easily and in a controlled manner , a straightforward screening of the library in the human cell . This pragmatic approach allows to overcome the generation of false positives variants , a frequent problem encountered upon in vitro screening of libraries . The application of the Retrovolution system led to the identification of a dCK mutant that fulfils the long-sought feature of increasing sensitivity of tumour cells to Gemcitabine treatment , based on the presence of mutations unpredictable on a rational basis . The screening step was performed following a protocol that ensures that the isolated mutant sensitizes the cells independently from the integration site of the provirus , therefore constituting a good candidate as “suicide gene” in gene therapy . The mutagenic potential of Retrovolution relies on the error-prone nature of the replication machinery of HIV-1 , and on the frequent occurrence of recombination , that reshuffles the pre-existing mutations contributing to increase the diversity of the library . We show here that these two sources of genetic diversification prompt enough diversity as to lead to the identification of mutants of interest . With a relatively reduced number of cycles the system allowed the creation of a library in which each variant of the target gene was characterized by 1 . 5 mutated positions on average . A higher proportion of mutations would have increased the chances of producing a high percentage of non-functional proteins . An alternative approach for generating a library of mutants of cellular genes contained in lentiviral vectors would have been constituted first by a mutagenesis step of the cellular gene through conventional methods as error-prone PCR , followed by the insertion of the library in the lentiviral gRNA plasmid . With respect to this approach , the method we describe , while slower in the generation of a complex library , provides the advantage of circumventing the unavoidable loss of complexity inherent to the step of cloning the library in the lentiviral gRNA plasmid . Another way to accelerate the acquisition of mutations would have been the use of error-prone reverse transcriptases or mild mutagens . These procedures were not privileged , though , due to the inherent drawback of reduced yield of vector production [27] , [28] . Depending on the nature of the target gene , the experimental settings of Retrovolution can be adapted to introduce selection ongoing during the mutagenesis steps . In the present work this was not possible since we targeted a gene for which a negative selection , as the induction of cell death , needed to be applied . The targeting of the dCK gene was aimed , despite the intrinsic difficulty of the screening procedure , at the isolation of variants of a gene with a relevant biological interest for biomedical applications . The properties of the mutant identified , indeed , constitute the second major point emerging from this work . Improving the efficiency of the currently clinically employed anticancer drugs and overcoming resistances arising during treatments is a major goal of cancer research . To improve the efficiency of the treatments with nucleoside analogues like Gemcitabine , gene therapy approaches aimed at sensitizing cancer cells through the introduction of a transgene that improves the efficacy of drug activation have been proposed . The hdCK is the best candidate transgene for this purpose as the kinase is responsible for the limiting step in intracellular activation of deoxycytidine analogues currently used in clinical treatments . The competition between the natural substrate of dCK and the prodrug inside the cell constitutes an obstacle to the efficient activation of the prodrug itself . The rational design of mutants in the catalytic site of the kinase attained the goal of improving phosphorylation of the prodrug in the past , but the mutants isolated displayed a concomitant , and stronger , increase in the efficiency of phosphorylation of the natural substrate [19] , [20] . As a result , phosphorylation of the prodrug is expected to remain disfavoured with respect to the natural substrate and these mutants fail , as we confirmed ( Figure 5C ) , to confer a drug sensitization phenotype to the cell . The G12 mutant , on the contrary , has an increased specificity for the phosphorylation of Gemcitabine due to the fact that it completely lost the ability to phosphorylate its natural substrate . This results in a reduction of the competition between the two substrates in vivo and in a sensitization of the cells to the prodrug . The biochemical mechanism through which an increased sensitization to the drug is obtained is therefore opposed to what generally tried to obtain by engineering the dCK on a rational basis . The screening for the isolation of the G12 mutant was performed on Messa 10K cells , where the most striking sensitization was observed . These cells constitute a model of tumoral cells acquiring resistance to deoxycytidine analogues due to a loss of dCK activity , a fundamental problem that has to be bypassed for an improvement of cancer treatment . Nevertheless , the G12 mutant has an effect also on tumour cells presenting an endogenous dCK activity , like HT29 and HEK-293T . Therefore , besides a potential application in cancer treatment , G12 also constitutes a promising suicide gene to use as negative selection marker in cells used in transplantation medicine . Suicide genes are inserted in cells transplanted for therapeutical purposes to specifically ablate them in the event they would undergo uncontrolled proliferation in vivo . To this end exogenous genes are generally used , the most frequently employed being the Herpes Simplex Virus thymidine kinase gene in association with gancyclovir treatment . A problem of increasing relevance in clinical gene therapy , though , is constituted by the immune response raised by the patient against the protein encoded by the suicide gene [29] , [30] . A suicide gene of human origin , as the mutant described here , is expected to have a faint possibility of being highly immunogenic and therefore represents an ideal candidate for this application . Besides conferring an increased sensitivity to Gemcitabine through an unexpected mechanism , the G12 mutant is characterized by mutations localized in regions unpredictable on a rational basis . The human dCK is a globular , dimeric protein in which each monomer is formed by a core of 5 beta-sheets surrounded by 10 alpha-helices [19] ( Figure 6A ) . G12 dCK carries the mutations E171K , E247K , and L249M , the first two of which induce a charge change in conserved residues ( Figure 6B ) . Amino acids 247 and 249 are located in the “base-sensing loop” ( alpha-helix 10 ) , which influences the folding of the protein upon binding of ATP or UTP as phosphate donor [26] , [31] and , consequently , also affects substrate binding . Indeed , a variant containing only these mutations , which was identified during the screening of the library ( mutant E8 , Figure 3B ) , slightly increased Messa10K cells sensitivity to Gemcitabine treatment , although with a cell death ratio that did not significantly differ from that of the wt-dCK ( p>0 . 05 ) . The marked sensitization observed with G12 thus requires the presence of the additional mutation , E171K . Residue 171 is located at the base of alpha-helix 7 that , with alpha-helix 4 , is involved in the generation of the interface of the dCK dimer [19] , and its mutation in a residue of opposite charge could potentially influence the efficiency of formation of the active dimer itself . Gel filtration analyses , though , revealed that the extent of dimerization of G12 does not differ from that of the wt-dCK ( Figure S2 ) . It is therefore likely that , as observed for other proteins [32] , [33] , the mutation rather triggers long-distance changes in the quaternary arrangement of the protein , possibly involving the region of the active site . In conclusion , Retrovolution allowed the identification of a promising suicide gene to use in cancer treatment but also as negative selection marker in transplantation medicine by allowing isolating a variant of the dCK that would not have been predictable on a rational basis . The property of producing the library and allowing its screening directly in human cells , ensuring that each mutant emerging from the procedure will be relevant for modifying the phenotype of the cells , is central to these findings . Retrovolution thus opens new avenues for the modification of genes conferring complex phenotypes of interest for a broad field of applications in basic and applied research . With the exploitation of its combination of genetic flexibility and ability to deliver transgenes to human cells , the lifestyle of one of the most important pathogens of our recent history appears far from being fully exploited . HEK-293T cells were obtained from the American Type Culture Collection ( ATCC ) and grown in Dulbecco's Modified Eagle's Medium ( Gibco ) supplemented with 10% FBS and 100 U/ml pennicillin-100 mg/ml streptomycin . Messa10K cells were kindly provided by LP . Jordheim and grown in RPMI medium supplemented with 10% FBS and penicillin-streptomycin . HT29 cells were kindly given by JN . Freund and grown in DMEM+ 10% FBS . BxPC3 cells were kindly provided by M . Dufresne and grown in RPMI+ 10% FBS . All cells were incubated at 37°C+5% CO2 . For the creation of the first generation of viral vectors , 8 plates containing 5×106 HEK-293T cells were transfected using the calcium phosphate protocol with 10 µg of pCMVΔR8 . 91 plasmid [34] , 5 µg of pHCMV-G plasmid [34]–[36] and 10 µg of genomic plasmid sdy-dCK ( encoding the RNA outlined in Figure 1A ) . 48 h after transfection , the virus-containing supernatant was collected , filtered through a 0 . 45 µm filter and concentrated 40-fold in Vivaspin 20 columns ( MWCO 50 KDa , Sartorius Stedim Biotech ) . Transfections for the creation of subsequent generations were performed with 10 µg of pCMVΔR8 . 91 and 5 µg of pHCMV-G plasmid . Transductions during the evolution procedure were performed on 5 million HEK-293T cells with 1 ml of 40× viral vectors ( MOI>100 ) . Transductions for the isolation of single clones were performed with 1 ml of 1∶500 dilution of viral vectors ( MOI<1 ) on 3 . 5×106 HEK-293T cells; after transduction , cells were seeded at 100 cells/well in 96-well plates in DMEM+ puromycin . For the generation of polyclonal populations of target cells , transduction was performed on 1×106 cells with 1 ml of 1∶5 diluted viral vectors rescued from the isolated clones . Transduced cells were selected by adding puromycin 24 h after transduction using 0 . 6 µg/ml to HEK-293T cells , 0 . 5 µg/ml to Messa10K cells , 0 . 8 µg/ml to HT29 cells and 0 . 4 µg/ml to BxPC3 . Clones of the F8 and F16 generation of the dCK library that had been isolated by limiting dilution were expanded in 10 cm plates and pelleted . Cells were lysed with 250 µl of Cell Direct PCR ( VIAGEN ) . The dCK transgene was amplified from 1 µl of the lysate with oligonucleotides on the vector , EF1 ( 5′-gatgtcgtgtactggctccg-3′ ) and PGK ( 5-gatgtggaatgtgtgcgagg-3′ ) , flanking the transgene . Sequencing of the PCR fragment was performed by the GATC sequencing service . The curves for the calculation of the number of Retrovolution cycles needed to have one mutation per target gene ( Figure 2A ) were drawn based on the formula p/m×n where p = n°of mutations wanted per target gene , m = mutation rate/nt , n = size of the target gene in nt . Cells were seeded at 5000 cells per well in a 96-well plate and grown overnight at 37°C . Two rows were used for each population . 12 h after seeding , increasing concentrations of Gemcitabine ( 0–400 nM or 0–100 µM ) were added to each well and incubation was continued for an additional 72 h . Cell viability was measured by MTT test ( CellTiter 96 Non-Radioactive Cell Proliferation Assay , Promega ) and the number of living cells in each well was evaluated by measuring the OD at 570 nm . For each population , the fraction of living cells was calculated as OD570 concentration X Gem/OD570 concentration 0 Gem , to estimate the sensitivity of the population to the prodrug . Messa10K cells transduced with the different dCK variants were lysed in 1X RIPA buffer , and 6 , 12 , 24 µg of total protein ( evaluated by Bradford ) for each cell type were loaded on a 12% bis-tricine gel ( Invitrogen ) . After transfer on a PVDF membrane , the dCK proteins were analysed by western Blot with 1∶4000 dilution of a polyclonal anti-dCK antibody ( rabbit , Sigma-Aldrich ) and 1∶3000 anti-rabbit HRP ( BioRad ) conjugated secondary antibody and detected by autoradiography . To evaluate whether the average amounts of cell death occurring in the Messa10K populations containing different dCK variants were significantly different from the value shown by Messa10K wt-dCK , a two sample t-test was applied for the different concentrations of Gemcitabine . The wt-dCK and G12 sequences were cloned in the pET14b plasmid and expressed in E . coli cells BL21 DE3 pLysE . Protein expression was induced by adding 0 . 1 mM IPTG , and cells were collected after 4 h of growth at 37°C . His-tagged proteins were eluted with 250 mM Imidazole from His-Trap TM FF Columns ( GE HealthCare ) , the Histidine tag was removed using the S-Tag Thrombin Purification Kit ( Novagen ) , and dCK and G12 were further purified by gel filtration on S-200 Sephacryl columns ( GE Healthcare ) . Purified proteins were used for the dCK activity assay or stored at −80°C in 30% glycerol . The efficiency of phosphorylation of the natural substrate deoxycytidine and of the prodrug Gemcitabine were measured for purified wt-dCK and G12 in a NADH-based assay as previously described [37] . All reagents were purchased from Sigma ( France ) except Gemcitabine ( Lilly France SAS ) . Enzymes were assayed at RT at a concentration of 0 . 3 µM with Gemcitabine or 0 . 9 µM with dC . Gemcitabine was used at concentrations between 10 µM and 1 mM and dC at concentrations between 5 and 50 µM . ATP was 4 mM . All experiments were performed in triplicate .
We exploited the error-prone replication machinery of HIV-1 and its ability to stably introduce transgenes in human cells to develop a novel system , Retrovolution , to generate libraries of mutants of cellular genes . When libraries are screened to isolate variants that modify the phenotype of the human cell for biomedical applications or basic research , false positives often arise from the classical screening procedures performed in vitro or in bacteria . Retrovolution allows an easy screening of the libraries directly in the human cell , where they are generated . We describe the creation and screening of a library of the hdCK ( a human kinase activating several anticancer compounds ) gene , to identify variants increasing the sensitivity of cancer cells to treatment with low , poorly toxic doses of the anticancer drug Gemcitabine . We isolated a dCK variant inducing death in tumour cells at doses up to 300 times lower than those required for killing non-engineered cells . The mutant presents mutations unpredictable on structural basis and revealed a change in enzymatic properties that accounts for the observed cellular effect . Besides the intrinsic interest of the mutant identified , these results fully validate Retrovolution as a mutagenesis system with broad applications in applied and basic research .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "biotechnology", "mutagenesis", "biochemistry", "genetic", "mutation", "molecular", "cell", "biology", "mutation", "types", "nucleic", "acids", "proteins", "genetics", "biology", "genetics", "and", "genomics", "genetic", "engineering", "gene", "function" ]
2012
Retrovolution: HIV–Driven Evolution of Cellular Genes and Improvement of Anticancer Drug Activation
Much of the research on cannabinoids ( CBs ) has focused on their effects at the molecular and synaptic level . However , the effects of CBs on the dynamics of neural circuits remains poorly understood . This study aims to disentangle the effects of CBs on the functional dynamics of the hippocampal Schaffer collateral synapse by using data-driven nonparametric modeling . Multi-unit activity was recorded from rats doing an working memory task in control sessions and under the influence of exogenously administered tetrahydrocannabinol ( THC ) , the primary CB found in marijuana . It was found that THC left firing rate unaltered and only slightly reduced theta oscillations . Multivariate autoregressive models , estimated from spontaneous spiking activity , were then used to describe the dynamical transformation from CA3 to CA1 . They revealed that THC served to functionally isolate CA1 from CA3 by reducing feedforward excitation and theta information flow . The functional isolation was compensated by increased feedback excitation within CA1 , thus leading to unaltered firing rates . Finally , both of these effects were shown to be correlated with memory impairments in the working memory task . By elucidating the circuit mechanisms of CBs , these results help close the gap in knowledge between the cellular and behavioral effects of CBs . Recent years have seen a resurgence of interest in the therapeutic role of cannabinoids ( CBs ) for several diseases and neurophyschiatric disorders such as psychosis , anxiety disorders , PTSD , and multiple sclerosis [1 , 2] . In particular , CB agonists have shown promising but mixed results in the treatment of epilepsy , as various types of agonists at various doses have been shown to be both pro- and anticonvulsant [3–8] . Parallel to increasing therapeutic research , much work has been done on the chemical structure of various cannabinoids and cannabinoid receptors , along with their cellular interactions and pharmacology [9] . Nonetheless , between the large bodies of literature on cannabinoids from chemical , disease , and behavioral perspectives , much less work has been done to explore the effects of cannabinoids on the neural circuit level . This is particularly important since a wide range of complex and often opposing effects have been attributed to cannabinoids on a molecular and cellular level . For example , cannabinoid activation of CB1 receptors , which are found on both pyramidal cells and interneurons , reduces the quantity of neurotransmitter released during an action potential; consequently , increased extracellular cannabinoid levels reduce both excitatory ( glutamatergic ) and inhibitory ( GABAergic ) transmission [10] . Furthermore , cannabinoids have been shown to interact with astrocytes [11] , mitochondria [12] , glycine receptors [13] , vanilloid receptors [14] , potassium ion channels [15] , and to reduce GABA and glutamate reuptake [16 , 17] . Consequently , it is very difficult to extrapolate the emergent network level changes simply from a catalogue of effects cannabinoids have a cellular/molecular level . Here , we studied the effects of Δ9-tetrahydrocannabinol ( THC ) on hippocampal networks during memory encoding using spiking activity recorded in rodents in-vivo performing the Delayed-NonMatch-to-Sample ( DNMS ) working memory task . Multivariate autoregressive ( MVAR ) models were used in both control and THC sessions to estimate feedforward and feedback dynamical filters , which are akin to the waveform shapes of the CA3→CA1 EPSP and CA1 afterhyperpolarization , respectively [18] . MVAR models , which are a type of linear nonparametric model , are ‘data-driven’ in the sense that they estimate model parameters directly from recorded neural spiketrains and , unlike more biologically realistic models , make very few a priori assumptions on the nature of the neural dynamics [19 , 20] . This characteristic makes them particularly well suited for this study , since as previously mentioned the emergent effects of THC on neural circuits are highly complex and unclear . Overall our results suggest that cannabinoids impair memory encoding by functionally isolating CA1 from CA3 via reduced theta information flow and altered excitatory-inhibitory balance across the Schaffer collateral synapse . To evaluate the effects of exogenous cannabinoids on the hippocampal network 1 mg/kg THC was injected intraperitoneally into N = 6 rodents during certain sessions while they were performing a DNMS task ( S1 Fig ) . All data was previously used in a study on the effects of cannabinoids on hippocampal multifractality [21 , 22] . Briefly , in the sample phase , the rats were presented one of two levers . After a variable length delay , both levers were presented in the match phase and the rat had to choose the opposite lever to receive a reward . On the behavioral level , it was found that THC reduced rodent-performance on the DNMS task by about 12 . 2 ± . 6% ( Fig 1a , [23] ) . This corresponds to a 24 . 4% impairment relative to baseline performance at 50% . While performing the DNMS task , single-unit activity was recorded from the hippocampal CA3 and CA1 regions using a multi-electrode array . There were no significant mean firing rate ( MFR ) differences between THC sessions and control sessions in either CA3 or CA1 cells ( P = . 502 , Fig 1b ) . No MFR differences were seen whether considering the entire session or only times around the DNMS sample phase , or whether considering all cells or only sample-presentation cells ( see below ) . The lack of any cannabinoid-induced changes in firing rates at this dosage has been observed in previous studies [24 , 25] . Two types of temporal coding were identified in the recorded spiketrains . First , on slower timescales , several neurons fired preferentially in response to lever presentation in the sample phase of the DNMS task [26] . It was found that THC reduced the proportion of sample-presentation cells in both CA3 and CA1 by roughly equal amounts ( Δ = 13 ± 4% , P < . 001; Fig 1c ) . Interestingly , some sample-presentation cells lost all of their preferential firing in THC sessions ( Fig 1d ) ; this contrasts with place cells whose receptive field stays largely intact under cannabinoids [27] . There was an insignificant trend connecting sample-presentation cell reduction with behavioral deficits ( R2 = . 27 , P = . 052 , S3a Fig ) . On faster timescales , it was found that several CA3 and CA1 neurons had theta band rhythmicity ( 4–7 Hz ) . Hippocampal theta oscillations are known to be intimately related to cognitive function [28–30] and have previously been linked to performance in the DNMS task [31]; furthermore , theta oscillations are known to be reduced by systemic injections of cannabinoids on both the single unit [24] and network level [32] . It was found that CA1 theta power was slightly but significantly reduced in THC sessions ( Δ = 2 . 52% , CI: [ . 61 , 4 . 4]%P = . 004; Fig 1e ) . A similar , albeit slightly weaker , theta power reduction was seen in CA3 cells ( Δ = 1 . 94% , P = . 045; S2 Fig ) . Interestingly , in both cases , the significant reduction of theta power occurred at 5–6Hz , which is lower than the observed theta peak . Unlike previous results in a different task [24] , the reduction in CA1 theta power was not found to be correlated with behavioral deficits in the DNMS task ( P = . 674 , S3b Fig ) . Overall , these results show that THC has minor effects on the actual neuronal spiketimes: quantity of spikes ( MFR ) was not affected and spike rhythmicity ( theta oscillations ) were only slightly affected . Furthermore behavioral deficits induced by cannabinoids could not be explained by any of these factors , which are the traditional markers of rate and temporal coding in the hippocampus . The remainder of the study will focus on systems analysis of the Schaffer collateral synapse connecting CA3 to CA1 , and how this synapse is affected by THC . Systems analysis aims to identify the input-output “blackbox” by which the input spiketrains are transformed into the output spiketrain . On a more abstract level , it aims to identify how the information encoded in CA3 is propagated into CA1 . This is distinct from the signal analysis done in the previous section which only looks at features of individual spiketrains rather than the causal relationship between multiple spiketrains as done in systems analysis . The relationship between an arbitrary number of input CA3 spiketrains and the output CA1 spiketrain was modeled using a multivariate autoregressive model described by Eq 1 and an example of which is pictured in Fig 2a . Each system consists of N input CA3 neurons and N feedforward filters describing the dynamical input-output relationship between the given CA3 and CA1 neurons ( Fig 2b ) . Intuitively , these filters can be thought of as the EPSP elicited in the output CA1 neuron in response to an action potential ( AP ) in the input CA3 neuron . However , unlike EPSPs which traditionally only encapsulate ion-conductances from neurotransmitter-gated ion channels , the “blackbox” nature of the feedforward filters means they also include more complex dynamical effects such as dendritic integration , spike generation , active membrane conductances , and feedforward interneuronal inhibition ( thereby allowing the filters between two pyramidal cells to be inhibitory ) . Each model also includes a feedback ( autoregressive ) filter which describes the effects of past output spikes onto the output present . This filter , which can be intuitively thought of as the afterhyperpotential ( AHP ) [33] includes intracellular processes such as the absolute and relative refractory periods , slow potassium conductances , and Ih conductances . It also includes more complex intercellular processes such as the recurrent connections between CA1 pyramidal cells and interneurons [34] . Neuronal connectivity was estimated using a stepwise input selection procedure . Filter parameters were estimated with Laguerre basis regression using neuronal activity around the sample phase . Model significance was verified using ROC plots and shuffling methods ( see supplementary methods ) . A representative connectivity grid from a recorded THC session with 10 recorded neurons ( 4 CA3 , 6 CA1 ) is shown in Fig 2a . Fig 2b shows a sample system from this session between 3 CA3 pyramidal cells and 1 CA1 pyramidal cell . Note that two of the feedforward filters are excitatory ( above the x-axis ) while the third has both excitatory and inhibitory components , presumably arising through feedforward inhibition involving interneurons [35 , 36] . The system also involves a feedback filter which shows a relatively long refractory period ( ∼40ms ) followed oscillatory bursting activity . Oscillations in the CA1 pyramidal cell AHP are a well known phenomena caused by slow K+ and Ih conductances , and these oscillations are known to lead to theta resonances [18 , 37 , 38] . In order to study the filter oscillations more closely , the filter frequency spectra were plotted in Fig 2c . Both feedforward excitatory filters were found to have peaks in the high theta range ( 8–9 Hz ) . Intuitively , this can be understood to mean that information encoded in the theta range in these input neurons is preferentially transmitted to the output CA1 neuron . Furthermore , the feedback filter has a low theta resonance of 3 . 5 Hz . Significance metrics for the displayed system is shown in S4 Fig , and additional systems are shown in S5 Fig . All together 66% ( 707/1068 ) of all systems were found to be significant and 2139 feedforward and 707 feedback filters were obtained . THC was found to reduce the number of significant models per session ( Δ = −7 . 4% , P = . 011 ) , but the predictive power of significant models , as measured by AUC ( see supplementary methods ) , was unaltered ( P = . 24 ) . To study how THC affects system dynamics on a population level , we examined how features change in the entire sample of control and THC filters . The average filter frequency profile for both control and THC sessions is shown in Fig 3a and 3b ( top ) . Both feedforward and feedback spectra are found to have clear theta band peaks , thus generalizing the trend seen in the example system of Fig 2 . This is consistent previous reports which show that CA3 propagates strong theta rhythms to CA1 [39 , 40] and also that CA1 is capable to generating endogenous theta rhythms [41] . THC produced a significant decline in the theta power of the feedback filters ( Δ = 20 . 8% , P < . 001; Fig 3b ) . Note that the feedback filter theta reduction is about 10x stronger than the theta reduction found in the CA1 spiketrain signals ( Fig 1e ) . No reduction in theta power was found in the feedforward filters ( P = . 61 , Fig 3a ) . This result suggests that cannabinoid-induced theta desynchronization results primarily from altered feedback properties rather than changes in CA3→CA1 dynamics . Cannabinoids have been reported to affect network excitation-inhibition balance ( EIB ) [10 , 42] . Particularly , there is much debate whether cannabinoids are pro- or anticonvulsants [4 , 6 , 8 , 43 , 44] . In order to examine the effects cannabinoids have on network EIB , we quantified the excitation of the estimated filters using a metric called the excitatory index ( EI ) , which is the ratio between positive filter area and total filter area . It was found that THC had no significant effect on feedforward EI ( P = . 14 ) ; however , there was an insignificant trend showing that THC-induced decreases in feedforward EI were correlated with behavioral deficits ( R2 = . 27 , P = . 063 , Fig 3c ) . Additionally , THC reduced the number of casually connected CA3-CA1 neuronal pairs ( Δ = −8 . 9% , P < . 001 ) . These findings , together with the THC-induced decrease of CA3→CA1 significant models , suggest that THC reduces the causal influence CA3 neurons have on CA1 spiketimes . In other words , THC can be said to functionally isolate CA1 from CA3 . It was also found that THC significantly increased feedback EI ( Δ = 3 . 5% , P = . 022 ) and that the increased feedback EI was correlated with behavioral deficits ( R2 = . 38 , P = . 007 , Fig 3d ) . The large quantity ( >2800 ) and variability of the obtained filters describing the CA3→CA1 dynamic transformation presents a challenge of interpretation . Namely , how could one identify features from the entire filter population which are representative of the CA3→CA1 transformation rather than just the input-output relationship found in this or that particular pair of neurons . In essence this is an unsupervised learning problem which aims to identify hidden structure within the filter population for the purpose of knowledge discovery . Our group has developed the concept of the global principal dynamic modes ( gPDMs ) towards this effort [19 , 45 , 46] . The gPDMs are a system-specific and efficient basis set which contain the essential dynamic components of the filter population and are meant to be amenable to biological interpretation . One set of gPDMs were estimated from all ( control and THC ) obtained filters with the hypothesis that THC would primarily change the expression strength of the gPDMs rather than their specific shapes . Fig 4a and 4b shows the obtained feedforward and feedback gPDMs in both the time and frequency domain . Once again , the feedforward and feedback gPDMs represent the dominant independent componenets of feedforward and feedback kernels , respectively . The first feedforward gPDM was found to have almost all its energy in the 1st time bin , with an immediate decline thereafter . This gPDM represents near concurrent firing between CA3 and CA1 neurons and presumably results from both direct CA3→CA1 connections via the Schaffer collateral synapse [47 , 48] and common inputs from the entorhinal cortex [49 , 50] . The third feedforward gPDM , which is characterized by an initial inhibitory phase , presumably represents feedforward interneuronal inhibition which is prevalent in the CA3→CA1 connection [35 , 36] . THC was not found to influence the strength of either of these gPDMs ( P = . 76 , P = . 60; S6 Fig ) . The second feedforward gPDM which is characterized by sustained and oscillatory excitation was found to have a strong theta peak in the frequency domain . Furthermore , it was found that THC-induced declines in the strength of this gPDM were correlated with behavioral deficits ( R2 = . 30 , P = . 032; Fig 4c ) . The three obtained feedback gPDMs are shown in Fig 4b . These gPDMs express the essential feedback dynamics found in CA1 neurons . As previously mentioned , these dynamics arise through the combination of intracellular processes such as the AHP and extracellular processes such as recurrent connections between CA1 pyramidal cells and interneurons . It was found that THC-induced increases in the third feedback gPDM were correlated with behavioral deficits ( R2 = . 39 , P = . 005; Fig 4d ) . This correlation was not seen in either of the first two feedback gPDMs ( P = . 32 , P = . 75; S6 Fig ) . Notably , the 3rd feedback gPDM was seen to be “theta-blocking” in the frequency domain due to its trough at 8 Hz . This gPDM counteracts the 1st “theta-promoting” feedback gPDM and disrupts theta oscillations in the CA1 neuron . The THC-induced changes in the feedforward and feedback theta gPDMs paint a more complete picture of the CA1 theta reductions seen in Fig 1e . Namely , they attribute the theta losses to specific feedforward and feedback dynamical filters which may potentially be traced to specific biophysical mechanisms . Furthermore , changes in these dynamical filters have been specifically correlated with behavioral deficits , which could not be done with theta reductions in the CA1 signal ( S3 Fig ) . The current study uses ‘data-driven’ nonparametric system dynamics modeling tools to study the effects of THC on the Schaffer Collateral synapse in rodents . The chief findings of the study can be summarized as: ( 1 ) THC induced little or no change in traditional rate and temporal coding metrics such as MFR and theta power , ( 2 ) THC altered the CA1 excitatory-inhibitory balance by reducing feedforward influence from CA3 while increasing feedback excitation from CA1 , ( 3 ) THC reduced theta information flow through the Schaffer collateral synapse , and ( 4 ) The magnitudes of both of the previous effects were directly correlated with the severity of behavioral deficits induced by THC . Overall these results suggest the conclusion that THC impairs memory encoding by functionally isolating CA1 from CA3 . From a computational perspective , the nonparamteric modeling methods used in this study proved successful in studying the network level effects of cannabinoids since , unlike biophysical models , all model parameters where estimated directly from recorded data and very few a priori assumptions were made about the effects of THC [19 , 20 , 51] . The global principal dynamic modes ( gPDMs ) , which were derived from MVAR filters of the entire population of neurons , further extracted hidden dynamical structure from ‘noisy’ neuron-neuron variability . Importantly , THC-induced changes in the gPDMs were directly correlated with behavioral impairments , thus justifying their utility . Furthermore , while most in-vivo studies on THC analyze macro level signals such as ECoG and EEG , this work adds to a relatively small body of literature which analyzes the effects of THC on neuronal population spiking activity . Finally , to our knowledge , this is the first work which examines the effect of THC on neuronal systems dynamics , or the causal interactions between signals , rather than on neuronal signals themselves . It was found that THC increased feedback excitatory index in CA1 and that the magnitude of this effect was correlated with behavioral deficits . We hypothesize that this is due to reduced feedback inhibition from CA1 cholecystokinin ( CCK ) -containing cells . While CCK cells only make up only 13 . 9% of interneurons [52] , they express significantly more CB1 receptors than any other cell in the hippocampus [53] , and their primary output is to CA1 pyramidal cells [52] . Increased THC concentrations would reduce CCK interneuron output by ( 1 ) reducing the amount of GABA they release per action potential ( 2 ) reducing their MFR due to reduced glutamatergic input from principal cells in both CA3 and CA1 [54 , 55] . It was also found that THC reduced the number of casually connected CA3-CA1 neuronal pairs; furthermore there was an interesting but insignificant trend for THC-induced deficits in feedforward excitation to lead to behavioral deficits . This trend may prove to be significant given a higher sample size . We hypothesize that this reduced feedforward influence is caused by decreased glutamate release from CA3 pyramidal cells due to CB1 receptor activation by THC [56] . Even though pyramidal cells have much lower densities of CB1 receptors than interneurons [53 , 57] , there is evidence that CB induced reduction of excitation is larger than these relative densities suggest . Principal cells outnumber interneurons 20:1 in CA1 [50] and their CB1 receptors were found to be several fold more efficacious than those of interneurons [58] . Further , lower baseline activation levels of CB1 receptors on principal cells than on interneurons suggest they would be disproportionately activated by CB agonists [59] . Altogether , the decreased feedback inhibition and feedforward excitation amount to a functional isolation , or breakdown in information flow between CA3 and CA1 . We suggest that this functional isolation is responsible for the behavioral impairments seen in the DNMS task . The ‘functional isolation’ hypothesis is further supported by previous work which showed that the behavioral impairments caused by cannabinoids in the DNMS task were similar to those seen with a full pharmacological lesion of the hippocampus [60] Given the centrality of CA3→CA1 information flow to hippocampal function , a functional isolation of these areas could indeed presumably lead to impairments similar to that of a full lesion . Relatedly , Goonawardena et al . , 2010 [25] injected THC intraperitoneally at low 1 mg/kg doses as in this study and in higher doses of 3 mg/kg . They found that while both doses disrupted hippocampal synchrony , only the higher dose resulted in a reduction in pyramidal cell MFR . This suggests that at the lower dose both previously described phenomena are at a net balance , while at the higher dose , the decrease in feedforward excitation overpowers the increase in feedback excitation and results in lower MFR . Finally , the hypothesis predicts a breakdown in the normal spiketime coordination between pyramidal cells and interneurons in CA1 circuits . The breakdown of this coordination , which has been extensively implicated in hippocampal oscillations [61 , 62] , could be responsible for the observed decrease in theta oscillations and information flow . Although the current results only suggest this hypothesis , several experiments could be done to further substantiate it . Feedforward and feedback kernels and gPDMs could be estimated at different doses of THC; the hypothesis would predict that different doses would effect the two processes independently , with one of the two processes potentially being more dominant at different THC levels . Significant developments in in-vivo synaptic patch clamping [63] and calcium imaging in recent years could be used to directly measure the drive of CCK cells and CA3 pyramidal cells onto CA1 pyramidal cells under THC . Much research has been done investigating the effects THC and other cannabinoids have on seizures and epilepsy . Results so far have been mixed , with various studies showing that THC is both pro- and anticonvulsant [3–8] . The results from this study and the presented hypothesis suggest that THC inherently is not pro- or anti-convulsant but that its effects will depend on the dosage and the unique circuitry of every epileptic focus . Interestingly , a study by Rudenko et al . , 2012 [6] has shown that indeed the effects of a CB1 agonist were dose dependant , with lower doses being anticonvulsant and higher doses being proconvulsant . Finally , this study suggests that in order to truly understand the effects of THC on epileptic circuits , one must study the systems level changes in circuit dynamics rather than taking a reductionist approach and studying the effects of THC on any particular receptor or cell type . The present study analyzed the effects of THC from both a signals and systems perspective—and found that systems analysis yielded much richer results . For example , while analysis of CA1 spiketrain signals showed a slight ( 2% ) reduction in theta frequency , analysis of system kernels showed that the theta loss was primarily due to CA1 feedback dynamics whose kernels lost over 20% of their theta power , while theta power in feedforward kernels was unaffected . Furthermore , only systems analysis allows one to analyze predictive power , feedforward and feedback excitation , and EPSP and AHP waveform shape . Notably , the finding that feedforward influence decreased while feedback excitation increased could not have been observed using only signal analysis which would have only detected a constant MFR . The present study also employed gPDMs as a means to extract the most significant information from the kernel dynamics estimated from several animals over several sessions [18 , 19 , 48 , 64] . The utility of the gPDM method was justified by the finding that reductions in theta related gPDMs in a given session were directly correlated with behavioral deficits , showing that the gPDMs can isolate the particular dynamics which are most affected by THC . Furthermore , THC-induced theta power losses in spiketrain signals were not found to be correlated with behavioral deficits . Although in the present study , kernels and gPDMs were restricted to being linear in order to more easily quantify their overall strength and excitation ( via the EI ) , future work will aim to identify the effects of THC on hippocampal nonlinear dynamics [51 , 65] . N = 6 Male Long-Evans rats were trained to criterion on a two lever , spatial Delayed NonMatch-to-Sample ( DNMS ) task ( see S1 Fig ) . Briefly , during the sample phase the rat was presented one of two levers ( left or right ) . After a delay phase ranging from 1–30 seconds , the rat was presented both levers and had to choose the opposite level in order to attain a reward . Each rodent underwent 16–25 sessions of the task , which were roughly evenly divided between control and THC sessions , wherein the rodent was intraperitoneally administered 1 mg/kg of body weight Δ9-tetrahydrocannabinol ( THC ) , an exogenous cannabinoid found in marijuana . During the task , spike trains were recorded in-vivo with multi-electrode arrays implanted in the left and right CA3 and CA1 regions of the hippocampus . In an effort to acquire a consistent cognitive state , only spiking activity around the sample phase of the task was used . Spikes from multiple trials were sorted , time-stamped , and concatenated into a discretized binary time series using a 4ms bin . For more details on the experimental setup , see supplementary methods . Nonparametric multiple-input linear autoregressive models were used to model the dynamical transformation between input and output spike trains ( see Figs 2 and 5 ) [18 , 51] . Each model consisted of a feedforward component , reflecting the effect of the N input cells on the output cell and a feedback ( autoregressive ) component reflecting the subthreshold and suprathreshold effects the output cell has on itself . Thus , the output y ( t ) is calculated as: y ( t ) = ∑ n = 1 N ∑ τ = 0 M k n ( τ ) x n ( t - τ ) + ∑ τ = 1 M + 1 k A R ( τ ) y ( t - τ ) ( 1 ) where kn reflects the feedforward filter of input xn ( t ) , and kAR reflects the feedback filter . In order to reduce the number of model parameters and thereby increase parameter stability , we applied the Laguerre expansion technique to expand the feedforward and feedback filters over L Laguerre basis functions ( see supplementary methods ) . Effective connectivity between neurons was assessed using a Granger causality-like approach . For each output CA1 neuron , input CA3 neurons were selected in a forward stepwise procedure whereby only neurons which help predict the output CA1 spike activity were included in the model . After all input neurons were selected , a Monte Carlo approach was used to assess model significance . A model was deemed significant if the CA3 inputs could predict the output CA1 activity significantly better ( P < . 0001 ) than randomly permuted versions of the inputs . See supplementary methods for more details . The global principal dynamic modes ( gPDMs ) were obtained in a two step process: first , all filters of each input from every animal were concatenated in a rectangular matrix . Then singular value decomposition ( SVD ) was performed on the rectangular matrix to obtain all the significant singular vectors , which are the gPDMs . It was found that 3 gPDMs were sufficient to describe the linear dynamics both the population of feedforward and feedback filters . gPDM strength in a given filter was computed by taking the dot product between the gPDM and the filter . gPDM strength in a given session was computed by taking the average gPDM strength in every filter of that session .
Research into cannabinoids ( CBs ) over the last several decades has found that they induce a large variety of oftentimes opposing effects on various neuronal receptors and processes . Due to this plethora of effects , disentangling how CBs influence neuronal circuits has proven challenging . This paper contributes to our understanding of the circuit level effects of CBs by using data driven modeling to examine how THC affects the input-output relationship in the Schaffer collateral synapse in the hippocampus . It was found that THC functionally isolated CA1 from CA3 by reducing feedforward excitation and theta information flow while simultaneously increasing feedback excitation within CA1 . By elucidating the circuit mechanisms of CBs , these results help close the gap in knowledge between the cellular and behavioral effects of CBs .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[ "medicine", "and", "health", "sciences", "action", "potentials", "nervous", "system", "drugs", "membrane", "potential", "electrophysiology", "neuroscience", "ganglion", "cells", "interneurons", "pharmacology", "animal", "management", "animal", "performance", "cannabinoids", "animal", "cells", "neural", "pathways", "agriculture", "cellular", "neuroscience", "neuroanatomy", "cell", "biology", "pyramidal", "cells", "anatomy", "synapses", "physiology", "neurons", "biology", "and", "life", "sciences", "cellular", "types", "neurophysiology" ]
2017
Cannabinoids disrupt memory encoding by functionally isolating hippocampal CA1 from CA3
Surface recognition and penetration are critical steps in the infection cycle of many plant pathogenic fungi . In Magnaporthe oryzae , cAMP signaling is involved in surface recognition and pathogenesis . Deletion of the MAC1 adenylate cyclase gene affected appressorium formation and plant infection . In this study , we used the affinity purification approach to identify proteins that are associated with Mac1 in vivo . One of the Mac1-interacting proteins is the adenylate cyclase-associated protein named Cap1 . CAP genes are well-conserved in phytopathogenic fungi but none of them have been functionally characterized . Deletion of CAP1 blocked the effects of a dominant RAS2 allele and resulted in defects in invasive growth and a reduced intracellular cAMP level . The Δcap1 mutant was defective in germ tube growth , appressorium formation , and formation of typical blast lesions . Cap1-GFP had an actin-like localization pattern , localizing to the apical regions in vegetative hyphae , at the periphery of developing appressoria , and in circular structures at the base of mature appressoria . Interestingly , Cap1 , similar to LifeAct , did not localize to the apical regions in invasive hyphae , suggesting that the apical actin cytoskeleton differs between vegetative and invasive hyphae . Domain deletion analysis indicated that the proline-rich region P2 but not the actin-binding domain ( AB ) of Cap1 was responsible for its subcellular localization . Nevertheless , the AB domain of Cap1 must be important for its function because CAP1ΔAB only partially rescued the Δcap1 mutant . Furthermore , exogenous cAMP induced the formation of appressorium-like structures in non-germinated conidia in CAP1ΔAB transformants . This novel observation suggested that AB domain deletion may result in overstimulation of appressorium formation by cAMP treatment . Overall , our results indicated that CAP1 is important for the activation of adenylate cyclase , appressorium morphogenesis , and plant infection in M . oryzae . CAP1 may also play a role in feedback inhibition of Ras2 signaling when Pmk1 is activated . Magnaporthe oryzae , a heterothallic ascomycete , is the causal agent of rice blast , which is one of the most destructive diseases of rice in the world . It infects the rice plant in a manner typical of many foliar pathogens . Germ tubes differentiate into specialized infection structures called appressoria . The fungus generates turgor pressure as high as 8 MPa in mature appressoria for plant penetration [1] . In M . oryzae , appressorium formation can be induced by attachment to hydrophobic surfaces that mimic the rice leaf surface . On hydrophilic surfaces , treatments with cAMP , cutin monomers , or primary alcohols induce the formation of melanized appressoria . To date , at least three putative sensor genes , PTH11 , CBP1 , and MSB2 , have been reported to be involved in the recognition of physical and chemical signals of plant leaves [2] , [3] , [4] . In M . oryzae , cyclic AMP ( cAMP ) is known to be the intracellular secondary messenger . Molecular studies have further confirmed the role of the cAMP-PKA pathway in surface recognition [5] , [6] , [7] . The CPKA gene encoding a catalytic subunit of PKA is required for normal appressorium formation and plant infection [8] , [9] . The Mac1 adenylate cyclase responsible for the synthesis of intracellular cAMP is required for appressorium formation and pathogenesis [5] , [6] . The surface attachment and recognition signals must be somehow relayed from the surface sensors to the activation of MAC1 , which may involve changes in the actin cytoskeleton because germ tube tip deformation is the initiation stage of appressorium formation . In mammalian cells , adenylyl cyclase AC8 orchestrates its downstream activity to yield a high output signaling by its association with actin cytoskeleton [10] . Although it is dispensable for surface recognition , the Pmk1 MAP kinase pathway is known to regulate late stages of appressorium formation , appressorial penetration , and invasive growth in M . oryzae [11] , [12] . In a number of plant pathogenic fungi , the orthologous MAPK cascade is conserved for the regulation of various infection or developmental processes [13] . In M . oryzae , the surface recognition signals must be conveyed from cAMP signaling to the Pmk1 MAPK pathway for appressorium formation and plant penetration , but the exact molecular mechanism is not clear . The adaptor protein of the Pmk1 MAPK cascade , Mst50 , is known to interact with both Mst11 and Mst7 . It also interacts with the two Ras proteins in M . oryzae , Ras1 and Ras2 , in yeast two-hybrids assays . Ras2 is likely involved in the activation of both cAMP-PKA and Pmk1 MAP kinase pathways [12] , [14] . Transformants expressing the dominant active allele of RAS2 form melanized appressoria on both hydrophilic and hydrophobic surfaces . In the budding yeast Saccharomyces cerevisiae , Ras2 regulates the activation of adenylyl cyclase ( AC ) . The yeast Cyr1 adenylyl cyclase is associated with a 70-kDa cyclase-associated protein ( CAP ) that was identified in a RAS-responsive adenylyl cyclase complex [15] . The CAP gene also was identified as SRV2 in a genetic screen for suppressors of a constitutive active RAS2G19V allele [16] , [17] . The yeast Srv2 protein has two distinct functional domains . Whereas the N-terminal adenylyl cyclase-binding ( ACB ) domain is sufficient for full cellular responsiveness to Ras2 and Cyr1 , the C-terminal actin-binding domain interacts with actin monomers . Loss of the C-terminal region of Srv2 caused morphological and nutritional defects that are not related to adenylyl cyclase or Ras2 activity . The C-terminal region of Srv2 plays a critical role in binding to G-actin in vivo and directing actin organization and polarized cell growth . However , recent studies suggested the N-terminus is equally important for Srv2 in driving actin turnover [18] , [19] , [20] . The CAP proteins are well conserved in eukaryotic organisms . Orthologs of Srv2 have been characterized in Schizosaccharomyces pombe , Candida albicans , Cryptococus neoformans , and Ustilago maydis [17] , [21] , [22] , [23] . In C . albicans , CAP1 is involved in initiation of the transition of yeast cells to hyphal growth . It also interacts with Ras and AC proteins to regulate the intracellular cAMP level . The cap1/cap1 mutant was reduced in virulence in a mouse model system and was defective in the yeast-hyphae transition , which can be stimulated by cAMP treatments [24] , [25] . In C . neoformans , Cap1 is a positive regulator of Cac1 and is required for the cAMP-mediated capsule formation and virulence [23] . The CAP1 ortholog in U . maydis , as an additional component of the cAMP/PKA signaling pathway , interacts with adenylate cyclase Uac1 and is important for morphogenesis and pathogenesis [26] . Although the cAMP signaling pathway has been shown to play important roles in various plant pathogenic fungi , to date no CAP genes have been functionally studied in filamentous ascomycetes . In this study , we identified and characterized the CAP1 ( for cyclase-associated protein 1 ) gene in M . oryzae . Cap1 interacts with Mac1 in yeast two-hybrid and co-immunoprecipitation ( co-IP ) assays . It was found to be important for the activation of Mac1 and invasive growth . The Δcap1 mutant was defective in appressorium formation , germ tube growth , and plant infection . Like LifeAct [27] , Cap1 localized to apical patches in vegetative hyphae but not in invasive hyphae , indicating that invasive hyphae may lack a typical actin cytoskeleton at the tip . Domain deletion analysis revealed that the AB domain of Cap1 was dispensable for its subcellular localization but plays a role in the proper regulation of appressorium formation and full virulence . Overall , our results indicate that CAP1 is important for the activation of adenylate cyclase , appressorium morphogenesis , and plant infection in M . oryzae . To identify genes interacting with MAC1 in M . oryzae , we generated the 3×FLAG knock-in construct and transformed it into the wild-type stain 70-15 ( Table 1 ) . In the resulting transformant MFG1 , a 242-kDa band of expected size of Mac1-3×FLAG fusion was detected by an anti-FLAG antibody ( Fig . S1A ) . For affinity purification , proteins bound to anti-FLAG M2 beads were eluted and digested with trypsin . Proteins that co-purified with Mac1-3×FLAG were identified by mass spectrometry ( MS ) analysis as described [28] . One of the Mac1-interacting genes ( Table 2 ) is MGG_01722 . 6 that encodes a protein highly similar to Srv2 in S . cerevisiae [29] . Srv2 is a cyclase-associated protein ( CAP ) involved in the Ras/cAMP pathway . Other proteins that co-immunoprecipitated with Mac1 included several components of the protein phosphatase PP2A and a number of hypothetical proteins unique to M . oryzae ( Table 2 ) . In mammalian cells , catalytic and regulatory subunits of PP2A also were co-immunoprecipitated with the N-terminal region of adenylyl cyclase 8 [30] . MGG_01722 . 6 was named CAP1 ( for cyclase- associated protein 1 ) and selected for further characterization in this study . It has an AC-binding ( ACB ) domain ( 2–166 aa ) , two proline-rich regions P1 ( 257–290 aa ) and P2 ( 355–377 aa ) , and a C-terminal actin-binding ( AB ) domain ( 378–534 aa ) ( Fig . 1A ) . In S . cerevisiae , Srv2 interacts with the C-terminal alpha helix region of Cyr1 adenylyl cyclase via its tandem repeats of the heptad motif aXXaXXX [31] , which are conserved in CAP1 and its orthologs in other fungi ( Fig . S1B ) . In M . oryzae , Cap1 interacted with the C-terminal region of Mac1 ( 1926–2160 aa , Mac1CT ) in yeast two-hybrid assays ( Fig . 1B ) , suggesting the direct association between Cap1 and Mac1 . To confirm their interaction , the MAC1CT-3×FLAG and CAP1-GFP fusion constructs were co-transformed into protoplasts of strain 70-15 . One of the resulting transformant was CMT8 ( Table 1 ) . In western blot analysis with total proteins isolated from transformant CMT8 , the anti-FLAG and anti-GFP antibodies detected a 30-kDa and a 85-kDa band , respectively . In proteins eluted from anti-FLAG M2 beads , the 85-kDa Cap1-GFP band was detected with an anti-GFP antibody in transformant CMT8 but not in transformant MCF12 ( Fig . 1C ) . Transformant MCF12 expressing the MAC1CT-3×FLAG construct only ( Table 1 ) was the negative control . These results indicate that Cap1 interacts with Mac1 in M . oryzae . To determine whether the N-terminal ACB domain is important for the interaction of Cap1 with Mac1 , the CAP1ΔACB-GFP and MAC1CT-3×FLAG constructs were co-transformed into the wild type strain 70-15 . In the resulting transformants , both Cap1ΔACB-GFP and Mac1CT-3×FLAG were expressed . In proteins eluted from anti-FLAG M2 beads , the 30-kDa Mac1CT-3×FLAG band was detected by an anti-FLAG antibody . However , we failed to detect the 67-kDa Cap1ΔACB-GFP band ( Fig . S2 ) , indicating that deletion of the ACB domain eliminated the interaction between Cap1 and Mac1 . These results indicate that the N-terminal ACB domain of Cap1 is essential for its interaction with Mac1 . To further characterize the Cap1-Mac1 interaction in vivo , the CAP1-NYFP and MAC1CT-CYFP fusion constructs were generated and transformed into protoplasts of the wild type strain 70-15 for bimolecular fluorescence complementation ( BiFC ) assays . In the resulting transformant CMB14 ( Table 1 ) , weak YFP signals were observed in the cytoplasm of vegetative hyphae and conidia . In appressoria , stronger YFP signals were observed in the cytoplasm and on cytoplasm membrane ( Fig . 2 ) . The presence of YFP signals on the cytoplasm membrane is consistent with the membrane association of adenylyl cyclase in mammalian cells , yeast , and other organisms [32] , [33] , [34] . These results confirmed that Cap1 interacts with Mac1 in vivo . The interaction between Cap1 and Mac1 may be weak during vegetative growth but enhanced during appressorium formation . The amino acid sequence of M . oryzae Cap1 shares 40% identity with yeast Srv2 . To test whether CAP1 is functional in S . cerevisiae , the CAP1 ORF was cloned into pYES2 and transformed into the srv2 mutant , which is hypersensitive to osmotic and oxidative stresses [35] . The resulting Trp+ yeast transformants containing the pYES2-CAP1 construct grew better than the original srv2 mutant on medium containing 5 mM H2O2 or 1 M NaCl ( Fig . 1D ) . In contrast , strains transformed with the empty pYES2 vector were as sensitive as the srv2 mutant to hyperosmotic or oxidative stresses ( Fig . 1D ) . Thus , when it was expressed in yeast , CAP1 could suppress the defects of the srv2 mutant in stress responses . To characterize its function in M . oryzae , the CAP1 gene replacement construct ( Fig . S3A ) was generated by ligation PCR [36] and transformed into protoplasts of Ku80 . Seven putative Δcap1 mutants were identified and confirmed by Southern blot analysis ( Fig . S3B ) . All seven mutants had the same phenotype , although only mutant HC83 ( Table 1 ) is described below for detailed analysis . The Δcap1 mutant was reduced in vegetative growth ( Fig . S3C ) and conidiation ( Table 3 ) . We also generated the cap1 deletion mutant in the Guy11 background . The resulting Δcap1 mutant HF12 ( Table 1 ) had similar defects as mutant HC83 . We then cloned the full length CAP1 gene into pYK11 as pXY63 and transformed it into the Δcap1 mutant HC83 . The resulting Δcap1/CAP1 transformant CH07 ( Table 1 ) was normal in growth and conidiation ( Table 3 ) . It also was normal in appressorium formation and plant infection ( see below ) , indicating that the reintroduction of the wild-type CAP1 allele fully complemented the defects of the Δcap1 mutant . On hydrophilic surfaces , like the wild type , the Δcap1 mutant failed to form appressoria . It still formed melanized appressoria on hydrophobic surfaces but the efficiency of appressorium formation per germ tube was significantly reduced ( Table 3 ) . In addition , germ tubes of the Δcap1 mutant had abnormal morphology and growth patterns . Besides producing branching germ tubes with uneven width ( Fig . 3A ) , 54 . 0±1 . 9% of the conidium compartments produced more than one germ tubes in the Δcap1 mutant . Subapical swollen bodies similar to those observed in the pmk1 mutant [37] were often formed by germ tubes of the Δcap1 mutant ( Fig . 3A ) . We also assayed appressorium formation with hyphal tips on hydrophobic surfaces . Whereas less than 10% of the hyphal tips in the Δcap1 mutant developed appressoria , majority of hyphal tips formed appressoria by 48 h in Ku80 ( Fig . 3B ) . Most of the Δcap1 hyphae that attached to hydrophobic surfaces formed irregular apical and intercalary swellings ( Fig . 3B ) , further indicating the importance of CAP1 in appressorium formation . Expressing the dominant active allele RAS2DA in the wild type resulted in the over-activation of the cAMP and Pmk1 pathways and formation of appressoria on both hydrophobic and hydrophilic surfaces [38] . Because SRV2 was first identified as a suppressor of the dominant active mutation of Ras , we introduced the RAS2DA allele into the Δcap1 mutant . The resulting Δcap1 RAS2DA transformant XY22 ( Table 1 ) failed to form appressoria on hydrophilic surfaces ( Fig . 4 ) . In contrast , the RAS2DA transformant of wild type Guy11 [14] , [38] formed appressoria on hydrophilic surfaces . On hydrophobic surfaces , the Δcap1 RAS2DA transformant , similar to the original Δcap1 mutant , formed melanized appressoria . To determine whether CAP1 is functionally related to the activation of MAC1 in M . oryzae , we assayed the intracellular cAMP content in vegetative hyphae . The Δcap1 mutant had a higher cAMP level than the Δmac1 mutant . However , its intracellular cAMP level was significantly reduced in comparison to that of strain Ku80 or Guy11 ( Fig . 5A ) . MAC1 is the sole adenylate cyclase gene in M . oryzae . Detection of cAMP in the Δmac1 mutant likely was from the background . These results indicate that Cap1 is not essential for Mac1 activation but plays an important role in the full activation of Mac1 in M . oryzae . Because the Δcap1 mutant had a reduced intracellular cAMP level , we assayed the effect of exogenous cAMP . On hydrophobic surfaces , cAMP treatment suppressed germ tube branching and growth defects of the Δcap1 mutant ( Fig . 5B ) . Melanized appressoria were formed on short germ tubes in the presence of 5 mM cAMP . On hydrophilic surfaces , 43 . 0±2 . 3% of the mutant germ tubes formed appressoria by 24 h in the presence of 5 mM cAMP . Under the same conditions , 80 . 1±3 . 7% of the wild-type germ tubes formed appressoria . In addition , the majority of appressoria induced by cAMP treatment in the Δcap1 mutant had abnormal morphology ( Fig . 5C ) . These results indicate that the Δcap1 mutant still responds to cAMP treatment . CAP1 must function upstream of cAMP by regulating the activation of Mac1 . However , the defects of the Δcap1 mutant in appressorium morphology suggest that CAP1 also plays a role in the proper regulation of responses to exogenous cAMP for appressorium morphogenesis . In spray infection assays with two-week-old rice seedlings of cultivar CO-39 , numerous blast lesions were observed on leaves sprayed with Ku80 or the complemented transformant at 7 dpi ( Fig . 6A ) . Under the same conditions , fewer and smaller lesions were observed on leaves sprayed with the Δcap1 mutant ( Fig . 6A ) . The number of lesions caused by the Δcap1 mutant was reduced about 6-fold in comparison with Ku80 ( Table 3 ) . More importantly , typical blast lesions with extensive necrotic zones were rarely or not observed on leaves inoculated with the Δcap1 mutant ( Fig . 6A ) . On wound-inoculated leaves , the Δcap1 mutant also caused fewer and smaller lesions outside the wound sites than Ku80 and the complemented transformant ( Fig . 6A ) . Even at the wounding sites , the Δcap1 mutant caused no or only limited necrosis . These results indicate that the CAP1 gene plays a critical role in lesion development and spread of blast infection . To further characterize the defects of the Δcap1 mutant in plant infection , we conducted penetration assays with rice leaf sheaths . Penetration by appressoria formed by the Δcap1 mutant was reduced over 4 folds . Even for those appressoria that successfully penetrated , invasive hyphae formed by the Δcap1 mutant inside plant cells were narrower than those of Ku80 . At 48 hpi , approximately 15% of the appressoria penetrated into rice leaf sheath cells and had limited growth of unbranched , non-bulbous invasive hyphae . Most of invasive hyphae were restricted to the penetrated cells . Under the same conditions , invasive hyphae formed by Ku80 grew extensively in the penetrated and neighboring cells ( Fig . 6B ) . Even at 72 hpi , invasive hyphae of the Δcap1 mutant were restricted to the initial penetrated cells at 41 . 2±1 . 6% of penetration sites examined ( Fig . 6B ) . At the sites where invasive hyphae had spread into neighboring plant cells ( 56 . 2±2 . 9% ) , the Δcap1 mutant still had limited growth and rarely branched ( Fig . 6B ) . These data indicate that the Δcap1 mutant was defective in invasive growth , which may be directly responsible for the reduced lesion size or limited necrotic zones . Failure to develop branching and bulbous invasive hyphae indicates that switching to pseudohyphal growth from primary invasive hyphae may be blocked or significantly attenuated in the Δcap1 mutant . A CAP1-GFP fusion construct was generated and transformed into the Δcap1 mutant . In the resulting transformant XY61 ( Table 1 ) , GFP signals were mainly in the apical regions of hyphae or germ tubes ( Fig . 7A ) . The localization pattern of Cap1-GFP was similar to that of actin patches in A . nidulans [39] . In developing appressoria , GFP signals were distributed in the peripherial regions ( Fig . 7A ) . However , in mature appressoria , Cap1-GFP fusion proteins were mainly localized in globular structures in the appressorium pore area ( Fig . 7A; Video . S1 ) . The transition of Cap1-GFP localization from the periphery to the base of the appressoria may be related to actin cytoskeleton rearrangement that occurs during appressorium morphogenesis . Because CAP proteins bind to G-actin , the globular structures at the base of appressoria may serve as the deposits of G-actins prior to penetration . When the CAP1-GFP transformant XY61 was treated with cytochalasin A ( CytA ) , an inhibitor of actin elongation , the subcellular localization pattern of Cap1-GFP in hyphae was completely disrupted . Instead of forming apical cortical patches , GFP signals were mainly observed in large cytoplasmic aggregates in the presence of CytA ( Fig . 8A; Video S2 , S3 ) . CytA treatment also disrupted the normal localization pattern of Cap1 during appressorium formation ( Fig . 8B; Video S4 , S5 ) , indicating that subcellular localization of Cap1 was changed from the actin-like pattern to the cytoplasm in both hyphae and appressoria . We then examined the localization of Cap1-GFP in invasive hyphae . In rice leaf sheath cells , invasive hyphae formed by the CAP1-GFP contained bright spots of GFP signals in the cytoplasm ( Fig . 7B ) . However , we failed to observe actin-like patches at the apical region of invasive hyphae . The difference between vegetative and invasive hyphae in the localization of Cap1-GFP may be related to differences in the role of actin cytoskeleton in hyphal growth . Bulbous invasive hyphae are morphologically distinct from vegetative hyphae and may lack the typical hyphal tip elongation mechanism . To further confirm whether Cap1 has the actin-like distribution pattern , we generated a LifeAct-GFP construct and transformed it into the wild-type strain Guy11 . In the resulting transformant LA31 ( Table 1 ) , LifeAct-GFP mainly localized to patches in the apical region of germ tubes and hyphal tips . It also displayed a localization pattern similar to Cap1-GFP in developing and mature appressoria ( Fig . 7A ) . GFP signals were mainly observed in the periphery of developing appressoria and globular structures at the base of mature appressoria ( Fig . 7A ) . In invasive hyphae formed by the transformant LA31 inside rice leaf sheath cells , we failed to observe actin-like patches or Cap1-like bright spots , further indicating that vegetative and invasive hyphae differ in the actin cytoskeleton organization at the tip . While LifeAct binds to F-actin , the C-terminal domain of CAP proteins binds to G-actin . In this study , we observed that Cap1-GFP and LifeAct-GFP had a similar localization pattern . To distinguish their subcellular localization patterns , we generated the CAP1-mCherry construct and co-transformed it with LifeAct-GFP into Guy11 . In the resulting transformant MC20 ( Table 1 ) expressing both constructs , LifeAct-GFP and Cap1-mCherry had similar localization patterns , but they were not co-localized in most of the cytoplasmic regions in vegetative hyphae , conidia , and appressoria ( Fig . 9 ) . To determine its function , we generated the AB domain knock-in deletion construct in which residues 378–533 of CAP1 were deleted . After transforming into the wild-type strain 70-15 , the resulting CAP1ΔAB transformant HC10 ( Table 1 ) was confirmed by Southern blot analysis ( Fig . S4 ) . On OTA plates , transformant HC10 grew faster than the Δcap1 mutant but still slower than Ku80 ( Fig . 10A ) . Conidiation was normal in the CAP1ΔAB transformant HC10 ( Table 3 ) , suggesting that the actin-binding domain is dispensable for the function of CAP1 in conidiation . In infection assays with rice seedlings , transformant HC10 caused more lesions than the Δcap1 mutant , but it was still reduced in virulence compared with Ku80 ( Fig . 10B , Table 3 ) . Similar results were observed in infection assays with barley seedlings . Therefore , although it is dispensable for conidiation , the AB domain of CAP1 is important for the normal function of Cap1 and plays a role in plant infection and hyphal growth . We also generated and transformed the CAP1ΔACB knock-in deletion construct into protoplasts of the wild-type strain Ku80 . The phenotype of the CAP1ΔACB transformants XY111 and XY002 ( Table 1 ) was similar to that of the Δcap1 mutant . All the defects associated with deletion of CAP1 were observed in the CAP1ΔACB transformants , including reduced appressorium formation , production of branching germ tubes ( Fig . 10A ) , and reduced virulence ( Fig . 10B ) . Thus , the ACB domain is essential for Cap1 function in M . oryzae . We then generated the CAP1ΔAB-GFP construct and transformed it into protoplasts of the Δcap1 mutant . In the resulting CAP1ΔAB-GFP transformant XY105 ( Table 1 ) , GFP signals had the Cap1WT-like localization pattern in appressoria ( Fig . 11A ) and hyphal tips ( Fig . 11B ) . Therefore , the C-terminal AB domain is not essential for the subcellular localization of Cap1 in M . oryzae . Because the AB domain is dispensable for Cap1 localization , we transformed the CAP1ΔACB- , CAP1ΔP1- , and CAP1ΔP2-GFP constructs that were deleted of residues 2–166 , 257–290 , and 355–377 , respectively , into protoplasts of the Δcap1 mutant . In transformants expressing the CAP1ΔACB-GFP or CAP1ΔP1-GFP constructs ( Table 1 ) , the localization of GFP signals was the same as the Cap1WT-GFP transformant in appressoria ( Fig . 11A ) and hyphal tips ( Fig . 11B ) . In the CAP1ΔP2-GFP transformants XY244 and XY001 ( Table 1 ) , GFP signals were observed in the cytoplasm in appressoria ( Fig . 11A ) and vegetative hyphae ( Fig . 11B ) . The Cap1WT-like localization pattern of GFP-signals was not observed in conidia or invasive hyphae of the CAP1ΔP2-GFP transformants ( Fig . S5 ) . These results indicate that the P2 region is responsible for the subcellular localization of Cap1 in M . oryzae . On hydrophobic surfaces , CAP1ΔP2/Δcap1 transformants produced long , branching germ tubes and were reduced in appressorium formation efficiency ( Fig . 12A ) . In spray infection assays , the CAP1ΔP2/Δcap1 transformant XY244 formed 17 . 3±4 . 3 lesions/5 cm leave tip , which was significantly lower than 51 . 3±4 . 5 lesions/5 cm leave tip caused by the wild type ( Fig . 12B ) . These results indicated that the P2 region of Cap1 is required for normal appressorium formation and virulence . Because the AB domain of CAP1 is important for full virulence , we assayed appressorium formation in the CAP1ΔAB transformant HC10 ( Table 1 ) . On hydrophobic surfaces , melanized appressoria were formed by transformant HC10 by 24 h ( Fig . 13A ) . On hydrophilic surfaces , it produced branching germ tubes and subapical swollen bodies but not melanized appressoria ( Fig . 13B ) . On hydrophilic surfaces , cAMP treatment induced appressorium formation on short germ tubes in the CAP1ΔAB strain ( Fig . 13C ) . Some of the melanized appressoria were formed immediately adjacent to the germinating conidia without visible germ tubes ( Fig . 13C ) . Interestingly , approximately 40% of the CAP1ΔAB conidia became melanized in one of the conidium compartments in the presence of 5 mM cAMP ( Fig . 13C ) . In hundreds of conidia examined , only one conidium compartment was melanized . The other two compartments in the same conidium often were empty and dead by 24 h , which is similar to what has been observed in conidia after appressorium formation [40] . However , exogenous cAMP failed to induce the formation of melanized conidium compartments in the CAP1ΔAB transformant on hydrophobic surfaces ( Fig . 13C ) , indicating that the surface recognition signal could suppress the effect of cAMP treatment on improper melanization of conidium compartments . To determine whether melanized conidium compartments were similar to appressoria , we introduced the GAS2-GFP construct [41] into the CAP1ΔAB transformant . GAS2 encodes a cytoplasmic protein specifically expressed during late stages of appressorium formation ( Fig . S6A ) [41] . The resulting transformant CGS ( Table 1 ) expressed GFP signals in the melanized conidium compartments , suggesting that those were appressorium-like structures ( Fig . 14A ) . When stained with DAPI , melanized conidium compartments induced by 5 mM cAMP contained a single nucleus ( Fig . 14B ) . The other two compartments were devoid of DAPI staining at 24 h . These results indicate that exogenous cAMP induced the formation of appressorium-like structures in conidia without germination . We also transformed the CAP1ΔAB construct into the Δpmk1 mutant nn78 that was defective in appressorium formation on hydrophobic surfaces [37] . In the resulting Δpmk1/CAP1ΔAB transformant HC1-39 , cAMP treatment failed to induce the formation of melanized conidium compartments ( Fig . 14C ) . Exogenous cAMP also failed to induce appressorium formation in the Δpmk1/CAP1ΔAB transformant on hydrophobic surfaces ( Fig . S6B ) . Therefore , like appressorium formation , the development of melanized conidium compartments is under the control of Pmk1 . The cAMP-PKA pathway is involved in surface recognition , appressorium turgor generation , and invasive growth in M . oryzae [8] , [15] , [37] . In other plant pathogenic fungi , cAMP signaling also has been implicated in the regulation of various differentiation and infection processes [42] , [43] . In this study , proteins associated with the Mac1 adenylate cyclase in M . oryzae were identified by affinity purification . One of the Mac1-interacting proteins was the adenylyl cyclase associated protein Cap1 . CAP proteins are well conserved from yeast to humans [44] . The Cap1 protein of M . oryzae has the typical structural features ( Fig . 1A ) of CAPs . The well-conserved N- and C-terminal regions of CAP1 are rich in α-helix and β-sheet structures , which are known to interact with adenyly cyclase and actin monomers , respectively [45] , [46] . Recently , the N-terminal region of yeast and mammalian CAP proteins were reported to bind to the cofilin-actin complex , which promotes the actin turnover [19] , [47] . We also noticed that the C-terminal region of Mac1 has the heptad motif that is sufficient for coiled-coiled interactions with Cap1 ( Fig . 1A; Fig . S1B ) . Unlike their association with actin cytoskeleton , the function of CAP proteins in regulating the adenylyl cyclase activity was not conserved in animals and plants [10] , [44] , [48] . Adenylyl cyclase was reported to be an integral transmembrane protein in mammalian cells [32] and a peripheral membrane protein in the budding yeast [33] . In BiFC assays , Cap1 and Mac1 appeared to weakly interact with each other in vegetative hyphae , conidia , and appressoria . Their interaction was enhanced during appressorium formation and localization of YFP signals to the cytoplasmic membrane was observed ( Fig . 2 ) . In S . cerevisiae , Cyr1 adenylate cyclase is activated by Ras2 and enriched in a subset of G-protein-containing fractions of cytoplasm membrane [49] . Although it is not essential , the interaction of Srv2 with Cyr1 and posttranslationally modified Ras2 [50] is important for the Ras-dependent activation of Cyr1 . Ras proteins are known for their membrane-anchoring functions in recruiting effector molecules to the cytoplasma membrane [51] . In yeast , disruption of the IRA1 Ras GTPase gene released the majority of adenylate cyclase activities ( 90% ) to the cytosol , which was significantly more than 20% in the wild-type cells [52] . In M . oryzae , the interaction of Cap1 with Ras2 and Mac1 may also play a role in the recruitment of Mac1 to the cytoplasm membrane and its activation . The Δcap1 mutant had a reduced intracellular cAMP level , indicating that CAP1 is involved in the cAMP-PKA pathway , possibly by regulating the activity of adenylate cyclase . Cap1 interacted with Mac1 in yeast two-hybrid and co-IP assays ( Fig . 1B , 1C ) . In S . cerevisiae , a small N-terminal region of Srv2 is sufficient for its association with Cyr1 and function in the Ras-adenylyl cyclase pathway [31] . L16P and R19T mutations of Srv2 resulted in attenuated cAMP signaling and reduced cortical actin patch localization [16] . Our data showed that the N-terminal region of CAP1 is essential for its interaction with Mac1 but not its actin-like localization pattern in M . oryzae ( Fig . S2 ) . The phenotype of the CAP1ΔACB/Δcap1 transformant was similar to that of the Δcap1 mutant , further confirming the importance of the ACB domain in Cap1 function . Because the Δcap1 mutant had a higher intracellular cAMP level than that of the mac1 mutant and it still recognized hydrophobic surfaces for appressorium formation , we conclude that CAP1 is not essential for surface recognition . The full activation of Mac1 was reduced but not completely blocked in the cap1 mutant . In S . cerevisiae , Srv2 plays an important role in normal activation of Cyr1 and the direct interaction between Cyr1 and Ras2 . Srv2 was first identified as a suppressor of a hyper-activated RAS2V19 allele in yeast , indicating their genetic association [29] . We also found that deletion of CAP1 suppressed the improper formation of melanized appressoria on hydrophilic surfaces in the RAS2DA transformant ( Fig . 4 ) . Therefore , Cap1 may also facilitate the interaction of Ras2 with Mac1 in M . oryzae . On hydrophobic surfaces , most of the germ tubes of the Δcap1 mutant were morphologically abnormal ( Fig . 3A ) . It appeared that these germ tubes attempted to form appressoria but failed to be arrested in tip growth . In addition , branching germ tubes were often observed in the mutant . Because exogenous cAMP suppressed these defects , maintaining the normal intracellular cAMP level must be important for the regulation of germ tube growth and branching ( Fig . 5C ) . It is likely that Cap1 plays a critical role in the proper activation of Mac1 and regulation of normal germ tube growth and appressorium formation . The Δcap1 mutant may be defective in actin cytoskeleton reorganization in response to surface recognition signals and maintaining polarized tip growth in germ tubes . However , the CAP1ΔAB transformant produced relatively normal germ tubes and appressoria ( Fig . 13 ) . It did not form branching germ tubes , suggesting that the AB domain is dispensable for the suppression of germ tube branching after the initiation of appressorium formation . Deletion of the AB domain had some minor effects on but did not eliminate the cortical patch localization of Cap1 ( Fig . 11 ) . Domain deletion analysis indicated that the P2 region but not the AB domain is responsible for the subcellular localization of Cap1 in M . oryzae . Furthermore , the P2 deletion mutant produced branching germ tubes on hydrophobic surfaces and had a reduced virulence , indicating that loss of the actin-like localization pattern resulted in phenotypes similar to those of the CAP1ΔACB transformant although to a less degree . In S . cerevisiae and C . albicans , the AB domain of CAPs also is dispensable for its association with actin cortical patches and cytoskeletal organization [24] , [53] . The P2 region likely binds to the SH3 domain of Abp1 , which may facilitate the localization of Cap proteins to sites of actin rearrangement . In yeast , the P2 motif is important for the binding of Srv2 with G-actin and directing its localization to cortical actin patches [47] , [54] . The P1 region of Srv2 may be responsible for profiling-binding but dispensable for actin-binding [16] , [19] . Consistent with these reports , we found that the P1 region was dispensable for the actin-like localization pattern of Cap1 in M . oryzae . In human pathogens , the cap1/cap1 mutant of C . albicans and the Δaca1 mutant of C . neoformans are non-pathogenic [23] , [24] . The Δcap1 mutant of M . oryzae was significantly reduced in virulence but still incited few small lesions on rice and barley leaves . In the rice blast fungus , MAC1 is essential for virulence . It is likely that reduced virulence of the Δcap1 mutant is directly related to the function of CAP1 in the cAMP-PKA pathway . Lesions caused by the Δcap1 mutant rarely had extensive necrosis indicating its defects in invasive growth in plant cells ( Fig . 6B ) . M . oryzae is a hemibiotrophic pathogen that has been used as a model for studying fungal–plant interactions [55] . However , it is not clear when and how the transition from the biotrophic phase to necrotrophic growth occurs . CAP1 may play a role in regulating the biotrophic-necrotrophic transition . It is also possible that CAP1 plays a role in maintaining normal invasive growth . Bulbous invasive hyphae of M . oryzae are morphologically distinct from vegetative hyphae and may involve different hyphal growth mechanisms . In yeast , Srv2 functions as an adaptor protein for the translocation of adenylate cyclase to actin cortex patches but this translocation is not essential for the cAMP signaling pathway [16] . In higher eukaryotes , the localization of CAP is species specific . In Dictyostelium discoideum , CAP localizes near the plasma membrane in resting cells but is remobilized during cell movement [56] . In mammalian cells , hCAP1 is distributed throughout cytoplasm but concentrated at actin-rich membrane ruffles and lamellipodia of migrating fibroblasts [20] , [57] . In cells induced for apoptosis , Cap1 was rapidly translocated to the mitochondrium independent of caspase activation [58] . In this study , we found that Cap1 in M . oryzae also had an actin-like localization pattern in hyphae and germ tubes . Cap1 mainly localized to patches in the apical region of vegetative hyphae or germ tubes and the base of appressoria surrounding the appressorium pore area . LifeAct had similar localization patterns with Cap1-GFP , but they usually were not co-localized to the same cytoplasmic regions , which is consistent with their differences in binding to different forms of actins [27] , [47] . Cap1 may also function as an adaptor protein for translocating Mac1 to the actin cortex in M . oryzae . Interestingly , localization of Cap1 or LifeAct to the apical region of invasive hyphae was not observed in the same transformants that displayed actin-like patterns in vegetative hyphae . The role of the actin cytoskeleton in hyphal tip growth may differ between vegetative and invasive hyphae . Expression of the CAP1ΔAB allele partially suppressed the Δcap1 deletion mutant in germ tube growth and virulence . Interestingly , the CAP1ΔAB transformant formed melanized conidial compartments when treated with 5 mM cAMP on hydrophilic surfaces . Because GAS2 [41] was specifically expressed in appressoria , those melanized compartments appeared to be appressorium-like structures . In M . oryzae , mitosis is known to be a prerequisite for appressorium development [59] , and DNA replication is necessary for the initiation of appressorium formation [40] , [60] . In the CAP1ΔAB transformant , only one nucleus was observed in the melanized compartments , suggesting that mitosis may not occur before these compartments become melanized in responses to exogenous cAMP . In S . cerevisiae , activation of the cAMP-dependent pathway causes cells to undergo unipolar growth , a process coupling with elongated growth that is controlled by the filamentous MAPK pathway [61] . It is likely that PMK1 was over-activated by exogenous cAMP when the AB domain of CAP1 was deleted in M . oryzae . Deletion of the AB domain of CAP1 in the pmk1 mutant failed to cause the formation of melanized conidium compartments , indicating that the formation of melanized conidium compartments requires the presence of functional Pmk1 ( Fig . 13C ) . Melanized conidium compartments were not observed in the Δcap1 mutant when treated with 5 mM cAMP . When the entire CAP1 gene was deleted , exogenous cAMP stimulated appressorium formation on germ tubes . Therefore , deletion of the AB domain or the entire CAP1 gene had different effects on responses to cAMP treatment . One possible explanation for this puzzling observation is that , in addition to being involved in the activation of Mac1 and Ras2 for appressorium formation , Cap1 is involved in the feedback inhibition or down-regulation of Ras2 signaling when Pmk1 is activated ( Fig . 15 ) . Ras2 likely functions upstream from both the cAMP signaling and Pmk1 MAPK pathway in M . oryzae . The Cap1ΔAB protein may be defective in the feedback inhibition of Mac1 and Ras2 , but retains the ability to regulate appressorium morphogenesis in response to exogenous cAMP . Therefore , cAMP treatment may overstimulate Ras signaling in the CAP1ΔAB transformant and result in the inappropriate activation of the Pmk1 pathway , which may be responsible for the formation of melanized conidium compartments and appressoria without visible germ tubes ( Fig . 15 ) . Interestingly , serine residue 450 of Cap1 was predicted to be a putative MAPK phosphorylation site ( over 50% probability ) by Kinasphos2 . 0 . This serine residue is conserved in its orthologs from S . cerevisiae , C . albicans , mammalian cells , and other few filamentous fungi examined . It is possible that Pmk1 , when activated , phosphorylates or interacts with Cap1 at the AB domain , which is involved in the down-regulation of Ras signaling . Therefore , further characterization of the functional relationships among Ras2 , Mac1 , and Cap1 may provide necessary information to better understand the interactions between the cAMP and PMK1 signaling pathways during appressorium morphogenesis and plant infection . The wild type and mutant strains of M . oryzae ( Table 1 ) were cultured on oatmeal agar ( OTA ) plates at 25°C under fluorescent light for conidiation and preserved on filter paper at −20°C as described [37] . For transformation selection , hygromycin ( Calbiochem , La Jolla , CA ) and zeocin ( Invitrogen , Carlsbad CA ) were added to final concentrations of 250 µg/ml and 200 µg/ml , respectively . Monoconidial culture isolation , measurement of growth rate and conidiation were performed as previously described [14] , [62] . Vegetative hyphae from 3-day-old 5×YEG ( 0 . 5% yeast , 1% glucose ) cultures were used for DNA , RNA , and protein isolation . Intracellular cAMP was assayed as described [12] , [63] with the cAMP enzyme immunoassay system ( Amersham Pharmacia Biotech , Piscataway , NJ ) . The upstream and downstream flanking sequences of CAP1 were amplified with primer pairs 1F/2R and 3F/4R ( Table S1 ) , respectively . The resulting PCR products were digested and ligated with the hygromycin-phosphotransferase ( hph ) cassette released from pCX63 as described [36] . The final gene replacement construct was amplified with primers 1F and 4R and directly transformed into protoplasts of Ku80 [64] . The putative Δcap1 mutants were identified by PCR and confirmed by DNA blot analysis . For complementation assays , the full-length CAP1 gene amplified with primers CF and CR was cloned between the NotI and XhoI sites of pYK11 [36] as pXY63 . The same ligation PCR approach [36] was used to generate the CAP1ΔAB construct and transformants . The upstream and downstream flanking sequences were amplified with primer pairs AB1F/AB2R and 3F/4R , respectively . After ligation with the hph gene , the ligation PCR product was transformed into Ku80 . The CAP1ΔAB transformants were confirmed by Southern blot analysis to be deleted of the actin-binding domain ( 378–533 aa ) . Conidia were harvested from 10-day-old OTA cultures and resuspended to 5×104 conidia/ml in H2O . For appressorium formation assays , 50-µl droplets of conidial suspensions were placed on glass cover slips ( Fisher Scientific , St . Louis , IL ) or GelBond membranes ( Cambrex , East Rutherford , NJ ) and incubated at 25°C . To assay its stimulatory effect on appressorium formation , cAMP was added to the final concentration of 5 mM to conidium suspensions . Penetration assays with onion epidermis and rice leaf sheaths were performed as described [8] , [65] , [66] . The growth of invasive hyphae was examined 48–72 h post-inoculation ( hpi ) . For infection assays , conidia were resuspended to 5×104 conidia/ml in 0 . 25% gelatin . Two-week-old seedlings of CO-39 were used for spray or injection infection assays as described [62] , [67] . Lesion formation was examined 7-day post-inoculation ( dpi ) . The S . cerevisiae srv2 deletion mutant was derived from MATα strain BY4741 ( Open Biosystems , Huntsville , AL ) . The entire CAP1 open reading frame ( ORF ) was amplified from 1st strand cDNA of strain 70-15 and cloned into pYES2 ( Invitrogen ) . The resulting construct , pYES2-CAP1 , was transformed into the srv2 mutant with the alkali-cation yeast transformation kit ( MP Biomedicals , Solon , OH ) . Ura3+ transformants were isolated and assayed for invasive growth and sensitivity to 5 mM H2O2 or 1 M NaCl as described [35] . The entire CAP1 gene was amplified and cloned into the pHZ126 and pDL2 [4] , [68] vectors by the yeast gap repair approach [69] . Similar approaches were used to generate the CAP1ΔACB , CAP1ΔAB , CAP1ΔP1 , and CAP1ΔP2 alleles that were deleted of amino acid residues 2–166 , 378–534 , 257–290 , and 355–377 , respectively . All of the primers used in the construction of these mutant alleles are listed in Table S1 . All of the resulting CAP1-3×FLAG ( pXY60 ) , CAP1-GFP ( pXY61 ) , CAP1ΔACB-3×FLAG ( pXY109 ) , CAP1ΔACB-GFP ( pXY94 ) , CAP1ΔAB-3×FLAG ( pXY110 ) , CAP1ΔAB-GFP ( pXY105 ) , CAP1ΔP1-GFP ( pXY95 ) , and CAP1ΔP2-GFP ( pXY244 ) fusion constructs were confirmed by sequencing analysis and transformed into the Δcap1 mutant HC83 . To confirm the interaction between MAC1 and CAP1 in vivo , the C-terminal region of MAC1 ( 1898–2016 aa , Mac1CT ) was cloned into pHZ126 by the yeast gap repair approach [69] . The resulting construct , pXY165 , was co-transformed with pXY61 into protoplasts of strain 70-15 . It also was co-transformed with pXY94 ( CAP1ΔACB-GFP ) into 70-15 to detect the Mac1-Cap1 interaction . Total proteins were isolated from transformants expressing both Mac1CT-3×FLAG and Cap1-GFP/Cap1ΔACB-GFP and incubated with anti-Flag M2 affinity resins ( Sigma ) . Proteins bound to M2 resins were eluted after a series of washing steps as described [67] , [70] . Western blots of total proteins and elution from the M2 resins were detected with anti-FLAG ( Sigma-Aldrich ) and anti-GFP ( Roche ) antibodies using the ECL Supersignal system ( Pierce , Rochford , IL ) . For affinity purification , total proteins were isolated from transformants expressing the CAP1-3×FLAG , CAP1ΔACB-3×FLAG , and CAP1ΔAB-3×FLAG constructs and incubated with anti-FLAG M2 affinity resins . After washing three times with 1×TBS ( pH 7 . 4 ) buffer , three times with 50 mM TMAB , and three times with ddH2O , proteins were eluted with 0 . 1% Rapigest as described [28] . The elution proteins were digested with trypsin [71] and analyzed by Nanoflow liquid chromatography tandem mass spectrometry ( nLC-MS/MS ) as described [70] . The resulting MS/MS data were used to search against the non-redundant M . oryzae protein database at NCBI . The first 17 amino acid residues ( MGVADLIKKFESISKEE ) of Abp140 of S . cerevisiae , named LifeAct , is an F-actin marker for higher eukaryotes [27] . The LifeAct sequence was amplified with primers LifeActF and LifeActR ( Table S1 ) and cloned into pDL2 and pFL1 by yeast gap repair [69] . The resulting LifeAct-Gly ( 10 ) -GFP constructs pXY198 ( HygR ) were confirmed by sequencing analysis and transformed into Guy11 . The CAP1-mCherry construct pXY210 was generated by cloning the CAP1 fragment into pXY201 , which was a vector generated in this study by replacing the GFP sequence on pYP1 [68] with the mCherry sequence . The bait construct of CAP1 was generated by cloning full-length CAP1 ORF amplified with primers CAP1-YF and CAP1-YR ( Table S1 ) into pAD-GAL4 ( Stratagene , La Jolla , CA ) . The C-terminal region of MAC1 ( 1926–2160 aa ) was amplified with primers MAC1-YF and MAC1-YR and cloned into pBD-GAL4-2 . 1 as the prey construct . The resulting prey and bait constructs were transformed in pairs into yeast strain YRG-2 ( Stratagene ) . The Trp+ and Leu+ transformants were isolated and assayed for growth on SD-Trp-Leu-His medium and the expression of LacZ reporter gene was detected according to the instruction provided by Stratagene . Yeast transformants expressing the MST11-bait and MST50-prey , PMK1-bait and MST50-prey constructs [12] , [14] were used as the positive and negative controls , respectively . The CAP1-CYFP fusion construct pXY178 was generated by cloning the CAP1 fragment amplified with primers FlagF and CBFR into pHZ65 [13] . The C-terminal region of MAC1 was amplified with primers Mac1-FLF and MBR and cloned into pHZ68 [13] to generate the MAC1-NYFP fusion construct pXY179 . Plasmids pXY178 and pXY179 were co-transformed into protoplasts of 70-15 . Transformants resistant to both hygromycin and zeocin were isolated and confirmed by PCR and Southern blot analyses . YFP signals were examined with a Nikon 800 epifluoresence microscope . Sequence data for genes described in this article can be found in the GenBank under the following accession numbers: M . oryzae CAP1 ( XM_363796 . 2 ) , MAC1 ( XM_365053 . 1 ) , S . cerevisiae SRV2 ( AAA35094 . 1 ) , Homo sapiens CAP1 ( CAG33690 ) , C . albicans CAP1 ( AAD42978 . 1 )
In Magnaporthe oryzae , cAMP signaling is known to play an important role in surface recognition and plant penetration . The Mac1 adenylate cyclase is essential for plant infection . To better understand Mac1 activation mechanisms , in this study we used the affinity purification approach to identify proteins that are associated with Mac1 in vivo . One of the Mac1-interacting protein is the adenylate cyclase associated protein ( CAP ) encoded by the CAP1 gene . Results from our study indicated that Cap1 is important for Mac1 activation and plant infection in M . oryzae . The Δcap1 mutant was defective in germ tube growth and appressorium formation and failed to cause typical blast lesions . Like LifeAct , Cap1 localized to apical patches in vegetative hyphae but not in invasive hyphae . The P2 proline-rich region was important for Cap1 localization but the actin-binding domain played a role in feedback inhibition of Ras signaling . To our knowledge , functional characterization of CAP genes has not been reported in filamentous fungi . Our results indicate that CAP1 is important for regulating adenylate cyclase activities , appressorium morphogenesis , and plant infection . Further characterization of CAP1 will be important to better understand the interaction between cAMP signaling and the PMK1 pathway in M . oryzae .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "mycology", "plant", "microbiology", "biology", "microbiology" ]
2012
The Cyclase-Associated Protein Cap1 Is Important for Proper Regulation of Infection-Related Morphogenesis in Magnaporthe oryzae
The relative contributions of neutral and adaptive substitutions to molecular evolution has been one of the most controversial issues in evolutionary biology for more than 40 years . The analysis of within-species nucleotide polymorphism and between-species divergence data supports a widespread role for adaptive protein evolution in certain taxa . For example , estimates of the proportion of adaptive amino acid substitutions ( α ) are 50% or more in enteric bacteria and Drosophila . In contrast , recent estimates of α for hominids have been at most 13% . Here , we estimate α for protein sequences of murid rodents based on nucleotide polymorphism data from multiple genes in a population of the house mouse subspecies Mus musculus castaneus , which inhabits the ancestral range of the Mus species complex and nucleotide divergence between M . m . castaneus and M . famulus or the rat . We estimate that 57% of amino acid substitutions in murids have been driven by positive selection . Hominids , therefore , are exceptional in having low apparent levels of adaptive protein evolution . The high frequency of adaptive amino acid substitutions in wild mice is consistent with their large effective population size , leading to effective natural selection at the molecular level . Effective natural selection also manifests itself as a paucity of effectively neutral nonsynonymous mutations in M . m . castaneus compared to humans . Several approaches have revealed evidence for adaptation at the molecular level . Local reductions in diversity , indicating selective sweeps , have been identified in populations of a number of species [1] , [2] . In Drosophila , reductions in neutral diversity have also been associated with increased divergence at amino acid sites , indicative of recurrent selective sweeps of advantageous amino-acid changing substitutions [3] . There is also evidence for a general reduction in diversity close to conserved sequence features in protein-coding genes and noncoding elements of hominids [4] . These reductions can be attributed either to genetic hitchhiking of positively selected alleles or background selection against negatively selected alleles . Higher FST within genic than nongenic regions of the human genome suggests that genic regions are subject to local adaptation within populations [5] , ( although see [6] ) . Further evidence for adaptation has come from attempts to identify sites or genes subject to recurrent positive selection by looking for an excess of substitutions in sites of interest over that expected . For example , a nonsynonymous to synonymous divergence ratio exceeding one at a locus may be evidence for positive selection at nonsynonymous sites . An approach that tests for an excess of substitutions at selected sites is the McDonald-Kreitman test [7] , which contrasts levels of polymorphism with divergence at selected sites ( e . g . , nonsynonymous sites ) and linked putatively neutral sites ( e . g . , synonymous sites ) . A recent extension of this test quantifies molecular adaptation as the fraction of substitutions driven to fixation by positive selection ( α ) [8] , [9] by comparing the observed number of selected substitutions to the number expected , based on levels of polymorphism and divergence at neutral sites . The application of derivatives of the McDonald-Kreitman test to amino-acid changing sites has resulted in a wide range of estimates of α , the causes of which may be multifaceted . Relatively high estimates of α have been obtained for enteric bacteria [10] and consistently high estimates have been obtained for Drosophila [9] , [11]–[13] , suggesting that α may be 50% or more in these species . On the other hand , estimates from yeast [14] , [15] , Arabidopsis [16] and hominids have been low . In the case of hominids , several independent estimates have found α for hominids to be 13% at most ( with the exception of an estimate by Fay et al . [8] ) [17]–[20] . Some of the observed variation in estimates of α may also be attributable to differences in the methods used . Specifically , some estimates of α are compromised by slightly deleterious mutations , since these contribute proportionately more than neutral polymorphisms to diversity than divergence . If slightly deleterious mutations are prevalent and not properly accounted for then α could be substantially underestimated [21] . This may partially explain low estimates of α obtained using methods that do not incorporate explicit population genetics models ( e . g . yeast [14] , [15] and Arabidopsis , [16] ) . Recently , improved methods to estimate α have been developed that model the contribution of slightly deleterious mutations to polymorphism and divergence [20] , [22] . However , estimates of α obtained from hominids are low , even when based on methods that attempt to model for slightly deleterious mutations [20] , [22] , whereas estimates from Drosophila are high [22] . It is possible that this observation is a consequence of differences in effective population size ( Ne ) . Drosophila melanogaster is estimated to have an Ne of 1–2 million , whereas hominids seem to have an unusually low recent Ne [23] . In D . miranda , which is estimated to have a lower Ne than D . melanogaster , Ne is probably still an order of magnitude larger than that for hominids , and a high estimate of α was observed [13] . The proportion of adaptive substitutions is expected to depend on Ne for two reasons . Firstly , a higher proportion of both advantageous and deleterious mutations are expected to be effectively neutral in species with low Ne , because selection for/against slightly advantageous/deleterious mutations becomes less effective ( see [24] ) . In such species , a higher proportion of slightly deleterious mutations , and conversely , a smaller proportion of slightly advantageous mutations , are expected to contribute to divergence . Secondly , if the rate of adaptation is limited by the supply of mutations , then species with low Ne will adapt more slowly simply because they have to wait longer for each new advantageous mutation to appear in the population . Currently there are no estimates for rates of adaptive evolution of protein-coding genes in mammals other than hominids , particularly for species with higher Ne . M . m . castaneus populations from NW India , a region believed to be part of the ancestral range of the house mouse sub-species complex [25] , have silent-site diversity for the X-chromosome of the order of 1% [26] . When combined with an estimate of the per nucleotide mutation rate per generation for murids [27] , this level of diversity suggests that M . m . castaneus Ne is two orders of magnitude higher than recent Ne of hominids , and comparable to Ne typically seen in Drosophila . We hypothesised that the murid protein-coding genes would therefore show pervasive natural selection , whereas its impact is much reduced in hominid orthologs . We tested the hypothesis by estimating α from protein-coding genes of murid rodents by comparing nucleotide polymorphism data of M . m . castaneus sampled from the NW Indian population with nucleotide divergence to M . famulus and the rat . To infer levels of negative and positive selection in murid protein-coding genes , we analysed nucleotide diversity within a sample of 15 wild , unrelated M . m . castaneus from the NW Indian population together with the nucleotide divergence between M . m . castaneus and either M . famulus or the rat . We sequenced amplicons from a sample of 77 autosomal loci that are part of the Environmental Genome Project ( EGP ) [28] ( details of the genes sequenced are presented in Table S1 ) . These loci are not a random sample , since they are associated with human genetic diseases whose susceptibility is influenced by environmental challenge . However , they show low rates of adaptive amino acid substitution that are typical of hominids [17]–[20] . Summary statistics concerning nucleotide diversity at intronic , 4-fold degenerate , 2-fold degenerate and 0-fold degenerate sites are shown in Table 1 and Table S2 and the allele frequency distributions ( or site frequency spectra , SFS ) are plotted in Figure 1 . As expected , zero-fold degenerate sites have the lowest nucleotide diversity , lowest divergence , the most negatively skewed SFS , and the most negative estimate of Tajima's D , a statistic related to the skew in the distribution of allele frequencies [29] . This is consistent with purifying selection keeping most amino acid mutations at low frequencies and reducing the number of fixations . Nucleotide diversity is higher for synonymous than intronic sites , as is Tajima's D . Together with a slightly higher synonymous than intronic divergence between M . m . castaneus and M . famulus ( Table 1 ) , this suggests somewhat weaker purifying selection acting on synonymous than intronic sites in murids , and that synonymous sites are likely to be the most appropriate neutral reference . Recent Ne in wild mice , humans and Drosophila can be compared by equating synonymous site nucleotide diversity ( θπ ) to 4Neμ , where μ is an estimate of the mutation rate per site per generation ( Table 2 ) . Using estimates of μ based on synonymous site divergence and an assumption of two generations per year , our estimate of Ne for wild mice of 580 , 000 is similar to that obtained for African D . melanogaster , whereas , in African populations of humans , Ne is nearly two orders of magnitude smaller . Our estimate for M . m . castaneus is consistent with , although marginally higher than , a recent estimate of 400 , 000 ( also assuming two generations per year ) [30] , based on smaller sample of loci . Nucleotide diversity in NW Indian M . m . castaneus is approximately one order of magnitude higher than observed in derived populations of M . m . domesticus and M . m . musculus from Europe and two orders of magnitude higher than among laboratory inbred mouse strains [26] , [30] , [31] . To estimate parameters of the distribution of fitness effects of deleterious amino acid-changing mutations we used a maximum likelihood ( ML ) procedure [32] that contrasts the SFS at putatively neutral sites ( four-fold degenerate or intronic sites in this case ) with sites assumed to be subject to purifying selection ( nonsynonymous sites ) . The procedure fits a gamma distribution of deleterious mutational effects to the nonsynonymous SFS , and a demographic model to both the neutral and nonsynonymous SFSs that allows a step change in population size at some time in the past . The method assumes that positively selected mutations make a negligible contribution to polymorphism . Selective effects ( s ) of new amino acid mutations are estimated as the product of Ne and s ( see Materials and Methods for details of the method ) . Assuming four-fold sites as the neutral reference , estimates of proportions of amino acid mutations that have fitness effects in different Nes ranges under the best-fitting mutation effect distributions are compared in Table 3 for our M . m . castaneus data set , three African or African-American human data sets ( [28]; the “Seattle SNPs” Programs for Genomic Applications ( PGA ) [33]; the dataset of Boyko et al . [20] ) and an African D . melanogaster data set [11] . Similar results are obtained if intronic sites are used as the neutral reference ( Table 3 ) . Nearly neutral deleterious amino acid mutations ( i . e . , mutations for which Nes<1 ) , which have an appreciable chance of drifting to fixation , are relatively uncommon in both mice and Drosophila ( 10% and 6% of amino acid mutations , respectively ) , whereas they make up ∼20% of amino acid mutations in humans ( maximum P = 0 . 038 for mouse versus human comparison , see Table 4 for details; P = 0 . 25 for mouse vs . Drosophila comparison ) . Strongly deleterious mutations ( Nes>10 ) , which essentially never become fixed , are inferred to be somewhat more frequent in mice and Drosophila ( 79% and 87% , respectively; P = 0 . 21 ) than humans ( ∼70%; P<0 . 05 for all contrasts with mice except one , see Table 4 for details ) . Whilst it is possible that these differences between the species in the relative frequencies of mutations in different Nes categories reflect differences in the distribution of absolute selection coefficients ( s ) between species , it is more likely that they reflect differences in Ne . For example , a lower long term Ne in humans would allow more deleterious mutations to segregate at higher frequencies than in either mice or Drosophila . ML estimates of the demographic parameters of the model using four-fold sites as the neutral reference imply that there has been a recent increase in Ne in M . m . castaneus ( Table S3 ) , as well as African D . melanogaster [32] . A lack of neutral diversity in fast evolving genes has previously been interpreted as evidence for the effects of selective sweeps and therefore adaptation in Drosophila [13] , [34] , [35] . However , in contrast to these results , we found a nonsignificant positive correlation between synonymous site diversity and nonsynonymous divergence ( Spearman r = 0 . 21 , p = 0 . 084 for dN vs . θπ and r = 0 . 16 , p = 0 . 084 for dN vs θS ) . Therefore , unlike in Drosophila , our data do not suggest that selective sweeps in genes undergoing high rates of adaptive evolution reduce local neutral diversity , although our relatively small data set limits the power of this analysis . To further investigate evidence for adaptation we estimated the fraction of adaptive amino acid substitutions , α , between M . m . castaneus and either M . famulus or the rat by a method related to the McDonald-Kreitman test for adaptive evolution [7] that contrasts polymorphism with divergence [22] . The method attempts to account for nearly neutral amino acid mutations , which , when compared to strongly deleterious mutations , contribute proportionately more to polymorphism than divergence . The parameters of the distribution of effects of deleterious amino acid mutations , estimated by ML from polymorphism within M . m . castaneus , and the neutral site divergence between M . m . castaneus and the outgroup ( M . famulus or rat ) are used to compute the number of amino acid substitutions expected between M . m . castaneus and the outgroup . The estimated fraction of adaptive substitutions is the difference between this expected number and the observed number of amino acid substitutions , scaled by the observed number ( see [22] and Methods ) . Simulations suggest that the method produces close to unbiased estimates of α if the assumptions of the model are met , and is robust to substantial departures from the model assumptions , including complex demographic scenarios and linkage between sites [22] . However , in common with all McDonald-Kreitman based approaches , it is sensitive to long-term population size changes , a point that is discussed later . Our estimate of α for wild mice , assuming four-fold sites as the neutral reference , is 57% ( Table 3 ) . This is somewhat higher than , but non-significantly different to , an estimate of 52% for African D . melanogaster with divergence to D . simulans ( P = 0 . 54 ) . However , the estimate for α in mice is very much higher than estimates for hominids for all the human polymorphism data sets , including the EGP data ( P = 0 . 014 for a comparison only involving the subset of gene orthologs sequenced in mice using divergence to macaque ) and PGA ( P = 0 . 11 using divergence to macaque ) , and the data set of Boyko et al . [20] ( P = 0 . 020 using divergence to chimpanzee ) . CpG dinucleotides have an elevated mutation rate in mammals and differ in frequency between coding and non-coding DNA . However , using only sites that are unlikely to be part of a CpG dinucleotide ( non-CpG-prone sites ) yields estimates of α that are similar to those based on all sites ( Table 5 ) . Estimates of α could also be affected by more complex demographic scenarios , such as admixture between differentiated sub-species and/or population subdivision , that are not modelled in our algorithm . We tested for evidence of population structure or admixture using the program Structure [36] using one randomly sampled four-fold degenerate or intronic SNP per sequenced locus . For both intronic and 4-fold degenerate synonymous sites we found no evidence for population subdivision in the M . m . castaneus sample , since the “no-admixture” model gives P = 1 in all but one case for a number of populations parameter K = 1 ( see Table S4 for details ) . However , under the “admixture” model there is better support for two populations ( K = 2 ) than one population ( Table S4 ) , suggesting population subdivision . Figure 2 shows an ancestry plot for one randomly selected run that provided support for K = 2 . In this plot each individual shows ancestries in both putative populations ( with roughly equal proportions in the population ) , suggesting that they are admixed . However , we do not find any individuals that have ancestries in just one of the two populations ( i . e . , there are no individuals that are purely from population 1 or purely from population 2 ) , suggesting that this result can be explained by a violation of an assumption in Structure , namely that genotype frequencies are at Hardy-Weinberg equilibrium . House mice are known to inbreed in the wild [37] , potentially causing an elevated inbreeding coefficient ( Fis ) . To determine whether such an effect can be observed in our sample , we calculated Fis values [38] for each SNP using the program Genepop ( http://genepop . curtin . edu . au/ ) . We found a substantial excess of loci showing positive Fis values , indicating a deficiency of heterozygotes and a deficiency of negative Fis values , indicating an excess of homozygotes ( Figure S1 ) . Thus , our sample shows evidence for inbreeding . In summary , our interpretation of these results is that the M . m . castaneus population shows no evidence for hidden population substructure or admixture between differentiated subspecies , but there is evidence that inbreeding is a feature of all individuals used in the study . We attempted to account for the effect of inbreeding , and therefore the possibility that alleles from the same individual are not independent samples from the population , by repeating the analysis for 20 datasets created by randomly selecting one allele from each individual for each site , such that each sequence analysed in each data set was a composite , derived from a single individual . We then calculated mean estimates of α by averaging over the 20 randomly generated datasets . When calculated by this method , our estimates of mean α are only marginally lower than estimates using the complete data set ( see Table S5 ) . Estimates of α obtained using rat as the outgroup are generally somewhat lower than those using M . famulus , but are still close to 40% ( estimates range from 0 . 33 to 0 . 51 , see Table 5 ) , suggesting that the estimate of α in murids is robust to the choice of outgroup . An earlier method to estimate α [8] attempts to remove the influence of nearly neutral deleterious mutations by excluding polymorphisms at a frequency below an arbitrary threshold ( e . g . , 10% ) . Estimates of α produced by this method are somewhat lower than the estimates from our method ( Table 6 ) , but this is expected because estimates are likely to be downwardly biased [21] . However , they are substantially higher than estimates of α using this method in hominids [8] . If all sites are included in this analysis , irrespective of their frequency , estimates of α in wild mice are close to zero , or even negative ( Table 6 ) . This is likely to be due to slightly deleterious mutations , which contribute low frequency polymorphisms but have little chance of fixation , and lead to downwardly biased α estimates . Indeed even when low-frequency , segregating at a frequency of <10% , are excluded , analyses suggest that estimates of α may still be downwardly biased . Currently available estimates of α from a variety of species vary widely . The estimates of the fractions of adaptive substitutions in microbes [10] , Drosophila [9] , [11] , [12] and now mice present a serious challenge to the neutralist view of protein evolution . Taken together these results suggest that most amino acid substitutions are caused by positive selection , and that genetic drift is therefore not the most important cause of protein evolution . However , estimates of α obtained for yeast [14] , [15] , Arabidopsis [16] , and hominids ( which are ∼10% at most [17]–[20] ) suggest the opposite . There are several possible explanations for these discrepancies . One possibility is that the estimates obtained for yeast and Arabidopsis are not based on an explicit population genetics model , and even though attempts have been made to reduce the impact of slightly deleterious mutations , the estimates may still be downwardly biased . Nevertheless , these results are hard to reconcile with estimates from microbes , Drosophila and mice , since both of these species are thought to have a relatively high Ne . On the other hand , the low estimated proportion of adaptive substitutions in hominids may reflect their low Ne , since this will increase the proportion of effectively neutral advantageous and deleterious mutations . Low Ne will also reduce the rate of adaptive evolution if the rate is limited by the supply of mutations . This is consistent with the low recent Ne estimates for humans [23] , chimpanzees [39] and gorillas [40] . It is also possible that most adaptive evolution occurs in noncoding regions in primates [41] . Alternatively , changes in effective population size can lead to bias in the estimate of α [7] , [42] . It can be shown that if the true value of α is independent of Ne , but that the current Ne ( which affects the level of polymorphism ) is different to the average Ne over the evolution of the species ( which affects the level of divergence ) then the relationship between the true ( αtrue ) and estimated ( αest ) values of α is given by ( 1 ) [22] , if the distribution of fitness effects is gamma , where λ is the ratio of the current and ancestral Ne and b is the shape parameter of the gamma distribution of mutational effects . Thus , a contraction in Ne will lead to an underestimate of α and an increase will lead to an overestimate . It is therefore possible that the difference in the estimate of α between hominids and rodents is due to recent demography; if the current Ne of humans was much smaller than the ancestral population size , and/or the current Ne of M . m . castaneus was much larger than the ancestral , then αtrue could be very similar in the two species . Recent evidence suggests that ancestral great ape Ne may have been substantially bigger than current [43] . So , for example , assuming b for humans is 0 . 2 [20] , [32] , [44] , αest = 0 . 1 implies αtrue = 0 . 35 and 0 . 43 for 5- and 10-fold reductions in long-term Ne , respectively ( equation 1 ) . However , we also infer from our polymorphism data that M . m . castaneus has undergone a recent increase in Ne , although our evidence for this is modest . Nevertheless , assuming that current M . m . castaneus Ne is 5- and 10-fold larger than the ancestral Ne , our estimates of αest = 0 . 57 and b = 0 . 31 ( see Table S3 ) would imply that αtrue = 0 . 29 and 0 . 12 respectively ( consistent with the estimates from humans ) . More estimates of α from other murid and mammalian species will help to determine whether the high rate of adaptive evolution we have inferred is widespread amongst murid species and therefore not an artefact of demography . The 15 M . m . castaneus individuals were collected in 2003 in 4 localities ( 2–5 individuals per locality ) along a 130 km transect ( from lat 32 . 244987° , lon 77 . 188181° to lat 30 . 977139° , lon 76 . 986026° ) south of the Himalayas in the North-West Indian state of Himachal Pradesh . Each locality extended over an area covering 5 km2 . To avoid collecting related individuals , we analysed only one individual per trap site within each locality and trap sites had to be separated by >500m . An individual M . famulus , originating from India ( locality Kotagiri ) , was obtained from the Montpellier wild mice genetic repository ( http://www . isem . cnrs . fr/spip . php ? article477 ) . DNA sequences were generated for genes sampled from a set whose human orthologs have been sequenced in the Environmental Genome Project ( EGP ) [28] . All genes sequenced as part of EGP are autosomal and many had polymorphism data available for African human populations at the time of searching , allowing us to make a direct comparison of the results we obtained in mice with results based on the same set of genes in a human population . Loci sequenced in Africans ( 618 as of 7th August 2007 ) whose orthologs could be identified in the mouse genome ( 585 genes , using NCBI Homologene ) were considered . For 77 loci , DNA sequences were generated for the 15 M . m . castaneus individuals and one M . famulus individual . Primers were designed to amplify regions that captured coding and intronic DNA using Primer3 [45] . PCR reactions were performed using GoTaq DNA polymerase ( Promega ) using a touchdown program consisting of 95°C for 15 minutes , followed by 28 cycles of 95°C for 30 seconds , 62°C for 45 seconds ( reducing by 0 . 5°C every cycle ) , 72°C for 2 minutes , then 12 cycles of 95°C for 30 seconds , 52 °C for 45 seconds and 72°C for 2 minutes , with a final extension at 72°C for 10 minutes . Following evaluation on 1% agarose gels , products were purified using ExoSAP-IT ( USB ) , or , if product indicated non-specific priming , the appropriate band was cut from a gel and extracted using Qiaquick gel extraction kit ( Qiagen ) . Forward and reverse sequences were generated using Big Dye Terminator Sequencing Kits ( Applied Biosystems ) on an ABI Prism 3730 DNA Analyzer . Sequence analysis and variant detection was carried out using CodonCode Aligner version 2 . 0 . 6 ( http://www . codoncode . com/aligner/ ) . Sequences had an average Phred score of >60 . All sequence traces were manually checked . CodonCode was set to highlight any site with a Phred score <30 ( which could include low quality sequence or heterozygous sites ) . All such sites were manually checked , but in order to avoid excluding heterozygotes , were not automatically excluded . Alignments between the 15 M . m . castaneus , the M . famulus individual and the M . m . musculus reference sequence were obtained . All alignments were manually checked before further analysis . We extended a maximum likelihood approach [32] to estimate parameters of the distribution of fitness effects of new amino acid mutations using the allele frequency distributions ( the site frequency spectra , SFSs ) for 0-fold and putatively neutrally evolving sites ( either 4-fold or intronic sites ) . We assumed that effects ( s ) of amino acid mutations are unconditionally deleterious , and sampled from a gamma distribution with shape and scale parameters a and b , respectively . These parameters were estimated along with the fraction of unmutated sites , f0 , and demographic parameters N1 , N2 and t , corresponding to ancestral population size , current population size and the number of generations since a population size change , respectively . Polymorphism data were summed across loci using folded SFSs . We extended the method to allow variation in the number of alleles at each site . We generated SFSs for sites with the same numbers of alleles , computed the log likelihood for each SFS , and summed these to compute the overall log likelihood . Selective effects are estimated on a scale Ns , where N is a measure of the population size at the time that the polymorphism data are censured . Under the assumption of a single step change in population size , there may be little information to estimate the relative values of the population size before ( N1 ) and after ( N2 ) the size change if , for example , t≫N2 or t≪N2 . We therefore computed a weighted recent population size fromwhere and . We estimated α for 0-fold substitutions using a method that attempts to account for the segregation of slightly deleterious mutations and recent population size changes [22] . The divergence at neutrally evolving sites ( dS ) is proportional to the mutation rate per site . At selectively evolving sites , the expected divergence due to fixation of deleterious mutations is proportional to the product of the mutation rate and the fixation probability , u ( N , Nes ) , of a new mutation [50] . Defining dN as the observed divergence at the selectively evolving sites , and equating N with Ne ( because these are equivalent under the transition matrix method under which population size is estimated ) α is proportional to the difference between the observed and expected divergence: ( note that dS and ds are different quantities ) . α was initially estimated using all sequenced alleles ( i . e . a total of 30 alleles , two per individual ) . To test whether our estimates of alpha could be biased as a result of assuming that different alleles from the same individual are independent , we also estimated α using only a selected and neutral reference sequence for each individual . Specifically we created 20 data sets in which , for each individual and for every site , we randomly picked a single base from the individuals' two alleles . Each data set therefore consisted of up to 15 selected/neutral reference sequences , each one corresponding to a different individual . Mean estimates of α were then computed by averaging over the 20 randomly generated datasets . We also used Fay , Wyckoff and Wu's [8] , [9] extension of the McDonald-Kreitman test to estimate the fraction of 0-fold substitutions driven to fixation by positive selection , αFWW:where DN ( PN ) and DS ( PS ) are numbers of divergent ( polymorphic ) sites for selected and neutral classes , respectively , and the summation is over genes . To reduce the influence of nearly neutral alleles on αFWW , we excluded sites where the rare variant was at a frequency of 10% or less . This method assumes that the number of alleles sampled is constant , so we again sampled the alignments to give 20 alleles per site and ignored sites that lacked an orthologous base in the outgroup . 95% confidence intervals for all statistics were obtained by bootstrapping 1 , 000 times by locus . House mice , especially those originating from the ancestral range , could exhibit a complex genetic composition , reflecting either incomplete lineage sorting or admixture between the different subspecies [30] , [51] . We used our multilocus SNP dataset to determine if the population sample of M . m . castaneus that was used in our study shows evidence for admixture or hidden population substructure . We randomly selected one SNP from each amplicon , excluding SNPs that cover known splice sites and only including SNPs where at least 10 of the 15 individuals could be sequenced . We excluded any SNPs covering indel sites . We separately analysed SNPs from intronic sites and SNPs from 4-fold degenerate sites . Altogether 84 ( intronic data ) or 82 ( 4-fold degenerate sites ) unlinked SNP loci were included in the analysis . We used the program Structure [36] to identify the presence of different subpopulations in the sample , if any , and to estimate the ancestry of the sampled individuals in each of these subpopulations . The number of subpopulations is inferred by calculating the probability P ( X|K ) of the data given a certain prior value of K ( number of subpopulations ) over a number of Monte Carlo Markov Chain ( MCMC ) iterations . The posterior probabilities P ( K|X ) can be calculated following Bayes' rule . The subpopulations are characterised by different allele frequencies , and , according to their multilocus genotypes , individuals are probabilistically assigned to one or more subpopulations . The scores of individuals in the subpopulations correspond to the probability of ancestry in any one of them . In this study we assumed prior values of K from 1 to 4 . We considered two models for the ancestry of individuals . In the first , the “no-admixture model” , individuals are assumed to be drawn purely from one of K populations . In the second , the “admixture model” , individuals are allowed to have mixed ancestry , that is , some fraction of an individual's genome comes from different subpopulations . Both of those models assume that all the markers are unlinked and provide independent information on an individual's ancestry . Inferences of the number of subpopulations and ancestries of individuals are based on 1 , 000 , 000 iterations of the MCMC , after a “burn-in” period of 100 , 000 iterations . We ran the program without incorporation of prior population information . We performed 3 independent runs of the Markov chain for each parameter set to check for convergence of the chains .
The prevalence of natural selection at the DNA level remains a controversial issue in evolutionary biology . In particular , estimates of the proportion of adaptive amino acid changes ( α ) vary greatly between taxa , being 50% or more in bacteria and fruit flies , but at most 13% in hominids . Here , we infer the frequencies of polymorphisms in protein-coding genes of 15 Mus musculus castaneus individuals sampled from the ancestral range of the house mouse species complex . By combining the polymorphism data with nucleotide divergence to the related murid species M . famulus and the rat , we obtain an estimate for α of 57% . This represents the first estimate of α for a mammal other than humans . The high rate of adaptive protein evolution in wild mice and other taxa implies that hominids may be somewhat unusual in having low rates of adaptive protein evolution . One possible cause of this is the low effective population size in humans , which is predicted to lead to less effective natural selection and fewer adaptive mutations . This is consistent with the higher frequency of nearly neutral deleterious amino acid mutations in hominids than murids that we infer in our analysis .
[ "Abstract", "Introduction", "Results/Discussion", "Materials", "and", "Methods" ]
[ "evolutionary", "biology", "genetics", "and", "genomics" ]
2010
Evidence for Pervasive Adaptive Protein Evolution in Wild Mice
Mucosal transmission of HIV is inefficient . The virus must breach physical barriers before it infects mucosal CD4+ T cells . Low-level viral replication occurs initially in mucosal CD4+ T cells , but within days high-level replication occurs in Peyer's patches , the gut lamina propria and mesenteric lymph nodes . Understanding the early events in HIV transmission may provide valuable information relevant to the development of an HIV vaccine . The viral quasispecies in a donor contracts through a genetic bottleneck in the recipient , such that , in low-risk settings , infection is frequently established by a single founder virus . Early-transmitting viruses in subtypes A and C mucosal transmission tend to encode gp120s with reduced numbers of N-linked glycosylation sites at specific positions throughout the V1-V4 domains , relative to typical chronically replicating isolates in the donor quasispecies . The transmission advantage gained by the absence of these N-linked glycosylation sites is unknown . Using primary α4β7+/CD4+ T cells and a flow-cytometry based steady-state binding assay we show that the removal of transmission-associated N-linked glycosylation sites results in large increases in the specific reactivity of gp120 for integrin- α4β7 . High-affinity for integrin α4β7 , although not found in many gp120s , was observed in early-transmitting gp120s that we analyzed . Increased α4β7 affinity is mediated by sequences encoded in gp120 V1/V2 . α4β7-reactivity was also influenced by N-linked glycosylation sites located in C3/V4 . These results suggest that the genetic bottleneck that occurs after transmission may frequently involve a relative requirement for the productive infection of α4β7+/CD4+ T cells . Early-transmitting gp120s were further distinguished by their dependence on avidity-effects to interact with CD4 , suggesting that these gp120s bear unusual structural features not present in many well-characterized gp120s derived from chronically replicating viruses . Understanding the structural features that characterize early-transmitting gp120s may aid in the design of an effective gp120-based subunit vaccine . Despite widely available prevention modalities against HIV transmission , 2 . 6 million individuals are newly infected with HIV every year . Thus , there exists an urgent need for an effective HIV vaccine . A number of studies that have focused on the earliest events in HIV transmission raise the possibility that new strategies for an effective vaccine immunogen can be developed . HIV transmission following mucosal exposure is inefficient[1] . The virus must first breach physical barriers in the mucosa , and then infect suitable target cells . In one study of heterosexual couples discordant for HIV infection , the frequency of transmissions per coital act averaged ∼0 . 01 [2] . One can therefore infer that , following deposition on the mucosal surface of the genital tract , HIV very frequently fails to establish infection . Both human and an SIV/macaque model studies indicate that during the first days of infection , termed the “eclipse phase” , low levels of viral replication occur , primarily in suboptimally activated memory CD4+ T cells in the genital mucosa[3] , [4] , [5] , [6] , [7] , [8] . Although these cells are metabolically active , they do not express classical activation markers[4] , [7] . Subsequently , HIV-1 infects fully activated memory CD4+ T cells . These events represent a critical point in transmission because they lead to high-level replication and the migration of virus into draining lymphoid tissue and ultimately gut-associated lymphoid tissue ( GALT ) where activated CD4+ T cells are plentiful , viral replication amplifies and the high level viremia that is associated with acute infection is established[9] . The best opportunity to prevent or abort establishment of HIV infection is likely during the eclipse phase following transmission , before HIV-1 migrates into the GALT . A striking feature of sexual transmission of HIV is the extreme restriction in the genetic diversity of the viral quasispecies shortly after infection . The genetically diverse viral swarm replicating in an infected transmitting partner constricts through a genetic bottleneck in the course of sexual transmission such that transmission is usually the result of a single infectious event; the productive infection that follows , in most instances , reflects an expansion from a single founder virus [10] , [11] , [12] , [13] , [14] , [15] . Importantly , this restriction occurs in the recipient , rather than in the transmitting partner . The early progeny of the transmitted founder virus show relative uniformity until adaptive immune responses drive the founder to diversify into a quasispecies . An instructive exception to this pattern occurs in recipients harboring certain sexually transmitted diseases ( STDs ) . Inflammation in the genital mucosa , mediated by STDs , can promote transmission of multiple founder viruses due predominantly to the ready availability of activated CD4+ T cell targets [10] , [12] , [16] , [17] , [18] . This underscores the crucial role that metabolically activated CD4+ T-cells likely play in transmission[19] . Considering that the early stages of HIV infection represent a possible window of opportunity for preventing infection , the structural , functional and immunogenic characteristics of founder/early-transmitting gp120s are highly relevant to the design of a preventive HIV vaccine . These viruses invariably utilize CCR5 , which is a key phenotype of early-transmitting gut-tropic isolates[20] , [21] . The genotype of early-transmitting viruses has also been a subject of great interest . Although a genotypic signature of early-transmitting viruses has proved difficult to identify , two key features have emerged . In studies of both heterosexual and mother to child transmission , early-transmitting gp120s have been found to be shorter in length , and encode fewer potential N-linked glycosylation sites ( PNGs ) than typical chronically replicating isolates[11] , [22] , [23] . These features have thus far only been found in the context of infection with HIV subtypes A and C . Length shortening has been observed in the V1/V2 region , as well as V4 and flanking regions of gp120 . PNGs absent from early-transmitting isolates ( PNGΔs ) appear in somewhat specific positions around the N- and C-terminal stems of V1/V2[11] , [22] and in C3/V4[24] . These characteristics apparently provide early-transmitting isolates with increased transmission fitness[22] , [24]; however , the nature of this fitness-advantage is unknown . The antigenic properties of early-transmitting isolates are the subject of ongoing studies [25] , [26] . V1/V2 , along with V4 and flanking regions are frequently an early target of autologous neutralizing antibodies ( Nabs ) [23] , [27] , [28] . The viral quasispecies escapes neutralization through amino acid substitutions , insertions/deletions ( INDELs ) , and also by adding/shifting PNGs . In this way the bulk of the quasispecies drifts away from the genotypic features that distinguish early-transmitting isolates . The V1/V2 domain of HIV-1 gp120 mediates binding to integrin α4β7 ( α4β7 ) on CD4+ T-cells[29] . α4β7 has been termed the gut homing integrin[30] . It is upregulated on lymphocytes in Peyer's patches and mesenteric lymph nodes , and then mediates , in concert with chemokine receptors , the homing of these lymphocytes into GALT through interactions with its natural ligands , MadCAM and VCAM , which appear on gut endothelial cells[31] . α4β7 appears in close association with the CD4 receptor on mucosal CD4+ T cells[32] . The biochemical characteristics of the interactions between α4β7 and gp120 closely mimic those of α4β7 with MadCAM and VCAM . There is a strong dependence of this interaction on divalent cations , and the α4β7 binding-site in V1/V2 shares close sequence homology with the binding sites of α4β7's natural ligands . This type of structural mimicry suggests that the specific affinity of gp120 for α4β7 provides increased fitness to HIV . However , unlike CD4 and CCR5 , α4β7 is not required for viral entry or replication in vitro . α4β7+/CCR5+/CD4+ memory T cells appear in the rectal and vaginal mucosa[32] , [33] . In this way , this subset of CD4+ T cells links the portal of entry during sexual transmission and the inductive and effector sites of the gut that provide a permissive environment for near-exponential replication . We have proposed that HIV evolved a specific affinity for α4β7 as a means of insuring that productive target cells with gut-homing potential will be infected shortly after transmission . We noted , however , that the α4β7-reactivity of gp120s varied widely among isolates we analyzed [29] , suggesting to us that strong α4β7-reactivity might provide increased transmission fitness over those isolates with lower α4β7 –reactivity in the context of mucosal transmission . In the present study we provide evidence that the apparent selection at the time of mucosal transmission for viral envelopes exhibiting transmission-linked PNGΔs coincides with dramatic increases in α4β7-reactivity . We conclude that α4β7-reactivity is one of the phenotypes that can contribute to the genetic restriction that occurs during mucosal transmission . Longitudinal studies of cohorts of couples discordant for HIV infection have been used to isolate and characterize genotypic , phenotypic and immunogenic properties of early-transmitting gp120s . In studies involving heterosexual transmission of subtype A and C viruses , early-transmitting viruses tend to encode more compact variable loops of gp120 with reduced numbers of PNGs relative to isolates replicating during the chronic phase of infection[11] , [22] , [23] . This pattern represents a bias rather than a rule , and although it is most frequently associated with the V1/V2 domain of gp120 , the V4 and flanking regions of early replicating viruses can also display these characteristics . In the V1/V2 domain , transmission-linked PNGΔs are clustered at the N- and C-termini of V1 and V2 respectively , while two central PNGs are well conserved ( Figure 1A ) . The apparent advantage that these characteristics confer upon early-transmitting viruses is not known . The V2 PNGs lie close to a tripeptide ( LDV/I – Figure 1A ) that mediates gp120 binding to α4β7 [29] . We determined whether removing PNGs in V1/V2 altered reactivity with α4β7 . Two recombinant gp120s , 92Ug037 ( subtype A ) , and 93MW959 , ( subtype C ) , were employed . Both envelopes were derived from viruses obtained from asymptomatic females who contracted HIV-1 through heterosexual transmission . 93MW959 was isolated ∼12 months post-seroconversion , while the time between seroconversion and isolation of 92Ug037 is unknown . PNGs were removed by site-directed mutagenesis in which asparagines were replaced with glutamines ( Figure 1A ) . α4β7-reactivity was measured using a steady-state binding assay that employs primary α4β7+/CD4+ T-cells ( Figure S1 ) . Each PNG mutant was compared to its wild type ( w . t . ) parent . For 92Ug037 , PNGΔs near the N-terminus of V1 or the C-terminus of V2 mediated increases in α4β7-reactivity of up to 20-fold ( Figure 1B ) . For 93MW959 , a single PNGΔ at the C-terminus of V2 mediated a ∼3 . 5-fold increase in α4β7-reactivity , while PNGΔs near the N-terminus of V1 had little effect on α4β7-reactivity ( Figure 1C ) . Length-shortening in V1/V2 is also associated with early transmitting gp120s . However , because those deletions often result in PNGΔs , it is difficult to determine whether short V1/V2s are favored independently of PNGΔs [34] . We examined the influence of V1/V2 length shortening on α4β7-reactivity by taking advantage of a previously characterized pair of subtype A envelopes isolated from an individual at ∼1-month and ∼41-months post-infection[23] . The month 1 V1/V2 ( QA203M1 ) encodes 63 amino acids and 5 PNGs , while the month 41 V1/V2 ( QA203M41 ) encodes 70 amino acids and 8 PNGs such that the 41 month envelope encodes two additional PNGs near the N-terminus of V1 and one additional PNG near the C-terminus of V2 ( Figure 2A ) . Both V1/V2s were grafted into a subtype A gp120 backbone isolated ∼1year post-infection from a second patient . QA203M1 gp120 displayed ∼20× greater α4β7-reactivity than did w . t . QA203M41 ( Figure 2B ) . To distinguish the influence of V1/V2 length from the influence of the number of PNGs on α4β7 reactivity we constructed a variant of the month-41 V1/V2 ( QA203M41variant1 ) lacking the two V1 PNGs that were missing in QA203M1 without changing its length ( Figure 2A ) . QA203M41 variant 1 , exhibited relatively strong binding to α4β7 that was nearly identical to that of QA203M1 gp120 . These results indicate that increased α4β7-reactivity mediated by the early-transmitting QA203M1 relative to QA203M41 was due to PNGΔs rather than to a shorter V1/V2 . We extended this analysis by removing additional PNGs from QA203M41 . QA203M41 variant 2 in which PNGs were removed from the C-terminal region of V2 mediated a small increase in α4β7-reactivity , while QA203M41 variant 3 , which combines variant 1 and variant 2 PNGΔs mediated an increase in α4β7-reactivity that was intermediate between variant 1 and variant 2 ( Figure 2A , C ) . These results demonstrate that PNGΔs do not necessarily enhance α4β7-reactivity in an additive manner . Taken together , analysis of all of the PNGΔs described above leads us to conclude that the removal of transmission-linked V1/V2 PNGs can mediate large increases in the α4β7-reactivity of both subtypes A and C gp120s . The pattern in which these increases were mediated is complex such that no single PNGΔ at a given position in V1/V2 mediated increased α4β7-reactivity in all three gp120s . PNGΔs near both the N- and C-termini of the V1/V2 of 92Ug037 increased α4β7-reactivity , while only one PNGΔ near the C-terminus of the V2 of 93MW959 increased α4β7-reactivity . For QA203M41 , the opposite was the case , removal of V1 PNGs , but not V2 PNGs mediated a large increase in α4β7-reactivity . These data suggest that increased α4β7-reactivity mediated by PNGΔs is due to changes in V1/V2 conformation rather than through steric occlusion , a subject that will be addressed below . A study of early-transmitting viruses following mother to child transmission reported that PNGs in the C3/V4 region of gp120 are also underrepresented in early replicating isolates in a manner similar to that described above for V1/V2 PNGs [24] . Overbaugh and colleagues found that PNGs at specific positions throughout C3/V4 were underrepresented in early-transmitting isolates of subtype A viruses isolated from infants shortly after birth . Although this region of gp120 is far removed from the known α4β7-binding site in V1/V2 , we determined whether removing these PNGs would also increase α4β7-reactivity . The C3/V4 region of 92Ug037 was aligned with subtype A sequences from the study noted above and five transmission-linked 92Ug037 PNGΔ gp120s were analyzed ( Figure 3A ) . 92Ug037N333Q , 92Ug037N362Q , and 92Ug037N393Q gp120s mediated increases in α4β7-reactivity of ∼18 to 21-fold . Interestingly , Ug037N355Q mediated a ∼27-fold increase in α4β7-reactivity . This increase was greater than that mediated by Ug037N144Q , which among the 92Ug037 V1/V2 PNGΔs mediated the largest increase in α4β7-reactivity . 92Ug037N385Q mediated an ∼8-fold increase in α4β7-reactivity ( Figure 3B ) . We conclude that , as with V1/V2 , PNGΔs in C3/V4 can mediate large increases in α4β7-reactivity . The increased α4β7-reactivity achieved by PNGΔs in C3/V4 supports the proposition that the manner in which glycans inhibit gp120 binding to α4β7 involves conformational changes in gp120 rather than simple steric occlusion . Data supporting this interpretation will be presented below . To better understand the role of glycan deletion in gp120-α4β7 interactions we compared the α4β7-reactivity of AN1 w . t . gp120 , a subtype B ancestral/consensus gp120[35] , expressed in three different cell lines known to glycosylate gp120 in different ways . AN1 w . t . was expressed in CHO S cells ( a nonadherent subclone of CHO K1 ) , the same cell line in which all of the gp120s described in the present study thus far were produced . gp120s expressed in CHO K1 cells present a heterogeneous pattern of oligo-mannose and complex carbohydrate type glycans [36] . Complex carbohydrates tend to appear on the solvent-exposed loops , including V1/V2 of recombinant gp120 proteins [37] . We also expressed AN1 gp120 in CHO lec1 cells , a CHO derivative that lacks N-acetylglucosamine ( GlcNAc ) glycosyl transferase activity , so that N-linked carbohydrate trimming is blocked at the Man5-GlcNAc2-Asn intermediate ( where Man is Mannose ) [38] . gp120s produced in CHO lec1 cells are devoid of complex carbohydrate , and are instead enriched with oligo-mannose type glycans . Finally , we expressed AN1 w . t . gp120 in 293F cells ( a nonadherent subclone of HEK 293T cells ) , which differ from both CHO cell lines in the manner in which it modifies complex carbohydrate . 293T derived cells sialylate the terminal galactose moieties of complex carbohydrates through both α-2 , 3 and α-2 , 6 linkages . CHO cells , which lack α-2 , 6-sialyltransferase , establish these linkages only at the α-2 , 3 position [39] . CHO S expressed AN1 w . t . gp120 reacted with α4β7 at an intermediate-low level ( Figure 4A ) , similar in magnitude to many of the gp120s that we have previously reported[29] . CHO lec1-derived AN1 w . t . gp120 reacted ∼100× more efficiently with α4β7 than did CHO S AN1 w . t . gp120 ( Figure 4A ) . It is likely that this increase in α4β7-reactivity results from the substitution of complex carbohydrates in V1/V2 with oligo-mannose type glycans . In contrast , 293F expressed AN1 w . t . gp120 showed no detectable α4β7-reactivity . We subsequently analyzed six additional 293F and T derived gp120s , including CAP881m . c17 , a gp120 that exhibits very strong α4β7-reactivity when derived in CHO S cells , and without exception they all exhibited low or undetectable α4β7-reactivity ( Figure S2 and data not shown ) . Considering the differential sialylation mediated by CHO S and 293 cells , we digested 293F AN1 w . t . gp120 with neuraminidase , which catalyzes the hydrolysis of terminal sialic acid residues; however , this failed to rescue α4β7-reactivity ( data not shown ) . We next expressed AN1 w . t . gp120 in 293F cells cultured in the presence of kifunensine , a mannosidase I inhibitor that restricts the processing of N-linked glycosylation beyond the Man9GlcNac2 intermediate and swainsonine , a mannosidase II inhibitor that restricts the processing of N-linked glycosylation beyond hybrid glycan intermediates [40] . Expressing gp120 in 293F cells in the presence of these drugs should , to some degree approximate the complex carbohydrate deficient glycan characteristics of CHOlec1 derived gp120 . Neither drug rescued α4β7-reactivity ( Figure 4A and data not shown ) . Although we do not know what the post-translational defect is in 293F and 293T-expressed gp120s , we conclude that utilizing gp120s produced in this manner may not be ideal for studies involving α4β7 . We next determined what effect restricting the glycans on AN1 w . t . gp120 to the oligo-mannose type , as occurs in CHO lec1 cells , would have compared to the effect of the deletion of transmission-linked glycans in AN1 derived from CHO K1 cells . Three AN1 PNGΔ derivatives: AN1 NN140/143QQ , AN1 N204Q , and the combined mutant AN1 NN140/143QQ , N204Q , were expressed in CHO-S cells ( Figure 4B ) . These PNGs appear near the N- and C-termini of V1 and V2 and their positions correspond to transmission-linked PNGs . Compared to AN1 w . t . expressed in CHO-S cells , AN1 NN140/143QQ increased α4β7-reactivity by ∼25-fold ( Figure 4A ) . AN1 N204Q increased α4β7-reactivity by ∼23-fold . The combined mutant NN140/143QQ , N204Q increased α4β7-reactivity by ∼96-fold , and resulted in a CHO-S derived gp120 with α4β7-reactivity ∼equivalent to CHO lec1 gp120 . To formally demonstrate that increased α4β7-reactivity can result from reduced N-linked glycosylation we carried out a brief digestions of CHOlec1 derived AN1 gp120 with endoglycosidase H ( Endo-β-N-acetylglucosaminidase H ) and determined that the enzymatic removal of PNGs from AN1 gp120 leads to increased α4β7-reactivity ( Figure S3 ) . Overall these results demonstrate that different glycosylation patterns mediated by different cells can exert a strong influence on α4β7-reactivity . In the case of AN1 gp120 , replacing complex carbohydrate with oligo-mannose type glycans mediates stronger α4β7-reactivity , but this same strong binding can be achieved by deleting a small number of PNGs in V1/V2 . A pseudotyped virus encoding the month 1 QA203M1 V1/V2 , described above ( Figure 2 ) , was efficiently neutralized by autologous serum taken at month 40 , while a pseudotyped month 41 QA203M41 virus escapes neutralization from this same serum[23] , indicating that neutralization-escape in the 40 month isolate was mediated by V1/V2 . This pattern of escape is somewhat typical of early- vs . chronic-replicating HIV-1 viruses and reflects the fact that in subtype A and C viruses V1/V2 is frequently a direct target of early autologous neutralizing antibodies[23] , [27] , [28] . Additionally , V1/V2 can mediate a second type of neutralization-escape that affects epitopes throughout gp120 . Mutations in V1/V2 can enhance a structural property of gp120 that has been termed conformational masking[41] , [42] , [43] , [44] . Both mechanisms of escape are mediated by a combination of sequence variation , INDELS , and the addition and/or position-shifting of PNGs . We questioned whether changes in gp120 leading to neutralization-escape would disrupt α4β7-reactivity . Two recent studies [27] , [28] , in which quasispecies evolution and autologous neutralizing Ab responses were followed longitudinally , beginning shortly after transmission , provide perhaps the best-defined examples of V1/V2-mediated escape . Rong , Derdeyn and colleagues[28] , demonstrated that the antibodies present in an HIV-1 subtype C infected female patient ( patient 205F ) at ∼38 months post-infection neutralized the early-replicating founder virus isolated within the first month following sexual transmission . Importantly , virtually all of the neutralizing activity in the 38-month sera targeted V1/V2 dependent epitopes . In a second study Moore , Morris and colleagues describe a more complex pattern of neutralization-escape[27] , in which the early-transmitting month 1 virus isolated from a female patient , CAP88 ( subtype C ) , was sensitive to autologous serum taken after 13 months , while viruses replicating at month 12 were not . Month 12 isolates were able to escape the month 13 sera via INDELS and PNG additions in both V1/V2 and C3 . We produced the neutralization sensitive 205F 0-month founder gp120 ( Z205F . ENV1 . 1 ) , and four escape mutants . The 0- and 8- month escape viruses are highly sensitive to neutralization by 38-month plasma and this escape was mediated by sequence changes that lie entirely within V1/V2 ( Figure S4 ) . The two 38-month viruses are partially resistant to neutralization by 38-month autologous plasma . The 205FENV1 . 1 0-month founder exhibited ∼8-fold greater α4β7–reactivity than did the 0-month escape , and at least 17-fold greater than did the 8- and 38-month escape gp120s ( Figure 5A ) . The level of α4β7-reactivity exhibited by the 205F . ENV1 . 1 0-month founder appeared to be substantially greater than many of the w . t . gp120s we previously reported[29] . We conclude that the high level of α4β7 reactivity in the 0-month founder was lost concomitant with sequence changes that mediated neutralization-escape . We next compared the α4β7-reactivity of CAP88 gp120s isolated at 1 month and 12 months post infection . We chose CAP88 . 1m . c17 , which reflects the predominant circulating isolate early in the first month post-infection . We also expressed the 12 month isolate CAP88 . 12m . c2 that diverged from CAP88 . 1m . c17 by adding PNGs in both V1/V2 and C3 at positions that correspond to transmission-linked glycans ( Figure S4 ) . The addition of these PNGs was shown to contribute to the resistance of CAP88 . 12m . c2 to autologous serum taken after month 13[27] . Like the 205F 0 month founder , CAP88 . 1m . c17 exhibited strong α4β7 reactivity ( Figure 5B ) . Surprisingly this high level of reactivity was maintained in CAP88 . 12m . c2 despite the fact that PNGs were added at transmission-linked PNG positions in both V1/V2 and C3 . This result demonstrates that viruses with relatively strong α4β7-reactivity can escape from autologous neutralizing antibodies without losing their α4β7-reactivity . Considering the complex pattern of changes in α4β7-reactivity we observed with the various N/Q substitutions described in Figures 1 , 2 and 4 , in which no single PNGΔ generated the same effect on each of the proteins analyzed , we conclude that the glycans that can mediate neutralization-escape may impact α4β7 reactivity ( e . g . 205F ) , but not in all gp120s , and not in a way that is easily predictable . In summary , we find that both of the early-replicating gp120s we analyzed showed high levels of α4β7-reactivity . Escape from neutralizing antibodies disrupted this activity in gp120s derived from patient 205F , but the α4β7-reactivity of a gp120 derived from patient CAP88 persisted 12 months post-infection despite sequence changes in V1/V2 and C3 that mediated escape from autologous neutralizing antibodies . The affinity of HIV-1 gp120s for CD4 varies over a wide range and these differences can influence the cell-tropism of a viral isolate . For example , high CD4 affinity facilitates replication in macrophages by compensating for the low density of CD4 appearing on the macrophage membrane[45] . Changes in CD4 affinity could theoretically impact the transmissibility of a viral isolate; however , studies of early replicating gp120s have not , to date , found any clear correlation between CD4 affinity and transmission fitness [25] . Numerous studies have , however , shown that both amino acid substitutions and glycan additions/deletions in V1/V2 can influence the conformation of gp120 in a global way [44] , [46] . Because V1/V2 plays an important role in α4β7 recognition we determined whether there was any relationship between α4β7-reactivity and CD4-reactivity . We first employed a steady-state binding assay to compare the reactivity of gp120s for α4β7 and CD4 . In this assay , retinoic acid-cultured CD4+ T cells were differentially masked with either a CD4 mAb or an α4 mAb , as described in supporting Figure S1 . For gp120s derived from early-transmitting viruses we found that α4β7 mediated a greater degree of binding to the cell surface than did CD4 . The amount of 205F 0-month founder gp120 bound to α4β7 was 37-fold greater than that bound to CD4 , but this differential disappeared in each of the 205F escape gp120s ( Figure 6A ) . Such differences could be mediated entirely by V1/V2 , as demonstrated by comparing the CD4- and α4β7-reactivities of the chimeric QA203M1 and QA203M41 gp120s , which are identical in all domains other than V1/V2 . 1-month QA203M1 bound 5-fold more to α4β7 than to CD4 while the 41 month QA203M41 was captured primarily by CD4 ( Figure 6B ) . The deletion of a single transmission-linked PNG was sufficient to achieve a binding profile in which more binding to the cell membrane was mediated by α4β7 than by CD4 . For example , while the majority of w . t . 92Ug037 binding to the cell surface was mediated by CD4 , PNGΔs at the N terminus of V1/V2 altered this pattern such that more binding was now mediated by α4β7 ( Figure 6C ) . Finally , we compared the two CAP88 gp120s to a panel that included several widely studied gp120s and subtypes B and C ancestral/consensus gp120s[35] , [47] . Similar to the 205F 0 month founder and the chimeric QA203M1 , both CAP88 proteins showed preferential binding to α4β7 over CD4 ( Figure 6D ) . In contrast to the CAP88 gp120s , all other gp120s in the panel , with one exception , exhibited CD4 binding that was equal to or greater than α4β7 binding . The one exception was SF162 , a highly neutralization-sensitive gp120 whose gut tropic characteristics have been well-documented[20] , [21] . Because levels of α4β7 expression on the cells we employed were equivalent to , or less than CD4 expression levels ( Figure 6 inset and Materials and Methods ) we can conclude that the steady-state affinity of gp120s , like the two CAP88 proteins , for Mn++ activated α4β7 is greater than their steady-state affinity for CD4 . In addition , this comparison underscores the strong α4β7-reactivity of the early-replicating gp120s we analyzed , relative to a number of well-studied gp120s . In this way , these gp120s appear to be better adapted to interact with CD4+ T cells that express a gut-homing receptor . It is noteworthy that the large changes in steady-state binding to α4β7 , mediated by sequence changes in V1/V2 , were coupled with relatively small changes in steady-state binding to CD4 ( Figure 6 ) . In some HIV-1 isolates , insensitivity to sCD4 and CD4 binding-site antibody neutralization is mediated by V1/V2 [44] , [46] , a phenomenon termed conformational masking[41] , [42] , [43] . Of note , this masking-effect can be modulated by the same transmission-linked PNGs that exert a strong influence on α4β7-reactivity [44] , [46] . To better understand the relationship between CD4-reactivity , V1/V2 masking , and α4β7-reactivity we employed a surface plasmon resonance-based kinetic CD4 binding assay . gp120s were immobilized on the surface of a biosensor chip and either monomeric sCD4 ( D1D2 ) , or D1D2-Ig αtp , a highly oligomerized ( dodecameric ) CD4-Ig derivative[48] , [49] , [50] , was passed over the surface , allowing us to measure reaction kinetics ( Figure 7A ) . In this format D1D2-Ig αtp can bind >1 gp120 in a near-simultaneous manner , allowing avidity-effects to contribute to interactions between it and gp120 . However , in this format , avidity cannot contribute to monomeric sCD4-gp120 interactions . Sensorgrams of nine well-studied gp120s , including JR-FL gp120 ( Figure 7B ) , and eight additional gp120s ( Figure S5 ) are provided for reference . 205FENV1 . 1 0-month founder gp120 failed to recognize monomeric CD4 , but did exhibit high affinity for dodecameric D1D2-Igαtp ( Figure 7C ) . The failure of this gp120 to react with monomeric CD4 distinguishes it from the nine standard gp120s we analyzed ( Figure 7B and Figure S5 ) . However , we observed the same phenomenon with the1-month QA203M1 chimeric gp120 ( Figure 7D ) , and CAP88 1m . c17 ( Figure 7E ) . In patient 205F and QA203 reactivity with monomeric sCD4 reappeared as the viral quasispecies evolved away from the early transmitting isolate . Both the 38 month 205F . ENV5 . 1 gp120 ( Figure 7C ) and 41 month QA203M41 ( Figure 7D ) , which bind weakly to α4β7 , did bind monomeric sCD4 with high-affinity . In summary , all three early-transmitting gp120s failed to react with monomeric sCD4 , which distinguishes them from nine standard gp120s . Because the chimeric QA203M41 and QA203M1 gp120s differ only in V1/V2 we conclude that the failure of QA203M1 to bind monomeric sCD4 was mediated by V1/V2 , and it seems likely that this also holds for 205F . ENV1 . 1 0-month founder , and CAP88 1m . c17 . All three of these gp120s did , however , react with dodecameric D1D2-Igαtp ( Figure 7 C , D , E ) , and with CD4 displayed on the surface of a T cell ( Figure 6 ) . The failure to interact with monomeric sCD4 in this SPR based assay does not indicate that the viruses from which these gp120s were derived are CD4-independent , nor can one conclude that these viruses would be resistant to sCD4 neutralization . This observation does indicate that the CD4-reactivity of these early-replicating gp120s is more dependent on avidity-effects than is the case for the other gp120s that we analyzed , and further suggests that , in addition to high-level α4β7-reactivity , these gp120s share structural features that distinguish them from many gp120s . Although these distinguishing biochemical features involve gp120 monomers , it is reasonable to suggest that these features will in some manner impact the stability and immunogenicity of trimeric spikes[51] . The search for a consistent genotypic signature of early-transmitting viruses has proven difficult . The most consistently observed genotypic marker thus far identified is a more compact V1–V4 with a reduction in the number of PNGs , relative to the length and average number of PNGs in chronically replicating isolates[11] , [22] , [23] . Early-transmitting viruses bearing this genotype do not appear to use CD4 or CCR5 in any way that distinguishes them clearly from chronically replicating isolates[25] , [52] . However , we found that removing transmission-linked PNGs from multiple gp120s consistently resulted in greatly increased α4β7-reactivity . Additionally , the early-transmitting gp120s that we analyzed showed notably higher levels of α4β7-reactivity than many of the chronically replicating isolates that we assayed . These results suggest that increased α4β7-reactivity is likely to be part of the phenotype underlying PNGΔs in some early- transmitting viruses and that this increased reactivity provides , under certain conditions , increased fitness in the process of transmission . Under this scenario , engaging α4β7 could provide an advantage at an early stage of transmission , but may not always be required . We should emphasize that additional analyses , that include greater numbers of early-transmitting gp120s are needed to better estimate the importance of α4β7+/CD4+ T-cells in mucosal transmission . Moreover , we do not yet know the specific point in the process of transmission where α4β7-reactivity may have an impact , nor do we know the relative importance of α4β7-reactivity on transmission fitness relative to other phenotypic features of gp120 . However , the fact that HIV has evolved an affinity for α4β7 reflects the important role that α4β7+/CD4+ T-cells likely play in the process of mucosal transmission . This is of potential importance since it follows logically that blocking viruses from infecting these cells will reduce the frequency of successful transmission . We analyzed only two early-transmitting gp120s and one early-transmitting chimera . Clearly , we cannot generalize from this small number , yet it is noteworthy that all three envelopes bind more efficiently to activated α4β7 than to CD4 . Thus , these gp120s are better adapted to interact with α4β7+/CD4+ T cells as opposed to α4β7lo-neg/CD4+ T cells . We previously reported that metabolically activated cells are enriched in the α4β7+ subset of CD4+ T cells in rectal mucosa[32] , and a similar observation has been made for cervical α4β7+/CD4+ T cells ( R . Kaul , personal communication ) . CD4 is not a marker of cellular activation or of gut-homing . When the affinity of a gp120 for CD4 is high , e . g . JR-FL gp120 , it engages all CD4 cells regardless of their metabolic state or homing potential . CCR5 expression can correlate with metabolic activity; however , prior to CD4-binding it is hidden from HIV . By increasing α4β7-reactivity , which is CD4-independent , we find that the early-transmitting gp120s that we analyzed are better adapted to engage with a specific subset of CD4+ T cells that are highly susceptible to infection by both HIV and SIV[32] , [53] , and located in anatomical sites relevant to HIV transmission . The context in which the interaction between HIV-1 and α4β7+/CD4+ T cells takes place remains unclear . It may involve cell-free virions , but it might also occur in the context of dendritic cell-CD4+ T cell interactions that are potentially important in mucosal transmission[54] . With one exception ( SF162 gp120 ) , none of the commonly studied gp120s that we characterized here or in our previous report exhibited this phenotype[29] . Although these standard gp120s may not necessarily be representative of chronic HIV isolates , it is noteworthy that several showed near-undetectable α4β7-reactivity . This is consistent with our previous observation that α4β7 interactions are not required for viral replication[29] and underscores the fact that α4β7-reactivity can diminish rapidly in a newly infected individual ( e . g . patient 205F ) . We note , however , that the introduction of a small number of transmission-linked PNGΔs into a gp120 derived from a chronically replicating virus ( e . g QA203M41 variant 1 ) can generate a strong α4β7-reactive phenotype such that this gp120 is now better adapted to engage CD4+ T cells that express α4β7 . Because these transmission-linked PNGΔs have been found in a somewhat consistent manner in early-transmitting gp120s in multiple studies involving transmission of subtypes A and C viruses [11] , [22] , [23] , [24] , [55] , it will be important in future studies to determine the extent to which strong α4β7-reactivity is a phenotype that is overrepresented in early-transmitting isolates . It will also be important to determine the frequency with which early-replicating viruses bind more efficiently to activated α4β7 than to CD4 . The frequency and circumstances under which viruses encoding fewer PNGs in V1–V4 are preferentially established in a newly infected individual provides clues concerning the possible role of gp120-α4β7 interactions in facilitating mucosal transmission . We note that transmission-linked PNGΔs have thus far been found only in cohorts of heterosexual couples discordant for HIV infection and a mother to child transmission cohort , all of which involve subtypes A and C , but not B[56] . The failure to observe this genotypic pattern in subtype B cohorts may reflect differences in the predominant modes of transmission i . e . heterosexual transmission versus transmission among men who have sex with men and intravenous drug users . It may also reflect other acquisition related risk factors , such that gp120-α4β7 interactions may play a greater role under conditions in which the risk of acquisition is low , but less of a role when that risk is high . One risk factor that can influence susceptibility to infection is the presence of STDs that cause inflammation of genital tissues in a recipient [10] , [12] , [16] , [17] , [18] . To the extent that inflammation increases the availability of activated CD4+ T cells near the site of infection , the selection pressure for a virus with strong α4β7-reactivity may be diminished . However , an opposite dynamic may also be operative under different circumstances . Chlamydia infection ( and possibly other STDs ) in the female genital tract has been shown to increase the number of antigen-specific α4β7+ T cells migrating through the female genital tract and into GALT[33] , [57] , [58] , [59] . Under these conditions , isolates that can bind α4β7 are more likely to engage activated CD4+ T cells . Further phenotypic characterization of α4β7+ T cells in mucosal tissues , as well as the early-transmitting viruses derived from different types of transmission cohorts , will be necessary to clarify the relationship between acquisition risk factors and α4β7-reactivity . We found that the type of glycans that decorate gp120 can exert a strong influence on α4β7-reactivity . By producing gp120 in a cell that fails to process oligomannose glycans into larger complex glycans we were able to increase the α4β7-reactivity of AN1 gp120 by >100-fold . In this regard , it is well established that different cell types , and even the same cell type , under different metabolic conditions , glycosylate proteins differently [60] , [61] . We know that gp120s produced in macrophages are glycosylated differently than the same gp120s produced in CD4+ T cells[62] . It follows then that α4β7-reactivity , and possibly the transmissibility of a virion , may be influenced by the type of cell in which that virion is produced in vivo . It would be useful if we could tailor our in vitro gp120 expression systems in a way that reflects in vivo glycosylation; however , although a recent study has characterized the glycan content presented on virion-associated spikes generated in primary PBMCs in vitro [63] , we do not yet know the true glycan profile of any in vivo derived HIV-1 gp120 . In addition to glycosylation , other post-translational modifications may also influence α4β7-reactivity . We found that gp120s produced in 293F and 293T cells showed greatly reduced α4β7-reactivity , which could not be explained entirely by differences in glycosylation . These observations may prove useful in evaluating the limitations of 293-derived gp120 for transmission-related studies . Overall these data indicate that post-translational modifications in gp120 have the potential to influence the transmissibility of HIV-1 , and that the type of carbohydrate that appears on an envelope may play a critical role in the selection of early-transmitting isolates . Although certain PNGs in V1/V2 may sterically interfere with gp120 binding to α4β7 , the manner in which transmission-linked PNGΔs throughout V1–V4 increased α4β7-reactivity suggests that increased α4β7-reactivity resulted from a change in the conformational state ( s ) of gp120 . In this regard , it is well established that V1-V4 glycans can influence gp120 conformation globally[64] , [65] . In addition , it is noteworthy that the naturally occurring gp120s that we analyzed exhibited either strong or weak α4β7-reactivity , since it suggests that there is a distinct conformational or structural feature associated with some early-transmitting gp120s that allows them to engage α4β7 in an efficient way . It is highly likely that sequence changes independent from PNGs will also influence α4β7-reactivity . In fact , the relatively high level of α4β7-reactivity exhibited by the 205F 0-month founder gp120 when compared to later isolates from this patient cannot be explained by a reduced number of PNGs . This underscores the limitation of using a genotypic feature i . e . the number of PNGs , to predict a phenotype . As more sequence information involving early-replicating isolates becomes available , and additional genotypic signatures are identified , it will be interesting to determine how they influence α4β7-reactivity . The number of gp120s analyzed here is small; therefore , it will be important to analyze the α4β7-reactivity of additional early-transmitting gp120s . Definition of the distinguishing features of early-transmitting gp120s is clearly relevant to subunit-based vaccine design . The influence that PNGs can exert on gp120 structure and immunogenicity has been well studied[66] . The transmission-linked glycans in V1/V2 can , when they are present on a gp120 , promote conformational masking . This term is used to describe the capacity of V1/V2 to increase the resistance of HIV-1 isolates to neutralizing antibodies with specificities throughout gp120[41] , [42] , [43] , [44] . This effect has also been referred to as “entropic masking” or “flickering” , both of which describe the property of envelope spikes to rapidly transition between alternate conformations . Rapidly alternating conformations has the net effect of reducing the apparent affinity of neutralizing gp120 antibodies . We find that PNGs , which have been shown to promote conformational masking [44] , [46] , are among those that can influence α4β7-reactivity . The manner in which high affinity for α4β7 is related to conformational masking is an important question that requires further investigation . In any case , our demonstration that the early-transmitting gp120s we analyzed ( 205F . ENV1 . 1 0 month founder , CAP88 . 1m and QA203M1 ) rely on avidity effects to engage sCD4 suggests that there are structural properties that distinguish these gp120s from many commonly studied gp120s . Finally , apart from the early-transmitting gp120s , strong α4β7-reactivity was infrequent among the panel of gp120s that we analyzed . This raises the question as to how frequently viruses bearing this phenotype appear in the viral quasispecies throughout the course of HIV disease . It may be that they appear spontaneously . This seems to be the case in patient 205F in which only the early-transmitting 0 month founder exhibited strong α4β7-reactivity . It is also possible that viruses bearing this phenotype are compartmentalized in mucosal tissues such that sampling virus from peripheral blood may lead to an underestimation of their frequency . In this regard , it is important to understand how frequently these viruses appear in the donor genital fluids from which virus is transferred to a recipient during sexual transmission . In conclusion , the transmission-linked PNGΔs that characterize some early-transmitting HIV-1 gp120s mediate a phenotype that includes increased α4β7-reactivity , suggesting that virions that react strongly with α4β7 may possess increased transmission-fitness . Further studies will be required to establish a direct link between α4β7-reactivity and increased transmission across mucosal surfaces . Our observations also suggest that α4β7+/CD4+ T-cells are an important target population in the process of transmission . The gp120s we analyzed that exhibited strong α4β7-reactivity interacted with sCD4 in a way that distinguishes them from many gp120s , and suggests that early-transmitting gp120s bear structural features that might be exploited in the context of a subunit vaccine . Additional genotypic , phenotypic and structural analyses of early-transmitting viral envelopes are essential . PBMCs were collected from healthy donors through a NIH Department of Transfusion Medicine protocol that was approved by the Institutional Review Board of the National Institute of Allergy and Infectious Diseases , National Institutes of Health . Informed consent was written and was provided by study participants and/or their legal guardians . Freshly isolated PBMCs were obtained from healthy donors and separated by Ficoll-Hypaque . Purified CD4+ T cells were obtained by negative selection using magnetic beads ( StemCell Technologies ) . Cultured CD4+ T cells were activated with OKT3 , IL2 ( 20 IU/ml ) and retinoic acid ( 10 nM ) unless otherwise specified . RA was obtained from Sigma Chemical and discarded 1-month after reconstitution . CHO lec1 cells and HEK293T cells were obtained from ATCC . CHO-S and 293F cells were obtained from Invitrogen . Integrin antibodies were purchased from BD Biosciences , Beckman Coulter and R&D . Leu3A ( SK3 ) and CD45RO were purchased from BD Biosciences . The genbank accession numbers of all gp120s used in this study are listed below . The QA203M1 and QA203M41 chimeric gp120s were constructed by inserting the V1/V2 sequences of each of these gp120s into a Q23 backbone . QA203M1: VTLECSNVNVTNNVTNDMGEEIKNCSFNMTTELRDKKQKTYSLFYKLDVVPFNNRSQYRLIN . QA203M41: VTLECSNVNVTNNVNVTNNVNVTNNVTNDMTGEIKNCSFNMTTELRDKKQKVYSLFYKLDVVPVNNNSSQYRLIN . All gp120 coding sequences were synthesized and codon-optimized for expression in mammalian cells ( DNA2 . 0 ) . All proteins were produced and purified in an identical manner unless noted otherwise . The mature coding sequences of each envelope protein , from +1 to the gp120-gp41 junction were inserted into a mammalian expression vector downstream of a synthetic leader sequence . Vectors were transiently transfected into either 293F or CHO-S cells ( Invitrogen ) using FreeStyle MAX Reagent ( Invitrogen ) per the manufacturers instructions . gp120s expressed in CHOlec1 cells were transfected by CaPO4 and stable cell lines were selected in media containing 1 mg/ml G418 . Clonal cell lines were established and subsequently seeded into hollow-fiber cartridges ( 30 kD MW cutoff ) ( Fibercell systems , Frederick MD ) . Protein containing supernatants were harvested daily from the extra-capillary space . Protein containing supernatants were harvested and passed over a column of galanthus nivalis lectin sepharose ( Vector Labs ) which was diluted 1∶5 with unliganded sepharose 4B to minimize avid binding . gp120 was eluted with 20 mM Glycine-HCl , pH 2 . 5 , 150 mM NaCl , 500 mM α-methyl-manno pyranoside ( Sigma ) , in 5 mL fractions directly into 1 mL M Tris-HCL , pH 8 . 0 . Low pH elution was found to be necessary with certain gp120s which bound too tightly to the lectin to be efficiently eluted using α-methyl-manno pyranoside . Peak fractions were pooled , concentrated with a stirred cell concentrator ( Millipore ) and dialyzed exhaustively against HEPES , pH 7 . 4 , 150 mM NaCl . Proteins were quantitated by UV adsorption at O . D . λ280 ( extinction coefficient 1 . 1 ) and values were confirmed by a bicinchoninic acid protein assay ( Pierce ) . Purified recombinant gp120s were biotinylated using amine-coupling chemistry . Proteins were reacted with a 100-fold molar excess of EZ-Link NHS-Biotin ( Pierce ) for 30 min , and reactions were quenched by rapid buffer-exchange into HBS . Biotin incorporation was determined by reacting gp120s with 4′-hydroxyazobenzene-2-carboxylic acid-avidin conjugates ( HABA ) per the manufacturers instructions ( Pierce ) . Protein preparations exhibiting a 1 . 0–1 . 2 mol/mole , biotin/gp120 incorporation were used in comparative semi-quantitative flow-cytometric binding assays . CD4+ T cells were cultured in RA for at least 6 days prior to use , and were stained with fluoresceinated anti β7 mAb FIB27 ( ATCC ) to insure RA-mediated upregulation of α4β7 . Receptor quantitation using anti-mouse IgG coated microbead analysis of CD4 and β7 expression levels of these cells from multiple donors indicates that levels of β7 expressed on the surface of these cells typically lower and less uniform than CD4 expression levels ( data not shown ) . The entire staining procedure , including wash steps was carried out in a 10 mM HEPES , 150 mM NaCl ( HBS Buffer ) buffer containing 100 µM CaCl2 and 1 mM MnCl2 . Cells were pre-blocked with normal mouse IgG and human IgG ( 5 µg each per 106 cells . 3×105 cells were stained in a volume of 50 µl on ice . Where indicated CD4-gp120 interactions were masked by preincubating cells for 15′ with 5 µg Leu3A ( SK3 ) ( Becton Dickinson ) . Where indicated α4β7-gp120 interactions were masked with 5 µg unlabeled α4 mAb HP2/1 ( Beckman Coulter ) . Masking antibodies were not washed away prior to gp120 staining . Biotin gp120 was added for 25′ on ice , after which cells were washed twice with staining buffer . In some assays CD45RO FITC and CCR5 APC were included , Neutravidin PE ( Pierce ) was then added , and incubation proceeded for an additional 30′ on ice . Cells were washed three times in staining buffer and then fixed in a 1% paraformaldehyde solution . Data were acquired using a BD FACSCalibur and mean fluorescence intensity measurements were obtained from the CD45RO+/CCR5+ gate . Analysis was performed on a Biacore 3000 instrument ( GE Life Sciences ) using CM5 sensor chips . The data were evaluated using BIAevaluation 4 . 1 software ( GE Life Sciences ) . The chip surface was activated by injecting 35 µl of a 1/1 mixture of 0 . 05 M N-hydroxysuccinimide and 0 . 2 M N-ethyl-N- ( dimethylaminopropyl ) carbodiimide at 5 µl/min . Purified gp120 ( 5 µg/ml in 10 mM NaOAc ( pH 5 ) ) was immobilized to a density of approximately 750 resonance units ( RU ) and blocked with 35 µl of 1 M Tris-HCl ( pH 8 . 0 ) . Human IgG was immobilized to one flow cell and used as a background control . Running buffer was HBS ( pH 7 . 4 ) , 0 . 005% Tween p20 . To evaluate Env: CD4 interactions , increasing concentrations of sCD4 ( D1D2 ) ( 6 . 25–400 nM ) were sequentially injected over surface-bound gp120 at a flow rate of 25 µl/min . After a 2 min injection of sCD4 , the surface was washed for 2 min in running buffer to follow the dissociation of sCD4 from gp120 . The surfaces were regenerated by injection of 4 . 5 M MgCl2 at a flow rate of 50 µl/min . The following gp120s were employed in these studies: 93MW959 ( GenBank accession # U08453 , R5-tropic ) , 92TH14-12 ( GenBank accession #U08801 , R5-tropic ) , 93Ug037 ( GenBank accession # U51190 , R5-tropic ) , AN1 gp120 [35] ( sequence available at http://ubik . mullins . microbiol . washington . edu/HIV/Doria-Rose2005/ , R5-tropic ) , 92Ug21-9 ( GenBank accession # U08804 , X4-tropic ) . ) , ADA ( GenBank accession #AF004394 , R5-tropic ) , YU-2 ( GenBank accession #M93258 , R5-tropic ) , NL4-3 ( GenBank accession # AF003887 , X4-tropic ) , 92Ug21-9 ( GenBank accession #AY669753 , X4-tropic ) , and SF162 ( GenBank accession # AY669736 , R5-tropic ) . Z205F . ENV1 . 1 0Mfounder ( GenBank accession #GQ485415 , R5-tropic ) Z205FENV6 . 3 0Mescape GenBank accession #GQ485419 , R5-tropic ) , Z205FENV2 . 3 8Mescape ( GenBank accession # GQ485425 , R5-tropic ) , Z205FENV5 . 1 38Mescape GenBank accession #GQ485445 , R5-tropic ) , Z205FENV4 . 1 38Mescape GenBank accession # GQ485447 , R5-tropic ) QA203M1 ( GenBank accession #DQ136332 , R5-tropic ) QA203M41 ( GenBank accession #DQ136341 , R5-tropic ) , Q23 ( GenBank accession # AF004885 , R5-tropic ) .
In the first days following sexual transmission , HIV replication occurs initially at relatively low levels in mucosal tissues because of a paucity of CD4+ T cell targets for the virus to infect . After a period of days , virus accesses specific gut tissues that are enriched in activated CD4+ T cells , where near-exponential replication ensues . The period of time before HIV accesses gut tissues represents a window of opportunity where a microbicide , pre-exposure and/or post-exposure antiretroviral prophylaxis or a vaccine-induced immune response could block infection . We previously reported that the HIV envelope protein gp120 binds to integrin α4β7 on the surface of CD4+ T cells . α4β7 mediates the homing of CD4+ T cells into the gut tissues where HIV can replicate exponentially . Here we report that the genotypic features that distinguish viruses isolated within the first month after infection , termed early-transmitting isolates , promote increased steady-state reactivity with α4β7 . This property likely provides these viruses with enhanced transmission-fitness . These results suggest that the infection of α4β7+/CD4+ T cells can play an important role early in HIV transmission . These findings have potentially important implications in the design of interventions to block the mucosal transmission of HIV .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "virology/virus", "evolution", "and", "symbiosis", "virology/vaccines", "virology/immunodeficiency", "viruses", "virology", "infectious", "diseases/viral", "infections", "infectious", "diseases/sexually", "transmitted", "diseases", "virology/host", "invasion", "and", "cell", "entry" ]
2011
The Genotype of Early-Transmitting HIV gp120s Promotes α4β7 –Reactivity, Revealing α4β7+/CD4+ T cells As Key Targets in Mucosal Transmission
Breaks at common fragile sites ( CFS ) are a recognized source of genome instability in pre-neoplastic lesions , but how such checkpoint-proficient cells escape surveillance and continue cycling is unknown . Here we show , in lymphocytes and fibroblasts , that moderate replication stresses like those inducing breaks at CFSs trigger chromatin loading of sensors and mediators of the ATR pathway but fail to activate Chk1 or p53 . Consistently , we found that cells depleted of ATR , but not of Chk1 , accumulate single-stranded DNA upon Mre11-dependent resection of collapsed forks . Partial activation of the pathway under moderate stress thus takes steps against fork disassembly but tolerates S-phase progression and mitotic onset . We show that fork protection by ATR is crucial to CFS integrity , specifically in the cell type where a given site displays paucity in backup replication origins . Tolerance to mitotic entry with under-replicated CFSs therefore results in chromosome breaks , providing a pool of cells committed to further instability . Accurate genome duplication is required at each cell generation to maintain genetic information . However , mammalian genomes contain regions that challenge the replication process , such as common fragile sites ( CFS ) . CFSs are loci that recurrently exhibit breaks on mitotic chromosomes following moderate slowing of replication fork movement [1] . To date , there is a consensus considering that such stresses delay completion of CFS replication more than the rest of the genome , and that breaks occur at under-replicated sequences upon chromosome condensation at mitotic onset . This delay was believed to result from replication fork blockage arising when forks encounter secondary structures formed at particular nucleotide sequences , notably AT-rich repeats [1] . However , the instability of FRA3B , the most active CFS in human lymphocytes , was recently shown to result from paucity in initiation events along a large region overlapping the most instable part of the site . This paucity forces replication forks emanating from flanking origins to cover long distances before merging in late S or G2 phases , leaving the sites incompletely replicated upon fork slowing [2] . Strikingly , FRA3B is weakly fragile in fibroblasts , in which initiation events are evenly distributed all along the locus [2] . Conversely , the two major CFSs in fibroblasts , that are not fragile in lymphocytes , display origin paucity in fibroblasts and a normal distribution of initiation events in lymphocytes [3] . Thus , the tissue-dependent organization of replication initiation controls the epigenetic setting of CFSs [4] . CFSs are a recognized source of the genomic instability driving oncogenesis from early steps of the process [5] . Indeed , CFS instability was repeatedly observed in pre-neoplasic lesions [5] , [6] , [7] . How pre-neoplasic cells , that generally retain wild-type checkpoints , escape surveillance by the DNA damage response ( DDR ) remains unclear . Central to DDR are two related protein kinases , ATM and ATR , that respectively sense double strand breaks ( DSB ) and RPA-coated single stranded DNA ( ssDNA ) accumulated upon fork slowing [8] . ATR and ATM activation then leads to phosphorylation of a large panel of substrates , including Chk1 and Chk2 , which triggers a second wave of phosphorylations that amplifies and spreads the signal [9] . Among these downstream targets is the major tumour suppressor p53 , a transcription factor that integrates signals from many different pathways [10] . Not surprisingly , inactivation of key DDR components leads to various diseases , including cancer [11] . In vertebrate cells , like in yeasts , the ATR/Mec1 pathway was mostly studied under conditions imposing a complete block to fork progression . Among other effects , such stresses lead , in cis , to stabilization of damaged replication forks and , in trans , to delayed mitotic onset [12] . In contrast , little is known about the cell response to moderate stresses such as treatments with low concentrations of aphidicolin , a well-known inhibitor of DNA polymerases , commonly used to induce breaks at CFSs . Several reports suggested that the frequency of breaks at CFSs increases in cells deprived of ATR , TopBP1 , Hus1 or Chk1 [4] . However , while the role of ATR has been largely confirmed , notably in vivo in human patients and in mutant mice [13] , [14] , the impact of other proteins , including Chk1 , in the maintenance of CFS integrity remains more controversial . Here we compared the response of human lymphoblastoid cells and normal fibroblasts to various levels of fork slowing . We showed that a two- to ten-fold reduction of fork speed ( called below moderate stress conditions ) leads to global chromatin recruitment of sensors and mediators of the ATR pathway without substantial activation of Chk1 , ATM or p53 . Analysis of the phenotype of cells depleted of ATR or Chk1 and submitted to moderate levels of stress shows that ATR , but not Chk1 , is crucial to fork protection and CFS integrity specifically in cell types where the site is fragile . These observations shed light on how pre-neoplastic cells continue cycling under inappropriate conditions . We used DNA combing to determine how increasing concentrations of aphidicolin impact fork movement in JEFF cells ( B lymphocytes immortalized by Eptsein-Barr virus ) and MRC-5 cells ( normal embryonic human fibroblasts ) ( Figure 1A ) . In untreated JEFF cells , forks progress at approximately 1 . 85 kb/min . In cells grown with aphidicolin 1 . 2 or 2 . 4 µM or with HU 1 mM , fork movement is too slow to be accurately measured with the labelling conditions we used . These treatments are considered below to block fork progression . For aphidicolin concentrations between 0 . 038 and 0 . 6 µM , the medians of fork speeds range between 1 and 0 . 2 kb/min ( Figures 1B upper panel and S1 ) , which is considered as moderate speed reduction . Similar results were obtained upon treatment of MRC-5 cells ( Figure S2A ) . The status of the ATR pathway was determined by western blot analysis of chromatin-bound ATR and RPA2 in cells treated for 4 h with different concentrations of aphidicolin or HU 1 mM ( Figure 1B lower panels ) . We found that chromatin loading of ATR starts to increase in cells treated with aphidicolin 0 . 075 µM and reaches a maximum at 0 . 15 µM , namely when fork speed is reduced approximately by a factor of two . Noticeably , the amount of chromatin-bound RPA2 also starts to increase upon treatment with aphidicolin 0 . 075 µM , but remains lower in cells treated with up to 2 . 4 µM of the drug than in cells grown with HU 1 mM . The study of MRC-5 cells also shows that a two-fold decrease in fork speed triggers recruitment of ATR to the chromatin and that , like in JEFF cells , ATR binding behaves essentially as an all-or-nothing phenomenon ( Figure S2B ) . A time course analysis of the status of the ATR pathway in JEFF cells treated with aphidicolin 0 . 6 µM shows that ATR , TopBP1 , Claspin , RPA2 and Rad9 , a subunit of the 9-1-1 complex , are rapidly loaded on the chromatin ( Figure 1C ) . Noticeably , Rad17 is loaded and phosphorylated on Ser645 , a recognized ATR phosphorylation site [15] . We observed that the amount of chromatin-bound proteins decreases upon prolonged treatment , a phenomenon previously observed by others [16] , [17] . Short treatments with aphidicolin 0 . 6 µM or HU 1 mM lead to comparable levels of recruitment for all proteins but RPA2 , which chromatin amount remains stable and lower in the presence of aphidicolin than in the presence of HU . Thus , chromatin loading of checkpoint sensors and mediators of the ATR pathway surprisingly appears poorly correlated to the amount of chromatin-bound RPA . To reinforce this conclusion , we studied ssDNA accumulation in JEFF cells treated as above using the procedure schematized in Figure 2A . As expected , no CldU labelling was observed in untreated JEFF cells ( Figure 2B ) . Following 1 h of treatment with HU 1 mM , 104 out of 110 ( 95% ) cells in S-phase , identified by the presence of PCNA foci , display CldU foci while cells not in S-phase remain unlabelled . In addition , these CldU foci generally co-localize with PCNA foci ( Figure 2C ) , showing that ssDNA forms at blocked forks . Strikingly , CldU foci were absent from the vast majority of S-phase cells following up to 16 h of treatment with aphidicolin 0 . 6 µM . Thus , under moderate speed reduction , the amount of sensors and mediators of the ATR pathway loaded on the chromatin is not proportional to the amount of RPA-coated ssDNA exhibited at the forks . Chk1 phosphorylation on Ser317 and Ser345 was analyzed ( Figure 3A ) . Both residues appear phosphorylated in JEFF cells treated with aphidicolin 2 . 4 µM or HU 1 mM but not , or weakly , in cells treated with up to 0 . 6 µM of aphidicolin . Chk1 status was confirmed by western blot analysis of extracts from S-phase cells ( Figure S3A ) and by immunofluorescence ( Figure S3B ) . Similar observations were made with MRC-5 cells ( Figure S2C ) . These results agree with some reports [17] , [18] but others found Chk1 phosphorylated in some cancer cell lines upon treatment with aphidicolin 0 . 6 µM [19] . These discrepancies most probably reflect differences in genetic backgrounds leading to cell-type variations in aphidicolin sensitivity . Unfortunately , the absence of fork speed measurement in previous works prevents further comparison of the results . In addition , we observed that phosphorylation of p53 on Ser15 and of RPA2 on Ser33 remains undetectable under moderate replication stress ( Figure 3B ) . Therefore , chromatin-bound ATR fails to trigger activation of several DDR effectors under these conditions . To further analyze the cellular response to replication stress , we studied the status of the ATM pathway ( Figure 4A ) . We found that a slight phosphorylation of ATM on Ser1981 starts to appear upon treatment with aphidicolin 2 . 4 µM , but remains much lower than in cells treated with HU 1 mM . Chromatin accumulation of ATM and phosphorylation of histone variant H2AX ( γH2AX ) occur only in cells treated with HU 1 mM . Time course analyses of ATM ( Figure S3C ) and H2AX ( Figures 4B and S3D ) in cells treated with aphidicolin 0 . 6 µM confirmed these observations . Thus , the ATM pathway is not triggered in cells experiencing moderate stress for up to 16 h . The fact that moderate stresses fail to activate the whole DDR cascade raises the question as whether mitotic onset is restrained under these conditions . The mitotic flow was determined following treatment with various aphidicolin concentrations and nocodazole to block mitotic exit , in DDR-proficient cell populations and in populations of cells depleted of either ATR or Chk1 by RNA silencing , ( Figure 5A ) . We found that the percentages of JEFF cells entering mitosis decrease upon treatment with increasing aphidicolin concentrations in the three genetic backgrounds and that only ATR depletion modestly impacts these percentages ( Figure 5B ) . Similar results were obtained in the presence of z-VAD fmk , a pan-caspase inhibitor that blocks apoptosis ( Figure S3E ) . Thus , the deficit in mitotic cells we observed upon aphidicolin treatment does not result from mitotic death but rather from delayed mitotic entry . We then determined how mitotic flow correlates with the degree of fork slowing . In good agreement with previous works [20] , [21] , we observed that ATR or Chk1 depletion “per se” reduces fork speed by a mechanism not yet elucidated . Not surprisingly , aphidicolin treatment further reduces fork movement ( Figure 5C ) . Plotting the percentage of mitotic cells against fork speed in the different conditions reveals a linear relationship between the two parameters , regardless of the transfection conditions ( Figures 5D and S3F ) . In addition , the curve obtained for Chk1-depleted cells aligns with that of control cells , indicating that the decrease in mitotic flow resulting from Chk1 depletion is completely accounted for by fork slowing . Therefore , Chk1 plays no direct role in the control of mitotic onset under these conditions , which agrees with the absence of Chk1 phosphorylation upon moderate fork speed reduction . The curve corresponding to ATR-depleted cells does not strictly align with the other two , which suggests that ATR plays a role , though modest , in the control of mitotic entry . The prominent mechanism that correlates mitotic flow to fork speed in these conditions is presently unknown . To evaluate the contribution of the ATR pathway to chromosome maintenance under moderate stress , we determined the percentage of metaphase plates displaying chromosome breaks in DDR-proficient cells and in cells depleted of either ATR or Chk1 following 16 h of treatment with various aphidicolin concentrations ( Figures 6A and S4A ) . We found that this percentage is the highest in cells depleted of ATR , although Chk1 depletion also alters chromosome stability . The percentage of metaphases displaying chromosome breaks was then plotted against fork speed for the different conditions of cell depletion and treatment ( Figure 6B ) . The curve obtained for Chk1-depleted cells is identical to that of control cells , indicating that the increase in chromosome breakage resulting from Chk1 depletion is completely accounted for by fork slowing . Strikingly , the curve for ATR-depleted cells coincides with the other two only when fork speed stands higher than 0 . 85 kb/min . Further slowing leads to percentages of cells with broken chromosomes greatly exceeding those observed in control or Chk1-depleted cells . Breaks at CFSs usually account for most of the breaks induced by low aphidicolin concentrations . We thus used fluorescent in situ hybridization ( FISH ) ( Figure S4B ) to determine the percentage of metaphase plates displaying chromosome breaks at FRA3B ( Figures 6C and 6D ) , the most active CFS in JEFF cells [2] . In either transfection conditions , breaks at FRA3B represent approximately 50% of total breaks and evolve like total breaks . Thus , ATR plays a major role in the maintenance of CFS stability when fork speed falls below a threshold of 0 . 85 kb/min . This value is consistent with the threshold for chromatin recruitment of ATR ( Figure 1B ) . Noticeably , this high level of CFS instability is not explained by the weak effect of ATR depletion on mitotic entry ( Figure 5D ) . To determine whether break frequencies in the different genetic contexts correlate with the presence of ssDNA at the fork , we studied cells displaying foci of ssDNA and/or PCNA in the three conditions of transfection ( Figure 7A ) . In the absence of aphidicolin treatment , we found that the percentage of cells displaying foci is low ( approximately 2% ) in either condition . Upon treatment with aphidicolin 0 . 3 µM , some 2% of DDR-proficient and 5% of Chk1-depleted cells exhibit CldU foci . This percentage reaches 15% in ATR-depleted cells and those foci strikingly do not co-localize with PCNA foci ( Figure 2D ) , suggesting that ssDNA takes place at collapsed forks . To determine whether these foci result from resection of collapsed forks by the Mre11-Rad50-Nbs1 ( MRN ) complex , the cells were depleted of Mre11 by RNA silencing or treated with mirin , an inhibitor of Mre11 nuclease activity [22] . Strikingly , compared to ATR depleted cells treated with mirin 100 µM or depleted of Mre11 , ATR-depleted cells treated with both aphidicolin 0 . 3 µM and mirin or co-depleted of ATR and Mre11 show similar percentages of cells with ssDNA foci ( Figure 7B ) . These results suggest that ssDNA foci we observed result from Mre11-mediated resection of collapsed forks . This conclusion agrees with the results of a recent work showing that MRN activity leads to the formation of RPA foci in checkpoint deficient U2OS cells [23] . We then asked whether ssDNA foci formed in ATR-depleted cells treated with aphidicolin 0 . 3 µM co-localize with CFSs ( Figure S4C ) . We found that FRA3B co-localizes more often with ssDNA foci ( 50% ) than non-fragile regions replicating late ( 25% ) or replicating early ( 15% ) ( Figure 7C ) . The relative co-localization of CFSs and ssDNA foci was also studied in MRC-5 fibroblasts where FRA3B is not fragile . In these cells the major CFS lies at 3q13 . 3 [3] . We found that this CFS also associates at higher frequency with ssDNA foci ( 45% ) than other tested sequences , including FRA3B that behaves like non-fragile and late replicating sequences ( Figure 7C ) . Noticeably , the non-fragile region identified by BACs 875H7 , that replicates early in JEFF cells and late in MRC-5 cells [24] , displays a percentage of co-localization with ssDNA foci that reflects its replication timing in each cell type , 12% and 24% respectively . The frequency of co-localization of a sequence with ssDNA foci is thus timing-dependent and , in the case of CFSs , correlates with their level of fragility in the tissue under study . Finally , we asked whether the preferential association of FRA3B with ssDNA foci in JEFF cells reflects the fact that forks collapse or stall more often along the site than in the bulk genome . Collapse and/or stalling should lead to asymmetrical forks , namely to individual forks presenting unequal IdU and CldU tracks [2] . Therefore , we calculated asymmetry as the ratio of the longest to the shortest track in cells treated or not with aphidicolin 0 . 3 µM and depleted or not of ATR or Chk1 ( Figures 7D and S5 ) . At the genome-wide level , fork asymmetry increases significantly only in cells depleted of ATR and treated with aphidicolin . We also studied fork asymmetry along FRA3B in the latter conditions . Strikingly , the distributions and the medians of fork asymmetries along FRA3B and in the bulk genome appear remarkably similar . We determine here how various levels of fork speed reduction impact DDR activation and genome integrity in human lymphoblastoid and fibroblastic cells . Surprisingly , ATR is loaded on the chromatin at roughly similar levels whether forks are completely blocked or only slowed by a factor of 2 to 3 . This loading thus behaves as an all-or-nothing phenomenon . Other sensors and mediators of the ATR pathway accumulate on the chromatin with similar kinetics while a modest increase in RPA-coated ssDNA is seen in cells experiencing moderate stress . These results suggest a model in which speed thresholds dictate the DDR status . Upon moderate fork impediment , stretches of RPA-coated ssDNA may be generated at a small fraction of the forks , which would escape detection by the techniques we used . Fork impediment may occur at random in the genome , or preferentially along regions difficult to replicate such as micro- and mini-satellites [8] , [25] , [26] , telomeric repeats [27] , [28] , [29] , [30] , [31] or highly transcribed genes [32] , [33] , [34] , [35] . We postulate that blocked forks elicit an alarm signal that commits cells to take steps against replisome disassembly , notably by recruiting sensors and mediators of the ATR pathway at all on-going forks . ATR contributes importantly to the stabilization of blocked forks in the yeast S . cerevisiae [36] , [37] , [38] and in higher eukaryotes [36] . We show here that ATR is also crucial to stabilize forks experiencing moderate slowing . Indeed , ssDNA foci form at high frequency in S-phase cells depleted of ATR and treated with low doses of aphidicolin . Those ssDNA foci do not co-localize with PCNA foci , which strikingly contrasts with the co-localization of ssDNA and PCNA foci we observed in DDR-proficient cells upon fork blockage . In addition , we observed that Mre11 depletion or Mre11 inhibition by mirin suppresses ssDNA foci formation in cells depleted of ATR and treated with aphidicolin . Together , these observations show that moderate stresses strongly impact fork stability in ATR-deficient cells , then resection of collapsed forks by the Mre11 nuclease activity gives rise to large amounts of ssDNA that form the foci we observe . Although phosphorylation of Chk1 , p53 or RPA2 does not occur in DDR-proficient cells submitted to moderate stress , the observation that Rad17 is phosphorylated indicates that chromatin-bound ATR is active . In addition , neither formation of γ-H2AX foci nor activation of ATM and/or Chk2 takes place under these conditions . How DDR is committed to complete activation remains unknown . Considering the many physiological situations that may lead to global or local replication fork slowing , such adaptation of the checkpoint response to the degree of stress might be crucial to cell proliferation , but this tolerance is detrimental to CFS integrity because mitotic entry with under-replicated sites triggers DNA breaks [39] . In agreement with the fact that Chk1 is not activated under moderate replication stress , the increase in chromosome instability observed in Chk1-depleted cells does not rely on its classical checkpoint function . Indeed , the frequencies of cells displaying ssDNA foci and breaks at CFSs are similar in checkpoint-proficient and in Chk1-depleted cells for similar fork speed reduction . In contrast , the frequencies of cells displaying ssDNA foci and chromosome breaks , notably at CFSs , increase considerably in cells depleted of ATR and submitted to moderate stress . This increase is accounted for neither by fork slowing nor by unscheduled mitotic entry . Noticeably , the specific increase in CFS breakage starts at the exact aphidicolin concentration that triggers chromatin recruitment of ATR in checkpoint-proficient cells . Together , these results confirm that ATR , and possibly other sensors and mediators of the pathway , plays a major role to prevent CFS instability under moderate fork slowing . These results do not fit with a previous model postulating that , upon replication stress , helicases tend to travel uncoupled from polymerases along CFS , giving rise to long stretches of ssDNA . In sub-regions able to adopt secondary structures , ssDNA would evolve into fork barriers that cause DNA breaks . In this model , checkpoint-proficient cells are supposed to be protected against these deleterious events because local accumulation of ssDNA triggers the ATR signalling pathway , resulting in delayed mitotic onset and activation of the repair machinery [1] . We show here that under moderate stress conditions , similar to those used to induce breaks at CFSs , forks stall and/or collapse in ATR-deficient cells at the same frequency along FRA3B as in the bulk genome . These results support data previously obtained in checkpoint-proficient cells showing that forks do not encounter sequence-specific obstacles along the sites [2] . In addition , we show here that , in ATR-deficient cells , ssDNA foci result from Mre11-dependent resection of collapsed forks , suggesting that long stretches of ssDNA are a consequence rather than a cause of CFS instability . Together , these results strongly argue against the model above . In contrast , the remarkable sensitivity of CFSs to moderate replication stress in ATR-deficient cells is well explained by paucity in initiation events along large regions nested in the sites [2] , [3] . Indeed , the lack of backup initiation events might prevent fast rescue of forks collapsing at random and normal frequency along CFSs , favouring extensive resection . This conclusion is supported by the fact that co-localization of ssDNA and CFSs increases specifically in the cell type in which a given site displays paucity in initiation events , namely FRA3B in lymphocytes and 3q13 . 3 in fibroblasts [2] , [3] . Tumor suppressor genes are rarely mutated in sporadic cancers , notably in preneoplasic lesions [5] . Surprisingly , chromosome instability has been repeatedly observed in cells of these lesions , while DDR activation should block their proliferation . We show here in two cellular models that a moderate reduction of fork speed does not fully activate the DDR , fails to block mitotic entry and elicits breaks at under-replicated CFSs . Strikingly , some of the genes hosted by CFSs behave as tumor suppressor genes in human and mouse models [40] . FRA3B , for example , hosts the FHIT gene that has been involved in the cellular response to DNA damage . Thus , FRA3B expression not only triggers local instability but also favors genome-wide DNA damages through FHIT inactivation [41] , [42] . The origin of the replication stress leading to CFS expression in pre-neoplastic lesions is still debated . It has been proposed to result from oncogene activation [7] , from alteration of metabolic process or from exposure to environmental stress [42] , [43] . Whatever the cause of CFS instability , we propose that imperfect repair of these damages in cells of pre-neoplastic lesions creates a pool of cells committed to further instability , from which the selection of cells with mutations affecting genes causally implicated in cancer development is facilitated . Lymphoblastoid cells were grown in RPMI 1640+GlutaMAX-I medium ( GIBCO ) and MRC-5 cells in MEM plus Earle's salts without L-glutamine medium ( GIBCO ) , 1% MEM nonessential amino acids ( GIBCO ) , 1 mM sodium pyruvate ( GIBCO ) and 2 mM L-glutamine ( GIBCO ) . In addition , all cells were grown with 10% foetal calf serum ( PAN-Biotech GmbH ) and 100 µg/mL of penicillin and streptomycin ( GIBCO ) . For transfections , 2×106 lymphoblastoid cells were resuspended in 100 µL of Nucleofector C solution ( Lonza Cologne AG ) with 0 . 6 µM RNAi and transfected with the Z-001 program according to manufacturer's instructions . MRC-5 cells were resuspended in Nucleofector R solution ( Lonza Cologne AG ) and transfected with the V-020 program . A mixture of 3 RNAi directed against ATR ( HSS100876 , HSS100877 , HSS100878 , Invitrogen ) or Chk1 ( HSS101854 , HSS101855 , HSS101856 , Invitrogen ) was used for transfection . The AllStars Negative Control siRNA ( 1027281 , QIAGEN ) was used for NONsi . Combing was performed as described [2] , [44] , [45] . An epifluorescence microscope ( Axio Imager . Z2; Carl Zeiss ) equipped with a 63× objective lens ( PL APO , NA 1 . 4 Oil DIC M27 ) connected to a charge-coupled device camera ( Cool-SNAP HQ2; Roper Scientific ) , and MetaMorph software ( Roper Scientific ) was used for image acquisition . Metaphase spreading and DNA fluorescent in situ hybridization for FRA3B detection were performed as described [39] . BACs were selected from the human genome project RP11 library . FRA3B has been assigned to band 3p14 . 2 and was probed with BAC 641C17 . Cells were spread on slide with a cytospin ( Shandon ) and processed for immunostaining as previously described [46] . Combined immunofluorescence and DNA fluorescent in situ hybridization was performed according to Chaumeil et al . [47] . Microscopic images were acquired using an upright motorized microscope ( Axio Imager . Z2; Carl Zeiss ) . Acquisitions were performed using an oil immersion objective 100× ( PL APO , NA 1 . 46 Oil DIC M27 ) and a high-sensitive cooled interlined CCD camera ( Cool-SNAP HQ2; Roper Scientific ) . For colocalization studies , rapid and precise Z-positioning was accomplished by a piezoelectric motor ( P-725 . 1CD ; Physik Instrumente ) mounted underneath the objective lens . Image stacks were acquired without camera binning , with a plane spacing of 0 . 2 µm , and MetaMorph software ( Roper Scientific ) was used for image acquisition . Chromatin fractionation was performed as described [48] . For total extract of cells in S-phase , DNA was stained with 5 µg/ml Hoechst 33342 ( Molecular Probes ) . Cells were sorted with a standard FACSVantage DiVa ( Becton Dickinson Immunocytometry Systems , San Jose , CA ) equipped with a 488 nm laser used at 250 mW and a multiline UV ( 351–363 nm ) laser used at 200 mW . Linear Hoechst fluorescence was acquired using a 424/44 filter . Doublets were excluded using pulse area vs width . 105 sorted cells were washed in cold PBS , re-suspended in 50 µL 1X SDS sample buffer ( New England BioLabs Inc . ) , sonicated 5 min ( Bioruptor - Diagenode ) and boiled before loading for western blotting . Cells were treated with aphidicolin and nocodazole as indicated and washed with buffer GM ( 1 . 1 g/L Glucose , 8 g/L NaCl , 0 . 4 g/L KCl , 0 . 37 g/L Na2HPO4 , 0 . 15 g/L KHPO4 , 0 . 5 mM EDTA ) . The cells were re-suspended in 1 mL of buffer GM and fixed by addition of 3 mL of ethanol 100% . The cell pellet was incubated with a mouse anti-MPM2 ( Mitotic Protein Monoclonal 2 ) antibody ( 05-368 , Upstate ) in buffer PBT ( PBS 1X , 1% BSA and 0 . 05% Tween20 ) , washed with PBS and incubated with a goat Alexa488-conjugated anti-mouse antibody ( Molecular Probes ) in PBT . Cells were washed with PBS and finally re-suspended in PBS containing 50 ng/µL propidium iodide and 40 µg/mL RNase A ( USB ) . The quality of the mitotic cell preparations was verified by FACScan analysis of MPM2-positive cells with CellQuest software ( Beckton Dickinson ) . The following primary antibodies were used in this study: mouse anti-PCNA ( PC10 ) antibody ( MAB424 , Chemicon International ) , goat anti-ATR ( N-19 ) antibody ( sc-1887 , Santa Cruz Biotechnology , Inc . ) , rabbit anti-Claspin antibody ( A300-266A , Bethyl Laboratories , Inc . ) , mouse anti-Rad17 ( sc-17761 , Santa Cruz Biotechnology , Inc . ) , rabbit anti-phospho-Rad17 ( Ser645 ) antibody ( #3421 , Cell Signaling Technology ) , rabbit anti-Rad9 ( sc-8324 , Santa Cruz Biotechnology , Inc . ) , goat anti-Hus1 antibody ( sc-30543 , Santa Cruz Biotechnology , Inc . ) , rabbit anti-RPA32 antibody ( GTX70258 , GeneTex , Inc . ) , mouse anti-Chk1 antibody ( sc-8408 , Santa Cruz Biotechnology , Inc . ) , rabbit anti-phospho-Chk1 ( Ser317 ) antibody ( #2344 , Cell Signaling Technology ) , rabbit anti-phospho-Chk1 ( Ser345 ) antibody ( #2348 , Cell Signaling Technology ) , rabbit γ-H2AX ( phospho S139 ) antibody ( ab2893 , Abcam ) . In immunochemistry experiments , primary antibodies were visualized with Alexa594-conjugated goat anti-rabbit and Alexa488-conjugated goat anti-mouse antibodies ( Molecular Probes ) . Detection of ssDNA was performed as previously described [49] , with a rat anti-BrdU ( OBT0030 , AbD Serotec ) and Alexa594-conjugated goat anti-rat antibodies ( Molecular Probes ) .
Accurate genome duplication is crucial at each cell generation to maintain genetic information . However , replication forks routinely face lesions on the DNA template and/or travel through sequences intrinsically difficult to replicate , such as common fragile sites ( CFS ) . To help the fork to proceed , the cells have evolved the DNA damage checkpoint that senses different types of damage and triggers well-adapted cellular responses . We have studied the DNA damage response of human lymphoblastoid cells and normal fibroblasts to various levels of fork slowing . We showed that a two- to ten-fold reduction of fork speed leads to global chromatin recruitment of sensors and mediators of the ATR pathway without substantial activation of Chk1 , ATM or p53 . Analysis of the phenotype of cells depleted of ATR or Chk1 and submitted to moderate levels of stress shows that ATR , but not Chk1 , is crucial to CFS integrity . We propose a model explaining how fork speed thresholds direct fine-tuned checkpoint responses that protect genome integrity without blocking cell cycle progression upon moderate replication fork impediment . Tolerance to mitotic entry with under-replicated CFSs therefore results in chromosome breaks , providing a pool of cells committed to further instability .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "mutagenesis", "cellular", "stress", "responses", "cancer", "genetics", "genetic", "mutation", "dna", "replication", "nucleic", "acids", "genetics", "dna", "dna", "repair", "biology", "dna", "recombination", "molecular", "cell", "biology" ]
2013
Stepwise Activation of the ATR Signaling Pathway upon Increasing Replication Stress Impacts Fragile Site Integrity
Unlike for other retroviruses , only a few host cell factors that aid the replication of foamy viruses ( FVs ) via interaction with viral structural components are known . Using a yeast-two-hybrid ( Y2H ) screen with prototype FV ( PFV ) Gag protein as bait we identified human polo-like kinase 2 ( hPLK2 ) , a member of cell cycle regulatory kinases , as a new interactor of PFV capsids . Further Y2H studies confirmed interaction of PFV Gag with several PLKs of both human and rat origin . A consensus Ser-Thr/Ser-Pro ( S-T/S-P ) motif in Gag , which is conserved among primate FVs and phosphorylated in PFV virions , was essential for recognition by PLKs . In the case of rat PLK2 , functional kinase and polo-box domains were required for interaction with PFV Gag . Fluorescently-tagged PFV Gag , through its chromatin tethering function , selectively relocalized ectopically expressed eGFP-tagged PLK proteins to mitotic chromosomes in a Gag STP motif-dependent manner , confirming a specific and dominant nature of the Gag-PLK interaction in mammalian cells . The functional relevance of the Gag-PLK interaction was examined in the context of replication-competent FVs and single-round PFV vectors . Although STP motif mutated viruses displayed wild type ( wt ) particle release , RNA packaging and intra-particle reverse transcription , their replication capacity was decreased 3-fold in single-cycle infections , and up to 20-fold in spreading infections over an extended time period . Strikingly similar defects were observed when cells infected with single-round wt Gag PFV vectors were treated with a pan PLK inhibitor . Analysis of entry kinetics of the mutant viruses indicated a post-fusion defect resulting in delayed and reduced integration , which was accompanied with an enhanced preference to integrate into heterochromatin . We conclude that interaction between PFV Gag and cellular PLK proteins is important for early replication steps of PFV within host cells . As obligatory intracellular parasites , retroviruses depend on host cell machineries for successful replication . Over the years , many cellular proteins counteracting or aiding retroviral replication have been identified [reviewed in 1 , 2] . Expectedly , a large number of these factors are capsid protein interaction partners [reviewed in 3 , 4 , 5] . Since retroviral Gag proteins orchestrate numerous intracellular trafficking steps and viral budding , they naturally recruit numerous cellular factors and machineries to accomplish their functions . Similar to its orthoretroviral relatives , prototype foamy virus ( PFV ) , the best-studied member of the genus spumavirus in the Spumaretrovirinae subfamily , relies on Gag interactions with host cell proteins for successful replication [reviewed in 6] . However , unlike for orthoretroviruses such as human immunodeficiency virus type 1 ( HIV-1 ) , few PFV Gag interaction partners have been identified to date . This is surprising , given the peculiarities of the PFV replication cycle and their significance to understanding retroviral evolution and vector development [reviewed in 7 , 8] . To our knowledge , only six host factors that interact either directly or indirectly with PFV Gag have been identified , but given the multitude of roles PFV Gag accomplishes for viral replication in cells , this list is clearly far from complete . The tumor susceptibility gene 101 ( Tsg101 ) component of the ESCRT-I complex directly interacts with primate FV Gag proteins to promote efficient virus budding [9 , 10] . Furthermore , direct or indirect interactions with actin and core histones ( H2A and H2B ) were identified upon co-precipitation of these cellular proteins with ectopically expressed PFV Gag [11 , 12] . These interactions are thought to be involved in PFV Gag precursor protein processing regulation and facilitate docking of the PFV pre-integration complex ( PIC ) to host cell chromosomes , respectively . FVs are also known to be restricted in a species-specific manner by cellular Trim5α proteins [13–15] . The viral determinants essential for restriction were mapped to the N-terminus of FV Gag , which probably engages Trim5α in the context of the viral capsid . In addition , a member of the DEAD-box helicase family , DDX6 , was found to co-localize with PFV Gag at assembly sites and aid in viral RNA encapsidation into newly assembling FV particles [16] . Finally , a potentially direct interaction of dynein light chain 8 with a coiled-coil domain within the N-terminus PFV Gag was reported , which appears to facilitate transport of incoming capsids to the centrosome of host cells [17] . At the centrosome of non-dividing cells PFV capsids can remain intact and biologically active for several weeks [18] . PFV capsids seem to proceed only to further disassembly steps upon entry of the cell into mitosis , which is required for PFV genome integration into cellular chromatin . The cues and machineries coordinating viral uptake and intracellular trafficking are incompletely characterized . Until now , it has only been speculated that members of cell cycle regulatory pathways may aid these processes . One family of mitosis promoting proteins is the polo-like kinases ( PLKs ) , which are direct downstream effectors of cyclin/cyclin-dependent kinase complex ( Cyclin/CDK ) cascades . PLK proteins are cellular Ser/Thr protein kinases that in vertebrates comprise five paralogues , PLK1-5 [reviewed in 19 , 20 , 21] . They are crucial regulators of cell cycle progression , centriole duplication , mitosis , cytokinesis and DNA damage response . However , other functions unrelated to cell cycle control such as protein aggregation and cellular stress response have also been attributed to various PLKs [22 , 23] . Mammalian PLK proteins are characterized by a modular structure ( Fig 1A ) , comprising an N-terminal ( Ser/Thr ) catalytic kinase domain ( KD ) and a C-terminal polo-box domain ( PBD ) . The PBD , which harbors one or more polo-boxes ( PBs ) that coordinate substrate binding , regulates catalytic activities and subcellular localization of PLKs [24–26] . With the exception of PLK4 , which contains a unique PBD organization [27] , the PLK PBDs facilitate binding to phosphopeptides with an optimal Ser- ( pThr/pSer ) -Pro ( S- ( pT/pS ) -P ) substrate binding motif [28 , 29] . The best-studied member of the PLK family is PLK1 , which typically utilizes its PBD to target proteins that have been primed through phosphorylation at their PLK binding site by other cellular kinases , such as CDKs or MAPKs [29 , 30] . However , there are a few cases known where the PLK1 PBD binds non-phosphorylated peptides or phosphopeptides that were previously phosphorylated by PLK1 itself [30 , 31] . Thus , multiple mechanisms exist to modulate the function and interaction of PLK1 and other PLKs and their targets in space and time . Although replication of various viruses is influenced by cellular PLKs and viruses have been reported to modify the activity of PLKs for their replication , only a few direct interactions of viral proteins with PLKs have been reported . For example the P protein component of the viral RNA dependent RNA polymerase ( vRdRp ) of paramyxoviruses , such as parainfluenza virus 5 ( PIV5 ) or mumps virus ( MuV ) , harbors a PLK1 STP binding site and is phosphorylated at another site by PLK1 [32 , 33] . PLK1-mediated P protein phosphorylation downregulates viral gene expression and thereby avoids efficient induction of the host innate immune response [34] . Furthermore , the interaction of the E2 protein of human papillomavirus type 5 with PLK1 was reported to inhibit Brd4 phosphorylation by PLK1 , thereby interfering with cellular functions of Brd4 in promoting cell cycle progression from the G1 to the S phases [35] . Finally , binding of PLK1 to the NS5A protein of hepatitis C virus and its hyperphosphorylation by PLK1 is important for efficient virus replication [36] . To gain further insight into the cellular biology of PFV replication and to identify new potential novel Gag interactors , we conducted a yeast two-hybrid ( Y2H ) screen using Gag protein as bait against different human cDNA expression libraries . In addition to the known PFV Gag interactor Tsg101 , a member of the PLK family was identified as a new hit . A detailed characterization and analysis revealed that PFV Gag specifically interacts with cellular PLK1 and PLK2 and alters their intracellular distribution . PFV particles , unable to interact with cellular PLKs upon virus entry , display an infectivity defect manifested by a delayed and reduced capacity to integrate viral genomic DNA ( vgDNA ) into host cell chromatin . To elucidate intracellular steps of PFV replication in more detail , we performed a Y2H screen , aiming to identify novel Gag protein interaction partners . Full-length ( FL ) PFV Gag ( PG ) was used as bait , fused either to the N- or C-terminus of the GAL4 DNA binding domain ( DB , here referred to as pNGB PG/coPG and pCGB PG/coPG , respectively ) ( Fig 1B ) . These bait constructs were used and tested for interaction with a normalized HeLa cDNA expression library and two different full-length human open reading frame ( ORF ) expression libraries , where prey proteins were fused to the C-terminus of the GAL4 activation domain ( AD , pGADT7 , here referred to as pCGA ) . Several human proteins were identified as likely PFV Gag interactors . Among them was the known PFV Gag interactor , Tsg101 ( 7 different fragments , identified a total of 51 times ) , previously shown to interact with PFV Gag in Y2H experiments [10] . A potential new PFV Gag interactor identified multiple times in the screen was human PLK2 ( hPLK2; 7 different fragments spanning the C-terminal PBD , identified a total of 20 times ) . We note that the hPLK2 PBD scored in the assay only in combination with the C-terminally tagged FL PFV Gag protein bait ( pNGB PG/coPG ) , whereas Tsg101 was identified with both PFV Gag baits . Additional Y2H experiments were performed to validate the results , investigate whether PFV can interact with other members of the PLK family , and characterize essential interaction domains/motifs within both proteins . The analyses included various bait constructs encoding FL PFV Gag as well as point mutants or N- and C-terminal truncations , fused to the N- or C-terminus of the GAL4 DB ( Fig 1B ) . Prey constructs encoded FL ORFs of hPLK1-3 , 5 , mouse PLK5 ( mPLK5 ) , rat PLK2 ( rPLK2 ) and point mutants thereof , as well as the PBDs of hPLK1-4 , fused to the C-terminus of the GAL4 AD ( pCGA-PLKs ) ( Fig 1A and 1B ) . A prey encoding the FL Tsg101 ORF was utilized as positive control , whereas negative controls included the empty bait ( pNGB , pCGB ) and prey ( pCGA , pNGA ) vectors . The results of the Y2H analysis are summarized in Fig 2 and S1 Fig FL PFV Gag interacted with FL hPLK2 and 3 , rPLK2 as well as PBDs of hPLK1-3 ( Fig 2A and 2B ) . FL PFV Gag did not interact with FL hPLK1 , though PFV Gag baits with C-terminal deletions of 104 or more residues did interact with hPLK1 ( Fig 2A; S1A and S1B Fig ) . Interactions with other tested PLK proteins were not detected ( Fig 2; S1 Fig ) . These results indicate that FL PFV Gag can interact with various hPLKs and that these interactions are mediated via the kinase PBDs . The use of N- and C-terminal Gag bait truncations allowed characterization of a minimal essential interaction domain spanning amino acids 130–295 ( Fig 2A; S1 Fig ) . A closer inspection of this region revealed the presence of a Ser-Thr-Pro ( S-T-P ) motif at position 224–226 , which was unique within the FL PFV Gag protein ( Fig 1C ) . Interestingly , this motif , which matches the minimal consensus PLK recognition and binding site ( S-T/S-P ) [29] , is present in all primate FV Gag proteins ( Fig 1C ) . Exchange of any of the residues of the PFV Gag STP motif in context of the FL protein , including inactivating ( iSTP ) and phosphomimetic ( pmSTP ) mutations at T225 , abolished interactions with all PLK preys that interacted with the wild type ( wt ) Gag protein , without affecting the interaction with Tsg101 ( Fig 2A and 2B ) . Hence , a preserved PFV Gag STP motif seems to be essential for interaction with cellular PLK proteins . In order to determine which domains and enzymatic properties of the PLK proteins are important for the interactions with PFV Gag , we focused on FL rPLK2 ORFs with point mutations in critical residues as preys ( Fig 2B; S1C Fig ) . As expected , none of the iSTP or pmSTP Gag constructs interacted with any of these rPLK2 variants , confirming interaction dependency on the PFV Gag STP motif . FL wt Gag protein interacted with wt and constitutively active T236E ( caKD ) rPLK2 , whereas no interaction was detected with kinase inactive K108M ( iKD ) rPLK2 or its PBD mutant W504F ( iPBD ) ( Fig 2B ) . These results suggest that the kinase activity of rPLK2 is required for stable interaction with FL PFV Gag , which is mediated by the rPLK2 PBD . Examination of the amino acid sequence of the other PFV structural proteins revealed the presence of two additional consensus STP peptide motifs in Pol but none in the cytoplasmic domains of Env . Both putative Pol PLK binding sites are located in the integrase domain of Pol surrounding T961 and T1058 ( S2A Fig ) . However , an extensive Y2H interaction analysis using bait constructs expressing either a PFV Pol precursor protein with enzymatically inactive PR or the mature integrase subunit failed to demonstrate signs of interaction with any of the PLK prey constructs that showed interactions with PFV Gag ( S2B Fig ) . Principal protein interaction functionality of the employed Pol and integrase bait constructs was demonstrated through their positive interaction , probably as a consequence of Pol/integrase oligomerization when using respective bait and prey construct combinations ( S2B Fig ) . To confirm the interaction between Gag and PLK1-3 in human cells , we investigated localization patterns of ectopically expressed , fluorescently labeled proteins . Constructs encoding C-terminally mCherry-tagged FL PFV Gag proteins ( either wt or iSTP/pmSTP variants ) and N-terminally eGFP-tagged variants of either FL hPLK1-3 or rPLK2 proteins were co-transfected into 293T cells . Representative results of confocal fluorescence microscopy analysis for selected Gag-PLK combinations are shown in Fig 3 and S3 Fig . Introduction of single mutations into the PFV Gag STP motif had no effect on intracellular Gag localization . As described previously [37] , all PFV Gag variants bound to chromatin during mitosis , completely overlapping the DAPI chromatin signal ( Fig 3B and 3C ) . In line with previous observations , eGFP-tagged PLKs displayed speckled nuclear and diffuse cytoplasmic distributions when expressed on their own ( Fig 3A and 3D; S3A Fig ) [38] . Importantly , none of the PLKs showed the overall chromatin association that was observed for PFV Gag ( Fig 3A and 3C; S3A Fig ) . Strikingly , the subcellular localization of hPLK1 , 2 and rPLK2 , and to less of an extent also hPLK3 , was dramatically altered during mitosis in the presence of wt but not iSTP PFV Gag ( Fig 3B; S3B and S3C Fig ) . Wt PFV Gag recruited PLK1 and PLK2 proteins to mitotic condensed chromatin resulting in an almost complete overlap of both signals , whereas for hPLK3 only a weak co-localization was observed . Distinct co-localization was not observed between any of the Gag variants with the examined eGFP-PLK fusions in interphase cells ( Fig 3D and 3E; S3 Fig ) . In some cases , overlap of mCherry and eGFP signals was observed in a diffuse pattern throughout the cytoplasm of interphase cells co-transfected with wt Gag and hPLK1 , hPLK2 , or rPLK2 ( Fig 3E; S3B Fig ) . Occasionally , co-localization between eGFP-hPLK1 and Gag-mCherry variants was noted at a perinuclear region , presumably the centrosome ( Fig 3E , white arrow ) . However , this event could not be ascribed to a specific co-localization pattern , as it was independent of the functional Gag STP motif and both partners individually localized to the microtubule-organizing center ( MTOC ) ( Fig 3D ) . As these results established that an intact STP motif in PFV Gag was required for co-localization with hPLK1 and hPLK2 in human cells , we addressed next whether the kinase requirements would correspond to the Y2H data using rPLK2 as a representative candidate . Hence , the same set of rPLK2 variants involved in the Y2H interaction analysis was employed in combination with wt and T225A PFV Gag . Whereas wt Gag , but not its T225A variant , co-localized with wt ( wt ) and constitutively active rPLK2 ( caKD ) , co-localization was not observed for any other rPLK2 protein variant including the kinase inactive ( iKD ) mutant ( Fig 3B; S3D Fig ) . Taken together , our results confirm that PLK1 and 2 and to a limited extent also PLK3 are cellular interaction partners of PFV Gag . The viral structural protein appears to recruit the PLKs to condensed , mitotic chromatin in an STP motif-dependent fashion , which requires functional PLK KD and PBDs . The interaction of PLKs with substrates is reportedly strongly dependent on PLK binding site phosphorylation status [29] . Nonetheless , low affinity binding to non-phosphorylated binding sites , and a few binding sites phosphorylated by PLKs themselves , have been reported [reviewed in 39] . To determine the phosphorylation-status of the putative PFV Gag PLK STP binding site , we examined particle-associated Gag or protein immunoprecipitated from cell lysates with various phosphopeptide-specific antibodies , including a custom-made antiserum specific for the PFV Gag STP motif ( Fig 4 ) . Detectable amounts of phosphorylated Gag protein were absent in immunoprecipitates of lysates of 293T cells transiently transfected with expression constructs of the four-component PFV vector system . This suggests that the majority of cellular PFV Gag is not phosphorylated at threonine residues in the context of T-P motifs , including the PLK binding motif . In contrast , phosphorylated Gag was readily detected in lysates of released wt PFV virions pelleted by ultracentrifugation , by either a commercial pThr-Pro ( pT-P ) –specific monoclonal or the custom-made PFV Gag PLK binding site phospho-specific ( Gag S-pT-P ) polyclonal antibody ( Fig 4B–4D ) . pT-P-specific signal intensity was reduced in particle lysate samples of all iSTP/pmSTP mutants , except S224A , which theoretically can still be phosphorylated at T225 by a non-PLK cellular priming kinase ( Fig 4B ) . Furthermore , pT-P-specific signals were almost undetectable in both wt and T225A mutant particle lysates pretreated with phosphatase ( Fig 4C ) . Hence , T225 phosphorylation appears to contribute to a large extent to the PFV Gag pT-P-specific signal detected in wt PFV particle samples . This was further substantiated by results using the custom-made PFV Gag S-pT-P-specific antiserum . A strong signal was detected in wt PFV particle samples , which was largely reduced by phosphatase pretreatment down to the background levels that were observed in PFV Gag T225A particle samples regardless of phosphatase treatment ( Fig 4D ) . In concordance with the results of Western blotting , mass spectrometric analysis revealed that a fraction of the PFV Gag protein recovered from SDS-PAGE-separated wt PFV particle samples is phosphorylated at the 223-TST-225 cluster ( S4A and S4B Fig ) . Obtained high-resolution fragmentation spectra do not exclude that within the 223–225 cluster the phosphogroup is localized at the position 225 ( S4C Fig ) . Taken together , these results indicate the presence of significant amounts of PFV Gag protein with phosphorylated PLK binding site motif in released wt PFV virions , which can potentially serve as a high affinity interaction partner for cellular PLK proteins upon PFV entry into newly infected host cells . The interaction between PFV Gag and human PLK proteins has not been implicated in previous studies , but may be of importance for the cell cycle-dependent replication characteristics of PFV [40–42] . Hence , we investigated whether introduction of Gag iSTP/pmSTP variants into PFV particles may exert an effect on virus replication . The infectivity of the PFV Gag mutants was assessed in the context of both replication-competent and single-round PFV expression systems , with similar results obtained between systems . Inclusion of any of the tested STP Gag mutations into the packaging construct decreased infectivity of respective single-round PFV vector supernatants on HT1080 , human primary fibroblast MRC-5 cells or murine embryonic fibroblasts by 50–70% ( Fig 5A; S5A Fig ) . A similar infectivity defect was observed early after target cell infection for replication-competent mutant viral particle preparations generated from full-length proviral expression constructs ( Fig 5B ) . Interestingly , the replication defect of the Gag STP motif mutant viruses became more prominent upon prolonged growth of the infected cell cultures ( Fig 5C ) . The infectivity of iSTP/pmSTP Gag-containing particles progressively mitigated down to 2 to 8% relative to wt after 10 days of co-cultivation . A similar effect was also observed in human primary fibroblast MRC-5 cells ( S5B Fig ) , indicating the phenotype is not cell-type specific . Thus , PFV particles harboring Gag unable to interact with cellular PLKs display a diminished infectivity in single-round infections and fail to efficiently spread in target cell cultures of different origin . Even though we were unable to detect any signs of PFV Pol interaction with cellular PLKs in the Y2H analysis , we nevertheless examined the influence of various PFV Pol STP motif mutants on viral infectivity in combination with either wt or iSTP mutant PFV Gag ( S2C Fig ) . Unlike Gag iSTP mutant PFV particles , the Pol iSTP mutant particles neither showed a significant difference in viral infectivity nor did they further promote the Gag iSTP infectivity defect ( S2C Fig ) . Taken together with the Y2H interaction analysis , these results strongly suggest that PFV Pol STP motifs located in the integrase domain did not mediate interactions with cellular PLKs and are dispensable for PFV-mediated infection of target cells . In order to understand the underlying causes for the replication defects of PFV particles with abolished Gag-PLK interaction , we characterized various features of the mutant viruses . We first harvested lysates of virus-producing cells after transient transfection of either four-component PFV vector system or full-length proviral PFV constructs into 293T cells . We also studied purified viral particles released into the supernatant to examine viral protein expression , particle release , viral genomic RNA ( vgRNA ) packaging and intra-particle reverse transcription ( Fig 6; S6 Fig ) . None of the PFV Gag STP motif mutations affected the efficiency of virus-production , as implicated by analysis of viral protein and nucleic acid content . Particle release efficiency , viral protein composition and processing of these mutants were comparable to wt , as were their vgRNA and vgDNA contents in the context of both viral production systems ( Fig 6; S6 Fig ) . Therefore , it can be concluded that the introduction of STP motif mutations into PFV Gag , which prevents interaction with cellular PLK proteins , does not have an effect on virus production , release , vgRNA packaging or intra-particle reverse transcription activity . The observed infectivity defects of PFV Gag STP motif mutant viruses pointed to an essential function of Gag-hPLK interactions at an early stage in viral replication . The results suggest that phosphorylation of a viral and/or cellular target , after binding of cellular PLKs to the phosphorylated Gag S-pT-P binding site , may be beneficial for PFV infection . To test this hypothesis we asked how enzymatic inhibition of hPLK1 might influence virus infectivity . For inhibition of cellular PLKs , in particular PLK1 , during FV entry , HT1080 target cells were incubated with the well characterized PLK inhibitor BI-2536 [43] in two different regimens as outlined in Fig 7A . Because BI-2536 treatment is known to arrest cells in prometaphase in a dose and time dependent manner , we used propidium-iodide ( PI ) staining to determine the cell cycle status 24 h post-infection , at the time when productive infection was quantified by flow cytometry ( Fig 7B ) . Short-term treatment of HT1080 target cells with 50 nM BI-2536 for 5 h ( 1 h prior to and 4 h during virus incubation ) did not alter the cell cycle profile at 24 h post-infection as compared to the mock-treated population . This indicates that 50 nM BI-2536 treatment yields a largely cycling cell population and that short-term hPLK1 inhibition did not result in cell cycle arrest . In contrast , the target cell cycle profile was dramatically changed when 10 nM BI-2536 was applied for the entire 25 h examination period ( long-term treatment ) . The target cell population was characterized by an altered cell cycle profile with highly increased G2-M population , indicating that hPLK1 inhibition by this inhibitor regimen led to a potent cell cycle arrest . The influence of PLK inhibition by BI-2536 treatment revealed interesting , differential phenotypes for Gag wt and Gag STP mutant containing FVs ( Fig 7C ) . When HT1080 target cells were treated with 50 nM inhibitor for only 5 h ( short-term ) the infectivity of wt PFV was reduced by 85% in comparison to mock treatment ( Fig 7C , orange bars ) . In contrast , the infectivity of STP motif mutant particles ( T225A , T225E , P226A ) were only reduced by 25 to 30% in comparison to respective mock treated samples . Thus , the short-term inhibitor treatment reduced the infectivity of wt virus nearly to the level of Gag STP mutant viruses . A more prominent inhibitory effect was observed for the long-term 10 nM BI-2536 treatment regimen ( Fig 7C , magenta bars ) . The infectivity of wt virus was again reduced to a much greater extent ( 65-fold ) than Gag STP mutant viruses ( 5–6 fold ) in comparison to the respective mock treatment controls , resulting in similar levels of residual infectivity . Taken together the data demonstrate that short-term PLK1 inhibition during entry of wt PFV , which does not induce a cell cycle arrest , can largely mimic the infectivity defect of virus mutants with Gag proteins unable to interact with cellular PLKs . The effect of BI-2536 on viral infectivity was also examined for murine leukemia virus ( MLV ) and HIV-1 ( HIV-1 ) vectors as representatives of retroviral genera that are dependent or independent of target cell mitosis for productive infection , respectively ( Fig 7C ) . Both types of vectors were pseudotyped with PFV Env and harbored an identical reporter gene as transmitted by FV vector particles . Whereas HIV-1 infectivity was only weakly ( 2-fold ) diminished under both treatment conditions , MLV infectivity was moderately inhibited ( 3-fold ) by short-term PLK inhibition and very strongly reduced ( 150-fold ) by long-term BI-2536 treatment . Hence , the data indicate that hPLK1 inhibition effects are not PFV-specific , which may encourage further insight into PLK proteins as a key link between retroviral infection and cell cycle progression . The reduced infectivity observed for PLK-interaction-deficient PFV or upon drug-mediated inhibition of PLKs during infection with wt PFV can result from defects at different steps during PFV entry and uptake into target cells . This includes early steps such as attachment , membrane fusion , capsid stability or later steps such as stability of the PIC or its tethering to host cell chromatin as well as vgDNA integration . To further narrow down which step might be affected we examined the uptake dynamics of wt and Gag T225A mutant virions containing fluorescently-tagged Gag proteins in synchronously infected HT1080 cells ( Fig 8 ) . At different time points after initiating virus uptake the amount of cell-associated , fluorescently labeled viral Gag protein was quantified by flow cytometric determination of the mean fluorescent intensity ( MFI ) of the whole cell population . As non-infectious control GFP-tagged Gag wt containing particles harboring a fusion-deficient PFV Env glycoprotein ( iFuse ) were used . All types of virions similarly attached to target cells by spinoculation at low temperature ( Fig 8 ) . Fusion-deficient virions ( wt iFuse ) were rapidly internalized upon raising the temperature and the Gag-GFP signal declined to background levels within 4 to 24 h depending on the amount of virus used for infection . This most probably is the consequence of rapid degradation of non-fused particles by endocytic uptake and targeting to lysosomes [44 , 45] . In contrast , Gag wt ( wt wt ) and Gag T225A ( T225A wt ) samples showed as early as 1 h post-infection significantly higher MFI signals than the samples of fusion-deficient virions ( wt iFuse ) , and this difference further increased over time . In line with previous studies [44 , 45] , this indicates rapid release of capsids into the cytoplasm for virions with fusion-competent , wt glycoprotein . Cytoplasmic PFV capsids are known to be quite stable and transported along microtubules to the MTOC of the host cell , where they await host cell entry into mitosis for progression towards virus genome integration into cellular chromatin [18] . In the examined time period of up to 24 h post-infection , no significant differences in the Gag-GFP MFI profiles of wt and Gag T225A mutant particles were observed , independent of the amount of virus initially attached to HT1080 cells . Taken together the analysis indicates similar attachment , fusion and uptake dynamics as well as Gag protein stability for wt and Gag T225A virions . This suggests a defect at a step post-disassembly of PFV capsids as responsible for the observed infectivity phenotype of iSTP mutant PFV virions . Closer examination of reporter gene expression in target cells transduced with wt or Gag STP motif mutant PFV particles over time revealed marked differences in the severity of STP mutant infectivity defect ( Fig 9A ) . The largest differences in STP motif mutant vector titers , up to 10-fold reduced compared to wt , were observed if the number of reporter gene expressing target cells used for virus titer calculation were quantified 1 day post-transduction in single-round infection experiments ( Fig 9A ) . Up to 3 days post-transduction this difference constantly declined until reaching a plateau of 3-fold reduced infectivity in comparison to wt . This phenotype of a decrease in viral infectivity and the time-dependent quantitative differences may suggest a combination of delayed and reduced integration potential of iSTP mutant PFV particles in comparison to wt virus . To gain support for this hypothesis , we extracted genomic DNA from HT1080 cells 10 days post-infection with different virus supernatants and determined integrated PFV genome copy numbers by quantitative Alu PCR ( Fig 9B ) . The number of integrated vector genomes of cells infected with PFV Gag T225A containing virus supernatants was reduced to 45% of wt and correlated well with their reduced infectivity . Thus , the reduced viral titer of PFV Gag T225A containing particles determined by the reporter gene transfer assay and flow cytometric analysis late after transduction is mainly the consequence of a reduced number of proviral integration events . Attenuated integration would suffice to explain the defects in mutant virus titers , but not the delay in the number of viral DNA-derived reporter gene expressing cells observed for the STP mutants during the first three days after infection . To further characterize the iSTP/pmSTP integration dynamics , we made use of a well-characterized HIV-1 integrase inhibitor , dolutegravir ( DTG ) [46] , previously shown to also be active against PFV integrase [47] . The rationale behind this set of experiments was that if wt and iSTP mutant viruses integrated their genomes at different dynamic rates during the initial 72 h post-infection , then they should be differentially sensitive to DTG treatment over time . Following this notion , we synchronously infected HT1080 cells with wt ( wt wt ) or T225A mutant ( T225A wt ) PFV and mock solvent ( DMSO ) or 2 μM DTG at different time points post-infection and maintained the drug until flow cytometry analysis was performed at day 10 post-infection ( Fig 9C ) . As control , cells were infected with supernatants harboring viral particles containing wt PFV Gag in combination with Pol variants with either enzymatic inactive reverse transcriptase ( wt iRT ) or integrase ( wt iIN ) domains . As expected , DTG addition to samples infected with the latter viruses ( wt iRT , wt iIN ) did not influence their infectivity in comparison to respective mock treated controls ( Fig 9D ) . In contrast , the infectivity of both Gag wt ( wt wt ) and Gag T225A ( T225A wt ) containing viral particles was efficiently inhibited ( to ~1% of respective mock treated samples ) when DTG was added at the time of virus addition ( 0 h time point ) ( Fig 9D ) . Strikingly , the infectivity of both types of viruses was differently affected when DTG was added at later time points . Gag wt ( wt wt ) containing particles became resistant to DTG-mediated inhibition quickly and acquired 50% drug resistance at ~12 h post-infection whereas Gag T225A ( T225A wt ) containing particles reached this value much later at ~27 h post-infection . These results indicate that approximately half of wt PFV proviruses establish within 12 hours post-infection and confirm that Gag STP mutations cause a significant delay in PFV integration . In sharp contrast to orthoretroviruses such as HIV-1 and MLV that predominantly integrate within or in the vicinity of actively expressed genes , PFV disfavors integration into transcriptionally active chromatin and within transcription units [48–50] . To test if the delayed integration dynamics of PFV Gag iSTP mutants influences their integration site preferences , we interrogated integration site distributions of PFV vectors produced using S224A and T225A mutants of the Gag packaging vector ( pcoPG4 ) . The integration sites produced by infection of HT1080 cells were amplified using ligation-mediated ( LM ) -PCR as described [48] . Pools of amplified fragments containing junctions of downstream ( U5 ) ends of vgDNA and chromosomal DNA were sequenced using the Illumina MiSeq platform , and precise positions of integration sites were mapped to the hg19 version of the human genome . To assess data reproducibility , PFV wt integration sites were determined from three independent viral infection experiments whereas mutant sites were collected from two independent sets of infections . In total , 185 , 481 WT , 44 , 918 S224A , and 40 , 455 T225A sites were mapped across experiments ( Table 1 ) . A previously described dataset of ~2 . 2 million unique integration sites obtained using recombinant PFV intasomes and isolated human genomic DNA in vitro was used as a reference [48] . Such a reference incorporates the innate preferences of PFV integrase for chromosomal DNA sequence , whatever they may be , as well as variations of PCR amplification efficiency depending on local G/C content or repeat density . In agreement with previous observations , wt PFV targeted genes 31 . 4% of the time across datasets , which was highly statistically significant from the matched in vitro frequency of 47 . 5% ( Fig 10A; S7 Fig ) . Due to the relatively large numbers of analyzed integration sites , the levels at which genes were targeted yielded statistical significance across some wt datasets . Given this , gene targeting by the iSTP Gag mutants was largely indistinguishable from the wt . This is especially evident by comparing the WT-1 and mutant datasets , which were derived from viral infections performed on the same day ( P values of 0 . 87 for S224A-1 and 0 . 58 for T225A-1; S7 Fig ) . By contrast , the targeting of heterochromatin , which is naturally preferred by PFV , was reproducibly hyper-targeted by the Gag mutants . Thus , mutant viral targeting of lamina-associated domains ( LADs ) was significantly enhanced ( Fig 10B; P values 4 . 1 x 10−46 for matched S224A and 8 . 4 x 10−47 for matched T225A infections–S7 Fig ) whereas the targeting of gene dense regions of chromosomes was significantly decreased ( Fig 10C; matched S224A and T225A P values of 1 . 2 x 10−55 and 1 . 3 x 10−61 , respectively ) . The timing of integration of some retroviruses including MLV and FVs has long been recognized to coincide with mitosis [reviewed in 51] , but until now no explanation on how this process is coordinated has been offered . Since host cell division is a major limiting factor for the broad use of FV ( and some other retroviral ) vectors for transduction of resting target tissues , knowledge of the cause of this limitation may help to improve the existing gene therapy vector systems [8] . In this study , we identified members of the hPLK family as functional co-factors of PFV replication . By using both yeast and mammalian cell based systems , we show that PFV Gag interacted with PLK1-3 in yeast and recruited PLK1-2 to condensed human mitotic chromatin . The failure of FL PFV Gag to interact with FL PLK1 but not with FL PLK2 or FL PLK3 in Y2H assays may indicate a differential structural requirement for Gag to interact with FL PLK1 compared to FL PLK2 or FL PLK3 . FL PLK1 only interacted with C-terminally truncated PFV Gag baits lacking at least 104 aa encompassing half of the GR-rich region , which is known to be functionally relevant for capsid morphology and other functions . Therefore it may be speculated that a differential oligomeric structure of PFV Gag mutants may be responsible for the differential interaction phenotype with individual FL PLKs in yeast . The interactions of PFV Gag with PLKs in yeast and mammalian cells depended on a central domain in PFV Gag containing a consensus S- ( T/S ) -P PLK binding site [29] , which is evolutionary conserved in all primate FVs . Similar as reported for STP binding sites of other PLK substrates [33 , 52 , 53] , phosphomimetic mutations of the central STP motif residue did not restore interaction of PFV Gag with the examined PLK proteins . This study for the first time identified this sequence in PFV Gag as a motif important for interaction with cellular PLKs and its conservation across primate FV Gag species points to a potentially general role of this interaction in primate FV replication . Interestingly , Gag proteins of other retroviruses , such as MLV , Rous sarcoma virus , gibbon ape leukemia virus or human T-lymphotropic virus 4 harbor S- ( T/S ) -P peptide motifs that upon phosphorylation may function as PLK binding sites . In line with this , the MLV vector used in this study showed a similar response profile towards target cell PLK inhibitor treatment as PFV . PFV Gag of released virions was phosphorylated to a significant extent at T225 , whereas Gag phosphorylated at threonine residues was not detectable in immunoprecipitates from virus producing 293T cells . This may be simply a consequence of overpowering the cells by transient overexpression of PFV Gag . Alternatively it may suggest that only a minor proportion of cellular PFV Gag is modified by threonine phosphorylation , potentially in a cell-cycle dependent manner , and that Gag phosphorylated at T225 and other residues ( as detected by the pT-P antibody staining ) is enriched in secreted virions . It will be interesting to determine whether only Gag in preassembled capsids or also monomeric Gag harbor phosphorylated PLK binding sites and if Gag phosphorylation is dependent on the cell cycle status of the expressing cell . While the phenomenon of the apparent enrichment of phosphorylated Gag in secreted virions awaits mechanistic characterization , it is plausible to speculate that Gag phosphorylation , not only at T225 but also other residues , may influence Gag protein oligomerization and functional capsid formation as well as their trafficking and/or their interaction with the cytoplasmic domain of the Env LP subunit , which is essential for PFV budding [54] . It has been shown previously that PFV Gag is phosphorylated on multiple serine residues [55] , but no kinase ( s ) responsible for this process had been identified . Therefore , it will be important to decipher the cellular kinase ( s ) that mediate Gag phosphorylation at T225 and/or other serine/threonine residues . Likely candidates are Cdk1 , Cdk2 or PLKs themselves . One speculation concerning the biological function of FV Gag phosphorylation could come from the analogy of their replication strategy to hepadnaviruses . Serine phosphorylation of the HBV core protein has important functions in reverse transcription and virion morphogenesis [56–59] . However , the analysis of consequences of PFV Gag T225 phosphorylation suggests rather a role during the early events of PFV replication than for virus morphogenesis processes as observed for HBV . On the PLK side , the assays using rPLK2 as a representative revealed that in addition to a functional PBD , kinase activity is required for the observed protein-protein interactions . This is reminiscent of reports for some other PLK2 substrates in neuronal tissues , for which the PLK2 ATP-binding pocket in addition to the integrity of the STP phosphorylation site are important for driving interaction between these two partners [23 , 60] . It was suggested that phosphorylation and ATP binding induce structural and conformational changes in PLK2 and its binding partner to facilitate their interaction , which may also be the case for PFV Gag-PLK interactions . Alternatively , potential PFV Gag phosphorylation by PLKs in cis , at yet to be identified sites , or the trans phosphorylation of other currently unknown components complexed with Gag , may stabilize the observed interactions and be crucial for the PLK-mediated role in PFV replication . However , it has to be determined whether similar requirements in respect to a functional KD as observed for rPLK2 are essential for the interaction of PLK1 and 3 with PFV Gag as well . The functional analysis of the FV Gag—PLK interaction for viral replication suggests a significant role in promoting timely and efficient PFV vgDNA delivery into the nucleus , as revealed by a detailed comparison of wt virions to those containing Gag with STP mutations . While the single amino acid exchanges of each of the residues in the Gag PLK recognition and binding motif did not alter virus particle production , vgRNA encapsidation or reverse transcription , they resulted in diminished infectivity of Gag STP mutant particles , underlined by the delayed and decreased vgDNA integration efficiency compared to wt . The particle-associated Gag STP mutants were likely unable to interact with cellular PLK proteins after entry into target cells , due to the deficient STP motif phosphorylation and the resulting lack of recognition by the PLKs . Along this line , future studies should address the details and mechanistic insights into PFV Gag phosphorylation as mentioned before and should characterize the spatial-temporal requirements of the PFV capsid interaction with one or more of the mammalian kinase families . By dissecting various individual steps in PFV uptake and entry , we determined that the iSTP mutant viruses display wt-like attachment , entry and Gag protein stability characteristics , but that their vgDNA genome integration is delayed and reduced compared to wt virions . The successful diminishment of wt PFV infectivity to the level of the STP mutants by drug-mediated PLK enzymatic inhibition in infected cells suggests that the reduced STP mutant infectivity is not induced by the lack of effective PFV Gag interaction with host cell PLK proteins per se . The abolished association of PLK enzymatic activity with FV capsids appears to be responsible for the STP mutant virus phenotype , because the observed effects of the PLK inhibitor on viral infectivity were dependent on the presence of wt Gag protein in PFV virions , although we currently cannot formally exclude an inhibition of the Gag—PLK interaction by the enzymatic inhibitor . Furthermore , the similar reduction of iIN mutant virus infectivity upon drug-mediated inhibition of PLK activity suggests that the replication step that the PLK-Gag interaction likely potentiates is nuclear import and/or tethering to host chromatin . Importantly , these experiments revealed that MLV , but not HIV-1 , virions were sensitive to PLK inhibition in a PFV wt-like manner . Interestingly , the MLV Gag contains a unique STP motif at amino acid positions 123–125 in the MA subunit , the significance of which for virus replication to our knowledge has not yet been characterized . Thus , we suspect that the ability for similar Gag domains to confer interaction with one or more cellular PLKs may contribute to the mitosis dependence for PFV and MLV infection . As BI-2536 has the highest specificity for PLK1 , but also inhibits PLK2 and PLK3 , it is currently unclear whether interactions with a single or multiple PLKs are involved in PFV replication . Although PLK1 appears the most likely candidate , the exact role of PLK2 ( and possibly also PLK3 ) needs to be addressed in the course of further studies . Furthermore , it will be important to resolve the spatio-temporal involvement of different PLKs in PFV replication ( e . g . by applying live cell imaging studies of PFV-infected cells ) to better understand the mechanistic details of these interactions . Furthermore , since mutant PFV unable to interact with cellular PLKs retain a residual infectivity it may be speculated that likely currently unknown interactions of the virus and/or capsid with other cellular components of the cell cycle regulatory machinery exist , which provide functions redundant to or compensatory for the Gag—PLK interaction characterized in this study . Finally , we observed that PFV Gag STP mutant viruses achieved only 45% of wt levels of proviral integrations , suggesting that 55% of vgDNA copies taken up are lost . Furthermore , statistically significant alterations in integration site profile , with a shift towards an even higher preference for heterochromatic regions than wt PFV , was detected for mutant viruses . Together these two features may account for the roughly 3-fold reduced infectivity of PLK-binding deficient viruses . The underlying mechanism responsible for this phenotype are currently unknown , but may be due to untimed mutant capsid disassembly and/or faster degradation of mutant PICs in host cells . Alternatively cellular PLKs may represent the first tethering- and/or integration-promoting host cell factors involved in PFV replication . Perhaps the PFV Gag-PLK interaction and subsequent phosphorylation of Gag itself or other proteins in PFV PICs influences their integration site selection profile . Even though it is uncertain whether PLK protein ( s ) may mediate roles similar to those of LEDGF/p75 and BET proteins in HIV and MLV integration , respectively [reviewed in 61] , the presented data offer initial insight into how PFV may coordinate its timely genome nuclear delivery with host cell mitosis . Analysis of the amino acid sequence of other PFV structural proteins revealed that PFV Pol also harbors two potential STP motif consensus binding sites , which are located in the integrase domain . However , follow up analyses , which included Y2H using Pol or its processed integrase subunit as baits as well as functional analysis of single-round vector particles with point mutations in single or both Pol STP binding motifs , failed to indicate a biologically relevant interaction between Pol/integrase and PLK or a role in PFV infectivity . This highlights that it is the interaction of cellular PLKs with PFV Gag and not Pol that are important for timely and efficient proviral integration . The capsid protein of HIV-1 has similarly been shown to play a significant role in integration through its cellular binding partner CPSF6 [62] . Although our data indicate that the two STP motifs in PFV Pol appear not to be of functional relevance for viral infectivity , they do not formally exclude the possibility that cellular PLKs in complex with PFV Gag mediate phosphorylation of PFV integrase residues upon virus entry and thereby influence integrase function . Which functionally relevant residues in integrase thereby might be posttranslationally modified needs to be determined , however , our data suggest , that it is not T961 or T1058 within the integrase STP motifs . The human embryonic kidney cell line 293T ( ATCC CRL-1573 ) [63] , the human epithelium HeLa cell line ( ATCC CCL-2 ) [64] , the human primary fibroblast MRC-5 cell line ( ATCC CCL171 ) , immortalized mouse primary embryonic fibroblasts ( obtained from M . Trilling , Essen , Germany ) and the human fibrosarcoma cell line HT1080 ( ATCC CCL-121 ) [65] as well as the clonal variant HT1080 PLNE thereof containing a PFV LTR driven EGFP reporter gene expression cassette [66] were cultivated in Dulbecco’s modified Eagle’s medium ( DMEM ) supplemented with 10% heat-inactivated fetal calf serum and antibiotics . HeLa cells used for confocal laser scanning microscopy were cultivated in phenol red free media . For hPLK1 and PFV integrase inhibition experiments , BI-2536 ( Selleckchem ) or Dolutegravir ( Selleckchem ) dissolved in DMSO was used at final concentrations as indicated . A four-component PFV vector system , consisting of the expression-optimized packaging constructs pcoPG4 ( PFV Gag ) , pcoPE ( PFV Env ) , pcoPP ( Pol ) , and the enhanced green fluorescent protein ( eGFP ) -expressing PFV transfer vectors pMD9 or puc2MD9 , has been described previously [67–69] . In some experiments previously described variants of the PFV Pol packaging construct , containing ORFs with catalytically inactive reverse transcriptase ( pcoPP2 , Pol iRT , YVDD312–315GAAA mutation ) or catalytically inactive integrase ( pcoPP3 , Pol iIN , D936A mutation ) , were used [66 , 67] . All PFV Gag packaging constructs used in this study are based on the parental pcoPG4 vector or its C-terminal mCherry and eGFP tagged variant pcoPG4 CCh and pcoPG4 CEG [68] . The PFV Gag packaging constructs encoding mutant Gag protein with alterations in a putative PLK STP binding motif ( pcoPG4 S224A , pcoPG4 T225A , pcoPG4 T225D , pcoPG4 T225E , pcoPG4 P226A ) generated for and used in this study are depicted in Fig 1D . All PFV Pol packaging constructs used in this study are based on the parental pcoPP vector [67] . PFV Pol packaging constructs encoding mutant Pol protein with alterations in the two putative PLK STP binding motifs , either individually ( pcoPP6 [S960A] , pcoPP7 [S1057A] ) or in combination ( pcoPP8 [S960A+S1057A] ) , were generated by recombinant PCR techniques and used in this study . The CMV-driven proviral expression vector pczHSRV2 ( wt ) and its variants pczHSRV2 M69 ( iRT ) , expressing a Pol protein with enzymatically inactive RT domain ( YVDD312-315GAAA mutation ) , and pczHSRV2 M73 ( iIN ) , with enzymatically inactive IN domain were described previously [70 , 71] . For this study the variants with altered putative PLK binding motif , pczHSRV2 S224A ( S224A ) , pczHSRV2 T225A ( T225A ) , pczHSRV2 T225E ( T225E ) , pczHSRV2 P226A ( P226A ) , were generated ( Fig 1E ) . Lentiviral PFV Env pseudotypes were generated using the mutant PFV Env packaging vector pcoPE01 , the HIV-1 Gag-Pol expressing packaging vector pCD/NL-BH and the HIV-1 based transfer vector p6NST60 [44 , 72 , 73] . MLV PFV Env pseudotypes were produced using pcoPE01 , the MLV Gag-Pol expressing packaging vector pHIT60 and the MLV based transfer vector pczCFG2 fEGFPf [44 , 74] . Y2H expression constructs containing the GAL4 activation domain ( GA ) with C-terminal ( pCGA ) or N-terminal ( pNGA and pNGA2 ) ORF fusions were based on pGADT7 ( Clontech ) . Y2H constructs containing the GAL4 DNA-binding domain ( GB ) with C-terminal ( pCGB ) or N-terminal ( pNGB ) ORF fusions were based on pGBKT7 ( Clontech ) ( Fig 1B ) . The respective Gag sequence for authentic , full-length wt Gag Y2H constructs was amplified from pcziPG4 [69] whereas constructs containing sequences of human expression-optimized Gag were based on pcoPG4 [68] . The Gag sequence for PFV Gag truncation Y2H constructs ( Gag 1–621 , Gag 1–594 , Gag 1–544 , Gag 1–482 , Gag 1–450 , Gag 1–350 , Gag 1–310 , Gag 1–297 , Gag 1–295 , Gag 1–212 , Gag 1–129 , Gag 130–648 , Gag 156–648 , Gag 183–648 , Gag 230–648 , Gag 130–295 , Gag 183–295 , and Gag 296–648 ) was amplified from the respective pcziPG4 truncation mutant expression constructs , which have been described previously [75] . Y2H PFV Pol expression constructs were generated inserting either the ORF of pcoPP1 ( iPR ) encoding a full-length Pol protein with enzymatically inactive PR domain or aa 752–1143 of pcoPP comprising the mature PFV Pol IN subunit into pCGA , pNGA , pCGB or pNGB expression vectors [66 , 67] . Y2H pCGB and pCGA variants with full-length hTsg101 ORF were obtained from Wesley Sundquist and have been described previously [76] . CMV-driven mammalian expression constructs with ORFs encoding human PLK1 to PLK4 ( hPLK1 to 4 ) were obtained from Rene Medema [77] , and encoding wt rat PLK2 ( rPLK2 wt ) and point mutants thereof from Daniel Pak [60] ( Fig 1D ) . ORFs encoding human ( hPLK5 ) and mouse PLK5 ( mPLK5 ) were purchased from Imagenes and subcloned into CMV-driven expression vectors . For some experiments N-terminally eGFP-tagged variants of hPLK1 to 3 ( peGFP-PLK1 to 3 ) were used ( Fig 1D ) . For Y2H interaction analysis the full-length ORFs of hPLK1 to 3 , 5 , mPLK5 , rPLK2 and point mutants thereof , as well as the polo-box domains of hPLK1 to 4 were inserted C-terminal to the GAL4 activation domain ( pCGA ) into pGADT7 ( Clontech ) ( Fig 1B ) . All constructs were verified by sequencing analysis . Primer sequences and additional details are available upon request . Cell culture supernatants containing recombinant viral particles were generated by transfection of the corresponding plasmids into 293T cells using polyethyleneimine ( PEI ) as described previously [37 , 66] . For subsequent Western blot analysis the supernatant generated by transient transfection was harvested , passed through a 0 . 45-μm filter and centrifuged at 4°C and 25 , 000 rpm for 3 h in a SW32Ti rotor ( Beckman ) through a 20% sucrose cushion . The particulate material was resuspended in phosphate-buffered saline ( PBS ) . For Western Blot analysis using phosphopeptide-specific antibodies viral particles were resuspended in TBS . Subsequently , half of the sample was digested for 1 h at 30°C with Lambda Phosphatase ( NEB , P0753; 7 . 15 U/μl Lambda PP , 1% Triton X-100 , 1 mM MnCl2 , 1x Lambda Phosphatase buffer ) whereas the other half was mock digested prior to addition of protein sample buffer . Transduction efficiency of recombinant , eGFP-expressing PFV vector particles by fluorescence marker-gene transfer assay was analyzed at various time points post-transduction as described previously [44 , 68 , 78] . For synchronized infections a modified spinoculation protocol was used as described previously [44] . Virus particles generated by use of proviral expression plasmids were titrated on HT1080 PLNE cells harboring a Tas-inducible nuclear egfp ORF in their genome as described previously [66] . All transduction experiments were performed at least three times . In each independent experiment the values ( eGFP focus forming units per ml , eGFP ffu/ml ) obtained with the wt construct pcoPG4 and pczHSRV2 , respectively , were arbitrarily set to 100% and values obtained with other constructs were normalized as a percentage of the wt values . For evaluation of the effect of Dolutegravir ( DTG ) inhibition of PFV integrase enzyme activity on the infectivity of various PFV viruses HT1080 cells were plated at 2 x 104 cells per well in 12-well plates 24 h prior to infection . Target cells were synchronously infected PFV Gag wt or Gag T225A mutant containing viral supernatants by spinoculation as described previously [44] . Briefly , target cells loaded with virus particles by incubation with 1 ml virus and centrifugation at 960x g and 10°C for 30 min . Subsequently , unbound virus was removed by a wash step and uptake and infection of virions attached to the cell surface initiated by shifting samples to 37°C in fresh growth medium . Two different amounts of PFV Gag wt or Gag T225A containing four-component vector system supernatants were used , which yielded in about 50% or 25% GFP positive cells , respectively , determined by flow cytometry analysis at 10 days post-infection in mock treated samples . As controls , cells were infected with supernatants harboring PFV Gag wt in combination with Pol variants with either enzymatic inactive RT ( iRT ) or IN ( iIN ) domains , at the highest amount of virus supernatant used . DTG stock ( 2 mM in DMSO ) or an identical amount of solvent alone was added at different time points during infection as indicated in the experimental outline in Fig 9C to obtain a final concentration 2 μM DTG in the growth medium . Once added , DTG was replenished every second or third day until 10 days post-infection when the percentage of GFP expressing target cells of the individual samples was determined by flow cytometry analysis . These values were used to calculate viral titers of individual samples as described previously [78] and the effect of the time point of DTG addition on relative viral infectivity of the individual viruses plotted in respect to their respective mock treated control whose viral titers were arbitrarily set to 100% . Thereby a bias of the roughly 3-fold reduced infectivity of T225A Gag containing virions in comparison to wt was avoided . For synchronized infections using virions with GFP-tagged Gag proteins a modified spinoculation protocol was used as described previously [44] . Briefly the protocol involved preincubation of target cell ( 5 x 104 cells/well plated in 12-well dishes one day in advance ) at 10°C for 10 min . Subsequently the growth medium was replaced with 1 ml cold virus supernatant with the plates kept on cool pads . Next the viral supernatant containing tissue culture plates were centrifuged for 30 min at 960 x g at 10°C in a tissue culture centrifuge for loading of viral particles onto the target cells but preventing their endocytic uptake or glycoprotein-mediated membrane fusion . Uptake and infection were then initiated by replacing the viral supernatant containing all non-adsorbed vector particles with fresh warm growth medium and incubation at 37°C , 5% CO2 for the time periods indicated . Duplicate target cell wells plated were simultaneously transduced by spinoculation and samples were harvested for flow cytometry analysis consecutively at different time points . For analysis of the level of Gag-GFP protein cell association the mean fluorescent intensity ( MFI ) in the GFP channel of cultures transduced with different amounts viral supernatants ( undiluted or 1:10 diluted cell-free 293T virus supernatant ) was determined from all living cells as gated by their FSC/SSC profile . Cells from a single transfected 100-mm cell culture dish were lysed in detergent-containing buffer and the lysates were subsequently centrifuged through a QIAshredder column ( QIAGEN ) . Protein samples from cellular lysates or purified particulate material were separated by SDS-PAGE on a 10% polyacrylamide gel and analyzed by immunoblotting as described previously [54] . Polyclonal rabbit antisera specific for PFV Gag [79] or the amino acids ( aa ) 1 to 86 of the PFV Env leader peptide ( LP ) , [54] as well as hybridoma supernatants specific for PFV PR-RT ( clone 15E10 ) or PFV IN ( clone 3E11 ) [80] , or a GAPDH-specific monoclonal antibody ( G8795; Sigma ) , or a commercial phospho-specific mouse monoclonal antibody ( 9391 , Cell Signaling ) recognizing phospho-Thr-Pro peptides ( pT-P ) in a highly context-independent fashion , were employed . Furthermore , a custom-made ( Abmart ) , affinity purified , phosphopeptide specific , rabbit polyclonal antiserum recognizing the phosphorylated PFV PLK STP binding site ( PFV Gag S-pT-P ) was generated by immunization with PRATS ( pT ) PGNIP peptides . After incubation with species-matched horseradish peroxidase ( HRP ) -conjugated secondary antibody , the blots were developed with Immobilon Western HRP substrate . The chemiluminescence signal was digitally recorded using a LAS-3000 ( Fujifilm ) imager and quantified using ImageGauge ( Fujifilm ) . For mass spectrometric analysis ( performed by the Mass Spectrometry Facility at the Max Planck Institute of Molecular Cell Biology and Genetics , Dresden ) 200 ml cell culture supernatants of 293T cells transfected either with wild type constructs of the 4-component PFV vector system ( wt ) or with pUC19 ( mock ) were subjected to a 2-step purification . In the first step 5 ml sucrose cushions ( 20% in PBS ) were overlaid with 35 ml of cell-free cell culture supernatant ( 0 . 45 μm sterile filtrate ) and centrifuged for 90 min in SW-32Ti rotors at 25 , 500 rpm ( ~82 , 600 gave ) and 4°C . The pellets were then gently resuspended in 150 μl of PBS buffer per bucket ( ~200 fold volume concentration ) and the content of 6 buckets of each supernatant type were pooled together . In the next step , 900 μl of pooled concentrated supernatants were successively loaded in 450 μl aliquots on Amicon Ultra 0 . 5 ml 100 K concentrators and centrifuged at 14 , 000 g ( ~5 min ) until a final volume of ~90 μl ( ~2000 fold concentration ) . Two 20 μl aliquots of the concentrated supernatant from virus producing ( wt ) and control cells ( mock ) were then separately loaded on NuPAGE Novex 4–12% Bis-Tris Midi Protein Gels ( Thermo Fisher Scientific ) , separated by gel electrophoresis and visualized with Coomassie staining . The gel regions at around 70 kDa corresponding to PFV Gag were then excised as indicated in S4A Fig and in-gel digested with trypsin [81] . The resulting peptide mixtures were analysed by LC MS/MS on an Ultimate3000 nanoLC system interfaced on-line to a Q Exactive HF hybrid Quadrupole-Orbitrap mass spectrometer ( both Thermo Fisher Scientific , Bremen , Germany ) . MS/MS spectra were matched using MASCOT v . 2 . 2 . 04 ( Matrix Sciences Ltd , London , UK ) and MS Amanda v . 1 . 0 . 0 . 7503 [82] programs with 5ppm mass accuracy for precursor and fragments , and finally manually evaluated . The analysis of the intracellular distribution of C-terminal mCherry-tagged Gag constructs and N-terminal eGFP-tagged PLK constructs using confocal microscopy was done as described previously [37] . Briefly , 293T cells were co-transfected with 1 . 7 μg of Gag-mCherry-encoding and 2 μg of eGFP-PLK encoding construct , using PEI-transfection method in 100 mm dishes as described above . One day post-transfection , cells were replated at a concentration of 3 x 105 cells per well on poly-L-lysine coated cover slips in 12-well plates . At 48 h post-transfection the cells were washed with cold PBS , fixed with 3% paraformaldehyde , and the cell nuclei were stained with DAPI for 5 min . Finally the cells were covered with Mowiol . Confocal laser scanning images were obtained on a Leica SP5 , using a Leica HC PL APO 40x 1 . 25 oil immersion objective . Fluorescence images were evaluated using ImageJ software . For quantitative analysis of co-localization at least 100 cells showing co-expression of individual Gag-PLK protein combinations were evaluated . Preparation of particle and cellular samples for qPCR analysis was performed as previously described [37 , 44] . Primers , Taqman probes and cycling conditions for specific quantification of PFV genome , eGFP , or human GAPDH are summarized in S1 Table . All sample values obtained using a StepOne Plus ( Applied Biosystems ) qPCR machine were referred to a standard curve consisting of 10-fold serial dilutions of respective reference plasmids containing the target sequences ( puc2MD9 for eGFP sequences , pczHSRV2 for viral genomic sequences , pCR2 . 1-TOPO-GAPDH for GAPDH sequences ) . All sample values included were in the linear range of the standard curves with a span from 10 to 109 copies . The values for the DNA or RNA content of viral particle samples obtained by the qPCR analysis are expressed as percentage of the wt ( generated by transfection of cells with pcoPG4 , pcoPP , pcoPE and puc2MD9 or alternatively pczHSRV2 ) . Determined RNA values of cellular samples were calculated as copies / ng total RNA and expressed as percentage of the wt . After infection of HT1080 target cells with various viral supernatants generated by transient transfection of different combinations of plasmids of the four-component PFV vector system into 293T cells , cellular genomic DNA was extracted one week later and PFV integration sites were amplified by linker-mediated PCR , as described previously [48] . Bioinformatic and statistic analyses of RefSeq genes , LADs , and gene density were as described [48] . The initial large scale Y2H screen was performed at the Preclinical Target Development , and Genomics and Proteomics Core Facilities of the German Cancer Research Center , Heidelberg , using PFV Gag Y2H bait constructs pCGB-PG and pNGB-PG and a commercially available HeLa cDNA library ( Clontech ) . For all other yeast two-hybrid analyses , the Y2H Matchmaker System based on the AH109 S . cerevisiae strain ( Clontech ) was used . Double-dropout medium was used for selecting the yeast transfected with both the plasmid containing GAL4 activation domain and the plasmid bearing GAL4 DNA-binding domain . Quadruple-dropout medium was used for selecting yeast in which GAL4 activity was present due to interactions of transfected proteins/peptides . For production of transformation-competent yeast cells , a standard protocol entailing lithium acetate was used [83 , 84] . Briefly , yeast overnight culture was set up by resuspending a few colonies , grown on complete-YPD agar plates , in 100 ml of YPD media and overnight incubation at 30°C . The main yeast culture was setup the following day by adding 400 ml of YPD media to the overnight culture and incubating the culture at 30°C until its optical density measured at 600 nm ( OD600 ) reached a value of 0 . 6 . A culture of the appropriate OD600 was pelleted at 3000 rpm for 8 min and the yeast pellet was washed once with 200 ml of 1x LiAc solution ( 0 . 1 M lithium acetate , 10 mM Tris-HCl pH 8 . 0 , 1 mM EDTA ) and once with 30 ml of 1x LiAc solution , before being resuspended in 1 . 5 ml 1x LiAc solution . After 2 h of incubation at room temperature , 0 . 375 ml of sheared salmon sperm DNA ( 10 mg/ml , Ambion ) and 2 . 275 ml of PEG solution ( 40% ( w/v ) PEG 3350 , 0 . 1 M lithium acetate , 10 mM Tris-HCl pH 8 . 0 , 1 mM EDTA ) were added . 67 μl of competent yeast cell solution and 1 . 5 μg DNA of each Y2H construct were mixed and incubated overnight at 30°C . On the following day , the yeast was subjected to a 15 minute heat shock at 42°C , after which 15–20 μl of transfected yeast cells were plated on double-dropout ( -Leu/-Trp ) agar plates . The plates were incubated for three to four days at 30°C . Colonies growing on double-dropout plates were resuspended in a mixture of 55% H2O + 25% glycerol + 20% double-dropout medium and plated on quadruple-dropout ( -Leu/-Trp/-Ade/-His ) agar plates . Yeast growth was analyzed after two to four days incubation at 30°C .
Viruses are masters at exploiting host cell machineries for their replication . For human immunodeficiency virus type 1 ( HIV-1 ) , the best-studied representative of the Orthoretrovirinae subfamily from the genus lentiviruses , numerous important virus-host interactions have been described . In contrast , only a few cellular proteins are known to influence the replication of foamy viruses ( FVs , also known as spumaviruses ) , an intriguing type of complex retrovirus of the Spumaretrovirinae subfamily that combines features of both retroviruses and hepadnaviruses in its replication strategy . Given the increasing interest in FVs as gene transfer tools and their unique status within the retrovirus family , this discrepancy urged the identification of novel host cell interaction partners of FV structural components . This study focused on prototype FV ( PFV ) , the best-characterized member of FVs , and its capsid protein , Gag , as the central player of viral replication . Members of the mitosis-regulatory , polo-like kinase ( PLK ) family were identified as novel Gag binding partners . The Gag interaction with PLK1 ( and possibly also PLK2 ) facilitated efficient PFV genome integration into host chromatin , ensuring successful replication and viral spread in infected target cell cultures . Collectively , our results elucidate the first link between cell cycle regulatory networks and the mitosis-dependent PFV integration process .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "phosphorylation", "protein", "interactions", "viral", "transmission", "and", "infection", "293t", "cells", "cell", "cycle", "and", "cell", "division", "biological", "cultures", "cell", "processes", "microbiology", "viral", "structure", "research", "and", "analysis", "methods", "proteins", "cell", "lines", "viral", "replication", "virions", "biochemistry", "host", "cells", "cell", "biology", "post-translational", "modification", "virology", "biology", "and", "life", "sciences", "capsids" ]
2016
Interactions of Prototype Foamy Virus Capsids with Host Cell Polo-Like Kinases Are Important for Efficient Viral DNA Integration
Transcription factors are a main component of gene regulation as they activate or repress gene expression by binding to specific binding sites in promoters . The de-novo discovery of transcription factor binding sites in target regions obtained by wet-lab experiments is a challenging problem in computational biology , which has not been fully solved yet . Here , we present a de-novo motif discovery tool called Dispom for finding differentially abundant transcription factor binding sites that models existing positional preferences of binding sites and adjusts the length of the motif in the learning process . Evaluating Dispom , we find that its prediction performance is superior to existing tools for de-novo motif discovery for 18 benchmark data sets with planted binding sites , and for a metazoan compendium based on experimental data from micro-array , ChIP-chip , ChIP-DSL , and DamID as well as Gene Ontology data . Finally , we apply Dispom to find binding sites differentially abundant in promoters of auxin-responsive genes extracted from Arabidopsis thaliana microarray data , and we find a motif that can be interpreted as a refined auxin responsive element predominately positioned in the 250-bp region upstream of the transcription start site . Using an independent data set of auxin-responsive genes , we find in genome-wide predictions that the refined motif is more specific for auxin-responsive genes than the canonical auxin-responsive element . In general , Dispom can be used to find differentially abundant motifs in sequences of any origin . However , the positional distribution learned by Dispom is especially beneficial if all sequences are aligned to some anchor point like the transcription start site in case of promoter sequences . We demonstrate that the combination of searching for differentially abundant motifs and inferring a position distribution from the data is beneficial for de-novo motif discovery . Hence , we make the tool freely available as a component of the open-source Java framework Jstacs and as a stand-alone application at http://www . jstacs . de/index . php/Dispom . Gene regulation is a complex process controlled by many influential components such as the binding of proteins to DNA or the binding of miRNAs to mRNA , RNA editing , splicing of pre-mRNA , mRNA degradation , or post-translational modification . One of the fundamental regulatory steps is the binding of transcription factors ( TFs ) to the promoters of their target genes . TFs influence the initiation of transcription , which in turn affects many subsequent regulatory processes . TFs bind to their binding sites ( BSs ) via a DNA binding domain , and one challenge in computational biology is the identification of transcription factor binding sites ( TFBSs ) in the promoters of target genes . Target regions of TFs can be obtained by a combination of different wet-lab experiments including electrophoretic mobility shift assays ( EMSA ) [1] , DNAse footprinting [2] , ELISA [3] , [4] , ChIP-chip [5] , [6] , ChIP-seq [7] , or expression profiling [8] . However , the regions identified by these methods are large and not limited to TFBSs solely , so de-novo motif discovery tools are typically used for predicting putative TFBSs . These tools take a set of target promoters with unknown binding motif and unknown BSs as input and predict putative binding motifs and the corresponding putative BSs simultaneously . A wealth of de-novo motif discovery tools has been developed over the last decades including , for example , Gibbs Sampler [9]–[11] , MEME [12] , Weeder [13] , Improbizer [14] , DME [15] , DEME [16] , or A-GLAM [17] . These tools differ by the learning principle employed to infer the model parameters and by their capability of learning the position distribution of the BSs from the data . Many de-novo motif discovery tools including Gibbs Sampler [9]–[11] , MEME [12] , Weeder [13] , Improbizer [14] , and A-GLAM [17] use generative learning principles for discovering statistically over-represented motifs from a set of target promoters , i . e . motifs with the highest abundance in the target promoters . However , the discovered motifs often turn out to be similarly over-represented in the rest of the genome , diminishing the specificity of these motifs for the target promoters . In order to overcome this limitation , de-novo motif discovery tools using discriminative learning principles such as DME [15] and DEME [16] have been developed during the last years . These tools utilize an additional control data set expected to contain no or only few BSs of the motif of interest for discovering differentially abundant motifs , i . e . motifs with a high abundance in the set of target promoters and a lower abundance in the control data set . Many de-novo motif discovery tools including Gibbs Sampler [9]–[11] , MEME [12] , Weeder [13] , DME [15] and DEME [16] use a fixed position distribution , chosen to be a uniform distribution in most cases . Motivated by the observation that TFBSs often occur not uniformly distributed along the promoters [10] , [18] , [19] , tools such as Improbizer [14] and A-GLAM [17] have been developed that are capable of learning the positional distribution from the data . In Table 1 , we categorize the above-mentioned tools according to their capability of ( i ) finding differentially abundant motifs and ( ii ) learning the position distribution from the data . None of these tools works perfectly [20] , [21] , but typically de-novo motif discovery tools utilizing a discriminative learning principle outperform those utilizing a generative learning principle [22] , and de-novo motif discovery tools capable of learning the positional preference of TFBSs typically outperform those with a fixed distribution [17] . No algorithm has been developed that combines both features . Here , we introduce Dispom , a discriminative de-novo position distribution motif discovery tool that is capable of modeling the positional preference of TFBSs . Although we focus on the application of Dispom to the de-novo discovery of motifs of TFs in promoter sequences , Dispom may also be used for the discovery of differentially abundant motifs of other origin such as enhancers , silencers , insulators , or miRNA target sites . Similar to other discriminative tools such as DEME or DME , Dispom utilizes a control data set assumed to contain no or few BSs of interest in addition to the target data set . And similar to Improbizer and A-GLAM , Dispom learns the distribution of binding positions from the data simultaneously with the parameters of the motif model . In addition , Dispom uses a heuristic during parameter learning for adapting the length of the binding motif , which is often unknown in advance , and for compensating phase shifts [9] , which frequently occur in many de-novo motif discovery tools . The remainder of this paper is structured as follows . In the section Methods , we describe Dispom and the data used in the subsequent case studies . In section Results , we compare the performance of Dispom based on the motif and on the BS level to that of commonly used de-novo motif discovery tools . For the motif level , we use the metazoan compendium proposed by Linhart et al . [23] , while for the BS level we use 18 benchmark data sets with planted BSs investigating whether the tools are capable of finding motifs with and without positional preference . Finally , we apply Dispom to a data set of promoters of auxin-responsive genes in a cell suspension culture of Arabidopsis thaliana . We compare the motif found by Dispom with the canonical auxin-responsive element and test how specific these motifs are at predicting auxin-responsive genes for an independent data set . Denote a DNA sequence of length by , the nucleotide at position by , the subsequence from position to by , and the reverse complement of by . Dispom is based on the Zero or One Occurrence Per Sequence ( ZOOPS ) model used in many de-novo motif discovery tools [12] , [14] , [16] , [17] . The ZOOPS model uses two hidden variables: Based on any motif model with motif length , any start position distribution , and any flanking sequence model , we obtain the likelihood for sequence given parameters ( 1 ) where the sum runs over all possible combination of values of and , and denotes the vector of model parameters . The probability is defined as ( 2 ) where denotes the probability of using the start position distribution . If the sequence contains no BS , i . e . , if , it is assumed that is generated by ( 3a ) If the sequence contains a BS , then it is assumed that the nucleotides upstream and downstream of the BS are generated by , while the BS is generated by . This yields ( 3b ) Similar to other tools , Dispom uses a position weight matrix as motif model for both DNA strands and a homogeneous Markov model of order 0 as flanking sequence model . In contrast to other tools , Dispom utilizes a mixture of a skew normal and a uniform distribution as position model . The choice is motivated by the observation that a Gaussian distribution decays quite rapidly , and hence , BSs further apart from the mean of the Gaussian are often overlooked . Similarly , the choice of the skew normal instead of a Gaussian distribution is inspired by the expectation that if the mean of the Gaussian is close to the transcription start site ( TSS ) there might be a skew of the distribution . Further details about the model can be found in Text S1 . For predicting BSs in a sequence , we compute the probability ( 4 ) for each possible position of . We also compute these probabilities for each possible position in each sequence of the control data set yielding a background distribution of probabilities . We define the -value of position being erroneously predicted as a BS as the fraction of the probabilities that exceed the probability at position according to the background distribution . We finally define a threshold on the -values , and predict all positions of a sequence with as starting positions of a BS . The goal of de-novo motif discovery is to infer proper parameters of the motif model from a set of target regions and , in case of discriminative approach , an additional set of control regions . We use a labeled data set of sequences where we denote the -th sequence by and its class label by . While tools like MEME and Improbizer use the generative maximum a-posterior ( ) principle for learning the parameters based on a target data set , DME , DEME , and Dispom use a discriminative learning principle , and , hence , utilize an additional control data set . Dispom uses the maximum supervised posterior ( ) principle [24] , [25] , a discriminative Bayesian learning principle . The estimator of is defined by ( 5 ) where the first summand is the logarithm of the conditional likelihood , and second summand is the logarithm of the prior on the parameters with hyper-parameters . For the distribution we choose the ZOOPS model described above , and for the distribution we follow the proposal of [16] and use a homogeneous Markov model of order . As prior , we choose a composite prior that utilizes Gaussian and Gamma distributions for the parameters of the position distribution and Dirichlet priors [26] for the sequence model . The hyper-parameters of these priors use mild assumptions , as for instance uniform pseudo-data for the motif model . Further details about the prior and the hyper-parameters can be found in Text S1 . We obtain estimates of the parameters of Dispom by numerical maximization [27] of Equation ( 5 ) . Since the ZOOPS model implements a non-convex supervised posterior it may get trapped in local optima or saddle points . One prominent type of local optima are so-called phase shifts where the BSs are only covered by a part of the motif model . Besides starting Dispom multiple times , we implement a heuristic that helps reducing this problem and at the same time allows to adjust the motif length . Similar to other models , the ZOOPS model is prone to phase shifts . For this reason , we allow the motif model to be shifted , truncated , or expanded using a heuristic . The complete parameter learning including heuristic steps consists of the following four steps . We ensure that these four steps do not lead to cycles by keeping a history of the performed steps . Text S1 contains further details about the heuristic . For non-convex functions , it is clear that the optimization algorithm can get trapped in local optima or saddle points . Hence , we start the optimization algorithm including the heuristic steps 50 times , and we choose those parameters with the highest supervised posterior . Due to these repeated starts of the numerical optimization , the runtime of Dispom is considerable . In Text S1 , we present a comparison of the runtimes of Dispom and other tools for different data sets with varying numbers of sequences and with varying lengths . A single run of Dispom needs approximately the same run time as Weeder of up to several hours . Conceptually , it is important to note that Dispom , like several other tools , is limited to model at most one BS per sequence , since it is based on the ZOOPS model . In Text S1 , we investigate whether the assumptions of the ZOOPS model hamper Dispom in cases where these assumptions are not met . Second , Dispom only works on sequences of identical length , since the position distribution of the BSs is learned from the data . The length of the sequences can be defined by the user . We successfully tested different promoter lengths up to 1 , 200 bp , but typically the algorithm tends to work better for short sequences than for longer ones . Third , Dispom , like other discriminative de-novo motif discovery tools , requires a control data set for discriminative learning . If no specific control data set is available , one can randomly draw a control data set from the remaining promoters . Typically , we choose a control data set with at least as many sequences as in the target data set . For small target data sets , it is often useful to choose a larger control data set containing e . g . 1 , 000 sequences . Much larger control data sets typically yield only a marginal improvement of accuracy but increase the runtime unnecessarily . For the target data sets , we tested several sizes starting from a few dozen up to few thousand sequences . Typically , larger data sets yield better results than smaller data sets if for each sequence the probability of containing a BS is similar in both data sets . Dispom is implemented in Jstacs ( http://www . jstacs . de ) , an open-source and object-oriented Java framework for statistical analysis and classification of biological sequences . This enables users to apply and extend Dispom easily , e . g . by other sequence or position models , parameter initialization methods , learning principles , or heuristic steps . Prediction performance of different de-novo motif discovery tools is usually compared using the nucleotide recall ( ) and the nucleotide precision ( ) , which are also referred to as nucleotide sensitivity and nucleotide positive predictive value , respectively [20] . Let the true positives be the number of positions correctly predicted to be covered by BSs according to the annotation , let be the number of positions covered by BSs , and let be the number of positions predicted to be covered by BSs . Then , is defined as the fraction of correctly predicted nucleotides out of all nucleotides of all annotated BSs , , and is defined as the fraction of correctly predicted nucleotides out of all nucleotides of all predicted BSs , . and depend on parameters of the tools , such as the threshold . For this reason , the values of and may be very different , and it is hard to compare the performance of different tools using only a single pair of and . Typically , some tools have high values of and low values of , while other tools have low values of and high values of , complicating a one-to-one comparison of their accuracy . Hence , we vary the threshold , which is connected to the number of predictions , and obtain a series of pairs of and for each tool . Plotting these values of against yields the nucleotide precision recall curve , which is more suitable for assessing imbalanced data sets than the commonly used ROC curve [28]–[31] . For the comparison , we use the predictions reported by the tools themselves . All of the tools provide some score or measure of significance together with their predictions , which we use to rank these prediction when computing and for different thresholds . Since , in contrast to Dispom , most tools operate with fixed internal thresholds resulting in a limited maximum , we can only obtain partial curves for these tools , which still provide more information than single pairs of and values . In this subsection , we describe the data sets used for de-novo motif discovery in the results section . First , we briefly describe the metazoan compendium which is initially used for evaluating the performance of Amadeus [23] . Second , we describe the data sets used for comparing the prediction performance of Dispom with existing de-novo motif discovery tools . Third , we describe two data sets of auxin-responsive genes of Arabidopsis thaliana [32] that we use for applying Dispom to a real-life problem where the true motif and the true BSs are unknown . We evaluate the performance of Dispom on the motif level for the 24 data sets of the metazoan compendium with at least one matrix available in TRANSFAC 7 . 0 . To allow for an evaluation of Dispom using more recent versions of TRANSFAC , we make the motifs reported by Dispom for each of the 32 TFBS data sets of the metazoan compendium available at http://www . jstacs . de/index . php/Dispom . In the original benchmark study [23] , the performance of six tools , namely AlignACE , MEME , YMF , Trawler , Weeder , and Amadeus , is compared on the data sets of the metazoan compendium . Each tool is allowed to report two motifs of length 10 and two additional motifs of length 8 . Out of these four motifs , the motif with the smallest normalized euclidean distance [33] is chosen to assess the performance of a tool [23] . The results achieved by the six tools with this procedure are available at http://acgt . cs . tau . ac . il/amadeus/suppl/results_metazoan . html , and we use the reported accuracies in the following comparison . Since Dispom is capable of learning the length of the motif from the input data , we allow Dispom to report two different motifs of learned lengths as opposed to the four motifs considered for the other tools . We obtain the two motifs reported by Dispom for the two different types of control data sets described in subsection Data sets . In Figure 1 , we present the results of this comparison . We find that Dispom discovers the correct motif for 19 of the 24 data sets , whereas Amadeus correctly discovers 17 motifs , Weeder and Trawler discover 11 motifs , YMF and AlignACE discover 7 motifs , and MEME discovers 1 motif . While most of the motifs are discovered by at least three of the tools including Dispom , there are the following notable exceptions . For the data sets “Human-ERa-Kwon-498” , “Human-HNF4a-Odom-1485” , and “Fly-MEF2-Sandmann-211-mapped” , none of the tools considered is capable of discovering the correct motif , which demonstrates the importance of developing improved algorithms for de-novo motif discovery . For the data set “Human-HCC-G2M-Whitfield-350” , Amadeus is the only tool that finds the correct motif , and the correct motif of “Human-p53-Kannan-38” is found only by Weeder . Finally , in two cases , namely “Human-HSF1-Page-333” and “Mouse-MEF2-Blais-26” , Dispom is the only tool that finds the correct motif . Considering the accuracy of the motifs reported by Dispom as measured by the normalized euclidean distance [33] , we find a greater distance compared to other tools for some of the data sets . One explanation for this observation might be that for most of the data sets not all matrices that were used in the original benchmark [23] are available in TRANSFAC 7 . 0 . Summarizing these results , we may state that Dispom performs at least comparable to the best of the existing approaches on the metazoan compendium . Since Dispom is the only tools that finds the correct motif for the data sets “Human-HSF1-Page-333” and “Mouse-MEF2-Blais-26” , we may conclude that Dispom might be a valuable tools for discovering new motifs in data sets for which other tools failed in the past . For testing the efficacy of Dispom , we compare it with commonly used available methods on the same data sets . First , we consider three different aspects of de-novo motif discovery for all tools . We consider the capability of de-novo motif discovery tools of For each of these issues , we consider only one specific example , and we present the remaining results in Figures S1 , S2 , S3 , and S4 . Finally , we provide an overview of the performance of the different de-novo motif discovery tools applied to each benchmark data set . We run all of the programs using default parameters with the following exceptions: if available and not the default , we use switches for searching on both strands , for enabling a position distribution , and for using the ZOOPS model instead of the OOPS model . We start each of the programs – including Dispom – once specifying the correct length of the motif and once with switches for the automatic adaption of motif length . If such a switch is not available , we set the length of the motif to . A list of the calls for all programs is given in Text S1 . In the previous subsection , we compared the performance of Dispom and seven commonly used tools based on 18 data sets , suggesting that Dispom might be useful for finding differentially abundant BSs and their positional preference . In this subsection , we apply Dispom to promoters of auxin-responsive genes with the goal of finding putative TFBSs . Auxin-responsive genes are regulated by a set of TFs commonly called auxin-responsive factors ( ARF ) , which bind to auxin responsive elements ( AuxREs ) that occur in the promoters of those genes . The canonical AuxRE TGTCTC has been identified as a sequence specifically bound by ARF1 using gel mobility shift assays [38] . However , the ARF multi-gene family consists of 23 members [39] , suggesting that AuxREs might differ for different members of ARFs . Indeed , subsequent analyses of 10 members of the ARF family indicate that only the first four nucleotides TGTC are essential for ARF-binding [40] . Analyses of genome-wide expression data are based on the assumptions that co-expressed genes are regulated by the same TFs and the majority of their promoters contains BSs of these TFs . We use expression data sets for searching for a refined AuxRE . We apply Dispom to a set of promoters of genes up-regulated by the plant hormone auxin in Arabidopsis thaliana grown in a cell suspension culture [32] . Figure 6 visualizes the results of Dispom as a sequence logo [41] and the positional preference corresponding to this motif . We find a motif of length 8 bp predominately positioned in the -bp region upstream of the transcription start site . The core motif can be described as TGTSTSBC and can be interpreted as an elongated and modified version of the canonical AuxRE TGTCTC . The presence of the canonical AuxRE TGTCTC in the promoters of a gene is often used as an indicator that this gene is auxin-responsive . For avoiding parameter overfitting , we use an independent test data set for evaluating the discriminative power of the found consensus sequence . We use the seedling data set described in the section Methods as target test data set , and we use the promoters of all remaining genes on the chip as control test data set . Interestingly , the restriction to the first four nucleotides TGTC , considered by some authors to be an improvement over the canonical ARF motif [40] , decreases rather than increases the specificity . In Table 3 , we summarize the results for the canonical AuxRE motif and the TGTSTSBC motif for the 500-bp upstream regions and the 250-bp upstream regions . For a more detailed analysis , we refer the reader to Table S1 . First , we compare the sensitivities and false positive rates of the different consensus sequences using the 500-bp region . We find ( Table 3 , lines 1 and 3 ) that the sensitivity decreases from 32% to 23% when replacing the canonical AuxRE by the refined motif TGTSTSBC . This decrease is clearly visible , but statistically non-significant , with a -value of using the one-sided binomial proportion test . Turning to the false positive rate , we find that it decreases from 23% to 11% when replacing the canonical AuxRE by the refined motif TGTSTSBC . This decrease is highly significant with a -value of using the one-sided binomial proportion test . Hence , the refined motif is slightly less sensitive but significantly more specific than the canonical AuxRE . Next , we compare the sensitivities and false positive rates for the canonical AuxRE in the 500-bp region and the refined motif TGTSTSBC in the 250-bp region . We find ( Table 3 , lines 1 and 4 ) that the sensitivity decreases from 32% to 19% when replacing the canonical AuxRE and the 500-bp region by the refined motif and the 250-bp region , yielding a -value of using the one-sided binomial proportion test . Turning to the false positive rate , we find that it decreases from 23% to 6% , yielding a -value below . This very small -value states that replacing the canonical AuxRE by the refined motif and replacing the 500-bp region by the 250-bp region yields a highly significant decrease of the false positive rate corresponding to a highly significant increase of the specificity . Finally , we assess the two consensus sequences and the two upstream regions using the F-measure and the -value of Fisher's exact test , which both consider the complete contingency table and combine sensitivity and false positive rate , for each of the four lines in Table 3 . We find that combining the canonical motif TGTSTSBC and the 500-bp region yields an F-measure of , which is increased to in case of the refined motif TGTSTSBC and the refined 250-bp region . This reflects the reduction of false predictions by a factor of due to the refined motif and the refined upstream region detected by Dispom . In addition , we find the lowest -value of for the refined motif combined with the refined region . These observations illustrate the potential of combining discriminative de-novo motif discovery with the approach of simultaneously learning the positional distribution . Gene regulation and specifically the binding of TFs to their BSs is of fundamental interest in many areas of genome biology . A combination of experimental and computational methods are typically used for finding putative TFBSs . For computational approaches , two fundamental improvements have been proposed in the last years . On the one hand searching for differentially abundant motifs , and on the other hand learning a position distribution have been shown to be promising in several experiments separately . However , up to now there is no tool combining both improvements . We present Dispom , a new computational tool for the de-novo motif discovery that combines the capability of searching for differentially abundant BSs with the capability of learning the positional preference of the BSs . Dispom includes a heuristic for finding motifs of unknown length . We evaluate Dispom on benchmark data sets of the metazoan compendium and find that Dispom discovers two motifs that could not be found by any of the other tools considered . Additionally , we compare the performance of Dispom with seven commonly used de-novo motif discovery tools based on 18 data sets , and we find that Dispom outperforms these tools . Especially in cases where the correct motif length is not provided , the predictions of Dispom are substantially more accurate than those of traditional de-novo discovery tools indicating that the combination of discriminative learning , inferring a position distribution from the data , and utilizing a heuristic for finding the motif length is beneficial for de-novo motif discovery . Finally , we use Dispom on a set of auxin-responsive genes where the true motif is unknown . We find the motif TGTSTSBC , which can be interpreted as an refined AuxRE , predominantly located in the promoter region of to . Both the refined motif as well as the refined promoter region lead to an improved discrimination of auxin-responsive and non-responsive genes on an independent genome-scale test data set . community as part of the open-source Java library Jstacs ( http://www . jstacs . de ) , which allows an easy application , automation , and extension .
Binding of transcription factors to promoters of genes , and subsequent enhancement or repression of transcription , is one of the main steps of transcriptional gene regulation . Direct or indirect wet-lab experiments allow the identification of approximate regions potentially bound or regulated by a transcription factor . Subsequently , de-novo motif discovery tools can be used for detecting the precise positions of binding sites . Many traditional tools focus on motifs over-represented in the target regions , which often turn out to be similarly over-represented in the entire genome . In contrast , several recent tools focus on differentially abundant motifs in target regions compared to a control set . As binding sites are often located at some preferred distance to the transcription start site , it is favorable to include this information into de-novo motif discovery . Here , we present Dispom a novel approach for learning differentially abundant motifs and their positional preferences simultaneously , which predicts binding sites with increased accuracy compared to many popular de-novo motif discovery tools . When applying Dispom to promoters of auxin-responsive genes of Arabidopsis thaliana , we find a binding motif slightly different from the canonical auxin-response element , which exhibits a strong positional preference and which is considerably more specific to auxin-responsive genes .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results/Discussion" ]
[ "developmental", "biology", "computational", "biology/sequence", "motif", "analysis", "computational", "biology/transcriptional", "regulation", "cell", "biology/cell", "signaling", "genetics", "and", "genomics/plant", "genetics", "and", "gene", "expression", "molecular", "biology/transcription", "initiation", "and", "activation", "plant", "biology", "molecular", "biology/bioinformatics", "plant", "biology/plant", "growth", "and", "development", "computational", "biology", "mathematics/statistics", "genetics", "and", "genomics/bioinformatics" ]
2011
De-Novo Discovery of Differentially Abundant Transcription Factor Binding Sites Including Their Positional Preference
The Notch signaling pathway controls a large number of processes during animal development and adult homeostasis . One of the conserved post-translational modifications of the Notch receptors is the addition of an O-linked glucose to epidermal growth factor-like ( EGF ) repeats with a C-X-S-X- ( P/A ) -C motif by Protein O-glucosyltransferase 1 ( POGLUT1; Rumi in Drosophila ) . Genetic experiments in flies and mice , and in vivo structure-function analysis in flies indicate that O-glucose residues promote Notch signaling . The O-glucose residues on mammalian Notch1 and Notch2 proteins are efficiently extended by the addition of one or two xylose residues through the function of specific mammalian xylosyltransferases . However , the contribution of xylosylation to Notch signaling is not known . Here , we identify the Drosophila enzyme Shams responsible for the addition of xylose to O-glucose on EGF repeats . Surprisingly , loss- and gain-of-function experiments strongly suggest that xylose negatively regulates Notch signaling , opposite to the role played by glucose residues . Mass spectrometric analysis of Drosophila Notch indicates that addition of xylose to O-glucosylated Notch EGF repeats is limited to EGF14–20 . A Notch transgene with mutations in the O-glucosylation sites of Notch EGF16–20 recapitulates the shams loss-of-function phenotypes , and suppresses the phenotypes caused by the overexpression of human xylosyltransferases . Antibody staining in animals with decreased Notch xylosylation indicates that xylose residues on EGF16–20 negatively regulate the surface expression of the Notch receptor . Our studies uncover a specific role for xylose in the regulation of the Drosophila Notch signaling , and suggest a previously unrecognized regulatory role for EGF16–20 of Notch . Notch signaling is a juxtacrine cell-cell communication pathway with broad roles in animal development and adult tissue homeostasis [1] , [2] . Both gain- and loss-of-function mutations in Notch pathway components cause human disease [3]–[5] , and therapeutic approaches to alter the activity of Notch signaling are a subject of intense research and development [6] . The extracellular domains of Notch receptors contain a large number ( up to 36 ) of EGF repeats . Each EGF repeat contains six cysteine residues , which are linked to each other via three disulfide bonds [7] . Of the several forms of O-linked carbohydrates found on Notch EGF repeats [8]–[10] , two have been shown to be required for Notch signaling in both flies and mammals: O-fucose and O-glucose [11]–[16] . Addition of N-acetylglucosamine ( GlcNAc ) to O-fucose by Fringe glycosyltransferases modulates Notch signaling in several contexts [17] , [18] . O-linked glucose is attached to serine residues in the consensus sequence C1-X-S-X- ( P/A ) -C2 by the protein O-glucosyltransferase Rumi/POGLUT1 [8] , [14] , [19] . Rumi is required for both fly and mammalian Notch signaling at a step downstream of ligand-binding [14] , [15] , [20] . In vivo structure-function analyses indicate that all of the 18 Rumi target sequences in Drosophila Notch ( dNotch ) contribute to Notch activation , with O-glucose sites on EGF10-15 playing a more important role than others [20] . O-glucose can serve as a substrate for additional sugar modifications to generate xylose-xylose-glucose trisaccharides [8] , [19] , [21] . The human enzymes responsible for the addition of xylose to O-glucosylated Notch EGF repeats have recently been identified: glucoside xylosyltransferase ( GXYLT ) 1 and GXYLT2 add the first xylose and xyloside xylosyltransferase ( XXYLT1 ) adds the second ( Figure 1A ) [22] , [23] . Thus far , no functional studies have been performed to analyze the role of xylosylation in Notch signaling . Here , we identify the Drosophila glucoside xylosyltransferase which adds xylose to O-glucosylated EGF repeats and show that this enzyme negatively regulates Drosophila Notch signaling in certain contexts . We use a combination of mass spectrometry , genetic and in vivo mutational studies to show that the functionally important sites of xylosylation reside in EGF repeats 16–20 of Notch , and that xylose negatively regulates the surface expression of Notch . Given that O-glucose positively regulates Notch signaling in all contexts studied so far [20] , negative regulation of Notch signaling by xylosylated O-glucose glycans provides an example of how the strength of a signaling pathway can be fine-tuned by stepwise addition of carbohydrate molecules to a receptor . Using homology searches , we identified two novel Drosophila proteins ( CG9996 and CG11388 ) homologous to human Notch xylosyltransferases . Sequence comparison between these two fly proteins and human Notch xylosyltransferases indicated that CG9996 is the only close fly homolog of human GXYLT1 and GXYLT2 , whereas CG11388 bears much more sequence identity with XXYLT1 ( Figure 1B and Figure S1 ) . We named CG9996 Shams , a companion and muse for the poet Rumi . To test whether Shams can add xylose to glucose similar to its human homologs , we first performed in vitro glycosyltransferase assays by using purified , recombinant Shams and synthetic lipophilic acceptors ( Figure 1C ) . We found that Shams can indeed add xylose specifically to a synthetic acceptor harboring a glucose residue ( Glc-β1-R ) but not to one harboring a xylose-glucose disaccharide ( Xyl-α1 , 3-Glc-β1-R ) ( Figure 1C ) . To further examine the substrate specificity of Shams , we used glucose or xylose residues attached to para-nitrophenol ( pNP ) in α- or β-linkage as the acceptor for the xylosyltransferase activity of Shams and found that Shams can only transfer xylose to Glc-β1-pNP ( Figure 1C ) . We also performed similar assays to determine the donor substrate specificity of Shams and found that Shams is able to transfer xylose , but not glucose or galactose , to the Glc-β1-R acceptor ( Figure 1D ) . To examine whether Shams is able to add xylose to O-glucosylated EGF repeats of Notch , we assayed purified Shams using a fragment of Drosophila Notch ( EGF16–20 ) harboring several O-glucosylation sites , expressed in and purified from Sf9 cells . Mass spectrometric analysis of a glycosylated peptide of EGF16 showed that Sf9 cells produce a mixture of glucose and xylose-glucose in a ratio of about 3 to 1 at this site . Sf9 cells apparently highly glucosylate EGF16 , but show limited xylosylation capacity . In vitro , both Shams and human GXYLT1 can transfer xylose to O-glucosylated EGF16 , changing the ratio in favor of the xylosylated form ( Figure 1E ) . Taken together , these experiments show that Shams functions as a glucoside xylosyltransferase capable of adding the first xylose to O-glucose on Notch EGF repeats , similar to its human homologs GXYLT1 and GXYLT2 [23] . Mass spectrometry on mouse Notch1 expressed in several mammalian cell lines has shown that xylose-xylose-glucose trisaccharide is the dominant form of O-glucose glycans on all mouse Notch1 EGF repeats harboring an O-glucosylation site , although the stoichiometry varies among different EGF repeats [8] , [19] . To determine the distribution of xylosylated O-glucose glycans on Drosophila Notch , we performed systematic mass spectrometric analyses of Drosophila Notch expressed in Drosophila S2 cells . We find that while O-glucose is found at all predicted sites analyzed , xylose is only detected on EGF14–20 , and xylose-xylose-glucose trisaccharides are only detected at EGF16 and EGF18 ( Figure 1F and 1G and Figure S2 ) . Therefore , unlike mammalian Notch1 , addition of xylose to O-glucose appears to be limited to a subset of EGF repeats of the Drosophila Notch . To examine the role of xylosylation in Drosophila Notch signaling , we performed genetic experiments on two independent alleles of shams ( Figure 2A ) . Flies homozygous for the piggyBac insertion shamse01256 ( shamsPB/PB ) are viable at 25°C and do not exhibit any adult phenotypes besides a loss of the posterior cross-vein in 20% of the flies ( Figure 2C; compare to 2B ) . When raised at 30°C , 56% of shamsPB/PB flies lose the distal portion of the L5 wing vein ( Figure 2D ) , similar to the phenotype observed in gain-of-function Abruptex alleles of Notch ( NAx ) [24] . Precise excision of this piggyBac insertion fully reverts the phenotype , indicating that the observed loss of the wing vein is due to the insertion ( Figure 2E ) . We also generated a null allele lacking 97% of the Shams coding region by using FLP/FRT-mediated recombination on two piggyBac insertions in the region ( Figure 2A ) [25] . Animals homozygous or hemizygous for the null allele shamsΔ34 survive to adulthood and exhibit a 100% penetrant loss of the L5 wing vein at 25°C ( Figure 2F ) . At 30°C , shamsΔ34/Df animals are semi-lethal and all the escapers exhibit partial loss of multiple wing veins ( Figure 2G ) . The wing vein loss phenotype can be rescued by overexpression of shams cDNA ( Figure 2H and 2I ) , by providing a shamsgt-wt genomic transgene that contains the shams locus ( Figure 2J ) , or by an HA-tagged version of the shams genomic transgene ( shamsgt-wt-HA , data not shown ) . However , even though the shamsΔ34 allele also affects CG11836 ( Figure 2A ) , a genomic transgene containing this gene ( CG11836gt-wt ) does not rescue the wing vein phenotypes of the shamsΔ34/Df animals ( Figure 2K ) . Of note , each of the shamsgt-wt and CG11836gt-wt genomic transgenes partially rescues the semi-lethality of these animals , indicating that both transgenes are functional . These observations indicate that loss of shams results in a wing vein loss phenotype . shamsΔ34/Df animals raised at 30°C also lose the ocellar and postvertical bristles in the head ( Figure 2M; compare to 2L ) similar to NAx alleles [26] . Again , this phenotype can be fully suppressed by a shams genomic transgene ( Figure 2N ) . Together , these data indicate that loss of shams results in phenotypes reminiscent of Notch gain-of-function phenotypes . To examine whether the observed phenotypes are indeed due to increased Notch signaling , we performed genetic interaction studies . As reported previously , loss of one copy of Notch results in wing margin defects and wing vein expansion ( Figure 3A ) [27] . Homozygosity for the shamsPB allele suppresses the N55e11/+ haploinsufficient phenotypes ( Figure 3B and Figure S3 ) . In a reciprocal experiment , we find that shamsPB enhances the wing vein loss observed in the NAx-E2 dominant gain-of-function allele ( Figure 3C and 3D ) . Together , these observations indicate that Shams decreases the activity of Notch , potentially by adding xylose to O-glucose residues on Notch EGF repeats . To examine the effects of increased xylosylation on Notch signaling , we overexpressed an HA-tagged version of human GXYLT1 in developing fly wings . We observed margin scalloping in 40% of adult wings at 30°C ( Figure 3E ) and a mild wing vein expansion at 25°C and 30°C ( Figure 3E and Figure S4 ) . Increasing the Notch gene dosage suppressed the wing margin loss caused by GXYLT1-HA overexpression ( Figure 3F ) , indicating that the phenotype is due to decreased Notch signaling . Overexpression of GXYLT1-HA in a shamsPB/PB background resulted in the mutual suppression of the wing scalloping and vein loss phenotypes observed in each genotype ( Figure 3G; compare to Figure 3E and Figure 2D ) . Simultaneous overexpression of GXYLT1-HA and Shams resulted in an enhancement of the wing vein and margin defects ( Figure 3H ) . Together , these data indicate that Shams and GXYLT1 are functionally homologous , although GXYLT1 seems to be more potent than Shams ( see Figure 2H ) , and suggest that increased xylosylation negatively regulates Notch signaling . Ectopic expression of human XXYLT1-HA resulted in a dramatic loss of margin and thickening of the wing veins , consistent with severe loss of Notch signaling ( Figure 3I and Figure S4 ) . In accordance with their enzymatic functions ( Figure 1 ) [22] , the XXYLT1-HA overexpression phenotypes are completely suppressed in a shamsPB/PB background with 100% penetrance ( Figure 3J; n = 10 ) . This indicates that the observed phenotypes are due to the enzymatic activity of XXYLT1 , as they are dependent on Shams to generate xylose-glucose disaccharide substrates . The data also agree with the homology searches which indicate that Shams is the only GXYLT in flies . To more directly show that the phenotypes caused by overexpressing human xylosyltransferases are due to a loss of Notch signaling , we overexpressed these enzymes along the antero-posterior axis of the developing wing discs by using the patched ( ptc ) -GAL4 driver . Again , both enzymes showed phenotypes compatible with loss of Notch signaling in the adult wings , with XXYLT1-HA phenotypes being stronger than the GXYLT1-HA phenotypes ( Figure 3K and 3L ) . Overexpression of XXYLT1-HA resulted in loss of wing margin tissue and a collapse of the L3 and L4 wing veins without affecting proliferation or apoptosis ( Figure 3L and S5A–D′ ) . These phenotypes are specific to XXYLT1 overexpression , because they are fully rescued in a shams background ( Figure S5E ) . Antibody staining of the third instar wing imaginal discs in these animals showed that the Notch downstream target Cut is either decreased ( GXYLT1-HA ) or lost ( XXYLT1-HA ) in the ptc-GAL4 domain upon human xylosyltransferase overexpression ( Figure 3M–3N′ ) . Altogether , these observations indicate that xylosylation negatively regulates Notch signaling . Notch transgenes harboring serine-to-alanine mutations in all or most O-glucosylation sites show a temperature-sensitive loss of Notch signaling , similar to rumi animals [14] , [20] . However , when smaller subsets of the O-glucosylation sites are mutated and the animals are raised at 25°C or lower , the negative effects of loss of O-glucose on Notch is significantly decreased [20] . If loss of xylose on specific EGF repeats results in increased Notch signaling , these mutations should recapitulate the shams mutant phenotypes , as loss of O-glucose precludes the addition of xylose . To test this , we generated animals that lack endogenous Notch but are rescued by one copy of Notch genomic transgenes carrying mutations in various subsets of O-glucosylation sites ( Figure 4A ) [20] . Serine-to-alanine mutations in EGF10–15 or EGF24–35 did not result in loss of wing vein , similar to a wild-type Notch transgene ( Figure 4B , 4C and 4E ) . However , O-glucose mutations in EGF16–20 resulted in a partial loss of wing veins L2 , L4 , and L5 ( Figure 4D ) , similar to but somewhat stronger than the shams null phenotypes ( Figure 2F and 2G ) . N54l9/Y; Ngt-16_20/+ animals also exhibited head bristle defects similar to shams mutants ( Figure S6 ) . These observations nicely match our mass spectrometry data ( Figure 1G ) and indicate that xylosylation of EGF16–20 plays a negative regulatory role in Drosophila Notch signaling . One prediction from the above conclusion is that mutations in EGF16–20 should suppress the loss of Notch signaling caused by the overexpression of human xylosyltransferases . To test this , we overexpressed GXYLT1-HA and XXYLT1-HA in genetic backgrounds lacking endogenous Notch and rescued by one copy of a wild-type or O-glucose mutant Notch transgenes . Overexpression of these enzymes in animals with one copy of a wild-type Notch transgene raised at 25°C results in phenotypes very similar to their overexpression in a wild-type background raised at the same temperature ( Figure 4F and 4J and Figure S4 ) . Among the mutant Notch transgenes , Ngt-16_20 is the only one which could fully suppress both GXYLT1-HA and XXYLT1-HA overexpression phenotypes ( Figure 4G–4I and 4K–4M; n = 10 for each genotype;100% penetrance ) . Of note , in Ngt-10_15 and Ngt-24_35 backgrounds the XXYLT1-HA overexpression phenotype is even enhanced , most likely due to the negative effect of loss of these O-glucose residues on Notch signaling ( Figure 4K and 4M ) [20] . Altogether , these data indicate that Shams regulates Notch signaling via an enzymatic mechanism , and that the O-glucosylation sites in Drosophila Notch EGF16–20 are the biologically-relevant targets of xylosylation by Shams . To examine the effects of loss of shams on Notch localization , we performed Notch surface staining on larval wing imaginal discs and pupal wings harboring shams mutant clones . Loss of shams does not affect the surface expression of Notch in third instar wing imaginal disc ( Figure S7A–A″ ) . However , more Notch protein is present in and at the surface of shams mutant cells in the pupal wing ( Figure 5A and 5A′ and Figure S7B–B′ ) . We also sought to determine whether mutating the O-glucose sites on EGF16–20 of Notch results in increased cell surface levels of Notch . To this end , we generated Mosaic Analysis with a Repressible Cell Marker ( MARCM ) clones [28] of a protein-null allele of Notch in a background harboring one copy of the wild-type or an EGF16–20 mutant Notch , similar to what we have described before for other mutant Notch transgenes [20] . In these animals , heterozygous cells have both endogenous Notch and a copy of our transgene , but cells in the mutant clones only harbor a copy of the transgene . Accordingly , the level of Notch expressed from the wild-type transgene in the clones is less than that in the heterozygous cells ( Figure 5B and 5B′ ) , in agreement with our previous report [20] . Clones of Notchgt-16_20 in larval wing disc do not show an increase in Notch surface expression ( data not shown ) . However , in the pupal wing , the level of surface Notch expressed from the Notchgt-16_20 transgene in the clones is significantly increased compared to that expressed from the wild-type Notchgt-wt transgene ( Figure 5C and 5C′; compare to 5B′ ) . One potential explanation for the difference between the effects of loss of Notch xylosylation in pupae versus larvae could be different levels of Shams expression at these stages . Indeed , Western blot confirmed higher Shams expression in pupae compared to third instar larvae ( Figure 5D ) . In agreement with a role for xylose in surface expression of Notch , overexpression of human XXYLT1 results in a significant decrease and overexpression of human GXYLT1 results in a mild and partially penetrant decrease in Notch surface expression in the larval wing imaginal discs ( Figure S8 ) . Altogether , these observations suggest that addition of xylose residues to EGF16–20 of Notch decreases the availability of Notch at the cell surface . Our data indicate that there are significant differences between the ways O-glucose monosaccharides and their extended ( xylosylated ) form regulate Drosophila Notch signaling . First , O-glucosylation promotes Notch signaling [14] , [20] , but xylosylation inhibits Notch signaling . Secondly , loss of O-glucosylation affects Notch signaling in all contexts studied so far [14] , [20] , [29] , but loss of xylosylation only affects Notch signaling in certain contexts , i . e . wing vein development and head bristle formation . Finally , O-glucose residues on all EGF repeats contribute to the Notch signal strength in redundant and/or additive fashions [20] , but xylose residues seem to be only required on a subset of Notch EGF repeats . In other words , unlike glucose residues which function globally on the Notch extracellular domain to promote Notch signaling in various contexts [14] , [20] , [29] , xylose residues function locally on a specific region of Notch to decrease signaling in certain contexts . Thus , our data show that the strength of the Notch signaling pathway can be fine-tuned by altering the distribution and relative levels of O-glucose monosaccharide and their xylose-containing extended forms on the Notch receptor . Since xylose-xylose-glucose glycans on EGF repeats are the only known glycans in animals with a terminal xylose [30] , [31] , our data strongly suggest that at least in Drosophila , terminal xylose residues play a fairly specific role in regulating the Notch signaling pathway . Our observations provide compelling evidence that the negative effects of xylosyltransferases on Drosophila Notch signaling are primarily mediated via their enzymatic activity . First , serine-to-alanine mutations in the O-glucosylation sites in EGF16–20 of the Drosophila Notch recapitulate the wing vein and head bristle phenotypes of shams . Since loss of the protein O-glucosyltransferase Rumi results in loss of Notch signaling [14] , [20] , the observed gain of Notch signaling phenotypes cannot be due to loss of O-glucose from these EGF repeats , but are very likely due to the loss of xylose normally attached to the O-glucose . Secondly , mutating the O-glucose sites of EGF16–20 fully suppresses the Notch loss-of-function phenotypes caused by GXYLT1 and XXYLT1 overexpression , strongly suggesting that these phenotypes result from the addition of xylose by these enzymes to O-glucose on these EGF repeats . Lastly , although the Notch loss-of-function phenotypes of XXYLT1 overexpression are much more severe than those caused by GXYLT1 overexpression , loss of shams fully suppresses the XXYLT1 overexpression phenotypes but only partially suppresses the GXYLT1 overexpression phenotype . These data fully match the enzymatic activities of these proteins: GXYLT1 and Shams both add the first xylose to O-glucose and function in parallel , but XXYLT1 adds the second xylose and therefore completely depends on the activity of GXYLT1/Shams . A hallmark of Notch receptors is the presence of many EGF repeats in their extracellular domain . However , the functional importance of only a handful of Notch EGF repeats has been elucidated . Specifically , EGF11 and 12 are necessary for ligand-binding [32] , EGF 24 , 25 , 27 and 29 negatively regulate Notch signaling , likely through opposing ligand binding [33] , and EGF8 is required for binding of Notch to Serrate but not Delta [34] . By showing that EGF16–20 ( or a subset of them ) negatively regulate Drosophila Notch signaling in a xylose-dependent manner , our data assign a function to these EGF repeats . Loss of shams or loss of xylosylation on EGF16–20 affects surface expression of Notch in the pupal wing and results in defects in wing vein formation , a process primarily regulated in the pupal stage . In contrast , loss of shams does not affect Notch expression in the larval wing discs , or the wing margin formation , which primarily occurs at the larval stage . Together , these observations suggest that Notch signaling in the pupal wing is more sensitive to loss of xylosylation compared to the larval wing disc , likely because of the relatively higher levels of Shams expression in the pupal stage . Nevertheless , given the loss of Notch signaling observed upon GXYLT1 and XXYLT1 overexpression , the larval wing discs likely have the machinery required to recognize and respond to xylose on Notch . Increased availability of Notch at the cell surface of shams and Notchgt-16_20 mutant clones in the pupal wing suggests a molecular mechanism for the role of xylose in regulating Notch signaling . However , it remains to be seen whether other steps of Notch signaling , including ligand-binding and cis-inhibition are also affected by the loss of xylose . Although cell type variability exists , all O-glucosylated EGF repeats of the mouse Notch1 harbor xylose-xylose-glucose trisaccharides at high stoichiometry [19] . Our previous work [14] and current mass spectrometry experiments indicate that all of the EGF repeats of the Drosophila Notch with a consensus Rumi target motif analyzed thus far ( 16 out of 18 ) harbor O-glucose . However , we find that xylose is only added to a subset of O-glucosylated Notch EGF repeats in Drosophila . This difference might result from different efficiency of the mammalian versus Drosophila xylosyltransferases in vivo despite their similar in vitro levels of activity ( Figure 1E ) . Indeed , overexpression of human GXYLT1 , but not its Drosophila homolog Shams , results in Notch loss-of-function phenotypes in the wing , suggesting that GXYLT1 is more potent in adding xylose to Notch in vivo . Moreover , the difference between the distribution of ( xylose ) -xylose-glucose saccharides on fly and mammalian Notch could in part be due to the presence of two GXYLT enzymes in mammals instead of only one ( Shams ) in flies . Regardless of the mechanisms underlying the observed differences , it will be of great interest to determine whether alteration of the level or distribution of xylose-xylose-glucose saccharides on Notch receptors can modulate the strength of mammalian Notch signaling as well . The following strains were used in this study: y w , y w; D/TM6 , Tb1 , w; nocSco/CyO , w; nocSco/CyO; TM3 , Sb/TM6 , Tb1 , w; CyO , P{FRT ( w+ ) Tub-PBac\T}2/wgSp-1 , y w hsFLP; Dr/TM3 , apterous-GAL4 , patched-GAL4 , y w N54l9 FRT19A/FM7 , N55e11/FM7c , NAx-E2 , Df ( 3R ) BSC494/TM6C , Sb1 , w; PBac{RB}CG999601256 , attVK22 , UAS-CG8::GFP ( Bloomington Stock Center ) , w; PBac{RB}CG11836e01985 ( Exelixis ) , shamsrev , shamsΔ34/TM6 , Tb1 , shamsgt-wt-attVK22 , shamsgt-wt-HA-attVK22 , CG11836gt-wt-attVK22 , UASattB-shams-HA-VK22 , UASattB-GXYLT1-HA-VK22 , UASattB-XXYLT1-HA-VK22 , Ubx-FLP FRT19A/FM7; Act-GAL4 UAS-CD8::GFP/CyO ( this study ) , Ngt-wt-attVK22 , Ngt-10_15-attVK22 , Ngt-16_20-attVK22 , Ngt-24_35-attVK22 [20] , y w Ubx-FLP tub-GAL4 UAS-GFPnls-6X-Myc; FRT82B y+ tub-GAL80/TM6 , Ubx [14] , vas-int-ZH-2A; attVK22 [35] , nubbin-GAL4 ( Georg Halder ) . All crosses were performed on standard media and incubated at described temperatures . To test the effect of Notch mutations on the adult wing , w; Ngt-wt , Ngt-10_15 , Ngt-16_20 , or Ngt-24_35 males were crossed to N54l9/FM7 , B females . B+ male progeny were scored for wing and head bristle defects . To determine the effect of Notch mutations on GXYLT1-HA and XXYLT1-HA overexpression in the wing , B+ , Cy+ male progeny were scored from apterous-GAL4 UAS-GXYLT1-HA/CyO or nubbin-GAL4 UAS-XXYLT1-HA/CyO males crossed with N54l9/FM7; Ngt-wt/+ , N54l9/FM7; Ngt-10_15/+ , N54l9/FM7; Ngt-16_20/+ , or N54l9/FM7; Ngt-24_35/+ females . To generate MARCM clones of Notch harboring a Notch genomic transgene , Ubx-FLP FRT19A/Y; Act-GAL4 UAS-CD8::GFP/CyO males were crossed to Notch54l9/FM7; Ngt-wt/+ or N54l9/FM7; Ngt-16_20/+ females . All MARCM crosses were set at room temperature and transferred to 30°C at L1–L2 instar stage . For further details on Drosophila genetics and other techniques used in this study , please see Text S1 . For recombinant expression of Shams ( CG9996 ) the predicted C-terminal lumenal domain starting from Gln23 was amplified from Drosophila w1118 cDNA and cloned into the pFast-Bac1 vector ( Invitrogen ) encoding the HBM-secretion signal followed by the Protein A coding sequence , as described for human GXYLT1 [23] . Constructs were expressed in Sf9 insect cells using the Bac-to-Bac System ( Invitrogen ) and secreted proteins were purified by IgG-Sepahrose-6 Fast Flow beads ( GE-Healthcare ) as described [23] . Activity assays were performed on bead-coupled enzyme in the presence of radiolabeled UDP-sugars as described before [23] , except that samples were incubated for 2 h at 27°C . To measure activity on EGF repeats , the Notch EGF16–20 fragment was expressed in Sf9 insect cells , purified by Nickel affinity chromatography , and used in an in vitro assay followed by mass spectrometric analysis as described [22] . A construct encoding EGF1–36 from Drosophila Notch with a C-terminal 3X-FLAG tag ( EGF1–36-FLAG3 , generously provided by Dr . Ken Irvine ) was expressed in Drosophila S2 cells . EGF1–36-FLAG3 was purified from medium , reduced and alkylated , and subjected to in-gel protease digests as described [36] . O-Glucose modified glycopeptides were identified by neutral loss of the glycans during collision-induced dissociation ( CID ) using nano-LC-MS/MS as described [19] . Dissection and staining were performed by using standard methods . For surface staining , third instar larval imaginal discs and pupal wings were dissected and incubated with anti-Notch antibody in the absence of detergent . Antibodies used were mouse α-Cut ( 2B10 ) 1∶500 , mouse anti-Notch ( C458 . 2H ) 1∶100 , mouse α-Delta ( C594 . 9B ) 1∶100 ( Developmental Studies Hybridoma Band ) , goat α-HA 1∶50 ( GenScript ) , mouse α-phosphorylated Histone H3 ( Ser10 ) 1∶100 , rabbit α-Cleaved Caspase 3 ( Asp175 ) 1∶50 ( Cell Signaling ) , donkey-α-goat-Dylight549 1∶500 , donkey-α-mouse-Cy5 1∶500 , goat-α-mouse-Cy3 1∶500 , donkey-α-Rabbit-Cy5 1∶500 , donkey-α-Guinea Pig-Cy3 1∶500 ( Jackson ImmunoResearch Laboratories ) . Confocal images were scanned using a Leica TCS-SP5 microscope and processed with Amira5 . 2 . 2 . Images were processed with Adobe Photoshop CS5; Figures were assembled in Adobe Illustrator CS5 . Third instar larvae and pupae aged 22 hours after puparium formation ( APF ) were collected from VK22 control animals or animals harboring one copy of the shamsgt-wt-HA genomic transgene . Protein extracts were generated using RIPA buffer ( Boston BioProducts ) and a Protease Inhibitor Cocktail ( Roche ) . Protein extracts were separated on 12% acrylamide gel and transferred to PVDF membrane . Blots were probed with goat α-HA 1∶300 ( GenScript ) , mouse α-Tubulin 1∶1000 ( Santa Cruz Biotech ) , donkey-α-goat-HRP 1∶2000 and goat α-mouse-HRP 1∶2000 antibodies ( Jackson ImmunoResearch Laboratories ) . Western blots were developed using Pierce ECL Western Blotting Substrates ( Thermo Scientific ) and imaged with an LAS4000 GE ImageQuant Imager . Three independent experiments showed the same result .
In multi-cellular organisms , neighboring cells need to communicate with each other to ensure proper cell fate decisions and differentiation . Signaling through the Notch receptors is the primary means by which local cell-cell communication is accomplished in animals . Given the broad usage of Notch signaling in animals and the host of human disease caused by Notch pathway misregulation , sophisticated mechanisms are required to adjust the strength of Notch signaling in each context . We have previously shown that addition of glucose residues to the Notch receptor promotes Notch signaling . Since these glucose residues on Notch can be extended by addition of xylose residues , we sought to determine whether xylose also plays a role in the regulation of Notch signaling . In contrast to glucose , we determine that xylose residues decrease Notch signaling in certain contexts by controlling Notch surface levels . Moreover , the xylose residues reside in a specific domain of Notch , unlike the glucose residues which are distributed throughout the Notch extracellular domain . Our data provide an example of signaling pathway regulation by altering the distribution of the short or elongated forms of a saccharide on a receptor protein , and offer a potential avenue for modulating Notch signaling as both a therapeutic modality and a tool in regenerative medicine .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "animal", "models", "biochemistry", "developmental", "biology", "pattern", "formation", "cell", "fate", "determination", "animal", "genetics", "model", "organisms", "organism", "development", "drosophila", "melanogaster", "signaling", "molecular", "development", "gene", "identification", "and", "analysis", "genetics", "molecular", "genetics", "biology", "glycobiology", "cell", "differentiation" ]
2013
Negative Regulation of Notch Signaling by Xylose
This study systematically reviews the literature on the occurrence , incidence and case fatality rate ( CFR ) of invasive nontyphoidal Salmonella ( iNTS ) disease in Africa from 1966 to 2014 . Data on the burden of iNTS disease in Africa are sparse and generally have not been aggregated , making it difficult to describe the epidemiology that is needed to inform the development and implementation of effective prevention and control policies . This study involved a comprehensive search of PubMed and Embase databases . It documents the geographical spread of iNTS disease over time in Africa , and describes its reported incidence , risk factors and CFR . We found that Nontyphoidal Salmonella ( NTS ) have been reported as a cause of bacteraemia in 33 out of 54 African countries , spanning the five geographical regions of Africa , and especially in sub-Saharan Africa since 1966 . Our review indicates that NTS have been responsible for up to 39% of community acquired blood stream infections in sub-Saharan Africa with an average CFR of 19% . Salmonella Typhimurium and Enteritidis are the major serovars implicated and together have been responsible for 91%% of the cases of iNTS disease , ( where serotype was determined ) , reported in Africa . The study confirms that iNTS disease is more prevalent amongst Human Immunodeficiency Virus ( HIV ) -infected individuals , infants , and young children with malaria , anaemia and malnutrition . In conclusion , iNTS disease is a substantial cause of community-acquired bacteraemia in Africa . Given the high morbidity and mortality of iNTS disease in Africa , it is important to develop effective prevention and control strategies including vaccination . Nontyphoidal Salmonellae ( NTS ) are a major cause of food borne infections throughout the developed and developing world [1] . Although infection most often results in self-limited acute gastroenteritis , NTS have been identified as a major cause of invasive bacterial infections in infants and young children in sub-Saharan Africa and HIV-infected individuals of all ages [2 , 3] . Invasive NTS ( iNTS ) disease is recognized as a problem in developed countries in young infants , the elderly and immunocompromised [4] . iNTS disease is caused mainly by Salmonella enterica serovars Typhimurium and Enteritidis [5 , 6] . NTS gastroenteritis is generally understood to be acquired from animal reservoirs , unlike Salmonella Typhi and Salmonella Paratyphi , where the only recognized reservoir is man . Transmission of gastroenteritis-causing NTS to humans can occur by many routes , including consumption of animal food products , especially eggs , poultry , undercooked meat , produce contaminated with animal waste , contact with animals or their environment , and contaminated water [7 , 8] . The African strains responsible for iNTS disease are characterized by genome degradation and appear to be increasingly adapted to an invasive lifestyle [9] . The relative role of animal reservoirs and human to human transmission of strains causing iNTS disease is unclear [10 , 11] . iNTS disease is diagnosed definitively by blood or bone marrow culture , usually with low sensitivity [12] . It is also impossible to diagnose using clinical symptoms alone due to the lack of pathognomonic features [5 , 6] . There is currently no commercially-available rapid diagnostic test for iNTS disease . Bacteraemia is an important cause of severe and often fatal disease globally , especially in developing countries ( where it substantially contributes to childhood deaths ) [2 , 13] . Invasive forms of Salmonella disease include enteric fevers ( typhoid and paratyphoid fevers ) and NTS bacteraemia and are important causes of morbidity and mortality in Asia and Africa [3 , 14 , 15] . In a review article of bacteraemia in Africa from 2010 , NTS was found to be responsible for 29 . 1% of all bloodstream infections [16] . Nevertheless , the data on NTS bacteraemia in Africa are limited with no current aggregate data , so it has been difficult to estimate the burden of NTS bacteraemia in Africa . Such information is vital for planning and implementing cost-effective solutions [17] to tackle iNTS disease , especially in Africa with its fragile health-care systems . Infectious diseases are a major obstacle to human development in the African region with people suffering from an extensive range of potentially preventable and treatable conditions including invasive Salmonella disease . Adequate evidence on the burden of Salmonella infections would enable health policy makers to make informed decisions on the need for vaccines against Salmonella infections in their country and region . The aim of the present study is to address this knowledge gap by conducting a review of the available reports on NTS bacteraemia in Africa and to describe the epidemiology of the disease; in order to inform effective prevention and control strategies , including vaccination . Specifically , the study describes all the geographical locations in Africa that have reported iNTS disease cases , and determines incidence and proportion of bacteraemia caused by NTS ( by whole population and patient groups ) , together with risk factors and CFR of NTS bacteraemia . We searched PubMed ( including MEDLINE ) and Embase . The search string used was: ( algeria OR angola OR benin OR botswana OR “burkina faso” OR burundi OR cameroon OR “cape verde” OR “central african republic” OR chad OR comoros OR “ivory coast” OR “cote d ivoire” OR congo OR djibouti OR egypt OR “equatorial guinea” OR eritrea OR ethiopia OR gabon OR gambia OR ghana OR guinea OR “guinea bissau” OR kenya OR lesotho OR liberia OR libya OR madagascar OR malawi OR mali OR mauritania OR mauritius OR morocco OR mozambique OR namibia OR niger OR nigeria OR rhodesia OR rwanda OR “sao tome” OR senegal OR seychelles OR “sierra leone” OR somalia OR “south africa” OR sudan OR swaziland OR tanzania OR togo OR tunisia OR uganda OR “western sahara” OR zambia OR zimbabwe OR Africa ) AND ( fever OR fevers OR bacteremia OR bacteremias OR bacteremic OR bacteraemia OR bacteraemias OR bacteraemic OR septicemia OR septicemias OR septicemic OR septicaemia OR septicaemias OR septicaemic OR salmonella OR salmonellas OR salmonellae OR “blood stream infection” OR “blood stream infections” OR “blood stream pathogen” OR “blood stream pathogens” OR febrile ) AND ( infant* OR child* OR adolescent* OR adult* OR patient OR patients OR human OR travel* OR communit* OR village* OR participant* OR volunteer* OR subject* OR incidence OR hospital OR man ) . We used the United Nations list of 54 African Sovereign states as the basis for searching , modified by including Rhodesia ( name had changed during the search period ) and both the English and French names for Ivory Coast . Initial tests showed that “Congo” found all references to the Democratic Republic of the Congo as well as Republic of the Congo . Similarly , “Sudan” found both Sudan and South Sudan . Although it is not recognized by the United Nations as a sovereign state , we also included “Western Sahara” in case there were reports from this area . PubMed and Embase have , an English translation for all paper titles regardless of language , English key words ( i . e . MESH terms in PubMed ) and usually an English Abstract . These were the only fields searched in the initial searching for papers of any language and assumes that the English terms used in the search would be sufficient to identify papers in any language for the first pass . Initial searches failed to find any papers prior to 1966 , search results were limited to publications from 1st Jan 1966 up to 31st December , 2014 . We imported the full texts/abstracts of the search result into Quosa Information Manager software ( Quosa ) [18] and a full text search was done through Quosa for articles containing the term “Salmonell*” . To ensure a comprehensive search of the literature , especially for reports within the last five years , we did an independent search on Embase database using a similar search string and strategy as above ( with limit publication dates from 2009 to 2014 ) . Additional reports that were not retrieved from the PubMed search were obtained and reviewed for inclusion . The full text of the search results of online articles/abstracts were reviewed independently by the study authors ( IVU , CAM and AS ) with the aim of including articles that used blood culture to isolate NTS from humans in Africa Inconsistences between the papers included by the different authors were resolved by consensus . The full text version either obtained on line or ordered , when only the abstract was available on line , of potentially relevant articles was retrieved and reviewed critically using predetermined inclusion and exclusion criteria for the study . This was done for articles , regardless of the published language . The reference sections of retrieved full text articles were reviewed critically in search of further potential articles for inclusion . The online database search performed on PubMed ( completed in February , 2015 ) , using the search string and limiting the results from 1st January 1966 to 31st December 2014 , yielded 16 , 638 articles . These were retrieved in Quosa and a full text search for the term ‘Salmonella’ using Quosa yielded a subset of 1 , 979 articles . The abstract and full text ( where available ) of the 1 , 979 articles obtained were reviewed manually for relevance based on our criteria . 177 articles were finally selected from PubMed and entered into the database ( Fig 1 ) . We found more publications using our search string with all the countries in Africa included by name , than with “Africa” alone , and by not restricting publications by using ‘human’ as a search term . The articles retrieved in Quosa were published in 16 languages of which French ( 216 articles ) was the most common following English . A similar search strategy was used on the Embase database with the initial search results limited to articles in Embase only , humans and the period 2009 to 2014 , yielding 8034 articles ( in February , 2015 ) . 736 articles were obtained following a full text search for the term ‘Salmonella’ in Quosa . Thereafter , the full texts of these articles were reviewed manually and 13 additional unique articles not obtained from the initial PubMed search were added to the database ( including 9 abstract-only articles ) . One hundred and ninety articles [2 , 3 , 10 , 11 , 13 , 19–202] were obtained following the literature search and used for the descriptive analysis in our review ( Fig 1 ) . Following a further review of these articles , 14 ( S2 Table ) reporting cases or cohorts of iNTS disease already reported in another published article were excluded from the quantitative analysis , leaving 176 articles including 12 studies with possibly overlapping data ( S2 Table ) . These 176 studies included 223 distinct subject cohorts . The reports varied in their methodology . Three of the 176 articles were reports from ill subjects in a longitudinal , community-based surveillance . 159/176 ( 90% ) of all articles retrieved are reports of isolates obtained from studies conducted on ill subjects presenting to a health facility setting ( hospital or clinic ) . In fourteen of the reports , we could not determine the original basis for selecting the subjects ( E . g . , retrospective analysis of microbiology laboratory samples , follow up analyses of bacteraemia in patients previously selected for another study ) . However , most of these 14 were probably from ill subjects presenting to a health facility . Only 9 ( 5% ) were designed to derive an estimate of the population-based incidence of iNTS disease in Africa; 76 ( 43 . 2% ) were prospective hospital-based studies with patients recruited during the course of the study; 22 ( 12 . 5% ) were retrospective studies with analysis of existing hospital or laboratory records; and the remaining 69 ( 39 . 2% ) are case reports , series , conference abstracts or outbreak reports . A total of 18 , 931 isolates of NTS were reported in this review . These included 9 , 084 isolates of Salmonella enterica serovar Typhimurium ( S . Typhimurium ) ( 48% ) , 2 , 801 isolates of S . Enteritidis ( 15% ) , 1 , 215 other serovars ( 6% ) and 5 , 831 ‘not further typed’ ( 31% ) . The category ‘other serovars’ includes , but is not limited to , less common serovars: 197 Infantis , 155 Wien , 88 Dublin , 71 Newport , 71 Bovis-morbificans , 92 Isangi , 24 Heidelberg , 8 Havana and 9 Ordonez [11 , 40 , 49 , 65 , 68 , 71 , 91 , 104 , 108] . Salmonella Typhimurium is therefore the most common serovar reported for iNTS disease in Africa and approximately three times ( 48% vs 15% ) more common than S . Enteritidis . iNTS disease was reported in 33 countries across the five geographical regions of Africa based on the United Nations classification ( Fig 2 ) . The majority , 53% ( 94/176 ) of the reports were from the Eastern Africa region ( including 28 reports from Kenya , 24 from Malawi , 16 from Tanzania and 10 from Uganda ) ; 26% ( 46/176 ) from the Western Africa region ( including 9 reports from Cote d’Ivoire , 9 from Nigeria and 8 from Ghana ) ; 10% ( 17/176 ) from Central Africa ( including 11 reports from Democratic Republic of Congo , 3 from Gabon and 2 from Central African Republic ) ; 7% ( 12/176 ) from Northern Africa ( including 5 reports from Tunisia and 2 each from Morocco and Algeria ) ; and 4% ( 7/176 ) from Southern Africa ( South Africa ) ( S3 Table ) . Details of the study population or study site setting could be extracted from only 113 of the 176 reports obtained . 50 . 4% ( 57/113 ) of the reports were from urban sites , 38 . 9% ( 44/113 ) from rural sites and 10 . 6% ( 12/113 ) from both urban and rural sites ( S4 Table ) . 59 . 2% ( 11 , 211/18 , 931 ) of the total NTS isolates were reported from the Eastern Africa region ( 6 , 057 isolates from Malawi , 2 , 976 isolates from Kenya and 767 isolates from Uganda ) . 14 . 9% ( 2 , 819/18 , 931 ) and 13 . 5% ( 2 , 558/18 , 931 ) of isolates were from Southern ( 2 , 819 isolates from South Africa ) and Western Africa ( 788 isolates from Mali , 371 from Cote d’Ivoire and 345 from Ghana ) respectively . 6 . 2% ( 1 , 180/18 , 931 ) and 6 . 1% ( 1 , 163/18 , 931 ) of isolates were from Central ( 1 , 003 isolates from Democratic Republic of Congo , 138 from Gabon and 34 from Central African Republic ) and Northern Africa ( 1 , 023 isolates from Algeria , 100 from Tunisia and 36 from Morocco ) respectively ( Fig 3 ) . The earliest report of iNTS disease was published in 1966 [69] and the numbers of reports of iNTS disease increased with time to a peak of 18 reports in 2011 ( Fig 4 ) . 55 . 7% ( 98/176 ) of the reports obtained were published within the last decade ( 2005 to 2014 ) , with 35 . 8% ( 63/176 ) were published in the last 5 years ( 2010 to 2014 ) . A similar trend can also be observed in the total number of NTS blood culture isolates with time . 67 . 5% ( 12770/18931 ) of the total isolates were reported within the last ten years ( 2005 to 2014 ) , while 22 . 3% ( 4218/18931 ) were reported in the last 5 years ( 2010 to 2014 ) . Fourteen reports on the incidence of iNTS disease in Africa were obtained from eight countries–the Gambia , Ghana , Kenya , Malawi , Mozambique , South Africa , Tanzania & Uganda , spanning three regions , Eastern ( ten studies ) , Southern ( one ) and Western ( three ) ( Table 1 ) . The incidence data were obtained using different methods of estimation and were from different population groups including specific age groups , HIV-infected patients and subjects with sickle cell disease . Overall , the estimated incidence of iNTS disease ranged from 1 . 4 per 100 , 000 population/year ( in South African individuals of all ages in 2003 to 2004 ) to 2 , 520 per 100 , 000 population per year in children < 5 years of age from Ashanti , rural Ghana , in 2007 to 2009 . The incidence rates were much higher in Eastern and Western Africa compared to Southern Africa . Across the studies , the estimated incidence rates were higher amongst HIV-infected subjects [123 , 131] , subjects with sickle cell disease[174] , young children[122 , 160 , 164] and in a rural setting compared to an urban setting [164] . Three studies estimated iNTS disease in the community . These studies were conducted on subjects from all age groups , and without bias for risk factors such as HIV infection , malaria and anaemia . Estimates obtained were 1 . 4 , 164 and >600 per 100 , 000 population per year from South Africa , Malawi and Ghana respectively [11 , 83 , 184] . Fifty six studies , describing a total of 114 , 634 blood cultures ( S5 Table ) from four regions were eligible for analysis: hospital-based studies from Africa investigating the organisms causing bacteraemia in the community and not in any specific risk group . Eligible studies were from four of the five regions ( all except the Northern region ) and sixteen countries in Africa . NTS proportions varied between and within regions . Regional averages ranged from 8% in Southern Africa to 38% in Central Africa with an overall average of 25% . In Eastern Africa , with an average of 27% , the range was 9% in Ethiopia to 39% in Malawi while in Western Africa , with an average of 18% , the range was 8% in Nigeria to 34% in Burkina Faso . Across countries , NTS was a cause of about 8% ( lowest value ) of community acquired bacteraemia in both Nigeria and South Africa to a peak value of 45% in one study from Central African Republic involving 131 blood cultures ( Fig 5 ) . Variations in NTS proportion within country were found in Malawi , Kenya , Tanzania and Uganda . This suggests that it might be difficult to directly extrapolate the burden of iNTS disease within a country as well as from one country to another . An analysis of the relative frequencies of different organisms causing community acquired bacteraemia in Africa was performed and it showed that NTS , Staphylococcus aureus and Streptococcus pneumoniae are most prevalent organisms isolated over time in Africa ( Fig 6 ) . Since the selection of articles included in this review excluded blood culture series without NTS , there is an element of bias favouring NTS compared to other organisms . However , the review of Reddy et al [16] also indicated the importance of iNTS as a cause of community acquired blood stream infection amongst adults and children in Africa [16] . The risk factors known to be associated with iNTS disease in Africa include HIV infection , malnutrition , malaria , young age , anaemia and rural setting ( Table 2 ) . Seven studies from Africa showed a positive association between HIV and iNTS disease in Africa . The estimated odds ratio ( OR ) from the studies for HIV-infected individuals developing iNTS disease compared with HIV-uninfected individuals ranged from 3 . 2 to 48 . 2 . The strong association between HIV infection and iNTS disease in Africa was described in 1990 when Gilks et al reported an OR of 48 . 2 ( confidence interval 13–176 ) in Nairobi , Kenya [81] . Malnutrition had a positive association with iNTS disease in five studies ( Table 2 ) , with the earliest study published in 2005 . The OR for iNTS disease occurring in children with malnutrition compared with children without malnutrition ranged from 1 . 44 to 2 . 42 . Plasmodium falciparum malaria had a positive association with iNTS disease in three African studies ( Table 2 ) since the year 2000 . The estimated OR for iNTS disease in individuals with Plasmodium falciparum malaria compared with those without Plasmodium falciparum malaria ranged from 1 . 5 to 4 . 1 . In one study , recent history of malaria was positively associated with iNTS disease[192] . Young age was associated with iNTS disease in three studies ( Table 2 ) . The OR for iNTS disease in young individual compared to older individuals ranged from 2 . 07 to 4 . 30 . In one study of children 1–4 years of age , iNTS disease was reported to be more associated with age when compared to other organisms causing bacteraemia[122] . Anaemia ( especially moderate and severe anaemia ) , was shown by five studies to be associated with iNTS disease in Africa with reported OR ranging from 1 . 86 to 35 . 6 . Rural settlement compared to urban was found by one study to be associated with iNTS disease [192] . The OR of iNTS disease occurring in a rural settlement compared to an urban settlement was 2 . 23 We were able to extract CFR data for iNTS disease in Africa from twenty four studies ( Table 3 ) , describing a total of 548 deaths among 2656 cases . The overall CFR was 20 . 6% . and ranged from 0% in individuals greater than or equal to 5 years old in Kenya [75] to 72 . 7% in another Kenyan study involving only HIV-infected patients [81] . The average CFR , derived from 8 studies [3 , 67 , 89 , 93 , 152 , 160 , 172 , 188] conducted among low risk populations ( not HIV-infected , anaemic , malnourished , or having malaria ) , and with >90 iNTS cases isolated , is 19% ( 276 fatalities from 1427 cases ) . Our review seeks to comprehensively document reports of iNTS disease in Africa: the number and location of the reported cases , and where available , the age of the subjects , risk factors , case fatality rates and incidence . As a result , we have included many more studies ( 199 ) than a recent review by Ao et al [204] that specifically only included reports of incidence . In their database , in addition to reports from other countries they list 10 reports from 6 African countries . All 10 reports from the Ao et al study are included in the larger set of 14 reports describing incidence from 8 African countries we include in Table 1 . Not surprisingly we find a similarly high incidence of iNTS disease in sub-Saharan Africa . Ao et al estimated the overall incidence of iNTS disease in Africa at 227 cases [range 152–341] per 100 , 000 population with 1 . 9 [range 1 . 3–2 . 9] million cases annually [204] . We sought to provide a comprehensive review of the published reports of NTS bacteraemia in the peer reviewed literature and found a substantial number of reports with the earliest in 1966 , interestingly about the same time as genetic studies suggest the evolution of the highly African-specific highly invasive NTS genotypes [205 , 206] . There are limits on the comprehensiveness of our study . In particular , many of the early and non-English reports in the initial screening stage were only available to us on line as abstracts , and in some cases , just the title with MESH headings and some of these may have contained details e . g . the use of blood cultures to determine Salmonella bacteraemia in the text but not in the abstract that resulted in these articles not being ordered and reviewed in full . Despite these limitations , this survey highlights important findings . As shown in Fig 2 , no reports were found in some countries in West Africa and in South West Africa . In view of the reported cases from neighbouring countries it seems unlikely that iNTS disease does not occur in these counties , but this highlights that gaps in the published literature almost certainly exist , e . g . due to lack of infrastructure in these countries to undertake the blood culturing required to obtain a definitive diagnosis , or the lack of researchers interested in publishing reports . Some of these gaps in the database may be addressed by examining health service records , but this is beyond the scope of the article . As shown in Table 1 in the papers we surveyed , the estimated incidence of iNTS disease ranged from 1 . 4 cases per 100 , 000 population per year over all age groups in South Africa to a yearly cumulative incidence of 2 , 520 per 100 , 000 among <5 years old children in Ghana . However , we caution against trying to over-interpret these incidence data or the range observed–a striking observation is the lack of consistency in the age groups reported , the inclusion criteria that may or may not select for populations with specific risk factors ( e . g . HIV , sickle cell anaemia ) and in the way in which the population denominator is determined for these largely facility-based studies . On the basis of the whole database of reported bacteraemia , we confirm the earlier findings that NTS accounts for a large proportion of the bacteria responsible for community acquired blood stream infections and is a substantial cause of morbidity and mortality with the serovars Typhimurium and Enteritidis responsible for 91% of NTS bacteraemia in Africa in the 69% of cases where the serotype was determined . A major knowledge gap exists in the incidence of iNTS disease in Africa owing to the paucity of reports from population-based surveillance of Salmonella in Africa . This is due to poor surveillance systems for infectious diseases in most parts of Africa . The majority of the studies on iNTS disease in Africa are hospital-based which can only provide an estimate of the minimal incidence of disease in the community , because not all febrile cases report to hospital . Denominators necessary for the detailed estimation of iNTS disease incidence in the reported publications are generally not available . This also makes it difficult to estimate the true population at risk of disease . Population-based incidence studies are best suited to estimate the incidence of an infectious disease in Africa . This can be done by either adjusting the data obtained from hospital-based study with data obtained from simultaneous health-seeking surveys carried out in the same community ( for example , as has been carried out by the Typhoid Surveillance in Africa Program ( TSAP ) and Severe Typhoid in Africa programme [207] or designing a study with a method that will allow for identification of all the cases of the disease within the target community . Epidemiological data on iNTS disease were available from publications in peer-reviewed journals from 33 of the 54 countries in Africa ( Fig 2 ) . Unavailable data from the remaining 21 African countries might be due to a lack of studies in these countries , or a true low burden of iNTS disease . These 21 countries are not involved in TSAP . It would therefore be valuable to conduct a thorough search of grey literature in the affected countries including reviewing existing databases of blood cultures carried out in major hospitals in various locations across the affected countries in the last five to ten years . Such an analysis would facilitate a better understanding of the relative importance of NTS as a cause of community acquired blood stream infection in Africa . New prospective hospital-based blood culture studies in these countries would be even more useful . The average case fatality rate of community acquired severe infections is 20 . 6% but with a wide range from 0 to 72% . Therefore , the development of an effective vaccine against NTS for Africa would be an important intervention to help reduce the burden of disease and deaths due to NTS . The exact mechanisms of transmission of iNTS disease are currently unclear and there are no rapid diagnostic tests available for its detection . The development of such diagnostics would greatly facilitate the study and management of iNTS disease in Africa .
Although Salmonella are a common global cause of mild gastro-intestinal illness that usually presents with self-limiting diarrhoea , the invasive nontyphoidal form of Salmonella disease manifests as bacteraemia which often presents only with fever . If left untreated , iNTS disease often results in death . iNTS disease is more common amongst people with impaired immunity , and in developing countries particularly in Africa . Data on the epidemiology of iNTS disease in Africa are sparse , making it difficult to estimate the true disease burden . This study aggregates all published cases of iNTS disease in Africa from 1 Jan 1966 up to 31 Dec 2014 and shows the relative importance of NTS as a cause of community acquired bacteraemia and deaths in Africa . The findings are important in raising awareness of iNTS disease and driving research to help develop effective control and preventive strategies to limit the disease . Important interventions needed for iNTS disease include rapid diagnostic tests and vaccines , as well as measures to improve hygiene .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[ "medicine", "and", "health", "sciences", "body", "fluids", "pathology", "and", "laboratory", "medicine", "pathogens", "geographical", "locations", "tropical", "diseases", "database", "searching", "microbiology", "parasitic", "diseases", "health", "care", "retroviruses", "viruses", "immunodeficiency", "viruses", "bacterial", "diseases", "inpatients", "rna", "viruses", "enterobacteriaceae", "patients", "bacteria", "africa", "bacterial", "pathogens", "research", "and", "analysis", "methods", "infectious", "diseases", "medical", "microbiology", "hiv", "microbial", "pathogens", "hospitals", "hematology", "salmonella", "people", "and", "places", "blood", "anatomy", "health", "care", "facilities", "physiology", "database", "and", "informatics", "methods", "viral", "pathogens", "biology", "and", "life", "sciences", "malaria", "lentivirus", "organisms" ]
2017
A Systematic Review of the Incidence, Risk Factors and Case Fatality Rates of Invasive Nontyphoidal Salmonella (iNTS) Disease in Africa (1966 to 2014)
Seeing the direction of motion is essential for survival of all sighted animals . Consequently , nerve cells that respond to visual stimuli moving in one but not in the opposite direction , so-called ‘direction-selective’ neurons , are found abundantly . In general , direction selectivity can arise by either signal amplification for stimuli moving in the cell’s preferred direction ( ‘preferred direction enhancement’ ) , signal suppression for stimuli moving along the opposite direction ( ‘null direction suppression’ ) , or a combination of both . While signal suppression can be readily implemented in biophysical terms by a hyperpolarization followed by a rectification corresponding to the nonlinear voltage-dependence of the Calcium channel , the biophysical mechanism for signal amplification has remained unclear so far . Taking inspiration from the fly , I analyze a neural circuit where a direction-selective ON-cell receives inhibitory input from an OFF cell on the preferred side of the dendrite , while excitatory ON-cells contact the dendrite centrally . This way , an ON edge moving along the cell’s preferred direction suppresses the inhibitory input , leading to a release from inhibition in the postsynaptic cell . The benefit of such a two-fold signal inversion lies in the resulting increase of the postsynaptic cell’s input resistance , amplifying its response to a subsequent excitatory input signal even with a passive dendrite , i . e . without voltage-gated ion channels . A motion detector implementing this mechanism together with null direction suppression shows a high degree of direction selectivity over a large range of temporal frequency , narrow directional tuning , and a large signal-to-noise ratio . Motion represents an essential visual cue , used for predator avoidance , prey capture and visual navigation throughout the animal kingdom . Accordingly , motion-sensitive neurons are found in various brain areas of vertebrates and invertebrates alike . Prominent and well-studied examples are retinal ganglion cells of the rabbit [1 , 2] , retinal starburst cells [3] and ganglion cells of the mouse [4 , 5] , cortical neurons of the mouse [6] , cat [7] , ferret [8] and monkey [9 , 10] , neurons of the accessory optic system of birds [11] and the lobula plate tangential cells of flies [12–15] ( for review , see [16] ) . To describe the response properties of these neurons , a number of partially equivalent models [17] have been developed , e . g . the Hassenstein-Reichardt detector [18] , the Barlow-Levick detector [2] , the F-model [19] , the elaborated Reichardt model [20] and the energy model [21] . As a common feature , these models compute the local direction of motion by correlating the luminance values of adjacent image pixels after asymmetric temporal filtering . With one exception [22] , however , the biophysical implementation of such a correlative , multiplicative-like interaction has so far not been elucidated . In the fruit fly Drosophila , visual signals are processed in the optic lobe , a brain area comprised of the lamina , medulla , lobula , and lobula plate , each arranged in a columnar , retinotopic fashion [23–25] ( for review , see [26 , 27] ) . In striking parallel to the vertebrate retina [28] , the direction of visual motion is computed within the optic lobe separately in parallel ON and OFF motion pathways [29–33] . Within each column , four T4 and four T5 cells represent the local output signals of the ON ( T4 ) and the OFF ( T5 ) channel , each one of them tuned to one of the four cardinal directions projecting accordingly to one of the four layers of the lobula plate [34] . There , T4 and T5 cells provide direct excitatory cholinergic input onto the dendrites of wide-field , motion-sensitive tangential cells [35 , 36] as well as onto glutamatergic lobula plate interneurons that inhibit wide-field tangential cells in the adjacent layer [37 , 38] . Electrophysiological [32] , optical voltage [39] and Calcium recordings [33 , 40–43] from presynaptic medulla neurons revealed that none of them is directionally selective . Therefore , T4 and T5 cells are the first neurons in the visual processing chain that respond to visual motion in a direction selective manner [34 , 44] . Different studies provided evidence that T4 and T5 cells become selective for the direction of motion mainly by preferred direction enhancement [44 , 45] , by null direction suppression only [46] , and by a combination of both mechanisms [47–49] . As shown by apparent motion stimuli placed precisely on the hexagonal lattice of the columns via a telescope , individual stimuli interact in a supralinear way on the preferred side of the dendrite , while they suppress each other when delivered on the null side of the dendrite [47 , 48] . An electron microscopy study revealed that T4 cells receive input on their dendrites from columnar neurons in a topographic order that follows their directional preference [50]: Mi9 cells contact T4 cells’ dendrites on their preferred side , Mi1 and Tm3 cells provide input in the central part , while Mi4 , C3 and TmY15 are presynaptic on the null side of the dendrite . Intriguingly , 2-Photon Calcium imaging showed an OFF center receptive field for Mi9 , while Mi1 , Tm3 and Mi4 all exhibit an ON center [42] . With respect to their temporal filter properties , the same study found Mi9 and Mi4 to be slow and sustained , well described by a temporal low-pass filter . In agreement with previous electrophysiological studies [32] , Mi1 and Tm3 turned out to be fast and transient , with temporal band-pass properties [42] . Taken together , the above results suggest that a preferred direction enhancement is realized by a supralinear interaction between Mi9 and the central inputs Mi1 and Tm3 on the preferred side and a null direction suppression by Mi4 on the null side of the T4 cells’ dendrite ( Fig 1A ) . Indeed , multiplying the positive , i . e . sign-inverted Mi9 signal with the one from Mi1 and dividing the result by the one from Mi4 results in a tuning characteristic of the postsynaptic T4 cell that matches the experimental data in detail [42] . In response to gratings drifting along the preferred and null direction at temporal frequencies over two orders of magnitude , the preferred direction response of T4 cells peaks at around 1 Hz while the null direction response is close to zero over the whole range ( Fig 1B , [34 , 47] ) . In response to gratings drifting along various directions , T4 cells exhibit a rather narrow directional tuning ( Fig 1C , [34 , 47] ) . A similar model has been proposed in [43] . However , in contrast to the one explained above , this one derives its direction selectivity mainly from a multiplicative interaction between Mi1 and Tm3 impinging on different parts of the dendrite with the result being attenuated by flanking inputs from Mi9 and Mi4 [43] . How should such operations be realized in terms of biophysics of nerve cells ? How can neurons multiply or divide , and what is the role of the OFF center receptive field of the enhancing neuron ? Here , the following evidence exists with respect to the transmitter phenotype of the various input neurons: ( 1 ) Mi9 is immuno-positive for the vesicular glutamate transporter VGluT [50] . Interestingly , in insects , glutamate can also exert hyperpolarizing , inhibitory action on the postsynaptic neuron via the glutamate-gated chloride channel GluClα [51 , 37 , 38] . Indeed , the transcript for GluClα has been found in mRNA pooled from T4 and T5 cells [52] . ( 2 ) Mi1 and Tm3 are immuno-positive for the acetylcholine-synthesizing enzyme choline-acetyl-transferase ( ChAT , [50] ) and also express the vesicular acetylcholine-transporter ( VAChT , [53] ) . ( 3 ) Mi4 are immuno-positive for the GABA-synthesizing enzyme glutamic-acid-decarboxylase GAD1 [50] . These observations suggest that the two flanking neurons , Mi9 and Mi4 , are inhibitory , while the central inputs are excitatory on the postsynaptic T4 cell , and will provide the substrate of a biophysical implementation of the fly motion detector in the ON pathway proposed in the following . Following a previous suggestion [54] , I will first describe how even a passive membrane model can reveal positive , multiplicative-like signal amplification . In order to extract the nonlinearity from the circuit , the responses to two inputs delivered simultaneously are compared with the sum of the responses to each individual stimulus presentation ( ‘linear expectation’ ) . Let us first consider a simple electrical equivalent circuit of a passive isopotential neuron that receives two excitatory input signals x and y acting on the excitatory conductances gexc1 and gexc2 ( Fig 2A ) . The steady-state postsynaptic membrane potential Vm is given by: Vm=Eexc ( gexc1+gexc2 ) +Eleakgleakgexc1+gexc2+gleak ( 1 ) When we express Vm as the difference between Vm and Eleak and all conductances relative to gleak , this becomes: Vm=Eexcgexc1+gexc2gexc1+gexc2+1 ( 2 ) For x = gexc1 and y = gexc2 , the responses to each individual input are ( Fig 2B and 2C ) : R1=Eexcxx+1;R2=Eexcyy+1 ( 3 ) For x = y , the linear expectation becomes ( Fig 2B and 2C ) R1+R2=Eexc ( 2xx+1 ) ( 4 ) The membrane response to both inputs given simultaneously equals ( Fig 2B and 2C ) : R1 , 2=Eexc ( 2x2x+1 ) ( 5 ) From this , we calculate the nonlinear response component as the difference between the response to both inputs and the linear expectation ( Fig 2D ) : RNonlin=Eexc−2x22x2+3x+1 ( 6 ) We see that the response of an electrically passive neuron is always sublinear , i . e . the response to the simultaneous activation of two excitatory inputs is smaller than the sum of the responses to each excitatory input individually . Hence , the nonlinear response component is negative and decreases with increasing input signals ( Fig 2D ) . We next investigate the situation for a combination of an excitatory and an inhibitory input . The membrane potential becomes: Vm=Eexcgexc+Einhginhgexc+ginh+1 ( 7 ) Let us now assume that the excitatory conductance follows the input signal x , while the inhibitory conductance follows 1-y ( with 0 ≤ y ≤ 1 ) , i . e . the inhibitory conductance becomes the smaller the larger the input signal y ( Fig 2E ) . As we will see , this has interesting consequences for the nonlinearity of the postsynaptic membrane voltage . The resting membrane potential , i . e . when x = y = 0 , now is Vrest = Einh/2 , and , as before , all membrane responses will be expressed relative to this resting potential . The individual responses become ( Fig 2F and 2G ) R1= ( 2Eexc−Einh ) x2 ( 2+x ) ;R2= ( −Einh ) y2 ( 2−y ) ( 9 ) For x = y , the linear expectation becomes ( Fig 2F and 2G ) R1+R2= ( Eexc ( 2−x ) −2Einh ) x4−x2 ( 10 ) The membrane response to both inputs given simultaneously equals ( Fig 2F and 2G ) R1 , 2= ( Eexc−Einh ) x2 ( 11 ) From this , the nonlinear response component is calculated as ( Fig 2H ) RNonlin= ( Eexc ( 2−x ) +Einhx ) x22 ( 4−x2 ) ( 12 ) For 0 ≤ x ≤ 1 , i . e . positive conductances smaller or equal to the leak conductance , and abs ( Eexc ) > abs ( Einh ) , this expression is always positive . Therefore , a passive membrane reveals a signal amplification if one of the two inputs decreases an inhibitory input conductance . Intuitively , this is because the input reduces the input resistance of the postsynaptic neuron and therefore leads to an increased response to an excitatory input as compared to when the latter is given in isolation ( Fig 2H ) . In order to explore whether such a biophysical mechanism is indeed useful to extract the direction of motion , I simulated a two-dimensional array of 40 x 40 motion detectors covering a visual space of 180 by 180 deg at a temporal resolution of 10 msec . Each detector received input from 3 neighboring locations in visual space ( Fig 3A ) . Input from the left and right location became processed by a 1st order low-pass filter with a time-constant of 50 msec , while the central input was 1st order high-pass filtered with a 250 msec time-constant , plus a DC component of 10% [30] . These signals acted on an electrical equivalent circuit of a passive piece of membrane . Since the dynamics of the input signals were assumed to be large against the membrane time-constant , the capacitive current could be neglected . Importantly , the left input was simulated as an OFF channel controlling an inhibitory conductance , i . e . this signal became the smaller the larger the local luminance . The central input was treated as an ON channel controlling an excitatory conductance and the right input as an ON channel controlling an inhibitory conductance . This way , preferred direction enhancement , as proposed in the Hassenstein-Reichardt detector [18] , was implemented by the interaction between the left and the central input , and null direction suppression , as proposed in the Barlow-Levick detector [2] , by the action of the inhibitory right input . Fig 3B–3D show the input signals ( Fig 3B ) , their filtered versions ( Fig 3C ) as well as the resulting conductances ( Fig 3D ) in response to a sine-grating moving at 1 Hz along the preferred and the null direction of the motion detector . The peak of the excitatory conductance coincides with the trough of the inhibitory conductance during preferred direction motion and with the peak during null direction motion ( Fig 3D ) . Therefore , the resulting membrane voltage of a single motion detector depolarizes periodically up to 10 mV during preferred direction motion , while it remains hyperpolarized during null direction motion ( Fig 3E ) . In analogy to a lobula plate tangential cell receiving excitatory input from T4 cells [55 , 37 , 38] , the output voltages of all local motion detectors were rectified at a membrane potential of 0 mV and averaged across the population . This signal reveals maximum direction selectivity with sustained depolarization during preferred direction and zero response to null direction motion ( Fig 3F ) . This motion detector ( Fig 4A , top ) was tested under various conditions . In a first series of simulations , a sine-grating was used drifting from 0 . 1 Hz up to 10 Hz along either the preferred ( Fig 4B ) or the null direction ( Fig 4C ) of the detector . To assess the directional tuning of the motion detector , the models were next stimulated by a sine-grating , again with 100% contrast , drifting at 1 Hz in various directions in steps of 30 deg ( Fig 4D ) . Finally , I tested the noise susceptibility of the motion detector in two different situations . In one case , photon noise was simulated by presenting a sine-grating first drifting along the preferred and then along the null direction of the motion detector . At each time point , the actual pixel value was drawn from a Poisson distribution with a mean value λ proportional to the value of each image pixel independently ( Fig 4E; for details , see Methods ) . Different noise levels were achieved by changing the overall mean luminance in logarithmic steps of 2 . In the other case , motion noise was simulated by presenting moving dots , a certain percentage of which moved coherently first along the preferred and then along the null direction of the motion detector , whereas the remaining dots moved randomly into any other direction ( [9]; Fig 4F ) . The noise susceptibility was quantified as the signal-to-noise ratio of the detector response , i . e . the difference between the average output signals during preferred and null direction simulation , divided by the square root of the response variance . Such a motion detector revealed a strong directional selectivity over a large range of temporal frequencies ( Fig 4B and 4C , green traces ) : With a peak of the preferred direction response at 2 Hz , the model responded virtually not at all during null direction motion . It also showed a narrow directional tuning ( Fig 4D , green trace ) with a response amplitude falling to less than 50% for gratings drifting 60 deg away from its preferred direction . For photon noise , the signal-to-noise ratio of the motion detector declines gradually over two orders of magnitude , from about 100 to 1 ( Fig 4E , green trace ) . For motion noise , the signal-to-noise ratio is already at values above 1 at 20% coherence reaching a maximum value of almost 1000 at 100% coherence ( Fig 4F , green trace ) . In order to assess the contribution of each mechanism individually , model simulations were repeated with either the left or the right arm blocked ( Fig 4A , middle: Null direction suppression = NDS only , brown traces . Fig 4A , bottom: Preferred direction enhancement = PDE only , orange traces ) . For gratings drifting along the preferred direction , the model with only NDS reached maximum responses at higher temporal frequencies , whereas the model with only PDE performed in an almost identical way to the full model ( Fig 4B ) . Both partial models showed significant responses to gratings drifting along the null direction ( Fig 4C ) . With respect to their directional tuning , the model with only NDS revealed some broadening compared to the full model , while the model with only PDE performed significantly worse ( Fig 4D ) . When confronted with photon noise , the signal-to-noise ratio of both partial models dropped more steeply with increasing input noise than the full model , with the model relying only on null direction suppression performing better than the one relying only on preferred direction enhancement ( Fig 4E ) . In a similar way , both partial models turned out to be less sensitive to the number of dots moving coherently ( Fig 4F ) . What is the reason for the full model to exhibit a higher signal-to-noise at its output , compared to both partial models , when confronted with noisy input ? In order to investigate this question , the responses of all three models ( Fig 5A ) are shown in the presence of photon noise ( Fig 5B , mean luminance level = 4 ) and in response to moving dots ( Fig 5C , coherence level = 80% ) . As can be seen in the histograms , shown next to the time-dependent response traces , the response variances , i . e . the widths of the response distributions , are somewhat smaller for the full model than for each of the partial models , both during preferred and during null direction motion . This is due to the fact that both partial models share the central input which , due to its high-pass properties , is more noise-sensitive than the left and the right input signals which are low-pass filtered . This way , the interaction between both mechanisms can reduce the noise more effectively than each partial model alone . In addition , the difference between the mean preferred direction and null direction response is larger for the full model , as compared to both partial model . Both effects lead to an increase of the signal-to-noise ratio . In summary , a motion detector that implements both preferred direction enhancement and null direction suppression in a biophysically plausible way by passive ionic conductances reveals a strong direction selectivity , narrow directional tuning and turns out to be rather noise insensitive . This behavior is not attributable to any one of the two mechanisms alone but rather rests on the combination of both of them . How neurons in the visual system compute the direction of motion from non-directional input signals , i . e . the emergence of direction selectivity , has been a prime example for neural computation in general for long and a field of intense studies in both vertebrates and invertebrates ( for review , see [16] ) . In the fruit fly Drosophila , our current understanding has reached a level where not only the exact location of this computation is known , i . e . the dendrite of T4 cells in the ON pathway , but also the identity and visual response properties of the input neurons , their transmitter phenotype as well as their precise placement on the dendrite . This opens the door to ask for the biophysical implementation of the underlying computations . At an algorithmic level , these computations comprise a signal amplification for motion along the preferred direction and a signal suppression for motion along the null direction of T4 cells , with both mechanisms at work at different locations within the receptive field [47–49] . Null direction suppression can be implemented by an inhibition opening a Chloride conductance several times larger than the leak conductance ( ‘shunting inhibition’ , [56] ) or , alternatively , by a modest inhibitory conductance change followed by rectification . The latter would reflect the voltage threshold of a Calcium channel . In contrast to null direction suppression , preferred direction enhancement seems more complex to understand in biophysical terms . Here , several proposals have been made in the past ( reviewed in [55] ) that include simple threshold nonlinearities [57] , log-exp transforms exploiting the relation of x ⋅ y = exp ( log x + log y ) [22] , NMDA receptors and chemical cooperativity . In contrast to most of the above , the mechanism advocated here relies purely on passive membrane properties of the postsynaptic neuron . Inspired by the fact that the input neuron in place for signal amplification has an OFF center receptive field and is potentially inhibitory , the proposed mechanism involves a decrease of inhibition for preferred direction signals leading to subsequent signal enhancement via an increase of the postsynaptic input resistance ( see also [55] ) . As is shown by a variety of tests ( Fig 4 ) , a motion detector employing such a mechanism leads to substantial degree of direction selectivity even in the absence of null direction suppression . It does so within a physiological range of conductance changes that are always smaller or at most equal to the resting conductance . The advantage of this mechanism over others relying purely on thresholds or other types of output nonlinearities implemented by the supra-linear behavior of voltage-gated ion channels lies in the fact that signal amplification is obtained over a large range of input signal amplitudes , as are expected to occur in natural environments ( see Fig 2H ) . Using membrane potential recordings from T4 cells together with apparent motion stimulation , a recent publication reports null direction suppression only , with no sign of preferred direction enhancement [46] . At first sight , this seems to be in conflict with several Calcium imaging studies that reported clear signs of preferred direction enhancement underlying direction selectivity in T4 cells [44 , 45 , 47–49] . This discrepancy could be explained by the action of a voltage-activated Calcium channel leading to a supralinear increase of Calcium concentration based on a current that is too small to be detectable in voltage recordings . A more likely explanation relies on the observation that null direction suppression is more sensitive at smaller stimulus sizes or intensities than preferred direction enhancement ( see Fig 2F in [48] ) . Based on these different sensitivities , one expects no preferred direction enhancement at 2 deg bar width as was used in Gruntman et al [46] . Obviously , more experiments are required to distinguish between these alternatives and to answer the fundamental question whether the membrane potential behaves linearly and supralinearity is expressed only at the level of the Calcium concentration . Apart from that , the specific role of the different input neurons to T4 cells can be tested by permanent blocking [58] , optogenetic hyper- [59] or depolarization [43] as well as removal of postsynaptic transmitter receptor via RNAi techniques [43] or genome editing [53 , 60 , 61] . Given that many models of motion detection imply a multiplicative-like interaction ( see Introduction ) , a detailed understanding of the biophysical mechanism underlying preferred direction enhancement in the fly T4 neuron will be of general interest beyond fly motion vision . Stimulus movies were generated as an array of 200 x 200 pixels and 1000 time points corresponding to 180 x 180 degree of visual space and 10 seconds of time . Sine-gratings had a spatial wavelength of 36 degree , an average luminance of 0 . 5 and a modulation from 0 to 1 , i . e . the contrast was 100% . Photon noise was simulated as shot noise by drawing random numbers from a Poisson distribution with the mean value λ proportional to the value of each image pixel I ( x , y , t ) of the sine grating . Different noise levels were achieved by different overall mean levels of luminance , i . e . by multiplying the image with a given factor ( 1 , 2 , 4 , 8 , 16 , 32 ) before drawing values from the Poisson distribution . Subsequently , all images were normalized to have the same mean luminance and contrast . Motion noise was generated by randomly placing 500 dots of 1 pixel size onto the visual scene . A certain percentage of these dots moved coherently into one direction while the rest moved into a random direction , which changed every 50 msec . Subsequently , the stimulus movie was spatially low-pass filtered by a Gaussian function of 4 . 5 degree half-width . First , the visual input was down-sampled to an array of 40 x 40 photoreceptors . This way , each photoreceptor received input from 4 . 5 x 4 . 5 degree of visual space . Next , the signal was high-pass filtered with 250 msec time-constant . 10% of the photoreceptor signal was added to the output of the filter . In the ON pathway , this signal was rectified at 0 . In parallel , the photoreceptor signal was also low-pass filtered with a time-constant of 50 msec . In the OFF pathway , the high-pass signal was sign-inverted and rectified at 0 , the low-pass signal was obtained by low-pass filtering 1 minus the photoreceptor signal . These filtered signals were then passed onto an array of 38 x 40 local motion detecting units simulating T4 cells tuned to rightward motion . Each such unit received as its left input , corresponding to the Mi9 cell , the low-pass signal from the OFF pathway and as its central and right input , corresponding to the Mi1 and Mi4 cell , the high-pass and the low-pass signal from the ON pathway , respectively . The excitatory conductance gexc was set to the Mi1 signal , the inhibitory conductance ginh was set to the sum of the Mi9 and Mi4 signals . The resulting membrane potential was then calculated as Vm=Eexcgexc+Einhginhgexc+ginh+gleak with Eexc = +50 mV , Einh = −20 mV , gleak = 1 . These signals were spatially averaged after rectification at 0 mV . All simulations were written in Python . The software is available as supplemental information .
Seeing the direction of motion is essential for survival of all sighted animals . Consequently , nerve cells that respond preferentially to visual stimuli moving in a certain direction are found abundantly . However , directional information is not represented at the level of single photoreceptors but rather has to be computed by subsequent neural circuits . Algorithmic models have been proposed in the past that calculate the direction of motion by multiplying and/or dividing the input signals from neighboring photoreceptors after asymmetric temporal filtering . But how can neurons multiply or divide ? Inspired by recent data from fly motion-sensitive neurons , I present a biophysical model of a nerve cell that is based on purely passive conductance changes . I show that such a model can reveal a high degree of direction selectivity over a large range of temporal frequency , narrow directional tuning , and a large signal-to-noise ratio .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[ "medicine", "and", "health", "sciences", "engineering", "and", "technology", "motion", "detectors", "signal", "processing", "membrane", "potential", "electrophysiology", "light", "neuroscience", "signal", "filtering", "electromagnetic", "radiation", "luminance", "visible", "light", "neuronal", "dendrites", "animal", "cells", "biophysics", "physics", "cellular", "neuroscience", "detectors", "cell", "biology", "equipment", "physiology", "neurons", "signal", "to", "noise", "ratio", "biology", "and", "life", "sciences", "cellular", "types", "physical", "sciences" ]
2018
A biophysical mechanism for preferred direction enhancement in fly motion vision
Many signaling proteins and transcription factors that induce and pattern organs have been identified , but relatively few of the downstream effectors that execute morphogenesis programs . Because such morphogenesis genes may function in many organs and developmental processes , mutations in them are expected to be pleiotropic and hence ignored or discarded in most standard genetic screens . Here we describe a systematic screen designed to identify all Drosophila third chromosome genes ( ∼40% of the genome ) that function in development of the tracheal system , a tubular respiratory organ that provides a paradigm for branching morphogenesis . To identify potentially pleiotropic morphogenesis genes , the screen included analysis of marked clones of homozygous mutant tracheal cells in heterozygous animals , plus a secondary screen to exclude mutations in general “house-keeping” genes . From a collection including more than 5 , 000 lethal mutations , we identified 133 mutations representing ∼70 or more genes that subdivide the tracheal terminal branching program into six genetically separable steps , a previously established cell specification step plus five major morphogenesis and maturation steps: branching , growth , tubulogenesis , gas-filling , and maintenance . Molecular identification of 14 of the 70 genes demonstrates that they include six previously known tracheal genes , each with a novel function revealed by clonal analysis , and two well-known growth suppressors that establish an integral role for cell growth control in branching morphogenesis . The rest are new tracheal genes that function in morphogenesis and maturation , many through cytoskeletal and secretory pathways . The results suggest systematic genetic screens that include clonal analysis can elucidate the full organogenesis program and that over 200 patterning and morphogenesis genes are required to build even a relatively simple organ such as the Drosophila tracheal system . Elucidating the genetic programs of organ formation and maintenance is a central goal of developmental biology and medicine . Many organogenesis genes have been isolated in systematic genetic screens in model organisms , and many others have been identified by their organ-selective expression patterns and by candidate gene analysis . These approaches have been very successful at discovering the signaling pathways and transcription factors that induce and pattern organs and specify cell fates , but they have been much less successful at identifying the downstream effectors that execute morphogenesis programs , what we call morphogenesis genes [1] , [2] , [3] , [4] . A similar abundance of signaling and transcription factor genes and dearth of morphogenesis genes has obtained from the pioneering genetic dissection of Drosophila body axis formation and other early developmental events [5] . We reasoned that many morphogenesis genes would function in multiple organs and developmental processes , so mutations in these genes would be pleiotropic and hence discarded in most genetic screens . We therefore designed systematic , saturation screens for genes required for Drosophila tracheal system organogenesis that included clonal analysis of gene function in the tracheal system , to identify all tracheal genes including those with pleiotropic phenotypes . The results of a screen of the third chromosome , representing ∼40% of the Drosophila genome [6] , are described here , and the results of a first ( X ) chromosome screen initiated earlier will be described elsewhere ( [7]; M . Metzstein and M . A . K . , unpublished data ) . The Drosophila tracheal ( respiratory ) system is a branched tubular network that transports oxygen throughout the body [8] . It is one of the most intensively studied and best understood organogenesis programs [9] , [10] , and it has emerged over the past decade as a paradigm of branching morphogenesis , the developmental process that gives rise to many organs including the lung , vascular system , kidney , and pancreas . Understanding how branching networks are patterned and how cellular tubes are made , shaped , and maintained is of fundamental importance in cell and developmental biology , and in medicine for understanding and treating tubular diseases such as aneurysms and polycystic kidney disease . The tracheal system develops from 10 pairs of tracheal sacs that arise by invagination of the embryonic ectoderm [8] , [11] . Each sac is an epithelial monolayer composed of ∼80 cells . Primary tracheal branches are formed by groups of 3–20 cells that bud from the sacs in different directions and successively sprout secondary and terminal branches . Some specialized primary and secondary branches grow towards and fuse with branches from neighboring sacs to interconnect the tracheal network [12] . The transformation of the simple epithelial sacs into an extensively branched tubular network occurs without cell proliferation , and is mediated by cell migration , rearrangement , and dramatic changes in cell shape [11] , [13] , [14] , [15] . During embryogenesis , the lumens of the developing tracheal branches are filled with a complex and changing matrix , which is cleared and replaced with gas just before the embryo hatches and the tubes become functional in respiration [8] , [16] . In the larva , terminal cells ramify extensively to form many new terminal branches ( tracheoles ) , long cytoplasmic extensions that grow toward oxygen-starved cells and then form a cytoplasmic , membrane-bound lumen , creating tiny tubes ( <1 um diameter ) that supply the targets with oxygen ( Figure 1B ) [14] , [17] . Unlike primary ( multicellular ) and secondary ( unicellular ) branches , tubes sealed by intercellular and autocellular junctions ( Figure 1B ) , terminal branches lack cell junctions and resemble the “seamless” endothelial tubes of the mammalian microvasculature [18] , [19] , [20] and C . elegans excretory system [21] . The first important tracheal gene identified was trachealess , isolated in the classical screens for embryonic patterning mutants by the complete and selective absence of the tracheal system [22] and later shown to encode a bHLH-PAS transcription factor , the earliest expressed tracheal-specific gene and a master regulator of tracheal identity [23] , [24] . Ten years after the discovery of trachealess , a Drosophila homolog of mammalian FGFRs was isolated and named Breathless because it is selectively expressed in the developing tracheal system and required for branching [25] , [26] , [27] . Around this time the first systematic screens for tracheal mutants were conducted , screens of P[lacZ] insertions that identified about 50 tracheal genes that subdivided embryonic tracheal development into genetically distinct processes including primary , secondary , and terminal branching , branch fusion , and tube size control [11] , [28] . Mapping and molecular characterization of these genes identified many components and modulators of the Breathless FGFR signaling pathway . These include Branchless FGF , which activates Breathless FGFR and plays a central role in controlling and coupling each of these processes by guiding outgrowth of primary branches and inducing expression of key genes encoding transcription factors such as pointed , blistered/pruned , and escargot required , respectively , for secondary and terminal branching and branch fusion [12] , [14] , [26] , [29] . Many other important genes have been identified by their tracheal expression patterns , analysis of candidate genes , and serendipitous discovery of a tracheal function for genes initially studied in other contexts [30] , [31] , [32] , [33] . And , over the past several years , several screens of chemically-induced mutations for tracheal morphogenesis defects in embryos and larval tracheal and air sac primordium clones have been conducted [34] , [35] , [36] along with more targeted genetic and genomic screens for genes that are expressed or function downstream of some of the key early signaling pathways ( branchless , breathless ) and transcription factors ( trachealess , ribbon ) [37] , [38] , [39] , [40] . Together these approaches have implicated ∼100 genes in tracheal development , most of which encode transcription factors or components of signaling pathways ( FGF , TGFα/EGF , TGFβ , Wnt , Notch , Slit/Robo , Jak/Stat , and Hedgehog ) [1] , [2] , [4] , [9] ( Table S1 ) . However , the downstream targets of the signaling pathways and transcription factors , the morphogenesis and maturation genes that create , shape , and stabilize the tubes , have only recently begun to be identified . And , although expected to be a large class , they are substantially under-represented among characterized tracheal genes ( Table S1 ) [4] . We conducted a large-scale screen of chemically-induced mutations to assess the function of nearly all Drosophila third chromosome genes , including early essential genes and genes with pleiotropic phenotypes . We sought to identify most or all of the genetically separable steps in tracheal development; to identify new tracheal genes associated with each step; and to provide an estimate of the total number of genes required to build an organ . We were especially interested in identifying tracheal morphogenesis genes . Our approach involved clonal analysis in the tracheal system of all chemically-induced mutations that did not survive late enough in development as homozygotes to assess their tracheal function , and a secondary screen to exclude general “house-keeping” genes . We isolated mutations representing ∼70 genes , 14 of which we identified molecularly , implicating most of the genes as morphogenesis genes and revealing new cell biological pathways in tracheal development . Many of the mutations affect terminal branch morphogenesis , genetically subdividing this poorly understood process into five major morphogenetic steps including an integral cell growth step . To identify tracheal morphogenesis genes , we screened the third chromosome for EMS-induced mutations that affect larval tracheal morphology . Approximately 4 , 300 mutagenized third chromosomes were generated , and balanced lines were established for each . Three-quarters ( 73% ) of the lines were homozygous lethal . Assuming a Poisson distribution , there were ∼1 . 3 lethal mutations per mutagenized third chromosome and a total of 5600 lethal mutations screened . Because there are ∼3600 essential Drosophila genes [6] , with roughly 40% ( ∼1370 ) on the third chromosome [41] , we expected to obtain an average of about four ( 5600/1370 ) mutations per gene , with at least one mutation in 97% of all third chromosome genes . The mutants were screened for tracheal defects in two steps . First , homozygous third instar larvae of the F3 generation ( Figure 1E ) , which carried btl-GAL4 and UAS-GFP transgenes ( abbreviated btl>GFP ) to label tracheal cells , were scored for tracheal defects by fluorescence microscopy . The balancer chromosome carried a Tub-GAL80 transgene that inhibits Gal4 and blocks expression of UAS-GFP so only homozygous mutant animals expressed GFP , facilitating screening ( Figure 1C ) . For 40% of lines , GFP+ F3 third instar larvae were not recovered , presumably because the homozygous mutations caused early lethality . These pre-pupal lethal lines were analyzed in a second step of the screen , using a genetic mosaic strategy in which we examined clones of homozygous mutant tracheal cells in otherwise heterozygous larvae ( Figure 1F ) . We devised a variant of the MARCM clone marking strategy [42] employing a UAS-GFP ( RNAi ) transgene on the homologous chromosome , in trans to the mutation of interest , which allowed us to label all tracheal cells with btl>DsRed and homozygous mutant tracheal cells ( lacking UAS-GFP ( RNAi ) ) with btl>GFP ( Figure 1D ) . This facilitated comparison of homozygous mutant cells ( DsRed+ , GFP+ ) with surrounding wild type tracheal cells ( DsRed+ , GFP- ) , enhancing the sensitivity of the screen and detection of cell non-autonomous effects in the tracheal system . Over 600 mutants with highly penetrant and expressive tracheal defects were identified . However , the vast majority were lethal mutations in genes required cell autonomously for tracheal cell growth and survival , and later found in a secondary screen ( see below ) to be presumptive housekeeping genes and discarded . 133 tracheal mutations were saved ( Table 1 and Table S3 ) , 18 from the F3 screen ( alleles with two letter prefix , e . g . PC213 ) and 115 from the genetic mosaic screen ( no prefix ) . Most of the mutations could be placed into one of five phenotypic classes ( Table 1 and Table S3 ) : ( 1 ) cell selection/specification; ( 2 ) cell size; ( 3 ) branch number , pattern size and shape; ( 4 ) tube formation , number , position , and shape; and ( 5 ) lumen clearance/gas filling . Within each broad class , phenotypic subgroups were defined and representative mutations in each subgroup were selected and subjected to detailed phenotypic characterization and genetic mapping as detailed below . Some mutations had more than one defect and were placed into more than one subgroup or class . Only a few mutants with cell non-autonomous effects were recovered from the clonal screen ( see below ) , implying that such tracheal mutations are rare . Complementation tests allowed assignment of 68 mutations to 24 loci ( Table 1 ) . In addition , four mutations that were mapped to specific chromosomal deficiencies were found to be new alleles of extant genes in the mapped intervals ( see below ) . In addition to these 72 definitively assigned mutations in 28 loci , we also characterized and named 30 other mutations with interesting tracheal phenotypes ( Table 1 ) . The rest of the saved mutations ( Table S3 ) were not extensively characterized; 12 of these are associated with mapped lethal mutations that complement extant tracheal mutations in the mapped interval , so may represent additional essential tracheal genes . It is difficult to estimate the number of mutations we obtained in previously known tracheal genes because the mosaic loss of function phenotype is not known for most tracheal genes , and the number of complementation tests necessary to determine this number directly is prohibitive . However , the apparent absence of mutations in two known tracheal genes ( stumps and trachealess ) whose mosaic phenotype we determined , and the lower than expected allele frequencies ( mean 2 . 6 ) obtained for the 28 definitively identified loci , indicate that the screen did not achieve the degree of saturation predicted by a Poisson distribution . Nevertheless , the screen was extensive so we think it is likely it identified mutations in most processes and molecular pathways involved in tracheal tube morphogenesis . Below , we describe each of the major phenotypic classes and subcategories , and representative mutations in each . Most of the mutations are homozygous lethal and all caused highly penetrant and expressive tracheal phenotypes . For ease in presentation , we treat the strongest phenotype in each complementation group as the null phenotype; however , we do not know for most if they truly represent the null condition because it is not readily possible to generate hemizygous ( mutant/deficiency ) clones for comparison or to exclude partial masking of phenotypes due to perdurance of wild type protein in mutant cells . These mutations eliminated specific tracheal cell types or blocked their differentiation . These mutants had their most profound affects on tracheal cell size . For nearly all mutations , the effect on terminal cell size correlated with branch number: larger cells had more terminal branches and smaller cells had fewer . One exceptional mutant , sprout , is presented below . These mutations affected the number of terminal branches , and in some cases also the position at which new branches bud from the parental branches . Many also affected the shape of the buds and mature branches . Mutations caused different but characteristic spectrums of defects so that , for example , mutations that reduced terminal cell branching to a similar extent could reproducibly give rise to terminal cells of very different morphology , such as short and thick versus elongate and wispy . These mutations prevented lumen formation , or altered the number , placement , or shape of the lumens that formed . Most of these mutations did not affect all tracheal tubes , but rather structurally distinct subsets of tubes . We start with a description of mutations that affect the seamless tubes of terminal branches ( see Figure 1B; Figure 5A–5F; Figure 6B ) . We expected that some mutants from the screen that appeared to lack lumens would instead be lumen clearance and gas-filling mutants in which the lumen was present but difficult to detect by brightfield optics because it remained filled with matrix , which has a similar refractive index as the surrounding cytoplasm . To identify such mutants , lines from the screen that appeared under brightfield optics to lack lumens were subjected to a secondary screen using a transgene expressing a fusion protein containing the signal peptide of p23 [50] linked to GFP . The fusion protein ( lumenal-GFP or lumGFP ) was designed to transit the secretory pathway , as it does in mammalian cells [50] , entering the tracheal lumen and remaining there to mark the lumen of mutants that affect lumenal clearance , such as ichorous and asthmatic ( Figure 6C ) . By contrast , no lumenal accumulation of lumGFP was observed in mutants such as impatent described above that lack or have seriously compromised lumens ( Figure 6B ) , or in wild type control clones because lumGFP is cleared from the lumen during the normal lumenal maturation process ( Figure 6A ) . The only lumGFP that remained in impatent mutant clones was the large puncta already described ( Section 4A1 ) , and the only lumGFP detected in control clones was the rare puncta near branch tips or the junctions between branches ( Figure 6A ) . In mutations such as scrub ( 659 ) , lumGFP was not detected in either matrix-filled lumens or cytoplasmic puncta ( data not shown ) ; these mutations might alter lumenal targeting such that lumGFP is secreted from other positions in the cell so does not accumulate intracellularly . To begin to define the molecular functions of tracheal genes identified in the screen , we mapped representative mutations and molecularly identified 14 of the genes ( Table 2 ) . Six of the identified genes ( no tc clones-L , no tc clones-R , short of breath , dyspneic , lopped and failed fusions ) were previously implicated in tracheal development . However , new functions were revealed for each by our clonal analysis . Two are allelic to canonical tracheal genes in the branchless/breathless FGF pathway , the breathless FGFR itself [26] and pointed [11] , [59] , which we showed are differentially required for competition during tip cell selection [13] . Two others , short of breath and dyspneic , which our results implicate in lumen clearance and gas-filling and as cell autonomous promoters of tube expansion , are allelic to krotzkopf verkehrt ( kkv ) and knickkopf ( knk ) , chitin synthesis pathway genes previously shown to coordinate the behavior of cells in the tracheal epithelium during tube expansion [53] , [54] , [60] , [61] . Our results demonstrate that chitin synthesis genes also have an unexpected , cell autonomous function in lumen clearance and gas filling of autocellular and seamless tubes . lopped784 is allelic to fatiga that encodes Drosophila Hif1 prolyl hydroxylase , and appears to be required in terminal cells for normal branching . Previous studies with hypomorphic fatiga alleles gave opposite results [56] , although new studies indicate that early exposure to hypoxia ( mimicked by loss of fatiga ) result in stunted tracheal development while later exposure stimulates branching [64] . failed fusions is allelic to polychaetoid , which has been implicated in branch fusion and in tracheal cell intercalation , but our mosaic analysis , along with other new data from our lab , suggest another function for polychaetoid in tip cell selection ( E . Chao , A . S . G . and M . A . K . , unpublished data ) . The other eight molecularly identified genes had not been previously implicated in tracheal development; indeed , three ( vine , moon cheese , and whacked ) had not been genetically defined ( Table 2 ) . All eight identify new cell biological pathways in tracheal development . jolly green giant , which encodes the Drosophila ortholog of TSC1 [55] , [65] , and miracle-gro ( see below ) [57] , [58] implicate general growth control pathways in tracheal growth and branching . tendrils , which is allelic to rhea and encodes talin [46] , and vine , which encodes the Drosophila ortholog of CCTgamma , show that talin-dependent integrin adhesion and a component of the TriChaperonin complex are required for maintenance of terminal branches and lumenal organization . The four other genes implicate membrane and vesicle trafficking genes in tracheal development . Such genes have been speculated to function in tube morphogenesis but few have been genetically identified . winded , essential for terminal branching , encodes the Drosophila homolog of CdsA [63] , an enzyme that converts phosphatidic acid to cds-diacyl glycerol in the production of the membrane lipid , phosphatidyl inositol . moon cheese , another terminal branching gene also implicated in lumen continuity , and burs , a terminal branching gene selectively required for side branches , encode the Drosophila homolog of the ER-Golgi t-SNARE membrin , and TSG101/erupted , a component of the ESCRTI complex that sorts endocytic vesicles to the multivesicular body , respectively [52] , [66] , [67] . whacked , which promotes the growth and proper shape of terminal cell lumens , encodes a putative RabGAP ( A . S . G . and M . A . K . , unpublished data ) . The identification of membrane lipid and vesicle trafficking genes in terminal branching supports the idea that outgrowth of cellular processes and lumen formation require targeting of apical and basolateral membrane components at a distance from the cell soma . It will be important to determine the number of trafficking pathways involved , how the pathways are activated at the appropriate times and places , and how the identified t-SNARE , Rab-GAP , and ESCRTI component function in the pathways . Thus , all 14 molecularly characterized genes from the screen reveal new cell biological pathways in tracheal development or new functions for established pathways . The identities of the molecularly characterized genes allowed us to assess the success of the screen in identifying morphogenesis effectors . Although it was not possible to unambiguously classify all 14 genes in this way from their sequence alone , eight very likely function as morphogenesis effectors: the vesicle trafficking genes moon cheese/membrin , burs/TSG101 , and whacked/RabGAP; the cell junction and cytoskeletal genes failed fusions/polychaetoid/ZO-1 , rhea/tendrils/talin , vine/cctγ; and the chitin synthesis genes short of breath/kkv/chitin synthase and dyspneic/knk . Three others are established patterning genes: the receptor btl/no-terminal cell clones-L , the transcription factor pnt/no terminal cell clones-R , and the transcription factor regulator lopped/fatiga/Hif prolyl hydroxylase . The remaining three are more difficult to categorize because they encode enzymes that likely couple patterning signals to cytoplasmic outgrowth ( winded/ ( CdsA ) and cell growth ( Tsc1/jolly green giant and miracle-gro ) , as discussed below . Thus , over three-quarters of the identified genes ( 11 of 14 , 79% ) appear to be downstream effectors/morphogenesis genes ( 8 of 14 , 57% ) or genes that couple patterning signals to morphogenesis ( 3 of 14 , 21% ) , supporting our hypothesis that systematic clonal analysis is an effective way of identifying such genes . In addition to identifying new tracheal genes and pathways , the screen suggested new functional connections between pathways . One example came from characterization of the miracle-gro cell overgrowth mutations ( Section 2A ) . In terminal cells , not only was the soma enlarged but there were many ectopic seamless tubes coursing through it ( Figure 4C , 4C′ and Figure 7B ) . This phenotype is nearly unique: it is seen otherwise only upon hyperactivation of the Breathless FGFR pathway ( Figure 7C ) . However , miracle-gro mutations did not map near breathless or any other extant loci in the pathway . Mapping and complementation tests ( Table 2 ) demonstrated that miracle-gro is allelic to warts/lats-1 [57] , [58] , which encodes a kinase that suppresses cell growth in a well-established general growth control pathway . Thus , loss of a key growth regulator in terminal cells leads not only to excessive cell growth but excessive lumen formation , revealing an unexpected coupling between cell growth control and tubulogenesis . The striking similarity of the warts/miracle-gro loss of function phenotype and the btl pathway gain-of-function phenotype suggested that FGF signaling might stimulate terminal cell growth and tubulogenesis by inactivating Warts function . To test this , we sought to define the genetic epistasis relationship between warts/miracle-gro and breathless-FGFR pathway mutations . Because mutations that disrupt FGF signaling abrogate terminal cell specification [11] , [29] , it is not possible to generate terminal cells doubly mutant for warts and breathless , so we examined terminal cells doubly mutant for warts/miracle-gro and blistered/pruned/SRF , the downstream transcription factor in the breathless FGFR pathway required for terminal cell growth and branching . Doubly mutant cells were unbranched and small , similar or slightly bigger than blistered mutant terminal cells , and with a single lumen in the soma ( Figure 7D ) . However , the lumen diameter was larger than normal and similar in size to those in warts/miracle-gro mutant terminal cells . Thus , the cell growth and excessive lumen formation seen in warts/miracle-gro mutant terminal cells are dependent on blistered , whereas lumen diameter can be modulated independently of blistered . This supports a model in which Warts/Miracle-gro functions downstream of Breathless FGFR but upstream of Blistered/SRF in the regulation of terminal cell size and lumen number , and upstream of another , as yet unidentified , transcription factor that controls lumen diameter ( Figure 7E ) . The earliest step of cell selection and specification in terminal branching is a well-characterized process that previous genetic studies have shown is controlled by the Branchless FGF pathway ( Bnl/Btl/RAS/Pointed ) that induces expression of the Blistered/Pruned SRF transcription factor that selects cells at the ends of budding branches for a terminal branching fate [9] , [10] . Re-expression of Bnl FGF later in hypoxic tissues is proposed to initiate terminal branch budding , at least in part by activation of the blistered SRF transcription complex and its downstream effector genes [14] , [45] . Our screen identified third chromosome genes previously implicated in the selection process , but revealed an interesting new aspect of the process because of the novel “no terminal cell clones” phenotype . These turned out to be mutations in breathless FGFR and pointed [13] , and lead to the discovery there is specialization among cells in a budding branch and only the leaders need to receive the Branchless FGF signal . Cells mutant for Breathless FGFR cannot receive the signal , and are relegated to trailing positions , never to be specified a terminal cell . We also discovered genes ( steeple , missing parts ) required for specification of a subset of terminal cells in specific regions or branches . These may encode region-specific enhancers of the Bnl-Btl pathway because sporadic failure of terminal cell formation is seen in animals in which this signaling pathway is partially compromised [11] , [13] , [34] . Although the Branchless pathway and Blistered SRF transcription complex are key regulators of branch budding and outgrowth [14] , [29] , little is known of the signal transduction pathway that connects them or of the downstream effectors . winded/cdsA mutations caused a severe , cell autonomous terminal branch pruning phenotype similar to that of blistered/SRF null alleles . winded/cdsA encodes an enzyme ( CDP-diacylglycerol synthase ) required for phosphoinositide ( PI ) synthesis , suggesting that a PI-dependent signaling process , presumably like those involved in other receptor tyrosine kinase ( RTK ) signaling pathways [68] , functions downstream of the Btl RTK in the control of Blistered/SRF and branch budding . Some of the other genes with pruned phenotypes ( e . g . , topiary , paltry , truncated ) might encode additional signal transduction components or targets of the SRF transcription factor required for polarized cell growth ( see below ) . We propose that other branching genes regulate bud site selection and the pattern of branching . spikes encodes a negative regulator of bud site selection because small ectopic buds form in mutant terminal cells . One appealing idea is that spikes restricts the normal budding response of terminal cells to the sites of maximal induction by Branchless FGF . TSG101/erupted/burs and denuded regulate branch pattern by promoting lateral and late rounds of terminal branching , perhaps by catalyzing the local disassembly or reorganization of the cytoskeleton within a maturing terminal branch . One set of genes ( lotus , oak gall , conjoined ) affected branch number but also dramatically altered the size and shape of the remaining branches ( see below ) . These were difficult to categorize purely as branching genes or growth regulators , so we place them in a special class at the boundary between those categories because they share features of both . They may function as integrators of branch outgrowth and size control signals . A new aspect of the tracheal developmental program highlighted by the mutants is cell size and growth regulation . Outgrowth of terminal branches requires not only chemoattractant signaling to induce and guide migration , but synthesis of cellular and membrane components to support cytoplasmic outgrowth . Terminal cells mutant for glutamyl-prolyl-tRNA synthetase or hundreds of other presumptive house keeping genes failed to form and extend terminal branch buds . Many such growth-promoting genes were presumably among those identified in a previous clonal screen for tracheal mutations [35] , but because there are many such mutations and their phenotypes are non-specific ( small , sick , or missing cells ) and difficult to distinguish from genes simply required for cell viability , it was hard to evaluate their developmental significance . Three types of data argue that growth control is an integral part of the tracheal developmental program . First , clonal analysis of the master regulator trachealess in terminal cells gave a similar phenotype ( Figure S1 ) , implying that terminal cell growth is a process actively regulated by Trachealess . etiolated is a particularly interesting mutant of this class because it resembled trachealess not just in its clonal phenotype in terminal cells , but in its tracheal specificity . Second , we obtained mutations in two canonical growth suppressor genes , warts/miracle-gro and Tsc1/jolly green giant , which gave the opposite phenotype: terminal branches and terminal cells were overgrown with particularly large somas that in warts/miracle-gro mutant cells contained multiple seamless tubes passing through them . This shows that a general growth regulator controls not only cell size but tubulogenesis , an essential step in the tracheal developmental program . Third , the phenotype of warts/miracle-gro mutant terminal cells is very similar to that of activated Btl , and genetic epistasis experiments suggest that warts/miracle-gro functions downstream of , and is negatively regulated by , btl FGFR but upstream of blistered SRF ( Figure 7E ) . sprout is the most intriguing undergrowth mutant because it was the only one that formed small but normally patterned terminal branches . sprout cleanly decouples branch size from branch budding and outgrowth , so we propose it is a key gene in branch size control . Three other genes , lotus , oak gall , and conjoined , also function in branch size control , but in a different way . In mutant cells , branches were much thicker and more variable than normal , but the diameter of the seamless tubes that form within them were normal . We propose that these genes function in the size control pathway by regulating the distribution of plasma membrane or other cell constituents among branches . When this process fails , branches become thicker and fewer in number . Most of the undergrowth mutations and all of the overgrowth mutations affected not only terminal cells but other tracheal cell types and cells outside the tracheal system , implying that the affected genes encode general growth regulators . However , many undergrowth mutations had their most extreme effects on terminal cells , presumably because they are larger and grow more than other cells . But two mutations , cincher and corset , affected the growth of dorsal trunk cells and spared terminal cells . Thus , the growth control programs of these tracheal cell types are genetically separable . Because terminal cell growth appears to be controlled primarily or exclusively by Bnl-Btl signaling and operates selectively under hypoxic conditions , conditions that arrest the growth of most other cell types , cincher and corset might identify specific regulators or components of aerobic growth pathways or other general growth processes dispensable in terminal cells . The striking phenotype of impatent mutant terminal branches , branches that superficially appear normal but lack air-filled tubes and hence are nonfunctional , leads us to propose that impatent encodes a key regulator or component of terminal branch lumen formation . We further propose that lumen shape is governed by cystic lumens , perhaps in conjunction with whacked , mutations in which result in irregularly-shaped lumens , and that lumen length and position are controlled by wobbly lumens and wavy lumens , mutations that cause long and convoluted lumens . Long and convoluted lumens are also seen in terminal cells under hypoxic conditions and other conditions that cause excessive branchless FGF pathway activity and/or terminal cell growth ( e . g . warts/miracle-gro ) , so an appealing model is that these conditions and this signaling pathway inhibit the activity of wobbly lumens and wavy lumens , which themselves function to restrict the length or the position of terminal branch lumens . Lumen continuity requires carbuncle and membrin/moon cheese , suggesting that lumens of seamless tubes are made piecemeal and these genes promote their connection . We also identified four genes ( disjoined , lotus , conjoined , and oak gall ) required to make functional connections between seamless tubes and the adjacent , architecturally distinct autocellular tubes that form by wrapping . The short lumenal gap in these mutant terminal cells may result from a failure to connect the tubes or a structural defect that prevents clearance of the connection . For tracheal branches to become functional , the lumenal matrix must be cleared and replaced by gas , the molecular composition of which is unknown . ichorous and asthmatic are required for clearance and gas filling of most or all terminal branches , and littoral and panting and others are required to clear the tips of terminal branches . Recent studies highlight the importance of secretion into the lumen and subsequent endocytosis of lumenal matrix and liquid during tube expansion and air-filling of large multicellular tracheal tubes [16] , [69]; the genes identified in our screen may mediate related processes in seamless tubes . These genes maintain the elaborate shape and structure of terminal branches under mechanical stress such as muscle contraction . In the mutant rhea/tendrils/talin [46] , terminal branches begin to form normally but branches break down and their lumens retract as the larva begins to move and the developing branches are subjected to the stress of stretch . The phenotypes of cctgamma/vine , creeper , braided , and ivy are similar to tendrils , suggesting that they function in the same integrin/talin cell adhesion and cytoskeletal support system . For example , the cctgamma/vine chaperonin may facilitate the folding or assembly of talin or some other component in the support system , an appealing hypothesis given that CCT chaperonins have been shown in other systems to mediate the folding of cytoskeletal proteins [70] . These maintenance genes emphasize the importance of analyzing the onset and evolution of a mutant phenotype when elucidating gene function , because similar phenotypes can arise from early developmental aberrations or later defects in maintenance . Other elaborate cell types and organs likely also require maintenance genes . For example , mutations in mouse Dlg5 perturb delivery of adherens complex proteins to the plasma membrane of brain and kidney epithelial tubes , resulting in cyst formation [71] . Many such structural maintenance genes are expected to function late in life so would be missed in typical developmental screens . A major effort should be aimed at their identification and isolation because of their importance in medicine and disease . How are the genetically separable terminal branch morphogenesis processes described above coordinated and controlled in time and space ? An important part of this control and coordination almost certainly involves the Branchless FGF pathway . Expression of both the ligand branchless and the receptor breathless are induced by hypoxia during larval life , and terminal cell outgrowth and branching are stimulated and directed to hypoxic cells by local production of Branchless FGF [56] . One way Bnl-Btl signaling stimulates branching is likely through transcriptional induction of morphogenesis genes via modification and activation of the Blistered /SRF transcription complex . In other systems , the SRF transcription complex has been shown to be regulated by the actin cytoskeleton and to regulate cytoskeletal genes [72] , and such genes are almost certainly required for growth of the actin-rich terminal branch buds . It will be important to determine which of the identified morphogenesis genes are regulated by Branchless signaling and SRF , and to identify the full set of downstream targets by transcriptional profiling . Because Branchless functions as a chemoattractant , it must also provide a spatial cue that guides terminal branch outgrowth . One appealing idea is that the ligand-bound Breathless FGFR at the surface of the terminal cell generates a spatial cue that directs polarized growth of cytoplasmic extensions toward hypoxic , FGF-secreting cells . Such a spatial cue could be used to direct vesicular traffic to the growing ends of terminal cell extensions , both for polarized growth of the new branch and construction of a lumen within it . The spatial cue might be a modified membrane PI because PI signaling functions downstream in many RTK signaling pathways [68] and can regulate vesicle trafficking [73] and tubulogenesis [74] , and mutations in the PI synthesis gene cdsA/winded severely abrogated terminal branching . Although Branchless-Btl signaling likely controls and coordinates many of the events in terminal branch morphogenesis , it is unlikely to be the sole control and coordination mechanism because not all of the morphogenesis events occur at the same time and place . Lumen formation ( tubulogenesis ) occurs after cytoplasmic outgrowth , and lumen maturation including clearance and gas filling occur even later , in some cases days after the lumen has formed . Likewise , branch and lumen maintenance are late steps in the process . Although there may be delays built into some of the effector pathways downstream of Breathless to stagger its effects , other factors likely also contribute to the timing and spatial organization of the events . For example , ecdysone signaling may gate the timing of lumen clearance and gas filling , and there are presumably cell intrinsic cues that direct transport vesicles carrying integrins and other basolateral markers to the plasma membrane of growing buds , and vesicles carrying apical markers and lumenal components to internal positions . One surprising finding of our clonal analysis of known tracheal genes was that terminal cell clones of the tracheal master regulator trachealess gave a pruned phenotype . This implies that trachealess is required not only for its well established role in the initiation of tracheal development [23] , [24] , but also for much later steps in the developmental program such as terminal cell growth and branching . Perhaps it functions in conjunction with SRF and other cell type and stage-specific transcription factors in the program to impart tracheal specificity in the control of downstream effector genes , as shown for the C . elegans pharyngeal master regulator pha-4 [75] . We identified a number of genes required for proper formation of the larger branches of the tracheal system that form earlier than terminal branches and from which they arise . For example , we identified mutants required for proper shape of multicellular tubes , including mutations that cause tracheal dilatations ( small potatoes , bulgy , balloon ) and others that cause local constriction of multicellular and autocellular tubes ( kkv/short of breath , knk/dyspneic , constricted ) . An especially intriguing set of genes ( lotus , conjoined , oak gall ) are those required to form autocellular junctions and lumens . These may encode specialized components or regulators of autocellular junctions and tubes , such as proteins required for a cell to wrap on or seal to itself . Although we identified some primary and secondary branch morphogenesis genes , there was a surprising paucity of such genes relative to the large number of terminal cell branching genes identified; a similarly skewed distribution obtained in a second chromosome screen , if the large number of putative housekeeping genes is excluded [35] ) . Although it is possible that morphogenesis of these larger branches requires fewer genes , more likely such genes were just not as efficiently identified in our screen . One reason is that some such genes only show a tube phenotype when most or all cells in the branch are mutant , as with breathless ( Figure 2B ) and grainyhead mutations [76] . Another reason is that perdurance of maternally expressed gene products likely obscures early functions of some genes . Finally , terminal cells have an elaborate structure that may make them more sensitive to mutations and makes phenotypes easier to detect . Our systematic genetic dissection of an organogenesis process , including a clonal analysis to identify tracheal genes with pleiotropic functions , allows an estimate of the number of genes required to build an organ–an important question not just for developmental biology but for medicine and tissue engineering . Because we identified ∼70 tracheal genes on the third chromosome ( Table 1 and Table S3 ) , which represents ∼40% of the genome , the full genome likely contains roughly two hundred genes required to construct the larval tracheal system . This almost certainly represents a lower limit because our screen did not achieve full saturation and , as described above , the screen would miss essential embryonic genes required non-cell autonomously and genes with a significant maternal contribution . Genomic profiling of developing and mature organs indicates that there are hundreds of differentially expressed genes among different organs , and genetic profiling to identify downstream genes of organ master regulators such as the C . elegans pharyngeal regulator pha-4 [77] , the mouse pancreas regulator Pdx1 [78] , [79] , and the Drosophila tracheal regulator Trachealess ( E . Chao and M . A . K . unpublished data ) , suggests that there are 110–240 genes dependent on the master regulator for expression , at least at certain stages of development . Although it is not known how many of these downstream genes are required for organ morphogenesis , or what fraction of organ morphogenesis genes are both selectively expressed and downstream targets of organ master regulators , these genomic results are in line with the estimate from our genetic studies that organ morphogenesis programs , even ones for relatively simple organs like the Drosophila tracheal system , are likely to involve several hundred genes . The approach used here , involving a clonal analysis in the tracheal system of all mutations that do not survive late enough in development as homozygotes to assess their tracheal function , has begun to be extended to the other major chromosomes to identify the rest of the tracheal morphogenesis program [7] , [35] ( Metzstein M . and M . A . K . , unpublished data ) ; most of the identified mutations fit with the genetic scheme described here , with the exception of a novel set of mutations on the second chromosome that compromise terminal branch mutual avoidance and spacing [35] . The clonal approach could easily be adapted to other organs to systematically dissect additional organogenesis programs . D . melanogaster strains used in the screen and meiotic mapping experiments were: ( 1 ) btl-GAL4 , UAS-GFP; Pr , Hs-hid/TM3Sb , Tub-GAL80 , ( 2 ) a newly isogenized btl-GAL4 , UAS-GFP; FRT2A , FRT82B , ( 3 ) y w FLP122; btl-GAL4 , UAS-DsRED; FRT82B cu UASi-GFPhp/TM6B , ( 4 ) y w FLP122; btl-GAL4 , UAS-DsRED; UASi-GFPhp th st FRT2A/MKRS ( 5 ) y w ey-FLP; cell-lethal , GMR-hid FRT2A/MKRS , ( 6 ) y w ey-FLP; FRT82B , cell-lethal , GMR-hid/MKRS , ( 7 ) y w FLP122; breathless-GAL4 , UAS-lumGFP , UAS-DsRED; FRT82B TubGal80 , ( 8 ) y w FLP122; breathless-GAL4 , UAS-lumGFP , UAS-DsRED; TubGal80 FRT2A; ( 9 ) a newly isogenized ru h th st cu sr e ca; ( 10 ) ru h th st cu sr e Pr ca/TM6B . Other strains were: FRT2A , FRT82B ( from Trudi Schüpbach ) ; Hs-hid ( on chromosome III; from Ruth Lehman ) , onto which Pr was recombined; TM3Sb , Tub-GAL80 ( from Stefan Luschnig ) ; and strains used in complementation tests and deficiency mapping experiments ( see Table S3 ) . All other strains , except the mutants isolated here , have been described ( http://flybase . bio . indiana . edu ) and are available from http://flystocks . bio . indiana . edu . UAS-DsRed . This Gal4-dependent DsRed transgene was constructed by inserting a 0 . 7 kb Kpn I-Xba I restriction fragment containing the DsRed coding sequences from pDsRed ( Clontech ) between the corresponding sites of the vector pUAST [80] . The resultant plasmid , pUAST-DsRed , was used to establish transgenic lines on the X , second , and third chromosome by P element mediated transformation of w1118 embryos . The second chromosome insertion ( line 5A ) was recombined with breathless-GAL4 and used here . pUASTi . This P element vector for generating RNAi transgenes was constructed by PCR amplification ( primers Xho+trh-intron F , Kpn+trh-intron B; see Table S2 for primer sequences ) and TA cloning of the 221 bp third intron from the trachealess gene into the vector , pCRII-TOPO ( Invitrogen ) . The intron fragment was then excised with Xho I and Kpn I and inserted at those sites in pUAST . Note that the trachealess intron contains an Eco RI site , leaving Bgl II , Not I and Xho I as the only unique restriction sites 5′ of the trachealess intron , and Kpn I and Xba I as the only unique sites 3′ of the trachealess intron . To generate RNAi constructs , a ∼500 bp fragment from the gene of interest is inserted in the forward orientation just upstream of the intron , and in the reverse orientation downstream of the intron , as described below for UAS-GFP ( RNAi ) . Gal4-driven expression of this transgene results in tissue specific transcription of the self-complementary RNA , which is predicted to form a double-stranded “hairpin” conformation that initiates the RNAi response . UAS-GFP ( RNAi ) . This Gal4-dependent GFP ( RNAi ) transgene was created by PCR amplification ( primers Not-GFP-F , Xho-GFP-R ) and insertion of an ∼500 bp fragment of GFP ( in the forward , sense orientation ) between the Not I and Xho I sites upstream of the intron in pUASTi , and amplification of the same fragment ( primers Xba-GFP-F , Kpn-GFP-R ) and insertion ( in the reverse orientation ) between the Kpn I and Xba I sites downstream of the intron , to create plasmid pUASTi-GFP ( RNAi ) . Transgenic flies were generated as above , and insertions were identified on all major chromosomes . For this study , an insertion on 3L ( insertion B ) was recombined onto ru h th st FRT2A ( from Stefan Luschnig ) to generate the UASi-GFPhp th st FRT2A chromosome , and an insertion on 3R ( insertion 4A ) was recombined onto FRT82B cu sr e ca ( from S . Luschnig ) to generate the FRT82B cu UASi-GFPhp chromosome . UAS-lumGFP . This Gal4-dependent transgene expressing secreted ( lumenal or “lum” ) GFP was constructed by inserting an Nhe I-Kpn I restriction fragment with the GFP coding sequence from plum-GFP [50] between the Xba I and Kpn I sites of pBS-KS ( Stratagene ) , and then subcloning the Not I/Kpn I lum-GFP fragment into those same sites in the pUAST vector . Transgenic flies were generated as above , and second and third chromosome insertions were recovered . A second chromosome insertion was recombined onto a breathless-GAL4 bearing second chromosome for use here . A standard F3 EMS mutagenesis screen was performed ( Figure 1C ) . Strains used were homozygous for breathless-GAL4 [81] and UAS-GFP transgenes on chromosome II . Males homozygous for an isogenized FRT2A , FRT82B chromosome III were fed 25 mM EMS as described [82] and mass mated to breathless-GAL4 , UAS-GFP; Pr , Hs-hid/TM3 , Sb , Tub-GAL80 virgin females . F1 males were each mated to two virgins of the genotype used in the P cross . After five days , parents were removed from the F1 cross and on days five and six F2 larvae were heated to 38°C for 1 . 5 hours to induce Hs-hid and eliminate animals carrying that chromosome . In the few cases ( ∼2–3% ) where animals carrying Pr , Hs-hid survived , virgins of the appropriate genotype were selected to generate a stock of FRT2A , FRT82B*/TM3 , Sb , Tub-GAL80 ( * , newly induced mutation ) . If animals homozygous for the treated third chromosome ( non-Sb ) were not detected in the F3 or subsequent generations , a lethal mutation was assumed to be present . Two hundred mutagenized chromosomes were assayed in three small-scale pilot screens , and 4100 mutagenized chromosomes were assayed in a final large-scale screen . Sibling F2 flies ( described above ) were allowed to mate and were brooded to produce two clutches of F3 individuals , one-quarter of which should be homozygous for the mutagenized third chromosome . The first F3 brood ( after five days ) was used to maintain the stock and assess for presence of a lethal mutation; the second F3 brood ( at 12 days ) was screened for tracheal phenotypes ( see below ) . If less than four pairs of flies were obtained in the F2 generation , screening was postponed for a generation . For tracheal phenotype screening , F3 larvae were washed out of their food vials with distilled water and examined under an M2 Zeiss or a Leica fluorescence stereomicroscope . Homozygous third instar larvae were identified by tracheal expression of GFP , and tracheal morphology was analyzed; at least three homozygous larvae were examined for each mutant line . Animals that appeared to have tracheal defects were heat-killed ( 70–75°C for 3–5 s ) , mounted in 50% glycerol and examined under a Zeiss Axioplan 2 compound fluorescence microscope . Lines in which a tracheal defect was detected were retested to confirm the phenotype . If no reproducible phenotype was found , the line was discarded . Mutant lines that did not give rise to viable homozygous third instar animals were analyzed in genetic mosaics as follows . Males from the mutant stock established in the F3 screen were crossed to y , w , FLP122; breathless-GAL4 , UAS-DsRED; UASi-GFPhp , th , st , FRT2A/MKRS virgin females , and to y , w , FLP122; breathless-GAL4 , UAS-DsRED;FRT82B , cu , UASi-GFPhp/ TM6B virgin females , to test mutants on 3L or 3R , respectively . For each arm ( 3L and 3R ) of every mutant stock , a cross with 20–40 pair matings was done . Embryos ( 0–4 hr old ) were heat-treated as above for 0 . 75–1 hr to induce FLP-mediated recombination; animals 2 hrs old or less at the time of heat treatment typically do not survive . Crosses were maintained at 25°C for five days and then mosaic animals were examined under a fluorescence stereomicroscope . All tracheal cells are marked by expression of UAS-DsRED; mutant tracheal cells also express GFP . This GFP marking of mutant cells was achieved by inducing recombination between a chromosome arm carrying the mutation of interest and the homologous chromosome arm carrying UASi-GFPhp ( see above ) . Daughter cells homozygous for the mutagenized chromosome arm lack the GFP ( RNAi ) transgene and thus express GFP . Animals of the correct genotype were selected , heat killed , and analyzed for tracheal defects as above . Under these clone induction conditions , ∼60±8 ( mean +/- SEM ) tracheal clones were generated per animal ( n = 5 animals ) . Among dorsal branch clones ( n = 127 clones ) , 50% appeared to be composed of a single cell , 37% of two cells , 10% of three cells , and 3% of four or five cells . Mutations that caused undergrowth defects in tracheal clones were tested for growth defects in eye development using the EGUF/HID technique that generates eye imaginal discs composed exclusively of mitotic clones of a single genotype [43] . Males from mutant lines were crossed to virgin females of genotype y , w , ey-FLP; * , GMR-hid FRT2A/MKRS , or y , w , ey-FLP; FRT82B , * , GMR-hid/MKRS , where * is the undergrowth mutation . Adult progeny lacking Sb ( carried on both of the balancer chromosomes used ) were scored for eye size . Mutations unable to support normal eye development were presumed to affect general cell growth and viability genes ( “housekeeping” genes ) and were discarded . Mutants with no detectable lumen under brightfield optics were tested for presence of liquid-filled or discontinuous lumens using the lumGFP transgene described above , which expresses GFP with a signal peptide that we found accumulates in liquid-filled tracheal lumens but is not detectable in normal , gas-filled lumens . Males from the mutant stocks were crossed to y w FLP122; breathless-GAL4 , UAS-lumGFP; TubGal80 FRT2A/MKRS or y w FLP122; breathless-GAL4 , UAS-lumGFP; FRT82B TubGal80/MKRS virgin females , and mosaic analysis was carried out as described above in the Genetic Mosaic Screen . Effect of miracle-gro338 on cell size was determined using ImageJ software to measure the maximal cross-sectional area in a stack of 2D optical sections through the soma of mutant and wild type control terminal cell clones ( see Figure 3 ) . Effect on dorsal trunk lumen diameter was determined by comparing lumen diameter between a section of tube containing a single mutant cell and the average diameter of the immediately anterior and posterior regions containing no mutant cells . Initial mapping was carried out by complementation tests against a panel of chromosomal deficiencies spanning the third chromosome , and by meiotic recombination mapping using visible recessive markers ru h th st cu sr e ca . Fine scale mapping was carried out using available single nucleotide polymorphisms ( SNPs ) [83] , [84] and new ones specifically identified in this study , in conjunction with complementation tests with chromosomes carrying small , molecularly characterized deletions . The affected gene in the mapped interval was then identified by sequencing candidate genes and comparing their sequences to those in the isogenized parental chromosome to reveal new EMS-induced nonsense mutations or other mutations predicted to compromise gene function , and by complementation tests with extant mutations in the interval . To compare the miracle gro/warts terminal cell phenotype to that of activated Breathless FGFR , the MARCM system [42] was used to generate marked ( GFP+ ) clones of tracheal cells expressing λ-Breathless [85] , a constitutively dimerized form of the protein , and marked cells at terminal cell positions were examined using a compound fluorescence microscope . To determine the genetic epistasis relationship between miracle-gro/warts and blistered/pruned/SRF , a downstream transcription factor in the Breathless pathway , virgins of the genotype y , w , hsFLP122; bsl ( 2 ) 3267 , btl-Gal4 , UAS-GFP/CyO; FRT82B cu , UAS-GFP ( RNAi ) were crossed to males of the genotype bsl ( 2 ) 3267 , btl-Gal4 , UAS-GFP/CyO; FRT82B miracle gro/warts338/MKRS . Mutant animals homozygous for bsl ( 2 ) 3267 were identified by the strong pruned phenotype of unmarked ( GFP- ) terminal cells , and marked ( GFP+ ) wts338 mutant terminal cell clones in these animals were examined and scored as above .
Elucidating the genetic programs that control formation and maintenance of body organs is a central goal of developmental biology , and understanding how these programs go awry in disease has important implications for medicine . Many such organogenesis genes have been identified , but most are early-acting “patterning genes” encoding signaling proteins and gene regulators that control expression of a poorly characterized set of downstream “morphogenesis genes , ” which encode proteins that generate the remarkable organ forms and structures of the constituent cells . We screened ∼40% of the fruit fly Drosophila genome for mutations that affect tracheal ( respiratory ) system development . We included steps to bypass complexities from mutant effects on other tissues and steps to exclude mutations in general cell “housekeeping genes . ” We isolated mutations in ∼70 genes that identify major steps in the organogenesis program including an integral cell growth control step . Many of the new tracheal genes are “morphogenesis genes” that encode proteins involved in cell structure or intracellular transport . The results suggest that genetic screens can elucidate a full organogenesis program and that over 200 patterning and morphogenesis genes are required to build even a relatively simple organ .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "developmental", "biology", "organism", "development", "genetic", "screens", "genetics", "organogenesis", "biology", "morphogenesis", "genetics", "and", "genomics" ]
2011
A Systematic Screen for Tube Morphogenesis and Branching Genes in the Drosophila Tracheal System
To characterize the consequences of eliminating essential functions needed for peptidoglycan synthesis , we generated deletion mutations of Acinetobacter baylyi by natural transformation and visualized the resulting microcolonies of dead cells . We found that loss of genes required for peptidoglycan precursor synthesis or polymerization led to the formation of polymorphic giant cells with diameters that could exceed ten times normal . Treatment with antibiotics targeting early or late steps of peptidoglycan synthesis also produced giant cells . The giant cells eventually lysed , although they were partially stabilized by osmotic protection . Genome-scale transposon mutant screening ( Tn-seq ) identified mutations that blocked or accelerated giant cell formation . Among the mutations that blocked the process were those inactivating a function predicted to cleave murein glycan chains ( the MltD murein lytic transglycosylase ) , suggesting that giant cell formation requires MltD hydrolysis of existing peptidoglycan . Among the mutations that accelerated giant cell formation after ß-lactam treatment were those inactivating an enzyme that produces unusual 3->3 peptide cross-links in peptidoglycan ( the LdtG L , D-transpeptidase ) . The mutations may weaken the sacculus and make it more vulnerable to further disruption . Although the study focused on A . baylyi , we found that a pathogenic relative ( A . baumannii ) also produced giant cells with genetic dependencies overlapping those of A . baylyi . Overall , the analysis defines a genetic pathway for giant cell formation conserved in Acinetobacter species in which independent initiating branches converge to create the unusual cells . In spite of controlling the most fundamental biological processes , essential genes are usually missing from loss-of-function mutant screens because strains carrying null mutations are not represented . Although essential functions can be studied using conditional alleles , such as temperature-sensitive or regulated expression alleles , suitable alleles that support normal growth under permissive conditions while fully eliminating activity under non-permissive conditions can be difficult to isolate [1 , 2] . In the work reported here , we describe the use of gene deletions generated by natural transformation as an alternative to conditional alleles for studying essential functions . We employed the approach to examine the consequences of inactivating genes needed for peptidoglycan synthesis in Acinetobacter baylyi , a Gram-negative bacterium belonging to the Gammaproteobacteria [3] . We focused on peptidoglycan because it is the major determinant of cell shape and mutations affecting it can produce dramatically altered cell morphologies [4–6] . Peptidoglycan is an essential glycopeptide mesh situated between the two membranes of Gram-negative bacteria , which helps protect cells from lysis due to turgor pressure . Peptidoglycan is constructed by complex mechanisms in which a lipid-linked disaccharide pentapeptide precursor is incorporated into peptidoglycan through the action of transglycosylase and transpeptidase activities ( Fig 1 ) . Parallel sets of enzymes are required for cell elongation and division [7–11] . Peptidoglycan growth reflects a balance between synthetic and degradative activities [7 , 8] . When new peptidoglycan biosynthesis is blocked , hydrolytic activities degrade the existing sacculus and bacteria usually lyse due to osmotic swelling . However , under osmoprotective conditions , cells are prevented from lysing and may proliferate as pleiomorphic wall-deficient cells called L-forms [12–15] . These bacteria have been an object of fascination for decades due to their capacity for growth without a peptidoglycan sheath , striking cell morphologies and antibiotic resistance [14 , 15] . It appears that the formation of L-forms requires not only loss of peptidoglycan synthesis , but also additional mutations that allow the wall-deficient forms to proliferate by membrane tubulation and blebbing [5] . In some cases , loss of peptidoglycan synthesis leads not to L-forms , but to non-proliferating cells that have lost their normal shape and may enlarge considerably [5 , 12 , 13 , 16–19] . For example , in one well-characterized example , mutations inactivating an E . coli cytoskeletal protein needed for elongation peptidoglycan synthesis ( MreB ) indirectly inactivate division by sequestering a division protein ( FtsZ ) in internal membranes [13] , leading to giant cells . In other cases , blocking peptidoglycan synthesis converts the bacteria into small non-dividing spherical cells [16 , 17 , 20] . In Vibrio cholera , formation of such spheroids requires murein endopeptidase activity [16] . The small spherical cells can grow when peptidoglycan synthesis resumes; their formation thus represents a novel antibiotic tolerance mechanism . Genetic determinants of this tolerance have been defined [20] . In this study , we examined the consequences of disrupting peptidoglycan synthesis in Acinetobacter baylyi . There are several advantages of studying peptidoglycan synthesis in this bacterium . First , the species undergoes natural transformation at high efficiency [21 , 22] , making it straightforward to generate deletion mutations in essential genes . Second , peptidoglycan synthesis associated with cell elongation is fully dispensable , making it simple to genetically manipulate septal synthesis in the absence of elongation synthesis . Third , like E . coli , A . baylyi belongs to the Gammaproteobacteria , and the detailed understanding of peptidoglycan synthesis and cell division in E . coli should provide a good foundation for understanding the processes in A . baylyi . Finally , A . baylyi peptidoglycan metabolism should be similar to that of the related nosocomial pathogen Acinetobacter baumannii . Understanding the consequences of disrupting peptidoglycan synthesis in A . baylyi may thus suggest approaches for enhancing the efficacy of antibiotics targeting the process in its clinically relevant relative . We found that A . baylyi forms giant cells in response to a variety of deletion mutations and antibiotics blocking peptidoglycan synthesis . We exploited the powerful genetic analysis possible in A . baylyi to characterize the requirements for formation of giant cells and to formulate a pathway for their creation . We sought to characterize the cellular consequences of disrupting conserved essential processes like peptidoglycan synthesis . To do this , we generated essential gene deletions by natural transformation of A . baylyi and examined the cells that resulted . Bacteria were exposed to PCR DNA fragments that replace targeted genes with a kanamycin resistance determinant , followed by plating on agar containing kanamycin ( Fig 2 ) ( Materials and methods ) . Transformed cells incorporate the mutagenic DNA , deleting the corresponding essential gene , and then grow and divide , depleting the essential product . Proliferation stops when cells run out of the targeted essential gene product . The resulting microcolonies are made up of dead cells whose morphologies reflect loss of the targeted essential product , and whose size reflects how rapidly depletion of the product blocks growth . Typically 5–10% of the cells are transformed to generate deletions in such experiments , and kanamycin-sensitive untransformed cells are readily distinguished because they stop dividing largely as singlets and doublets with vegetative cell morphology ( see below ) . The principle unwanted background event in the generation of essential gene deletions ( occurring at ~10−6 frequency ) appeared to be due to transformation of partially diploid cells ( presumably due to tandem duplications ) , which generated fast growing cells carrying copies of both deletion and wild type alleles of targeted genes ( Materials and methods ) [23–26] . In a large-scale screen of microcolonies of cells carrying deletion mutations affecting different essential processes , the most dramatic phenotypes resulted from disruption of peptidoglycan synthesis ( Fig 1 ) . For example , when the gene encoding the first step of lipid-linked disaccharide pentapeptide precursor synthesis ( murA ) was deleted by transformation with a ΔmurA::kan PCR fragment , microcolonies of polymorphous giant cells formed ( Fig 3 ) . The giant cells were stabilized by high osmolarity medium and typically enlarged for ~12–24 hours . Giant cells also formed when wild-type bacteria were treated with fosfomycin , an antibiotic that targets MurA ( see below ) . Giant cells could reach diameters greater than ten times that of vegetative cells ( see below ) , and often contained one or more vacuoles at their peripheries ( Fig 3 , 12 and 24 h ) . The vacuoles failed to fluoresce in cells expressing cytoplasmic green fluorescent protein ( S1 Fig ) , indicating that they are shielded from the cytoplasm , e . g . , as if they were derived from the periplasm . Developing giant cells were sometimes joined to each other by membranous bridges ( Fig 3 , 8 h bottom panel ) or exhibited wispy filaments with vesicles extending from their surfaces ( S1 Fig ) . The giant cells did not proliferate , and thus are distinct from L forms . Time-lapse imaging of giant cell formation following murA deletion shows a process in which cells transform from rods into amorphous amoeboid cells that enlarge and eventually burst ( S1 Movie ) . The development of the giant cells proceeds by enlargement without apparent midcell ( preseptal ) blebbing [27] . Deletions of other genes required for peptidoglycan precursor biosynthesis also produced microcolonies of giant cells ( Fig 4A ) . The microcolonies appeared similar in wild type ( wt ) and in a peptidoglycan elongation–minus triple mutant genetic background ( “ΔE” ) ( see below ) . Mutations in three of the genes ( murG , murJ and mraY ) led to microcolonies similar in size to those produced by ΔmurA mutants . The deletion of the fifth gene ( ispU ) , required for cofactor undecaprenol synthesis , led to larger microcolonies than the others . Since undecaprenol is recycled rather than consumed by peptidoglycan synthesis , it may require more growth to deplete it than the precursor intermediates , leading to the larger mutant microcolonies . Overall , the results indicate that disrupting peptidoglycan precursor synthesis at different steps has a similar consequence , the formation of small microcolonies of giant cells . This phenotypic consistency suggests that intermediates in the peptidoglycan precursor synthetic pathway are not particularly toxic , in contrast , for example , to the lipopolysaccharide biosynthetic pathway [28] . The lipid-linked disaccharide pentapeptide precursor is incorporated into peptidoglycan through the action of transglycosylase and transpeptidase activities , with different machineries responsible for cell elongation and division ( Fig 1 ) [7] . Enzymes making up the elongation complex are nonessential in A . baylyi , although elongation–minus mutants grow as spheres rather than short rods [25] . Septal peptidoglycan synthesis functions are essential in A . baylyi . To evaluate whether blocking disaccharide pentapeptide precursor incorporation into peptidoglycan led to giant cell formation , we created mutants defective in both elongation and septal peptidoglycan synthesis . These strains were created from a parent ( “ΔE”; MAY106 ) carrying deletion mutations eliminating three elongation functions ( PBP2 , RodA and PBP1a ) , combined with different mutations blocking cell division ( Fig 4B ) . In all cases , elimination of elongation and division functions together led to the formation of giant cells , whereas the division mutations alone led to long filaments ( Fig 4B ) . The results show that like eliminating precursor synthesis , disrupting incorporation of the precursors into peptidoglycan produces giant cells . We also examined whether treating cells with antibiotics targeting peptidoglycan synthesis produced giant cells . In agreement with the mutant studies , we found that antibiotics inhibiting precursor synthesis or peptide cross-linking induced giant cells ( Fig 5 ) . The antibiotics examined were fosfomycin , which targets MurA , aztreonam , which targets the division-specific transpeptidase FtsI ( PBP3 ) , and meropenem , which apparently targets both the PbpA ( PBP2 ) and FtsI ( PBP3 ) transpeptidases [29] . As expected , fosfomycin and meropenem treatments induced giant cells in both wild type and in an elongation-minus genetic background , whereas aztreonam induced giant cells only in the elongation-minus background . For all three antibiotic treatments , the giant cells were comparable in size , with median dimensions ~10 times that of untreated cells ( Table 1 ) . Treatments with two additional antibiotics , cycloserine , which targets Ddl ( D-ala-D-ala racemase ) and Alr ( alanine racemase ) , and the ß-lactam mecillinam , which apparently targets multiple cross-linking enzymes in A . baylyi ( unlike in E . coli ) [11] , also led to giant cells . The findings indicate that , as was seen for deletion mutations , antibiotic inhibition of peptidoglycan precursor synthesis or incorporation into the sacculus produces giant cells . A straightforward model for giant cell formation is that after inhibition of peptidoglycan synthesis blocks cell enlargement , the activity of hydrolytic functions ruptures the peptidoglycan sheath , allowing the growing cytoplasm to break out of it and expand [5 , 13] . To identify functions potentially involved in this process , we screened for mutations altering giant cell recovery using saturation-level transposon mutant sequencing ( Tn-seq ) . We assumed that the representation of mutations that either blocked or accelerated giant cell formation and subsequent lysis would be changed compared to growth without giant cell induction . We carried out the screens after inducing giant cell formation by fosfomycin or aztreonam treatment ( Materials and methods ) . For the fosfomycin treatment screens , we created a genome saturation-level mutant pool in wild type by transposon-transposase complex electroporation mutagenesis [30] . The pool was exposed to fosfomycin on protective medium to induce giant cells , and DNA isolated from the cells after 24 h growth was subjected to Tn-seq ( Materials and methods ) . Mutations in 35 genes reduced recovery and in 56 genes increased recovery in the presence of fosfomycin compared to no antibiotic ( S1 Database ) . A second set of Tn-seq screens employed a ΔpbpA ( PBP2 ) mutant pool created by natural transformation of the wild-type pool used for the fosfomycin screen by ΔpbpA::kan ( Materials and methods ) . The ΔpbpA mutant pool was exposed to aztreonam on protective medium , and DNA isolated at two different times for Tn-seq . In these screens , mutations in 54 genes reduced recovery and in 27 genes increased recovery relative to no treatment ( S1 Database ) . Among the mutations depleted in one or both screens were those inactivating peptidoglycan penicillin binding proteins , recycling functions , and other proteins involved in peptidoglycan metabolism . Two caveats in interpreting the phenotypes of transposon insertion mutants are that strains may carry unlinked mutations , and that insertions may have polar effects on downstream genes in operons . Unlinked mutations responsible for phenotypes are unlikely in our analysis because the phenotypes identified in the Tn-seq screens are seen for multiple independent insertions ( >35/gene on average ) . In most cases , polar effects are also unlikely to account for mutant phenotypes , because the saturation-level genome coverage Tn-seq provides includes all of the ( nonessential ) genes in an operon . Thus , a polar effect would be seen as a downstream gene ( as well as the upstream gene ) having a mutant phenotype . Although there were four such genes in our top set , validation studies for three of them with constructed mutants designed to be nonpolar indicate that their phenotypes were not due to polarity ( next section ) . Since Tn-seq assays involve cells grown in competition , weak growth phenotypes can lead to significant mutant representation changes . To distinguish the subset of genes with strong mutant phenotypes , we carried out validation experiments with individual mutants . We constructed and analyzed 38 deletion mutants corresponding to genes identified in the Tn-seq screens ( 32/38 ) or considered candidates based on their annotated functions ( 6/38 ) . The mutations were created by replacing the targeted genes with a kanamycin resistance determinant oriented the same as the deleted gene to support transcription of any downstream genes and reduce polar effects [1] ( Materials and methods ) . A total of 29 of the 38 deletions were confirmed to affect giant cell formation ( S1 Table ) , eleven of them leading to particularly strong phenotypes ( Table 2 ) . Mutations in four of the eleven genes blocked giant cell formation at intermediate stages . Deletions of two of the four ( mrcB , encoding transglycosylase-transpeptidase PBP1b , and lpoB , encoding an MrcB regulator ) [31] , had similar phenotypes , leading to much stronger defects in fosfomycin induction than aztreonam induction , with intermediate effects on meropenem induction ( Fig 6 ) . The mutations blocked fosfomycin induction at an early stage at which cells rounded but did not enlarge significantly before lysing . Deletions of the other two genes ( mltD , encoding a membrane lytic transglycosylase , and gcf ( giant cell formation ) , encoding an exported protein of unknown function ) blocked induction by all three antibiotics . However , the phenotypes differed depending on the condition . The blocks in fosfomycin induction were severe , with modest but detectable enlargement before lysis . In contrast , both mutations allowed more enlargement following aztreonam and meropenem induction , producing clusters of “small giants” . Bridges between cells in such clusters were common . Similar clusters were seen as an intermediate in the formation of giant cells after aztreonam treatment of the ΔPBP2 strain . Mutations in four other genes accelerated giant cell formation and lysis . Mutations in three of them were specific to aztreonam induction ( ldtG , encoding a peptidoglycan DAP-DAP ( 3->3 ) cross-linking enzyme; dacA , encoding peptidoglycan carboxypeptidase PBP5; and ponA , encoding transglycosylase-transpeptidase PBP1a ) ( S2 Fig ) . We suspected that the mutations may lead to a weakened peptidoglycan with reduced cross-linking that is more sensitive to further reduction upon aztreonam exposure . Indeed , all three mutations reduced the aztreonam minimal inhibitory growth concentration ( MIC ) ( S2 Fig legend ) . Mutations in a fourth gene ( nagZ , encoding a peptidoglycan recycling function ) led to a complex phenotype . They affected fosfomycin induction more strongly than aztreonam induction , with some early giant cell formation and rampant lysis ( S3 Fig ) . Other recycling pathway mutants also reduced giant cell recovery in Tn-seq ( S1 Database ) . The recycling pathway provides a precursor that bypasses the fosfomycin-inhibited step of the de novo pathway ( MurA ) . As thus expected , the nagZ mutation reduced the fosfomycin MIC ( S3 Fig legend ) . Most mutations increasing Tn-seq recovery moderately to severely reduced vegetative growth rate and may simply delay giant cell lysis by slowing down progression through the pathway ( S1 Database ) . Three examples of such genes with strong mutant phenotypes were ettA , encoding a regulator of translation , ACIAD1396 , encoding a histidine triad protein of unknown function , and zapE , encoding a division ATPase ( Table 2 ) . We did not identify any mutations allowing giant cells to propagate as L forms . In the course of these validation experiments , we identified an unusual phenotype associated with loss of zipA , a gene encoding a division protein that is essential in E . coli but not in Acinetobacter species [32] . A ΔzipA mutation alone caused A . baylyi cells to propagate as elongated rods , some very long ( S4 Fig ) . When a ΔzipA mutation was combined with a mutation blocking elongation peptidoglycan synthesis ( ΔPBP2 ) , the double mutant bacteria propagated as mixed colonies of enlarged spherical cells and giant cells ( S4 Fig ) . We suspect that the zipA mutation hobbles septal peptidoglycan synthesis such that ΔzipA ΔPBP2 double mutant cells divide less frequently than ΔPBP2 mutant alone , leading to larger spheres , and that occasional outright division failure in the double mutant leads to the production of the giant cells . Acinetobacter species produce an outer membrane lipooligosaccharide ( LOS ) that corresponds to lipopolysaccharide lacking an O side chain [33] . Since the outer membrane contributes significantly to envelope stability [34] , we sought to examine whether lipooligosaccharide ( LOS ) is needed for giant cell formation . Although LOS is nonessential in Acinetobacter species [33] , its absence slows growth and the corresponding mutants were poorly represented in the transposon mutant pools we used for Tn-seq analysis . To test the requirement for LOS in A . baylyi giant cell formation , we therefore generated deletions of genes required for LOS precursor synthesis ( lpxA ) or transport to the outer membrane ( lptAB ) . Both classes of mutations affected giant cell formation . The LpxA mutant , which grew as enlarged spheres without antibiotic , formed giant cells after fosfomycin treatment that frequently lysed earlier than usual ( S5 Fig ) . The LptAB deletion mutant grew as smaller spheres and had a more dramatic defect , with massive lysis under giant cell induction conditions . Recent studies have found that LOS transporter mutations cause toxic precursors to accumulate in cells [28 , 35] , and we suspect the ΔlptAB phenotype may be accentuated by such toxicity . Overall , the studies thus indicate that the loss of LOS does not block giant cell formation , but that the cells formed are more fragile than those of wild type . If the antibiotics that induce giant cell formation act by inhibiting their established targets ( MurA for fosfomycin and FtsI ( PBP3 ) for aztreonam ) , inducing giant cells by deleting the target genes should show the same genetic dependencies as those seen for antibiotic induction . To test this prediction , we examined whether three of the mutations blocking giant cell formation after antibiotic treatment also blocked it after deletion of the target genes . We saw congruent effects in all three cases ( S6 Fig ) , a result supporting the conclusion that the two antibiotics induce giant cells by inhibition of their established targets . The mutations blocking fosfomycin-induced giant cell formation appeared microscopically to cause early , wholesale lysis ( Fig 6 ) . In order to quantify this lysis and death , we exposed bacteria grown in liquid protective medium to fosfomycin and followed recovery of viable bacteria ( colony forming units ) ( Fig 7 ) . Under these conditions , the wild type and a control mutant ( carrying a neutral kanamycin resistance marker ) exhibited good recovery for eight hours , whereas five mutants defective in forming giant cells all showed >100-fold reductions in recovery . The mutants and wild type showed comparable growth and survival in the absence of fosfomycin . The results indicate that unimpaired giant cell formation helps protect cells from rapid death when peptidoglycan precursor synthesis is blocked , i . e . , contributes to fosfomycin tolerance . Is the response of A . baylyi to inhibition of peptidoglycan synthesis seen for other Acinetobacter species ? A previous study by Doerr et al . found that meropenem treatment of a clinical isolate of the nosocomial pathogen A . baumannii converted the bacteria into non-dividing spheres that resemble small giant cells [16] . We examined giant cell formation by A . baumannii strain AB5075 , and found that like A . baylyi , the strain formed giant cells upon exposure to fosfomycin or meropenem ( Fig 8 ) . In addition , transposon insertion mutants in mrcB ( ABUW_1358 ) , mltD ( ABUW_2840 ) and gcf ( ABUW_3408 ) interfered with the process in a manner similar to that seen for A . baylyi , albeit more weakly for fosfomycin induction ( Fig 8 ) . The finding that A . baumannii forms giant cells upon inhibition of peptidoglycan synthesis with some of the same genetic dependencies as A . baylyi suggests that the processes are similar in the two species . Our results make it possible to formulate a genetic pathway for giant cell formation in A . baylyi ( Fig 9 ) . The pathway is initiated by blocking either the de novo biosynthesis of peptidoglycan precursor ( upper left ) or incorporation of the precursor into the existing murein sacculus ( lower left ) , with convergence of the two initiating branches . The biosynthesis block can be achieved by null mutations in genes encoding biosynthetic enzymes or by an antibiotic that targets one of the enzymes ( fosfomycin , inhibiting MurA ) . The precursor incorporation block can be achieved in the absence of elongation synthesis by null mutations in division genes or treatment with an antibiotic targeting septal peptidoglycan synthesis ( aztreonam , inhibiting FtsI ) . Other mutations accelerate giant cell formation but do not induce it . Inactivating a peptidoglycan recycling gene ( nagZ ) accelerates their formation and lysis after fosfomycin treatment , presumably because recycling provides an intermediate later in the precursor biosynthetic pathway than the product of the MurA step [36 , 37] . Mutations inactivating three genes ( ldtG , dacA and ponA ) accelerate giant cell formation following inhibition of precursor incorporation ( aztreonam treatment of an elongation-minus mutant ) . LdtG shows homology to enzymes ( L , D-transpeptidases ) that produce an unusual class of peptide cross-links ( DAP-DAP ) in peptidoglycan [38 , 39] , and reducing their level could accelerate giant cell formation by making cells more sensitive to loss of the more abundant cross-links ( DAP-D-ala ) . DacA is a carboxypeptidase that is needed in E . coli to provide the substrate for L , D-transpeptidase [40] , and may play an analogous role in A . baylyi . In E . coli , the transglycosylase activity of MrcB ( PBP1B ) is also required for L , D-transpeptidase-dependent growth , and the ponA ( PBP1A ) transglycosylase might play an analogous role in A . baylyi . Mutations in four genes block giant cell formation following initiation . Two of the genes are relatively specific for the precursor synthesis branch ( mrcB and lpoB ) , with mutants rounding up and lysing without enlarging significantly following initiation . The mrcB gene encodes a transglycosylase-transpeptidase ( PBP1b ) , whereas the lpoB gene encodes an outer membrane protein that activates PBP1b [31] . In E . coli , PBP1b is needed for the generation of L forms [15 , 41] and the conversion of spheroplasts to vegetative cells [42] , indicating that the protein stabilizes peptidoglycan-deficient cells . The protein may thus also stabilize early intermediates in A . baylyi giant cell formation when precursor synthesis is blocked . Mutations in two other genes alter giant cell induction by both initiation branches of the pathway ( mltD , gcf ) . However , the blocks are different for the two modes of initiation . When precursor synthesis is blocked , the mutants enlarge somewhat and then lyse . When precursor incorporation is blocked , they form small giant cells that also have a tendency to lyse . The mltD gene encodes a membrane lytic transglycosylase predicted to hydrolyze the peptidoglycan glycan backbone . MltD activity may contribute to giant cell formation by allowing the growing cytoplasm to emerge from the constraining murein sacculus when new peptidoglycan synthesis is blocked . We think it is unlikely that MltD carries out a function analogous to Slt in E . coli of selective elimination of uncross-linked precursors [11] because it is needed for giant cell induction when precursor synthesis is blocked . Gcf is a protein of unknown function , but is predicted to be an exported protein with a Sel1 tetratricopeptide protein interaction module [43] . Perhaps Gcf interacts with and activates MltD . A Δgcf ΔmltD double mutant exhibits a giant cell induction defect no greater than that of a ΔmltD single mutant , supporting this possibility . It remains to be determined how well the Acinetobacter pathway for giant cell formation represents the generation of wall-deficient forms of other bacteria . However , an intriguing potential link to the recovery of peptidoglycan-deficient V . cholerae spheroids is provided by the observation that a lytic transglycosylase ( MltA ) is required for the process [20] . What is the relationship between giant cells and L forms ? Both types of cells result from inhibition of peptidoglycan synthesis and are pleomorphic , but L forms proliferate and giant cells do not . We hypothesize that giant cells represent a primary consequence of growth without peptidoglycan synthesis , and that additional mutations are required for them to proliferate as L forms . This model readily accounts for overlap in functions needed for production of the two types of cells ( e . g . , mrcB ) and the low yield of L forms when peptidoglycan synthesis is inhibited ( e . g . , ~10−5–10−4/cell ) [41] . Natural variation in the capacity of different bacteria to generate L forms may reflect differences in other factors needed for L form growth , e . g . , the nature and amount of polysaccharide capsule [41 , 44] . In this study we used a new procedure for examining the terminal phenotypes of bacteria deleted of essential genes to analyze mutations disrupting peptidoglycan synthesis . Mutations blocking the process in different ways led to the formation of pleomorphic giant cells , and the phenotypes of mutants defective in making the unusual cells suggested a genetic pathway for their formation . Mutant strains were derivatives of Acinetobacter baylyi ADP1 ( MAY101 ) [3] ( the gift of C . Harwood ) and A . baumannii AB5075-UW [45] . A . baylyi MAY106 ( “ΔE” ) ( ΔpbpA ΔrodA ΔponA ) is an unmarked peptidoglycan elongation-deficient triple mutant constructed from ADP1 . A . baylyi MAY116 is a “wild type” control strain carrying a kanamycin resistance marker ( nptII ) in place of an IS element ( IS1236_1 ) [3 , 24] . MAY103 carries a ΔpbpA allele marked with nptII . A complete list of strains is provided in S2 Database . Growth media were TYE ( 10 g tryptone , 5 g yeast extract , 8 g sodium chloride and 15 g agar per liter ) , LB ( TYE lacking agar ) and minimal-succinate ( M9 medium [46] supplemented with 15 mM sodium succinate , 2 mM magnesium sulfate , 0 . 1 mM calcium chloride and 1–3 μM ferrous sulfate ±15 g agar/l . Protective medium was minimal-succinate supplemented with 0 . 4 M sucrose and 10 mM magnesium sulfate , and was solidified with 1 . 5% agar ( “protective agar” ) or 2% agarose ( “protective agarose” ) . Supplements were routinely used at the following concentrations: kanamycin , 10 μg/mL ( TYE ) or 20μg/mL ( minimal media ) ; fosfomycin , 192–360 μg/mL; aztreonam , 120–190 μg/mL; and meropenem , 5–25 μg/mL . A . baylyi strains were routinely grown at 30° C whereas A . baumannii strains were grown at 37° C . We created deletion mutations by natural transformation of linear DNA fragments constructed by PCR using extension overlap [22 , 47 , 48] . The transformed fragments carried sequences of homology of ~2 kb flanking targeted genes that were either directly joined ( for unmarked deletions ) or flanked a kanamycin resistance determinant ( for kan-marked deletions ) . Unmarked deletions were in-frame and included 18 bp insertions carrying diagnostic restriction sites at the deletion junctions . Marked deletions carried the nptII gene ( “kan” ) from plasmid pACYC177 [49] encoding kanamycin resistance , with nptII in the same orientation as the deleted gene . In creating deletions , we designed primers ( and often employed the same primers ) as described previously [25] ( S2 Database ) . Thermocycling reactions employed Q5 Polymerase ( New England Biolabs ) , and DNA fragments were routinely purified using Qiaquick columns ( Qiagen ) prior to transformation . The mutagenic DNA fragments were transformed into A . baylyi cultures grown overnight in minimal-succinate with 1uM ferrous sulfate , diluted 1:5 in fresh medium and grown one hour with shaking at 30° . DNA was added at 1μg/mL , followed by incubation for 3 hours with shaking and plating on selective ( marked deletions ) or non-selective ( unmarked deletions ) media . Unmarked deletion mutations were identified by screening individual colonies by PCR using primers flanking targeted genes . Essential gene kan-marked deletion mutations were selected by plating on protective medium containing kanamycin ( 20 μg/mL ) . All unmarked and the majority of marked deletion mutations were verified by PCR . Microcolonies of the deletion mutants were generally imaged after 18–24 hours incubation at 30° C . In a typical experiment creating kan-marked essential gene deletions , 5–10% of the cells were transformed forming microcolonies of cells carrying the deletion . There was commonly a background of ~10−6 fast-growing colonies that carried both deleted and undeleted versions of the targeted genes [25] , presumably arising from cells with tandem duplications . Bacterial microcolonies were routinely imaged after growth on protective agar in 15 X 60 mm diameter Petri plates under bright field illumination using a Nikon Eclipse 90i with an ELWD 20X objective equipped with 2X digital zoom . For high-resolution imaging , microcolonies were grown on thin 2% agarose protective medium pads under cover slips in Gene Frames ( Thermo Scientific ) prior to phase contrast imaging using a 100X oil immersion objective . The microcolonies of giant cells grown in Gene Frames tended to be somewhat smaller and lyse somewhat earlier than those formed on plates . ADP1 was mutagenized by insertion of the tetracycline resistance-marked transposon T26 using a modification of a previously described procedure [45] . An overnight LB culture of ADP1 was diluted 1:200 into fresh medium without NaCl and grown with shaking to OD600 0 . 8 . Cells were then pelleted and washed three times in decreasing volumes of cold 10% glycerol until cells had been concentrated approximately 150-fold . Aliquots ( ~0 . 5 μL ) of transposon-transposase complex were mixed with 50 μl concentrated cells for electroporation ( 1 . 8kV , 200 Ω , 25μF using a Biorad Gene Pulser ) . Insertion mutants were selected on TYE media supplemented with tetracycline ( 5–7 . 5 μg/mL ) by overnight growth at 30° , and then harvested and pooled . Two independent pools were created , each made up of approximately 80 , 000 independent mutants . A pbpA-minus transposon mutant pool was created by transforming one of the ADP1 mutant pools with a PCR fragment corresponding to the ΔpbpA::kan mutation found in MAY103 , with selection for kanamycin resistance on minimal-succinate agar . Tn-seq screens were carried out for cells grown on fosfomycin or aztreonam . For the fosfomycin screen , one of the ADP1 transposon mutant pools was plated on protective medium supplemented with 360 μg/mL fosfomycin at approximately 5x107 and 5x108 cells/plate , grown for 24 hours at 30°C and harvested ( two Tn-seq assays total ) . For the aztreonam Tn-seq screen , the ΔpbpA::kan transposon mutant pool was plated on protective medium supplemented with aztreonam ( 120 or 192 μg/mL ) at ~5X105 and 1X107 cells/plate , and cells were harvested at 24 and 48 hours ( eight Tn-seq assays total ) . As controls , mutant pools were grown on protective medium lacking antibiotic . Tn-seq analysis was carried out using a terminal deoxynucleotidyl terminal transferase-based procedure [45] . To identify genes whose inactivation affected giant cell formation after fosfomycin or aztreonam treatment , combined read counts for insertions in nonessential genes ( 5 to 90% of predicted coding regions , normalized for total reads/sequencing run ) were evaluated for each time point analyzed , and histograms of the ratios of the log2-transformed read counts of antibiotic-treated to the corresponding antibiotic-untreated cultures plotted . Genes whose mutants were significantly depleted or enriched under giant cell induction conditions were identified using normal distribution thresholds specified for each condition in Data Set S1 , with genes identified in multiple independent experiments chosen for subsequent validation studies using constructed deletion mutations .
Although essential genes control the most basic functions of bacterial life , they are difficult to study genetically because mutants lacking the functions die . We have developed a simple procedure for creating bacteria in which different essential genes have been completely deleted , making it possible to analyze the roles of the missing functions based on the features of the dead cells that result . When genes needed for the production of the cell wall were inactivated , the bacteria formed bizarre giant cells . It was possible to identify the functions responsible for forming the giant cells , and to formulate a model for how they form . Since cell wall synthesis is one of the most important antibiotic targets , understanding how bacteria respond to its disruption may ultimately help in developing procedures to overcome antibiotic resistant bacterial infections .
[ "Abstract", "Introduction", "Results", "and", "discussion", "Materials", "and", "methods" ]
[ "antimicrobials", "deletion", "mutation", "medicine", "and", "health", "sciences", "evolutionary", "biology", "cell", "cycle", "and", "cell", "division", "drugs", "cell", "processes", "population", "genetics", "microbiology", "precursor", "cells", "gene", "pool", "physiological", "processes", "mutation", "antibiotics", "materials", "science", "pharmacology", "population", "biology", "macromolecules", "polymers", "polymer", "chemistry", "tissue", "repair", "animal", "cells", "chemistry", "lysis", "(medicine)", "cell", "biology", "peptidoglycans", "physiology", "genetics", "microbial", "control", "biology", "and", "life", "sciences", "cellular", "types", "giant", "cells", "physical", "sciences", "materials" ]
2019
Essential gene deletions producing gigantic bacteria
Homologous recombination is a universal mechanism that allows repair of DNA and provides support for DNA replication . Homologous recombination is therefore a major pathway that suppresses non-homology-mediated genome instability . Here , we report that recovery of impeded replication forks by homologous recombination is error-prone . Using a fork-arrest-based assay in fission yeast , we demonstrate that a single collapsed fork can cause mutations and large-scale genomic changes , including deletions and translocations . Fork-arrest-induced gross chromosomal rearrangements are mediated by inappropriate ectopic recombination events at the site of collapsed forks . Inverted repeats near the site of fork collapse stimulate large-scale genomic changes up to 1 , 500 times over spontaneous events . We also show that the high accuracy of DNA replication during S-phase is impaired by impediments to fork progression , since fork-arrest-induced mutation is due to erroneous DNA synthesis during recovery of replication forks . The mutations caused are small insertions/duplications between short tandem repeats ( micro-homology ) indicative of replication slippage . Our data establish that collapsed forks , but not stalled forks , recovered by homologous recombination are prone to replication slippage . The inaccuracy of DNA synthesis does not rely on PCNA ubiquitination or trans-lesion-synthesis DNA polymerases , and it is not counteracted by mismatch repair . We propose that deletions/insertions , mediated by micro-homology , leading to copy number variations during replication stress may arise by progression of error-prone replication forks restarted by homologous recombination . Maintenance of genome stability requires the faithful and accurate replication of the genetic material . Genome instability is a hallmark for most types of cancer and it is strongly associated with predisposition to cancer in many human syndromes ( for a review , see [1] , [2] ) . Genome instability is manifest at two levels: at the nucleotide level , resulting in base-substitutions , frame-shifts or in micro-insertions/deletions ( referred to herein as mutations ) ; and at the chromosomal level , resulting in duplications , deletions , inversions and translocations ( referred to herein as gross chromosomal rearrangements or GCRs ) . Genome instability during cancer development and in other human genomic disorders may be consequences of failures in chromosome replication ( for a review , see [3] , [4] ) . Origin spacing has recently been shown to cause chromosomal fragility at some human fragile sites [5] , [6] . Impediments to replication fork movements per se may also cause genome instability [7]–[9] . Indeed , both slowing down and blockages to fork progression can lead to chromosomal fragilities or GCRs in human cells and yeast models [10]–[14] . However , how a blocked replication fork leads to genetic instability remains poorly understood . In eukaryotes , DNA replication is initiated at numerous origins along linear chromosomes , and impediments to fork progression appear unavoidable during each S-phase ( for a review , see [9] , [15] ) . Impediments to fork progression can be caused by DNA lesions , by non-histone proteins tightly bound to DNA , by sequence-caused secondary structures such as cruciform structures and possibly G-quadruplexes , by nucleotide pool imbalance and by conflicts with transcription machinery ( for a review , see [16] , [17] ) . In case of failures in fork progression , DNA replication can be completed either by the recovery of the arrested fork by fork-restart mechanisms , or as a result of the progression of a converging fork which can be ensured by activation of dormant origins [7] , [15] , [18] . Fork restart is presumably essential in unidirectional replication regions , such as the rDNA locus , in regions of low densities of origins , such as some human fragile sites , and when two converging forks are both impeded [5] , [19] , [20] . To ensure faithful and complete DNA replication , cells coordinate DNA synthesis restart with specific pathways , including DNA replication checkpoint and homologous recombination mechanisms [17] . The integrity of replication forks is guaranteed by the DNA replication checkpoint that maintains the replisome in a replication-competent state to keep DNA polymerases at the site of nucleotide incorporation [21] . It remains unclear how the DNA replication checkpoint modulates replisome activities to maintain its function [21] , [22] . The DNA replication checkpoint also regulates nuclease activities ( e . g . Exo1 or Mus81 ) which contribute to preserving the integrity of stalled forks [23] , [24] . If replisome function is lost or the replisome dissociates at broken replication forks , the resumption of DNA synthesis appears to require the replisome to be rebuilt . In E . coli , restart of a collapsed fork involves homologous recombination and the PriA helicase that allows replisome components to be loaded de novo on joint-molecule structures [25] , [26] . In eukaryotes , the restart of collapsed or broken replication forks is dependent upon homologous recombination , but the mechanism of origin-independent loading of the replisome remains to be described [20] , [27]–[30] . It has been proposed that the repair of a double-strand break ( DSB ) by recombination ( break-induced replication , BIR ) in budding yeast similarly involves the assembly of a replication fork ( for a review , see [30]–[32] ) . When BIR occurs outside S-phase , recombination-dependent replication fork assembly can synthesise hundreds of kilobases ( Kb ) . However , this DNA synthesis is highly inaccurate due to frequent template switching of nascent-strands and frame-shift mutations [33] , [34] . We previously reported a system that displays replication fork arrest at a specific locus in the fission yeast S . pombe . The system is a polar replication fork barrier ( RFB ) regulated by the Rtf1 protein binding to its RTS1 binding site [35] . The RTS1-RFB causes fork arrest because of a non-histone protein complex binding to the DNA . As proposed for other polar RFBs , the RTS1-RFB is thought to block fork progression by directly ( contact between proteins and the replisome ) or indirectly ( topological constraint ) affecting the replicative helicase activity and thereby preventing DNA unwinding [36] , [37] . Recovery of the arrested fork occurs by a DSB-independent mechanism and involves the recruitment of recombination proteins at the RTS1-RFB site . We proposed that recombination proteins associate with unwound nascent strands that then anneal with the initial template to allow DNA synthesis to restart [11] , [20] . The causative protein barrier then has to be removed either by DNA helicase or by the recombination machinery itself to allow fork-progression to resume [38]–[40] . Occasionally , the unwound nascent strand can mistakenly anneal with a homologous template in the vicinity of the collapsed fork , resulting in the restart of DNA synthesis on non-contiguous template . This incorrect template switch of nascent strands results in inversions and iso-acentric and dicentric chromosomes in ∼2–3% of cells/generation [11] , [20] . Error-free template switching between sister-chromatids provides an efficient mechanism for filling-in single-stranded gaps left behind damage-induced stalled forks [41] . Inverted chromosome fusions in yeast and rare-genome rearrangements in human genomic disorders , may both be consequences of template switching between ectopic repeats associated with impeded replication forks [8] , [14] . Here , we used the RTS1-RFB to investigate the consequences of fork collapse on genome instability . We report that recovery from a collapsed fork is associated with a high frequency of instability , with a single fork arrest increasing the rates of mutation , deletion and translocation by 10 , 40 and 5 fold , respectively . We show that genetic instability associated with fork arrest is dependent on homologous recombination . Fork-arrest-induced GCRs ( deletion and translocation ) result from inappropriate ectopic recombination at the site of the collapsed fork . We also demonstrate that restoration of fork progression by homologous recombination results in error-prone DNA synthesis due to frequent replication slippage between short tandem repeats . We investigated the molecular mechanisms of this replication slippage and found that post-replication repair , including ubiquitination of PCNA or trans-lesions-synthesis ( TLS ) DNA polymerases , is not involved in fork-arrest-induced replication slippage . Micro-deletions/insertions flanked by micro-homology associated with copy number variations ( CNVs ) in cancer cells or in response to replication stress may therefore be scars left following the restoration of forks progression by homologous recombination . We generated fork arrest constructs by manipulating the polar RTS1-RFB ( Figure 1A ) . We introduced the RTS1 sequence on the centromere-proximal ( cen-proximal ) side of the ura4 locus , 5 kb away from the strong replication origin ( ori ) 3006/7 on chromosome III . This created the t-ura4<ori locus , in which “t” and “ori” refer to the telomere and the origin 3006/7 , respectively; and “<” and“ >”refer to the RTS1-barrier and its polarity that is whether it blocks replication forks travelling from the ori 3006/7 towards the telomere or forks travelling from the telomere towards the ori 3006/7 , respectively . We previously confirmed that forks moving from ori 3006/7 towards the telomere ( tel ) are efficiently blocked by the RTS1-RFB at the t-ura4<ori locus [35] . In this model system , fork arrest is activated by inducing the expression of rtf1+ gene that is under control of the thiamine repressible promoter nmt41 . Thus , the RTS1-RFB is inactivated by adding thiamine to the media and it is activated in thiamine-free media . Efficient induction of Rtf1 expression requires incubation for 12–16 hours in thiamine-free media . Replication intermediates were analysed by native 2-dimensional gel electrophoresis ( 2DGE ) . In conditions of Rtf1 expression , more than 95% of replication forks were blocked by the RTS1-RFB at the t-ura4<ori locus ( see black arrow on Figure 1B , t-ura4<ori ON ) . Arrested forks were not detected without Rtf1 induction ( Figure 1B , t-ura4<ori OFF ) [20] . The RTS1 sequence was inserted on the tel-proximal side of ura4 creating the t<ura4-ori locus . 2DGE analysis of this construct revealed a strong fork arrest signal on the descending large Y arc ( Figure 1A and 1B , t<ura4-ori ON ) . The ura4+ gene , used in this system as a reporter to score genetic instability , is located behind the arrested fork when the RTS1-RFB is active at the t<ura4-ori locus and ahead of the arrested fork at the t-ura4<ori locus . This explains the distinct position of the arrested fork signal on the Y arc . Inversion of the RTS1 sequence at the tel-proximal side of ura4 created the t>ura4-ori locus and no fork arrest signal was detected for this construct by 2DGE when Rtf1 was expressed ( Figure 1A and 1B , t>ura4-ori ON ) . Thus , RTS1 behaves as a polar RFB at the ura4 locus , and replication across this locus is strongly unidirectional due to the relative positions of the origins [42] . Introducing a second RTS1 sequence , such that the two RTS1 sequences are inverted repeats ( IRs ) , created t>ura4<ori and t<ura4>ori loci ( Figure 1A and 1B , t>ura4<ori and t<ura4>ori ON ) . Given the orientation of the polar RTS1-RFB in the t<ura4>ori strain , converging forks cannot be blocked . Whereas block of converging forks can virtually occur in the t>ura4<ori strain , 2DGE in this construct revealed that forks arrested on the cen-proximal side of ura4 were efficiently recovered by recombination before forks are arrested on the tel-proximal side . Indeed , joint-molecules ( JMs ) resulting from recombination between RTS1 repeats were detected by 2DGE ( see red arrows on Figure 1B , t>ura4<ori and t<ura4>ori ON ) . Resolution of these JMs gives rise to chromosomal rearrangements [20] . In the absence of homologous recombination ( i . e . in a rad22-d mutant ) , JMs were not detected and termination signals accumulated ( see green arrow on Figure 1B , t>ura4<ori rad22-d strain ) . Similarly , termination signals accumulated in the rad22-d t-ura4<ori strain ( see green arrow on Figure 1B , t-ura4<ori rad22-d ) , showing that , when arrested forks are not restarted by homologous recombination , the RTS1-RFB behaves as a hot spot for replication termination [20] . We investigated fork-arrest-induced genome instability by selecting for cell resistance to 5-FOAR , the result of loss of ura4+ function . Inducing fork-arrest at t-ura4<ori increased ura4 loss 3 fold ( Table 1 ) . Rtf1 expression in the t-ura4-ori and t>ura4-ori strains did not cause site-specific fork-arrest at ura4 as assessed by 2DGE and did not increase the rate of ura4 loss . Thus , ura4 loss results from the RTS1-RFB activity and not simply from the presence of RTS1 and/or Rtf1 expression ( Table 1 ) . To investigate the nature of this genetic instability , primers were designed to amplify the ura4 coding sequence and , as a control , the essential rng3 gene , mapping 30 kb tel-proximal to ura4 , that should not be rearranged ( Figure 2A and 2B ) [35] . The absence of ura4 amplification was classified as a deletion event; sequencing of amplified ura4 sequence was used to identify point mutation events ( Figure 2B ) . A single arrested fork at the t-ura4<ori locus was sufficient to increase the rate of genomic deletion up to 40 times over spontaneous events ( i . e . in the t-ura4-ori strain , p = 0 . 006 ) ( Figure 2C and Figure S1A ) . Fork-arrest-induced deletion was recombination-dependent . Spontaneously ( i . e . when the RTS1-barrier was inactivated ) , the rate of genomic deletion in rad22-d or rhp51-d strains was higher than that in the wild-type strain ( Figure S1B ) . Nonetheless , no further increase in the rate of genomic deletion was observed in the surviving rad22-d or rhp51-d cells upon activation of the RTS1-barrier ( Figure S1B , t-ura4<ori ) . Frequent spontaneous genomic deletion in the rad22-d or rhp51-d strains is consistent with previous reports showing that mutations in recombination genes are associated with an increase level of GCRs [14] , [43] , [44] . Deleting the natural RTS1 sequence from chromosome II abolished deletion events at collapsed forks , indicating that fork-arrest-induced deletion was also mediated by inter-chromosomal recombination ( Figure 2C and t-ura4<ori RTS1-d on Figure S1A ) . Thus , these data are consistent with the view that homologous recombination makes a major contribution to suppressing genome instability , but can occasionally drive non allelic recombination events leading to GCRs [35] , [45] . We detected no fork-arrest-induced deletion in the t<ura4-ori strain , in contrast to the t-ura4<ori strain ( Figure S1A and Figure 2C ) . The ura4 marker is located behind and ahead of collapsed forks in the t<ura4-ori and t-ura4<ori strains , respectively ( Figure 1A ) . Therefore , replicated regions , located behind collapsed forks , do not display instability , and fork-arrest-induced deletion occurs within unreplicated regions immediately in front of arrested forks . Overall , our data establish that genomic deletion at collapsed forks results from inappropriate recombination between ectopic sequences during the process of fork recovery by recombination proteins . Inverted repeats ( IRs ) are structural elements often associated with genome rearrangements [11] , [14] , [46] , [47] . We investigated the effects of IRs in the vicinity of the RTS1-RFB on fork-arrest-induced genomic deletion . We first compared the t>ura4<ori strain ( IRs flanking ura4 ) to the t-ura4<ori strain ( no IRs near the RTS1-RFB ) . The rate of fork-arrest-induced genomic deletion was 200 times higher in the t>ura4<ori than that in the t-ura4<ori strain ( p = 0 . 009 , Figure 2C and Figure S1A ) . Thus , intra-chromosomal ectopic recombination permitted by the RTS1 sequence on the tel-proximal side of ura4 accounted for 99 . 5% of the genomic deletions observed in the t>ura4<ori strain ( Figure 2C , compare with t-ura4>ori ) . Preventing inter-chromosomal recombination by deleting RTS1 from the chromosome II ( t>ura4<ori RTS1-d ) abolished 90% of deletion events ( Figure 2C and Figure S1A ) . Thus , genomic deletions induced by fork-arrest near IRs are due to inter- and intra-chromosomal recombination events . In support of this , stimulation of fork-arrest-induced deletion by IRs is mediated by homologous recombination . Indeed , the rate of genomic deletion was not increased upon induction of the RTS1-RFB in the surviving population of t>ura4<ori rad22-d and rhp51-d strains ( Figure S1B ) . These data indicate that IRs favour genomic deletion at collapsed forks by promoting inappropriate inter- and intra-chromosomal recombination during fork recovery by recombination proteins . We verified that our data were not influenced by the orientation of IRs or by rare blocking of converging forks in the t>ura4<ori strain . We analysed the t<ura4>ori construct in which RTS1 repeats are in the opposite orientations relative to the t>ura4<ori construct , such that forks converging towards ura4 cannot be blocked ( Figure 1 ) . The rate of fork-arrest-induced genomic deletion was 1 , 000 times higher in the t<ura4>ori than that in the t<ura4-ori strain , that does not contain IRs near the RTS1-RFB ( p = 0 . 008 , Figure 2C and Figure S1A ) . Thus , intra-chromosomal recombination , permitted by the RTS1-RFB sequence on the cen-proximal side of ura4 , accounted for nearly 100% of the genomic deletions observed in the t<ura4>ori strain ( Figure 2C , compare with t<ura4-ori ) . Preventing inter-chromosomal recombination by deleting RTS1 from the chromosome II ( t<ura4>ori RTS1-d ) abolished 90% of deletion events ( Figure 2C and Figure S1A ) . Importantly , the deletion rates for the t<ura4>ori and t>ura4<ori strains were not significantly different ( Figure 2C ) , showing that IRs cause genomic deletion at collapsed forks irrespective of their orientations and independently of blockage of converging forks . Fork-arrest at t>ura4<ori results in translocations between ectopic RTS1 repeats on chromosomes II and III [35] . We investigated the influence of IRs on fork-arrest-induced translocation . We designed primers to amplify the predicted translocation junction between chromosomes II and III ( TLII and TLIII on Figure 2A and 2B ) . A single arrested fork at the t-ura4<ori locus was sufficient to increase the translocation rate to 5 times higher than the spontaneous rate ( p = 0 . 002 , Figure 2D and Figure S1C ) . The translocation rate for the t>ura4<ori construct ( containing IRs ) was 1 , 500 fold higher than that for the t-ura4<ori strain that does not contain IRs near the RTS1-RFB ( p = 0 . 009 , Figure 2D and Figure S1C ) . Thus , intra-chromosomal recombination accounted for nearly 99% of translocation events observed in the t>ura4<ori construct ( Figure 2D and Figure S1C , compare with t-ura4<ori ) . No translocation events were detected when inter-chromosomal recombination was prevented by deleting RTS1 from the chromosome II ( t>ura4<ori RTS1-d on Figure 2B ) . Therefore , as reported for genomic deletions , fork-arrest-induced translocation associated with IRs is due to inter- and intra-chromosomal ectopic recombination . No translocations were detected in the t<ura4>ori strain ( data not shown ) , so we cannot formally exclude the possibility that fork-arrest-induced translocations in the t>ura4<ori strain was caused by blocking of converging forks . However , as no translocation event occurred in the absence of Rad22Rad52 or Rhp51Rad51 , it is most likely that translocations occur during fork recovery by recombination ( Figure 2B and [35] ) . Overall , our data indicate that recovery of a single collapsed fork causes translocations and IRs near the site of fork-arrest stimulate translocations by promoting inappropriate inter- and intra-chromosomal recombination . Fork-arrest-induced GCRs are caused by inter- and intra-chromosomal recombination . We noticed a slightly greater contribution of intra- than inter-chromosomal recombination ( Figure 2C ) . This is consistent with ectopic recombination preferentially occurring at the most proximal homologous sequence , as previously reported [48] . Nonetheless , the rate of fork-arrest-induced deletion in the t>ura4<ori strain ( 8 . 4 10−7 ) was not simply the sum of the rates of intra-chromosomal recombination events ( 9 . 9 10−8 in the t>ura4<ori RTS1-d strain ) and inter-chromosomal recombination events ( 4 10−9 in the t-ura4>ori strain ) . Similar reasoning can be applied for the t<ura4>ori strain . Thus , independent intra- and inter-recombination events cannot themselves explain high rate of GCRs induced by fork arrest near IRs . Therefore , we infer that there is interplay between inter- and intra-chromosomal recombination such that fork-arrest-induced GCRs may involve recombination between three homologous sequences ( tri-parental recombination ) . To confirm that fork-arrest-induced GCRs are the result of inappropriate ectopic recombination during fork recovery , we analysed the involvement of the RecQ helicase Rqh1 . We previously reported that Rqh1 limits inappropriate template switching of stalled nascent strands without affecting the efficiency of fork restart [20] . In the t-ura4<ori construct ( in which only inter-chromosomal recombination is possible ) , fork-arrest-induced deletion and translocation rates were 31 and 109 times higher in the rqh1-d strain than that in the wild-type control , respectively ( p = 0 . 0003 , Figure 2D–2E and Figure S1C ) . For the t>ura4<ori construct ( containing IRs near fork-arrest ) , fork-arrest-induced deletion and translocation rates were 5 times higher in the rqh1-d than that in the wild-type control ( p = 0 . 0007 , Figure 2D–2E and Figure S1C ) . Thus , Rqh1 limits GCRs at collapsed forks by preventing inappropriate ectopic recombination during the process of fork recovery by recombination proteins . We analysed the effects of collapsed forks on the mutation rate . We sequenced the ura4 coding sequence from 5-FOAR isolated cells and identified base-substitutions , frame-shifts and small insertions and duplications between short tandem repeats ( Table 2 ) . A single collapsed fork in the t-ura4<ori strain increased the overall mutation rate up to 10 times over spontaneous events ( Figure 3A , p = 0 . 003 ) . Similar increases in the overall mutation rate were found for the strains with IRs near the arrested fork and those with RTS1 deleted from chromosome II ( Figure 3A and Figure S2A ) . Thus , fork-arrest-induced mutation is not mediated by inappropriate ectopic recombination . Induction of the RTS1-RFB in the t<ura4-ori strain did not increase the mutation rate of the ura4 gene . Thus , as for GCRs , replicated regions behind arrested forks are not prone to mutation . This observation rules out the hypothesis that fork-arrest-induced mutation is a consequence of the accumulation of damaged single-stranded DNA behind collapsed forks ( see discussion ) . Our data suggest that recovery from collapsed forks results in error-prone DNA-synthesis . We then analysed the spectra of mutations found in the ura4 ORF by sequencing the PCR products . The rates of base-substitutions and frame-shifts were not significantly increased by the RTS1-RFB activity over spontaneous events ( i . e . compare to t-ura4-ori strain , Figure 3C and Table 2 ) . In contrast , the rate of deletions and duplications ( Del/Dup ) flanked by short homology was increased by 7 times over spontaneous events in the t-ura4<ori strain , but not in the t<ura4-ori strain ( Figure 3C and Table 1 ) . These data further confirm that fork-arrest does not promote mutation events behind collapsed forks . We used reversed mutation assays to test if fork-arrest at the RTS1-RFB specifically induced Del/Dup mutations . We made use of strains harbouring a single mutation within the ura4 ORF: either a single base-substitution or a −1 frame-shift in homo-nucleotide ( Figure S2B ) . We also studied strains harbouring either a duplication of 20 or 22 nt flanked by 5 or 4 bp of micro-homology , respectively ( defined as ura4-dup20 and ura4-dup22 , Figure S2B ) . These non-functional ura4− alleles were inserted in front of the RTS1-RFB in the t-ura4<ori configuration and we then tested whether fork arrest could restore a functional ura4+ gene . Activation of the RTS1-RFB at ura4 increased the frequency of Ura+ revertants up to 15 and 7 times in strains harbouring ura4-dup22 and ura4-dup20 , respectively ( Figure 3D and Figure S2B ) . Thirty Ura+ colonies were studied by PCR and all gave a product of the same size as the wild-type ura4+ gene: they had therefore lost the duplication ( Figure 3E and data not shown ) . Sequencing the full ura4 ORF confirmed that Ura+ revertants contained an intact ura4+ sequence , showing that the reversion of these alleles was due solely to the precise deletion of 20 or 22 nt ( Figure 3F and data not shown ) . In contrast , activation of the RTS1-RFB did not increase the frequency of Ura+ revertants of strains harbouring ura4 alleles with a single base-substitution or a −1 frame-shift ( Figure 3D and Figure S2B ) . Thus , collapsed forks tend to induce deletion events between short tandem repeats rather than base-substitution or frame-shift mutations . Among Del/Dup events , deletions represented the two-third of events in the t-ura4<ori strain ( Table 2 ) . The median size of Del/Dup events was 24 and 22 nt respectively , and Del/Dup occurred between short direct repeats 1 to 10 nt long ( Figure S3 ) . Thus , the ura4-dup20 and ura4-dup22 alleles used in the reverse mutation assay were representative of the Del/Dup events observed . Del/Dup flanked by micro-homology result from intra-molecular template switching mechanisms in which nascent strands dissociate from the template and misalign with the template when restarting the elongation step . This leads to loop formation , either in the nascent strand or in the template , resulting in duplication or deletion events , respectively [49] . Consequently , we will hereafter refer to Del/Dup as replication slippage . Replication slippage was observed all along the ura4 ORF and up to 1 . 2 kb ahead of the arrested fork , even if a hot spot of deletion was present 500 bp away from the RTS1-RFB ( Figure 3G and Figure S3B ) . Thus , our data suggest that the DNA synthesis is prone to replication slippage at least for the first 1 , 200 nt synthetized during the recovery of collapsed forks . Inaccuracy of DNA synthesis on further distances was not directly addressed . To confirm that replication slippage occurs as forks recover , and not behind the fork in the DNA already replicated , we inserted the ura4-dup20 or the ura4-dup22 allele either behind ( in the t<ura4-ori configuration ) or in front of the RTS1-RFB ( in the t-ura4<ori configuration ) ( Figure 4 ) . This allows the analysis of the same event of replication slippage behind and ahead of collapsed forks . In the t-ura4<ori configuration , induction of the RTS1-RFB resulted in a 8 and 16 fold increases in the replication slippage frequency for the ura4-dup20 and ura4-dup22 alleles , respectively ( Figure 4A and 4B ) . Similar increases in the rate of replication slippages were observed ( Figure 4C ) . In contrast , in the t<ura4-ori background , the frequency of replication slippage was induced by only 2–3 fold by the RTS1-RFB ( Figure 4B–4C ) . These data confirm that DNA located ahead of collapsed forks is more prone to replication slippage than replicated DNA adjacent to arrested forks , further evidence that replication slippage arises during fork recovery . Replication slippage occurs in DNA in front of ( and not behind ) the arrested fork , this DNA being replicated only after restart of the fork . Thus , a defect preventing fork recovery would be expected to abolish the error-prone DNA synthesis during restart . We analyzed fork-arrest-induced mutation in recombination mutants in which collapsed forks at the RTS1-RFB cannot recover , resulting in cell death . Induction of the RTS1-RFB did not increase the overall mutation rate in the surviving populations of t>ura4<ori or t-ura4<ori rad22-d and rhp51-d strains ( Figure 3B ) . In addition , only 7% of mutation events in the survivors of the rad22-d t-ura4<ori strain were Del/Dup mutations , compared to 40% in the wild-type strain ( Figure 3C and Table 1 ) . We currently cannot assess mutation events associated with defects in fork recovery because this appears to be lethal in the absence of recombination . Nevertheless , our data are consistent with fork-arrest-induced replication slippage being dependent on homologous recombination . The rad22-d and rhp51-d strains are themselves spontaneously mutagenic . Consequently , any small increase in the fork-arrest-induced mutation rate might be masked by the high frequency of spontaneous 5-FOAR cells in rad22-d and rhp51-d strains . We therefore used a more specific mutation assay , based on the ura4-dup20 allele , to determine the rate of replication slippage induced by the RTS1-RFB over spontaneous events . Strains carrying mutations in recombination genes grow slowly , so replication slippage was scored as a function of the number of generations following thiamine removal ( i . e . generations subject to fork arrest at ura4 ) ( Figure 4D and 4E ) . In the wild-type strain , fork arrest at the RTS1-RFB resulted in a 10 fold-increase in the frequency of replication slippage , as expected . In recombination mutants ( rad50-d , rhp51-d and rad22-d ) , fork-arrest at the RTS1-RFB increased the frequency of replication slippage by only 2 times over spontaneous events: therefore , replication slippage occurs less frequently in survivors from recombination mutants than those from the wild-type strain ( Figure 4D–4F ) . Based on 2DGE analysis , fork-restart is severely impaired in the absence of Rad22Rad52 ( Figure 1B and [20] ) , such that even the two-fold induction in replication slippage by fork arrest in the rad22-d strain was surprising . The rad22-d strain accumulates suppressors involving the Fbh1 helicase that limits Rhp51Rad51- dependent recombination at blocked replication forks [50] , [51] . Therefore , we analyzed replication slippage in the rad22-d rhp51-d double mutant in which no homologous recombination event occurs . In this background , there was no detectable fork-arrest-induced replication slippage ( Figure 4D–4F ) . Thus , complete defect in fork restart results in a complete abolition of fork-arrest-induced replication slippage in the surviving population . Overall , our data establish that replication slippage results from inaccurate DNA synthesis during the restart of collapsed forks by recombination . We investigated the effects of replication stress , other than the replication block imposed by the RTS1-RFB , on replication slippage . Strains harbouring ura4− alleles ( base-substitutions , −1 frame-shift , and ura4-dup20 ) were exposed to replication-blocking agents or UV-C-induced DNA damages and the frequency of Ura+ revertants was scored . Three hours of treatment with either the topoisomerase I inhibitor camptothecin ( CPT ) or mitomycin C ( MMC ) , an inter-strand cross-linking agent ( ICls ) , increased the frequency of Ura+ revertants by 3 to 4 fold in the ura4-dup20 strain ( Figure 5A and 5B ) . At equivalent survival ( 70–90% ) , DNA-damages induced by a dose of 100 J/m2 of UV-C did not increase the frequency of Ura+ revertants in the ura4-dup20 strain . Increasing the UV-C dose ( 150 J/m2 ) resulted in an increased reversion effect . The other ura4 alleles exhibited an opposite behaviour pattern . As expected , UV-C-induced DNA damages , but not CPT or MMC treatment , increased the frequency of Ura+ revertants of the base-substitution and the −1 frame-shift mutants ( Figure 5A ) . Thus , replication slippage , unlike other point mutations , appears to be a mutation event specifically induced by replication stress . Hydroxyurea ( HU ) that prevents the bulk of dNTP synthesis during S-phase by inhibiting the ribonucleotide reductase , results in a slow-down of fork progression which did not induce replication slippage ( Figure 5A ) . In contrast , CPT and MMC treatments that lead to replication stress by causing fork collapse induced replication slippage . Homologous recombination is repressed during HU treatment and recombination proteins are recruited to collapsed but not stalled forks [52]–[54] . Consistent with this , we found that the rad22-d mutant is highly sensitive to acute exposure to CPT , but not to HU ( Figure S4 ) . Thus , acute exposure to HU results in stalled forks that recover without recombination , whereas recombination may be required for restarting forks that have collapsed due to CPT or MMC treatment . We confirmed that CPT-induced replication slippage results from collapsed forks and was thus S-phase specific: the ura4-dup20 strain was synchronized in early S-phase by HU treatment and released into S-phase with or without CPT . HU-synchronization and release into DMSO ( used as vehicle for CPT ) did not induce replication slippage . In contrast , the release of cells into S-phase in the presence of CPT stimulated replication slippage up to 12 fold ( Figure 5C ) . These data indicate that CPT-induced fork collapse results in error-prone DNA synthesis characterized by replication slippage . These experiments further support the view that replication slippage results from recovery of collapsed forks by recombination and point out that the RTS1-barrier is representative of collapsed forks restarted by homologous recombination . To investigate the inaccuracy of the DNA synthesis occurring immediately following the restart of collapsed forks , we analysed the involvement of TLS-DNA polymerases . In fission yeast , TLS pathways require either mono- or poly-ubiquitination of the clamp loader PCNA on lysine 164 [55] . We found that mutating this lysine to arginine residue did not affect replication slippage induced by the RTS1-RFB . None of Rev1 , Rev3 or DinB DNA polymerases were required for fork-arrest-induced replication slippage ( Figure S5 ) . Therefore , the error-prone DNA synthesis associated with fork recovery by recombination does not rely on TLS DNA polymerases activity . The mismatch repair ( MMR ) pathway is temporally coupled to DNA replication , and MMR components are associated with replication centres [56] . The heterodimer Msh2/Msh6 recognises mispaired DNA and Msh2/Msh3 recognises small DNA loops up to 31 bases long , arising from replication slippage [57] . The failure to repair small DNA loops results in more frequent insertions and deletions [58] . Therefore , MMR activity could potentially lead to an underestimation of the extent of fork-arrest-induced replication slippage . However , replication slippage induced by the RTS1-RFB activity was as frequent in msh2-d , msh6-d and msh3-d strains as in wild-type control . Also , spontaneous replication slippage at ura4 ( without RTS1-RFB ) was unaffected by the absence of MMR proteins ( Figure S5 ) . Therefore , there is no evidence that MMR repairs small DNA loops ( 20 nt ) in fission yeast and fork-arrest-induced replication slippage is not counteracted by MMR in our model system . Using conditional fork arrest constructs , we studied the consequences for genome instability of impediments to replication forks progression . A single fork arrest results in large-scale genomic changes and mutations that occur during recombination-mediated fork recovery ( Figure 6 ) . Inappropriate ectopic recombination at arrested forks results in GCRs , whereas appropriate restarting of the fork on the initial template results in error-prone DNA synthesis . GCRs and mutations at collapsed forks are genetically separable: Rqh1 limits fork-arrest-induced GCRs but not mutations ( Figure 2D and Figure 3B ) . We demonstrate here that collapsed forks whose progression resumes by recombination lose accuracy during DNA synthesis , resulting in frequent intra-template switches . Thus , homologous recombination contributes to completion of DNA replication when forks progression is impeded but also fuels genome modifications both at the chromosomal and nucleotide level . Non allelic homologous recombination ( NAHR ) between low copy number repeats ( LCR ) contributes to recombination-mediated GCRs in mitosis and meiosis . NAHR is responsible for translocations , deletions , inversions and loss of heterozygosity [59] . Ou et al . predicted 1143 LCR pairs in the human genome liable to mediate recurrent translocations [60] . In budding yeast , a single DSB is sufficient to mediate recombination-dependent translocation [61] . Here , we report that a single collapsed fork increases the rate of genomic deletion 40 fold , and that of translocation 5 fold . Fork-arrest-induced GCRs are mediated by NAHR between heterologous chromosomes . It is not clear whether fork arrest on both homologous repeats contributes to fork-arrest-induced GCRs . We could not address this question in our model system , because the RTS1 sequence close to the mat1 locus on chromosome II has a low RFB activity [62] . Also , the recruitment of recombination proteins at the RTS1 sequence near the mat1 locus is not regulated by the level of Rtf1 expression , showing that regulating Rtf1 expression was in itself insufficient to regulate the RTS1-RFB activity at the mat1 locus [35] . Inverted repeats ( IRs ) are structural elements that contribute to genome instability . Impediments to replication forks progressing through IRs favour their fusion to generate acentric and dicentric inverted chromosomes [11] , [14] . IRs in humans can also trigger the formation of inverted genomic segments and complex triplication rearrangements by a replication-based mechanism [47] . Here , we report that IRs near collapsed forks can increase the rate of GCRs by up to 1 , 500 fold . This high level of GCRs cannot be explained merely by the addition of independent inter- and intra-chromosomal recombination events . Rather , our analyses suggest that IRs may stimulate tri-parental recombination events induced by template switching of nascent strands at collapsed forks , such that three homologous sequences are involved . Similarly , recombination-dependent translocations induced by a single DSB in budding yeast is proposed to be the consequence of tri-parental recombination [63] , [64] . One possible mechanism is that IRs-induced GCRs result from successive template switches initiated by nascent strands at collapsed forks , reminiscent of the multiple template switches during break-induced-replication ( BIR ) in budding yeast [33] . Interestingly , Rqh1 prevents fork-arrest-induced GCRs by limiting both inter- and intra-chromosomal recombination without affecting fork restart efficiency . Thus , tri-parental recombination might correspond to multiple and successive template switches between homologous repeats . The high accuracy of DNA replication is compromised by impediments to fork progression , and recombination-mediated fork recovery results in decreased processivity of DNA synthesis ( Figure 6 ) . Recombination-induced mutations associated with DSBs or impeding DNA replication have been described previously . The formation of damaged single-stranded DNA during the resection of DSBs favours base-substitutions [65] . We detected the recruitment of the single-strand binding protein RPA up to 1 . 4 kb behind , but not ahead of the RTS1-arrested fork ( data to be published ) , showing that fork-arrest-induced mutation is not correlated with damaged single-stranded DNA exposed behind collapsed forks . Nevertheless , there were rare replication slippage occurred behind RTS1-arrested forks ( in the t<ura4-ori construct ) , suggesting that resumption of DNA synthesis can in some cases occur at a position behind the site of the collapsed fork . Recombination-dependent DNA synthesis occurring outside S-phase is also highly inaccurate during gene conversion and BIR , resulting in either template switches , base-substitutions or frame-shifts [33] , [34] , [66] . Elevated dNTP pools , due to activation of the DNA-damage checkpoint in G2 cells , contributes to the generation of mutations when hundreds of kbs are synthesised during BIR [34] . Here , we demonstrate that during normal S-phase progression , a single collapsed fork , restored by recombination , results in replication slippage up to 1 . 2 Kb away from the initial restarting point . Recombination-induced replication slippage has been reported previously . In fission yeast , a defect in pol alpha ( swi7-H4 mutant ) is associated with a recombination-mediated mutator phenotype characterized by an increased frequency of base-substitutions and Del/Dup between short direct repeats [67] . DNA-polymerase kappa ( DinB ) and zeta ( Rev3 ) are responsible for the increased base-substitution rate , but the DNA-polymerases involved in Del/Dup mutations were not identified [68] . In budding yeast , a defect in polymerase delta ( pol3-t mutant ) results in an increased level of replication slippage , mediated by homologous recombination [49] . In contrast , the increased rate of replication slippage in the absence of the accessory subunit of polymerase delta , Pol32 , does not depend on a functional recombination pathway [69] . Here , we report that recovery of collapsed forks by recombination is specifically associated with replication slippage . Nonetheless , spontaneous replication slippage events are also increased in strains mutated for recombination genes ( Figure 4D–4F ) . Pol32 is required for BIR and replication-induced template switches leading to segmental duplication [70] , [71] . Recombination is responsible for only half of these segmental duplications . Thus , it is possible that fork-restart mechanisms dependent on Pol32 and homologous recombination are prone to replication slippage and that in the absence of these pathways alternative micro-homology mediated mechanisms are revealed . We suggest that at least two steps of the recombination-based fork recovery mechanism can compromise genome stability ( Figure 6 ) . At an initial stage , recruitment of recombination proteins on stalled nascent strands favours both fork recovery and ectopic template switches leading to GCRs . At a later stage , once the replisome has been reconstituted and the fork has resumed its progression , the nascent strands are prone to intra-template switching leading to replication slippage . The reasons for the inaccuracy of DNA synthesis associated with restarted forks during scheduled DNA replication ( i . e . in S-phase ) remain to be determined . One possibility is that one or more factors are missing in the rebuilt replisome during recovery by recombination . Oncogene-induced replication stress results from unbalanced DNA replication that contributes to genome instability in precancerous cells [12] . Completion of DNA replication in such stress conditions is likely to rely on recombination-mediated fork recovery that in turn generates genome instability . Insertions/deletions flanked by micro-homology , responsible for copy number variations ( CNVs ) , have been identified both in cancer cells and also in response to replication inhibition [72] , [73]; their reported sizes are between 1 Kb and several tens of mega bases , but the analysis of these features has been limited by the resolution of array-based genomics approaches . Sub-microscopic insertions/deletions flanked by micro-homology have been also described at loci in which replication origins are scarce , including the human fragile site FRA3B , the instability of which is a consequence of replication stress [74] . Interestingly , homologous recombination contributes to the stability of fragile sites by facilitating complete replication or by repairing gaps and breaks at these sites . Thus , we propose that micro-homology-mediated CNVs could be viewed as scars left by error-prone replication forks restarted by recombination . Strains used were constructed by standard genetic techniques and are listed in Table S1 . 2DGE was performed as previously described [20] . To create ura4-dup20 and ura4-dup22 alleles associated or not with the RTS1-RFB , genomic DNA was isolated from selected 5-FOAR cells containing duplications identified by sequencing . A PCR fragment containing duplications within the ura4 ORF was amplified using the following primers: TTCTGTTCCAACA-CCAATGTTT and TCACGTTTATTTTCAAACATCCA . The PCR products were purified and used to transform strains SL206 ( ura4+ ) , SL350 ( t-ura4<ori ) and SL504 ( t<ura4-ori ) . Transformants were selected on 5-FOA-containing plates . Appropriate replacement of ura4+ by ura4-dup20 or ura4-dup22 was verified by PCR and sequencing . A minimum of 11 independent single colonies from appropriate strains growing with or without thiamine were inoculated in 10 ml of non-selective media ( with or without thiamine ) and grown to stationary phase . Appropriate dilutions were plated on supplemented YEA to determine plating efficiency and on 5-FOA-containing plates . Colonies were counted after 5–7 days of incubation at 30°C . The rate of ura4 loss was determined with the method of the median and data are presented on Table 1 . Statistical significance was detected using the nonparametric Mann-Whitney U test . At least 200 5-FOAR colonies per strain and condition were subjected to PCR analysis with the following primers: AAAACAAACGCAAACAAGGC and GTTTAACTATGCTTCGTCGG to amplify ura4 ORF , TGAATCCTCCGTTCAGTAGG and AAGGACTGCGTTCTTCTAGC to amplify rng3 and TTTCCTTTCACGGCTAACCC ( TLII ) and TGTACCCATGAGCAAACTGC ( TLIII ) to amplify the translocation junction . The amplified ura4 fragments were then sequenced on both strands , with the primers used to amplify the ura4 ORF . Only mutations present on both strands were considered to determine mutation spectra . Deletions , mutations and translocations were scored as percentages of all events and these values were used to balance the rates of ura4 loss to determine the respective rates of deletion , mutation and translocation ( see Figure S1 for deletion and translocation rates and Figure S2 for mutation rates ) . The fork-arrest-induced deletion , translocation and mutation rates ( Figure 2C–2E and Figure 3A ) were calculated by subtracting the rate obtained in presence of thiamine ( Rtf1 being repressed , OFF ) from the rate obtained in the absence of thiamine ( Rtf1 being expressed , ON ) . This method allows the spontaneous instability of IRs and the leakiness of the RTS1-RFB activity to be disregarded to determine strictly the rate of events induced by fork-arrest . The nonparametric Mann-Whitney U test was used to test for statistically significant differences . Exponentially growing cells were treated with 20 mM of HU , 40 µM of CPT or 1 mM of MMC . At indicated times , samples were taken and appropriate dilutions were plated on supplemented minimal media to determine plating efficiency and on uracil-free plates . Colonies were counted after incubation at 30°C for 5–7 days and the frequency of Ura+ colonies was determined . For strains showing a slow growth phenotype ( recombination mutants ) , the frequency of Ura+ revertants was determined as a function of the number of generations experiencing fork arrest at ura4 . Exponentially growing 5-FOAR cells were washed twice in water and used to inoculate uracil-containing media without thiamine . Every 24 hours , cells were counted to determine the number of generations , and appropriate dilutions were plated on supplemented minimal media and on uracil-free plates . Colonies were counted after incubation at 30°C for 5–7 days and the frequency of Ura+ colonies was determined . The slope of the curves presented on Figure 4F corresponds to the rate of replication slippage/generation . For strains showing similar growth to wild-type cells , a single 5-FOAR colony was grown on uracil-containing plates with or without thiamine for 2–3 days , and then grown in uracil-containing media with or without thiamine for 2 days at 30°C . Appropriate dilutions were plated on supplemented minimal media and on uracil-free plates . Colonies were counted after incubation at 30°C for 5–7 days and the frequency of Ura+ colonies was determined .
The appropriate transmission of genetic material during successive cell divisions requires the accurate duplication and segregation of parental DNA . The semi-conservative replication of chromosomes during S-phase is highly accurate and prevents accumulation of deleterious mutations . However , during each round of duplication , there are many impediments to the replication fork machinery that may hinder faithful chromosome duplication . Homologous recombination is a universal mechanism involved in the rescue of replication forks by rebuilding a replication apparatus at the fork ( by mechanisms that are not yet understood ) . However , recombination can jeopardize genome stability because it allows genetic exchanges between homologous repeated sequences dispersed through the genome . In this study , we employ a fission yeast-based arrest of a single replication fork to investigate the consequences of replication fork arrest for genome stability . We report that a single blocked fork favours genomic deletions , translocations , and mutations; and this instability occurs during fork recovery by recombination . We also report that a single arrested fork that resumes its progression by recombination is prone to causing replication slippage mediated by micro-homology . We propose that deletions/duplications observed in human cancer cells suffering from replication stress can be viewed as scars left by error-prone replication forks restarted by recombination .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "cellular", "stress", "responses", "microbiology", "mitosis", "model", "organisms", "dna", "replication", "dna", "dna", "synthesis", "mycology", "chromosome", "biology", "schizosaccharomyces", "pombe", "biology", "molecular", "biology", "yeast", "cell", "biology", "nucleic", "acids", "genetics", "yeast", "and", "fungal", "models", "dna", "recombination", "molecular", "cell", "biology", "genetics", "and", "genomics" ]
2012
Recovery of Arrested Replication Forks by Homologous Recombination Is Error-Prone
In the present study , we assessed the annual screening coverage ( i . e . , the percentage of dogs that are screened for anti-Leishmania antibodies annually ) in the municipality of Sobral , Ceará state , Brazil . Data on the number of dogs screened during 2008−2017 ( except 2010 ) were obtained from the Centre for Zoonoses Control of Sobral . The annual screening coverage during 2012−2017 was calculated . Data on human visceral leishmaniasis ( VL ) cases during 2008−2017 were compiled from the National Disease Notification System . Correlation analyses were performed to assess the correlation between canine and human data . During 2008−2017 , 73 , 964 dogs ( range , 0 to 13 , 980 dogs/year ) were serologically screened and 2 , 833 ( 3 . 8% ) were positive . The annual screening coverage during 2012−2017 ranged from 11 . 1% to 45 . 7% . There were no significant correlations between the number of dogs culled and the number of human VL cases , canine positivity and human VL incidence , number of dogs culled and human VL incidence , or between canine positivity and number of human VL cases . An inconsistent and relatively low annual screening coverage was found in the study area , with no dog being screened in 2010 due to the lack of serological tests . Our results highlight that many dogs potentially infected with Leishmania infantum have been virtually overlooked by public health workers in the study area , perhaps with a negative , yet underestimated , impact on the control of canine and human VL . Hence , the failure of the dog culling strategy in controlling human VL in Brazil may be due to the low screening coverage and low percentage of culled dogs , rather than the absence of associations between canine and human infections . Human visceral leishmaniasis ( VL ) is a neglected vector-borne disease of great public health significance . The disease is endemic in more than 60 countries , with estimated 200 , 000 to 400 , 000 human cases and 20 , 000 to 40 , 000 deaths occurring annually worldwide [1] . In the Americas , VL is a zoonosis caused by Leishmania infantum and Brazil concentrates most of the notified cases , with estimated 4 , 200 to 6 , 300 new cases per year [1] . Leishmania infantum is transmitted to susceptible hosts , including humans , through the bite of infected female phlebotomine sand flies [2] . While several animals can serve as a source of infection to phlebotomine sand flies , dogs are the most important reservoirs in domestic settings [3] . As such , the presence of infected dogs is reputed to be a risk for L . infantum infection in humans [4] . In this perspective , the culling of Leishmania-seropositive dogs has been recommended as a control measure in many countries where human VL is endemic , including in Brazil [5] . Currently , this measure is one of three main strategies of the VL surveillance and control program of the Ministry of Health of Brazil , which also includes early diagnosis and treatment of human cases , as well as vector control [6] . Nonetheless , the dog culling strategy has long been an issue of debate , as there is no convincing scientific evidence supporting its effectiveness [5 , 7 , 8] . Reasons for the failure of the dog culling strategy in controlling human VL in Brazil may include limited sensitivity of serological tests , long delay between diagnosis and the removal of infected dogs , and rapid replacement of culled dogs by new susceptible ones [5 , 9–13] . Another important factor that may negatively influence the effectiveness of the dog culling strategy is the annual screening coverage , i . e . , the percentage of dogs living in a given area that are screened for anti-Leishmania antibodies annually . In a recent study conducted in the city of Araçatuba , south-eastern Brazil , the authors reported an annual screening coverage ranging from 1 . 0% to 10 . 0% [14] . This study highlighted that the effectiveness of the dog culling strategy is likely compromised by the low annual screening coverage . In the present study , we assessed the annual screening coverage in the municipality of Sobral , a historical focus of human VL in north-eastern Brazil , where the first outbreak of the disease was detected in this country and where the dog culling strategy was firstly implemented [15] . In particular , our hypothesis was that the inconsistent annual screening coverage may be one of the reasons for the failure of the dog culling strategy in controlling human VL in an important urban focus of this disease in Brazil . The municipality of Sobral ( 03°40'26"S , 40°14'20"W , altitude: 70 m above the sea level ) is located in the north-western region of Ceará state , 240 km away from Fortaleza ( the capital city ) . Sobral is home to 205 , 529 residents spread over an area of 2 , 122 . 989 km2 ( including both rural and urban areas ) . Its urban area is divided into 35 districts ( Fig 1 ) . Sobral has the second best human development index ( HDI = 0 . 714 ) of Ceará and 75 . 6% of its territory has adequate sanitary sewers [16] . The climate is tropical semi-arid ( steppe climate ) , characterized by rainy and dry periods , with rains concentrated from January to May , monthly relative humidity ranging from 56 . 2% to 85 . 9% and monthly temperature ranging from 21°C to 39°C [17] . Data regarding the number of dogs serologically screened and the number of seropositive ones were obtained from the Centre for Zoonoses Control ( CZC ) of Sobral . Data from 2008 to 2017 ( except 2010 , when no screening was conducted due the lack of serological tests ) and from 25 out of 35 districts were obtained , representing 71 . 4% of urban area of Sobral ( there was no screening in 10 districts due to the CZC’s logistical reasons ) . At the CZC , dogs are serologically screened when the owners spontaneously bring their dogs to for testing or when public health agents visit each district to sample and screen both resident and stray dogs . Until 2009 , all dogs were screened using an indirect fluorescent antibody test ( IFAT ) ( IFI—Leishmaniose Visceral Canina , Bio-Manguinhos , Fiocruz , Rio de Janeiro , Brazil ) . Since 2011 , all dogs started to be screened using a rapid immunochromatographic test ( ICT ) ( TR DPP Leishmaniose Visceral Canina , Bio-Manguinhos , Fiocruz , Rio de Janeiro , Brazil ) and , if positive , retested using an enzyme-linked immunosorbent assay ( ELISA ) ( EIE—Leishmaniose Visceral Canina , Bio-Manguinhos , Fiocruz , Rio de Janeiro , Brazil ) . The CZC informed that all dogs positive by IFAT ( until 2009 ) and by both ICT and ELISA ( from 2011 onwards ) were humanely culled and then the carcasses were incinerated , as recommended by the Ministry of Health of Brazil [6] . Secondary data regarding human VL cases notified during 2008−2017 were obtained from Health Surveillance Secretariat of Sobral . Data were compiled from the National Disease Notification System ( SINAN ) database and processed anonymously . Data regarding the human population size during 2008–2017 were obtained from Brazilian Institute of Geography and Statistics ( IBGE ) [16] . The canine population size in the last six years ( 2012–2017 ) were obtained from CZC , which conducts annual censuses in Sobral . The maps were produced using QGIS software version 2 . 18 . 28 [18] and based on public geographical data obtained from OpenStreetMap [19] . The annual screening coverage was calculated by dividing the number of dogs serologically screened in a given year by the canine population size in the same year and then multiplied by 100 . Positivity was calculated by dividing the number of seropositive dogs by the number of dogs serologically screened and then multiplied by 100 . Results were expressed as percentages . Incidence of human VL was calculated by dividing the number of new cases in a given year by the human population size in the same year and multiplied by 100 , 000 . Results were expressed as the number of cases per 100 , 000 population . Prior to statistical analyses , normality of data was checked using Lilliefors . As data presented a non-normal distribution , the correlation between dog culling and human VL incidence was investigated using Spearman’s coefficient ( rs ) . The trend in the human VL incidence in Sobral over the years was assessed using Mann-Kendall trend test . The differences were considered statistically significant when P ≤ 0 . 05 . Statistical analyses were performed using BioEstat v . 5 . 3 ( Instituto Mamirauá , Tefé , Amazonas , Brazil ) and PAST 3 . 23 [20] . The Health Secretary of Sobral ( 0184/2018 ) and Research Ethics Committee ( 97934718 . 4 . 0000 . 5190 ) of the Aggeu Magalhães Institute ( Fiocruz ) approved the access and using of secondary data ( number of dogs serologically screened and human VL cases ) analysed in this research . From 2008 to 2017 , 73 , 964 dogs were serologically screened for anti-Leishmania antibodies , with an average of 8 , 218 dogs sampled per year ( range , 0–13 , 980 ) . The annual screening coverage from 2012 to 2017 ranged from 11 . 1% to 45 . 7% ( Table 1 ) . In total , 2 , 833 out of 73 , 964 dogs serologically screened resulted positive , representing an overall positivity of 3 . 8% . The annual positivity ranged from 0 . 5% to 8 . 1% ( Table 1 ) . There was no correlation between the annual screening coverage and the number of seropositive dogs detected each year ( rs ( 4 ) = 0 . 486; p = 0 . 3287 ) . Over the study period , the decrease in the canine positivity in a given year was usually preceded by a higher positivity in the previous year , resulting in a bimodal pattern , with peaks of positivity every two years in some districts ( Fig 2 ) . This bimodal pattern was observed in most districts ( 14/25 ) ( Fig 2A and 2B ) , with the remaining districts displaying no defined pattern ( Fig 2C ) . The positivity in each district ranged from 1 . 6% to 13 . 1% . In average , 113 dogs were culled per district ( range , 1–303 dogs/district ) during the study period . In addition , nine out of 25 districts showed positivity above average ( Table 2 ) . From 2008 to 2017 , 247 human cases of VL were notified in 17 districts of Sobral . Although the Padre Ibiapina district reported 14 human VL cases , no single dog was serologically screened in this district during the same period ( Fig 1 and Table 2 ) . The total number of cases per district ranged from one to 52 . In addition , 11 districts reported at least 10 cases from 2008 to 2017 ( Table 2 and Fig 3 ) . The incidence of human VL in Sobral ranged from 0 . 5 to 25 . 7 cases per 100 , 000 population ( Table 1 ) during the study period and no fatal cases were recorded . There was a significant decreasing trend in the human VL incidence in Sobral from 2008 to 2017 ( Mann-Kendall trend test , p = 0 . 001 ) . There were no statistically significant correlations between the number of dogs culled and the number of human VL cases ( rs ( 7 ) = –0 . 367; p = 0 . 332 ) , canine seropositivity and human VL incidence ( rs ( 7 ) = –0 . 067; p = 0 . 864 ) , number of dogs culled and human VL incidence ( rs ( 7 ) = –0 . 050; p = 0 . 898 ) , or between canine seropositivity and number of human VL cases ( rs ( 7 ) = –0 . 377; p = 0 . 318 ) . In conclusion , we found an inconsistent and relatively low annual screening coverage in the study area , with no dog being screened in 2010 due to the lack of serological tests . Our results highlight that many dogs potentially infected with L . infantum have been virtually overlooked by public health workers in the study area , perhaps with a negative , yet underestimated , impact on the control of canine and human VL . Hence , the failure of the dog culling strategy in controlling human VL in Brazil could also be , in some instances , a result of the low screening coverage and low of percentage of culled dogs , rather than the absence of an association between canine and human infections .
The euthanasia of Leishmania-seropositive dogs has been recommended for controlling human visceral leishmaniasis ( VL ) in some countries where this zoonosis is endemic . We assessed the annual screening coverage ( i . e . , the percentage of dogs living in a given area that are screened for anti-Leishmania antibodies annually ) in the municipality of Sobral , Ceará state , one of the main foci of human VL in Brazil . From 2008 to 2017 , nearly 74 , 000 dogs were screened and 3 . 8% of them were positive for anti-Leishmania antibodies . No statistically significant correlation was found between the number of dogs culled annually and the incidence of human VL . The annual screening coverage ranged from 11 . 1% to 45 . 7% . Our results highlight an inconsistent and relatively low annual screening coverage , indicating that dogs potentially infected with L . infantum have been virtually overlooked by public health workers in the study area .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[ "kala-azar", "medicine", "and", "health", "sciences", "pathology", "and", "laboratory", "medicine", "tropical", "diseases", "geographical", "locations", "vertebrates", "parasitic", "diseases", "dogs", "animals", "mammals", "parasitic", "protozoans", "protozoans", "leishmania", "neglected", "tropical", "diseases", "public", "and", "occupational", "health", "infectious", "diseases", "zoonoses", "serology", "south", "america", "protozoan", "infections", "brazil", "people", "and", "places", "health", "screening", "leishmania", "infantum", "eukaryota", "leishmaniasis", "biology", "and", "life", "sciences", "amniotes", "organisms" ]
2019
Failure of the dog culling strategy in controlling human visceral leishmaniasis in Brazil: A screening coverage issue?
Type 2 immune responses are essential in protection against intestinal helminth infections . In this study we show that IL-22 , a cytokine important in defence against bacterial infections in the intestinal tract , is also a critical mediator of anti-helminth immunity . After infection with Nippostrongylus brasiliensis , a rodent hookworm , IL-22-deficient mice showed impaired worm expulsion despite normal levels of type 2 cytokine production . The impaired worm expulsion correlated with reduced goblet cell hyperplasia and reduced expression of goblet cell markers . We further confirmed our findings in a second nematode model , the murine whipworm Trichuris muris . T . muris infected IL-22-deficient mice had a similar phenotype to that seen in N . brasiliensis infection , with impaired worm expulsion and reduced goblet cell hyperplasia . Ex vivo and in vitro analysis demonstrated that IL-22 is able to directly induce the expression of several goblet cell markers , including mucins . Taken together , our findings reveal that IL-22 plays an important role in goblet cell activation , and thus , a key role in anti-helminth immunity . Type 2 immune responses are essential in protection against intestinal helminth infections , including the rodent hookworm Nippostrongylus brasiliensis ( Nb ) [1] . Type 2 immunity involves the recruitment of effector cells such as eosinophils , mast cells and production of IgE antibodies [2] and it is well established that the activation of IL-4- and IL-13-producing CD4+ T helper 2 cells is central in the development of a successful anti-parasite response . One of the key components in the expulsion of intestinal helminths is secretion of mucus by goblet cells [2] . Intestinal goblet cells are found interspersed within the epithelial monolayer and are differentiated from epithelial progenitor cells . Goblet cells produce a number of effector molecules including a range of mucins and antimicrobial proteins , including trefoil factors and resistin-like molecules , which enable these to play a key part in innate defense mechanisms in the gut , against both bacterial and helminth infections [2] , [3] . IL-22 is a member of the IL-10 cytokine family and is produced by a wide variety of innate and adaptive immune cells including CD4+ T cells , most notably Th17 and Th22 cells , CD8+ T cells , natural killer ( NK ) cells , lymphoid tissue inducer ( LTi ) cells and other group 3 innate lymphoid cells ( ILCs ) [4] . The heterodimeric receptor for IL-22 , consisting of the IL-22R and the IL-10R2 chain , is exclusively expressed on non-hematopoietic cells , such as intestinal epithelial cells , and its signaling is mediated via Stat3 [5] , [6] , [7] , [8] , [9] . IL-22 has been shown to directly mediate epithelial defence mechanisms through the induction of IL-6 , IL-8 and various antimicrobial peptides [6] , [10] , [11] , [12] , [13] . Furthermore , a protective function of IL-22 has been demonstrated in some colitis models and a model of Concanavalin A induced liver damage [14] , [15] , [16] , [17] . Therefore it appears that the function of IL-22 is to strengthen epithelial barriers , mediate repair and wound healing mechanisms as well as participate in epithelial defence . Interestingly , IL-22 has been shown to be upregulated within the human gastrointestinal tract following infections with the whipworm Trichuris trichiuria and the hookworm Necator americanus [18] , [19] , but no studies have as yet demonstrated a role for IL-22 in intestinal helminth infections and the associated type 2 response . To clarify the role for IL-22 in the defence against intestinal helminth infection , we infected wild-type ( WT ) and IL-22-deficient mice with Nippostrongylus brasiliensis , a rodent nematode with a life cycle resembling that of human hookworm . Our data show that IL-22-deficient mice have reduced worm clearance , despite strong induction of IL-4 , IL-5 and IL-13 in the mesenteric lymph nodes and mucosal tissue . Despite this intact type 2 cytokine induction IL-22-deficient mice showed a defective goblet cell response and in vitro and ex vivo analyses revealed that IL-22 can directly regulate the expression of several goblet cell markers . Taken together , our data suggest that IL-22 plays a key role in driving intestinal goblet cell responses in vivo and thus acts as an important mediator of intestinal worm expulsion . All animal work was approved following local ethical review by MRC National Institute for Medical Research , NIMR , Animal Procedures and Ethics Committee and was performed in strict accordance with the U . K Home Office Animals ( Scientific Procedures ) Act 1986 ( approved H . O Project License 80/2506 ) . Six to nine week old male and female C57BL/6 and IL-22KO mice [20] were bred at the specific pathogen-free animal facility at the MRC National Institute for Medical Research ( NIMR , London , UK ) . Age- and sex-matched experimental animals ( 3–8 per group ) were infected with 500 infective Nippostrongylus brasiliensis ( Nb ) larvae by subcutaneous injection [21] or with 150 embryonated Trichuris muris ( Tm ) eggs by oral gavage [22] . Mesenteric lymph nodes were removed and single cell preparations were resuspended in RPMI 1640 supplemented with 10% heat-inactivated FCS , 2 mM L-glutamine , 100 U/ml penicillin , 100 µg/ml streptomycin and 0 . 05 mM β-mercaptoethanol ( Life technologies ) . Cells were cultured at 37°C and 5% CO2 in flat-bottomed 96-well plates at a final concentration of 5×106/ml in a final volume of 0 . 2 ml/well . Cells were stimulated with Nb or Tm antigen ( 25 µg/ml ) , or plate-bound anti-CD3 antibody ( mAb145-2C11 , 10 µg/ml , ATCC ) and cell-free supernatants were harvested after 48 hours and stored at −80°C . Cytokine analyses were carried out using a multiplex cytometric bead assay ( Flowcytomix , eBiosciences ) . Explants of small intestine were washed extensively in ice cold PBS and cultured overnight in the same medium as above , with or without the addition of recombinant IL-22 ( R&D systems ) . LS174T cells ( kindly provided by Dr AC Williams , University of Bristol , UK ) were cultured in DMEM supplemented with 10% heat-inactivated FCS , 2 mM L-glutamine , 100 U/ml penicillin and 100 µg/ml streptomycin . After removal of Peyer's patches the small intestine was cut into 5 mm pieces and epithelial cells and intraepithelial lymphocytes were first removed by shaking gut pieces in PBS with 10% FCS , 1 mM pyruvate , 20 µM Hepes , 10 mM EDTA , 100 U/ml penicillin , 100 µg/ml streptomycin , 10 µg/ml Polymyxin B and 2 mM DTT for 30 min at 37 C . The remaining gut tissue was washed and digested using Collagenase D ( Roche , 1 mg/ml ) and DNAse1 ( Sigma , 10 U/ml ) for 45 minutes at 37°C , before being subjected to Percoll centrifugation ( 37 . 5% ) , followed by washing and resuspension of the isolated lamina propria leukocytes in medium . To identify innate lymphoid cells ( ILC ) , isolated leukocytes were stained by using fluorochrome-coupled antibodies against CD45 , Thy1 . 2 , IL-7R ( CD127 ) and a combination of lineage markers ( Lin ) , including CD3 , CD8 , CD11b , CD11c , CD19 , CD49b , TCR-β , TCR-γδ , NK1 . 1 , GR-1 and Ter119 . ILC were defined as CD45+Lin−Thy1 . 2+IL-7R+ . For further characterization of ILC surface marker expression , antibodies against CD4 and NKp46 were used . For intracellular cytokine staining isolated leukocytes were restimulated with phorbol 12 , 13-dibutyrate ( PdBU ) and ionomycin ( both at 0 . 5 µg/ml ) in the presence of brefeldin A ( 1 µg/ml ) for 2 . 5 h , fixed with formalin ( 3 . 8% ) , permeabilised with octylphenyl-polyethylene glycol ( 0 . 1% , Sigma ) , and stained with fluorochrome-coupled antibodies against IL-17A and IL-22 . All samples were acquired on a LSRII flow cytometer ( BD Biosciences ) and analysed with the FlowJo software ( Treestar Inc . ) . Consecutive lengths of small intestine taken 10 cm distal to the pyloric sphincter were fixed in neutral-buffered formalin , histologically processed using standard methods , and 5 µm sections were stained for goblet cells ( Periodic Schiff ) . The number of cells per 20 randomly selected villus-crypt units ( VCU ) was determined under light microscopy from at least two sections per animal . Tissues were harvested and stored in RNAlater ( Qiagen ) at −80 C until processing . RNA was purified using Trizol ( Life technologies ) . Reverse transcription was performed using a Quantitect RT kit ( Qiagen ) and real time PCR was performed in an ABI 7500 sequence detection system ( Applied Biosystem ) using the SYBR Green PCR Master Mix ( Qiagen ) . Results were normalised to the housekeeping gene hypoxanthine guanine phosphoribosyl transferase ( HPRT ) . Significant differences ( P<0 . 05 ) between experimental groups were determined using Student's t-test . Upregulation of IL-22 expression in the gastrointestinal tract has been reported in human helminth infection [18] , [19] . To address the question if IL-22 may play a role in anti-helminth immunity , we utilized the mouse model of Nippostrongylus brasiliensis ( Nb ) . We first assessed IL-22 mRNA expression in the small intestine at different time points post Nb infection ( p . i . ) ( Fig . 1A ) . Elevated expression of Il22 mRNA was found from day 6 p . i . and further increased at day 10 p . i . Flow cytometric analysis of lamina propria lymphocytes isolated from the small intestine of infected mice showed IL-22 cytokine staining both in the lineage-negative ( Lin− ) and lineage-positive ( Lin+ ) compartments , with the majority ( ∼80% ) of the IL-22 coming from Lin+ cells ( Fig . 1B , C ) . Further characterization of the IL-22+Lin+ and IL-22+Lin− subsets revealed that amongst the Lin+ cells , CD4+ T cells were the predominant IL-22+ cell type , while the IL-22+Lin− population consisted predominantly of NKp46− innate lymphoid type 3 cells ( ILC3 ) ( Fig . 1D , E ) . Co-staining for IL-17A demonstrated that ∼70% of the T cell-derived IL-22 was produced by Th17 cells , whereas the majority of the IL-22+Lin− cells did not show co-production of IL-17A ( Fig . 1F ) . Thus , our data show that IL-22 is produced by both T cell and non-T cell populations in the lamina propria during Nb infection . This is in agreement with a number of studies demonstrating that IL-22 can be produced by a variety of cells in the intestine , including innate cells , such as LTi , NKp46+ and NKp46− ILC3 cells , as well as conventional CD4+ T cells [23] , [24] , [25] . In addition , Basu et al [13] recently demonstrated that early production of IL-22 during bacterial infection is mainly derived from innate sources , shifting to CD4-derived during later stages of infection , and that both sources play a vital role in protection at different stages against enteric infection , thus demonstrating the importance of both innate and adaptive sources of IL-22 in mucosal responses . Although the increased expression of IL-22 in human intestinal mucosa after helminth infections suggests a possible functional link between IL-22 and anti-helminth immunity [18] , [19] , the role of IL-22 in mouse models of gastrointestinal helminth infections has not been addressed . Therefore , we infected WT and Il22−/− mice with Nb and assessed the intestinal worm burdens at various time points after infection ( Fig . 2 ) . While similar numbers of parasites had reached the intestine by day 3 in WT and Il22−/− mice , indicating an unperturbed lung passage of the Nb larvae , we observed a marked delay in worm expulsion in Il22−/− mice with significantly increased worm burdens at day 6 and day 10 compared to WT animals ( Fig . 2 ) , demonstrating a key role for IL-22 in the clearance of intestinal helminth infection . Goblet cell hyperplasia in the intestinal epithelium is a hallmark of intestinal helminth infections and is crucial for worm expulsion in Nb infection [26] . Histological analysis of the small intestine of WT and Il22−/− mice at various time points after Nb infection revealed a significant reduction of goblet cell numbers in the small intestine of IL22−/− mice as compared to their WT counterparts ( Fig . 3A , B ) . Analysis of mRNA expression of the goblet cell markers Clca3 ( Gob5 ) [27] , Retnlb ( RELMβ ) , Tff2 ( trefoil factor 2 ) and mucins Muc1 , Muc2 and Muc3 [28] , showed that their upregulation observed in Nb-infected WT mice was almost completely abolished in absence of IL-22 ( Fig . 3C ) . Thus , the delayed worm expulsion observed in the absence of IL-22 is strongly correlated to reduced goblet cell hyperplasia and reduced expression of goblet cell mediators . Since type 2 responses are instrumental in anti-helminth immunity and goblet cell hyperplasia , we hypothesized that impaired type 2 cytokine production may be responsible for the reduced goblet cell function and delayed worm expulsion in Il22−/− mice . However , neither type 2 cytokine mRNA expression in intestinal tissues , or protein levels in supernatants from Nb antigen-restimulated mesenteric lymph node cells at day 10 p . i . were impaired in Il22−/− mice ( Fig . 4 A , B ) . In fact , the levels of IL-4 mRNA in the intestine and IL-5 protein from lymph nodes were even increased in Il22−/− mice , possibly as a result of overcompensation due to the increase in worm burden . Therefore , Il22−/− mice have impaired intestinal worm expulsion and reduced goblet cell function despite normal type 2 cytokine production . A possible explanation for the reduction in goblet cell hyperplasia detected in Il22−/− mice is a direct effect of IL-22 on the differentiation and/or activation status of goblet cells . To address this possibility , we treated small intestinal tissue ex vivo from uninfected WT mice with IL-22 and analysed expression of goblet cell markers . IL-22 treatment induced expression of Retnlb , Muc1 and Muc3 , but not Clca3 , Muc2 or Tff2 ( Fig . 5A ) . In addition , we cultured LS174T cells , a human mucus-secreting intestinal adenocarcinoma cell line , in the presence of human IL-22 , IL-13 , or a combination of both cytokines , and observed that IL-22 alone induced expression of several mucins including MUC1 , MUC3 , MUC4 and MUC5b , but not MUC5AC ( Fig . 5B ) . Similarly to our observations using mouse tissue ( Fig . 5A ) we did not observe induction of expression of Clca1 ( the human ortholog of mouse Clca3 ) by IL-22 , however , IL-13 alone induce expression of Clca1 and this induction was further amplified by IL-22 in a synergistic manner ( Fig . 5B ) . Taken together , this data demonstrate that IL-22 alone is able to induce the expression of several goblet cell mediators whilst also working in synergy with other mucogenic cytokines such as IL-13 , in the induction of other goblet cell mediators , such as Clca1/3 . This data is in agreement with the study by Sugimoto et al , where IL-22 via STAT3 signaling , was linked to goblet cell hyperplasia and mucin expression in a model of colitis [15] . In order to confirm the role for IL-22 in anti-helminth immunity we infected Il22−/− and WT mice with the murine whipworm Trichuris muris , another helminth model where goblet cells and mucus production is known to play an important role in resistance to infection [29] , [30] , [31] . We assessed worm burden , cytokine and goblet cell responses 21 days post infection . Similarly to Nb infection we found that Tm infected Il22−/− mice had higher worm burdens despite normal type 2 cytokines responses compared to WT mice ( Fig . 6 ) . Furthermore , Tm infected Il22−/− mice had reduced goblet cell hyperplasia and reduced expression of intestinal Retnlb , Muc1 , Muc3 , Clca3 and Tff2 , but not Muc2 ( Fig . 6 ) . Thus , our data demonstrate that IL-22 deficiency significantly impairs anti-helminth immunity and goblet cells function in two different murine nematode models . Previous studies have shown that IL-22 is particularly important in regulating inflammatory responses within the intestine through the production of antimicrobial peptides , as well as enhancement and regulation of epithelial wound repair and regenerative mechanisms [6] , [10] , [11] , [12] , [13] , [25] . In addition , IL-22 has also been shown to play a detrimental role in chronic mucosal inflammation and progression to colorectal cancer [32] , [33] demonstrating a pivotal role for IL-22 in balancing tissue regeneration and tumour development in the intestinal environment . Our data now suggest an important function for IL-22 in promoting other epithelial functions , particularly goblet cell-derived production of mucins and antimicrobial peptides , leading to anti-helminth immunity . A previous study by Wilson et al [34] found no role for IL-22 in the development of hepatic pathology during infection with the trematode helminth Schistosoma mansoni in mice , thus supporting the concept that IL-22 exerts organ and/or pathogen-specific functions . Mucus is believed to play an important part in intestinal anti-helminth mechanisms , partly through the generation of the mucus barrier , but also via other proteins in the secretions such as the antimicrobial peptide Relmβ ( FIZZ2 ) which may impair worm movement and feeding [26] , [31] . The mucins themselves are a large family of both secreted and surface bound glycoproteins , together forming the mucus layer . Quantitative and qualitative differences in mucus composition are evident during intestinal infections [35] , [36] although information is still limited regarding the contribution of specific mucins to the overall physical properties of mucus and how mucus composition may change with respect to different types of infections . Little is also known regarding the cytokine-mediated control of mucin expression , but it is clear that a number of cytokines are able to induce mucin expression , and other goblet cell products in vitro and in vivo . This includes type 2 cytokines such as IL-13 and IL-4 [37] , [38] , but also pro-inflammatory cytokines such as TNF and IL-1 [39] , [40] . Although IL-13 and IL-4 are believed to be the key driver cytokines for the induction of goblet cell hyperplasia during helminth infections [41] , [42] , [43] , some studies have demonstrated IL-4/13-independent goblet cell hyperplasia [44] . Furthermore , increase in intestinal Muc2 and Muc3 expression during infection with the nematode Trichinella spiralis has been reported in IL-4-deficient mice [45] . The data presented here , together with that of Sugimoto et al [15] , now show that IL-22 is central to goblet cell hyperplasia and function in the intestine . Our analyses of Nb and Tm-infected IL-22-deficient mice demonstrate that IL-22 increases the abundance of goblet cells and is required for upregulation of mucins and other goblet cell products in vivo . The fact that IL-22-deficient mice mount a strong type 2 cytokine response in both draining lymph nodes and intestinal tissue provides further evidence that IL-4/13 alone are not sufficient for promoting effective goblet cell functions during intestinal helminth infection . In contrast , our ex vivo and in vitro analyses suggest that IL-22 ( and not IL-13 ) alone might be sufficient to increase mucin production by the intestinal epithelium in certain settings . In vivo , however , a concerted action of the type 2 cytokines together with IL-22 , and possibly other inflammatory mediators , is most likely needed for an efficient anti-helminth response . The reported findings of upregulated IL-22 expression following human infections with the whipworm Trichuris trichiuria and the hookworm Necator americanus [18] , [19] indicate that IL-22 may play a similar role in human helminth infections . In conclusion , our study demonstrates a key role for IL-22 in goblet cell function and , thus , for anti-helminth immunity in the intestine . In addition , our data provide additional insight into the pivotal role played by IL-22 in mucosal immunity in protection against various types of infectious pathogens .
Intestinal helminth ( worm ) infections are some of the most common parasite infections in the world . Immunity to worm infection is dependent on the production of Type 2 cytokines , such as IL-4 and IL-13 , and the induction of mucosal defence mechanisms including production of mucus by intestinal goblet cells . Here we show that the cytokine IL-22 , which was previously known to be involved in the defence against bacterial infections in the gut , is also involved in the defence against intestinal worms . IL-22 deficient mice are unable to expel the rodent parasites Nippostrongylus brasiliensis and Trichuris muris from their intestines despite the fact that they make strong Type 2 cytokine responses . This inability to expel the worms correlates with a reduction in the number of goblet cells as well as a reduction in intestinal mucins and other goblet cell products . We also demonstrate that IL-22 is able to act directly on goblet cells to stimulate the secretion of mediators such as mucins . Taken together , our data show that IL-22 is a key mediator of anti-helminth immunity in the gut . Furthermore , our data provide additional insight into the pivotal role played by IL-22 in protection against various types of intestinal pathogens .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results", "and", "Discussion" ]
[]
2013
IL-22 Mediates Goblet Cell Hyperplasia and Worm Expulsion in Intestinal Helminth Infection
Interspecific and intervarietal hybridization may contribute to the biological diversity of fungal populations . Cryptococcus neoformans is a pathogenic yeast and the most common fungal cause of meningitis in patients with AIDS . Most patients are infected with either of the two varieties of C . neoformans , designated as serotype A ( C . neoformans var . grubii ) or serotype D ( C . neoformans var . neoformans ) . In addition , serotype AD strains , which are hybrids of these two varieties , are commonly isolated from clinical and environmental samples . While most isolates of serotype A and serotype D are haploid , AD strains are diploid or aneuploid , and contain two sets of chromosomes and two mating type alleles , MATa and MATα , one from each of the serotypes . The global population of serotype A is dominated by isolates with the MATα mating type ( Aα ) ; however , about half of the globally analyzed AD strains possess the extremely rare serotype A MATa allele ( Aa ) . We previously described an unusual population of serotype A in Botswana , in which 25% of the strains contain the rare MATa allele . Here we utilized two methods , phylogenetic analysis of three genes and genotyping by scoring amplified fragment length polymorphisms , and discovered that AD hybrid strains possessing the rare serotype A MATa allele ( genotype AaDα ) cluster with isolates of serotype A from Botswana , whereas AD hybrids that possess the MATα serotype A allele ( AαDa and AαDα ) cluster with cosmopolitan isolates of serotype A . We also determined that AD hybrid strains are more resistant to UV irradiation than haploid serotype A strains from Botswana . These findings support two hypotheses: ( i ) AaDα strains originated in sub-Saharan Africa from a cross between strains of serotypes A and D; and ( ii ) this fusion produced hybrid strains with increased fitness , enabling the Botswanan serotype A MATa genome , which is otherwise geographically restricted , to survive , emigrate , and propagate throughout the world . The impact of hybridization between fungal species and varieties on their evolution is unresolved . Hybridization may be considered an evolutionary disadvantage because some interspecies hybrids have reduced fitness [1 , 2] . Alternatively , natural hybridization may be beneficial because it can generate new evolutionary lineages that are able to occupy novel ecological niches [2–5] . In recent years , several examples of epidemiologically successful interspecific hybrids that were able to colonize new environments and infect new hosts have been described among fungal plant pathogens [5–7] and oomycetes [4] . These hybrids illustrate the effect of natural hybridization on the production of biological diversity in fungal populations . C . neoformans is an opportunistic human pathogen that is acquired exogenously and readily isolated from the environment worldwide [8] . Based on serological differences in capsular epitopes and molecular phylogenetic evidence , two varieties are recognized: C . neoformans var . grubii , which encompasses isolates of serotype A , and C . neoformans var . neoformans , which includes isolates of serotype D [8–10] . These varieties represent monophyletic lineages that diverged approximately 18 million years ago [11 , 12] , and according to the phylogenetic species concept , they may reflect cryptic species [13] . More than 90% of clinical isolates from patients with cryptococcosis are strains of serotype A . Strains of serotype D are also found globally , but they are more prevalent in Europe [14] . The clinical manifestations of human infections with serotype A or D appear to be similar , but experimental infections suggest that strains of serotype A are more virulent than strains of serotype D [8 , 15] . AD strains are hybrids of the two varieties . Whereas most isolates of serotypes A and D are haploid , AD strains are diploid or aneuploid , contain two sets of chromosomes , and possess two mating type alleles , one from each of the two serotype A and D haploid genomes [16–18] . Recent reports suggest that hybrid AD strains may be more common in clinical samples then previously appreciated . For example , a prospective survey of cryptococcosis in Europe from 1997 to 2001 found that up to 30% of all isolates of C . neoformans from patients in Europe were AD hybrids [14] . Strains of both serotype A and serotype D , as well as AD hybrids , are found in the environment , where they are primarily associated with avian feces . Our recent analysis of environmental and clinical populations of C . neoformans in North America revealed that approximately 7 . 5% of strains isolated from the environment are AD hybrids [19] . C . neoformans has a bipolar mating system with two alternative mating type alleles , MATa or MATα . When sexual reproduction is induced in the laboratory , haploid MATa and MATα strains of serotype A or serotype D are capable of plasmogamy , karyogamy , and meiosis , during which they produce dikaryotic hyphae , basidia , and chains of haploid basidiospores . Although most naturally occurring AD hybrid strains are incapable of mating with strains of the opposite mating type , some AD hybrid strains are self-fertile . That is , when stimulated by growth on mating medium in the absence of a mating partner , they produce hyphae , basidia , and basidiospores [16 , 20] . Most of the basidiospores produced by these self-fertile AD hybrid strains fail to germinate , which suggests that meiosis is impaired in these hybrids . However , approximately 5% of these spores germinate to produce viable diploid , aneuploid , and rarely , haploid cells [16 , 20 , 21] . In addition , postzygotic reproductive isolation of serotype A and serotype D is supported by phylogenetic analyses of multiple gene genealogies , which is consistent with the monophyletic origins of both serotypes [11 , 12 , 22] , and by a comparative genomics analysis of representative strains of serotype A and serotype D , which confirmed that recombination is rare between these serotypes [23] . Clinical and environmental populations of both serotype A and serotype D are dominated by isolates with the MATα mating type . Isolates of serotype D with the rare MATa mating type have been known for some time [8] , and recently , a few strains of serotype A with the MATa allele have been isolated from Tanzania [24 , 25] , Italy [26] , Hungary [27] , and Botswana [28] . However , with the exception of the unique subpopulation of serotype A in Botswana ( designated VNB strains ) , in which 25% of the isolates possessed the MATa allele , the prevalence of the MATa allele among global isolates of serotype A is less than 0 . 1% [22] . Although the MATa allele is extremely rare among global isolates of serotype A , the MATa serotype A allele is common among AD hybrids . Several reports indicate that from 20% to 80% of AD hybrids contain the AaDα genotype [16–19 , 29] . At least two hypotheses have been proposed to explain the rarity of serotype A MATa allele among strains of serotype A but its high prevalence in serotype AD hybrids [16 , 19 , 24]: ( i ) nearly all serotype A isolates with the MATa allele are extinct , and AD hybrids represent historic evidence that strains with the MATa allele were once more abundant; and ( ii ) isolates of serotype A with the MATa allele are extant , but they are geographically isolated , exist in a different natural reservoir , and/or are nonpathogenic; consequently , they are rarely recovered from environmental samples or clinical specimens . Our recent discovery of the VNB subpopulation of serotype A in Botswana , which includes a significant number of isolates with the MATa allele [22 , 28] , invokes an alternative hypothesis about the origin of AaDα hybrids . These hybrids may have originated from sexual encounters between one or more strains of serotype D and one or more VNB strains of serotype A with the MATa allele . The progeny of this union may have enjoyed increased fitness compared to the parental strains , enabling them to expand clonally and disseminate beyond Africa . To test this hypothesis , we investigated whether serotype A MATa alleles in non-Botswanan AD hybrids could have originated from the VNB subpopulation of serotype A , and whether hybrid strains are fitter than strains of serotype A or serotype D . To ascertain the distribution of the mating type alleles in the AD strains , serotype-specific PCR primers were used to amplify fragments of STE20 gene situated within the MAT locus [25] . Five strains , isolated from China , Italy , Kuwait , and the United States of America , generated characteristic products with the PCR primers specific to the serotype A STE20 MATα allele ( Figure 1C ) and serotype D STE20 MATa allele ( Figure 1A , lanes 2 , 5–7 , and 10 ) , and therefore were identified as AαDa strains ( Table 1 ) . Conversely , six strains isolated from China , Italy , and the US produced characteristic amplicons with the PCR primers specific to the serotype A STE20 MATa allele ( Figure 1B ) . A pair of PCR primers that were mating-type-specific but non-serotype-specific amplified the STE20 MATα gene from all of the strains ( Figure 1D ) . The DNA sequences of these fragments were compared with the available sequences of STE20 genes for both serotype A [30] . As expected , PCR products obtained from four strains from the US ( CDC228 , CDC304 , MMRL1365 , and nc34–21 ) , one strain from China ( ZG290 ) , and one strain from Italy ( it752 ) were identical to the reference serotype D STE20α sequence , while they exhibited 5% nucleotide polymorphisms with the reference serotype A STE20α sequence ( Table 2 ) . Thus , they were identified as AaDα strains . An environmental isolate from the US ( nc5–19 , Figure 1 , lane 4 ) contained two copies of the STE20α allele , one from serotype D and the other from serotype A , and therefore had the unusual AαDα genotype , which is consistent with our previous observations ( [19] and X . Lin , A . Litvintseva , K . Nielsen , S . Patel , A . Floyd , et al . , unpublished data ) . To examine the genetic relationships among the 12 AD strains , amplified fragment length polymorphism ( AFLP ) genotyping was conducted with two independent primer pairs [28] . Eighteen polymorphic bands were generated , which differentiated six unique AFLP genotypes . As shown in Figure 2 , AFLP band patterns of the AaDα strains are more similar to each other than to the AαDa strains: at least five distinct AFLP bands are characteristic of either one of the groups . Two different genotypes were observed among AaDα strains , and three genotypes were identified among AαDa strains . The AFLP banding pattern of the unusual AαDα genotype ( strain nc5–19 ) was similar to that of the AαDa strains . Previously , we developed a 12-gene multi locus sequence typing ( MLST ) genotyping system that allows unambiguous genotyping of strains of serotype A; however , this system is impractical for the AD hybrid population , because AD hybrids contain two copies of each loci , both of which are PCR-amplified using our MLST primers [22] . Consequently , we developed new PCR primers that selectively amplified serotype A alleles of three genes , CAP10 , URE1 , and GPD1 , and generated amplicons of 410 , 810 , and 220 base pairs , respectively . We used these primer pairs to amplify DNA from the 12 AD hybrids and a global sample of 45 strains of serotype A , which represent unique genotypes in a previously described sample of 1 , 057 global isolates of serotype A ( Table 1 , [22] ) . DNA sequences of these genes were determined and subjected to phylogenetic analyses ( Figure 3 ) . For the CAP10 locus , 308 nucleotides were alignable , and five were phylogenetically informative; for the URE1 locus , 605 nucleotides could be aligned , and seven were phylogenetically informative; and for the GPD1 locus , 123 nucleotides were aligned , and three were phylogenetically informative . Bayesian consensus trees were constructed for CAP10 , GPD1 , and URE1 ( unpublished data ) , as well as the combined data for all three loci ( Figure 3 ) , and rooted with JEC21 strain of serotype D . Gene genealogies of all three loci and the combined genealogy were congruent , and the phylograms recognized the three major subpopulations of serotype A: VNI , VNII , and VNB [22] . The VNI group includes the majority of global isolates of serotype A , as well as three previously identified strains of serotype A with the MATa mating type , 125 . 91 [25] , IUM 96-28-28 [26] , and bt130 [28] . The VNII group has a small number of strains , and all have the MATα allele . The VNB group contains most of the known isolates of serotype A with the MATa allele [22] . DNA sequences of the CAP10 , GPD1 , and URE1 genes were identical among the six AaDα strains . The DNA sequences of these genes were also identical among six AD strains that possess the MATα serotype A allele , AαDa and AαDα . In all of these phylograms , AαDa and AαDα strains cluster with the VNI group , whereas the AaDα strains cluster with the VNB group that contains most of the serotype A MATa isolates ( Figure 3 ) . We have not analyzed the origin of the serotype D counterpart in these AD hybrids because an MLST genotyping scheme for serotype D is not yet available . In addition , serotype D isolates are rare in the clinical samples and in the collection , which further complicates this analysis . Previous analyses of AFLP and MLST genotypes of numerous strains of serotype A showed that the VNB subpopulation is geographically confined to Botswana [22] . Yet , we have shown here that AaDα hybrid strains with the VNB genome are distributed globally . One possible explanation for this apparent paradox is that hybridization produces hybrid strains with increased fitness that are better able to tolerate stress and/or propagate in the environment . Most strains of C . neoformans are highly sensitive to direct sunlight and temperatures above 38 °C , and both conditions exist in much of Botswana . To test the hypothesis of hybrid vigor associated with the AD hybrid genome , we compared the effects of elevated temperature and exposure to UV irradiation on the growth of haploid VNB strains and AD hybrids . To compare fitness among strains of serotypes A , D , and AD hybrids with the same genetic background , AD hybrids were constructed in the laboratory by fusing ura5 mutants of bt65 and bt88 strains of serotype A from Botswana , and with an ade2 mutant of JEC21 strain of serotype D . Spontaneous ura5 mutants of bt65 and bt88 were obtained by selection on 5-fluoroorotic acid medium , which inhibits strains with the functional URA5 gene [31 , 32] . These ura5 mutants were coincubated with the ade2 mutant of JEC21 [33] on V8 medium ( X . Lin , A . Litvintseva , K . Nielsen , S . Patel , A . Floyd , et al . , unpublished data ) . Three prototrophic hybrid strains , designated XL1595 , XL1596 , and XL1597 , were obtained and confirmed to be diploid by fluorescence-activated cell sorting ( FACS ) analysis . Figure 4 illustrates that VNB isolates of serotype A are more sensitive to UV irradiation than laboratory-generated AD hybrids , naturally occurring AD hybrids , or control haploid strains of serotype A ( H99 ) and serotype D ( JEC21 ) . UV irradiation for 24–30 seconds almost completely inhibited the growth of VNB strains and dramatically impaired growth of haploid control strains of serotype A ( H99 ) and serotype D ( JEC21 ) ; however , a significant number of AD hybrid cells survived this treatment ( Figure 4A ) . The differences in survival between AD hybrids and haploid strains of serotypes A and D were statistically significant ( p < 0 . 01 , Figure 4B ) . Several lines of evidence suggest that C . neoformans is melanized in the natural environment [34 , 35] , and the amount of melanin is known to affect susceptibility of the fungus to environmental stress [36–38] . Therefore , UV irradiation experiments were performed under melanin-inducing conditions . Figure 4A shows that resistance to UV irradiation is not dependent upon the amount of melanin produced by each strain . Although most AD hybrids produced less melanin than haploid strains bt65 , bt109 , and H99 ( Figure 4A , right panel ) , they were significantly more resistant to UV irradiation ( Figure 4A , middle panels ) . Furthermore , the only heavily melanized hybrid , CDC228 , was highly susceptible to UV radiation ( Figure 4A , lower right panel ) . Recently , we demonstrated that laboratory-generated hybrids with the AαDα genotype are more resistant to elavated temperature than either haploid parent strain ( X . Lin , A . Litvintseva , K . Nielsen , S . Patel , A . Floyd , et al . , unpublished data ) . Here , we compared the growth of AaDα hybrids at 40 °C . All of the hybrid strains tested , including laboratory-constructed ( XL1595 , XL1596 , XL1597 ) and wild-type ( ZG290 and nc34–21 ) hybrid strains , were capable of growing at 40 °C ( unpublished data ) . Two of the haploid strains , JEC21 and bt65 , were severely inhibited at 40 °C , whereas strain bt88 was not affected and survived well at 40 °C ( unpublished data ) , which indicates that growth at high temperature can vary among VNB strains . All strains grew equally well on culture medium that contained pigeon excreta [35] as the sole source of carbon and nitrogen ( unpublished data ) . The high prevalence of the MATa serotype A allele among the AD hybrid strains has been puzzling because this allele is uncommon in the global population of serotype A . To investigate the origin of these strains , we designed serotype A–specific PCR primers that amplify three genomic regions situated on three different chromosomes of C . neoformans ( CAP10 , GPD1 , and URE1 ) and constructed gene genealogies . These gene genealogies were congruent and revealed that both AαDa and AαDα hybrid strains cluster with isolates of the VNI group of serotype A , which is dominated by isolates with the MATα mating type . In contrast , AaDα strains cluster with the VNB group from Botswana , which contains a significant proportion of strains with the MATa allele [19] . Our results confirm data presented by Xu et al . [39] , who used LAC1 and URA5 gene sequences to investigate phylogenetic relationships among serotypes A , D , and AD . Their LAC1 gene phylogeny revealed that AD hybrids are separated into two clusters: one included AD strains possessing the MATa serotype A allele , whereas the other included strains of serotype A and AD hybrids possessing the MATα serotype A allele . Their URA5 phylogram confirmed these clusters and also provided evidence of recombination between them , which is consistent with our observations [22 , 28] . The unusual clinical sample of serotype A isolates from Botswana consists of two genetically isolated subpopulations , VNI-Botswana and VNB [28] . Both groups are genetically unique and characterized by unusually high genotypic diversity . In addition , the VNB group contains an unprecedented proportion of fertile isolates with the MATa mating type and exhibits phylogenetic and population genetic evidence of recombination [22 , 28] . Here , we demonstrate that DNA sequences of the CAP10 , GPD1 , and URE1 genes in a global collection of AaDα hybrid strains are identical to those of two strains from Botswana , bt65 and bt88 ( Figure 3 , Table 1 ) , which possess the MATa mating type and are fertile in the laboratory [28] . Therefore , it is likely that these , or related Botswanan strains , represent one of the progenitors of the many existing AaDα hybrids . DNA sequences of the CAP10 , GPD1 , and URE1 genes were identical among the AD strains with the MATa serotype A allele ( AaDα ) , as well as among the six AD strains that possess the MATα serotype A allele ( AαDa and AαDα ) , which may indicate that the AD hybrids analyzed here are clonal descendents of only three ancestral strains with the AaDα , AαDa , and AαDα mating types . However , it is also possible that the loci we used here are insufficiently polymorphic to differentiate among these related strains . For example , we previously analyzed 12 MLST loci in over 100 global isolates of serotype A and demonstrated that two possible parental strains , bt65 and bt88 , have closely related , but nevertheless distinct , genotypes . Similarly , this analysis of three MLST loci did not reveal the phylogenetic structure among the VNI isolates that we previously observed using 12 MLST loci and a larger sample of serotype A strains [22] . AaDα and AαDa strains are products of hybridization between isolates of opposite mating types ( MATa and MATα ) , whereas AαDα is an apparent product of mating between isolates of the same mating type , which was recently described in C . neoformans ( [40] and X . Lin , A . Litvintseva , K . Nielsen , S . Patel , A . Floyd , et al . , unpublished data ) . Genetic diversity among the isolates was also estimated by using AFLP , and the phylogenetic analyses based on these data demonstrated that the AaDα strains are more closely related to each other than to strains with the AαDa or AαDα genotypes ( Figure 2 ) . In addition , AFLP analysis identified two distinct AFLP genotypes among the AaDα isolates , and three distinctive AFLP genotypes among the AαDa isolates . The only AαDα isolate ( nc5–19 ) had a unique AFLP genotype that was related to the AαDa strains ( Figure 2 ) . Both AFLP and MLST analyses suggest at least two , non-exclusive explanations for these data: Strains with identical MLST but distinct AFLP genotypes may have originated from multiple hybridization events between different serotype A and serotype D strains . Conversely , they may have originated from a single hybridization event , but their AFLP genotypes changed following evolution over subsequent generations . For example , some of the AD strains are diploid and others are aneuploid [16] . Therefore , different AFLP genotypes may reflect the loss of chromosomes after the initial hybridization or mitotic recombination and homozygosis . Because both populations of serotype A isolates and AD hybrids exhibit a high level of clonality , contemporary AD strains may be the descendents of a limited number of ancestral AD hybrid strains that proliferated clonally , or may have resulted from multiple hybridization events between clonal strains of serotype A and serotype D . MLST analyses using more loci and careful examination of the karyotypes of these strains are necessary to resolve this question . Our data indicate that serotype A isolates from the VNB group are geographically restricted to sub-Saharan Africa , whereas AaDα strains that inherited their serotype A genomes from the VNB group are distributed globally . Geographic isolation may have contributed to the unique genetic composition of this Botswanan population . For example , a major portion of Botswana is the Kalahari Desert , which is noted for its remarkably hot , arid climate and high levels of UV radiation . These conditions are likely to deter the spread of VNB strains . This hypothesis is supported by our experiments demonstrating that VNB strains are hypersensitive to UV irradiation and that global AaDα strains of VNB origin are more resistant to UV irradiation and in some cases are more thermotolerant . Several animal and plant hybrids have been shown to exhibit greater fitness than their parental genotypes in novel or perturbed habitats [1 , 2] . There are also several cases of the increased fitness of hybrid genotypes among phytopathogenic fungi and oomycetes [41 , 42] . For example , rapid expansion of interspecies hybrids of the rust fungi , Melampsora medusae and M . accidentalis , coincided with the introduction of a new poplar host to the habitat for these fungi , which suggests that these fungal hybrids are better adapted to this new host [7 , 41] . Similarly , the transmission of lethal infections of alders in Europe has been attributed to the emergence of heteroploid hybrids between Phytophthora cambivora and P . fragariae [4] . Therefore , it is possible that the abundance of the VNB AaDα hybrid strains of C . neoformans in the global population and scarcity of the VNB haploid strains outside Botswana may be attributed to the increased fitness of the hybrids , which enhanced their ability to propagate and disseminate to new environments . Other explanations for the worldwide distribution of the hybrid strains are also possible . For example , global spread of AD hybrid strains from Botswana may be attributed to human activity and the creation of a novel habitat that reduced competition with the parental species . In particular , an association between C . neoformans and pigeons ( Columbia livia ) is well documented [8 , 19] . Pigeons were introduced into southern Africa during the period of 1500 to 1700 [43] . The increased availability of pigeon excreta , and perhaps the excreta of other domesticated animals , which are rich substrates for the growth of C . neoformans , may have contributed to the global expansion of AD hybrid strains . However , since our experiments demonstrated that all strains grow well on pigeon excreta in the laboratory , other factor ( s ) may yet be discovered that favor hybrid over haploid VNB strains . It is also possible that VNB isolates will be discovered outside of Africa . For example , if VNB strains were historically more abundant and only recently became restricted to Botswana , hybridization might have occurred outside sub-Saharan Africa . Future analyses of the population structure of the C . neoformans species complex will investigate this possibility . This investigation confirmed previous observations that approximately half of the global isolates of AD hybrids contain the rare MATa serotype A allele ( [16 , 18 , 39] and X . Lin , A . Litvintseva , K . Nielsen , S . Patel , A . Floyd , et al . , unpublished data ) . We applied methods of AFLP genotyping and phylogenetic analysis of three genes to demonstrate that many AaDα strains possess serotype A genomes from the Botswanan population . We also demonstrated that AD hybrid strains are more resistant to UV irradiation and in some cases are more thermally resistant . These data suggest two hypotheses: ( i ) AaDα strains probably originated in sub-Saharan Africa from an intervarietal hybridization between isolates of serotype A from the VNB subpopulation and isolates of serotype D; and ( ii ) this hybridization resulted in increased fitness of the hybrid strain ( s ) , which allowed it to colonize new ecological niches , spread beyond Africa , and populate the world . Twelve AD hybrid strains isolated from four countries were analyzed and compared with a subset of 45 previously described strains of serotype A ( Table 1 ) [22] . Isolates were maintained on yeast extract-peptone-dextrose ( YPD ) agar medium ( Difco , http://www . bd . com/ds/ ) at 30 °C . Genomic DNA was extracted from each isolate and the AFLP analysis was performed as described [28] . Only pronounced and reproducible bands were scored for the analyses of population structure . Polymorphic AFLP bands were defined as bands of the same size that were present in some , but not all , isolates . To assess the reproducibility of the AFLP method , DNA was extracted and the AFLP reactions and analyses were performed on at least two separate occasions for each isolate . In comparing replicate analyses , 92% of the AFLP bands were identical ( unpublished data ) . We previously developed MLST genotyping for serotype A strains using 12 loci ( http://www . cgrubii . mlst . net/ ) . However , for this investigation , we did not amplify and sequence all 12 MLST loci in the AD hybrids , because unlike strains of serotype A or D , which are haploid , the AD hybrids are diploid , and the primers amplified both serotype A and D alleles of the loci . To overcome this problem , we developed new serotype A–specific primers for three of the MLST loci: CAP10 , GPD1 , and URE1 ( PCR primers and conditions are listed in Table 3 ) . Each PCR mixture contained 32 μl of 1X PCR buffer , 2 mM MgCl2 , 0 . 2 mM dNTPs , 1 μM each primer , 0 . 065 μl of iTaq DNA Polymerase ( Bio-Rad , http://www . bio-rad . com/ ) , and approximately 1 ng of genomic DNA . PCR products were purified using the QIAquick PCR purification kit ( Qiagen , http://www . qiagen . com/ ) and sequenced using an ABI 3700 sequencer with Big Dye terminators ( Applied Biosystems , http://www . appliedbiosystems . com/ ) . PCR primers used for the amplification of the fragments were also used for sequencing . Mating type– and serotype-specific primers were used to confirm the mating type alleles of the AD hybrid strains . The primer pairs , which amplify portions of the STE20a or STE20α alleles of serotype A or D , are designated STE20Aa , STE20Aα , and STE20Da , respectively [25] . Since the primers that were previously designed to amplify a serotype D–specific portion of the STE20α gene failed to generate any product in the AD strains , we designed new PCR primers specific to a more conserved region of the STE20α gene of serotype D: STE20α-F: 5′-GTAAGTGCAAAGGACCCATCTC; and STE20α-R: 5′-TGATCC CCAAAGACCAAATATC . The PCR conditions for amplification were as follows: 94 °C for 5 min , followed by 30 cycles of 94 °C for 30 sec , 52 °C for 30 sec , and 72 °C for 1 min , followed by 7-min extension at 72 °C ( PCR reactions were the same as those described for MLST ) . DNA sequences of the corresponding STE20α portions were obtained and compared with the DNA sequences in the typing strains , JEC21 , serotype D [30] , and H99 , serotype A ( C . neoformans H99 sequencing project , Duke Institute for Genome Sciences and Policy , Center for Applied Genomics and Technology , http://cgt . duke . edu/ ) . Sequences were automatically aligned using Sequncher 4 . 1 ( Gene Codes Corporation , http://www . genecodes . com/ ) ; the alignment was imported into MacClade 4 . 05 [44] and manually edited . Ambiguously aligned characters and gaps were excluded from the analysis . Optimal phylogenetic trees were constructed using maximum parsimony ( MP ) and Bayesian approaches . MP trees for the individual loci and for the combined data set were identified with heuristic searches based on 500 random sequence additions for each data set and implemented in PAUP version 4 . 0b10 [45] . Bayesian inferences were performed with MrBayes version 3 . 0B4 using a General Time Reversible model with a proportion of invariable sites and a gamma-shaped distribution of rates across sites ( GTR+I+G ) model of evolution [46] . Each Bayesian analysis consisted of two runs of 1 , 000 , 000 generations , each using the default uniform priors , and a sample frequency of 100 . Likelihood scores of each sampled generation were plotted by using Excel ( Microsoft , http://www . microsoft . com/ ) and visually analyzed; phylograms collected before the stationary phase was reached were discarded [47] . The phylograms remaining from both runs were combined , and a majority-rule consensus phylogram of each gene and for the combined data from all three genes were generated using PAUP and compared for topological congruence . To construct AaDα hybrid strains in the bt65/bt88 and JEC21 background , ura5 mutants of strains bt65 and bt88 were obtained . Wild-type strains bt65 and bt88 were grown in liquid YPD , washed three times in sterile distilled water , and plated on 5-fluoroorotic acid ( 5-FOA ) medium , which inhibits cells with a functional URA5 gene [32] . 5-FOA-resistant colonies were selected and tested on the yeast nitrogen base minimal medium ( YNB , Difco ) . Isolates that showed no growth on YNB ( Aa ura5 ) were mixed with cells of the XL342 ( Dα ade2 ) strain of serotype D [33] and incubated on V8 agar medium in the dark at 22 °C . After coincubation for 24 h , cells were collected , washed , and spread on YNB minimal medium to select for prototrophic fusion products . The ploidy of prototrophic AaDα strains was confirmed by FACS analysis . Yeast cells were processed for flow cytometry as described previously [48] . Briefly , cells were harvested from YPD medium , washed once in PBS buffer , and fixed in 1 ml of 70% ethanol overnight at 4 °C . Fixed cells were washed once with 1 ml of NS buffer ( 10 mM Tris-HCl [pH 7 . 6]; 250 mM sucrose; 1 mM EDTA [pH 8 . 0]; 1 mM MgCl2; 0 . 1 mM CaCl2; and 0 . 1 mM ZnCl2 ) and then stained with propidium iodide ( 10 mg/ml ) in 0 . 2 ml of NS buffer containing RNaseA ( 1 mg/ml ) at 4 °C for 4–16 h . Then 0 . 05 ml of the stained cell preparation was diluted into 2 ml of 50 mM Tris-HCl ( pH 8 . 0 ) and sonicated for 1 min . Flow cytometry was performed on 10 , 000 cells and analyzed on the FL1 channel of a Becton-Dickinson FACScan ( http://www . bd . com/ ) . The following strains were tested for their sensitivity to UV irradiation: ( i ) six wild-type AaDα hybrid strains; ( ii ) three AaDα hybrids obtained in the laboratory by fusing bt65/bt88 and JEC21 strains; ( iii ) four strains of serotype A from the VNB subpopulation , including the two strains ( bt88 and bt65 ) that were most closely related to the probable serotype A progenitor of AaDα strains; and ( iv ) the representative haploid strain of serotype A ( H99 ) and representative haploid strain of serotype D ( JEC21 ) . All strains were grown to logarithmic phase at 30 °C with constant agitation in the dark in low-glucose medium ( 0 . 1% glucose , 1 . 3 mM glycine , 2 . 2 mM KH2PO4 , 0 . 1 mM MgSO4-7H2O , 0 . 3 μM thiamine , 2 nM biotin , [pH 5 . 6] ) supplemented with DL-3 , 4-dihydroxyphenylalanine- ( DL-DOPA , 20mg/l ) . Yeast cells were counted in a hemocytometer , all cultures were adjusted to the same cell concentration ( 5 . 5 × 107 CFU/ml ) , serial 10-fold dilutions were prepared , and 2 . 0 μl of each strain was spotted onto YPD agar plates ( to assess sensitivity to the UV irradiation ) and low-glucose agar plates supplemented with 200 mg/l DL-DOPA ( to assess melanin production ) . The YPD plates were then irradiated with UV light at approximately 48 mJ/cm2 ( 24- and 30-second settings , UV Stratalinker 1800; Stratagene , http://www . stratagene . com/ ) [49] , and all plates were incubated at 30 °C for 72 h . To provide a quantitative assessment of UV resistance for each strain , the highest dilutions that yielded colonies after treatment with UV were identified . For such dilution , the number of individual colonies within a spot on the YPD plate was counted . The GenBank ( http://www . ncbi . nlm . nih . gov/Genbank/index . html ) accession numbers for the STE20 genes discussed in this paper are H99 serotype A ( AF542529 ) and JEC21 serotype D ( AF542531 ) . MLST genotyping for serotype A strains using 12 loci is under accession numbers DQ212527–DQ212692 . Unique MLST sequence types were deposited under accession numbers EF625826–EF625831 .
Hybridization between individuals of different species or varieties is common among fungi . However , the impact of hybridization on the evolution of pathogenic fungi is unresolved . Several hybrids of phytopathogenic fungi exhibit expanded host ranges . To our knowledge , this report is the first description of increased hybrid fitness ( hybrid vigor ) in a human pathogen , Cryptococcus neoformans , the most prevalent cause of fungal meningitis . We demonstrate that diploid hybrid strains are common among both environmental and clinical isolates of two varieties , represented by serotypes A and D . We determined that many globally isolated AD hybrid strains originated in sub-Saharan Africa and have increased resistance to ultraviolet radiation . We hypothesize that hybrid strains have increased fitness , which enabled them to emigrate from Africa and spread globally .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods", "Supporting", "Information" ]
[ "yeast", "and", "fungi", "microbiology", "evolutionary", "biology", "molecular", "biology" ]
2007
Many Globally Isolated AD Hybrid Strains of Cryptococcus neoformans Originated in Africa
Netrin is a key axon guidance cue that orients axon growth during neural circuit formation . However , the mechanisms regulating netrin and its receptors in the extracellular milieu are largely unknown . Here we demonstrate that in Caenorhabditis elegans , LON-2/glypican , a heparan sulfate proteoglycan , modulates UNC-6/netrin signaling and may do this through interactions with the UNC-40/DCC receptor . We show that developing axons misorient in the absence of LON-2/glypican when the SLT-1/slit guidance pathway is compromised and that LON-2/glypican functions in both the attractive and repulsive UNC-6/netrin pathways . We find that the core LON-2/glypican protein , lacking its heparan sulfate chains , and secreted forms of LON-2/glypican are functional in axon guidance . We also find that LON-2/glypican functions from the epidermal substrate cells to guide axons , and we provide evidence that LON-2/glypican associates with UNC-40/DCC receptor–expressing cells . We propose that LON-2/glypican acts as a modulator of UNC-40/DCC-mediated guidance to fine-tune axonal responses to UNC-6/netrin signals during migration . Directed migrations of developing axons are essential for the proper wiring of the nervous system . A host of guidance cues and their receptors instruct axon guidance decisions . However , how these cues and the growth cone’s responses to them are spatially and temporally regulated in vivo remains largely unknown . Answering this question is central to our understanding of how growing axons navigate in complex environments to reach their targets during development and regeneration . UNC-6/netrin is a highly conserved secreted guidance cue with structural similarity to the extracellular matrix protein laminin [1–3] . UNC-6/netrin directs attractive guidance through receptors of the UNC-40/DCC family and repulsive guidance through both UNC-40/DCC and UNC-5/UNC5 receptors [4–6] . Notably , whereas netrin receptors and downstream transduction pathways have been well characterized , how netrin signals are regulated extracellularly remains largely unknown . UNC-6/netrin was identified through genetic analysis in Caenorhabditis elegans [1] and biochemically purified and cloned from vertebrate embryos [2] . A second biochemical component that synergized with netrin to elicit axon outgrowth was termed “netrin synergizing activity” ( NSA ) [3] and remains unidentified . Vertebrate netrin-1 and its receptor DCC can bind heparin , a fully sulfated version of heparan sulfate ( HS ) , in vitro [3 , 7 , 8] , and a general disruption of HS chain synthesis is detrimental to netrin-1-mediated axon outgrowth in vitro [9 , 10] . While heparan sulfate proteoglycans ( HSPGs ) might be intriguing candidates for NSA , it is not yet known whether a specific HSPG is required for netrin signaling or how interactions with HSPGs might regulate netrin signals to direct axons during nervous system development . We addressed these questions using the nematode C . elegans , which has been instrumental for discovering major conserved axon guidance pathways . During larval development , the axon of the mechanosensory neuron AVM migrates ventrally as its growth cone integrates signals from two complementary guidance cues ( Fig 1A ) [1 , 4–6 , 11–13]: ( 1 ) UNC-6/netrin is secreted at the ventral midline and attracts the growth cone ventrally via the receptor UNC-40/DCC [5 , 14] , and ( 2 ) SLT-1/Slit is secreted by the dorsal muscles and repels the growth cone away from the dorsal side via the receptor SAX-3/Robo [12 , 13] . Animals null for the guidance cues unc-6/netrin or slt-1/Slit exhibit partial AVM ventral axon guidance defects , and loss of both cues in unc-6 slt-1 double mutants results in fully penetrant guidance defects ( S1 Fig , [13] ) . AVM axons defective in guidance fail to extend ventrally and instead migrate laterally in the anterior direction ( Fig 1 ) . In this study , we use the AVM axon as a model to elucidate mechanisms that regulate UNC-6/netrin signaling . Here we provide a missing link in understanding the modulation of UNC-6/netrin signaling in the extracellular milieu . We demonstrate that LON-2/glypican , a HSPG secreted from epidermal cells , acts as a modulator of the UNC-6/netrin signaling pathways to guide migrating cells and axons . We show that LON-2/glypican modulates UNC-6/netrin signaling in both attractive guidance mediated by the UNC-40/DCC receptor and repulsive guidance mediated by the UNC-40/DCC and UNC-5/UNC5 receptors . We provide evidence that LON-2/glypican associates with UNC-40/DCC-receptor-expressing cells . We show that the N-terminal globular region of LON-2/glypican , lacking the three HS chain attachment sites , is functional in UNC-6/netrin-mediated guidance . Our studies unravel a novel mechanism by which LON-2/glypican is produced by substrate epidermal cells and released from the membrane to likely associate with UNC-40/DCC-expressing neurons , enabling the modulation of their responses to UNC-6/netrin during axon migrations . To address whether a specific HSPG interacts with the netrin signaling system to guide axons , we first examined axon guidance in mutants lacking core HSPGs . HSPGs are composed of a core protein with covalently attached long unbranched HS chains [15] . HSPGs can be associated with the plasma membrane through either a transmembrane domain ( e . g . , syndecans ) or a glycerophosphatidylinositide ( GPI ) anchor ( e . g . , glypicans ) or be secreted into the extracellular milieu ( e . g . , perlecans and agrins ) . We examined the axon morphology of AVM in single , double , and triple mutants for several core HSPG proteins ( see S1 Table for alleles ) . These included the sole C . elegans syndecan ( sdn-1 ) , the two glypicans ( lon-2 and gpn-1 ) , perlecan ( unc-52 ) , and agrin ( agr-1 ) . We found that the mild AVM axon guidance defects of sdn-1/syndecan mutants , including a null , [16] were enhanced by the complete loss of lon-2/glypican in double mutants lon-2 sdn-1 ( Fig 1B ) , revealing a role for lon-2/glypican in AVM axon guidance . Similarly , loss of lon-2/glypican enhances sdn-1/syndecan mutants in motorneuron guidance [17] . Although the C . elegans genome encodes two glypicans , loss of function of the second glypican , gpn-1 , using two likely null mutant alleles ( see S2 Fig ) , did not enhance the defects of lon-2/glypican or sdn-1/syndecan null mutants in double or triple mutants ( Fig 1B ) . Moreover , we did not observe abnormal phenotypes in the single mutants for agr-1/agrin or unc-52/perlecan . These observations highlight the specificity of lon-2/glypican function in this axon guidance process and raise the possibility that lon-2/glypican might be a component of the pathways guiding the AVM axon towards the ventral midline . Considering that AVM axon guidance occurs via the unc-6/netrin and slt-1/Slit pathways , mutations in genes such as lon-2/glypican and sdn-1/syndecan that affect AVM axon guidance may point towards interactions with either of these two guidance systems . Since the AVM axon guidance defects in lon-2 sdn-1 double mutants are qualitatively similar to those of mutants lacking unc-6/netrin or slt-1/Slit , we determined how lon-2/glypican and sdn-1/syndecan impact unc-6/netrin and slt-1/Slit signaling . In animals that completely lack slt-1/Slit function , the complete loss of a gene functioning independently of slt-1/Slit is expected to enhance the AVM guidance defects , such as in the double null mutants unc-6/netrin slt-1/Slit ( see S1 Fig ) . We tested the interactions of lon-2/glypican with the slt-1/Slit pathway in AVM axon guidance and found that the complete loss of lon-2/glypican enhanced a presumptive null allele of slt-1/Slit in lon-2 slt-1 double mutants ( Fig 1C ) , suggesting that lon-2/glypican functions in a pathway separate from slt-1/Slit . Loss of lon-2/glypican also enhanced guidance defects when signaling through sax-3/Robo , the slt-1/Slit receptor , was disrupted in lon-2 sax-3 double null mutants , providing further evidence that lon-2/glypican functions in a pathway separate from that of slt-1/Slit ( Fig 1C ) . As an additional method to investigate the impact of lacking lon-2/glypican function when slt-1/Slit signaling is perturbed , we used a transgene that ectopically expresses slt-1/Slit from both ventral and dorsal body wall muscles ( using Pmyo-3::slt-1 ) and misguides the axon of AVM [18] . Loss of lon-2/glypican enhanced the defects caused by slt-1/Slit misexpression ( Fig 1C ) , consistent with the above findings that lon-2/glypican mediates its axon guidance effects independently of slt-1/Slit . The unc-6/netrin pathway functions independently of slt-1/Slit to guide AVM . To address whether lon-2/glypican functions in the unc-6/netrin axon guidance pathway , we examined the AVM axon in double mutants of lon-2/glypican and unc-6/netrin . In animals that completely lack unc-6/netrin function , the complete loss of a gene functioning in the same unc-6/netrin pathway is expected to not enhance the AVM guidance defects , such as in the double null mutants unc-6; unc-40 ( see S1 Fig ) . We found that the complete loss of lon-2/glypican did not enhance the guidance defects displayed by unc-6/netrin null mutants ev400 ( Fig 1D ) . Given that loss of lon-2 enhances the defects of other guidance mutants ( see doubles with sdn-1 , slt-1 , sax-3 , and Pmyo-3::slt-1 in Fig 1B and 1C and sqv-5 in S3 Fig ) , the lack of enhancement when combined with the unc-6 null mutation suggests that lon-2/glypican functions in the same pathway as unc-6/netrin . Consistent with this idea , we also found that complete loss of lon-2/glypican did not enhance the AVM guidance defects of two null mutant alleles of the netrin receptor unc-40/DCC in the double mutants unc-40; lon-2 ( Fig 1D ) , suggesting that lon-2/glypican functions in the same pathway as unc-40/DCC in AVM ventral guidance . These observations raise the interesting possibility that lon-2/glypican may be the HSPG dedicated to modulate unc-6/netrin signaling through unc-40/DCC during axon guidance . Since lon-2/glypican functions independently of slt-1/Slit ( Fig 1C ) and partly separate from sdn-1/syndecan ( Fig 1B ) , we tested whether sdn-1/syndecan and slt-1/Slit function together to guide the axon of AVM . We found that defects in slt-1 sdn-1 double null mutants were not enhanced compared to the single mutants ( Fig 1E ) , consistent with findings in Drosophila [19 , 20] and C . elegans [16] . We also found that double null mutants for sdn-1/syndecan and the slt-1/Slit receptor sax-3/Robo were not enhanced compared to the single mutants ( Fig 1E ) . Our results support the notion that sdn-1/syndecan acts in the same genetic pathway as slt-1/Slit to guide AVM . Consistent with this , we found that the double null mutants for sdn-1/syndecan and the netrin receptor unc-40/DCC were enhanced , indicating that sdn-1/syndecan functions in a pathway separate from unc-6/netrin . The analysis of axon guidance in double mutants of unc-6/netrin and sdn-1/syndecan was precluded by their lethality . Our results are consistent with the notion that unc-6/netrin and sdn-1/syndecan act in different pathways of axon guidance . In addition to unc-6/netrin acting as an attractive cue for cells expressing the unc-40/DCC receptor in ventral guidance , unc-6/netrin also acts as a repulsive cue for cells expressing both the unc-5/UNC5 and unc-40/DCC receptors , which together mediate dorsal guidance away from unc-6/netrin [4–6] . To address whether lon-2/glypican functions in unc-6/netrin-mediated repulsive guidance as well , we examined the dorsal migration of the distal tip cells ( DTCs ) and of the GABAergic motorneuron axons [4 , 11] . We found that lon-2/glypican single null mutants are defective in dorsal DTC migrations ( Fig 2A and 2B ) and that the complete loss of lon-2/glypican did not enhance the dorsal DTC migration defects of unc-6/netrin , unc-40/DCC , or unc-5/UNC5 null mutants ( Fig 2B ) , indicating that lon-2/glypican functions in the unc-6/netrin-repulsive guidance pathway as well . Similarly , complete loss of lon-2/glypican did not enhance the defects of unc-40/DCC mutants in the dorsal guidance of motorneuron axons ( Fig 2C ) . Given that loss of lon-2/glypican enhances the motorneuron axon guidance defects of sdn-1 mutants as shown in [17] , lon-2/glypican plays a role in the dorsal guidance of motorneuron axons . The lack of enhancement of the defects in the dorsal guidance of motorneuron axons of unc-40/DCC mutants by loss of lon-2/glypican further supports that lon-2/glypican functions in the unc-6/netrin pathway mediating dorsal guidance . Thus , lon-2/glypican may modulate unc-6/netrin signaling not only during attractive guidance but also during repulsive guidance . To complement the above loss-of-function approach , we next used a gain-of-function strategy to test the model that lon-2/glypican functions in the unc-6/netrin signaling pathway . We focused on the axon of the PVM neuron instead of AVM , because it could reliably be identified ( AVM cannot be distinguished from ALMR in these experiments ) . In wild-type animals , PVM , like AVM , expresses the receptor unc-40/DCC , and its axon grows ventrally towards unc-6/netrin ( Fig 3A ) . In mutants lacking unc-6/netrin signaling , PVM axons that fail to extend ventrally instead extend anteriorly ( never dorsally , see S4 Table ) . The PVM axon normally does not express the receptor unc-5/UNC5 that mediates repulsive guidance away from ventral unc-6/netrin [6] , but misexpression of the receptor unc-5/UNC5 ( using transgene Pmec-7::unc-5 [21] ) in PVM forces its axon to extend dorsally in an unc-6/netrin- and unc-40/DCC-dependent manner ( Fig 3A and 3B , [21] ) . We used this unc-6/netrin-dependent unc-5/UNC5-mediated abnormal dorsal migration to further investigate the function of lon-2/glypican in netrin signaling . By analyzing lon-2/glypican mutants carrying Pmec-7::unc-5 , we found that compete loss of lon-2/glypican function significantly suppressed the unc-6/netrin-dependent unc-5-mediated abnormal dorsal migration of the PVM axon , indicating that unc-6/netrin signaling is lon-2/glypican dependent ( Fig 3B ) . In contrast , the complete loss of sdn-1/syndecan , of slt-1/Slit , or of sax-3/Robo function did not suppress these PVM abnormal dorsal migrations ( Fig 3B , see S4 Table ) , highlighting the specificity of lon-2/glypican action on unc-6/netrin signaling . As expected , lon-2 sdn-1 double mutants lacking both lon-2/glypican and sdn-1/syndecan and expressing unc-5/UNC5 in PVM did not further suppress the abnormal unc-5/UNC-5-mediated dorsal migration of PVM as compared to lon-2 single mutants , further supporting the specificity of lon-2/glypican on unc-6/netrin signaling . To investigate whether lon-2/glypican functions in the same genetic pathway as known downstream mediators of unc-6/netrin signaling , we tested for genetic interactions between lon-2/glypican and unc-34/enabled . unc-34/enabled is a regulator of actin polymerization for axonal filopodia outgrowth [18 , 22–26] , and its role in both unc-6/netrin and slt-1/Slit guidance pathways renders the analysis of genetic interactions in the context of normal AVM axon guidance challenging . Therefore , we used the unc-6/netrin-specific gain-of-function approach as above , in which the dorsal migration of the PVM axon upon ectopic expression of unc-5/UNC5 is unc-34/enabled dependent ( Fig 3B , [21 , 27] ) . We asked whether loss of lon-2/glypican could enhance the extent of suppression of PVM dorsal migration induced by loss of unc-34/enabled . We found that the PVM dorsal migration was suppressed to the same degree in the double null mutants lon-2; unc-34 and the single mutant unc-34/enabled upon expression of unc-5/UNC5 in PVM ( Pmec-7::unc-5 , Fig 3B ) . These results support that lon-2/glypican functions with unc-6/netrin and unc-34/enabled during axon guidance . The AVM growth cone extends along a basement membrane that is located between the epidermis , which is referred to as the hypodermis , and body wall muscles [11] . lon-2/glypican is expressed in the hypodermis and the intestine [28] . We asked in which cell type lon-2/glypican needs to be produced to guide AVM . We found that wild-type lon-2 ( + ) transgenes expressed under the heterologous epidermal promoters Pdpy-7 and Pelt-3 ( that drive expression in the hypodermis underlying the AVM growth cone , hyp7 ) rescued lon-2 slt-1 double mutants back to slt-1 single mutant levels , as efficiently as when expressed under the endogenous promoter Plon-2 ( Fig 4A , S3 Table ) . Rescue was not observed when we expressed lon-2/glypican in other epidermal cells ( seam cells , Pgrd-10 ) , in the migrating neuron itself ( Pmec-7 ) , in the intestine ( Pelt-2 ) , or in body wall muscles ( Pmyo-3 ) ( Fig 4A , S3 Table ) . Our results suggest that lon-2/glypican is produced by the hypodermis underlying the growth cone of AVM to function in axon guidance . We found that expressing wild-type copies of sdn-1 ( + ) in the AVM neuron ( using the heterologous promoter Pmec-7 ) rescued axon defects of lon-2 sdn-1 double mutants ( Fig 4B ) . Accordingly , our examination of a transgene reporting sdn-1/syndecan expression ( sdn-1::gfp [16] ) revealed that sdn-1/syndecan is indeed expressed in the AVM neuron ( S4 Fig ) , at the time of its ventral migration during the first larval stage . Thus , sdn-1/syndecan appears to function in the migrating neuron in the slt-1/Slit-sax-3/Robo guidance pathway , whereas lon-2/glypican appears to function nonautonomously , as it is produced by the hypodermis underlying the migrating neuron to modulate the unc-6/netrin guidance pathway . Consistent with this , we found that sdn-1 ( + ) cannot replace the function of lon-2/glypican; expressing sdn-1/syndecan in either the cells that normally express lon-2/glypican ( using Plon-2::sdn-1 ) or the migrating neuron itself ( Pmec-7::sdn-1 ) did not rescue the loss of lon-2/glypican ( S5 Fig ) , supporting that lon-2/glypican and sdn-1/syndecan have specific roles in axon guidance . Glypicans are composed of a core protein moiety with covalently linked HS chains attached via a tetrasaccharide linker at specific Serine residues ( Fig 5A , [15] ) . Prior studies on the role of HSPGs in other developmental pathways indicate that both the identity of the HSPG core proteins and the heterogeneity of their HS chains modified by epimerization and sulfations [15] contribute to the specificity of the interactions between particular HSPGs and the proteins that they bind [15 , 29 , 30] . To address the importance of the HS chains linked to LON-2/glypican during axon guidance , we tested whether a mutated form of LON-2/glypican lacking its HS chains could still function in axon guidance . For this experiment , the three Serine residues serving as HS chain attachment sites were mutated to Alanine residues , generating the mutant LON-2ΔGAG [31] . Western blot analysis confirmed that LON-2ΔGAG severely reduced HS chains associated with LON-2 , in both worms and S2 cells ( Fig 5B and S6 Fig ) . We then expressed LON-2ΔGAG under the Plon-2 endogenous promoter and found that the AVM guidance defects of lon-2 slt-1 double mutants were rescued back to the level of slt-1 single mutants ( Fig 5C ) . Similarly , the DTC migration defects of lon-2/glypican mutants were rescued by LON-2ΔGAG expression ( Fig 5D ) . Our results indicate that LON-2/glypican devoid of its HS-chain attachment sites can function in unc-6/netrin-mediated guidance , suggesting that the core protein is the critical part of LON-2/glypican for its function in unc-6/netrin-mediated guidance of cell and axon migrations . Our above observations provide evidence that the HSPG lon-2/glypican functions in the same genetic pathway as unc-6/netrin to guide migrating axons . It has been shown in several models that HSPGs play multifaceted roles across various signaling pathways , such as facilitating ligand-receptor interactions and transporting morphogens , as well as localizing and stabilizing ligands [32 , 33] . We asked if the LON-2/glypican molecules might interact with either UNC-6/netrin or its receptor UNC-40/DCC , suggesting a potential mechanism of action for LON-2/glypican in unc-6/netrin-mediated guidance . To test these interactions , we generated epitope-tagged versions of LON-2/glypican , UNC-6/netrin , and UNC-40/DCC proteins , with human influenza hemagglutinin ( HA ) , superfolder-GFP ( SfGFP ) , and FLAG , respectively ( Fig 6A ) , and used cell-mixing experiments . We independently expressed each of these labeled proteins in separate populations of Drosophila S2 cells for 2 d , then cocultured them overnight , and detected the tagged proteins by western blot analysis ( see S7 Fig ) and by immunostaining ( Fig 6A ) . We observed that the HA::LON-2 signal filled the cytoplasm of HA::LON-2 producing cells ( indicated by white asterisks in Fig 6B experiment 1 and S8 Fig ) . Notably , HA::LON-2 was also found decorating the outline of UNC-40::FLAG-expressing cells ( Fig 6B and 6C experiments 1 , 6 , 7 , and 8 ) . This observation suggests that LON-2/glypican is released from the cells that produce it , diffuses in the extracellular medium , and associates with UNC-40/DCC-expressing cells . In contrast , HA::LON-2/glypican did not bind to cells expressing SfGFP::UNC-6 ( Fig 6B and 6C experiments 4 , 6 , and 7 ) or to cells expressing an unrelated type I transmembrane receptor , Evi ( see S9 Fig ) , or to untransfected cells ( Fig 6B and 6C experiments 1–8 ) . Furthermore , we found that another HSPG , SDN-1/syndecan , did not bind UNC-40-expressing cells ( see S9 Fig ) . These findings provide evidence for a specific interaction between LON-2/glypican and UNC-40-expressing cells . We tested whether the HS chains of LON-2/glypican were necessary for its association with UNC-40-expressing cells . We used a mutated form of LON-2/glypican lacking its three HS chain attachment sites , HA::LON-2ΔGAG ( see S6 Fig , [31] ) . Western blot analysis confirmed that LON-2ΔGAG severely reduced HS chains associated with LON-2/glypican ( S6 Fig ) . We found that LON-2ΔGAG associated with UNC-40/DCC-expressing cells ( Fig 6B and 6C experiment 2 ) , suggesting that the association of LON-2/glypican with UNC-40/DCC-expressing cells is HS-chain independent . The HA::LON-2 signal outlined the UNC-40/DCC-expressing cells ( Fig 6B , experiments 1 , 6 , 7 , and 8 ) suggesting a potential interaction at the cell surface . To further support this idea , we asked whether LON-2/glypican would associate with cells expressing a mutated form of UNC-40/DCC that lacks the extracellular domain and contains only the intracellular and transmembrane domains ( UNC-40ΔNt::FLAG ) . We found that HA::LON-2 did not associate with cells expressing the UNC-40ΔNt::FLAG ( Fig 6B and 6C experiment 3 ) , indicating that the extracellular domain of UNC-40/DCC is required for LON-2/glypican to associate , as would be predicted if LON-2/glypican and UNC-40/DCC interact , directly or indirectly , at the cell surface . Interestingly , HA::LON-2 was absent from cells expressing SfGFP::UNC-6 ( Fig 6B and 6C experiments 4 , 6 , and 7 ) , indicating that while LON-2/glypican interacts with cells expressing UNC-40/DCC , it does not bind to UNC-6/netrin-expressing cells in this assay . Moreover , the presence of SfGFP::UNC-6 did not reduce the ability of HA::LON-2 to associate with UNC-40/DCC-expressing cells in experiments in which the three singly transfected cell populations were mixed ( Fig 6B and 6C experiment 6 ) . These results suggest that if LON-2/glypican interacted directly or indirectly with UNC-40/DCC , then the interactions of LON-2/glypican and UNC-6/netrin would occur with different regions of UNC-40/DCC . Consistent with this possibility , we found that LON-2/glypican still associated with cells expressing UNC-40ΔFn4/5::FLAG , a mutated form of UNC-40/DCC that lacks the UNC-6/netrin-binding sites ( FnIII domains 4 and 5 ) ( Fig 6B and 6C experiment 7 , [34 , 35] ) . Our results indicate that for LON-2/glypican to associate with UNC-40/DCC-expressing cells , the FnIII domains 4 and 5 of UNC-40/DCC are dispensable and UNC-6/netrin does not need to be bound to UNC-40/DCC . Previous work has suggested that overexpression of DCC in cells overactivates DCC downstream signaling pathways , leading to cytoskeletal rearrangements that result in increased membrane extensions and cell surface area [36] . Similarly , expression of UNC-40/DCC leads to changes in cellular morphology in our cell assays ( Fig 6D ) . To test whether the association of LON-2/glypican with UNC-40/DCC-expressing cells results in an activation of signaling downstream of UNC-40/DCC , we examined the impact of LON-2/glypican on the morphology of UNC-40/DCC-expressing cells . For these experiments , we mixed mCherry-expressing cells with either untransfected control cells or LON-2/glypican-expressing cells , and we also mixed UNC-40/mCherry-expressing cells with either untransfected control cells or LON-2/glypican-expressing cells . Examination of the morphology of these cells 1 d after mixing revealed that UNC-40/mCherry-expressing cells mixed with LON-2/glypican exhibited an increased frequency of irregular shapes and membrane extensions , compared to UNC-40/mCherry cells mixed with control cells ( Fig 6D ) . Thus , consistent with a model in which LON-2/glypican functions in the UNC-6/netrin signaling pathway to guide developing axons , the association of LON-2/glypican with UNC-40/DCC-expressing cells leads to increased membrane extensions , suggestive of increased signaling downstream of the UNC-40/DCC receptor . While LON-2/glypican possesses a signature GPI anchor that mediates its attachment to plasma membranes ( Fig 5A ) , our experiments indicate that LON-2/glypican is released into the extracellular milieu through cleavage where it can diffuse to associate with UNC-40/DCC-expressing cells . This is consistent with prior work demonstrating that many glypicans are shed or cleaved into a soluble form [37] . To verify that LON-2/glypican is indeed released into the extracellular medium , we collected cell-free media from HA::LON-2 cultures ( HA::LON-2-conditioned medium ) and added it to cells expressing UNC-40::FLAG . We found that HA::LON-2-conditioned medium contained HA::LON-2 that associated with UNC-40::FLAG-expressing cells . As above , this interaction was specific , as no HA::LON-2 signal was found on adjacent untransfected cells ( Fig 6B and 6C experiment 8 ) . This result provides compelling evidence that LON-2/glypican can be released from the membrane of LON-2/glypican-expressing cells , diffuses , and associates with UNC-40/DCC-expressing cells . We propose that using a similar mechanism , LON-2/glypican may be shed from epidermal cells and may interact with migrating axons that express UNC-40/DCC . This is consistent with our finding that LON-2/glypican is produced by the hypodermis to function nonautonomously in unc-6/netrin-mediated AVM axon guidance . To provide evidence for the model that LON-2/glypican can function in axon guidance when detached from the plasma membrane , we used a form of LON-2/glypican lacking the GPI anchor , LON-2ΔGPI , which should be directly secreted into the extracellular milieu [31] . LON-2ΔGPI rescued the AVM guidance defects of lon-2 slt-1 double mutants back to the level of slt-1 single mutants ( Fig 5C ) . We also used a truncated form of LON-2/glypican ( N-LON-2 ) containing the N-terminal globular domain , but lacking the C-terminal region , thus removing the three HS attachment sites and the GPI membrane anchor . N-LON-2 also rescued the AVM guidance defects of lon-2 slt-1 double mutants back to the level of slt-1 single mutants ( Fig 5C ) . In contrast , a reciprocal construct containing only the C-terminus with the three HS attachment sites and the GPI anchor ( C-LON-2 ) did not rescue the AVM axon guidance defects of lon-2 slt-1 , consistent with the model that the N-terminal globular domain of LON-2/glypican is the key functional domain during guidance ( Fig 5C ) . A secreted form of LON-2/glypican is also functional in DTC guidance , as we found that DTC guidance defects of lon-2/glypican mutants could be rescued by expression of N-LON-2 , containing only the N-terminal globular domain ( Fig 5D ) . These findings also support the hypothesis that LON-2/glypican may normally be released from the hypodermis to interact with the unc-6/netrin pathway to direct the migrating growth cone during development ( Fig 7 ) . Our studies identify the HSPG LON-2/glypican as a component of the unc-6/netrin attractive and repulsive signaling pathways that guide axons during development . We show that LON-2/glypican specifically acts on unc-6/netrin signaling independently of slt-1/Slit . We demonstrate that lon-2/glypican functions from the hypodermis , the epidermal cells that secrete the substrate along which growth cones extend [11] , and that a secreted form of LON-2/glypican , containing only its N-terminal globular region and lacking its HS chains , guides cells and axons in vivo . In addition , we provide evidence that LON-2/glypican is released from cells producing it and associates with cells expressing UNC-40/DCC receptors . Taken together , our observations support a hypothetical model in which GPI-linked LON-2/glypican is produced by substrate epidermal cells , is released into the extracellular milieu , and binds growth cones expressing UNC-40/DCC receptors to regulate attractive and repulsive responses of the growth cone to UNC-6/netrin . The impact of lon-2/glypican on the unc-6/netrin signaling pathway is highly specific . First , loss of lon-2/glypican , but not of sdn-1/syndecan , suppresses the guidance phenotypes elicited by the gain-of-function condition in which unc-5/UNC5 was misexpressed . Second , the complete loss of lon-2/glypican does not enhance the guidance defects observed in null mutants for unc-6/netrin or its receptors unc-40/DCC and unc-5/UNC5 , whereas it does enhance the defects of several other axon guidance mutants , including sdn-1/syndecan , slt-1/Slit , misexpressed slt-1/Slit ( Pmyo-3::slt-1 ) , sax-3/Robo , and sqv-5 , suggesting that lon-2/glypican functions specifically in the unc-6/netrin pathway . Third , sdn-1/syndecan , cannot replace lon-2/glypican function , highlighting a requirement for lon-2/glypican that cannot be achieved by any HSPG . Given that the core protein of LON-2/glypican , devoid of its HS chains , is fully functional in guidance , the specificity of action of LON-2/glypican in netrin-mediated guidance appears to reside in the core protein itself . As a note , whereas lon-2/glypican mutants are defective in DTC migration , the lon-2/glypican mutant by itself does not show drastic alterations in AVM axon guidance as is observed with other modulators [32] . It is possible that in the absence of lon-2/glypican , another HSPG may provide compensation or that our scoring of strong alterations in pathfinding did not include more subtle phenotypes , as could be expected from a modulator of the signal [32] . We show that the LON-2/glypican core protein , devoid of HS attachment sites , is able to associate with UNC-40-expressing cells and is functional in unc-6/netrin-mediated guidance . Thus , the core protein is the critical region of LON-2/glypican for netrin-mediated axon guidance . This is in line with previous studies showing a contextual dependence of HS chains for glypican function . For instance , the core protein of C . elegans LON-2/glypican and of Drosophila glypican Dally do not require HS chains to function in the transforming growth factor beta ( TGFβ ) pathway [31 , 39] . Similarly , Drosophila glypican Dally-like interacts with Wg and Hh through their protein core in a HS-independent manner [33 , 40 , 41] , and mammalian Glypican-3 does not require HS chains for its role in Wnt and Hh signaling [42–45] . While the HS chains are not critical for the role of LON-2/glypican in guidance , a contribution of HS chains to modulate functionality , as observed for other glypicans in the context of BMP4 , Wnt3 , Wg , and Hh signaling [33 , 39 , 41 , 42] , cannot be ruled out . For instance , it is conceivable that the normal endogenous HS chains of LON-2/glypican may impact its trafficking , levels , release from the membrane , recruitment of binding partners , or recycling . LON-2/glypican is predicted to localize at the cell surface via its GPI anchor [31] . However , in our cell culture studies , we demonstrate that LON-2/glypican can be released as a soluble molecule from producing cells . We also show that two truncated forms of LON-2/glypican , LON-2ΔGPI and N-LON-2 , which are no longer associated with the plasma membrane and are secreted into the extracellular milieu , can function to guide axons in vivo . This indicates that LON-2/glypican is likely released from the epidermal cells to reach the growth cone to modulate its guidance . This finding raises the question of how LON-2/glypican is released from the cell membrane and how this process might be regulated during development . The release of LON-2/glypican from the surface of cells could involve phospholipases that cleave the GPI anchor and/or proteases that cleave its extracellular domain , such as at a predicted furin-cleavage site ( Fig 5A , [31] ) . Glypican cleavage by lipases and proteases has been demonstrated to occur and to be functionally important in other contexts , such as in regulating fibroblast growth factor ( FGF ) and Wnt signaling during morphogenesis [37 , 46] . For instance , the Drosophila glypican Dally-like protein is cleaved at the GPI anchor by the lipase Notum , to negatively regulate Wnts [47] . Similarly , several mammalian glypicans , including glypican-3 , are cleaved by Notum [48] . The functional importance of glypican proteolytic cleavage is illustrated by the processing of glypican-3 by a furin-like convertase to modulate Wnt signaling in zebrafish [49] . In addition , glypican-1 and glypican-4 are proteolytically cleaved to stimulate long-range FGF signaling in the Xenopus embryo [50] and increase the efficiency of myogenic differentiation in the presence of FGF in mammalian cells [51] , respectively . Our studies show that glypican processing also functions during axon guidance . We demonstrate that LON-2/glypican is secreted into the extracellular medium and decorates the outline of UNC-40/DCC-expressing cells . Deleting the extracellular domain of UNC-40 ( UNC-40ΔNt ) abrogated the association of LON-2/glypican with UNC-40/DCC-expressing cells , indicating that LON-2/glypican may interact with UNC-40/DCC at the cell surface . The association of LON-2/glypican with UNC-40/DCC may be direct or indirect through interactions with other molecules ( Fig 7 ) . Our experiments demonstrate that UNC-6/netrin binding to UNC-40/DCC was undisturbed by the association of UNC-40/DCC with LON-2/glypican , suggesting that the possible interaction of LON-2/glypican with UNC-40/DCC likely involves a region of UNC-40/DCC other that the netrin binding sites . Indeed , we found that LON-2/glypican associates with UNC-40/DCC-expressing cells even when the UNC-40/DCC receptors lack the UNC-6/netrin binding domains . We found that LON-2/glypican leads to increased irregular morphology of UNC-40/DCC-expressing cells . Ectopic expression of DCC in mammalian cells activates downstream signaling via Cdc42 and Rac1 , producing cytoskeletal rearrangements that lead to filopodia outgrowth and cell surface extensions [36] . Our finding that the presence of LON-2/glypican enhances the UNC-40/DCC-induced irregular cell morphology and filopodia-like extensions suggests that the association of LON-2/glypican with UNC-40/DCC-expressing cells may increase signaling downstream of UNC-40/DCC . Consistent with this notion , we show that lon-2/glypican functions in the same signaling pathway as the UNC-40/DCC downstream mediator unc-34/enabled during axon guidance . Our results suggest a possible regulatory mechanism in the extracellular space in which secreted LON-2/glypican modulates the activity of the receptor UNC-40/DCC . LON-2/glypican may directly bind UNC-40/DCC , or alternatively , LON-2/glypican may instead interact with other molecules to impact UNC-40/DCC to modulate its stability , distribution , or activity . Alternatively , LON-2/glypican could potentially function as a co-receptor for UNC-6/netrin , where it may facilitate the formation of UNC-6/netrin-UNC-40/DCC-LON-2/glypican signaling complexes , similar to the situation in FGF signaling [15] . It is also conceivable that LON-2/glypican could bind UNC-6/netrin directly as well , even if undetected in our assays , as netrin has been found to bind heparin in vitro [3 , 52 , 53] . Previous studies have also documented the binding of DCC to heparin in vitro [7 , 8] , and while we have found that the core protein is the critical portion of LON-2/glypican in netrin-mediated axon guidance , it remains possible that the endogenous HS chains contribute to the function of LON-2/glypican in axon guidance . In summary , our studies uncover a novel mechanism by which UNC-6/netrin signaling through its UNC-40/DCC receptor is modulated by the HSPG LON-2/glypican during axon pathfinding . Given the evolutionary conservation of the UNC-6/netrin pathway components ( UNC-6/netrin and its receptors UNC-40/DCC and UNC-5/UNC5 ) and of glypicans ( LON-2 is most similar to mammalian glypican-3 ) and that synthesis of HS chains is required for mammalian axons to respond to netrin-1 in vitro [9 , 10] , glypicans are likely to play a role in netrin-mediated axon pathfinding in mammals as well . Our findings provide a general mechanism for the extracellular regulation of growth cone responses to netrin during the development of nervous systems . Nematode cultures were maintained at 20°C on NGM plates seeded with OP50 bacteria as described [54] . Strains were constructed using standard genetic procedures and are all listed in S8 Table . Genotypes were confirmed by genotyping PCR or by sequencing when needed , using primers listed in S9 Table . Total RNA was extracted from worm samples using Trizol ( Invitrogen ) according to the manufacturer’s instructions . 500 ng RNA was used to reverse transcribe using the High Capacity cDNA Reverse Transcription Kit ( Applied Biosystems ) and random primers . PCR reactions were carried out with first-strand cDNA template , primers oCB834 ( ATCAAGACCGAGTGATAGTG ) and oCB1321 ( TGGCGAGTATTCCCGTTTAG ) were used for gpn-1 cDNA amplification , and primers oCB992 ( TCGCTTCAAATCAGTTCAGC ) and oCB993 ( GCGAGCATTGAACAGTGAAG ) were used for the control gene Y45F10D . 4 [55] cDNA amplification . Animals were mounted on agarose pads , anaesthetized with 100 mM sodium azide , and examined under a Zeiss Axio Scope . A1 or a Zeiss Axioskop 2 Plus . All inserts of finalized clones were verified by sequencing . Mixed-stage wild-type ( N2 ) , GFP control ( lqIs4 ) , LON-2::GFP ( TLG257 ) , and LON-2ΔGAG::GFP ( TLG199 ) worms were collected in buffer and protease inhibitors ( Roche ) . Worm pellets were subjected to repeated freeze-thaw cycles . Protein concentration was measured using the Pierce 660 nm Protein Assay on a Nanodrop . 70 μg of samples mixed with 2x Laemmli sample buffer ( Bio-Rad ) were boiled , separated by SDS-PAGE on a 4%–20% Mini-Protean TGX gel ( Bio-Rad ) , and transferred to PVDF membrane . Membranes were incubated in 1:3000 anti-GFP primary antibody ( Millipore #AB3080 ) and 1:9000 goat anti-rabbit HRP secondary antibody ( Bio-Rad #166-2408EDU ) . For the loading control , membranes were incubated in 1:5000 anti-HSP90 antibody ( CST #4874 ) and 1:10000 goat anti-rabbit HRP secondary antibody ( Bio-Rad #166-2408EDU ) . Signal was revealed using Clarity Western ECL Substrate ( Bio-Rad ) and imaged using film ( LabScientific ) . All inserts of finalized clones were verified by sequencing . S2 cells were transfected with HA::LON-2::myc ( pCB313 ) and HA::LON-2ΔGAG::myc ( pCB330 ) constructs . Cells were washed once with 1X Phosphate Buffered Saline and lysed for 30 min at 4°C in 1X Phosphate Buffered Saline , 0 . 5% Triton X-100 , and 1X Protease Inhibitor Cocktail ( Roche ) . Samples of supernatant and cell lysates were each mixed with 2X Laemmli sample buffer ( BioRad ) . Proteins were separated by SDS-PAGE and transferred to PVDF membrane . Membranes were incubated with rabbit anti-HA ( Life Technologies #715500 ) and rabbit anti-myc ( Santa Cruz #sc-789 ) primary antibodies and HRP-linked goat anti-rabbit ( Bio-Rad #166-2408EDU ) secondary antibody . Signals were revealed by chemiluminescence with Clarity Western ECL Substrate ( BioRad ) and imaged using the ChemiDoc System ( BioRad ) . S2 cells were independently transfected with HA::LON-2::myc ( pCB313 ) , UNC-40::FLAG ( pCB301 ) , or SfGFP::UNC-6 ( pCB292 ) constructs . 48 h after transfection , old culture medium was removed , and new medium was added to resuspend the cells . Equal volumes of resuspended cells that had been transfected with individual constructs were mixed and cocultured overnight . Cells were harvested , centrifuged , and combined with their corresponding supernatant from each of these cell mixes . 100 μL of supernatant of each mixture was saved and kept on ice . Cell pellets were washed once with 1X Phosphate Buffered Saline and lysed for 30 min at 4°C in 100 μL of ice-cold RIPA buffer ( 50 mM Tris HCl pH 7 . 5 , 150 mM NaCl , 1% Triton-X100 , 0 . 5% sodium deoxycholate , 0 . 1% SDS , and 1mM EDTA pH 8 . 0 ) supplemented with Protease Inhibitor Cocktail ( Roche ) and PMSF . Cell lysates were combined with their corresponding supernatant and mixed with 2X Laemmli sample buffer ( BioRad ) . Each sample was split into three in order to run three protein gels in parallel . Proteins were separated by SDS-PAGE and transferred to PVDF membrane . Membranes were incubated with rabbit anti-myc ( Santa Cruz #sc-789 ) , mouse anti-FLAG ( Sigma #F3165 ) , and rabbit anti-GFP ( Millipore AB3080 ) primary antibodies as well as HRP-linked goat anti-rabbit ( Bio-Rad #166-2408EDU ) and HRP-linked horse anti-mouse ( Vector Labs PI-2000 ) secondary antibodies . Signals were revealed by chemiluminescence with Clarity Western ECL Substrate ( BioRad ) and imaged using the ChemiDoc System ( BioRad ) . S2 cells were maintained in SFX Insect Media ( HyClone ) containing 10% Fetal Bovine Serum ( HyClone ) and Penicillin-Streptomycin ( 50 units-50 μg/mL ) ( Sigma ) . 70%–90% confluent S2 cells were transfected with 500 ng of each construct using Effectene ( Qiagen ) according to the manufacturer’s protocol . 48 h after transfection , old culture medium was removed and new medium was added to resuspend the cells . Equal volumes of resuspended cells that had been transfected with individual constructs were plated onto coverslips and cocultured overnight . Cells were then fixed with 4% paraformaldehyde and immunostained with rabbit anti-HA ( Life Technologies #715500 ) and mouse anti-FLAG ( Sigma #F3165 ) primary antibodies and Alexa594 donkey anti-rabbit ( Life Technologies #R37119 ) and Alexa647 goat anti-mouse ( Life Technologies #A21235 ) secondary antibodies . Confocal analysis was performed on a Zeiss LSM 5 Pascal confocal microscope . Confocal images were processed using ImageJ . Each experiment was repeated at least three times . For the experiment in which we use HA::LON-2-conditioned medium ( supernatant ) of cells expressing HA::LON-2 , the culture medium was also changed 48 h after transfection , fresh medium was added , and the cells were incubated for another 48 h . This medium was collected and centrifuged at 1 , 500 rpm to remove cells and debris . This supernatant was added onto cells expressing UNC-40::FLAG , incubated overnight , and as above , fixed , stained , and imaged . Independent populations of S2 cells were transfected with ( 1 ) 450 ng of pActin5 . 1::mCherry alone , ( 2 ) 50 ng of the UNC-40::FLAG construct plus 450 ng of the cotransfection marker pActin5 . 1::mCherry , or ( 3 ) 500 ng of HA::LON-2 . The medium was changed and cells were mixed 48 h after transfection . Control mCherry-expressing cells were mixed with untransfected cells or with HA::LON-2-expressing cells . Similarly , UNC-40::FLAG/mCherry-expressing cells were mixed with untransfected cells or with HA::LON-2-expressing cells . To maintain the total number of cells constant in our different mixes , one volume of UNC-40::FLAG/mCherry cells was mixed with either ( a ) one volume of control/untransfected cells or ( b ) one volume of LON-2-transfected cells . Cell mixes were cocultured overnight . Cells were then fixed with 4% paraformaldehyde and examined under a Zeiss LSM 5 Pascal confocal microscope . Control mCherry-expressing cells or UNC-40::FLAG/mCherry-expressing cells were identified by the cotransfection marker mCherry . 20 fields of ~300 cells each per mix per were photographed for each of three independent experiments . Cells were categorized as having the typical S2 cell round and smooth shape , irregular edges , and/or extensions protruding from the cell .
During the development of the nervous system , migrating axons are guided as they navigate through complex environments to reach their target destinations . These directed migrations are essential to ensure the proper wiring and function of the nervous system and are instructed by guidance cues and receptors . There is a remarkably small set of guidance cues and receptors relative to the large number of neuronal migrations , suggesting that the actions of these guidance cues might be diversified by regulatory mechanisms . We have addressed this question in the genetically tractable nematode Caenorhabditis elegans . We identify that the response of migrating neurons to a key guidance cue , UNC-6/netrin , is modulated by a specific proteoglycan , LON-2/glypican . We show that LON-2/glypican may carry out this regulation by interacting with the UNC-40/DCC netrin receptor on the cell surface . We propose that LON-2/glypican acts as an extracellular modulator of UNC-40/DCC-mediated guidance to fine-tune axonal responses to UNC-6/netrin signals during migration .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[]
2015
Glypican Is a Modulator of Netrin-Mediated Axon Guidance
The elongation phase of transcription by RNA Polymerase II ( Pol II ) involves numerous events that are tightly coordinated , including RNA processing , histone modification , and chromatin remodeling . RNA splicing factors are associated with elongating Pol II , and the interdependent coupling of splicing and elongation has been documented in several systems . Here we identify a conserved , multi-domain cyclophilin family member , SIG-7 , as an essential factor for both normal transcription elongation and co-transcriptional splicing . In embryos depleted for SIG-7 , RNA levels for over a thousand zygotically expressed genes are substantially reduced , Pol II becomes significantly reduced at the 3’ end of genes , marks of transcription elongation are reduced , and unspliced mRNAs accumulate . Our findings suggest that SIG-7 plays a central role in both Pol II elongation and co-transcriptional splicing and may provide an important link for their coordination and regulation . Transcription by RNA Polymerase II ( Pol II ) is a highly regulated process involving coordination of multiple processes that together modulate the level of gene expression and its temporal and spatial control [1–4] . Epigenetic mechanisms play important roles in both transcription initiation and elongation , with various histone modifications both guiding and resulting from these processes [5–8] . Kinases also regulate both stages by modifying the C-terminal domain ( CTD ) of Pol II’s catalytic subunit and phosphorylating other factors that regulate the transitions accompanying Pol II transcription , including promoter-proximal pausing [9–16] . The CTD is composed of a conserved heptapeptide repeat , and phosphorylation of specific serines and threonines within the repeats correlates with these transitions [2 , 17–25] . A connection between the modifications of the CTD and mRNA splicing has long been observed . It has been shown that the association of SR ( Serine/Arginine-rich ) splicing factors with the CTD requires phosphorylation of Ser2 of the heptapeptide repeat [26–28] . It has also been observed that some splicing factors are required for normal RNA Pol II elongation , suggesting a reciprocal mechanistic relationship between RNA processing and transcription elongation [29–32] . RNA processing in the nucleus is largely co-transcriptional , so an interdependency of splicing and Pol II elongation represents a potentially important mode of transcription regulation . In addition to kinases and histone modifying enzymes , peptidyl proline isomerases ( PPIs ) can regulate Pol II during transcription progression . The nuclear parvulin family of PPIs direct cis-trans isomerization of prolines in the context of Ser/Thr , such as those found in the Pol II CTD heptapeptide repeats , and the activity of these PPIs is affected by the phosphorylation of Ser/Thr [33–36] . These PPIs are thought to contribute to structural regulation of the CTD , participating in a “CTD code” that controls the recruitment of various factors to Pol II during elongation and transcript processing [33 , 37–39] . The nuclear cyclophilin PPI family , characterized by having an RNA-recognition motif ( RRM ) in addition to a PPI domain has also been implicated in regulation of Pol II through interactions with the CTD . Members of this highly conserved family include KIN241 in Paramecium tetraurelia , AtCyp59 in Arabidopsis thaliana and Rct1 in Schizosaccharomyces pombe [40–43] . AtCyp59 interacts with Pol II , and its overexpression causes defective regulation of Pol II CTD phosphorylation [41] . AtCyp59 also interacts with RNA through its RRM domain and has PPI activity , but whether the PPI domain is required for AtCyp59 function is unclear [42] . The S . pombe Rct1 also interacts with and affects Pol II CTD phosphorylation , and the effect on phosphorylation is dependent on Rct1’s PPI domain [43 , 44] . Here we present the first in vivo , genome-wide analysis of a C . elegans nuclear cyclophilin , SIG-7 , and show that it is essential for normal transcription and RNA processing during embryogenesis . Loss of SIG-7 results in a genome-wide decrease in mRNA production that is correlated with both defective elongation and defective co-transcriptional splicing . Our results identify SIG-7 as a conserved and important factor for both efficient Pol II transcription elongation and co-transcriptional splicing . A recessive mutation in sig-7 ( silencer in germline 7 ) was recovered years ago from an unpublished screen for defective silencing of repetitive transgenes in germ cells . The sig-7 ( cc629 ) allele exhibits multiple developmental defects in both soma and germline ( S1 Fig ) . The cc629 mutation was mapped to a small region of LG I , and a mutation in a predicted splice acceptor site in gene F39H2 . 2 was identified by sequencing ( S2B Fig ) . Sequencing of F39H2 . 2 cDNAs from cc629 animals showed the predicted splicing error in most , but a few correctly spliced cDNAs were also recovered , consistent with the lack of a strict requirement for a canonical AG dinucleotide at the 3' end of pre-mRNA introns [45] . RNAi targeting F39H2 . 2 resulted in embryonic lethality with rare “escapers” growing up to exhibit the same spectrum of pleiotropic phenotypes observed in the cc629 animals . The cc629 allele is thus hypomorphic , with partial maternal rescue of homozygous offspring produced by heterozygous mothers ( see below ) . An additional deletion allele , n5037 ( a deletion allele from L . Ma and R . Horvitz , S2B Fig ) , causes early larval arrest . The early embryonic arrest caused by F39H2 . 2 RNAi and the larval arrest of n5037 homozygous offspring from heterozygous mothers is indicative of both maternal and zygotic requirements for F39H2 . 2 function . Because F39H2 . 2 ( hereafter called sig-7 ) is the third gene in an operon ( CEOP1492; S2A Fig ) , a sig-7::gfp::3XFLAG translational fusion transgene was generated from a fosmid clone ( TransgeneOme Project ) encompassing the entire operon [46] . This transgene , when integrated as a single copy using MosSCI , rescues sig-7 ( n5037 ) animals to fertile adults . The rescued n5037 deletion strain expressing SIG-7::GFP::3xFLAG was used in the experiments described below . SIG-7 possesses an N-terminal peptidyl prolyl isomerase ( PPI ) domain and an adjacent RNA-recognition motif ( RRM ) ( S2C Fig ) . The C-terminal region is of low complexity , characterized by the presence of many charged residues including RS and RD dipeptides ( S3 Fig ) . SIG-7 is the sole C . elegans ortholog of a highly conserved family of PPI and RRM domain-containing nuclear cyclophilins found in most eukaryotes from fission yeast to humans , but notably absent from S . cerevisiae [47] . SIG-7 homologs share greater than 39% overall amino acid identity , most of which is concentrated within the PPI and RRM domains , with the highest degree of sequence identity in the RRM domain ( S2D and S3 Figs ) . A . thaliana AtCyp59 and S . pombe Rct1 interact with the CTD of Pol II and serve roles in Pol II regulation ( 41 , 43 ) . SIG-7::GFP::3XFLAG transgene localizes to the nucleus in all tissues including germ cells ( S4 Fig ) . In adult germ cells , the localization overlaps with DNA in mitotic and meiotic nuclei , but with a somewhat broader distribution than DAPI-intense chromatin ( Figs 1 and S4D ) . In diakinetic oocytes , SIG-7::GFP loses its chromatin association and becomes diffuse within the nucleoplasm ( S4D Fig ) . This transition correlates with the loss of Pol II from chromatin and the presumed global cessation of transcription in late oogenesis [48] . We further examined the correlation between SIG-7 staining and transcriptionally active chromatin in meiotic germ cells . Transcriptional activity is repressed on the X chromosomes relative to the autosomes during C . elegans meiosis: the X chromosomes are easily identified by their significantly lower levels of AMA-1 ( the catalytic subunit of C . elegans Pol II ) , and H3K36me3 and H3K4me2 , chromatin marks of transcription [48] . SIG-7::GFP localization in meiotic chromatin exhibits the same pattern as AMA-1 and H3K36 and H3K4 methylation i . e . , abundant on all autosomes and depleted from the X chromosomes ( Fig 1 ) . Thus , SIG-7 associates with chromatin and co-localizes with active transcription . We next examined sig-7 ( RNAi ) embryos for defects in gastrulation , a sensitive indicator of zygotic transcription defects in embryos . Gastrulation in C . elegans consists of the inward migration of a few peripheral cells , including the P4 cell , which is the progenitor of the two primordial germ cells , Z2 and Z3 [49 , 50] . Gastrulation is largely completed by the ~80 cell-stage , with P4 having migrated to the interior and subsequently divided to yield internally localized Z2/Z3 , which are readily identified using antibodies that recognize germline-specific P-granules ( Fig 2A , L4440 control ) . Zygotic gene activation is required for gastrulation , and disruption of embryonic Pol II or other essential transcription activities in embryos results in a failure of P4 to migrate internally , causing Z2 and Z3 to be born at the periphery ( Fig 2A; ama-1 ( RNAi ) ) . RNAi targeting of either ama-1 or sig-7 caused a highly penetrant gastrulation phenotype , yielding 92 . 5% and 86 . 15% gastrulation-defective embryos , respectively ( Fig 2A and 2B ) . Thus , SIG-7 is required for normal zygotic transcription during early embryonic development . The sig-7 ( RNAi ) gastrulation phenotype could be due to inactivation of one or a few genes specifically involved in gastrulation . As a first test for a more widespread defect , we quantified transcript levels of a panel of genes with strictly zygotic expression in sig-7 ( RNAi ) embryos [51–53] . Significant decreases for all tested zygotic transcripts were observed ( Fig 2C ) . The decreases were substantial , albeit not as dramatic as those observed in ama-1 ( RNAi ) embryos . Thus , both molecular and phenotypic data indicate that loss of SIG-7 activity leads to reduced levels of zygotic gene expression . We next tested for physical interactions between SIG-7 with Pol II in C . elegans by immunoprecipitating AMA-1 from transgene-rescued sig-7 ( n5037 ) animals , followed by probing western blots of the precipitated material with anti-FLAG antibodies to detect SIG-7::GFP::3xFLAG . SIG-7 co-precipitated with AMA-1 ( Fig 3 ) , and in reciprocal experiments AMA-1 co-precipitated with SIG-7 ( S5 Fig ) . These results indicate that SIG-7 interacts with Pol II in vivo . To further explore the extent of SIG-7’s role in gene expression , we next performed RNA-seq on sig-7 ( RNAi ) embryos and on L4440 RNAi control embryos ( Fig 4 ) . The results revealed that sig-7 RNAi causes a global defect in embryonic gene expression . Of the 45 , 627 annotated genes ( including non-coding RNAs , etc . ) , 10 , 703 had sufficient read representation for further analysis . Of these , 3 , 045 genes displayed significantly different RNA accumulation ( q≤0 . 05 ) between sig-7 ( RNAi ) and L4440 control RNAi samples ( S2 Table ) . Many more genes were down-regulated at least 2-fold ( 1549 ) than were up-regulated at least 2-fold ( 362 ) in sig-7 ( RNAi ) ( Fig 4A ) . We sorted these genes into gene categories based on published evidence for either zygotic expression during embryonic development ( “soma-specific” , “embryo–expressed” , and “X-linked” ) , exhibiting “ubiquitous” expression , or displaying expression enriched in or restricted to the germline ( “germline-enriched” and “germline-specific” , respectively ) [54] . X-linked genes show a distinct bias for either having weak expression in germ cells or only being expressed in somatic lineages [55] . Genes categorized as soma-specific , embryo-expressed , or X-linked were significantly over-represented among the down-regulated genes , and genes categorized as germline-expressed were significantly under-represented ( Fig 4B , left panel ) . This pattern was reversed for the up-regulated genes: germline-expressed genes , including ubiquitous and germline-enriched genes , were significantly over-represented ( Fig 4B , right panel ) . We investigated whether the different effects of loss of SIG-7 on germline versus somatic transcripts in embryos reflected a different requirement of those tissues for SIG-7 , or weaker germline RNAi effects using the standard feeding technique from L3 stage . Favoring the latter possibility , we found that extended RNAi for longer periods as adults resulted in significant reduction of germline-expressed genes in both sig-7 and ama-1 RNAi adult animals ( S6B Fig ) , and RNAi starting from earlier stages caused sterility . Thus , SIG-7 is required for efficient RNA production in larval and adult germ cells as well as in embryos . The differential effect of loss of SIG-7 from mothers on germline versus soma transcripts in embryos is probably due to lower efficiency of RNAi in adult germ cells using standard feeding protocols from the L3 stage . We also considered whether the embryonic arrest phenotype was skewing the effect on genes expressed in later-stage embryos , since we compared sig-7 ( RNAi ) embryos that mostly arrest at ~200–300 cells with control embryos that can continue to develop . The impact of stage differences on our results is probably low . The embryos used in our experiments were isolated from young adults with developing embryos in their uterus; these embryos are highly enriched for stages prior to the sig-7 ( RNAi ) arrest point . Analysis of the embryo stage distributions from independent RNAi experiments showed the expected bias for early stages ( e . g . S7 Fig ) . We further compared the affected genes from our experiments with those analyzed in a landmark study examining transcript dynamics at early C . elegans embryonic developmental time points , all of which are earlier than the sig-7 ( RNAi ) arrest point [51] . We focused on three narrowly defined gene sets: “strictly maternal” ( expressed in the ovary and degraded in the early embryo ) , “maternal/embryonic” ( expressed in both ovary and embryos ) , and “strictly embryonic” ( expressed only by the embryo with no maternal contribution ) . Of the genes classified as “strictly embryonic” that in our analyses showed >2-fold changes in sig-7 ( RNAi ) embryos , 328/339 ( 96 . 7% ) were down-regulated and only 11/339 ( 3 . 2% ) were significantly up-regulated ( S8 Fig ) . “Maternal/embryonic” and “strictly maternal” genes showed less bias , with 202/310 ( 65% ) and 77/163 ( 47 . 3% ) showing down-regulation , respectively , in sig-7 ( RNAi ) embryos . The increase of several strictly maternal genes in sig-7 ( RNAi ) embryos was verified by qRT-PCR ( e . g . S6A Fig ) . The increased abundance of strictly maternal RNAs in sig-7 ( RNAi ) embryos may be an indirect effect of defective zygotic transcription-driven development , causing impaired degradation of maternal RNAs [56] . Our RNA-seq analyses also revealed a role for SIG-7 in RNA splicing . 1431 of the 1549 genes down-regulated in sig-7 ( RNAi ) also registered “isoform differences” in our Cuffdiff analyses ( S2 Table ) . Upon closer examination , many of these sig-7 ( RNAi ) -dependent “isoform differences” appeared to be caused by decreases in exon reads without corresponding decreases in intron reads ( S9A Fig ) . Genome-wide analyses also revealed this trend: the average exon read coverage of genes showing decreased expression in sig-7 ( RNAi ) embryos showed the expected decrease , but the intron read coverage showed little change relative to controls ( S9B Fig ) . Thus , although the amount of RNA for these genes was decreased , the ratio of intron to exon reads for these RNAs increased . Importantly , many of the intron read sequences were linked to exon sequences , indicating they were from unprocessed transcripts , rather than from abnormal persistence of spliced-out intron segments . We confirmed an increase in intron abundance compared to exons for several affected genes by qRT-PCR ( Fig 5A ) . We used primer sets that span intron-exon junctions and exon-exon junctions to distinguish unspliced primary transcripts ( pre-mRNAs ) from spliced mature mRNAs ( mRNAs ) , respectively . The six embryonic genes tested ( sdz-27 , sdz-28 , epi-1 , sqd-1 , vet-2 , end-1 ) displayed the markedly reduced mRNA levels in sig-7 ( RNAi ) embryos observed by RNA-seq . All six also showed significantly increased levels of unspliced RNAs , confirming that the reduced transcripts present in sig-7 ( RNAi ) embryos are enriched for defectively processed RNAs ( Figs 5A and S9A ) . In contrast , RNAs from “upregulated” genes , such as strictly maternal genes , showed the opposite trend . RNA-seq results for these genes showed an increase in exon reads , yet their intron reads stayed relatively constant in sig-7 ( RNAi ) embryos , suggesting a relative enrichment for spliced RNAs relative to controls ( S9B Fig ) . Indeed , intron sequences could not be detected by qRT-PCR for two maternal genes tested ( Fig 5A ) . This result is also consistent with abnormal persistence of fully spliced maternal products , resulting in an apparent enrichment for exon reads relative to intron reads compared to controls . Reads from sequences 5’ to the first exon of many of the down-regulated genes also increased in sig-7 ( RNAi ) embryos relative to controls ( Figs 5A and S9A ) . These reads represent 5’ outrons , which like introns are removed from the primary transcripts . C . elegans exhibits co-transcriptional trans-splicing , in which a common spliced leader transcript serves as a 5’ splice donor , leading to a common 5’ exon that is present on the majority of mRNAs [57–59] . In C . elegans , approximately 70% of mRNAs are reported to be trans-spliced [60] . The outron reads thus represent 5’ nascent transcript sequences that are normally removed by trans-splicing and replaced by spliced leader sequences during transcription . Indeed the 5’ reads enriched in sig-7 ( RNAi ) embryos precisely mark the transcription start sites ( TSSs ) recently identified by GRO-seq and related methods [61–63] . The relative increase of RNA-seq reads corresponding to introns and outrons indicates that depletion of SIG-7 causes defects in both cis- and trans-splicing , the latter of which is only known to occur co-transcriptionally [64 , 65] . Since SIG-7 interacts with Pol II , this strongly suggests that SIG-7 plays an important role in transcription-coupled RNA processing events . A transcription defect could indirectly cause splicing defects by reducing the production of essential splicing factors . This seemed unlikely , since splicing factors in early embryos are available from maternal stores and thus would fall into the class of genes either unaffected or slightly enriched in sig-7 ( RNAi ) embryos . Indeed , our RNA-seq data confirmed this: of 18 conserved C . elegans splicing factors [66] for which significant RNA levels could be detected in control embryos , 8 factors showed a slight increase in sig-7 ( RNAi ) embryos , and 10 factors showed no significant difference between sig-7 ( RNAi ) and control ( S1 Table ) . Thus , the splicing defects in sig-7 ( RNAi ) embryos are unlikely to be due to reduced expression of splicing factors and are instead likely to be directly due to defects in transcription coupled processing . Numerous reports indicate that co-transcriptional splicing is mechanistically coupled to Pol II elongation , and it has recently been proposed that defects in co-transcriptional splicing can affect Pol II elongation [67–74] . We therefore analyzed the genome-wide distribution of Pol II by anti-AMA-1 ChIP-seq in sig-7 ( RNAi ) and control embryos . The ChIP-seq data showed a strong correlation with the RNA-seq data; i . e . , genes showing down-regulation by RNA-seq also showed decreased Pol II occupancy by ChIP-seq ( S10 and S11 Figs ) . We next performed metagene analyses of the Pol II distribution within the body of genes in five of the expression categories described above ( Fig 6 ) . Genes classified as either “soma-specific” or “ubiquitous” showed substantial changes . In these genes , the 3’ enrichment of Pol II observed in control embryos was significantly reduced in sig-7 ( RNAi ) embryos , with 3’ depletion the most obvious in the “soma-specific” class ( Fig 6 ) . 5’ localization was also reduced for the “soma-specific” class , but the effect was less marked than the 3’ reduction . In contrast , genes classified as “germline-enriched” showed little change in Pol II distribution . This result indicates that , as with the RNA-seq data , there is a disproportionate effect of sig-7 ( RNAi ) on genes expressed in embryos , including an effect on steady-state localization of Pol II within gene bodies . The lack of effect on “germline-enriched” loci is not as easy to ascribe to reduced sig-7 RNAi efficiency in parental germlines compared to embryos , since many of these genes include ubiquitously expressed genes transcribed in embryos . The reduced effect of SIG-7 depletion for these genes may be related to the different modes of Pol II regulation observed for germline- and ubiquitously-expressed genes compared to soma-specific genes , the latter of which involve tissue-specific modes of gene regulation [75] . The decrease in Pol II at the 3’ end of gene bodies observed by ChIP-seq suggested that sig-7 RNAi affects the elongation phase of transcription . The phosphorylation of specific residues in the Pol II CTD correlates with different stages of the transcription cycle; e . g . , Ser-5P correlates with initiation and Ser-2P increases with elongation [2 , 17 , 76–78] . We assessed the relative abundances of these CTD phospho-epitopes in sig-7 ( RNAi ) and L4440 control embryos using monoclonal antibodies specific for the different phosphorylated isoforms of AMA-1 ( Fig 7A ) . We observed similar levels of AMA-1 protein in experimental and control lanes , indicating that SIG-7 depletion has little effect on embryonic AMA-1 protein levels . The amount of hypo-phosphorylated Pol II ( hypo-phos; 8WG16 ) was variable between experiments but often higher in sig-7 ( RNAi ) embryos relative to controls . Pol II Ser-5P levels were also variable , but with a slight decrease often observed . In contrast , a significant decrease in levels of Pol II Ser-2P was consistently observed in sig-7 ( RNAi ) embryos ( Fig 7A and 7C ) . The decrease in Pol II Ser-2P is consistent with the decreased 3’ Pol II profile observed by ChIP-seq , and indicates that the elongation phase of transcription is altered in sig-7 ( RNAi ) embryos . In yeast , the addition of Ser-5P to the CTD by TFIIH correlates with recruitment of the histone H3K4-specific methyltransferase Set1 , which in turn leads to an enrichment of H3K4me3 in nucleosomes near the promoter [79–81] . Elongation and increased phosphorylation of Ser2 in turn correlates with recruitment of the H3K36 methyltransferase Set2 and a resulting enrichment of H3K36me3 within the body of the gene as elongation proceeds [13 , 81 , 82] . H3K79me2 is also added to gene body nucleosomes during Pol II elongation [83–85] . We examined sig-7 ( RNAi ) and control embryos by western blot analysis , using antibodies specific for H3K4me3 , H3K36me3 , or H3K79me2 and compared these to total histone H3 ( Fig 7B and 7C ) . We observed a slight decrease in H3K4me3 and a substantial decrease in H3K36me3 and H3K79me2 in sig-7 ( RNAi ) embryos . Thus , like elongation-dependent phosphorylation of Ser2 in the Pol II CTD , elongation-dependent histone modifications are also disproportionately affected by sig-7 ( RNAi ) . We also looked at H3K4me3 and H3K36me3 levels in embryos by immunofluorescence . Consistent with our western blot result , we observed slight decreases in the level of H3K4me3 , most notably in the ~250 cell stage embryos , while H3K36me3 levels were observed to be decreased in all stages after the ~60 cell stage ( Fig 8 ) . The H3K36me3 in the early embryo is predominantly provided by MES-4 , a transcription-independent H3K36 methyltransferase , whereas transcription-dependent H3K36me3 predominates in later stages . The minimal effect on H3K4me3 , a promoter-proximal mark , and the dramatic reduction in H3K36me3 , a mark enriched toward the 3’end of transcribed genes , is consistent with a role for SIG-7 in normal Pol II elongation . In summary , depletion of SIG-7 from C . elegans embryos causes a developmental arrest , likely due to widespread defects in splicing accompanied by a global decrease in transcription of genes required for normal embryogenesis . This transcription defect correlates with a marked decrease in Pol II at the 3’ end of genes and decreases in Pol II CTD phospho-epitopes and chromatin modifications that are hallmarks of elongating Pol II . SIG-7 physically associates with Pol II in vivo and is enriched in chromatin in patterns consistent with association with active transcription , and loss of SIG-7 causes defects in co-transcriptional splicing . SIG-7 is thus required for both transcription and splicing , and while it could directly impact just one process and indirectly the other , it’s possible it coordinates both processes to promote accurate and efficient mRNA production . We report the first genome-wide analysis of a highly conserved , multi-domain nuclear cyclophilin , SIG-7 , that is required for efficient transcription elongation and transcript splicing in C . elegans . Similar to the S . pombe and A . thaliana SIG-7 orthologs ( Rct1 and AtCyp59 , respectively ) , SIG-7 is an essential protein implicated in regulation of the phosphorylation status of important serines in the CTD of Pol II [41 , 43] . Both a maternal supply and zygotic production of SIG-7 are required for normal development at all stages . Depletion of SIG-7 by RNAi treatment results in a substantial decrease in embryonically produced transcripts in embryos . This decrease is accompanied by defective patterns of mRNA splicing , including co-transcriptional trans-splicing . A change in the distribution of Pol II within gene bodies is also observed that , along with reduced Pol II CTD Ser-2 phosphorylation and H3K36 and H3K79 methylation , are consistent with defects in Pol II elongation . The correlation between elongation defects and splicing defects could suggest an interdependency of these two processes in C . elegans , with SIG-7 providing an essential link . The extent to which mRNA processing and Pol II elongation are co-dependent in any organism is controversial , and indeed our results cannot rule out separable functions for SIG-7 in both processes [32 , 73 , 86–91] . This will likely remain controversial , because it is challenging to experimentally discriminate between a splicing defect directly causing an elongation defect versus an elongation defect causing a splicing defect . However , several reports have indicated that a primary defect in splicing can cause defective Pol II elongation . In cultured fibroblasts , depletion of a known splicing factor , SC35 , results in attenuation of Pol II elongation through gene bodies in mammalian cells [29] . In HeLa cells , inhibition of splicing using spliceostatin ( SSA ) or antisense oligos targeting snRNAs resulted in defects remarkably similar to those caused by SIG-7 depletion: early dissociation of Pol II leading to its 3’ depletion and decreases in Ser2P [92] . It thus seems likely that the primary defect in sig-7 mutants is defective splicing , which leads to defective elongation and Pol II dissociation from genes . Metagene analysis of our RNA-seq data showed no increased reads past the annotated TES in RNAi versus control embryos ( S12 Fig ) , indicating that in the reduced instances when Pol II completed elongation , termination was largely unaffected . While the mechanistic roles of nuclear cyclophilins in any organism remain to be determined , yet our results and those from other studies provide important clues . The SIG-7-type nuclear cyclophilins all have a conserved RNA-recognition motif ( RRM domain ) in addition to the peptidyl-prolyl cis-trans isomerase ( PPI ) domain . Studies in S . pombe and A . thaliana demonstrate a role for Rct1 and AtCyp59 , respectively , in the regulation of Pol II CTD phosphorylation [41 , 43] , and in vitro binding experiments show that the CTD of Pol II interacts with the PPI domain of Rct1 [44] . Furthermore , it was shown that Rct1-dependent effects on Pol II CTD phosphorylation is dependent on the PPI domain , indicating that this motif is important for the association of Rct1 with Pol II and regulation of Pol II phosphorylation [44] . The other motifs in the SIG-7 orthologs are involved with RNA interactions . The RRMs of both Rct1 and AtCyp59 were shown to bind a motif present in ~70% of all mRNAs , and AtCyp59 has been shown to interact with pre-mRNAs , supporting a general role in co-transcriptional RNA processing in vivo [42] . Thus , a potential model for SIG-7 is that it binds to the CTD of Pol II through its PPI domain and employs its RRM domain to capture emerging RNAs , perhaps to efficiently recruit them to the spliceosome machinery attached to the CTD . In the absence of SIG-7 , the coordinated interactions between emerging transcripts and the splicing machinery may be compromised , leading to decreased splicing efficiency and as a consequence , elongation may be disrupted through an as yet unknown mechanism . An alternative model is that the PPI domain’s catalytic function in isomerization of prolines may target the Pol II CTD repeat , which in turn may affect Pol II elongation via CTD structural alterations . PPI activity may be regulated by RNA binding and/or RNA processing , and this could provide a mechanistic link between elongation and splicing . Indeed , binding of RNA to AtCyp59 affects the isomerase activity of the PPI domain in vitro [42] . Future studies should investigate the importance of the PPI domain’s catalytic activity and a requirement for it to be structurally linked to the RRM domain . While SIG-7 homologs are found in most eukaryotes , there is no obvious homolog in budding yeast . Introns are relatively rare in budding yeast genes ( present in just ~4% of protein-coding genes ) and the few introns present are small in size [93] . Indeed , SIG-7/Rct1 is among a number of conserved spliceosome components and related proteins that are present in fission yeast , but have been lost from budding yeast [47] . SIG-7 function is thus dispensable in S . cerevisiae , as might be predicted for a protein responsible for coordinating transcription elongation with efficient splicing , as budding yeast have relatively few , small introns to process . The predominance of intron-less genes in budding yeast would presumably make maintaining a protein that was central to linking proper splicing with efficient elongation no longer essential . C . elegans strains were maintained at 20°C . Worms were grown on NGM ( Nematode Growth Medium ) plates unless stated otherwise . Strains: wild type N2 ( Bristol ) , KW1317: sig-7 ( cc629 ) I/hT2[bli-4 ( e937 ) let- ? ( q782 ) qIs48] ( I;III ) , KW2230: sig-7 ( n5037 ) I/hT2 [bli-4 ( e937 ) let- ? ( q782 ) qIs48] ( I , III ) , PD7271: pha-1 ( e2123ts ) III; ccEx ( pBK48 . 1::pC1 ) , KW2309: sig-7 ( n5037 ) I; ckIs33 ( unc-119 , sig-7:gfp:3XFLAG ) II ( below ) . Repetitive sequences within intron 1 of the sig-7 gene prevented PCR-based cloning of the whole gene . We used a fosmid clone ( construct ID: 15087717651452437 A06 ) containing sig-7 engineered with a 3’ GFP::3XFLAG tag obtained from the TransgeneOme project [46] . The fosmid was cut with SphI to generate an 18KB fragment containing the entire operon with sig-7 and two neighboring genes ( CEOP1492 ) . This fragment was blunt ended and inserted into the pCFJ151 MosSCI targeting vector cut with PvuII . This construct , pJA8 , was integrated into an LGII MosSCi targeting site ( ttTi5605 ) by standard Mos-SCI integration techniques [94] . RNAi was performed by feeding HT115 bacteria transformed with plasmids expressing dsRNA targeting the corresponding gene , or carrying the empty L4440 RNAi vector for controls . RNAi embryos: Adult worms were collected from plates and washed with M9 buffer ( 22mM KH2PO4 , 42mM Na2HPO4 , 86mM NaCl , 1 mM MgSO4 ) , bleached with sodium hypochlorite ( 5% bleach with 1 . 0N NaOH ) to isolate embryos . Embryos were placed on NGM ( Nematode Growth Medium ) plates without food overnight . The synchronized hatched L1s were transferred to plates with OP50 bacteria and grown for 36 hours until the L3 larval stage . L3s were washed with M9 buffer 3 times and transferred to induced RNAi plates ( NGM+1mM IPTG+1mM Ampicillin ) pre-seeded with bacteria expressing the desired dsRNA . The worms were grown on RNAi plates for 36 hours , after which the gravid adults were washed with M9 buffer and separated from any extruded embryos by filtration through a 40μm mesh cell strainer ( Fisher Scientific , #22363547 ) . The adults were bleached as described above to collect in utero embryos for analysis . RNAi Adults ( S6B Fig ) : L3 larvae prepared as above were fed dsRNA-expressing bacteria for 55 hours instead of 36 hours and directly processed for total RNA purification and analysis by qRT-PCR . Intact embryos were fixed in 2 . 5%PFA/ethanol [95] or methanol/acetone [96] for all immunofluorescence except for those probed with monoclonal antibody H5 , which was fixed in methanol/formaldehyde [95] . Primary antibodies used were: anti-H3K4me2 ( CMA30 ) 1:1000 END Millipore] , anti-P-granules [OIC1D4 , 1:5 [96 , 97] ) ] , anti-H3K36me3 [ ( CMA333 ) , 1:1000 [95]] , anti-Ser2p RNA pol II CTD ( H5 , 1:500 , Covance MMS-129R ) , anti-GFP ( 1:1000 , Novus NB600-308 ) , anti-AMA-1 ( 1:10 , 000 , Novus 38520002 ) , and anti-FLAG ( M2 , 1:1000 , Sigma F1804 ) . Secondary antibodies used were Alexa Fluor 488-conjugated donkey anti-mouse ( 1:500 , Invitrogen R37114 ) and Alexa Fluor 594-conjugated goat anti-rabbit ( 1:500 , Invitrogen R37117 ) . Samples were mounted in ProLong Gold anti-fade reagent ( Life technologies , P36934 ) and observed under a fluorescence microscope ( Leica DMRXA; Hamamatsu Photonics , Hamamatsu , Japan ) with Simple PCI software ( Hamamatsu Photonics ) . Image J was used for quantification of raw immunofluorescence intensity [98] . Embryos were collected as described above , washed , and resuspended in 3X volume of ice-cold Hypotonc Triton-X buffer [20 mM Tris–HCl ( pH 7 . 4 ) , 10 mM KCl , 10 mM MgCl2 , 2 mM EDTA , 10% glycerol , 1% Triton X-100 , 2 . 5 mM β-glycerophosphate , 1 mM NaF , 1 mM DTT , and Complete protease inhibitors; [99]] . Resuspended embryos were frozen in liquid nitrogen and ground into a fine powder using a mortar and pestle and thawed on ice for 10 min . The suspension was sonicated for 2 min at high setting using a Bioruptor sonicator ( Diagenode Inc . , Denville , NJ , USA ) . The salt concentration was then adjusted to either 150mM or 350mM NaCl , and incubated for 30 min with rotation at 4°C . After an additional 2 min sonication , the lysate was centrifuged for 15 min at 13 , 000g . The supernatant was transferred to new tubes , and 1ml of each lysate supernatant was pre-cleared by incubation with 60 μl of either Protein A ( Life Technologies , 10002D ) or Protein G Dynabead ( Life Technologies , 10004D ) for 30 min with rotation at 4°C . 100μl of each pre-cleared lysate was saved as input sample , and the remaining 900ul was used for immunoprecipitation . Either anti-FLAG ( Sigma , F1804 ) or anti-GFP ( Novus biological , NB600-308 ) for SIG-7 IP and anti-AMA-1 ( Novus biological , 38520002 ) for Pol II ( AMA-1 ) IP were added to the lysate ( 10ug of antibody/2 . 5mg of lysates ) and incubated for 12 hours at 4°C . 60μl of either Protein A or Protein G Dynabeads were added directly to the lysate/antibody mix and incubated at 4°C for 3 hours . Beads were separated from solution using a magnetic bar , washed 2 times for 5 min in Hypotonic Triton-X buffer , and washed twice more with 500mM NaCl hypotonic Triton-X buffer for 10 min at room temperature . For final elution , beads were incubated with 150μl of 2X SDS-PAGE sample buffer for 15 min at room temperature . The final eluates were further analyzed by SDS-PAGE and Western blot . RNAi-treated embryos were resuspended in 4X volume of RIPA buffer ( Thermo Scientific , #89901 ) and 2X volume of glass beads ( Sigma , G8772 ) , and homogenized using a Mini Beadbeater-16 ( BioSpec , Bartlesville , OK , USA ) for 1 min 3 times at 4°C . The homogenized embryos were incubated at 4°C on a rotator for 30 min and further processed in a Bioruptor sonicator ( Diagenode Inc . , Denville , NJ , USA ) at high setting for 10 min to fragment chromatin . The final lysates were centrifuged at 13 , 000g for 10 min at 4°C , and the supernatants were collected . The protein concentration was determined using the Bradford reagent ( Biorad , #500–0006 ) . Supernatants were mixed with an equal volume of 2X SDS-PAGE sample buffer and denatured for 5 min at 95°C , and equal protein amounts were loaded and run on a 4–20% precast SDS-PAGE gel ( Biorad , #456–1094 ) and transferred to PVDF membrane . Transferred proteins were blocked in 5% milk PBST for 1 hour , incubated with primary antibody overnight , and washed 3 times with 1X PBST for 10 min each . After incubation with secondary antibody for 2 hours at room temperature , the blot was washed 3 times with 1X PBST for 10 min each . The washed membrane was incubated with chemiluminescence reagent ( Thermo Scientific , #34087 ) for 5 min , and protein bands were visualized with autoradiography film ( Genesee Scientific , #30–100 ) . The primary antibodies used are the following: anti-FLAG ( Sigma , 1:2000 , F1804 ) , anti-Actin ( Millipore , 1:10 , 000 , MAB1501 ) , H14 ( Covance , 1:3000 , MMS-134R ) , H5 ( Covance , 1:3000 , MMS-129R ) , anti-AMA-1 ( Novus Biological , 1:5000 , 38520002 ) , 8WG16 ( Covance , 1:1000 , MMS-126R ) , anti-H3K36me3 ( Abcam , 1:1000 , Ab9050 ) , anti-H3K79me2 ( Abcam , 1:1000 , Ab3594 ) , anti-H3K4me3 ( Abcam , 1:1000 , Ab8580 ) and anti-H3 ( Abcam , 1:20 , 000 , Ab1792 ) . The following secondary antibodies were used: Goat Anti-Rabbit IgG , HRP-conjugated ( Millipore , 1:2500 , 12–348 ) , Goat Anti-Mouse IgG , Peroxidase-conjugated ( Millipore , 1:2500 , AP124P ) , AffiniPure Goat Anti-Mouse IgM , Peroxidase-conjugated ( Jackson ImmunoResearch Laboratories , Inc , 1:5000 , 115-035-075 ) . Embryos collected after RNAi exposure were washed with M9 , and pelleted embryos were resuspended in Trizol ( 50μl of embryos/300μl of Trizol , Invitrogen ) , snap frozen in liquid nitrogen , and subjected to 3 freeze/thaw cycles . 62μl of chloroform was added and mixed thoroughly by shaking 10 times and spun down for 15 min at 4°C . Nucleic acids were precipitated with 0 . 3M acetic acid in 100% isopropanol and resuspended in 100μl of nuclease-free water . Total RNA was purified using RNeasy kit ( Qiagen , Valencia , CA , USA ) as per the manufacturers’ instructions . cDNA was synthesized from 1μg of purified total RNA using iScript select cDNA synthesis kit ( Bio-Rad , 170–8896 ) . 50ng of cDNA was used for qPCR using SsoFast reagent ( Bio-Rad , 172–5201 ) on CFX96 Real-Time system ( C1000 Thermal Cycler , Bio-Rad ) . The transcript levels of genes analyzed were first normalized to 18S RNA for each sample , and the normalized transcript levels of either sig-7 ( RNAi ) or ama-1 ( RNAi ) experiments were then compared to the transcript levels of L4440 controls to generate ΔΔCT plots of relative transcript levels . The averages of two technical replicates from two biological samples were plotted with standard deviation . Total RNAs were purified as described for qRT-PCR and sent to Axeq Asia ( Seoul , Korea ) for transcriptome sequencing . 1μg of total RNA was used as starting material , and a sequencing library was prepared using TruSeq Stranded Total RNA Sample Prep Kit after treatment with Ribo-Zero ( Human/Mouse/Rat ) for rDNA depletion . Library QC was performed using Tapestation D1000 Screen Tape ( Agilent ) and quantified using KAPA Library Quantification Kit ( for Illumina platforms ) . Clusters were generated by HiSeq PE ( Paired-End ) Cluster Kit v3 cBot , and sequencing was done on a HiSeq2000 with 100bp paired-ends using TruSeq SBS v3-HS kit reagents . RNA-seq reads were quality-checked using FastQC version 0 . 5 . 2 to ensure per-base sequence quality , per sequence quality scores , per base sequence content , per base GC content , per sequence GC content , per base N content , sequence length distribution , sequence duplication levels , kmer-content , and that over-represented sequences were within accepted norms . FastQ Quality Trimmer version 1 . 0 was used to trim reads with less than optimal quality scores . The DE analysis protocol outlined in Trapnell et . al . was used to perform the DE analysis [100] . The quality filtered reads were mapped to C . elegans ( ce10 ) reference genome using TopHat2 version 0 . 5 . TopHat2 internally uses Bowtie2 to map the reads . Mapping results were used to identify splice junctions between exons . Cufflinks version 2 . 1 . 1 was used to assemble transcripts and estimate their abundance . The transcript assembly outputs from Cufflinks were merged into a unified list of transcripts using Cuffmerge . Cuffdiff version 2 . 0 . 2 was then used to quantify gene and transcript expression levels and test them for significant differences . Default parameters were utilized for all steps . This analysis was done in part by the Emory Integrated Genomics Core ( EIGC ) , which is subsidized by the Emory University School of Medicine and is one of the Emory Integrated Core Facilities . For Fig 7B , exon and intron coordinates were obtained from WS2220 . 99 , 830 exons with length ≥ 50bp and 69 , 762 introns of ≥ 50bp were obtained . Custom scripts were used to calculate the average read coverage of exons and introns per gene . For S12 Fig , all RNA-seq samples were scaled to 10 million mapped reads . The averaged values from two biological replicates were plotted using the same pipeline employed for the metagene analysis of RNA Pol II CHIP-seq data , described below . Chromatin immunoprecipitation ( ChIP ) was done as described in Ercan et al . with the following modifications [101] . 1 ) Worms were grown on 15cm peptone rich plates seeded with NA22 bacteria . 2 ) Samples were sonicated using a Bioruptor sonicator at high setting for 40min ( 40sec on/ 20sec off ) . After collection of immunoprecipitated DNA , DNA libraries were prepared as described in [101 , 102] . DNA libraries were sent to Axeq Asia ( Seoul , Korea ) for sequencing . Library QC was done using BioAnalyzer High sensitive DNA chip ( Agilent ) . Clusters were generated by HiSeq PE ( Paired-End ) Cluster Kit v3 cBot , and sequencing was performed on HiSeq2000 with 100bp paired-ends using TruSeq SBS v3-HS kit reagents . Gene set definitions used were as published in [54] . Briefly , ubiquitous genes are defined as genes expressed in 4 different tissue-specific SAGE data sets: germline , neuronal , muscle , and gut [103 , 104] . The germline-enriched category was defined by [55] , although spermatogenesis-specific genes were removed . Germline-specific genes were defined as expressed in the germline either in Reinke et al . or in Wang et al . , and were intersected with the strict maternal gene class in Baugh et al . ; any genes expressed in any of the somatic SAGE expression data of Meissner 2009 were removed [55 , 103–105] . Soma-specific ( any ) genes were defined as expressed in any of the somatic SAGE data sets of Meissner et al . but not in the germline SAGE data set of Wang et al . or the germline-enriched set of Reinke et al . [55 , 103 , 104] . Embryo-expressed is the “strict embryonic” class defined in Baugh 2004 [105] . chrX genes are all the X-linked genes in WS220 . Silent genes are mostly serpentine receptors and were defined in [106 , 107] . The hyper-geometric distribution was used to calculate the significance of the enrichment or depletion of any of the gene sets among the mis-regulated genes in Fig 4B . AMA-1 ChIP-seq data were mapped to WS220 using bowtie [108] . MACS2 was used to obtain peak calls for each replicate of L4440 and sig-7 ( RNAi ) [109] . The broad peak option was found to produce the most appropriate peak calls and a significance cutoff of q = 0 . 05 . The peak calls were mapped to WS220 gene annotations . If AMA-1 peak calls of both replicates of a condition overlapped with a gene body , the gene was called bound by AMA-1 . Meta-gene profiles were produced using custom R scripts . Genes were aligned at their Transcription Start Site ( TSS ) and Transcription End Site ( TES ) , and signal over the gene bodies was averaged in 50 bp windows . The 95% confidence interval of the mean is shown with error bars . To normalize reads between samples AMA-1 peak regions from both conditions were removed . The remaining read coverage was scaled genome-wide so the total number of reads was 2 million reads . Protein sequences of SIG-7 homologues were obtained from the NCBI protein database . The alignment of homologues was generated with ClustalW2 [110] . The conserved protein domain/motif search was done using ScanProsite web-based tool [111 , 112] . The accession numbers for proteins used for alignment are the following: SIG-7 ( CAB03088 . 2 ) , PPIL-4 ( NP_624311 . 1 ) , CG5808 ( AAF56342 . 1 ) , AtCyp59 ( NP_175776 . 2 ) , Rct-1 ( CAB52803 . 1 ) and KIN241 ( CAC35733 . 1 ) .
mRNA splicing can occur co-transcriptionally; i . e . , splicing occurs as the RNA emerges from the RNA Polymerase II holoenzyme during transcription elongation . Recent studies suggest that defective splicing can cause defective transcription elongation , suggesting an interdependency of the two mechanisms . The C . elegans gene sig-7 encodes a nuclear cyclophilin , a highly conserved protein family characterized by the presence of both a peptidyl isomerase domain and an RNA-recognition motif . Studies on sig-7 homologs in plants and fission yeast have shown these proteins to interact with RNA Polymerase II , and indicate they regulate the phosphorylation status of its C-terminal domain . We show that SIG-7 activity is essential for both efficient co-transcriptional splicing and normal RNA Polymerase II elongation and may provide an important link between the two processes .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "phosphorylation", "invertebrates", "rna", "interference", "caenorhabditis", "animals", "dna", "transcription", "animal", "models", "developmental", "biology", "caenorhabditis", "elegans", "model", "organisms", "epigenetics", "embryos", "research", "and", "analysis", "methods", "embryology", "genome", "complexity", "genetic", "interference", "proteins", "gene", "expression", "rna", "splicing", "biochemistry", "rna", "rna", "processing", "nucleic", "acids", "post-translational", "modification", "genetics", "nematoda", "biology", "and", "life", "sciences", "genomics", "computational", "biology", "introns", "organisms" ]
2016
A Conserved Nuclear Cyclophilin Is Required for Both RNA Polymerase II Elongation and Co-transcriptional Splicing in Caenorhabditis elegans
The pathways that trigger exacerbated immune reactions in leprosy could be determined by genetic variations . Here , in a prospective approach , both genetic and non-genetic variables influencing the amount of time before the development of reactional episodes were studied using Kaplan–Meier survival curves , and the genetic effect was estimated by the Cox proportional-hazards regression model . In a sample including 447 leprosy patients , we confirmed that gender ( male ) , and high bacillary clinical forms are risk factors for leprosy reactions . From the 15 single nucleotide polymorphisms ( SNPs ) at the 8 candidate genes genotyped ( TNF/LTA , IFNG , IL10 , TLR1 , NOD2 , SOD2 , and IL6 ) we observed statistically different survival curves for rs751271 at the NOD2 and rs2069845 at the IL6 genes ( log-rank p-values = 0 . 002 and 0 . 023 , respectively ) , suggesting an influence on the amount of time before developing leprosy reactions . Cox models showed associations between the SNPs rs751271 at NOD2 and rs2069845 at IL6 with leprosy reactions ( HRGT = 0 . 45 , p = 0 . 002; HRAG = 1 . 88 , p = 0 . 0008 , respectively ) . Finally , IL-6 and IFN-γ levels were confirmed as high , while IL-10 titers were low in the sera of reactional patients . Rs751271-GT genotype-bearing individuals correlated ( p = 0 . 05 ) with lower levels of IL-6 in sera samples , corroborating the genetic results . Although the experimental size may be considered a limitation of the study , the findings confirm the association of classical variables such as sex and clinical forms with leprosy , demonstrating the consistency of the results . From the results , we conclude that SNPs at the NOD2 and IL6 genes are associated with leprosy reactions as an outcome . NOD2 also has a clear functional pro-inflammatory link that is coherent with the exacerbated responses observed in these patients . Leprosy reactions affect up to 50% of patients [1] . They are episodes that disrupt the natural course of disease , characterized by a strong and abrupt reactivation of immune responses [2–4] . Generally , nerve injury is present , which may lead to permanent incapacities . At the onset of the reactions , high levels of cytokines such as IL-1 , IL-6 , IL-12 , IFN-γ , and TNF [2 , 5–12] have been detected either in serum or skin lesions , and recently , cytokine profiling has identified promising host biomarkers to reaction in patients from Bangladesh , Brazil , Ethiopia , and Nepal [13] . Also , anti-LID-1 antibody levels were found to be high and persistent in multibacillary leprosy patients who developed a reaction , therefore suggesting that this is a putative serological predictive tool [14] . Reactions can be observed prior to , during , or after multidrug therapy ( MDT ) and are classified as type 1 , or reversal reaction ( T1R ) , and type 2 , or erythema nodosum leprosum ( T2R ) . T1R is common among patients exhibiting borderline clinical forms of leprosy and consists of an increased cell-mediated immune response with intense skin and nerve inflammation [15 , 16] . T2R is mainly observed in patients at the lepromatous pole , and is related to acute cellular immune response activation , leading to severe systemic symptoms [17 , 18] . Epidemiological studies [19] suggest that comorbidities and pregnancy are risk factors for reaction outcome [20 , 21] . In addition , medical variables including clinical form , positive bacillary index , introduction of multidrug therapy , age , and gender also play key roles in leprosy reactions [22–24] . The genetic component influencing leprosy reactions , severe nerve impairment , and/or disabilities has been previously suggested [25] . Remarkable similarities were detected between granulomatous inflammatory responses and the presence of polymorphisms at genes such as NOD2 and LRRK2 associated with either leprosy reactions and other inflammatory diseases such as Crohn’s , ulcerative colitis ( UC ) , and inflammatory bowel disease ( IBD ) [26 , 27] . SNPs at NINJ1 , TLR1 , IL6 , and TNFSF8 are clearly associated with reactional phenotypes ( TR1 or TR2 ) [28–32] . Recently , the first GWAS assessing reaction outcome has identified lncRNA as a risk factor to T1R [33] . However , so far few of these genes have been replicated or functionally characterized to support the epidemiological findings . The present study was designed to investigate by means of survival curves the risk factors associated with reactional episodes by testing Brazilian leprosy patients and the effects of 15 SNPs at 8 genes . The markers included 11 SNPs selected to replicate the previously reported associations of the TLR1 , NOD2 , and IL6 genes and the remaining SNPs in 4 candidate genes TNF , LTA , IFNG , and IL10 , in reference to markers previously associated with leprosy per se as an outcome [30 , 31 , 34–37] . Here , we confirmed the association between IL6 and NOD2 SNPs and leprosy reactions among Brazilians , and demonstrated that the NOD2 genotype associated with a decreased risk for developing a reaction is also related to lower levels of IL-6 in the sera of non-reactional patients . Written informed consent was obtained from all individuals included in the study as required by the Research Ethics Committee at Fiocruz ( CEP-Fiocruz Protocol 151/01 ) . Our institutional ethical committee board allowed to disclose data as de-identified dataset information regarding clinical parameters and genetic/functional data specific to the research article . We performed a study including patients with a confirmed leprosy diagnosis between 1985 and 2008 that attended Souza Araújo Outpatient Reference Unit , Fiocruz , Rio de Janeiro , who were followed for the development of a leprosy reaction ( outcome ) . Experienced professionals performed the leprosy and leprosy reaction diagnoses after careful clinical examinations and histopathological analyses . Leprosy patients were classified according to Ridley–Jopling criteria [38]: I , indeterminate; TT , tuberculoid; BT , borderline tuberculoid; BB , borderline borderline; BL , borderline lepromatous; LL , lepromatous leprosy , and were treated as specified by the World Health Organization ( WHO ) according to the multibacillary ( MB ) and paucibacillary ( PB ) classifications . Reaction occurrence , as well as classification as either T1R or T2R , were determined by clinical examination and confirmed by histopathological evaluation [15 , 39] . Follow-up started upon initiation of leprosy treatment ( MDT/WHO ) and stopped on the day of the first reaction episode ( event ) or the date of the last follow-up available . When the date of the last follow-up was not available , we limited the observation time to 3 years after start of follow-up . Therefore , we defined the patients from this group as not reactional , considering that the great majority of the patients who develop reaction do so in the first year after the diagnosis . We excluded patients classified as TT since they are not at risk of developing reactions and also excluded individuals without available information regarding their treatment dates . Additional variables such as gender , age , ethnicity , and leprosy relapse were retrieved from each patient’s medical record . For the cytokine quantification ( functional study ) , we used serum from a group of 84 leprosy patients diagnosed at the same outpatient unit . All the characteristics of the first-time patients diagnosed with leprosy are detailed in the S1 Table . The present study was designed to investigate the association of TNF/LTA , IL10 , IFNG , IL6 , TLR1 , and NOD2 genes with leprosy reactions . The SNPs TNF -308G>A ( rs1800629 ) , LTA +252A>G ( rs909253 ) , IL10 -819C>T ( rs1800871 ) , and IFNG +874T>A ( rs2430561 ) were selected based on previously published data which showed their association to leprosy per se among Brazilians [34 , 35 , 40] . The SNPs rs2069832 , rs2069840 , and rs2069845 were selected as tags that covered haplotype bins at IL6 locus . The tag SNP search was based on data from the HapMap Genome Browser release #27 ( http://hapmap . ncbi . nlm . nih . gov/ ) using Caucasians or African reference populations ( CEU and YRI , respectively ) . The same strategy led to the selection of the SNPs rs751271 , rs2066843 , and rs748855 at the NOD2 gene . The remaining NOD2 SNPs ( rs7194886 , rs9302752 , and rs8057341 ) were marker candidates retrieved from the literature [41] . TLR1 SNPs rs5743592 and rs4833095 were included in this study in order to replicate previous findings regarding their association to leprosy per se and reactional episodes [29 , 30 , 42–44] . DNA was extracted from frozen blood samples using a salting-out precipitation method [45] . SNPs at TNF/LTA , IFNG , IL10 , and TLR1 loci were genotyped as described [34 , 35 , 40 , 42] . SNPs at the NOD2 and IL6 genes were genotyped using TaqMan® assays according to the manufacturer’s instructions ( Life Technologies , CA , USA ) . Briefly , amplifications were carried out in a final volume of 5 μL containing 2 . 5 μL of the TaqMan® Universal Master Mix , 0 . 125 μL of the TaqMan mix ( primers and probes ) , and 20–50 ng of template . PCR reactions were performed on ABI Prism 7000 and StepOne Plus Sequence Detection Systems ( Life Technologies , CA , USA ) . Individuals with missing genotypes were excluded from the statistical analysis . Serum cytokine levels from leprosy patients were quantified by ELISA , using Millipore’s MILLIPLEX® Human Cytokine/Chemokine panel commercial kit , according to the manufacturer’s instructions . The following cytokines were included in the kit: IL1-β , IL-4 , IL-6 , IL-10 , IFN-γ , IL-12p40 , IL-12P70 , IL-13 , IL-17 , TNF-α . First , we performed a descriptive analysis to compare the influence of the considered variables on the length of time before the development of a leprosy reaction using survival curves . Proportional hazards assumption for different covariates in the Cox regression model was tested using Schoenfeld residuals [46] . All variables were adequate in the model , except for the clinical forms variable , which was added as strata in the model to avoid the effect of non-proportionality over time in the analyses . The genotype curves were obtained by the Kaplan–Meier method and compared using the log-rank statistic . Age at leprosy diagnosis ( continuous variable categorized as ≤ 40 years old and > 40 years old ) , gender , ethnicity , leprosy relapse , and clinical forms were the analyzed variables . Leprosy clinical forms were grouped into three levels according to the risk of developing leprosy reactions: I/BT/BB , BL , and LL . The median time until event ( MST ) was calculated for each variable using Kaplan-Meir method . Then , to estimate the associations between genetic markers and leprosy reaction ( outcome ) , crude and adjusted hazard ratios ( HR and aHR , respectively ) at 95% confidence interval and p-value were calculated using the Cox proportional-hazards regression model . Variables that showed significance with a reaction outcome in the survival curves were used to adjust the HR in the Cox model ( possible confounders ) . For the cytokine measurements , the median values from each genotype group were compared by the Mann–Whitney U test ( two groups of comparison ) or by an ANOVA Kruskal–Wallis test ( three groups ) . Missing data was excluded from the analysis . The statistical analyses for the Kaplan–Meier and Cox proportional-hazards regression model , as well as the survival graphical curves , were performed in the R environment for Windows [47] , version 3 . 1 . 2 , using the packages “survival” and “genetics . ” The cytokine analysis was conducted using GraphPad Instat software ( GraphPad Software 3 . 0 for Windows , San Diego , CA ) and considered statistical significance to be a p-value < 0 . 05 . Of 567 potentially eligible patients , 120 were excluded due to missing follow-up information , which are described in S2 Table . Therefore , a total of 447 patients were enrolled in the genetic study among whom 222 developed leprosy reactions with an overall median survival time ( MST ) of 165 weeks until the reaction’s occurrence ( Table 1 , S1A Fig ) . The clinical characteristics of the patients enrolled in the study , including their age , gender , ethnicity , leprosy relapse , and clinical classification are summarized in S1 Table and Table 1 , along with the results of the log-rank test . As observed , the length of time before a leprosy reaction , as described by survival curves , shows a significant difference when considering the variables of gender ( p = 0 . 002 ) and clinical form ( p < 0 . 001 ) . Regarding the median time until event ( MST ) , men developed reactions earlier than women ( MST = 86 weeks and 182 weeks , respectively ) . According to the clinical forms , the lower median was observed among LL patients when compared to the other groups ( MSTLL = 39 weeks vs MSTBL = 59 weeks and MSTI/BB/BT = 182 weeks ) , as presented in Table 1 , S1B and S1C Fig . To investigate the association between genetic markers and reaction outcomes we constructed survival curves for each SNP ( stratified by genotype groups ) as a descriptive approach to evaluate the occurrence of reaction in leprosy patients over time . We estimated that our sample was sufficient in detecting a genetic risk effect of 1 . 8 with a statistical power of 79% , considering single nucleotide variants with a minor allele frequency of 0 . 10 under the additive model . From the 15 candidate SNPs ( total counts and frequencies detailed at S3 Table ) , 2 showed different Kaplan–Meier curves between genotype groups ( log-rank p-value < 0 . 05 ) . Patients with the AG genotype for IL6 rs2069845 developed reactions faster ( MST = 116 weeks ) than other genotypes ( MSTAA = 195 , MSTGG = 295 weeks; p = 0 . 023 ) , as illustrated in Fig 1A . In addition , NOD2 rs751271 homozygous patients ( TT ) developed reactions significantly earlier ( MST = 62 weeks ) than the GT ( MST = 165 weeks ) and GG ( MST = 194 weeks ) genotypes ( Fig 1B ) . The log-rank results from the Kaplan–Meier comparisons of all SNPs tested are presented in S4 Table . To estimate the SNP effects on reaction occurrence , the Cox proportional-hazards regression model was applied including adjusted analyses for gender and clinical forms ( strata ) . The NOD2 rs751271-GT genotype and G-allele carriers ( GT/GG ) were associated with protection against a reaction’s occurrence , even after adjustment ( aHR = 0 . 45 , p = 0 . 002 and aHR = 0 . 56 , p = 0 . 01 , respectively ) , as presented in Table 2 . On the other hand , AG genotype and AG/GG carriers of IL6 rs2069845 were associated with having an increased risk of developing a reaction ( aHR = 1 . 88 , p = 0 . 0008 and aHR = 1 . 73 , p = 0 . 002 , respectively ) . Genotypic information for individual patients is detailed in S5 Table . We evaluated serum cytokine levels to investigate whether the SNPs correlated with the inflammatory profile in leprosy patients . Of 84 individuals ( S1 Table ) , 39 developed reactions , including 22 T1R and 17 T2R subtypes . Overall , IL-6 , IL-10 , and IFN-γ dosage ( pg/mL ) confirmed the literature reports in which reactional patients exhibited higher pro-inflammatory and lower anti-inflammatory cytokine levels ( Fig 2 ) . The quantification showed that IL-6 ( median: reaction group 0 . 68 vs . no reaction group 0 . 25 , p = 0 . 016 ) and IFN-γ ( median: reaction group 0 . 61 vs . no reaction group 0 . 25 , p = 0 . 017 ) were higher in the reaction group ( Fig 2A and 2B ) . Also corroborating previous findings , IL-10 levels were lower in the reaction group ( median: reaction group 0 . 40 vs . no reaction group: 0 . 71 , p = 0 . 027 ) , as shown in Fig 2C . The cytokines quantification for each of the patients analyzed in the functional approach are described in S6 Table . The remaining cytokines from the kit were not analyzed because they did not reach the detection limit . Then , the cytokine levels from patients were stratified in accordance with IL6 and NOD2 SNPs ( Fig 3 ) . The cytokine dosage showed a heterogeneous distribution among the patients , and stratified analysis showed no differences between groups in either the patients overall ( S2 Fig ) or the reactional patients ( S3 Fig ) . Considering the non-reactional leprosy patients , IFN-γ and IL-10 levels showed no differences between genotype groups for either of the tested polymorphisms ( Fig 3C–3F ) . Neither was IL-6 production associated with IL6 rs2069845 SNP ( Fig 3A ) . However , individuals carrying NOD2 rs751271-GT showed a borderline difference ( p = 0 . 05 ) , suggesting lower IL-6 production ( GT median: 0 . 008 ) as compared to GG individuals ( GG median: 0 . 53 ) ( Fig 3B ) . Unfortunately , we did not have patients with the NOD2 rs751271-TT genotype for a functional analysis . In the present study , we have found two polymorphisms that , independently of other non-genetic risk factors , were associated with inflammatory reactions in leprosy . Considering that reactions are serious , and remain an unpredictable outcome , we decided to include the time of occurrence in the model using survival analysis , which tests the prognostic value of polymorphisms as genetic markers . As a result , we observed that patients with NOD2 rs751271-TT or carrying an IL6 rs2069845-G allele developed reactions faster than those with other genotypes/alleles , suggesting that these SNPs are good prognostic markers for reactional episodes . Moreover , NOD2 rs751271-GT individuals had a statistically significant decreased risk of developing reactions and , in a smaller group of patients , also showed lower serum IL-6 , but not IL-10 or IFN-γ , levels . SNPs at NOD2 were previously associated with protection against reaction outcomes in a population from Nepal [36] . NOD2 acts as a sensor of mycobacterial components and contributes to bacterial killing by activating the NF-kB pathway , an inflammation cascade , and an interleukin-32-dependent pathway . Also , NOD2 appears as a key player in autophagy , including its genetic association with Crohn’s disease [48–50] . Uncontrolled inflammation is driven by an inability to clear bacteria through autophagy in patients carrying specific alleles , and that is followed by exacerbated pro-inflammatory response . In fact , here , NOD2 rs751271-TT individuals have a higher risk of developing a reaction per se . None of the individuals enrolled in the functional analysis exhibited the TT genotype . Our results suggest that rs751271-GT individuals have the lowest risk of a premature reactional outcome and these individuals have lower productions of IL-6 during an unreactional state . It could be hypothesized that rs751271-GT carriers have the most balanced production of cytokines—in this case IL-6 that could be used as a readout—and during the natural course of leprosy ( in non-reactional patients ) , these patients showed a decreased risk of experiencing a reactional episode . On the other hand , rs751271-GG individuals have higher IL-6 and a moderate risk of developing reactions ( survival curves ) . TT individuals would possibly exhibit a similar high IL-6 production pro-inflammatory profile , although unfortunately we did not have any patients with this genotype in our sample so we could not perform such a comparison . This hypothesis is illustrated in Fig 4 . The results from different studies have demonstrated the emergence of an inflammatory immune response during reactional episodes [4] , and a pro-inflammatory signature to reactions was suggested recently by Khadge and colleagues [13] . Probably , polymorphisms are contributing to control the inflammatory profile in leprosy patients , influencing the reaction’s development , when it occurs , and its severity . IL6 has been previously associated with leprosy reactions in other populations and by using different study designs . The IL6 rs2069840-G allele was associated with protection against and IL6 rs2069845-G with risk of T2R in a population from Goiania , Brazil [31] , which is , at least in part , corroborated by our genetic results indicating that IL6 rs2069845-G is associated with an increased risk of reactional episodes per se . IL-6 is a pleiotropic cytokine that acts in the acute inflammatory response and activation of Th1 and Th17 lymphocytes . IL-6 also inhibits T regulatory ( Treg ) cells , mediating the balance between pro-inflammatory and immunosuppressive T-cells [51] , which is consistent with the central role of IL-6 as a prognostic molecule in chronic inflammation . Our data confirmed higher levels of IL-6 among reactional patients , but a stratification of IL-6 genotypes failed to confirm that IL-6 rs2069845-G carriers secrete more IL-6 in serum . In fact , SNPs only partly and transiently regulate gene expression and maybe the period for the detection of IL-6 in the sera that could be impacted by SNPs may be different from what we recovered in the present study . We have additionally investigated the classic candidate genes TNF/LTA , IFNG , and IL10 , which were consistently associated with leprosy per se [35 , 40 , 52] but , despite the central roles of these cytokines in leprosy reactions [5] , were not genetically associated with reaction outcomes in our study . We were also unable to detect any association of TLR1 SNPs , although the previously reported data regarding the association of this gene in other populations [30] . Variations in study designs and linkage disequilibrium patterns may explain these divergent results . Recently , SNPs at TNFSF pathways were consistently associated with T1R outcomes in populations from Vietnam and Brazil [32 , 53] . These and other possible candidates still need to be confirmed among our Southern Brazilian population . Also , it is necessary to understand the mechanisms that underlie the genetic susceptibility of reactional episodes since in this study we observed that the presence of risk alleles do not have additive or synergistic effects in a way that would make the combination of some of these most important SNPs as a genetic test feasible . We understand that the sample size could be considered a limitation of our study , mainly due to low patient compliance , however , the results confirm certain classical variables ( gender and clinical forms ) as risk factors for reaction , which implies that our genetic data is consistent and should be used in a score to estimate the risk of multibacillary patients developing reactional episodes . Also , we have to consider other possible sources of bias such as non-genetic variables that could determine the development of reactions for which we did not have any available information , amongst them pregnancy , modifications in therapeutic scheme , and comorbidities . Nevertheless , the specific variables for which we have information were utilized as covariates and adjusted for in the Cox model . The issue of correcting for multiple comparisons has been recently discussed in epidemiology . Even though it is desirable to control for type I error , there are many instances when it is not sound to consider that a priori all the hypotheses being tested are false . As it is our case we have performed a genetic study at candidate gene/SNPs level , we did not consider it applicable to perform corrections for multiple comparisons , since it could increase the Type II error and not finding a true association [54] . Despite the studies suggesting the influence of host and environmental factors on leprosy reactions , the complete mechanisms of their occurrence remain unclear . However , there is no doubt that reactions are the main cause of physical disabilities . Results obtained from a prospective study on leprosy patients from Brazil showed that 30% of the reactional cases were associated with persistent physical impairment [55] . The development of a prognostic panel with the capacity to predict the likelihood of progressing to reactions indicates a possible strategy that could contribute to the surveillance of patients with greater chances of developing clinical complications and possibly interfere with prophylaxis in order to prevent future disabilities .
Leprosy reactions are abrupt inflammatory episodes that can occur before , during or after treatment . Although reactional episodes are the main cause of physical disabilities in patients , there are few genetic studies evaluating this outcome . We studied the influence of both risk factors and genetic markers in leprosy reaction outcomes . For this , we used a population including patients that either did or did not develop a reaction , and we performed a characterization of the genetic markers indicating susceptibility to leprosy . As a result , we identified how variables such as gender and clinical forms influence the amount of time before the occurrence of a reaction . In addition , we highlighted genetic markers related to the timing of the reaction’s occurrence , and associated them with reaction susceptibility . These markers were also related to the production of anti- and pro-inflammatory cytokines . Our study describes the risk factors and genetic components associated with leprosy reactions and can help in the prognosis and monitoring of patients to prevent clinical complications .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[ "innate", "immune", "system", "medicine", "and", "health", "sciences", "immune", "physiology", "cytokines", "pathology", "and", "laboratory", "medicine", "enzyme-linked", "immunoassays", "immunology", "tropical", "diseases", "variant", "genotypes", "genetic", "mapping", "biochemical", "analysis", "bacterial", "diseases", "developmental", "biology", "genetic", "predisposition", "enzyme", "assays", "signs", "and", "symptoms", "molecular", "development", "neglected", "tropical", "diseases", "bioassays", "and", "physiological", "analysis", "immunologic", "techniques", "research", "and", "analysis", "methods", "infectious", "diseases", "inflammation", "immunoassays", "immune", "response", "immune", "system", "diagnostic", "medicine", "heredity", "physiology", "genetics", "leprosy", "biology", "and", "life", "sciences", "human", "genetics", "genetics", "of", "disease" ]
2017
Genetic polymorphisms of the IL6 and NOD2 genes are risk factors for inflammatory reactions in leprosy
Rift Valley fever ( RVF ) is a viral mosquito-borne disease with the potential for global expansion , causes hemorrhagic fever , and has a high case fatality rate in young animals and in humans . Using a cross-sectional community-based study design , we investigated the knowledge , attitudes and practices of people living in small village in Sudan with respect to RVF outbreaks . A special One Health questionnaire was developed to compile data from 235 heads of household concerning their knowledge , attitudes , and practices with regard to controlling RVF . Although the 2007 RVF outbreak in Sudan had negatively affected the participants’ food availability and livestock income , the participants did not fully understand how to identify RVF symptoms and risk factors for both humans and livestock . For example , the participants mistakenly believed that avoiding livestock that had suffered spontaneous abortions was the least important risk factor for RVF . Although the majority noticed an increase in mosquito population during the 2007 RVF outbreak , few used impregnated bed nets as preventive measures . The community was reluctant to notify the authorities about RVF suspicion in livestock , a sentinel for human RVF infection . Almost all the respondents stressed that they would not receive any compensation for their dead livestock if they notified the authorities . In addition , the participants believed that controlling RVF outbreaks was mainly the responsibility of human health authorities rather than veterinary authorities . The majority of the participants were aware that RVF could spread from one region to another within the country . Participants received most their information about RVF from social networks and the mass media , rather than the health system or veterinarians . Because the perceived role of the community in controlling RVF was fragmented , the probability of RVF spread increased . Global outbreak of zoonotic infections , not only affect human and animal health but also affect food security , and socio-economic stability . To control such outbreaks require local as well regional cooperation . Most zoonotic outbreaks begin and occur in settings where resources are poor and where the outbreak severely affect the local community [1 , 2] . These outbreaks spread both within and outside the country of origin , often with devastating consequences [3] . Zoonotic infections originate and spread at the interface between humans , animals , and their environments , making them candidates for the One Health approach to disease control [4–8] . Although international organizations , government authorities , and academic institutions , believe the One Health concept should be a part of a local community’s response to zoonotic infection , the One Health concept is rarely implemented at the community level . It is undeniable that community involvement is crucial in reducing the risk of zoonotic diseases at the interface between animal-human and their ecosystem . Rift Valley fever ( RVF ) is a mosquito-borne viral disease affecting both humans and animals . It can be transmitted by mosquito bites or by direct contact with infected animals , their fluids , or products derived from them [9] . The disease in humans varies from a mild influenza-like illness to more severe forms such as hemorrhagic fever , renal failure , encephalitis , retinitis and miscarriage [9 , 10] . Because there is no approved human vaccine or treatment available for RVF , RVF poses a major threat to public health [9] . The infection causes so called “abortion storm” in livestock and deadly epidemics in young animals , with severe consequences for local and national economies [11] . RVF is present in Africa , and as of 2000 , it has spread to the Arabian Peninsula [12 , 13] . Environmental changes , international travel , trade , and the spread of RVF virus ( RVFV ) and vectors outside of Africa highlights the potential for its global spread [14–17] . Sudan is an agricultural country with diverse ecology and has the second largest livestock population in Africa . Most of the Sudanese population depends on livestock for food and income . Our previous studies of the RVF 2007 outbreak in Sudan found a gap in knowledge regarding the role of the local community [2 , 18] . In this study , we used a bottom-up approach to evaluate the knowledge , attitudes and practices that affect how the local agro-pastoralist community in Sudan confronts an RVF outbreak . In accordance with the guidelines of strengthening and reporting of observational studies in epidemiology ( STROBE ) , a cross-sectional community-based study was conducted in March 2013 at the Mabroka Sagadi village in the Managil locality , Gezira State , Sudan ( Fig 1 ) . The Managil locality is close to irrigation canals . The local people are mainly farmers and shepherds and the locality is home to about 2 million livestock ( mainly cattle , sheep and goats ) . The annual rainfall in the region varies between 100 and 350 mm; however , in 2007 the rain level reached up to 400 mm [19] . The rainy season is from mid-June to September . Managil is located in the Savannah zone and the temperature ranges from 25°C to 46°C . The study area included all the states affected by the 2007 RVF outbreak ( Fig 1 ) . Most of the reported human cases ( n = 402 ) [20] were found in Gezira , so this state was selected . In Gezira , the Managil locality reported the highest number of cases and the village of Mabroka Sagadi had nine recorded human RVF cases during the 2007 outbreak . At the time of the study , there were 5 , 830 inhabitants in Mabroka Sagadi village . Of the 641 households in the village , 240 households were randomly selected , with a response rate of 98% ( n = 235 ) . The household head was defined as the person who in charge of the household and any dependents . We interviewed the household heads and all of them were men or women at least 15 years of age . To develop the One Health questionnaire to collect data about RVF at the human-animal-environment interfaces , we intensively searched the literature for information that would help us develop questions regarding the disease and we searched for the best possible answers to these questions supported by the medical literature . This research resulted in the One Health questionnaire that was designed to compile relevant data on RVF in humans , animals , and the environment . Originally written in English , the questionnaire was eventually translated into Arabic ( S1 Table ) . The questions were open ended and the participants were allowed to provide more than one answer . The participants’ answers were compared to the answers listed in the questionnaire but these answers were hidden from the participants to avoid leading questions . If the participant’s answer was not in the listed answers , it was added to the category named “other” . A two-day training workshop was held for data collectors ( public health officers acquainted with the study area ) to discuss the objectives of the study , the contents of the questionnaire , and the methods of the study . A pilot study was conducted in the village of Algila ( Fig 1 ) . Algila had similar socio-cultural and ecological characteristics to those of the final study area , and it was also affected by the 2007 RVF outbreak . The findings were analyzed and used to update the questionnaire for the full-scale study . A pre-study field visit to the study area was conducted to build trust , explain the study objectives , mobilize the community leaders , and ask the community to be involved in all parts of the study . This led to a sense of ownership and empowerment . Thus , successful face-to- face interviews with the heads of households at their home could take place in a friendly environment . The thematic areas covered by the One Health questionnaire were socio-demographic considerations , knowledge of RVF in animals and humans , attitudes and practices regarding RVF , and environmental aspects of RVF ( S1 Table ) . The data were coded , entered into Microsoft Access , and checked for data entry errors by re- entering 10% of the data from the questionnaires . The data were exported and analyzed using STATA version 12 ( Stata Corp LP . College Station , TX , USA ) . Ethical clearance was granted from the Ministry of Health , Gezira State , Sudan . All participants were informed about the objectives of the study and about the confidentiality of the information and results . The participants signed an informed consent document for participation and they were free to leave the study at any time . The study included slightly more women ( 53% ) than men ( 47% ) , and almost 67% had a low level of education ( less than higher secondary school level ) with no significant gender difference . Just over half ( 55% ) were above 35 years of age , the vast majority ( 87% ) of the heads of households were married , and their household had at least six members ( Table 1 ) . Most women were housewives , while men were mostly occupied with farming ( Table 1 ) . Most of the households bred animals ( 72% ) ( Table 1 ) . Cattle were most common , followed by goats and sheep . Animals were kept at home ( 44% ) , near the home ( 22% ) , or far away from home ( 6% ) . Around 25% of the households had members who worked as temporary herdsmen . Most of the households ( 71% ) used their own livestock products as the main source of food , and just under half ( 43% ) sold livestock as a source of income . The 2007 RVF outbreak negatively affected many households ( 46% ) , including disrupting the availability of food and livestock trade for about 33% of the households . The awareness about RVF in general was high: 80% of the heads of households heard about the disease . About 9% of the heads of households had seen people who had contracted RVF during the 2007 outbreak . More than half of those interviewed stated that RVF is a zoonosis that affects both animals and humans , and that RVF had been a serious health problem in the area during the 2007 outbreak . The most common sources of information in the community about an RVF outbreak was their social networks of relatives and friends ( 54% ) and mass media ( 23% ) ; a less common source of information was the health system ( 6% ) or others such as veterinarians ( 5% ) . According to the respondents , there was higher livestock mortality in the area during the 2007 RVF outbreak than the year before ( 2006 ) and the year after ( 2008 ) the outbreak . During the 2007 RVF outbreak , 34% of the heads of households had experienced death of livestock; in 2006 and 2008 only 17% had experienced death of livestock . This difference , however , could not be confirmed from official reports . Although the community experienced higher livestock mortality , only some mentioned known symptoms in livestock that died , such as nasal and ocular discharge and hemorrhagic diarrhea ( Table 2 ) . Regarding sick livestock , known RVF symptoms were not mentioned by the majority: only 20% stated hemorrhaging and less than 9% mentioned fever and refusal to eat ( Table 2 ) . Likewise , RVF symptoms in humans were not well known to the majority of the respondents: 25% stated that fever and hemorrhagic symptoms were most common . Most of the respondents did not know how livestock become infected with RVF . For human infection , ( 21% ) stated that humans are infected through uncooked meat while 13% suggested direct contact with livestock . Only a few respondents suggested other mode of transmission such as raw milk ( 9% ) , and mosquito bites ( 6% ) . The most common answer on how humans should avoid RVF , was to avoid uncooked meat and handling of sick livestock . Avoiding livestock that had suffered miscarriage was the least important according to the results of the survey . Although few thought that RVF is a contagious disease , more than half of the respondents expected to get RVF when it was present in their area―either due to animal-to-human contact or human-to-human contact ( Table 3 ) . One-quarter said that they would avoid contact with neighbors who they suspected of having RVF . About half of the participants ( 45% ) were in favor of patient isolation as a preventive measure during outbreaks and 12% stated that they would avoid contact with RVF patients . Mosquito nets on beds were used by 60% of the respondents , but less than half of these ( 40% ) were impregnated . The majority ( 60% ) had noticed an increase in mosquito population during 2007 . To prevent RVF in livestock , the respondents suggested isolation of sick livestock ( 21% ) and vaccination ( 14% ) . In addition , the majority of respondents ( 73% ) strongly recommended quarantine of RVF-infected livestock . The survey revealed that two-thirds of the participants had a positive attitude about medical treatment against RVF in humans . However , only one-third thought that medical treatment could be used for livestock . They also indicated that veterinarians should diagnose RVF in livestock and health workers should diagnose RVF in human . Around 75% of the heads of households knew where to seek medical treatment , either in the public or the private health sector . If an outbreak occurs , about 40% said they should notify the veterinary authorities about the death of livestock . Almost all the respondents ( 99% ) stressed that they would not get any compensation for their dead livestock if they notified the authorities . When asked if any of their own livestock had had RVF during the 2007 outbreak , 13% said yes . These suspected cases were not confirmed by the authorities . When we asked which disciplines need to work together in order to control RVF , only eight participants ( 3% ) stated that veterinary and health authorities should work together . The majority believed that human health authorities ( 50% ) rather than veterinary authorities ( 15% ) should work to control RVF . With regard to the community’s role in confronting RVF , the respondents suggested that the community should improve its hygienic measures ( 22% ) , its health education ( 18% ) and its vector control ( 8% ) . The majority of the study participants ( 70% ) were aware that RVF could spread from one region to another within the country . In addition , 66% of the participants revealed that they were aware of the livestock trade ban associated with RVF outbreak inside and outside the country . We used the 2007 RVF outbreak in Sudan as a case study to investigate , the knowledge , attitudes and practices , from the One Health perspective that affect the involvement of the local community in disrupting an emerging zoonotic infection . Most of the measures aimed to control RVF were formulated by authorities ( i . e . , health and veterinary ) to be implemented by the local communities ( a top-down ) . For RVF , it seems as these actions are not enough as disease emergence continues . We believed that insights from the local community ( a bottom-up approach ) about RVF prevention and control could help stop the spread of RVF outbreaks . This bottom-up approach could be a tool to better combat the transmission and the spread of RVF in the regions where the outbreaks are initiated , and could enhance a top down approach . To identify the important factors we used the unique One Health approach to gather information about livestock , people , and the environment . We also investigated whether the community considered the integration of the One Health approach of health , veterinarians , and environment authorities as the best strategy to confront RVF and how they could contribute to RVF control . Although the community studied had experienced a large outbreak , few had proper knowledge of RVF regarding cause , mode of transmission , symptoms , prevention , and control . This lack of knowledge increases the chances of RVF spreading to neighboring areas and would prevent the community from confronting the disease . The knowledge of risk factors for RVF was insufficient , because the community only practiced some of the measures known to prevent RVF [21] ( e . g . , eating cooked meat and staying away from sick livestock ) while exposing themselves to other risks ( e . g . , exposure to mosquitoes and handling miscarriage livestock ) . As in most rural areas , these communities in rural Sudan depend on their livestock as a source of food and income , and often breed and raise their livestock inside or near their homes [22] . The majority did not know that handling livestock that had miscarried was a serious risk factor for RVF [23] . This lack of knowledge could pose a significant risk , especially for rural women , as they tend to take care of livestock at home . Although the community lives near irrigation canals , which serve as a breeding site for mosquito vectors , the majority ignored mosquitoes as a source of infection . The lack of knowledge of risk factors for transmission could explain the high number of cases reported in the 2007 RVF outbreak in Sudan [18] . Like other mosquito-borne diseases , RVF is associated with heavy rains [23] , so the authorities could update communities about rain forecasts . This knowledge would encourage a participatory role of local communities in integrated vector control management similar to the malaria control programs that have been established in most countries where malaria is endemic . The infrastructure of such programs such as established vector control management and the use of impregnated mosquitoes bed nets could also help control other types of mosquito-borne viruses in endemic regions[24 , 25] . Therefore , human behavior contributes to the disease emergence . To control the spread of RVF , it is essential to understanding how local communities interact with livestock and the environment . We expect that social scientists , who are well equipped to deal with human behavior changes , will be able to find acceptable ways for rural communities to practice better animal husbandry [26] . For control of RVF , we found the main focus in the community was human health and access to regional hospitals , particularly in the rainy season when the roads are difficult to navigate . Notably , the animal dimension to confronting a zoonosis such as RVF was not well understood . To better implement the One Health approach , authorities could work together with communities to prevent and control emerging zoonotic diseases . This is particularly important because the veterinary services might not be able to cover a vast country like Sudan , where the veterinarians are based in the capital of each locality . The veterinarians visit remote villages for vaccination or investigation of suspected cases of abnormal livestock diseases and visits during rainy season are very difficult due to flooded and otherwise impassable roads . In such a context , voluntary animal health workers from the local communities could be trained to identify livestock diseases , including RVF . This work would take place in collaboration with veterinarians , who have an increasingly important role in global health [27] . Similarly , human voluntary health workers could be trained to identify human diseases . These volunteers could be selected in cooperation with community leaders , which would ensure successful collaboration and communication between health care providers and the community . The sustainability of such a system would depend on a rapid response and support from the authorities when needed by the local communities . These suggestions are in keeping with a participatory approach that regards farmers as being effective partners to curb zoonoses [28 , 29] . The rural household economy is affected by RVF outbreaks , regarding both food security and disrupted incomes . There were two opposing interests . The community was only interested in interventions to curb the disease that would not result in the culling of livestock without compensation . The absence of a compensation system weakened the motivation to report early cases in livestock to veterinary authority , a requirement if RVF is to be halted before infecting humans . This lack of compensation could be a possible explanation for the delay in reports of RVF in livestock . If RVF had been identified in livestock early , then livestock as well as human RVF outbreaks in Sudan in 2007 and in Kenya in 1997‒1998 might have been prevented [18 , 30] . To support the devastated rural economy due to RVF outbreaks [3] , a new policy of compensation for culled or dead livestock must be developed . The respondents stressed the importance of safe vaccination at the right time to prevent their livestock from contracting RVF and preventing the spread of RVF to humans , [31] . Normally , RVF vaccination of livestock is not free in Sudan , an expenditure that might impede locals from regularly vaccinating their livestock , bearing in mind that a new episode of RVF might take some years to re-emerge . Therefore , subsidizing vaccinations for emerging zoonotic diseases might encourage farmers to regularly vaccinate their livestock . In general , RVF livestock vaccines are either inactivated or live attenuated [32 , 33] . However , the inactivated vaccine needs multiple doses to booster immunity , making it more expensive and more difficult to distribute . Because the vaccine requires multiple shots , establishing immunity requires time and this vulnerability decreases the vaccine’s usefulness during outbreaks . The live attenuated vaccine is administered as a single dose , but it has shown some teratogenic effects that can lead to abortion among inoculated pregnant animals[34 , 35] . However , safe vaccine remains the effective way to protect animals and humans [33 , 36] . The respondents were aware of the possibility of RVF spreading inside the country , especially through the free mobility livestock grazing system . They also knew about the economic consequences of a ban on livestock trade after an RVF outbreak . This awareness is important to consider when early warning systems are developed to avoid bias in disease surveillance . The community’s main sources of information on RVF were social network and mass media such as radio , not veterinarians or health workers , who were mainly involved in case notification rather than increasing public awareness [37] . The strong dependence on social networks , rather than on medical and veterinary professionals , could increase the risk of misconceptions if the wrong information is spread . Thus , the World Health Organization recommends that during zoonotic outbreaks interdisciplinary teams of health providers , veterinarians and environmentalists , provide main communication with the public [38] . These teams can communicate through social networks and mass media such as radio , which is one of the most common sources of information in remote areas of many countries . This local involvement will empower the community , allowing them to contribute to notification and control of the outbreaks , and lead them to play proactive roles . This cooperation could strengthen the national surveillance system , which depends mostly on passive notification , a system that might overlook health related events in remote areas . Empowering livestock owners is an opportunity to strengthen the surveillance system for zoonoses , including RVF [39] . Although our study was conducted in 2013 in an area that was affected during the 2007 RVF outbreak in Sudan , up-to-date questions about RVF were also asked at the time of the study . For the questions related to the 2007 RVF outbreak , we considered recall bias . Since the 2007 RVF outbreak , no other recorded hemorrhagic fever outbreaks had occurred in this area according to the participants , so the participants would not have mixed the information about RVF with other similar diseases . In addition , the 2007 outbreak was severe , affecting humans , livestock and the economy in a unique way , which made it easier to remember , decreasing the possibility of recall bias . In general , the results of this survey are generalizable for the agro-pastoralist regions of Sudan due to the similarity of the context as well as for other countries that experience endemic RVF with similar knowledge , attitudes and practices . This study addressed the challenges and opportunities of including local communities in controlling RVF outbreaks at the interface between humans , animals , and their environment . The suddenness of the outbreaks , the lack of treatment , the lack of vaccines , and the complex transmission cycle of RVFV highlights the need to increase community participation in disrupting RVF outbreaks . Crucial challenges include improving the knowledge and correcting misconceptions about RVF that result in risky behaviors . However , by empowering rural communities through education and motivating them to recognize cases early , the authorities could be notified and could act accordingly to support the local community . The willingness of the community to participate in curbing RVF outbreaks is an opportunity that can be effectively managed in a bottom-up approach: the more we know about a community’s knowledge , attitudes and practices related to the emergence of RVF , the better we will be embowering local communities with the best information and strategies to prevent the spread of RVF . That is , this bottom-up approach may result in mutually acceptable and cost-effective interventions that can be used to disrupt transmission of RVF in affected communities .
Rift Valley fever ( RVF ) , is a neglected , emerging , mosquito-borne disease that has caused outbreaks in Africa and the Arabian Peninsula . RVF outbreaks have a severe negative impact on livestock , human health and economy , placing further demands on communities already experiencing high levels of poverty . We believe there is an immediate need to develop new approaches that will tackle the ongoing spread of RVF . One such approach would prioritize outbreak prevention by involving local communities in the surveillance of emerging zoonotic diseases , empowering local communities as agents of change rather than relegating them as passive victims . RVF is a disease with global implication , but it starts at a local level . Therefore , to control zoonotic disease such as RVF , it is important to understand the local communities’ knowledge , attitudes and practices related to RVF . Using a bottom-up perspective , we investigated the factors that keep the local community from participating in the control of RVF outbreaks at the interface between humans , animals , and the environment . By devising acceptable and cost-effective interventions , we believe local communities can be encouraged to be the first line of defense against RVF outbreaks . Furthermore , policies aimed at curtailing RVF outbreaks would benefit from involvement of the local communities .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion", "Conclusions" ]
[ "livestock", "medicine", "and", "health", "sciences", "tropical", "diseases", "geographical", "locations", "sudan", "rift", "valley", "fever", "neglected", "tropical", "diseases", "infectious", "disease", "control", "veterinarians", "africa", "veterinary", "science", "veterinary", "medicine", "infectious", "diseases", "veterinary", "diseases", "zoonoses", "livestock", "care", "agriculture", "people", "and", "places", "professions", "biology", "and", "life", "sciences", "population", "groupings", "viral", "diseases" ]
2017
The One Health approach to identify knowledge, attitudes and practices that affect community involvement in the control of Rift Valley fever outbreaks
A role for variant histone H2A . Z in gene expression is now well established but little is known about the mechanisms by which it operates . Using a combination of ChIP–chip , knockdown and expression profiling experiments , we show that upon gene induction , human H2A . Z associates with gene promoters and helps in recruiting the transcriptional machinery . Surprisingly , we also found that H2A . Z is randomly incorporated in the genome at low levels and that active transcription antagonizes this incorporation in transcribed regions . After cessation of transcription , random H2A . Z quickly reappears on genes , demonstrating that this incorporation utilizes an active mechanism . Within facultative heterochromatin , we observe a hyper accumulation of the variant histone , which might be due to the lack of transcription in these regions . These results show how chromatin structure and transcription can antagonize each other , therefore shaping chromatin and controlling gene expression . Chromatin dynamics is now well recognized as key in the regulation of nuclear processes such as gene expression , DNA replication and DNA repair , therefore impinging on biological phenomena such as cellular differentiation as well as normal and cancer development . Cells have developed strategies to locally reconfigure chromatin structure , including i ) the use of ATP-dependent chromatin remodeling complexes to either modify the topology , slide or remove nucleosomes; ii ) the covalent modification of histone and; iii ) the incorporation of variant histones within nucleosomes . Incorporation of variant histones within chromatin regions is emerging as a way for cells to create specialized chromatin domains in order to mediate various cellular functions [1] . A large number of variant histones exist , especially for histone H2A and H3 , but the best characterized variant histone is perhaps H2A . Z [2] , [3] . H2A . Z is highly similar to canonical H2A . The two molecules diverge only by a few amino acids and the structure of a nucleosome containing H2A . Z is almost identical to that of a canonical nucleosome [4] . Despite this , H2A . Z clearly has a specialized function that cannot be complemented by an additional allele of H2A . H2A . Z is required for viability in most organisms [5]–[8] . It is necessary for early development in Xenopus , Drosophila and mouse [6]–[8] , is associated with cancer progression [9]–[12] , and is required for estrogen receptor signaling [13] and for embryonic stem cell lineage commitment [14] . H2A . Z occupies one or two nucleosomes in most promoters , a phenomenon first described in yeast [15]–[19] but recently shown to be conserved in Drosophila , chicken , plants and mammals [20]–[25] . In mammalian cells , H2A . Z also occupies pericentromeric heterochromatin domains [26] and regulatory elements such as enhancers and insulators [20] , [22] . The involvement of H2A . Z in gene regulation has been clearly established [24] , [27]–[30] but the mechanism underlying its function is not fully understood . Work in yeast suggest that H2A . Z regulates transcription either by creating unstable nucleosomes [19] , by positioning nucleosomes [15] , by making contacts with the transcription machinery [27] and/or by maintaining active genes to the nuclear periphery [31] . In mammalian cells , H2A . Z was shown to be important for transactivation by transcription factors such as p53 [24] and nuclear receptors [13] , [32] , leading to local chromatin reorganization and transcriptional regulation . A possible role for H2A . Z in transcriptional elongation has also been reported [30] . Despite these few documented cases , it is not known whether H2A . Z plays a more general role in transcription . Here we provide evidence for a global role of H2A . Z in gene expression in human cells . We also demonstrate that H2A . Z regulates early steps in transcription by helping in the recruitment or RNA polymerase II ( RNAPII ) to promoters . Surprisingly , we also show that a mechanism allows for the non-targeted incorporation of H2A . Z molecules in the genome , and that transcription results in the depletion of these randomly incorporated H2A . Z molecules from genes . The battle between these two phenomena shapes the chromatin landscape and may have implications in heterochromatin function . In order to gain some understanding about the function of H2A . Z in human cells , we performed ChIP-chip experiments in U2OS cells . The H2A . Z-enriched material was hybridized on tiling arrays covering the non-repetitive portion of chromosome 19 and 22 at an average of 250 base pairs resolution . Acetylated-H2A . Z , RNAPII , P-Ser2 RNAPII as well as various histone modifications were also profiled in the same conditions ( summarized in Table S1 and available in Dataset S1 , S2 , S3 , S4 , S5 ) . H2A . Z occupancy is high in the 5′ regions of many genes ( Figure 1A and Figure S1 ) , reminiscent of what is observed in yeast [15]–[19] and in agreement with work in human , plants and Drosophila by others [20] , [21] , [23] . H2A . Z is also enriched in distal regulatory elements ( as defined as regions enriched in H3K4me1 [33] ) ( Figure S2 ) as described before [20] , and in heterochromatin ( see below ) as previously suggested [34] . Interestingly , when genes were binned based on their RNAPII occupancy ( Figure 1B and Figure S3A ) , a clear correlation was observed between the level of RNAPII and that of H2A . Z at promoters ( Figure 1C and Figure S3B ) . This is different from what we and others have observed in yeast , where H2A . Z is present at most promoters regardless of their transcriptional activity [15] , [16] , [19] , but agrees with previous reports from metazoan systems [20] , [21] . Taken together , our data shows that H2A . Z associates with many kinds of regulatory elements in human cells , and that its association with promoters is linked to RNAPII . It has recently been observed that a large fraction of metazoan genes are associated with RNAPII molecules paused at their 5′ end [35] , [36] . Consequently , the level of RNAPII at promoters does not strictly correlate with transcription in mammalian cells . We used two independent strategies to separate active genes from genes poised by the presence of paused RNAPII . First , we binned the genes with high levels of RNAPII at promoters into four groups based on their level of histone H3K36me3 ( Figure S4 ) . Levels of H3K36me3 have been previously shown to represent a good proxy for transcription rate [37] . Genes with high levels of both RNAPII and H3K36me3 are considered as actively transcribed [37] since H3K36 is methylated co-transcriptionally . On the other hand , genes with high levels of RNAPII but low levels of H3K36me3 are generally considered as non-transcribed [37] , and most likely represent genes with paused RNAPII , a process that may poise genes for rapid activation [35] , [36] . Second , we profiles RNAPII phosphorylated on Ser2 . Because phosphorylation of Ser2 is occurring during elongation [38] , the level of P-Ser2 RNAPII on genes should be a better indication of transcription rate than that of total RNAPII at promoters . As expected P-Ser2 RNAPII correlates better with levels of H3K36me3 ( Pearson , r = 0 , 53 ) than with levels of RNAPII ( Pearson , r = 0 . 34 ) . Interestingly , H2A . Z is present at similar levels on transcribed genes and on genes poised for activation by the presence of paused RNAPII ( Figure 1E and Figure S5 , using P-Ser2 and Figure S4 , using H3K36me3 ) . This data demonstrates that H2A . Z incorporation at promoters does not require fully processive transcription . Our data rather suggests that H2A . Z is recruited early during the transcription cycle; that is , just prior to , or concomitantly with , RNAPII recruitment . Expression profiling experiments identified IL8 and CCL2 as two genes transiently induced upon daunorubicin ( dauno ) treatment , a DNA damaging agent known to induce p21 expression via the p53 pathway [39] ( Figure S6 ) . To investigate the dynamics of the interaction of H2A . Z with promoters , we looked at the association of the variant histone with the promoters of IL8 and CCL2 upon treatment of U2OS cells with dauno . A detailed time course looking at H2A . Z ( normalized for H2B ) and RNAPII recruitment at IL8 and CCL2 showed that H2A . Z is recruited to the promoter just prior to RNAPII and leaves the promoter as the polymerase gets recruited ( Figure 1F and Figure S7 for additional replicates ) . Interestingly , at CCL2 , RNAPII is present in a paused state at the promoter prior to dauno treatment . Upon activation , RNAPII is first released from the promoter; an event that is coupled to an increase in H2A . Z occupancy and followed by re-loading of RNAPII . Note that these variations in H2A . Z levels at promoters cannot be explained simply by fluctuations in nucleosome density since our data are normalized for nucleosome occupancy . Taken together our data suggests that H2A . Z associates with promoters prior to RNAPII and may therefore affect early steps in transcription such as RNAPII recruitment . In order to test for a role of H2A . Z in RNAPII recruitment , we made use of small hairpin-mediated RNA ( shRNA ) interference . Knockdown of H2A . Z crippled both the induction of IL8 ( Figure 2A and Figure S8A ) and the recruitment of RNAPII ( Figure2B and Figure S8B ) upon induction by dauno . This shows that H2A . Z plays a role in the regulation of IL8 by dauno by helping with the recruitment of RNAPII . This mechanism is consistent with what is observed in yeast where we showed that H2A . Z ( Htz1 in yeast ) assists in the recruitment of RNAPII and its associated co-factors [27] , [29] , [40] . Interestingly , knockdown of H2A . Z also resulted in reduction of RNAPII occupancy at the genomic level in steady state cells ( Figure 2C and Figure S9 ) , arguing for a broad role for H2A . Z in RNAPII recruitment in metazoans . This data is supported by the fact that the amount of RNAPII associated with bulk chromatin is reduced in H2A . Z knockdown cells , despite the fact that no change in total cellular RNAPII levels are observed ( Figure 2D ) . To our knowledge , this is the first report of such a broad effect of H2A . Z in transcriptional regulation , although a transcriptional role has been demonstrated for a few genes [24] , [32] . Close inspection of Figure 1E and Figure S4D revealed that H2A . Z is depleted from the body of actively transcribed genes compared to inactive genes . Figure 3A shows a zoomed out view of Figure 1E allowing to fully appreciating the observation that transcribed genes have lower H2A . Z levels than their flanking intergenic regions ( the regions upstream of the promoter and downstream of the terminator ) . This data shows that H2A . Z accumulates randomly in the genome at low levels and suggests that transcription may “clean” genes from these “miss-incorporated” histone variant molecules . Because our H2A . Z profiles are normalized for H2B occupancy , it seems unlikely that this result is solely due to nucleosome loss during transcription but we nevertheless investigated that possibility in more details . First , we looked at nucleosome occupancy by profiling histone H4 normalized for total genomic DNA . Histone H4 occupancy is not affected by transcription in these conditions ( Figure 3B ) , ruling out the possibility that transcription results in a depletion of H3/H4 tetramer . H2B occupancy , when normalized for histone H4 levels ( Figure 3C ) or for total genomic DNA ( not shown ) , is not dramatically affected by transcription either , suggesting that H2A/H2B dimers are not depleted by transcription . Since it is well known that transcription disrupts and reassembles H2A/H2B dimers [41] , we conclude that H2A . Z depletion likely results from a preferable re-association of H2A-containing dimers over H2A . Z-containing dimers in the wake of RNAPII ( see Discussion ) . Interestingly , revisiting our previously published ChIP-chip data on yeast H2A . Z suggested that this phenomenon is conserved throughout eukaryotes ( Figure S10 ) . We therefore took advantage of the tractable yeast system to investigate H2A . Z eviction in a dynamic fashion . Yeast cells were submitted to a heat shock and H2A . Z occupancy was measured by ChIP-chip over time . H2A was also profiled as a control . Heat shock was chosen because it strongly induces a significant number of genes in a transient manner , allowing us to look at both activation ( between 0 and 15 minutes ) and repression ( between 15 minutes and 2 hours ) . As shown in Figure 3D , H2A . Z levels ( normalized to H2B ) are reduced in the transcribed regions of heat shock-induced genes ( compare 0 minutes with 15 minutes ) . This data clearly shows that transcription directly or indirectly depletes H2A . Z molecules during elongation . In addition , the figure shows that the H2A/H2B ratio slightly increases as H2A . Z decreases . This clearly shows that the depletion is specific for H2A . Z . Quite strikingly , Figure 3D also shows that H2A . Z re-appears on the body of genes during or after repression ( compare 15 minutes with 2 hours ) suggesting that the variant histone is actively loaded within chromatin . We therefore propose that an active mechanism allows for non-targeted/random H2A . Z incorporation in the genome and that transcription antagonizes this recruitment . A prediction of the “transcription-dependent eviction model” proposed here is that genes that have not been transcribed for a long time should accumulate “abnormally” high levels of H2A . Z on their transcribed region . The best cases for such long-term-repressed genes lie within heterochromatin . Quite strikingly , and in agreement with the model , high levels of H2A . Z are associated with heterochromatic regions ( Figure 4 ) . Figure 4A shows an example of such a heterochromatic region ( defined here as a region with high level of histone H3K9me2 ) , while Figure 4B–4E shows the average signal of H3K9me2 , H3K36me3 , H2A . Z and acetylated-H2A . Z over a composite of the 30 heterochromatic regions we have identified on chromosome 19 . Heterochromatin-associated H2A . Z is hypo-acetylated relative to euchromatin-associated H2A . Z ( Figure 4A and 4E ) , and is not limited to promoter regions but rather covers transcribed regions ( Figure 4D and Figure S11 ) . This further suggests that H2A . Z accumulates in heterochromatin as a consequence of a lack of transcription . Also noteworthy is the fact that heterochromatic genes that manage to escape silencing ( that is that they have low H3K9me2 and high H3K36me3 ) do not show over-accumulation of H2A . Z on their gene body ( see red arrow in Figure 4A and Figure S11 ) . Taken together , this data suggests that H2A . Z accumulates on the body of non-transcribed genes , and that this may play a role in shaping heterochromatin . Our genomic data shows that human H2A . Z associates with promoters occupied by RNAPII . Time course experiments , however , showed that H2A . Z and RNAPII do not co-occupy the promoter since H2A . Z is recruited prior to RNAPII and is evicted as the polymerase is loaded . Human H2A . Z is therefore recruited to promoters as part of the activation process; which is different from what is observed in yeast , where the association of Htz1 to promoters does not require any activation cues . Htz1 is rather pre-bound to most promoters and it has been proposed that its loading actually occurs during transcriptional repression in order to prime the promoter for the next activation cycle [42] . The reason why mammalian cells have evolved ways to recruit H2A . Z “on demand” may reflect the fact that a small fraction of mammalian genes are actively transcribed at a given time . Yeasts , on the other hand , express a large fraction of their genome at any time and may therefore benefit from having most of their genes primed with Htz1 at all time . Human enhancers , however , are more like yeast core promoters in that they are occupied by H2A . Z regardless of the transcriptional statue of their associated gene ( see Figure S2 ) . While we see transient recruitment of H2A . Z upon activation of inducible genes such as IL8 and CCL2 , we also observe large amount of H2A . Z on promoters carrying paused RNAPII but no evidence for transcription . At these promoters , the ultimate activation cue ( the one that would stimulate the release RNAPII from the pause site and its entry into elongation ) is not present . Hence , the presence of H2A . Z at promoters does not require full transcriptional activity . We rather propose that H2A . Z is recruited upon transacting signals that triggers RNAPII recruitment . At some promoters , such a signal is sufficient to activate transcription . This is the case at IL8 , where we observe H2A . Z-dependent recruitment of RNAPII and active transcription upon dauno treatment . At CCL2 , however , a significant amount of RNAPII is pre-bound to the promoter in the absence of dauno . Upon dauno treatment , we observe release of RNAPII into elongation and subsequent association of H2A . Z followed by re-association of RNAPII to the promoter . We therefore conclude that transient association of H2A . Z with promoters always precedes ( and favors ) RNAPII recruitment; and that this may lead to transactivation or promoter poising depending on the promoter context . Interestingly , H2A . Z reappears on the promoter after the first wave of RNAPII has left the promoter , suggesting that H2A . Z is not only recruited during the initial round of transcription but is reloaded prior to each new round . This suggests that the presence of H2A . Z on the promoter may play an active role beyond that of creating unstable promoter chromatin . Our data show that the transient association of H2A . Z with human promoters helps with the recruitment of RNAPII . We observed the same in yeast [27] , suggesting that the function of the variant histone at promoters is conserved , although the strategy used to recruit it may be different between yeast and mammals . Many mechanisms have been proposed to explain how Htz1 may regulate transcription in yeast ( see Introduction ) . The knowledge that mammalian cells also use H2A . Z to help recruiting RNAPII to promoters will prompt us and others to test these mechanisms in higher eukaryotes as well . Sequence analysis of H2A . Z-bound promoters showed that these are enriched in CpG-island ( data not shown ) . Since H2A . Z has been shown to antagonize DNA methylation in Arabidopsis [23] , it is possible that CpG-rich promoters use the presence of H2A . Z to counteract DNA methylation–dependent silencing . Perhaps the most striking finding in this study is the fact that transcription results in the removal H2A . Z from transcribed regions . Transcriptional elongation was known to perturb chromatin structure but , to our knowledge , this study represents the first evidence that the elongation complex can discriminate between canonical histones and their variants . Most likely , elongation disrupts both H2A- and H2A . Z-containing dimers but a transcription-coupled mechanism would prevent H2A . Z-dimers to re-associate in the wake of RNAPII . What allows elongation complexes to discriminate between histone molecules is unknown but may be linked to chromatin assembly/disassembly complexes as well as the various histone chaperones that interact with the elongating polymerase . Activities such as FACT , Asf1 , Rtt106 , Spt6 , Spt2 , Chd1 and others have been shown or proposed to be important for the co-transcription chromatin assembly [41] , [43]–[46] . A preference of those machines for canonical H2A would make the transcription-dependent H2A . Z depletion possible . Also interesting , is the possibility that H2A . Z depletion may be linked to the well established incorporation H3 . 3 in transcribed regions [47] . Indeed , H3 . 3/H2A . Z nucleosomes are known to be relatively unstable [48] , [49] . It is therefore tempting to speculate that H2A . Z depletion may be coupled to H3 . 3 deposition . Alternatively , it is possible that the depletion of H2A . Z is required for the incorporation of H3 . 3 . The question of why cells have evolved mechanisms to remove H2A . Z during elongation will require additional work , but we envision that it may play a role in repressing cryptic initiation . Indeed , the fact that the presence of H2A . Z within promoters helps recruiting RNAPII suggests that removing H2A . Z molecules from transcribed region may be important to suppress cryptic initiation inside genes . Alternatively , removing H2A . Z from transcribed region may somehow help with subsequent transcription rounds , a model compatible with previous work suggesting that this variant histone affects elongation [30] . Also striking is the fact that H2A . Z molecules quickly re-appear on genes after transcription has ceased . Indeed , the level of H2A . Z molecule in recently transcribed genes reaches background level within 2 hours after cessation of transcription . This implies that random H2A . Z incorporation occurs via an active mechanism . Whether this non-targeted incorporation proceeds via SWR1 ( in yeast ) , p400 and SRCAP ( mammals ) like for promoter-targeted incorporation will require additional experiments . The H2A . Z molecules associated with promoters are acetylated while the ones that accumulate in heterochromatin are under-acetylated . Elegant immunostaining work by the Chueng laboratory showed that a ubiquitinylated form of H2A . Z associates with the inactive X chromosome [34] . It is not clear at this point whether the non-acetylated H2A . Z found in facultative heterochromatin is also ubiquitinylated but this possibility is attractive . It will be interesting to test if a cross-talk exists between H2A . Z acetylation ( or deacetylation ) and ubiquitination . It is tempting to speculate that the accumulation of under-acetylated H2A . Z in facultative heterochromatin plays a role in gene silencing . So far , H2A . Z was shown to be important for the formation of pericentromeric heterochromatin [50] but its role in the function of facultative heterochromatin remains unexplored . Using H2A . Z-specific shRNA , we were not able to show any defect in the expression of heterochromatic genes in U2OS cells ( data not shown ) . This data suggests that the presence of H2A . Z in heterochromatic regions is not required for maintaining silencing . This could mean that H2A . Z accumulation in heterochromatin is solely a bi-product of these genes not being transcribed . Alternatively , H2A . Z accumulation may be important for the de novo formation of heterochromatin rather than its maintenance . Finally , it remains possible that H2A . Z plays a role in other aspects of heterochromatin function . All these possibilities will require further experimentations but the data shown here clearly establishes the accumulation of hypo-acetylated H2A . Z in large domains as a landmark of facultative heterochromatin . In yeast , Htz1 was shown to restrict the spreading of Sir2-dependent silencing [51] . Despite the fact that Sir2-dependent silencing does not have its equivalent in mammals , the presence of H2A . Z in heterochromatin suggested that it might play as role in restraining heterochromatin spreading . Investigation of H3K9me2 profiles in H2A . Z knockdown experiments did not allow detecting any sign of heterochromatin spreading ( not shown ) . This result is not very surprising since mammalian heterochromatin is very different than Sir2-dependent “heterochromatin” and probably uses different spreading mechanisms . The work presented here provides clear evidence that H2A . Z is recruited to chromatin via at least two distinct mechanisms . First , a very dynamic recruitment of H2A . Z at promoters plays a role in the recruitment of RNAPII . Second , a non-targeted mechanism allows for the random incorporation of some H2A . Z molecules in the genome . Finally , we show that a process linked to transcriptional elongation allows for the depletion of the variant histone from active genes . This battle between active H2A . Z incorporation and transcription-dependent H2A . Z depletion shapes the euchromatin and heterochromatin landscape of mammalian cells . Because heterochromatin formation plays a key role in normal development and in tumorigenesis , the results presented here shall open the way to experiments aiming at better understanding the role of H2A . Z in these important aspects of biology . U2OS cells were grown in McCoy's medium in the presence of 10% FBS , 0 . 2 mM L-glutamine , 100 IU/mL penicillin and 100 µg/mL streptomycin at 37°C in the presence of 5% CO2 . For time course experiments , cells were serum-deprived ( by changing the medium to 0 . 5% FBS ) for 24 hours prior to addition of 250 nM dauno for various periods of time as indicated in figure legends . ShRNA-mediated knockdowns were performed as described previously [24] . Yeast cells ( yFR212: MATa , ade2-1 , trp1-1 , can1-100 , leu2-3 , 112 , his3-11 , 15 , ura3-1 , HTZ1::3myc; yFR134: MATa , ade2-1 , trp1-1 , can1-100 , leu2-3 , 112 , his3-11 , 15 , ura3 , HTA1::9myc::TRP1 ) carrying respectively three and nine myc epitopes in the C-terminal end of endogenous HTZ1 or HTA1 were used in ChIP-chip experiments . Heat shock was performed as follow . Cells were grown in YPD at 25°C until mid-log phase and an aliquot ( to be used as t0 ) was taken prior to adding an equal volume of 49°C YPD medium ( to instantly raise the temperature to 37°C ) . The cells were put at 37°C and an aliquot was taken at 15 minutes . Cells were then put back at 25°C and a last aliquot was taken at 120 minutes . All aliquots were used in ChIP as described below . For mammalian cells , chromatin immunoprecipitations were performed as described previously [24] starting from 108 U2OS cells . For yeast cells , chromatin immunoprecipitations were performed as described previously [15] . The antibodies used were as follow: H2A . Z ( Ab4174 ) , acetylated-H2A . Z ( Ab18262 ) , H3K9me2 ( Ab1220 ) , H3K36me3 ( Ab9050 ) , H3K4me3 ( Ab8580 ) , H3K4me1 ( Ab8895 ) , RNAPII ( 8WG16 ) , H2B ( Ab1790 ) , recombinant H4 ( a gift from Alain Verreault ) , Myc-tagged Htz1and Hta1 in yeast ChIPs ( 9E10 ) . DNA from mammalian or yeast ChIP and controls samples were labeled using a double or single amplification protocol respectively as described previously [15] , [52] . The data were normalized using limma's loess function [53] in BioConductor ( from the ArrayPipe Analysis Pipeline ( ref PMID: 15215429 ) and replicates were combined using a weighted average method as described previously [54] . The combined datasets are available in Dataset S1 , S2 , S3 , S4 , S5 , S6 . To interpolate between probes , a standard Gaussian filter ( SD = 200 bp ) was applied twice to the data as described previously [15] to generate a value each 10 bp . This “smoothed data” was used to calculate the average signal in the promoter ( defined as TSS±500 bp ) and on the complete gene length of each gene ( using the UCSC_knownGenes_hg18 from Mar . 2008 ) . Genes ( using unique TSS ) were then binned into groups as described in the figure legend . The non-smoothed data was mapped on the 5′ and 3′ boundaries into 200 bp windows for each half-gene and adjacent half-intergenic regions . A sliding window of 1 . 6 kb was then applied to the ratios . The same mapping procedure was used to map the yeast Htz1 dataset ( Figure S10 ) using 50 bp windows and a sliding window of 300 bp . ( Figure 4B–4E ) To identify heterochromatic regions , the combined H3K9me2 dataset was smoothed as described above , except that the standard deviation was 10 kb instead of 200 bp . Regions of interest were defined as consecutive positions above 0 covering at least 50 kb , with an average intensity of at least 0 . 1 . Regions separated by less than 50 kb were then merged , identifying 30 “heterochromatic regions” on chromosome 19 . The non-smoothed data was mapped on the 5′ and 3′ boundaries into 1 kb windows for each half-region and a sliding window of 20 kb was then applied to the ratios . ( Figure S2 ) To identify predicted enhancers , we applied a strategy developed by the Ren laboratory [33] . Regions specifically enriched in H3K4me1 ( p<0 . 0 . 1 ) were identified using an algorithm described previously [52] . The regions were then filtered to include only those located with intergenic regions and that are at least 5 kb from a known TSS ( to avoid contamination from proximal promoters ) . This led to the identification of 577 predicted enhancers . The middle of each region was used to map the different non-smoothed data into 50 bp windows and a sliding window of 1 . 5 kb was then applied to the ratios . Enhancers were then classified in three groups based on the transcriptional status of the closest TSS within the same chromosomal domain as defined by a regions flanked by two CTCF binding sites [55] . ( Figure 3D ) The data was smoothed as described above ( SD = 200 bp , resolution = 10 bp ) and the average signal was calculated in the middle of each protein-coding gene ( using SGD version Feb . 02 2008; http://www . yeastgenome . org/ ) in a window excluding the first and last 250 pb to avoid contamination by intergenic signal ( genes<500 bp were then discarded ) .
DNA in living cells is packaged into chromatin by histones and non-histone proteins . This packaging is very dynamic , allowing the controlled access of regulatory proteins such as transcription factors to DNA . Most chromatin is packaged with so-called canonical histones; namely H2A , H2B , H3 , and H4 . In some regions , however , variant histones replace canonical histones , creating special chromatin regions . Here we show that the variant histone H2A . Z is dynamically recruited to promoter regions where it helps in the recruitment of RNA polymerase II , the enzyme responsible for the first step of gene expression . In addition , we show that H2A . Z also associates randomly in the genome , but these molecules are removed during the passage of RNA polymerase II . In non-transcribed regions , H2A . Z accumulates in large domains called heterochromatin . We propose that a battle between random H2A . Z deposition and RNAPII-dependent H2A . Z eviction shapes the chromatin landscape .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "molecular", "biology/histone", "modification", "cell", "biology/gene", "expression", "molecular", "biology/transcription", "initiation", "and", "activation", "genetics", "and", "genomics/gene", "expression", "genetics", "and", "genomics/functional", "genomics", "molecular", "biology/transcription", "elongation", "genetics", "and", "genomics/epigenetics", "molecular", "biology/chromatin", "structure" ]
2009
The Euchromatic and Heterochromatic Landscapes Are Shaped by Antagonizing Effects of Transcription on H2A.Z Deposition
Systematic analysis of synthetic lethality ( SL ) constitutes a critical tool for systems biology to decipher molecular pathways . The most accepted mechanistic explanation of SL is that the two genes function in parallel , mutually compensatory pathways , known as between-pathway SL . However , recent genome-wide analyses in yeast identified a significant number of within-pathway negative genetic interactions . The molecular mechanisms leading to within-pathway SL are not fully understood . Here , we propose a novel mechanism leading to within-pathway SL involving two genes functioning in a single non-essential pathway . This type of SL termed within-reversible-pathway SL involves reversible pathway steps , catalyzed by different enzymes in the forward and backward directions , and kinetic trapping of a potentially toxic intermediate . Experimental data with recombinational DNA repair genes validate the concept . Mathematical modeling recapitulates the possibility of kinetic trapping and revealed the potential contributions of synthetic , dosage-lethal interactions in such a genetic system as well as the possibility of within-pathway positive masking interactions . Analysis of yeast gene interaction and pathway data suggests broad applicability of this novel concept . These observations extend the canonical interpretation of synthetic-lethal or synthetic-sick interactions with direct implications to reconstruct molecular pathways and improve therapeutic approaches to diseases such as cancer . Synthetic interactions between two mutations in different genes were first revealed in Drosophila by Dobzhansky in the 1940s [1] . Synthetic lethality ( SL ) describes that two viable single gene mutations lead to lethality ( synthetic-lethal ) or severely impair growth ( synthetic-sick ) when combined as a double mutant . This concept was implemented as a powerful research tool for molecular pathway analysis in yeast [2]–[5] . Further refinement introduced more quantitative measures of genetic epistasis [6] and lethality induced by gene overexpression in a mutant background ( synthetic dosage-lethality [7] ) . A genetic interaction is negative or aggravating , when the combined effect of two gene defects is more severe than it is expected from a simple multiplicative model . In a positive or alleviating interaction the effect is less severe than expected . These approaches and measures are now increasingly used in mammalian cells exploiting RNA-mediated gene knockdown technologies [8] , [9] . Following a proposal by Hartwell and colleagues [10] , SL has been utilized as a therapeutic approach in cancer treatment employing a combination of genetic ablation ( loss of tumor suppressor ) and chemical inhibition . The first paradigm was set in BRCA1/2-deficient tumor cells , which are synthetic-lethal with inhibition of PolyADP-Ribose Polymerase ( PARP ) [11]–[13] . Small molecule PARP inhibitors are currently being evaluated in clinical trials in BRCA1- and BRCA2-deficient cancers ( e . g . [14] ) . The canonical interpretation of SL stipulates two mutually compensatory , parallel pathways capable of performing the same essential function [2]–[4] . Thus , disrupting a single pathway is viable , while disrupting both pathways is lethal . This concept of between-pathway synthetic lethality ( bpSL ) ( Figure 1A ) led to the creation of computational approaches aiming at reconstructing interaction networks from pair-wise gene deletions or siRNA-induced gene knock-down screens in yeast and mammals [15]–[17] . However , recent genome-wide genetic interaction data revealed multiple negative interactions between mutations affecting the same molecular pathway or complex [3] , [5] , [16] , [18]–[21] . For example , it was estimated that ∼9% [19] and in another study 14% [17] of all negative genetic interaction clusters belong to the same biological pathway . Several mechanistic models were suggested to explain within-pathway SL ( wpSL ) [6] , [16] , [20] , [22] . The deletion of a gene might lead only to a partial degradation of an essential pathway which might be tolerable , whereas the double mutation leads to complete pathway degradation and lethality ( Figure 1B1 ) . This is especially relevant for the interpretation of siRNA-based screens where the efficiency of a particular gene knock down is uncertain . A second possible mechanism suggests that steps in an essential pathway are internally redundant ( Figure 1B2 ) . Lastly , two mutations may cumulatively degrade an essential protein complex , whereas they are individually viable ( Figure 1B3 ) . This mechanism is consistent with the observations that molecular complexes are frequently characterized by the dominance of negative over positive genetic interactions between their components [18] . wpSL interactions between defects in components of a single protein complex are highly enriched for complexes with an essential component [16] , [22] . It was estimated that the contribution from within-complex interactions to the total number of within-pathway negative interactions does not exceed 7% [19] . Common to these mechanistic explanations of wpSL is that they involve either an essential pathway or an essential protein complex . Here , we highlight a novel scenario of wpSL involving two components of a non-essential pathway . The view of molecular pathways as unidirectional , linear reaction cascades is too simplistic . Pathway steps can be reversible which leads to forward and backward propagation of molecular events along the pathway increasing robustness and fidelity of the overall process [23]–[28] . Affecting both forward and reverse steps of the pathway by abrogating the corresponding enzymes creates scenarios in which the pathway flow can be trapped in an intermediate state that may be toxic to the cell or deprive the cell of a limiting resource ( Figure 1C ) . This can create a genetic scenario we define as within-reversible-pathway synthetic lethality ( wrpSL ) , which is the subject of this study . Here , we study bpSL and wrpSL scenarios using mathematical modeling to better understand the system properties of these genetic relationships . We present a simplified model of the pathway applicable for its formal analytical study and performed in silico simulations for bpSL and wrpSL as well as synthetic dosage effects . Our main experimentally confirmed examples of wrpSL are in the homologous recombination DNA repair pathway . Homologous recombination ( HR ) is an important mechanism to maintain genome integrity [29] ( Section S1 and Figure S1 in for more discussion ) . Analysis of yeast gene interaction and pathway data suggests broad applicability of this novel concept . In order to assess the importance of within-pathway negative interactions we ranked all pathways from the KEGG database [30] according to their normalized proportion of negative interactions [5] within each KEGG pathway ( Figure 2 ) . This analysis confirms the previous conclusion [19] that only a minority of within-pathway negative interactions can be explained by negative interactions within a complex ( Figure 1B3 ) . In our analysis only 12% of all negative interactions were of this type ( compared to 7% in [19] ) . Interestingly , HR ranks at the top with 27 within-pathway negative interactions between 20 KEGG pathway components ( Figure 2 ) , of which only a single one affects components of the same protein complex . Recent studies show that individual reaction steps in HR are reversible [23]–[25] ( Figure S1 in Text S1 ) . The Rad51-ssDNA filament is a key intermediate in HR , as it performs the signature reactions of homology search and DNA strand invasion . The formation of this filament is catalyzed by specific co-factors ( see Section S1 and Figure S1 in Text S1 ) . The helicase Srs2 specifically targets the Rad51-ssDNA filament for disruption to reverse filament formation [31] , [32] , [33] . The reversibility of the Rad51-ssDNA filament sets a new paradigm and draws attention to additional reversible steps and their mechanisms in HR , other DNA repair processes , and unrelated molecular pathways . We derived the simplest linear mathematical model of a main DNA repair pathway with reversible steps and a toxic intermediate , and a compensatory pathway , which can recapitulate bpSL and wrpSL ( Figure 3 ) . Each state transition is catalyzed by an abstract enzyme , which may correspond to several biological entities ( compare Figure S1 in Text S1 with Figure 3 ) . In wrpSL trapping of toxic intermediate I is caused by defects in the first backward reaction ( I→S , R1 , k−1 ) and the second forward reaction ( I→P , F2 , k2 ) . The reversibility of the second step ( reaction P→I , R1 ) is not essential for wrpSL to occur , but might be important for quantitative pathway characteristics . Introduction of a final irreversible step ( Figure S2 in Text S1 ) would result in a kinetic proofreading mechanism [34] ( see Figures S2 , S3 in Text S1 and discussion there ) . Such a mechanism increases the robustness of DNA repair , as it avoids a futile P↔I cycle . However , in this simplest model we eliminated the final irreversible step to allow us analyzing the most essential features of wrpSL ( Section S2 in Text S1 ) . Figure 4 explores conditions for various cellular fates ( Normal Robust , NR: no single knockout leads to lethality ( Figure S4 in Text S1 ) ; Normal Fragile , NF: single knockout can lead to lethality ( Figure S4 in Text S1 ) ; Compensated , C: repair is performed by compensatory pathway; death due to DNA Damage , DD: steady state probability of DNA damage >50%; and Death due to Toxic intermediates , DT: steady state probability of toxic intermediate >50% ) . Figure 5 visualizes parametric conditions ( see Section S2B in Text S1 for discussion ) . Using analytical study and numerical simulations with some characteristic choices of kinetic rate values , we explored the dynamical behaviors of the simplest model ( see Figure 6 ) . Here , we discuss the qualitative results and interpretations , while the more formal derivation of these statements is found in Section S3 in Text S1 . To illustrate the static and dynamic properties of the toy model , we selected two typical positions ( Figure 5 #1 , #2 ) corresponding to NR and NF pathway states , respectively . From these “normal” conditions we simulated a number of single and double mutant/overexpression conditions as shown in Figure 5 ( see also Figure S4 in Text S1 ) . Synthetic lethality/sickness and synthetic dosage lethality are important genetic tools to assign individual gene functions into molecular pathways [2]–[4] , [7]–[9] . The canonical interpretation for two mutants found to be synthetically lethal or sick stipulates that the encoded gene products function in different parallel pathways that can mutually compensate ( bpSL ) [2]–[4] , [7]–[9] , [15]–[17] . However , computational analysis of genetic interaction data combined with protein interaction data revealed multiple negative interactions between mutations affecting functions in the same molecular pathway or complex ( wpSL ) [3] , [5] , [16] , [18]–[21] . Several mechanisms of wpSL have been proposed ( Figure 1B ) , and they all involve either essential pathways or essential protein complexes . In extension of this fundamental concept of wpSL , there are several cases of SL between mutants encoding proteins acting in HR , a pathway that is not essential in yeast [35] , [46]–[48] . We term this novel genetic interaction within-reversible-pathway Synthetic Lethality ( wrpSL; Figure 1C ) and provide a novel mechanistic explanation for wpSL , which can create SL within a non-essential pathway or between hypomorphic mutations in an essential pathway that is different from a model invoking sequential pathway degradation by accumulation of partial defects of successive steps ( Figure 1B1 ) . Here , we explore by mathematical modeling the system properties of wrpSL . The modeling must make assumptions about the system properties ( state transition rates , relative pathway efficiencies , etc . ) and identifies several conditions to be met for wrpSL . 1 ) Reversibility of pathway steps . In fact , only the first pathway step must be reversible , whereas reversibility of the second pathway steps enables additional genetic scenarios . 2 ) Possibility of kinetic trapping of an intermediate state of the pathway when both the backward and forward reactions are compromised . The trapping per se can be detrimental due to blockage of cell signaling , sequestering an essential compound , or toxicity . We have assumed lethal toxicity in our model . 3 ) The possibility of rescue by a parallel compensatory pathway may not be strictly required , but highlights the applicability of this concept to non-essential pathways . The mathematical model is validated by the experimentally observed recombination-dependent SL of the srs2 rad54 double mutant in budding yeast [35] ( Figure 3 , Figure S1 in Text S1 , Figure 6 , row 5 ) . Srs2-defective cells are unable to reverse Rad51-ssDNA filaments . These Rad51-ssDNA filaments represent toxic intermediates that accumulate in the cell due to kinetic trapping and interfere with cell viability . The key functions of the Rad54 protein are to assist in DNA strand invasion and allowing DNA synthesis off the invading 3′-end [36] . Hence , in the srs2 rad54 double mutant Rad51-ssDNA filaments and/or D-loops may accumulate forming a toxic intermediate that leads to cell death ( Figure S1 in Text S1; Figure 3 green pair and Figure 6 , row 5 ) . This interpretation is supported by the observation that lethality in this double mutant is suppressed by a defect in Rad51-ssDNA filament formation ( mutations in RAD51 , RAD55 , RAD57 , or RAD52 ) [49] ( see Figure S1 in Text S1 ) , what has been termed recombination-dependent lethality . Preventing Rad51-ssDNA filament formation allows bypass of recombination by alternative means of DNA repair ( for DSBs: Nonhomologous endjoining or single-strand annealing; for gaps: Translesion synthesis or fork regression [23]; see Figure S1 in Text S1 ) . The recombination-dependent lethality of srs2 rad54 is not unique and is also found in additional double mutants in recombinational repair genes including the double mutants mph1 mus81 , mph1 mms4 , srs2 sgs1 and sgs1 ( or top3 or rmi1 ) and mus81 ( or mms4 ) which likely reflect additional examples of wrpSL possibly involving different toxic intermediates [35] , [46]–[53] . As discussed in detail in Section S1 in Text S1 , the synthetic lethalities involving sgs1 are more complex , because of the multiple roles of Sgs1-Top3-Rmi1 in HR , and could be caused also by other mechanisms of SL . Further modeling revealed additional genetic conditions including overexpression of specific pathway enzymes that are predicted to lead to wrpSL ( Figure 6 ) . The mathematical modeling also reveals the importance of reversible pathway steps , which are validated by genetic and biochemical experiments in yeast [23]–[25] . First , the existence of reversible pathway steps does not affect normal pathway progression ( Figure 6 , rows 1–3 ) . Second , reversible pathway steps allow much more efficient and timely use of compensatory pathways ( Figure 6 , row 6 ) . Third , reversible pathways coupled with compensatory pathways avoid lethality of single mutations ( Figure 6 , row 7 ) . The existence of reversible intermediates in HR , and possibly other molecular pathways , has been proposed to increase the robustness of the overall DNA repair system [23]–[25] , and here we provide quantitative modeling evidence and formal analysis of this assertion . An important question is how general wrpSL might be or whether it is an idiosyncrasy of the recombinational repair pathway . Even if wrpSL were restricted to HR , this concept provides significant potential application in anti-cancer therapy . However , there is considerable evidence that many molecular pathways include reversible steps catalyzed by different enzymes in the forward and backward directions ( see Figure 8 ) . Any of those processes can be theoretically trapped into one of their intermediate states if two regulators of forward and backward steps are inactive . In these cases , the accumulating intermediate might be toxic , block proper signal propagation or prevent resource recycling . Focusing on three examples of reversible protein modifications ( phosphorylation by Cdc5/dephosphorylation by Cdc14 , sumoylation by Slx5–Slx8/desumoylation by Ulp1 , Nup60 , ubiquitylation by Rad6–Rad18/deubiquitylation by Bre5 , Ubp3 or degradation dependent on Doa1 , Rpn6; see Figure S6 in Text S1 for details ) , we found ample evidence in published genetic interaction data that are consistent with the wrpSL mechanism . These examples have not been fully explored , but are consistent with the wrpSL concept and amenable to test specific predictions . In summary , genetic and biochemical data strongly support the significance of the wrpSL mechanism in HR , and existing data are consistent with the notion that wrpSL could be a general , widely applicable type of genetic interaction . This may refine our understanding of relationships between gene products and will help to improve pathway reconstruction . In particular , our mathematical modeling provides a conceptual framework for guiding systematic exploitation of mutations and changes in the expression profiles of HR genes and potentially genes of other pathways to induce SL . The simplest mathematical model of Figure 3 was converted into a set of linear ordinary differential equations using the standard chemical kinetics formalism . The steady state model properties were analyzed analytically and exemplified with numerical simulations . Classification of the pathway states according to the extreme ( large or small ) values of the control parameters and the corresponding asymptotic solutions follow the methodology of the asymptotology of reaction networks [54] . All numerical simulations were performed using SBTOOLBOX package for Matlab ( Section S4 in Text S1 ) .
Organizing gene functions into molecular pathways is a major challenge in biology . The observation that two viable gene mutations become lethal when combined as a double mutant has been developed into a major genetic tool called synthetic lethality . The classic interpretation of synthetic lethality stipulates that the two mutations identify genes that work in parallel , mutually compensatory pathways that together perform an essential function . However , a significant number of negative interactions are caused by defects affecting a single molecular pathway . Here , we recapitulate by mathematical modeling recent experimental data that demonstrate synthetic lethality between mutations in genes acting in a single , non-essential molecular pathway . We propose a novel mechanism involving reversible pathways steps and trapping of an intermediate . The modeling also predicts that overexpression of certain genes functioning in reversible pathways will lead to synthetic lethality with gene defects in the same pathway . Our results significantly broaden the interpretation of synthetic lethal and synthetic dosage effects , which fundamentally impacts the assignment of genes to pathways . The concept of synthetic lethality has been applied to cancer therapy , and our modeling results suggest new approaches to how to target a single pathway to induce synthetic lethality in cancer cells .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "medicine", "biology" ]
2013
Synthetic Lethality between Gene Defects Affecting a Single Non-essential Molecular Pathway with Reversible Steps
In areas where health resources are limited , community participation in the recognition and reporting of disease hazards is critical for the identification of outbreaks . This is particularly true for zoonotic diseases such as monkeypox that principally affect people living in remote areas with few health services . Here we report the findings of an evaluation measuring the effectiveness of a film-based community outreach program designed to improve the understanding of monkeypox symptoms , transmission and prevention , by residents of the Republic of the Congo ( ROC ) who are at risk for disease acquisition . During 90 days , monkeypox outreach was conducted for ∼23 , 860 people in northern ROC . Two hundred seventy-one attendees ( selected via a structured sample ) were interviewed before and after participating in a small-group outreach session . The proportion of interviewees demonstrating monkeypox-specific knowledge before and after was compared . Significant gains were measured in areas of disease recognition , transmission , and mitigation of risk . The ability to recognize at least one disease symptom and a willingness to take a family member with monkeypox to the hospital increased from 49 and 45% to 95 and 87% , respectively ( p<0 . 001 , both ) . Willingness to deter behaviors associated with zoonotic risk , such as eating the carcass of a primate found dead in the forest , remained fundamentally unchanged however , suggesting additional messaging may be needed . These results suggest that our current program of film-based educational activities is effective in improving disease-specific knowledge and may encourage individuals to seek out the advice of health workers when monkeypox is suspected . Human monkeypox ( MPX ) is caused by infection with Monkeypox virus ( MPXV ) , a member of the Orthopoxvirus genus in the family Poxviridae . The clinical manifestations of severe MPX bear pronounced similarity to those of smallpox , though fatality rates are consistently lower for monkeypox [1]–[4] . Smallpox has now been eradicated but MPX , a zoonosis , continues to be an endemic disease threat in large areas of the Congo Basin , principally in forested regions of the Democratic Republic of the Congo . It also occurs sporadically in other parts of West and Central Africa [1] , [5] , [6] . A study published in 2010 reported the cumulative incidence of monkeypox in a disease endemic region of the Democratic Republic of the Congo ( DRC ) as 5 . 5 per 10 , 000—substantially higher than 2 decades prior—with children ( under 15 years of age ) constituting the greatest proportion of the population affected by this disease [6] , [7] . Contact with wildlife , which is routine throughout most of the Congo Basin , is believed to increase the risk for human infection with monkeypox virus [8] . The virus is also communicable between people . During the era of eradication , smallpox vaccine was shown to provide protection against monkeypox infection , but smallpox vaccinations ceased in Africa in 1980 . Conventional , second-generation smallpox vaccines ( such as ACAM2000™ , the vaccine currently licensed for use in the United States ) are not currently recommended for widespread use in the region , in part , due to safety concerns stemming from high rates of HIV prevalence . As smallpox vaccine-derived immunity wanes across communities throughout the Congo Basin , monkeypox disease incidence is anticipated to increase . Blindness due to corneal scarring ( typically unilateral ) , and superficial skin discolorations and scaring have been reported as sequelae of infection for monkeypox [9] , [10] . Severe complications of illness include secondary bacterial infections of the skin , bronchopneumonia , dehydration , and encephalitis . Complications appear to occur more frequently in children and in individuals who have not had prior vaccination against smallpox [3] , [11] . There are no pharmaceutical therapies approved to treat MPX , and use of conventional licensed vaccines ( such as ACAM 2000™ ) is limited due to safety concerns . Monkeypox remains a disease of persons in impoverished rural areas , and disease control hinges on deterring zoonotic exposure to the virus and , barring that , interrupting person-to-person spread [8] . The 2010 Integrated Disease Surveillance and Response technical guidelines for Africa—jointly produced by the World Health Organization ( WHO ) and the US Centers for Disease Control and Prevention ( CDC ) — stipulate that early identification and reporting infectious disease hazards such as monkeypox should be a core function of community-level health systems [12] . In many resource poor areas , however , community health workers are rare , which shifts the burden of early hazard detection and reporting to the community members themselves . There are many obvious challenges to the engagement of community members for this task , not the least of which is identifying effective and lasting means for imparting basic information about an infectious disease hazard , how to avoid the disease , and why it is important to alert the medical community when occurrence of the disease is suspected . In 2003 , there was an outbreak of monkeypox identified in the town of Impfondo in the Likouala region of the Republic of the Congo ( ROC ) [10] . A subsequent serosurvey performed in the region revealed seroprevalence levels for Orthopoxvirus antibodies ranging from 23–83% in various villages and towns throughout Likouala , suggesting high rates of exposure to Orthopoxviruses in a population that , for the most part has not benefited from the smallpox vaccine [13] . These observations highlighted the need for initiation of rash illness surveillance in the area , focusing primarily on the identification of human monkeypox . Along with this surveillance activity , a comprehensive program of community outreach and education was initiated by CDC and partners , with the dual intent of stimulating community participation in the surveillance activity and providing community members with basic knowledge to protect themselves against the disease . Film-based approaches have previously been shown to be effective in imparting health messages to members of rural communities in Africa [14]–[16] . Beginning in 2008 , a series of educational films addressing monkeypox were produced by the International Conservation and Education Fund ( INCEF ) , in conjunction with CDC . Two of these films were produced specifically for the purposes of educational outreach to members of communities at risk for monkeypox in the Likouala region . During the months of June–October in 2009 , INCEF educators performed community outreach for monkeypox in 16 towns and villages in Likouala , and conducted a real-time evaluation of the effectiveness of the education program in increasing community member's basic knowledge of the disease and their intended future behaviors relevant to risk reduction , such as diminishing contact with suspected zoonotic hosts of MPXV [17]–[19] . In addition , the basic retention of knowledge about monkeypox was assessed among persons who had participated in a pilot program of outreach performed 6 months to a year prior . We report the results of these evaluations . Two films addressing monkeypox were produced by INCEF personnel with technical input provided by personnel from the US Centers for Disease Control and Prevention , the Republic of the Congo's Ministry of Health ( Ministere de la Sante et de la Population ) , and the WHO office in Brazzaville . The films ( “Monkeypox Testimonies” and “Understanding Monkeypox” ) feature recognizable members of the local community ( i . e . , health professionals and individuals who had been affected by monkeypox ) . The first film ( 13 minutes in length ) covers topics related to monkeypox recognition , modes of acquisition , and consequences ( e . g . , costs and sequelae ) ; the second ( 12 minutes ) covers topics related to virus transmission , disease prevention and the importance of seeking medical care . Both films feature individuals speaking in , or dubbed in , Lingala ( a local language ) . Monkeypox educational materials can be viewed at http://www . incef . org/ . A team of 2 facilitators , one an experienced community outreach coordinator and one versed in health education , conducted each outreach session . Two teams performed all the outreach described in this work . The lead facilitator trained the other team members in outreach and interview methods . A typical outreach mission entailed a team's travelling , by foot , bicycle or boat , to towns or outlying communities , performing outreach at a single location for 2–4 nights at a time . The use of portable –rechargeable—batteries as power sources for projection equipment allowed a team to remain in the field for several days at a time , or weeks at a time if a portable generator was also utilized . Written permission from local government officials ( prefect , sous-prefect ) was obtained prior to the team's departure into the field , and census data for outreach locations were collected . Upon arrival at an outreach location , educators met with village leaders to obtain verbal permission to conduct the outreach . Outreach activities began with small-group sessions involving 10–100 persons per group ( depending on the size of the village ) , with participants separated into the following 5 groups: children <13 years old ( these individuals did not participate in the evaluation component of this activity ) , individuals 13–18 years old , males ≥18–35 years old , males >35 , and females ≥18 years old . Small-group education sessions were conducted individually ( one at a time ) over the course of two days , typically during the day in a suitable building or structure ( church , school , etc ) . Those who volunteered to participate in the small-group sessions were asked to form a queue and members were admitted until the relevant quota was obtained . After completion of the small-group sessions , additional outreach sessions were held at night for the entire village ( or neighborhood within larger towns ) ( Figure 1 ) . Both small and large-group sessions began with a general discussion of monkeypox . Then the first film in the module was shown , followed by a 15–45 minute discussion about the information presented . The second film was then shown , again with subsequent discussion . Facilitators recorded ( on paper ) anecdotes and comments made by participants during the question and answer period . Prior to viewing the films , approximately 10% of participants in the small-group discussions ( described above ) were selected to complete pre- and post-screening interviews . Volunteers for the interviews identified themselves by raising hands , after which the facilitators selected the appropriate number of interviewees based on convenience , quota- selection process . The interview tool consisted of a series of open-ended or yes/no questions addressing material presented in the films ( Text S1 ) . Each person was questioned individually out of earshot of others . Questions were read out loud in Linguala ( a local language ) by one facilitator , and responses were recorded on a paper questionnaire sheet by the second . Questions that were asked prior to screening the film addressed the individual's basic knowledge of monkeypox , including the principal disease symptoms , and modes of virus transmission . As well , questions were asked that pertained to the individual's current ( or past ) behaviors regarding health care-seeking practices and the handling of primate and rodent carcasses . After viewing the films , individuals were again questioned about their basic knowledge of monkeypox ( as above ) . In addition , interviewees were queried as to their intended future behaviors with respect to when they would seek health care for themselves or a family member , and how they would handle primate and rodent carcasses . Respondents also provided their age , sex , occupation or school attendance status , and current village of residence ( if other than the village or town in which the films were shown ) . The purpose of the proposed activity was to evaluate an educational film designed to deliver health messages about monkeypox to communities in ROC , in order to determine whether the messages were understood by the audience . A written protocol describing the evaluation was reviewed by Human Subjects Research Advisors at the National Centers for Zoonotic and Vector Borne Diseases at CDC , who determined that the work did not involve research as defined under 45 CFR 46 . 102 ( d ) . Prior to engaging volunteers , the evaluation methods , purpose , and voluntary nature of the evaluation was described by the health educators to prospective participants and verbal consent was obtained . Two hundred eighty-two questionnaires were generated , of which 271 ( 96 . 1% ) were determined complete for analysis ( i . e . , interviewees provided an answer for at least one question both prior to and again after viewing ) . Data analyses were based on affirmative responses . Comparisons of responses pre- and post- screening were calculated with a McNemar's Test to account for matched-pair data . Associations between groups were calculated using Pearson's Chi-Square or Fisher's Exact tests . A p-value of <0 . 05 was considered statistically significant . All statistical analysis was performed in PASW Statistics 18 software ( SPSS , Chicago , IL , USA ) . Intervention locations were geo-referenced using maps available from the UNHCR GIS and Mapping Unit in Kinshasa and Google Earth ( UNHCR [2010] DRC Refugees in the Republic of Congo . Kinshasa , DRC . ) Locations were divided into geographic sectors based on proximity to large towns ( Impfondo , Dongou , Enyelle ) and ease of travel by road or river . During a 47 day period between May and June , 2009 , educators performed outreach in 7 locations ( 4 towns and 3 villages ) to ∼19 , 000 people . ( In towns— locations with populations >1000 – multiple presentation were made in order to cover individual neighborhoods . ) A further ∼4500 persons , from 9 locations ( 2 towns , 7 villages ) , received outreach during a 43 day period between July and October of that same year . For both periods approximately half of the enumerated population residing in the area covered attended the outreach activities ( 47% for the former period and 64% for the latter ) . The geographic zone of outreach coverage extended from Enyelle in the north to Congomelembe in the south , encompassing 3 sectors within Likouala—north , south and central ( Figure 2 ) . The northern sector included the greatest number of outreach attendees ( 11 , 057 ) , followed by the southern sector ( 7 , 860 ) then the central sector ( 4 , 943 ) . Prior to conducting outreach for the entire village or town , roughly equal numbers of individuals representing the 5 age/sex categories ( described above ) were selected to participate in structured small-group discussions . Further , within four of these small groups , approximately 1 out of 10 persons was recruited ( selected among volunteers until the quota was met ) to complete pre and post-screening interviews ( Table 1 ) . The mean age of persons completing the pre- and post-test questionnaires was 33 years overall . This was consistent across the three geographic sectors . A somewhat higher percentage of males than females completed questionnaires ( 57% ) , though the distribution neared equivalence in the central sector . In each of the three sectors monkeypox ‘experienced’ interviewees were identified . These were individuals who at the time of the 2009 outreach professed a prior knowledge of monkeypox that had been obtained during the prior 12 months from a health professional or education specialist . Whether the interviewee recalled INCEF by name or not , all educational encounters were presumed to be INCEF encounters as no other outreach for monkeypox was being undertaken in the region . The northern sector had the lowest proportion of ‘experienced’ interviewees which accorded with our expectations based on the locations of pilot outreach activities ( performed 6 months to one year prior ) . The pre-and post-test questionnaires addressed the interviewee's knowledge of the signs and symptoms of monkeypox ( ‘disease recognition’ ) , and the principal modes of inter-human and zoonotic virus transmission ( ‘disease transmission’ ) , as well as his or her past and future ( intended ) behaviors with regard to found animal carcasses ( ‘zoonotic risk’ ) and seeking of health care ( ‘health seeking’ ) ( Table 2 ) . For most knowledge and behavior subject areas— with the exception of issues relating to ‘zoonotic risk’— the proportion of interviewees who exhibited enhanced knowledge or a stated intention toward a constructive behavior after the outreach was statistically significant . For example , the proportion of interviewees who could recognize lesions on the palms of the hands and soles of the feet as a sign of monkeypox increased from 14 to 51% ( p<0 . 001 ) . As well , the proportion who said that they would avoid touching an animal that they found dead in the forest ( with no known cause of death ) increased from 23 to 61% ( p<0 . 001 ) . A significant gain in the proportion of interviewees who could identify the possibility of monkeypox virus transmission via objects that had been used by a patient ( linens , clothing ) was observed , but the final proportion of individuals exhibiting the gain remained low ( 14% ) . For multiple subject areas , gains in knowledge exceeded 30% , and for 2 ( the ability to identify one sign or symptom of monkeypox , and the willingness to take a family member with monkeypox to the hospital ) , the final proportion of the interviewees who had improved knowledge or who professed a willingness to perform a constructive behavior ( such as willingness to take an ill family member to a hospital ) exceeded 85% . However , when queried about past behaviors and intended future behaviors relevant to collection of wild rodents and primates , some intended behaviors remained invariant . For example , while few interviewees ( n = 29 , 11% ) stated that they had in the past eaten a monkey that they had found dead in the forest , the number remained essentially unchanged after interviewees saw the films and were then asked if they would engage in the behavior in the future ( n = 30 , 11% ) . This trend was consistent for collection and for sale of found primate carcasses as well . Considerably more interviewees stated that they had eaten rodents or squirrels that they had found dead in the forest ( n = 88 , 33% ) , but here after viewing the films the proportion who said that they would do so again after diminished by a significant fraction ( n = 43 , 16% ) ( p<0 . 001 ) . The result was similar with regard to intended sales of rodent or squirrel carcasses . Of interest , both before and after viewing the films , very few ( n = 4 , 3 , respectively ) said that they would take an ill family member ( suspected of having monkeypox ) to a traditional healer . When looking at response patterns in light of various categories of interviewees—tabulated by sex , age or geographic location—we found that geographic location was the variable most often significantly associated with the response outcome . ( The association was found for 12 questions , asked either before or after film screening . ) In general , interviewees in the southern sector were significantly more knowledgeable than those in the central or northern sector about monkeypox symptoms and disease transmission , and this group had the lowest proportion of interviewees indicating prior and intended future behaviors relating to the consumption of found animal carcasses . Beyond this , few categorical differences were noted when responses were evaluated with regard to sex or age . However , after seeing the films , female interviewees were significantly more likely than male counterparts to say that they would avoid direct contact with an ill person ( 66% vs . 53% p = 0 . 034 ) . As well , prior to seeing the films juveniles ( <15 years ) were less aware than adults of the risk of disease transmission from direct contact with an ill person ( 15% vs . 33% , p = 0 . 011 ) and were less likely to avoid contact with a sick person ( 10% vs . 32% , p = 0 . 003 ) . However , after seeing the films juveniles reported being significantly less likely than adults to say that they would consume the carcass of a found rat or squirrel ( 6% vs . 19% , p = 0 . 032 ) . Responses to 4 exemplar questions , broken down by participant category , are shown in Figure 3 . One factor which seemed to influence the how interviewees answered questions prior to viewing the films , was whether he or she had had previous experience with monkeypox outreach . Within the pool of interviewees , 79 persons were identified as ‘experienced’ , the remaining ‘naïve’ ( Table 3 ) . Prior to viewing the films , experienced interviewees were significantly more likely to know the symptoms of monkeypox and that the virus is transmissible by direct contact with someone who is ill . Experienced interviewees were also significantly more likely to say prior to seeing the films that monkeypox could be avoided by avoiding direct contact with someone who is ill and by avoiding contact with animals found dead in the forest . For the most part , these differences became less apparent or unapparent after the current intervention took place . The open-end responses to interview questions offer additional insight into the film participants' knowledge and behavior regarding monkeypox . A summary of pertinent responses is given in Table 4 . The most common anecdotes recorded regarded the cultural norms associated with hunting , selling , and eating forest animal products ( bushmeat ) . These often involved perceptions and beliefs surrounding the disease including conspiracies involving the introduction of the virus to the area or a disbelief in the existence of disease . Also implicated were the roles of hunters in bringing disease into communities and the necessity to collect forest animal products the only source of protein or income for family members . Additional themes from these responses included a belief that the disease was not present in one's particular village , or it was not necessary to change behaviors until the disease emerged in that area . Confusion also seems to exist between the role of animals from the forest ( primates ) and villages ( rodents ) in spreading disease . Some respondents also expressed a desire for vaccination and treatment options in addition to basic education . As of 2007 , WHO's Global Health Observatory ( http://apps . who . int/ghodata/ ) estimated that the prevalence of community health workers in the Republic of the Congo was <0 . 5 per 10 , 000 persons . ( For purposes of comparison , Rwanda in 2004 reported 14 community health workers per 10 , 000 . ) In many areas of the country the concentration of community health workers is insufficient to sustain effective surveillance for communicable disease threats , suggesting that it may be necessary to mobilize the community members themselves to assist in identifying early instances of disease . However , there are obstacles to garnering effective participation by community members in these activities . The populations most at risk for monkeypox are likely to be hard to access , due to limited infrastructure for both communication and transportation , and they may be harder yet to mobilize because of intrinsic cultural and linguistic barriers . In this paper , we described a method for culturally-appropriate community-based monkeypox outreach that has been demonstrated to be scalable for large numbers of individuals over a broad geographic expanse , and which we demonstrate to be effective in imparting basic disease-specific knowledge to persons with relatively low levels of health literacy . The monkeypox outreach program described here was performed in 16 towns and villages in northern Republic of the Congo , and involved people many of whom had little formal education and most of whom had little access to health services . Despite this , among small-group participants we were able to demonstrate substantial gains—and high endpoints percentages – in people's ability to recount the major symptoms of monkeypox ( 95% could recount at least one major symptom after the outreach , and roughly a third identified both rash and fever ) and in their professed willingness to seek healthcare when they suspect a family member has the illness ( 87% said they would do so after the outreach ) . We would anticipate that in combination these two elements could have a considerable impact on disease reduction through early case identification and diminished opportunities for community-based transmission of monkeypox virus . However , there are multiple subject areas for which our messages about disease transmission or risk mitigation strategies resonated less successfully . For example , while the proportion of interviewees who could identify contaminated fomites ( bedding , clothing ) as a vehicle for transmission was significantly increased after seeing the films , the overall endpoint remained relatively low at 14% . As well , after seeming the films a similarly low proportion ( 13% ) identified avoidance of potentially contaminated items as a means to reduce the risk of contracting monkeypox . Also of note , a small proportion of individuals reported after seeing the films that they would continue to eat the carcasses of dead primates found in the forest ( carcasses for which no signs of trauma or cause of death was evident ) . These same individuals reported having done so in the past; their intended behaviors were essentially unchanged after participating in the outreach . While the numbers of individuals who purportedly engage in such behaviors is relatively small , the activities in question could be of measureable importance to the manner in which virus enters into human communities . Small-group participants seemed far more accepting of avoiding contact with ( i . e . , eating , selling ) the found carcasses of rodents or squirrels . Young people ( those under 15 years of age ) were significantly more apt to say that they would avoid doing such than older people , after having seen the films . During this outreach period which took place from May–October , 2009 , several villages and towns were included which had also participated in pilot outreach activities during the year prior . The pilot program was constructed with the same core of learning objectives and recommendations , and also utilized film and discussions , but the materials employed were modified by INCEF after the pilot to better conform to local needs . The fact that some interviewees had had prior experience afforded us with some insights into the durability of the knowledge imparted using these methods . Interviewees who we identified as having had prior experience with monkeypox outreach were in particular better able to answer questions addressing disease recognition than were those who had not received outreach during the pilot phase . Much of the pilot-phase outreach took place in the southern sector and indeed interviewees from the south in general displayed more knowledge about monkeypox and more often professed an intention to pursue risk-reduction behaviors than did those from the central or northern sectors . Attempts were made to minimize potential biases introduced to the survey during selection of subjects and questionnaire design , but the possible influence of volunteer bias ( a type of selection bias ) and attention bias ( a measurement bias ) toward inflation of positive findings cannot be entirely overlooked . Whether pronounced or subtle , these biases have the impact of reducing our confidence that findings from this evaluation can be directly extrapolated to the small-group participants or to the broader population of Likouala residents who participated in the outreach . We also cannot discount potential error stemming from the inexperience of this population in participating in interview-driven questionnaires . When interpreting interviewee responses pertaining to intended behaviors , it must be remembered that a person's stated intention of a future behavior may have little to no bearing on what that person is actually apt to do . Measuring these behaviors will be an important objective for future evaluation studies . The goal of this outreach effort was and is to assist people living in communities at risk for the monkeypox virus to participate in preventing the introduction and spread of disease . The results of this evaluation indicate that our current program of film-based educational activities has been effective in improving disease-specific knowledge in both juvenile and adult community members , and may have the effect of encouraging individuals to seek out the advice of health workers when they are confronted with suspicion of monkeypox . Our next steps involve improving the educational messaging to better inform individuals about all sources of virus transmission ( including fomite ) and to better encourage risk-mitigation strategies for preventing zoonotic infections . The ultimate measure of the impact of this program will be an increase in community-wide detection and reporting of disease alongside enhanced prevention efforts .
Human monkeypox is a potentially severe illness that begins with a high fever soon followed by the development of a smallpox-like rash . Both monkeypox and smallpox are caused by infection with viruses in the genus Orthopoxvirus . But smallpox , which only affected humans , has been eradicated , whereas monkeypox continues to occur when humans come into contact with infected animals . There are currently no drugs specifically available for the treatment of monkeypox , and the use of vaccines for prevention is limited due to safety concerns . Therefore , monkeypox prevention depends on diminishing human contact with infected animals and preventing person-to-person spread of the virus . The authors describe a film-based method for community outreach intended to increase monkeypox knowledge among residents of communities in the Republic of the Congo . Outreach was performed to ∼23 , 600 rural Congolese . The effectiveness of the outreach was evaluated using a sample of individuals who attended small-group sessions . The authors found that among the participants , the ability to recognize monkeypox symptoms and the willingness to take ill family members to the hospital was significantly increased after seeing the films . In contrast , the willingness to deter some high-risk behaviors , such as eating animal carcasses found in the forest , remained fundamentally unchanged .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[ "medicine", "infectious", "diseases", "public", "health", "and", "epidemiology", "zoonoses", "skin", "infections", "smallpox", "global", "health", "neglected", "tropical", "diseases", "viral", "diseases", "public", "health" ]
2011
Assessing the Effectiveness of a Community Intervention for Monkeypox Prevention in the Congo Basin
Onchocerca volvulus is the causative agent of onchocerciasis , or “river blindness” . Ivermectin has been used for mass treatment of onchocerciasis for up to 18 years , and recently there have been reports of poor parasitological responses to the drug . Should ivermectin resistance be developing , it would have a genetic basis . We monitored genetic changes in parasites obtained from the same patients before use of ivermectin and following different levels of ivermectin exposure . O . volvulus adult worms were obtained from 73 patients before exposure to ivermectin and in the same patients following three years of annual or three-monthly treatment at 150 µg/kg or 800 µg/kg . Genotype frequencies were determined in β-tubulin , a gene previously found to be linked to ivermectin selection and resistance in parasitic nematodes . Such frequencies were also determined in two other genes , heat shock protein 60 and acidic ribosomal protein , not known to be linked to ivermectin effects . In addition , we investigated the relationship between β-tubulin genotype and female parasite fertility . We found a significant selection for β-tubulin heterozygotes in female worms . There was no significant selection for the two other genes . Quarterly ivermectin treatment over three years reduced the frequency of the β-tubulin “aa” homozygotes from 68 . 6% to 25 . 6% , while the “ab” heterozygotes increased from 20 . 9% to 69 . 2% in the female parasites . The female worms that were homozygous at the β-tubulin locus were more fertile than the heterozygous female worms before treatment ( 67% versus 37%; p = 0 . 003 ) and twelve months after the last dose of ivermectin in the groups treated annually ( 60% versus 17%; p<0 . 001 ) . Differences in fertility between heterozygous and homozygous worms were less apparent three months after the last treatment in the groups treated three-monthly . The results indicate that ivermectin is causing genetic selection on O . volvulus . This genetic selection is associated with a lower reproductive rate in the female parasites . We hypothesize that this genetic selection indicates that a population of O . volvulus , which is more tolerant to ivermectin , is being selected . This selection could have implications for the development of ivermectin resistance in O . volvulus and for the ongoing onchocerciasis control programmes . Onchocerca volvulus is the filarial nematode , transmitted by Simulium flies , that causes human onchocerciasis , or “river blindness” . It is estimated that 37 million people , mostly in Africa , are infected with this worm [1] . At present , ivermectin ( IVM , Mectizan ) is the only safe drug available for mass treatment of onchocerciasis . IVM , administered at the standard dose of 150 µg/kg , has a rapid effect on the embryonic stage of the parasite , the microfilariae ( mf ) , which cause most of the ocular and cutaneous manifestations of the disease . As a result of this microfilaricidal effect , the skin microfilarial loads decrease by 95–99% within one month after treatment . The drug also blocks the production of new mf by the adult female worms , who only resume mf release 3–6 months after treatment . This “embryostatic effect” of IVM explains why the mf loads remain at very low levels for up to one year . Furthermore , IVM treatments repeated at 1- to 3-monthly intervals have some , though moderate , effect on the longevity of the adult worms ( “macrofilaricidal effect” ) [2 , 3] . The drug , when given repeatedly , is therefore acting on at least three components of parasite fitness: reproduction , microfilarial survival and adult parasite lifespan , which together affect morbidity and the intensity of transmission . Due to the limited macrofilaricidal effect of the drug , treatments must be repeated and sustained . Endemic communities in Africa receive annual IVM treatment , while those of Latin America receive semi-annual treatments . To date , more than 400 million treatments have been distributed in Africa [4] , with some individuals having received up to 18 annual treatments . Due to this enormous drug pressure on the parasite , there is a risk of resistance of O . volvulus to the drug [5–7] . This concern is justified by reports of suboptimal responses to IVM from Sudan [8] and Ghana [9 , 10] , although in the former report reduced immune responsiveness in some of the treated people has been suggested as a possible explanation for the suboptimal responses to IVM . And in the study in Ghana the poor responses have been attributed to the parasites , with adult female worms resuming microfilarial production earlier after treatment than classically described . More recently , another report in Ghana [11] shows the first unequivocal parasitological and epidemiological evidence of ivermectin resistance in O . volvulus populations . In addition to this evidence of IVM resistance , changes in the genetic structure of O . volvulus populations , associated with IVM treatments , have been observed in parasites from Ghana [12–16] . These changes occurred particularly on the β-tubulin gene [16 , 17] , which has been associated with IVM resistance in the sheep intestinal nematode Haemonchus contortus [17] . However , in these previous studies , O . volvulus from IVM-naïve and -treated human populations were collected from different individuals in different communities . It is important to assess whether the genetic changes reported in O . volvulus are associated with a reduced response to IVM in any of the three effects of IVM on parasite fitness , described above . Furthermore , to eliminate the possibility that differences in genotype frequencies between IVM-naïve and -treated populations could be due to geographical effects , due to separate individuals and communities being sampled , it is important to assess whether changes in genetic frequency could occur in parasites collected from the same individuals before and after exposure to IVM . Genetic changes clearly associated with treatment , which could not possibly be associated with other covariates , would provide unequivocal evidence of genetic selection by IVM on O . volvulus . Such treatment-induced selection would be heritable . Heritable genetic changes that could reduce the susceptibility of O . volvulus to any of the effects of IVM on the parasite could have long-term consequences for the control of onchocerciasis because there is currently no alternative drug available for mass treatment of this disease . In a previous study [18] , we reported that in an IVM-naïve O . volvulus population from Cameroon , adult female worms presenting a homozygous genotype for β-tubulin were more fertile than adult worms that were heterozygous at this locus . In the present study , we have analyzed genetic characteristics ( β-tubulin gene and two control genes , heat shock protein 60 ( hsp60 ) and acidic ribosomal protein ( ARP ) ) and phenotypic characteristics ( female worm fertility ) of parasites collected , in the same individuals , before and after 4 or 13 IVM treatments over a three-year period . These treatments were administered as part of a clinical trial conducted in Central Cameroon . The main objective of this trial was to assess the effects of different regimens of IVM treatment on the mortality of O . volvulus adult worms , and the results of this phase have been published elsewhere [3] . In the second phase , results of which are presented in this paper , we evaluated whether repeated treatment with IVM led to ( a ) genetic changes in the adult worm population and ( b ) any modification of the relationship between β-tubulin genotype of the female worms and their reproductive status . The study was carried out in the Mbam Valley , a region hyper-endemic for onchocerciasis , located in the Central province of Cameroon , where no IVM had been distributed at the beginning of the study and where no vector control activities have ever been performed . In this area , before the introduction of IVM , the intensity of infection in the population , as expressed by the Community Microfilarial Loads ( CMFL ) [19] ranged between 10 and 114 mf per skin snip ( mf/ss ) [3] . The full details of the clinical trial , which was approved by the Cameroonian Ministry of Public Health and by Merck and Co . , the manufacturer of IVM ( Mectizan ) , have been published elsewhere [3] . The study also subsequently received approval from the institutional review board of McGill University . Briefly , 657 individuals were selected using the following inclusion criteria: men between 18 and 60 years old , with at least two palpable nodules during the preliminary examination but otherwise in good health , who had not received any filaricidal treatment within the five previous years , and who agreed to participate in the trial by signing an informed consent form . These patients were randomly allocated to one of the four following IVM treatment groups: 150 µg/kg body weight/year ( standard group; group 1 ) ; 150 µg/kg/three-monthly ( group 2 ) ; 800 µg/kg/year ( group 3 ) ; and 800 µg/kg/three-monthly ( group 4 ) . Over the three-year study period , patients received either 4 or 13 IVM treatments . In order to assess the macrofilaricidal effect of IVM on O . volvulus , adult worms were collected , by nodulectomy , at the outset of the trial ( before the first IVM dose was administered ) and once again after three years of treatment in the four different treatment groups described above . The protocol used for parasite collection was identical for the two rounds of nodulectomy . Just before each round of nodulectomy , each person was carefully examined and all the sites on their body where a nodule or a group of nodules was palpable were recorded on a body chart . Subsequently , one of the sites was selected at random and all the nodules located at this site were removed from each person . The site selected for the second nodulectomy was one of those recorded at the outset of the study so that the worms collected at that time had probably been subjected to the IVM treatments administered over the previous three years . Just after the nodulectomy , all the nodules collected were immersed in fixative ( 70% ethanol , 20% water , 10% glycerol ) . One of the nodules was used for histological examination , as previously described [3] , to evaluate the status of the worms . Any additional nodules ( “extra nodule” ) from the excision site were stored in the fixative at room temperature and available for genotyping and phenotyping . Of the 657 individuals selected before treatment , 290 had more than one nodule at the first nodulectomy site , and thus at least one “extra nodule” available after the histological examination . Similarly , of the 541 patients present at the second round of nodulectomy ( following three years of treatment ) , 156 had at least one extra nodule available . Patients included in the present study were selected taking into account our objectives , which were to assess the genotypes of three polymorphic genes , including β-tubulin , in the adult worms , and any relationship between the genotype of female parasites and their reproductive status , before and after IVM treatment . To make the comparison more sensitive , we performed the genotyping and the phenotyping only on parasites obtained from those people for whom “extra nodules” , containing at least one adult worm , had been collected at both nodulectomy rounds ( pre-treatment and after three years of repeated treatments ) . The total numbers of individuals who met these inclusion criteria were 18 in group 1 , 16 in group 2 , 22 in group 3 and 17 in group 4 . Thus , the analyses were performed on the nodules collected from 73 individuals . This procedure has been described previously [18] . In 2002 , the nodules were washed with phosphate buffered saline ( PBS ) for 24 h with regular changes of medium in order to remove all residues of fixative . The nodules were then digested in collagenase [20] . Worms were collected and stored individually in labelled Eppendorf tubes , which were frozen at −80°C . Each female worm was phenotyped by microscopical examination of its reproductive status in terms of the presence of mf and embryos . Three phenotypes were defined: ( a ) non-fertile females , i . e . worms with empty reproductive organs , ( b ) females with low fertility , in which the reproductive organs contained only a few embryos , but no mf , and ( c ) fully fertile females , in which the reproductive organs were full of mf and embryos . After the phenotyping , each worm was disrupted and its DNA was extracted using a Dneasy kit ( Qiagen Inc . , Mississauga , Canada ) . Heat shock protein 60 ( hsp60 ) ( GenBank , AF121264 ) , which is a molecular chaperone that participates in the folding of proteins , was chosen as a control gene . It was known to be polymorphic and previously found not to be selected by IVM treatment in O . volvulus [16] . Two polymorphs ( “A” and “G” ) were found in the hsp60 gene partial sequence analyzed . The region analyzed started at position 214 on the cDNA and included 100 bp in the exon , followed by 276 bp in the intron . The A/G polymorphism was located in the intron region . The fragment of 376 bp was amplified by PCR from individual adult worms with the primers 5′CAA TCA TGG GGA AGT CCA AAG 3′ and 5′CTC AAA ACC TTC CTT TGC AAT 3′ at Tm = 53°C . PCR products were sequenced with the hsp60 anti-sense primer using the 3730XL DNA Analyzer system ( McGill University/Genome Quebec Innovation Centre ) . Platinum Taq DNA polymerase High Fidelity ( Invitrogen ) was used in the PCR reaction to avoid introduction of error during amplification . Each individual chromatogram was analyzed with Sequencher 4 . 7 software ( Gene Codes Corporation , Ann Arbor , MI , USA ) , to detect the homozygotes AA and GG and the heterozygotes AG . The acidic ribosomal protein ( ARP ) gene ( GenBank , AI130565 ) , which is involved in protein synthesis , was chosen as a second control because it was expected to be polymorphic [21] and not known to be sensitive to IVM treatment . Two polymorphs ( “C” and “T” ) were found in the acidic ribosomal protein gene partial sequence analyzed . The region of interest was from 1270 bp to 1488 bp of the complete gene . It was amplified by PCR from individual adult worms with the primers 5′ TGA AAA ACT GCT ACC GCA TA 3′ and 5′ AAA TTT TCG TTG GAA TTT GC 3′ at Tm = 54°C . PCR products were analyzed by restriction fragment length polymorphism , based on C/T polymorphism apparent in the EST database , using the restriction enzyme Mnl 1 for 2 hours , and subjected to electrophoresis on a 12% polyacrylamide gel ( 39∶1 ) for 2 hours at 130 V , stained with ethidium bromide and visualized using an ABI Imager ( Bio-Rad , Hercules , CA , USA ) . Two alleles ( “a” and “b” ) have been described for β-tubulin [16] . These two alleles have three single nucleotide polymorphisms in an exon region . These differences lead to changes in three amino acids in the putative protein sequence . The worms were genotyped individually for β-tubulin ( GenBank , F019886 ) by PCR amplification followed by amplicon length analysis [17] . The aim of the analysis was to assess whether a variety of covariates related to the worm , nodule or patient characteristics were associated with three different dependent variables: ( a ) the inability to genotype some of the worms from the preserved nodules; ( b ) the frequency of the various polymorphs analyzed; and ( c ) the degree of fertility of the worm . We considered the five following covariates: the age of the patient at the outset of the trial ( continuous variable ) ; the CMFL in the village of residence of the patient , defined in four categories: 10–40 , 41–60 , 61–70 , and 71–114 mf/ss; the treatment group ( for analysis of the worms collected post-treatment: 150 µg/kg/year , 150 µg/kg/three-monthly , 800 µg/kg/year , and 800 µg/kg/three-monthly ) ; the total number of females in the nodule; and the total number of palpable nodules on the patient at the outset of the trial . In addition , we also assessed the degree of fertility in relation to the genotype of the worms and to the total number of males in the nodule . The procedure for genotyping the worms failed with a significant number of worms obtained from the nodules that had been preserved at room temperature for 5 to 8 years . To test whether this inability to genotype some worms could be explained by sampling biases , we assessed , using multivariate logistic regression , whether the success in genotyping the worm ( genotyped vs . non-genotyped status ) was associated with one or the other of the possible covariates quoted above . All regressions analyses were performed using Stata v9 . 0 ( Stata Corporation , TX , USA ) , where parameters were estimated using the cluster option [22] accounting for intra-nodular correlation . Hardy-Weinberg equilibrium was tested using the χ2 test , unless the sample size was small . In this case , Fisher's exact test was used . The genotypic frequencies before and after treatment were compared using Fisher's exact test . To evaluate whether some host covariates or village characteristics may have influenced the heterozygosity of the worms , the association between heterozygous status and the five main possible covariates quoted above was assessed separately on pre- and post-treatment data , by multivariate logistic regressions . Potential intra-nodule clustering was accounted in the regression models . Logistic regression models were used to analyze the independent variables associated with the fertility of the female worms before and after treatment . The dependent variable “fertility” was defined , for this analysis , using two categories: no or low fertility versus high fertility [18] . This choice is based on the fact that only worms with mf have the possibility of having their progeny transmitted , at the time of sampling , and this may be relevant to the possible transmission of any “resistant” genotypes . However , any treatment group effect on fertility status could be due to either treatment frequency or to the fact that the worms were collected three months after the last treatment in the three-monthly treated groups ( groups 2 and 4 ) and twelve months post-treatment in the annual groups ( groups 1 and 3 ) . The possible covariates in the model included the five host-related independent variables defined above , and two other independent variables: the genotype of the worm at the β-tubulin locus ( homozygous versus heterozygous ) , and the total number of males present in the nodule . Here again , the intra-nodule clustering was considered in the logistic regressions . The χ2 and Fisher's exact test analyses were performed using VassarStats ( http://faculty . vassar . edu/lowry/VassarStats . html ) . We previously showed , in a sample of 320 female worms collected before treatment as part of the same trial , that the 90 worms that could not be genotyped for β-tubulin did not differ significantly , with regard to several host independent variables , from the 230 worms that could be genotyped [18] . Similar results were obtained when comparing the 65 non-genotyped females to the 183 genotyped ones , and the 63 non-genotyped males to the 56 genotyped males , collected before treatment , from the 73 people from whom nodules could be analyzed both before and after treatment . The proportion of female worms that could not be genotyped for β-tubulin was significantly higher after treatment ( respectively 26 . 2% and 35 . 9% before and after treatment; p = 0 . 043 ) . Among the 153 female worms collected after treatment , 55 could not be genotyped . According to multivariate logistic regression , the “genotyped” status was not associated with any of the five covariates included in the analysis . The proportion of non-genotyped male worms did not differ significantly before and after treatment ( respectively 52 . 9% and 63 . 4%; p = 0 . 17 ) . After treatment , we observed that a significantly higher proportion of male worms could be genotyped in the 800 µg/kg/year treatment group ( OR = 3 . 97 ( 95% CI , 1 . 05–15 . 08 ) ; p = 0 . 043 ) compared to the standard group . None of the other independent variables differed significantly between the genotyped and the non-genotyped male worms . Taken together , these results do not provide evidence of bias between the 45 non-genotyped and 26 genotyped male worms according to the tested covariates . Before treatment , β-tubulin heterozygous status was not influenced by age of host , total number of females in the nodule , total number of palpable nodules or CMFL in the village of residence ( Table 3 ) . After treatment , none of the tested covariates was significantly associated with the β-tubulin heterozygous status , except the fact of living in a village with a CMFL between 41 and 60 mf/ss . This weak association ( OR = 6 . 24 ( 95% CI , 1 . 09–35 . 59 ) ; p = 0 . 039 ) might indicate that the probability of being heterozygous was higher in the villages where infection rates were rather high ( Table 4 ) . The probability of being heterozygous tended to be higher in the groups treated three-monthly . Even if this was not significant in the analysis taking into account the various groups separately , this trend is consistent with the results presented above comparing the pooled annual treatment groups and the pooled three-monthly treatment groups . Before treatment , the homozygote genotype was the only independent variable associated with a high fertility phenotype ( p<0 . 002 ) ( Table 5 ) . After treatment , high fertility of the worms was still associated with the homozygous genotype ( p = 0 . 035 ) . In addition , high fertility of the worms was more likely to be observed amongst younger patients ( p = 0 . 018 ) . Finally , high fertility in the worms was more apparent in nodules containing higher numbers of male worms ( p = 0 . 030 ) ( Table 6 ) . As the intervals between the last IVM treatment and nodulectomy in the three-monthly groups and the annual treatment groups were different , they have also been considered separately ( Figure 2 ) . Twelve months after the last IVM treatment , analysis of the fertility ( non- and low fertility versus full fertility ) in relation to genotype showed that the β-tubulin homozygous worms remained more fertile than the heterozygous worms ( χ2 = 11 . 06 , p<0 . 001; Figure 2 ) . Because the sample size was small , we did not perform an analysis on the data collected on the three-monthly groups ( samples collected three months after the last IVM treatment; groups 2 and 4 ) . However , the figure shows that the proportion of fully fertile worms was higher in the homozygous worms ( 42% compared with 24% ) , but both genotype groups showed a similar proportion of non-fertile worms ( 50% and 48% , respectively , for the homozygous and heterozygous parasites ) . IVM has been used since the late 1980s , and more than 400 million doses have been distributed in Africa [4] . It remains the only safe drug for community treatment of onchocerciasis . Our results clearly show a genetic selection in O . volvulus caused by repeated IVM treatment . Since the parasites were collected before and after treatment from the same patients , these results cannot be explained as differences arising from different host populations being sampled . These results , together with other evidence of genetic selection and reports of sub-optimal responses to IVM , provide a warning that selection for IVM resistance could be occurring in some populations of O . volvulus . In view of these results , it is imperative that field studies be undertaken to characterize all treatment responses to IVM in O . volvulus , coupled to further genetic analysis , in order to confirm or not the possible emergence of IVM resistance . Such longitudinal studies , which would look at the repopulation of the skin of treated people by mf , should be undertaken without delay if the benefits that have been achieved by the onchocerciasis control programmes are not to be lost as a result of the spread of IVM resistance in O . volvulus .
Onchocerca volvulus is the causative agent of onchocerciasis , or “river blindness” . Ivermectin has been used for mass treatment of onchocerciasis for up to 18 years , and recently there have been reports of poor parasitological responses to the drug and evidence of drug resistance . Drug resistance has a genetic basis . In this study , genetic changes in β-tubulin , a gene associated with ivermectin resistance in nematodes , were seen in parasites obtained from the patients exposed to repeated ivermectin treatment compared with parasites obtained from the same patients before any exposure to ivermectin . Furthermore , the extent of the genetic changes was dependent on the level of ivermectin treatment exposure . This genetic selection was associated with a lower reproductive rate in the female parasites . The data indicates that this genetic selection is for a population of O . volvulus that is more tolerant to ivermectin . This selection could have implications for the development of ivermectin resistance in O . volvulus and for the ongoing onchocerciasis control programmes . Monitoring for the possible development and spread of ivermectin resistance , as part of the control programmes , should be implemented so that any foci of resistant parasites can be treated by alternative control measures .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[ "pharmacology/drug", "resistance", "genetics", "and", "genomics/population", "genetics" ]
2007
Genetic Selection of Low Fertile Onchocerca volvulus by Ivermectin Treatment
Although bubonic plague is an endemic zoonosis in many countries around the world , the factors responsible for the persistence of this highly virulent disease remain poorly known . Classically , the endemic persistence of plague is suspected to be due to the coexistence of plague resistant and plague susceptible rodents in natural foci , and/or to a metapopulation structure of reservoirs . Here , we test separately the effect of each of these factors on the long-term persistence of plague . We analyse the dynamics and equilibria of a model of plague propagation , consistent with plague ecology in Madagascar , a major focus where this disease is endemic since the 1920s in central highlands . By combining deterministic and stochastic analyses of this model , and including sensitivity analyses , we show that ( i ) endemicity is favoured by intermediate host population sizes , ( ii ) in large host populations , the presence of resistant rats is sufficient to explain long-term persistence of plague , and ( iii ) the metapopulation structure of susceptible host populations alone can also account for plague endemicity , thanks to both subdivision and the subsequent reduction in the size of subpopulations , and extinction-recolonization dynamics of the disease . In the light of these results , we suggest scenarios to explain the localized presence of plague in Madagascar . Although bubonic plague has marked human history by three pandemics ( Justinian plague in the – centuries , Medieval plague in the – centuries and Asiatic plague since 1894 [1] ) , this zoonosis caused by the coccobacillus Yersinia pestis is primarily a rodent disease . Its persistent circulation in wild reservoirs is responsible for occasional epidemics in human populations [2] , [3] . Each plague focus has distinct characteristics , but all have mammal hosts as reservoirs and fleas as vectors . Two main factors are suspected to explain the endemic persistence of plague despite its high virulence . The first one is the coexistence of plague resistant and plague susceptible rodents in many wild foci of the world . Susceptible hosts are assumed to allow plague transmission by developing the high septicemia needed for the disease to spread , while resistant hosts would help maintain the host and flea populations and would lower the effective rate of encounter between infectious fleas and susceptible hosts [4]–[6] . The second factor which could explain plague endemism is the host metapopulation structure [7]–[10] , which may allow for extinction-recolonization dynamics of plague in local foci , between which the disease spreads slowly [11] , [12] . Theoretical models have shown that these extinction-recolonization dynamics are involved in the persistence of various infectious diseases , e . g . measles [8] , [13] , [14] . A few other mechanisms are thought to favour the persistence of plague , such as the presence of multiple hosts [5] , [15] , the possible persistence of Y . pestis in soils [16] , [17] , the direct transmission between rats inside burrows [18] , [19] or the heterogeneity in the phenology of the host reproduction [15] . These alternative explanations will be discussed at the end of the article . Explaining the endemism of plague has been the objective of several theoretical studies [15] , [20]–[22] . However , the roles of resistance and of metapopulation structure of rodent reservoirs for plague persistence have rarely been explored separately . For instance , the model developed by Keeling & Gilligan [21] showed that metapopulation structure can explain the long-term persistence of plague via extinction-recolonization dynamics of the disease , but the theoretical population that they modelled included some resistant individuals , so that the roles of both factors cannot be disentangled . This is also the case for theoretical studies on the Kazakh focus [22] , [23] , where the hosts are modelled as partly resistant . In populations of susceptible hosts , such as prairie dogs ( Cynomys spp . ) in the United States , the link between spatial structure and plague persistence has been empirically observed [24]–[26] and theoretically confirmed , at least on short time scales ( Salkeld et al . found that plague transmission between adjacent coteries of a structured susceptible host population lead to enzootic phases which last for more than 1 year in about 25% of model runs [27] ) . Also , several studies pointed out the need to test for the different mechanisms involved in plague endemism [5] , [28] . Madagascar is one of the major plague foci in the world , accounting for 31% of the 50 , 000 reported human cases worldwide between 1987 and 2009 [29] . Bubonic plague was introduced in Madagascar in 1898 [30] and spread to central highlands in the 1920s [31] . Since that time , the disease persists in this region at the landscape level . In coastal areas and regions below 800 m of altitude , only sporadic urban epidemics occurred , due to human-mediated translocation of infected rodents from central highlands [30] , [32] . In Madagascar , the main host of plague is the black rat , Rattus rattus [30] , that is widespread throughout the island [30] , while two species of fleas are involved as vectors [31]: Xenopsylla cheopis , the oriental rat flea , which has a cosmopolitan distribution , and Synopsyllus fonquerniei , an endemic flea from Madagascar whose distribution is restricted to central highlands . Compared to many other natural plague foci , only a few species are involved in the transmission of the disease [3] . Nevertheless , and despite its importance regarding public health , the causes of plague persistence in Madagascar have never been explored . Preliminary population genetic studies suggested that rat populations from the highlands are more geographically structured than those of the coastal areas , probably because of the more rugged physical landscape that limits migration [33] . Also , consistent with the hypothesis of a causal relationship between plague persistence and host resistance , at least 50% of the rats caught in Malagasy highlands are plague-resistant , whereas they are all susceptible in low altitude plague-free areas [34] , [35] . However , as highlands were colonised by rats from coastal areas several centuries ago [34] , the evolution of plague resistance in Malagasy black rats may be recent and posterior to the spread of the disease . It might thus be a consequence rather than a primary cause of plague persistence in rural areas of central highlands . Building on the model of Keeling & Gilligan [20] , [21] , we developed a theoretical approach to evaluate independently the roles of host population structure and of host resistance in the long-term persistence of plague . The model was parameterized using data from plague ecology in Madagascar when they are available or from the literature otherwise . The sensitivity of the model to a range of parameters was tested . We evaluated the consistency of the hypothesis suggesting a recent evolution of plague resistance in Madagascar , and identified the parameters that need to be measured in order to test it . Our model of plague epidemiological dynamics is built on the framework developed by Keeling and Gilligan [20] , [21] and is parameterized using data from studies on plague ecology in Madagascar when available [18] , [34]–[36] . The system of differential equations ( 1 ) accounts for the number of individuals and epidemiological status of the rat ( host ) and fleas ( vector ) populations . The rodent host population is composed of three phenotypes: healthy , plague susceptible rats ( whose number is in system ( 1 ) ) , healthy , plague resistant rats ( ) , and infectious rats ( ) . Two categories of vectors are taken into account: the mean number of fleas living on a rat ( pulicidian index , ) and the number of free infectious fleas ( ) . The birth rate of rats is assumed to be density dependent [37] and modelled by a logistic equation , being the maximal birth rate and the carrying capacity of the rat population . The rats are assumed to die naturally at constant rate . We assume no direct cost to resistance , however only a proportion of the offspring of resistant rats are resistant ( is the heritability of resistance; ) , the other offspring ( ) being all susceptible to plague [36] . In contrast , all the offspring of susceptible rats are born susceptible to plague [36] . Susceptible rats ( ) can contract the disease and become infectious ( ) while resistant rats ( ) always remain uninfected , which is a realistic assumption in the context of the Malagasy plague [36] . Infection happens when free infectious fleas ( ) land on susceptible rats ( ) and transmit the bacillus according to the transmission parameter . Free infectious fleas ( ) come randomly in contact with rats with a probability of encounter [38] . The parameter measures the search efficiency of fleas . Following [39] , the infection of rats by fleas is modelled as a frequency-dependent process . We thus consider that the force of infection is . Infectious rats quickly die from septicemia , which results in an additional mortality term , also called the virulence of the bacillus . The death of each of these rats leads to the release of fleas in the environment , increasing the number of free infectious fleas ( ) . Free infectious fleas die at rate . Fleas on the rats are assumed to have a density-dependent growth , with maximal growth rate and a carrying capacity per rat . All these assumptions result in the following system of differential equations: ( 1a ) ( 1b ) ( 1c ) ( 1d ) ( 1e ) Our model includes several modifications compared to the model of Keeling & Gilligan [20] , [21] , in order to better depict wild plague foci , and specifically that of Madagascar . In our model , ( i ) infectious rats do not recover , as frequently observed [4] , [6] , [36] , ( ii ) free infectious fleas either find a host or quickly die from starvation , a more explicit modelling of two events that were not distinguished in [21] , and ( iii ) the descendants of resistant rats which are resistant also have a density-dependent birth rate , while they grew exponentially in [20] . Nevertheless , the above changes do not change the main characteristics of the outputs of the model ( comparison of Figures 1 and 2 with Supporting Figures S1 and S2 ) . In the first steps of this study , model ( 1 ) is also analysed without the class of resistant rats ( and no resistant rats initially in the system; see system ( S1 . 1 ) in the Supporting Text S1 ) , in order to investigate their role in plague persistence . The parameter values that we use come preferentially from experiments or field observations done in the context of the Malagasy plague focus . When relevant data are lacking , parameters are derived from values found in the plague literature ( see Table 1 ) . The basic reproductive number of a disease , , is the expected number of secondary cases caused by one infected individual introduced into a susceptible population [40] . A disease is expected to spread only if is greater than unity . We calculated in our model using the Next Generation Approach [40] , [41] . Details of the calculations are presented in the Supporting Text S2 . The system of differential equations ( 1 ) is then solved numerically , using the deSolve package [42] in R [43] . Deterministic simulations are run on a time long enough to ensure that equilibrium states are reached ( typically years with our parameters ) . When it is possible to find analytical solutions for the equilibria ( for example with system ( S1 . 2 ) without fleas , in the Supporting Text S1 ) , we can check the accuracy of the numerical integrations of the model . There are four qualitative types of equilibrium states for the rat populations: ( i ) whole population extinction , labelled ( , ) or ( , , ) for the models without and with resistant rats , respectively; ( ii ) persistence of susceptible rats only , labelled ( , ) or ( , , ) ; ( iii ) persistence of susceptible and infected rats but extinction of resistant rats ( in systems containing resistant rats initially ) , ( , ) or ( , , ) ; and finally ( iv ) , in the model with resistant rats , coexistence of susceptible , infected and resistant rats , ( , , ) . We consider a class to be extinct when the number of individuals drops at least once below during the last years of the numerical integration ( to avoid any influence of the initial state of the system on the extinction criteria ) . To study the effect of spatial structure on disease persistence , a metapopulation of susceptible hosts only ( i . e . without resistant hosts ) of total carrying capacity equal to rats is modelled as a set of subpopulations of equal sizes . We neglect the effect of distance by assuming that all subpopulations are equidistant ( a situation called island model [44] in population genetics ) . We consider ( i ) no spatial structure ( ) , ( ii ) a weak spatial structure ( ie , a low population subdivision ) ( ) and ( iii ) a higher population subdivision ( ) . The fraction of infections that occur between subpopulations is given by the parameter . Although the rats in Madagascar may move temporarily to other subpopulations , thereby spreading the disease , capture-recapture studies have shown that these movements are temporary [18] , [45] . We therefore model a migrating force of infection , instead of the migration of the rats themselves [46] . The value of was estimated to be around 1% [18] , [21] . The force of infection in system ( 1 ) thus becomes in the subpopulation , which includes the rats , and the fleas : ( 2 ) To assess the effect of population structure on the persistence of the disease , we use a stochastic version of our model without resistant rats ( system ( S1 . 1 ) in the Supporting Text S1 ) , based on the Gillespie algorithm [47] that is implemented in the GillespieSSA R package [48] . It simulates a Markov stochastic process in continuous time and with discrete state values . We ran simulations both with and without metapopulation structure , in order to compare the persistence of plague ( a hundred replications for each set of parameters ) . The number of simulations where susceptible rats or infectious rats persist was recorded over time , to obtain an estimation of the probability of extinction of rats and through time . The scripts of the simulations ( deterministic and stochastic ) are deposited in the Dryad Repository: http://dx . doi . org/10 . 5061/dryad . 55t60 . We first investigated the different outcomes of the model without resistance ( variable ) and without population structure ( ) , depending on the value of the transmission parameter and of the maximal birth rate . When there are no resistant rats and no population structure ( system ( S1 . 1 ) in the Supporting Text S1 ) , the rat population is viable if the maximal birth rate of rats is larger than their mortality rate ( Figures 1 and 2 ) . The disease propagation threshold , , sets the limit between the rat population equilibria ( , ) and ( , ) . Using the Next Generation Method , we found ( 3 ) The propagation of plague is favoured by a high transmission from fleas to rats ( ) , which increases the number of infectious rats , and by a high flea carrying capacity of rats ( ) , which increases the number of free fleas , the vectors of the disease ( equation ( 3 ) ) . It is also favoured by a high carrying capacity of the rat population , , and by a high search efficiency of fleas , , through their direct effect on the probability that a flea finds a host . Finally , a high mortality rate of free infectious fleas , , disadvantages disease propagation by limiting the number of vectors . However , note that the fact that plague can initially spread in a susceptible rat population , although necessary , is not a sufficient condition for the long-term persistence of the disease . The parameters and cannot for now be estimated from data collected in Madagascar , but the value of is not sensitive to changes in their values ( see sensitivity analysis of in Supporting Figure S3 ) . The parameters and have more effect on the value of but their range of possible values is better known [6] , [18] . As we did not have a precise estimate of the value of ( the transmission rate varies between fleas and depends whether these are blocked or not; transmission efficiency was found to be about or for blocked X . cheopis [49] , [50] , and for unblocked X . cheopis [51] ) and as it had a strong effect on the value of , we investigated the outcomes of the model for a range of values of ( to ) and calculated the critical transmission of the disease , , which is defined as the value of for ( see equation ( 4 ) below ) . In the deterministic model , the disease initially spreads if and only if , which is equivalent to ( 4 ) The threshold transmission parameter is plotted as a horizontal black line in Figures 1 , 2 and 3; its values match with the values of obtained by numerical simulations ( Figure 1 ) . However , as already mentioned , the condition does not imply long-term persistence: Figure 1 shows that the equilibrium states with disease persistence ( , ) disappears when increases further above the critical transmission threshold , especially for large host populations . For rats and values of just above , strong oscillations of the number of rats occur in each class , with low values between the peaks ( Supporting Figures S4 and S5 , ) . For higher values of , no oscillations happen but an epidemic wave decimates the host population ( Supporting Figure S4 , ) : both the disease and the rat population therefore go extinct . Without resistant rats , the disease can thus not persist in the long run within large host populations , except for a thin range of values . However , for smaller population sizes , the amplitude of the dynamics decreases ( Supporting Figure S5 ) , which prevents the extinction of the rat population and allows for disease persistence ( Figure 1 ( b ) ) . We thus observe that above the critical transmission , high host population sizes disadvantage the long-term persistence of plague . It is worth noting that the same system without the flea compartment and with a direct disease transmission instead ( system ( S1 . 2 ) in the Supporting Text S1 ) shows stable equilibria when is above the critical transmission ( Supporting Figure S6 ) : the vectors therefore play a major role in the observed high amplitude dynamics leading to plague extinction by prolonging the infection process after the rats' death ( free fleas infected by Y . pestis survive long enough to widely spread the disease ) . This point highlights the importance of accounting for the flea demography whenever studying the epidemiology of plague . The above results are not highly dependent on the values of all other parameters linked to the behaviour of fleas and rats: changing these values may have a quantitative effect on the critical transmission but it does not modify the qualitative behaviour of the system ( see the sensitivity analysis of the equilibrium states on Supporting Figure S7 ) . When resistant rats were included into the system ( system ( 1 ) ) , three stable equilibrium states existed: ( , , ) , ( , , ) , and ( , , ) ( Figure 2 ) . The threshold for disease propagation , i . e . , corresponds here to the limit between the equilibria ( , , ) and ( , , ) . The expression of remains the same as without resistant hosts ( see equation ( 3 ) ) . Contrasting with the system without resistant rats , changing the carrying capacity of the rat population does not influence the equilibrium states: when including resistant rats in the model , plague persists as long as ( Figure 2 and Supporting Figure S8 ) . The sensitivity analysis showed that this result is not sensitive to changes in parameter values ( Supporting Figure S9 ) . The pattern of short epidemics followed by disease extinction that we previously observed and which was due to a lack of surviving susceptible rats , does not occur anymore because resistant rats allow the maintenance of not only resistant but also susceptible phenotypes ( through partial heritability of resistance , ) in the population . In order to study the effect of spatial structure alone , we here assumed that resistant rats were absent ( see system ( S1 . 1 ) in the Supporting Text S1 ) . The deterministic analysis of the system shows that host population structure alone allows for disease persistence when and ( Figure 3 ) . Indeed , when the metapopulation is subdivided into enough subpopulations , oscillations in the number of healthy and infectious rats occur in each subpopulation , but the numbers of individuals stay above unity . Thus , host population structure allows plague persistence for parameter values where the disease would go to extinction in non-structured populations . Fragmentation turns large host populations which undergo high amplitude cycling dynamics ( rats , Figures 1 ( a ) and 3 ( a ) ) into small subpopulations which undergo dynamics of decreased amplitude ( rats , Figures 1 ( b ) , 3 ( c ) and Supporting Figure S5 ) . However , if the total carrying capacity of the metapopulation is strongly decreased ( rats ) , then the densities of rats in each subpopulation become too low to allow for disease persistence . Stochastic analyses with parameter values such that revealed that even a weak spatial structure increases the time of disease extinction by several decades ( Figures 4 ( a ) and 4 ( b ) ) . The effect of population structure is twofold . First , population structure introduces extinction-recolonization dynamics of the disease between local foci ( Supporting Figure S10 ) , due to the asynchrony of the dynamics between subpopulations . Secondly , consistent with our deterministic results in non structured populations of susceptible rats , as long as remains above the critical transmission the disease persists for longer in smaller populations ( and rats , no disease persistence after years across all our simulations; see Figure 4 ( c ) ) than in larger ones ( and rats , no disease persistence after years in all our simulations; see Figure 4 ( a ) ) . Thus , the longer persistence of the disease in the four-subpopulation metapopulation of rats ( no disease persistence after years; see Figure 4 ( b ) ) is due to both a population size reduction ( in each subpopulation ) and to extinction-recolonization dynamics of the disease . However , if population subdivision is too high , or the subpopulations too isolated ( tested with and ) , extinction time decreases again , as recolonization events become rare ( Supporting Figure S11 ) . In rat populations without resistant rats , our results show that the persistence of the disease is favoured by intermediate population sizes . This may seem surprising given that the propagation of many infectious diseases is known to be favoured by larger host population sizes [52] , [53] . However , disease invasion and persistence are two very different phenomenons [54] , and the classically reported effect of population size on [53] is a matter of invasion rather than persistence . Here , the presence of vectors , the fleas , amplifies the spread of the disease ( the fleas act as a very short-term external reservoir , [55] , [56] ) and thus triggers , after an intense epidemic , the extinction of the disease in large populations . Accordingly , other empirical studies reported that high host carrying capacities favour the invasion but not the persistence of plague . In Kazakhstan for instance , plague epidemics have been shown to be preceded by an increase in gerbil abundance over a minimum abundance threshold [22] , [27] , but the abundance of gerbils would predict plague endemicity ( ie , long-term persistence ) less than the probability of plague epidemics [23] . Heier et al . [23] suggested that if the initial spread of plague is faster when the population density of rodents is high , so is the extinction of the rodent population [23] . In large rat populations , we find that the presence of resistant rats alone may explain plague persistence . This confirms the hypothesis of a role for resistance in the endemism of plague [55] . Keeling & Gilligan [21] showed that if the initial proportion of resistant rats is below 20% , then short epizootics are more likely to occur than disease persistence . Previous theoretical studies [28] , [57] assumed that resistant rats could play the role of plague reservoirs , by carrying infectious fleas , or that infectious rats could recover [21] , whereby restoring the population of disease-sensitive rats . Our results show that these assumptions are not required to account for the persistence of this highly virulent disease , but that what matters most is the fact that resistant rats provide a source of new sensitive rats ( since resistance is not totally heritable , ) . Also , spatial structure alone may account for plague persistence . The possible recovery of infectious rats and the presence of resistant hosts were included in the model of Keeling and Gilligan [21] , but we show here that they are not necessary to induce the long-term persistence of plague . A weak structure is enough to explain decades of disease persistence . It confirms what was already suggested by Salkeld et al . [27] . The effect of spatial structure is related to the combined effects of reduced subpopulation sizes and asynchrony between subpopulations . As in [58] , [59] , we indeed find that plague extinction takes longer for an intermediate force of coupling , , between subpopulations . Interestingly , the extinction-recolonization dynamics we observe happen to have about the same tempo as chronic re-emergences that have been recorded in some plague foci , such as the Kazakh focus , where epizootics last two to five years and occur every two to eight years [60] . Our results on the role of spatial structure are supported by field observations on prairie dogs ( Cynomys sp . ) : in the United States , prairie dog colonies are on average smaller and separated by larger distances in regions where plague has historically been endemic than in regions where plague is historically absent [24] . Mortality due to Y . pestis is , for prairie dogs , close to 100%: the theoretical model we developed for susceptible rats may thus be applied to this example . Both population subdivision alone and the presence of resistant rats may thus contribute to promote the persistence of plague in natural foci . However , the host population structure allows the persistence of the disease for a duration depending on the degree of spatial structure and on the features of the host and flea populations , while the presence of resistant hosts may allow a stable persistence of plague . In Madagascar , the plague focus is restricted to the central highlands . The focal persistence of plague may be explained by two different ( non mutually exclusive ) mechanisms , both of which will need to be validated through further field studies . First , the differential persistence of plague may be due to different parameters in highlands and lowlands , such that is above unity in the highlands and below unity in the lowlands . The few comparative studies that exist have not shown any major difference yet in biological parameter values related to rats between lowlands and highlands [3] , [31] ( J . -M . Duplantier , unpublished data ) . However , most of the parameter values linked to the fleas that we used have not been measured in Madagascar ( , , , ) . Some of these parameters ( , ) are among the ones which influence the basic reproductive number of the disease most . Climate is different in highlands and lowlands and may influence plague transmission by fleas [26] , [61] . Moreover , flea communities are different , as one of the flea species ( S . fonquerniei ) is only in central highlands . The two flea species may not have the same demographic and transmission characteristics , and S . fonquerniei could play a role in the endemism of plague by being responsible for a high transmission: it has been shown to carry more bacillus Y . pestis during the plague season than X . cheopis [62] . Thus , further studies would be needed to experimentally compare the two flea species . Even if both regions had above unity ( being or not equal ) , our results suggest that the persistence or extinction of the disease in each area may be explained by differences in the dynamics of the system , due to the presence/absence of resistant hosts and of population structure . In Madagascar , highlands were colonized by rats from Malagasy coastal populations some 800 years ago [63] , long before the introduction of plague in the island . As no resistant phenotype , even at low frequency , has been found in rats from coastal populations [34] , it seems likely that resistance evolved secondarily , after the spread of plague in highland rat populations . Population genetic studies showed that Malagasy rat populations are more genetically structured in landscapes characterised by sharp topographical relief , such as those found in some regions of the highlands , than in flat areas ( Brouat et al . , in revision ) . Rat population structure may thus have been more favourable to the persistence of the disease in the highlands than in coastal areas , and for periods of time sufficiently long to select resistance alleles . The evolution of resistance in the highlands may have in turn led to a more long-term plague persistence in this area . Host population structure and host resistance could thus have had a synergic effect to maintain plague in the Malagasy highlands . Testing this scenario would require more thorough theoretical studies based on the estimation of numerous biological parameters , especially in Malagasy flea populations ( see above ) . Also , this requires the examination of the time needed for a resistance allele to invade a metapopulation depending on spatial structure . However , it is interesting to note that the scenario of a secondary evolution of host resistance in a restricted geographical range has already been identified for other diseases , such as malaria in Hawaii , for which resistance has only been selected in low altitude [64] . Also , an empirical analysis of the genetic structure of Y . pestis among prairie dogs in Arizona [25] suggests a dispersion dynamic of plague consistent with the above scenario . This latter study highlights the existence of two stages in plague propagation: first a phase of rapid expansion , through the encounter of a highly dense susceptible rodent population , and then a phase of decline of the host population and of extinction of the disease , except if slow and stable transmission cycles can arise either through resistant hosts , or spatially structured or low density susceptible populations [25] . Using a simple model of plague propagation , we showed that both resistance and population subdivision may explain plague endemism . Madagascar may be a good illustration of how these two factors may act together , in synergy , to favour the long-term persistence of this highly virulent disease . However , further comparative field studies should aim at testing our assumptions on plague establishment in Madagascar , by trying to better assess the parameter values in the lowlands and highlands . It is worth noting that several aspects of the cycle of plague transmission have been neglected in our study . Some of these could play an additional role in Madagascar , others should not have any impact , and all of them could be involved in plague persistence in other foci . Indeed , the existence of multi-plague reservoirs [5] , [15] , [55] seems unlikely in Madagascar , as R . rattus is margely dominant in rural communities , representing at least 95% of the captures [3] , [65] . Also , although plague persistence in soils may exist in very peculiar situations ( e . g . , [66] ) or in steppic environments [16] , [17] , it has never been demonstrated in Madagascar [17] . Alternatively , the heterogeneity in the phenology of the host reproduction [15] , the direct transmission of Y . pestis inside burrows ( through for example the release of the bacillus as aerosols [18] , [19] ) , or the fact that resistant rats might be infectious for a short period of time before recovering ( not shown for rats but observed for mice , [67] ) or might release infectious fleas at their death could play additional roles in Madagascar and remain to be tested .
Bubonic plague , known to have marked human history by three deadly pandemics , is an infectious disease which mainly circulates in wild rodent populations and is transmitted by fleas . Although this disease can be quickly lethal to its host , it has persisted on long-term in many rodent populations around the world . The reasons for this persistence remain poorly known . Two mechanisms have been invoked , but not yet explicitly and independently tested: first , the spatial structure of rodent populations ( subdivision into several subpopulations ) and secondly , the presence of , not only plague-susceptible rodents , but also plague-resistant ones . To gain insight into the role of the above two factors in plague persistence , we analysed a mathematical model of plague propagation . We applied our analyses to the case of Madagascar , where plague has persisted on central highlands since the 1920s and is responsible for about 30% of the human cases worldwide . We found that the long-term persistence of plague can be explained by the presence of any of the above two factors . These results allowed us to propose scenarios to explain the localized presence of plague in the Malagasy highlands , and help understand the persistence of plague in many wild foci .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results", "Discussion" ]
[ "computer", "science", "evolutionary", "ecology", "computer", "modeling", "theoretical", "biology", "ecology", "population", "modeling", "epidemiology", "infectious", "disease", "epidemiology", "theoretical", "ecology", "population", "biology", "infectious", "disease", "modeling", "biology", "evolutionary", "biology", "computational", "biology" ]
2013
Host Resistance, Population Structure and the Long-Term Persistence of Bubonic Plague: Contributions of a Modelling Approach in the Malagasy Focus
Identifying the forces that drive proteins to misfold and aggregate , rather than to fold into their functional states , is fundamental to our understanding of living systems and to our ability to combat protein deposition disorders such as Alzheimer's disease and the spongiform encephalopathies . We report here the finding that the balance between hydrophobic and hydrogen bonding interactions is different for proteins in the processes of folding to their native states and misfolding to the alternative amyloid structures . We find that the minima of the protein free energy landscape for folding and misfolding tend to be respectively dominated by hydrophobic and by hydrogen bonding interactions . These results characterise the nature of the interactions that determine the competition between folding and misfolding of proteins by revealing that the stability of native proteins is primarily determined by hydrophobic interactions between side-chains , while the stability of amyloid fibrils depends more on backbone intermolecular hydrogen bonding interactions . Defining the rules of protein folding , a process by which a sequence of amino acids self-assembles into a specific functional conformation , is one of the great challenges in molecular biology [1]–[3] . In addition , deciphering the causes of misfolding , which can often result in the formation of -sheet rich aggregates , is crucial for understanding the molecular origin of highly debilitating conditions such as Alzheimer's and Parkinson's diseases and type II diabetes [4] . Major advances in establishing the interactions that drive the folding process have been made by analysing the structures in the Protein Data Bank ( PDB ) , and particularly by examining the frequency with which contacts between the different types of amino acid residues occur [5] . In this statistical approach , interaction free energies are derived from the probability , , of two amino acids of types and being in contact in a representative set of protein structures using the Boltzmann relation . This operation defines a matrix that lists the free energies of interaction between amino acid pairs . One of the most studied matrices of this type has been reported by Miyazawa and Jernigan [5] . Three distinct analyses of this matrix ( Fig . 1A ) have all revealed that residue-water interactions play a dominant role in protein folding [6]–[8] . More recently , the same statistical potential method has been used to investigate aggregation of soluble proteins into the amyloid state , now recognised as a generic , alternative , stable and highly organised type of protein structure [3] . A method for predicting the stability of amyloid structure ( PASTA ) [9] extracts the propensities ( ) of two residues found on neighbouring strands in parallel or antiparallel -sheets in a representative set of PDB structures . The resulting parallel strand and antiparallel strand interaction free energy matrices ( referred to here as “parallel” and “antiparallel” respectively ) are shown in Fig . 1B and 1C . Owing to the absence of a large number of solved atomic resolution amyloid fibril structures in the PDB , the central assumption of the PASTA approach is that the side-chain interactions found in the -sheets of globular proteins are the same as those stabilising -sheets in the core of amyloid fibrils [9] . This assumption is supported by the observation that the PASTA matrices are highly successful at predicting the portions of a polypeptide sequence that stabilise the core regions of experimentally determined amyloid fibrils and the intra-sheet registry of the -sheets [9] . We therefore treat the PASTA matrices as statistical potentials for the parallel and antiparallel -sheets found in the core of amyloid fibrils [9] . In this work we carry out a comparative analysis of the interaction matrices for folding and amyloid formation , in order to reveal the nature of the interactions that drive these two processes , and to provide fundamental insight into the competition between them . Our results indicate that the balance between hydrophobic and hydrogen bonding interactions is inverted in these two processes . The contact approximation for the effective Hamiltonian , , used to describe a system of polypeptide chains usually takes the form ( 1 ) where is the residue type at position along the polypeptide chain , is the position of residue and is a function reflecting the fact that two amino acids interact with free energy when they are in spatial proximity to each other [10] . For random heteropolymers , the pairwise contact free energies can be approximated as a set of 210 independent random variables ( i . e . the 210 independent elements in a symmetric matrix ) . For the MJ matrix , a plot with the axes running from hydrophobic ( C , F , L , W , V , I , M , Y , A , P , black ) [11] to hydrophilic ( H , G , N , T , S , R , Q , D , K , E , magenta ) [11] residue types reveals three large blocks of hydrophobic interactions ( Fig . 1A ) . The most stabilising interactions are hydrophobic-hydrophobic ( Fig . 1A , top left corner , blue ) , followed by hydrophobic-polar ( Fig . 1A , bottom left corner and top right corner , yellow/green ) and polar-polar interactions ( Fig . 1A , bottom right corner , red ) . On closer inspection , analysis of these interactions in the form of a histogram shows that the distribution of contact free energies determined from the Miyazawa-Jernigan ( MJ ) matrix ( Fig . 1D ) can be represented as the sum of three Gaussian terms corresponding to hydrophobic-hydrophobic ( H-H ) , hydrophobic-polar ( H-P ) and polar-polar ( P-P ) contacts [6] ( Fig . 1D ) . This interpretation implies that globular proteins are stabilised mainly by side-chain hydrophobic interactions [6] since the sum of all H-H , H-P and P-P contacts captures the overall distribution of contact free energies extremely well ( Fig . 1D ) . In contrast to the MJ matrix , contour maps of the parallel and antiparallel -sheet contact matrices of the type characteristic of amyloid fibrils [4] show highly destabilising contact free energies between all Pro-X pairs ( Fig . 1B , C , proline row , proline column , red/yellow ) . Since proline cannot form inter-molecular backbone hydrogen bonds this observation suggests that the stabilisation of -sheets arises mainly from the dominance of backbone hydrogen bonding , with hydrophobic interactions ( Fig . 1B , C , top left corner , blue ) playing a secondary role . Furthermore , plots showing the distribution of the contact free energies from parallel and antiparallel -sheets ( Fig . 1E , F ) of the type found in amyloid structures [4] indicate , unlike the situation for native folds described above , a single narrow Gaussian distribution for polar and non-polar contacts alike . This result , combined with the significance of the destabilising Pro-X contacts , is consistent with the view that a major role in protein aggregation into amyloid fibrils is played by backbone hydrogen bonding interactions [12]–[14] , which are “generic” [3] to any polypeptide chain , although sequence-dependent effects are also important to modulate the propensity of specific peptides and proteins [15]–[17] . The difference in these probability distributions arises because we are examining the contact free energies that define the protein folding and misfolding free energy minima via the MJ and PASTA matrices respectively . It is clear that the possible number of ways of forming a given contact between amino acids and is greater in globular proteins than in fibrillar aggregates as the area of Ramachandran space available to -sheets ( 13 . 3% of the total space ) is much smaller than that accessible to native proteins . In addition , the type of amino acid and specific sequence patterns have varying degrees of globularity [18] or aggregation propensity [16] with certain amino acids , notably proline , appearing much more frequently in globular proteins than in the core region of amyloid fibrils [9] . To investigate the consequences of these differences in the conformational spaces relevant to folding and misfolding we consider the constrained sampling of the protein Hamiltonian over a subspace of conformational space , which is given formally by ( 2 ) where is the partition function sampled over the subspace . Interaction parameters to describe the folding process are usually defined by considering a subspace that includes the regions of conformational space corresponding to the native states of globular proteins [19] . By contrast , interaction parameters to describe the aggregation process are defined for a subspace that includes only the regions of conformational space corresponding to -sheet rich structures such as -helices or amyloid fibrils [19] . While the Hamiltonian , , is invariant , the space over which it is integrated will vary depending on the region of conformational space that is being explored . In our case , this leads to distinct “effective” Hamiltonians for the protein folding and misfolding minima; these Hamiltonians have the same general form as Eq . [1] but have different amino acid interaction matrices , according to Eq . [2] , depending on which process is involved . We thus conclude that there could be differences in the various effective energy terms stabilising globular proteins and amyloid fibrils and that such differences can be described by giving different weights to hydrophobicity and hydrogen bonding interactions in the two states . In this view , hydrophobicity and hydrogen bonding do not represent fundamental interactions but effective ones , which result from constrained sampling procedures such as those defined by Eq . [2] . We decomposed the MJ and PASTA matrices into a combination of the HP ( Hydrophobic-Polar ) model [11] and a backbone hydrogen bonding model in which all amino acids , except for proline , are capable of forming backbone hydrogen bonds ( by analogy , we term this the HB model ) . These two-body interactions are described by three interaction matrices , , and , with the following properties: if and are both hydrophobic residues and topological neighbours , and otherwise; if either or is a hydrophobic residue , and are topological neighbours , and otherwise; if and can both form backbone hydrogen bonds and are topological neighbours , otherwise . As a first approximation , we initially fit the MJ and PASTA matrices to an equation of the form: ( 3 ) where is the matrix of interest , , and are the weightings of the , and matrices , respectively , and is a constant ( the solvent-solvent interaction parameter ) [8] . The normalisation constant shifts the elements of the MJ and PASTA matrices along the free energy axis thus allowing comparison of , and between different matrices . It is used to set the free energy of forming a polar-polar contact , , to zero and all other weightings are measured relative to this reference , i . e . and measure the additional free energy of forming hydrophobic contacts and the free energy gained through hydrogen bond formation . Importantly , the adjustment of to give a non-zero free energy has no effect on the ratios of to listed in Table 1 . The weightings ( Table 1 ) should be , and are , approximately equal to the free energy of a single hydrogen bond ( 2 . 5 [20] ) . This simple decomposition given by Eq . [3] gives very good agreement with the MJ ( correlation coefficient 0 . 87 ) and parallel matrices ( correlation coefficient 0 . 77 ) and good agreement with the antiparallel matrix ( correlation coefficient 0 . 69 , or 0 . 70 if disulfide bonds are taken into account ) . This coarse-grained HP-HB model is therefore a good approximation to the original matrices , and can thus provide insight into the relative importance of the hydrophobicity and hydrogen bonding terms for the different types of structures ( Table 1 ) . Since , and are all binary matrices , it is straightforward to quantify the marginal effect of each of the regressors in our general linear model from the values of their coefficients , and . For the MJ matrix , the ratio of to is ( Table 1 ) indicating that for protein folding the hydrophobic term is twice as important as the hydrogen bonding term . This ratio was corroborated by decomposing three recent pairwise contact potentials for the native states of globular proteins [21]–[23] which gave a similar result ( values are 0 . 4 [21] , 0 . 7 [22] , 0 . 73 [23] and on average ) . This finding is in agreement with previous work suggesting that the HP model captures the essence of protein folding [11] . Nevertheless , hydrogen bonding does play an important role in protein folding since highly polar sequences can fold to form -helices , and “side-chain only” molecular dynamics simulations fail to capture crucial aspects of protein folding [24] . Indeed , protein folding simulations have shown that it is necessary to include a mainchain-mainchain hydrogen bonding term in order to obtain secondary structure [25] . For protein misfolding and amyloid formation , the ratio of to for both PASTA matrices is ( Table 1 ) suggesting that backbone-only hydrogen bonding is about 50% more important in stabilising amyloid fibrils than hydrophobic interactions . To demonstrate the robustness of this result , we tested the sensitivity of the ratio to the Pro-X elements of the PASTA matrices and calculated that the high values of the Pro-X side-chain interaction free energies in the parallel and antiparallel matrices would have to be reduced by 4 or 5-fold respectively to achieve the same ratio of found in the MJ matrix . Given that the side-chain interaction free energies are derived from the Boltzmann relation , and that the high Pro-X interaction free energies reflect the infrequent occurrence of proline residues in -sheets , a reduction of this magnitude would translate into a much greater number of Pro-X contacts being detected in the -sheets of the PDB dataset used by the authors of PASTA [9] . The increased weighting of the matrix relative to the matrix in the decomposition of the PASTA matrices shows that the destabilising effect of proline is more disruptive to the hydrogen bonded -sheet structure than to the native fold of globular proteins in which proline has evolved to play an important structural , and stabilising , role e . g . in Pro-induced -turns [26] . This result underscores the importance of sequence-independent hydrogen bonding in defining the amyloid structure . This “generic” view [12] is consistent with the observation that even hydrophilic and homopolymeric sequences of amino acids can form amyloid fibrils [13] . However , the amino acid sequences of individual peptides and proteins influence their specific propensity to aggregate [16] , [17] , and to form self-complementary side-chain packing interfaces between adjacent -sheets in the fibrils [15] , [27] , [28] . We also note that in the -sheets of globular proteins , the effects of backbone hydrogen bonding tends to be averaged out in Eq . ( 2 ) by the presence of other secondary structure motifs ( -helices , -turns and coil ) . A number of controls were performed to confirm that the ratio of to is inverted between folded globular proteins and amyloid fibrils . Firstly , the value of is only slightly affected by considering amino acids such as Proline and Alanine to be hydrophilic rather than hydrophobic . In our initial classification of hydrophobic and hydrophilic residues [11] , the ratios between the hydrogen bonding and hydrophobic terms , , are 0 . 48 , 1 . 59 and 1 . 39 for the MJ , parallel and antiparallel PASTA matrices respectively ( Table 1 ) . By considering proline residues to be hydrophilic , rather than hydrophobic , the ratios become 0 . 55 , 1 . 78 and 1 . 66 for the MJ , parallel and antiparallel PASTA matrices respectively . Furthermore , if we adopt the partitioning suggested by Li , et al . [6] in which both proline and alanine residues are considered to be hydrophilic rather than hydrophobic , the ratios become 0 . 61 , 2 . 14 and 2 . 27 for the MJ , parallel and antiparallel PASTA matrices respectively . This analysis shows that the ratio is inverted between the MJ and PASTA matrices using the most common classifications of amino acids into hydrophilic and hydrophobic sets . We also note that the MJ matrix is calculated by using the quasi-chemical approximation in which protein residues are assumed to be in equilibrium with the solvent . By considering water to be the reference state , all residue-residue interactions are attractive and so all elements of the MJ matrix are negative . By ignoring chain connectivity , it has been argued that this “connectivity effect” introduces a bias into the MJ matrix . However , a knowledge-based pair potential for describing amino acid interactions in the native folds of globular proteins developed by Skolnick , et al . [21] , which we refer to as the SJKG matrix , explicitly includes effects due to chain connectivity . Skolnick , et al . [21] conclude that ignoring chain connectivity does not introduce errors and that the quasi-chemical approximation is sufficient for extracting statistical potentials such as the MJ matrix . By virtue of using native reference states , the SJKG matrix has both positive and negative side-chain interaction free energies and is similar in this way to the PASTA matrices ( Fig . 1B , C ) . The SJKG matrix also has a mean free energy of approximately zero ( 0 . 08 ) like the PASTA matrices ( 0 . 51 and 0 . 13 for parallel and antiparallel respectively , Fig . 1B , C ) . However , like the MJ matrix , the SJKG is a statistical potential for the native folds of globular proteins and when we decompose this matrix using the HP-HB model we get a ratio of to of 0 . 4 , which is almost identical to the ratio found for the MJ matrix . Thus , this result strengthens our findings as the hydrophobicity term , , is even more dominant than the hydrogen bonding term , , in the decomposition of the SJKG matrix than in the MJ matrix ( ratios of 2 . 50 and 2 . 08 respectively ) . In addition , the comparison of the value of the normalisation constant ( 0 . 94 ) with the values of the and terms ( 0 . 49 and 1 . 24 , respectively ) in the HP-HB decomposition of the SJKG matrix confirms that the value of does not affect the ratio of for native proteins and that this ratio is reversed between folded globular proteins and amyloid fibrils . From the contour maps ( Fig . 1A , B , C ) and the histograms of contact free energies ( Fig . 1D , E , F ) it is clear that the free energy of forming hydrophobic-polar ( H-P ) side-chain contacts is stabilising for globular proteins although not nearly as important in the simple formation of -sheets . Thus , for protein folding we find that where is the free energy of forming a polar-polar contact and is not stabilising ( ) and and are the free energies of forming hydrophobic-polar contacts and hydrophobic-hydrophobic contacts respectively . These weightings are in excellent agreement with a modified form of the HP model [29] ( in the present study compared to 2 . 3 in the modified HP model [29] ) and so validate its use in protein folding simulations . The inclusion of the HP term in Eq . [3] has only a marginal effect on the regression to the parallel or antiparallel matrices as demonstrated by the relatively small coefficient 0 . 2 ( Table 1 ) . This result suggests that the segregation of hydrophobic and polar residues is not very important in -sheet formation and could lead to solvent exposed non-polar side-chains in prefibrillar aggregates , a feature that has been suggested to be closely linked to cytotoxicity [30] . The minor effect of the HP term is also in accord with our finding that hydrophobic interactions play a less significant role than inter-molecular hydrogen bonding in stabilising amyloid fibrils and again supports the idea that peptides and proteins are prone to forming amyloid structures irrespective of sequence [12] , [13] , although the relative propensities to form such structures will vary with sequence [16] , [27] . Previous analyses of the MJ matrix shows that two-body interactions are not sufficient to capture all of the details of the 210 independent amino acid interactions that describe the variety of native protein structures [6]–[8] . A one-body term , , describing the individual properties of each amino acid , is also required . Adding this additional term to our previous free energy expression Eq . [3] gives ( 4 ) The application of this equation to the MJ , parallel and antiparallel matrices gives correlation coefficients of 0 . 99 , 0 . 90 and 0 . 90 respectively ( Fig . 2A , B , C ) . This expression , therefore , describes the original data extremely well and suggests that the diverse and complex interactions stabilising both the native and fibrillar states are amenable to a low-dimensional representation using simple two-body and one-body terms [6]–[8] . It is remarkable that the same approach can be used to decompose both the MJ and PASTA matrices , indicating that the underlying interactions are the same but that the balance is different , and leads to a clear demarcation of the thermodynamic minima of the native and amyloid states of the protein free energy landscape . The three sets of 20 one-body parameters , , that are derived from the MJ , parallel and antiparallel matrices are listed in Table 2 . Previous work has shown that one-body components of the MJ matrix , known as q-values , are closely related to the interactions governing secondary structure formation [6] . We find that our equivalent one-body potentials , MJ ( Table 2 ) , correlate extremely well with ( correlation coefficient of 0 . 98 , Fig . 3A ) , and are numerically almost identical to this previously published q-scale ( Table 2 , column 4 ) provided that the hydrophobic and hydrophilic q-values are separated and have their respective mean values subtracted from each non-polar and polar element . This procedure removes an average hydrophobic penalty for non-polar residues ( +1 . 45 ) and an average hydrophilic gain for polar residues ( −0 . 07 ) . This residue-specific hydrophobic ( hydrophilic ) cost ( gain ) can be interpreted as an average free energy cost of placing in water the surface of a given residue plus the gain of attractive dipolar interaction between the residue concerned and water , with polar residues being more favourable than non-polar residues [7] . This effect is even more apparent in the simpler case of the one-body components of the parallel and antiparallel PASTA matrices ( Table 2 , parallel and antiparallel respectively ) . When existing parallel and antiparallel -sheet propensity scales [31] are converted into free energies ( Table 2 , column 5 and 6 respectively ) , grouped into polar and non-polar terms and then separately shifted to have zero mean , thus removing the average hydrophobic ( hydrophilic ) cost ( gain ) to water of forming a sheet ( the values are +0 . 32 ( −0 . 51 ) and +0 . 34 ( −0 . 25 ) for parallel and antiparallel -sheets respectively ) , the remainder correlates extremely well with ( correlation coefficients of 0 . 96 and 0 . 97 for parallel and antiparallel -sheets respectively , Fig . 3B , C ) , and is numerically almost identical to the one-body potentials of the parallel and antiparallel matrices ( parallel and antiparallel respectively , Table 2 ) . This result suggests that the one-body free energy components of the MJ , parallel and antiparallel matrices are given by ( 5 ) where represents the free energy to form hydrogen bonded secondary structure and is an average free energy of solvation . Hence , we suggest that the one-body free energy terms , , correspond to a stabilisation of the native or fibrillar state through a competition between hydrophilicity and the formation of hydrogen bonded secondary structure . The HP-HB-SS ( HP-HB-secondary structure ) model described above suggests therefore that both the globular and amyloid states of proteins are stabilised by hydrophobic interactions , hydrogen bonding and the formation of secondary structure , and that there is a common form for the effective Hamiltonian , , describing both protein folding and misfolding , given by the substitution of Eq . [4] into Eq . [1] ( 6 ) The two-body terms in the effective Hamiltonian are , and , which correspond to the relative strengths of hydrophobic interactions and hydrogen bonding , and take the values given in Table 1 . The effective energy function is further modulated by the additive residue specific terms ( Table 2 ) , which correspond to the free energy of secondary structure formation plus a free energy of solvation . It is important to note that there is a loss of translational and rotational entropy on going from native to fibrillar states [32] which we do not consider here . This loss of entropy would be expected to stabilise the native state in a sequence- and conformation-independent manner and would add a native-biasing term to the effective energy function given in Eq . [6] . Although the general form of the effective Hamiltonian is the same for protein folding and misfolding , the variables , , and are different for these two processes , with the result that the minima in the two cases will occur at different positions in conformational space . Fibrillar aggregates represent a well-defined region of the wider protein folding landscape characterised by the pervasiveness of generic intermolecular hydrogen bonding [12] . Since the Hamiltonian maps the sequence space on to the structure space , as the weights , and change so too does the shape of the resulting structure . The dominance of the collapse-inducing hydrophobic force in protein folding leads to a globular tertiary structure , with hydrophobic residues buried in the core and largely polar residues on the surface of the protein [33] . However , when unidirectional inter-molecular hydrogen bonding is in the ascendancy , the result is ordered protein self-association into elongated , rigid , rod-like aggregates [14] . By decomposing the MJ and PASTA matrices into two-body and one-body components , we have effectively decoupled the two-body non-local interactions from the one-body , local interactions entangled in these statistical potentials . This approach enables us to analyse quantitatively the relative importance of local and non-local interactions in determining the folding and misfolding of proteins . It is clear from Tables 1 and 2 that the magnitude of the non-local ( tertiary ) interactions are significantly greater than the local ( secondary ) interactions in stabilising the native protein or fibrillar aggregate . This result indicates that nonlocal inter-residue interactions are the major determinant of secondary structure in the HP-HB-SS model . This finding is in excellent agreement with a large body of experimental [34] and computational analyses [35] , which demonstrates that the sequence patterns of polar and non-polar amino acids dominate their intrinsic secondary structure propensities in determining the secondary structure motifs of a globular protein [36] or amyloid fibril [37] . Our prediction that hydrophobic patterning and sequence independent hydrogen bonding is more important than residue-specific identity in shaping secondary and tertiary structure helps explain why a wide variety of amino acid sequences can encode the same basic protein fold [38] . It is also consistent with the mutational robustness of functional proteins , which typically only fail to fold correctly following several mutations of individual amino acids [39] . In addition , globular proteins have evolved to mitigate against the non-local effect of polar/nonpolar periodicity by deliberately spurning alternating hydrophobic patterns which program amino acid sequences to form amphiphilic -sheets and amyloid fibrils [40] . This is further evidence that tertiary interactions overwhelm the intrinsic propensities of individual amino acids in real proteins , which agrees with our analysis . The mathematical form of the effective Hamiltonian of Eq . [6] describing protein folding and misfolding is analogous to that of a spin glass model in which competition between conflicting interactions leads to a rugged free energy landscape [41] . Apart from topological frustration , which arises due to chain connectivity , the three sources of energetic frustration in the HP-HB-SS model stem from the competition between intramolecular collapse and intermolecular self-association , the contest between frustrating nonlocal interactions and , finally , the inability to satisfy simultaneously all local secondary structure preferences . As discussed earlier , in our model the relative strengths of the hydrophobicity to hydrogen bonding terms governs the dichotomy between folding and misfolding ( Table 1 ) . The conflicting optimisation factors imposed by hydrophobic clustering , maximal backbone hydrogen bonding and the segregation of hydrophobic and polar residues prevent the native state or fibrillar aggregate from energetically satisfying all of these inter-residue interactions . Finally , since non-local interactions predominantly determine globular [36] and fibrillar protein structures [37] , there is an additional source of mismatch between the secondary structure motifs encoded by the hydrophobic patterning of the amino acid sequence as a whole and the secondary structure propensities of the individual amino acids . This intricate interplay of competing interactions gives rise to multiple local minima in the effective energy function of Eq . [6] but , in accordance with the principle of minimal frustration [2] , the sequence of a protein has evolved to reduce the number of alternative minima as much as possible and to have its native state as the global minimum of the protein folding free energy landscape [2] , [3] . However , the ruggedness of the folding free energy landscape increases the likelihood that excited native-like states exist , which may be transiently populated via thermal fluctuations , thus potentially leading to amyloid formation even under physiological conditions [42] . Moreover , frustration in the protein misfolding free energy landscape can lead to amyloid fibril polymorphs with different physical and biological properties [43] . Lowering the discordance between non-local ( Table 1 ) and local ( Table 2 ) interactions leads to more stable and cooperative native protein folds [35] , [44] , and has implications for the de novo design of proteins [44] and amyloid fibrils [45] , [46] . Indeed , knowledge of the residue-specific one-body terms ( Table 2 ) , and the understanding that they correspond to the free energy of secondary structure formation once a solvation free energy is taken into account , may aid in the rational design of globular folds through mutational screening of regions known to be critical for aggregation . The present work indicates that there are common intermolecular forces stabilizing both globular and fibrillar states of proteins , but that a different balance of these forces results in either folding or misfolding to non-functional and potentially toxic aggregates . This situation occurs as the competing processes of protein folding and misfolding are finely tuned in terms of their free energies . Upon folding , the protein minimises the free energy of the protein-water system by clustering hydrophobic groups and forming intramolecular hydrogen bonds in the globular interior . By contrast , upon aggregation into amyloid fibrils , the formation of an extensive intermolecular hydrogen bonding network compensates for any exposure of hydrophobic groups to water that results from the fibrillar structure of the aggregated state . It has been found in molecular dynamics simulations that the correct balance between hydrophobicity and hydrogen bonding must be attained for proteins to fold correctly or to self-assemble into the alternative well-defined amyloid structure rather than into amorphous aggregates [19] , [47] . For example , if hydrophobicity is too dominant , then an amorphous cluster of residues with few native contacts can be formed rather than a correctly folded protein [19] . Interestingly , these simulations suggest that hydrogen bonding is more than twice as important as hydrophobicity for aggregation into amyloid fibrils [19] , [48] , and that hydrophobicity is approximately twice as important as hydrogen bonding for protein folding [19] , findings that are in close agreement with those reported by the analysis in the present paper . Recent experimental evidence supports this interpretation of protein folding and misfolding . It has been found that the substitution of backbone ester groups for the amide linkage does not significantly affect the structure of native proteins [49] , suggesting that the folded core is mainly stabilised by hydrophobic interactions . Similar experiments for protein aggregation , however , reveal that peptides with removed backbone amide groups have a much reduced propensity to form ordered aggregates [50]; indeed such species are being explored as potential therapeutic inhibitors of amyloid fibril growth [51] . In addition , the large elastic modulus of amyloid fibrils stems mainly from generic inter-backbone hydrogen bonding indicating that this is a dominant interaction defining the amyloid state [14] . The weights , and are functions of physical [52] , [53] and chemical [54]–[56] parameters . Hydrophobic attraction , , and hydrogen bond interaction strength , , are both strongly environment-dependent intermolecular forces and vary in a complex manner as externally driven parameters such as temperature , pH , ionic strength and denaturant concentration are changed [32] . Despite the complicated nature of these interactions , experiments show that at low concentration , denaturants increase the monomer-monomer dissociation energy approximately linearly [54] . This suggests that the monomer-monomer association energy is a linear decreasing function of denaturant concentration under mildly denaturing conditions . In keeping with our model , we speculate that at low denaturant concentrations , is large , thereby promoting the native state by increasing residue-residue hydrophobic attraction , whereas at higher denaturant concentrations the lowering of leads to destabilisation of the hydrophobic core of the native structure , making intermolecular association much more likely [57] . Our analysis suggests that there is an optimal balance between hydrophobicity and hydrogen bonding for protein folding and a significant redistribution of these intermolecular forces for amyloid formation . Such a shift in balance can be seen as a jump between free energy landscape minima , and could occur , for example , at a critical concentration [58] , or pH [55] , or at a temperature sufficiently high to overcome kinetic barriers between the native and amyloid minima [46] . Overall , however , this balance appears to be very finely tuned for both protein folding and misfolding , and it is interesting to speculate on the role of this delicate balance of forces within the cell . It has been suggested that proteins have evolved to be expressed intra-cellularly at levels in the region of the critical concentration for aggregation [58] . While a plentiful abundance of a given protein in the cell optimises its function , being on the verge of insolubility leaves proteins susceptible to environmental changes and prone to aggregation [59] . Our findings are consistent with this hypothesis [58] , since elevated protein levels increase the likelihood of intermolecular as opposed to intramolecular interactions , and suggest that a precarious balance between hydrogen bonding and hydrophobic forces dictates whether peptides and proteins adopt normal or aberrant biological roles . In conclusion , we have reported an interpretation of statistical potentials for protein folding [5] and misfolding [9] by expressing them in terms of a model containing specific terms for hydrogen bonding and hydrophobicity . This approach has enabled us to describe complex and diverse interactions using specific values of three distinct two-body terms and intrinsic secondary structure propensities . We have explained the significance of each of these terms and derived a physically meaningful common form of effective Hamiltonian for both protein folding and amyloid formation . This approach suggests that while hydrophobicity , hydrogen bonding and the formation of secondary structure are important to both processes , the balance between hydrophobicity and hydrogen bonding is remarkably different in the two regimes . Our central finding is that the stabilities of correctly folded proteins are dominated by side-chain hydrophobic interactions and that amyloid fibrils are stabilised mainly by sequence-independent intermolecular hydrogen bonding . We have also quantified the relative importance of local and non-local interactions in determining the structure and stability of proteins in both their globular and fibrillar forms and find that inter-residue interactions are more influential than secondary structure propensities in shaping the final native or amyloid fold . This result shows that , in accordance with the principle of minimal frustration [2] , natural proteins have evolved to maintain a low ratio of local-to-non-local interaction strengths , thereby minimising the effect of a potent source of frustration and ensuring cooperative and stable folding [35] , [44] . In summary , we have found that the conflict between protein folding and misfolding is governed by the contest between a side-chain-driven hydrophobic collapse and a backbone-driven self-association . The almost infinite variety of outcomes of such a conflict gives rise to the rich and diverse behaviour exhibited by proteins and the resulting balance between health and disease . The weights of the two-body terms , , , , and the constant , , were determined by performing multiple regression in MATLAB . The twenty one-body terms , , were determined by performing a simulated annealing minimisation in MATLAB .
In order to carry out their biological functions , most proteins fold into well-defined conformations known as native states . Failure to fold , or to remain folded correctly , may result in misfolding and aggregation , which are processes associated with a wide range of highly debilitating , and so far incurable , human conditions that include Alzheimer's and Parkinson's diseases and type II diabetes . In our work we investigate the nature of the fundamental interactions that are responsible for the folding and misfolding behaviour of proteins , finding that interactions between protein side-chains play a major role in stabilising native states , whilst backbone hydrogen bonding interactions are key in determining the stability of amyloid fibrils .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[ "biophysic", "al", "simulations", "biology", "computational", "biology" ]
2011
Inversion of the Balance between Hydrophobic and Hydrogen Bonding Interactions in Protein Folding and Aggregation
Allosteric interactions in the Hsp70 proteins are linked with their regulatory mechanisms and cellular functions . Despite significant progress in structural and functional characterization of the Hsp70 proteins fundamental questions concerning modularity of the allosteric interaction networks and hierarchy of signaling pathways in the Hsp70 chaperones remained largely unexplored and poorly understood . In this work , we proposed an integrated computational strategy that combined atomistic and coarse-grained simulations with coevolutionary analysis and network modeling of the residue interactions . A novel aspect of this work is the incorporation of dynamic residue correlations and coevolutionary residue dependencies in the construction of allosteric interaction networks and signaling pathways . We found that functional sites involved in allosteric regulation of Hsp70 may be characterized by structural stability , proximity to global hinge centers and local structural environment that is enriched by highly coevolving flexible residues . These specific characteristics may be necessary for regulation of allosteric structural transitions and could distinguish regulatory sites from nonfunctional conserved residues . The observed confluence of dynamics correlations and coevolutionary residue couplings with global networking features may determine modular organization of allosteric interactions and dictate localization of key mediating sites . Community analysis of the residue interaction networks revealed that concerted rearrangements of local interacting modules at the inter-domain interface may be responsible for global structural changes and a population shift in the DnaK chaperone . The inter-domain communities in the Hsp70 structures harbor the majority of regulatory residues involved in allosteric signaling , suggesting that these sites could be integral to the network organization and coordination of structural changes . Using a network-based formalism of allostery , we introduced a community-hopping model of allosteric communication . Atomistic reconstruction of signaling pathways in the DnaK structures captured a direction-specific mechanism and molecular details of signal transmission that are fully consistent with the mutagenesis experiments . The results of our study reconciled structural and functional experiments from a network-centric perspective by showing that global properties of the residue interaction networks and coevolutionary signatures may be linked with specificity and diversity of allosteric regulation mechanisms . The evolutionary conserved and versatile 70-kilodalton ( kDa ) heat shock proteins Hsp70s play a central role in supervision of various protein folding processes and coordination of a broad range of cellular events–from maintenance of cellular homeostasis to regulation of the heat shock response [1–8] . The Hsp70 proteins cooperate with chaperones of other families to facilitate protein folding , prevent aggregation , and ensure stabilization and quality control of native regulatory proteins [9–12] . The Hsp70 biochemical cycle involves a precisely orchestrated succession of nucleotide-induced allosteric structural changes that are executed through complex and adaptive interactions with co-chaperones , particularly J-domain proteins DnaJ and Hsp40 which accelerate ATP hydrolysis , and nucleotide exchange factors assisting in a timely progression of the ATP-ADP exchange [13–16] . During this cycle , ATP binding in the nucleotide-binding domain ( NBD ) accelerates substrate dissociation from the substrate-binding domain ( SBD ) , while substrate binding synchronously stimulates ATPase hydrolysis in the NBD ( Fig 1 ) . Structural and biophysical investigations of various E . coli Hsp70 ( DnaK ) constructs [17–26] have established that the NBD and SBD are only loosely coupled in the extended ADP-bound and the nucleotide-free states , whereas ATP binding shifts thermodynamic preferences and stabilizes a compact domain-docked DnaK structure , leading to stimulation of the ATPase activity and substrate release ( Fig 1 ) . The solution NMR structures of the ADP-DnaK and apo DnaK ( Figs 1 and 2 ) have confirmed that the NBD and SBD could only weakly associate in the domain-undocked state of the chaperone [27] . Chemical-shift perturbation analysis of the DnaK states has provided evidence of ATP-induced rotational movements of the NBD subdomains that promote binding of the inter-domain linker and stabilization of the NBD-SBD interface [28] . The crystal structures of the ATP-bound DnaK [29 , 30] have revealed a synchronous docking of the SBD-β and SBD-α subdomains to the NBD in the ATP-DnaK , that causes accelerated substrate dissociation ( Figs 1 and 2 ) . The early biochemical studies [31] and subsequent electron spectroscopy experiments [32–35] have established the existence of multiple conformational states in the Hsp70 proteins , particularly showing that nucleotide exchange could promote formation of partially undocked meta-stable intermediates . Recent NMR studies have discovered that dynamic changes in the inter-domain and substrate binding regions are coupled and may coordinate ATP hydrolysis and substrate release via an entropy-driven allosteric mechanism [36] . While the ADP-DnaK structure with high substrate affinity may be stabilized by both enthalpy and entropy contributions , the thermodynamics of the ATP-bound DnaK with low substrate affinity may be mainly driven by entropy changes [37] . The crystallographic and NMR studies of the yeast Hsp110 ( Sse1 ) [38–41] have revealed that the ATP-bound state can adopt a similar docked conformation , but ATP hydrolysis in Sse1 may proceed without triggering any appreciable conformational changes ( Fig 2 ) . Consequently , a limited entropy-driven allostery could present a dominant “modus operandi” of the Sse1 biochemical cycle , where ATP hydrolysis is coupled to the substrate release via redistribution of conformational dynamics in the functional regions . Mutagenesis studies have quantified the contributions of functional residues to allosteric signaling and inter-domain communications of DnaK [42–48] . Several mutational variants of the NBD residues could maintain ATP binding and hydrolysis functions although they are severely deficient in allosteric signaling: Y145A , N147A , and D148A [42] , P143G and R151A [43] , K155D , R167D [44] and D326V [47] . Important functional sites were also found in the SBD regions , where mutations K414I [45] , P419A [46] , and N415G [47] may completely abolish or significantly weaken allosteric interactions in the DnaK chaperone . Mutations in the inter-domain linker region of DnaK could shut down functional activity of the DnaK-DnaJ chaperone systems [48] . Mutagenesis and functional experiments of DnaK have determined the role of the SBD-α lid helices and the hydrophobic regions of the SBD-β in coordinating substrate binding affinity and kinetics of substrate dissociation [49–52] . Recent breakthrough studies have discovered the previously overlooked divergence of allosteric signaling pathways in the DnaK chaperone by convincingly demonstrating how mutations of regulatory sites ( I438A , V440A , L484A , D481A , D481K ) could selectively interfere with direction-dependent steps of allosteric communication [53] . According to this seminal work , functional cycle in DnaK may be allosterically controlled through concerted rearrangements of the specific inter-domain interactions that couple nucleotide exchange with substrate binding and release . Mechanisms of allosteric signaling are ultimately determined by the thermodynamics of a system that can be described using ensemble-based computational models of allosteric interactions [54–56] . Computational studies have investigated various molecular factors underlying allosteric regulatory mechanisms in the DnaK chaperones by simulating dynamics and energetics of the crystal structures and allosteric pathways [57–68] . Molecular dynamics ( MD ) simulations and elastic network models have explored evolution and dynamics of the Hsp70 chaperones in binding with client proteins [57] and molecular aspects of nucleotide-induced conformational transitions in these chaperones [58] . The diversity of conformational sates observed in biophysical experiments of the human Hsp70 has been successfully reproduced in atomistic simulations [60] . The free energy landscape mapping of the DnaK structures has also suggested several mediating residues that may be instrumental in signal propagation between the NBD and SBD regions [61] . Biophysical simulations and mutagenesis experiments have characterized several hinge residues controlling the nucleotide-dependent allostery in DnaK [62] . Functional assays combined with perturbation response scanning analysis have identified a group of regulatory residues in subdomain IA of the NBD that promote allosteric interactions and inter-domain signal transmission [64] . According to this study , while allosteric coupling in Hsp70 proteins could be maintained through clusters of conserved interactions , binding to co-chaperones may be facilitated by coevolving flexible residues in the subdomain IIA . Coevolutionary analysis has allowed to capture large-scale conformational transformations of the Hsp70 chaperones and predicted functional dimer interactions between Hsp70 proteins [66] . Computational modeling of the residue interactions in combination with in silico alanine scanning of the Hsp70 residues has probed molecular determinants and specific role of functional sites in allosteric signaling and biochemical cycle [67] . MD simulations have elucidated the molecular determinants underlying ligand-induced modulation of conformational dynamics in the DnaK structures , showing that local dynamics changes in response to ligand binding may be coupled to allosteric structural rearrangements [68] . The relationships between protein dynamics and allosteric signaling can be conveniently explored using structural analysis of the residue interaction networks [69–71] . This approach can successfully identify functional residues [72 , 73] , describe ligand-induced shifts in conformational populations of allosteric proteins [74–78] and reconstruct allosteric communication pathways [79] . MD simulations and network modeling have been combined to explore conformational ensembles of the Hsp70 chaperones [80] . In our most recent investigation , dynamics-based network modeling and community detection approaches have joined forces in probing mechanisms of allosteric inhibition in the DnaK and human Hsp70 proteins [81] . According to this study , functional effects of allosteric modulators may be linked with the inhibition of specific interaction networks that alter structural environment of the regulatory sites . Despite advances in the experimental and computational studies of the Hsp70 mechanisms , the role and contribution of functional residues and specific interactions implicated in allosteric regulation are yet to be fully understood and properly quantified . Moreover , the outstanding questions concerning modular organization of the allosteric interaction networks and hierarchy of interactions that control functional cycle of the Hsp70 chaperones remained largely unexplored . Current understanding of allosteric communication pathways in the Hsp70 chaperones remains mostly mechanistic , lacking a proper physical foundation and atomistic insights that are required to rationalize latest experimental data on direction-specific pathways of allosteric regulation in DnaK [53] . In this work , we employed a computational framework that combined atomistic and coarse-grained simulations of the Hsp70 structures with coevolutionary analysis and network-centric modeling . The goal of this study was to elucidate in a systematic manner dynamic and evolutionary factors underlying allosteric structural transformations of the Hsp70 proteins . A novel methodological aspect of this work was the incorporation of dynamic residue correlations and coevolutionary residue dependencies in the construction and analysis of allosteric interaction networks . Statistical coupling analysis [82–84] , mutual information ( MI ) model [85–87] and other evolutionary approaches [88–91] have shown that functionally significant residues can be connected via coevolutionary relationships . Recent studies have revealed the important role of coevolving residues in mediating residue-residue contacts [92] , protein folding [93] and protein-protein recognition [94] . Coevolving residues are often close to each other in the protein structure [95 , 96] and may form independent structural modules ( or protein sectors ) with distinct biochemical functions [97 , 98] . Networks of residues with high mutual information can also characterize structural proximity of functionally important sites [99 , 100] . In the current study , we investigate the relationships between conformational dynamics , coevolutionary residue associations and hierarchy of allosteric interactions in the Hsp70 proteins . By performing a community decomposition of residue interactions , we find that regulatory sites can be distinguished by their high network centrality and integrating role in modular organization of allosteric interactions . We show that the inter-domain communities may be coordinated by key functional centers that are surrounded by coevolving flexible residues in order to facilitate conformational transitions . Cooperative rearrangements in these communities can adequately describe allosteric changes and population shifts in conformational ensembles of the Hsp70 proteins . By using allosteric residue propensities , we introduce a community-hopping model of communication pathways that explained the asymmetry of allosteric control in the DnaK chaperone . The results of this study reconcile a wide spectrum of functional experiments by providing a network-centric outlook of the conformational dynamics , coevolution and interaction networks in the Hsp70 proteins . We argue that these factors may act as synchronizing forces that shape up the efficiency and robustness of allosteric regulatory mechanisms in these chaperones . We began by investigating the relationships between conformational dynamics , sequence conservation and residue coevolution in the Hsp70 protein family . MD simulations were independently performed for full-length two-domain Hsp70 structures: DnaK_GEOKA ( pdb is 2V7Y ) DNAK_ECOLI ( pdb id 2KHO , 4B9Q , 4JNE , 4JN4 ) , HS7C_BOVIN ( pdb id 1YUW , 4FL9 ) , and HSP7F_YEAST ( Hsp homolog Sse1 ) proteins ( pdb id 2QXL , 3C7N , 3D2E , 3D2F ) . ( Fig 2 ) . We first analyzed evolutionary factors underlying allosteric regulation of the DnaK chaperone . By analyzing the sequence conservation profile ( Fig 3A ) we identified evolutionary features that may differentiate regulatory sites of the DnaK chaperone from nonfunctional conserved residues . The highly conserved residues were primarily assembled in the core regions of subdomains IA , IIA , IIB and the SBD-β subdomain . Functional residues involved in the nucleotide binding and allosteric regulation were highly conserved ( K70 , R71 , P143 , Y145 , F146 , R151 , E171 , D393 , and V396 ) , while several other regulatory sites ( D148 , K155 , R167 , I168 , K414 , D481 ) could experience small conservative modifications during evolution . Using MISTIC approach [99 , 100] we also characterized the extent of mutual information and coevolutionary couplings between residue pairs in the Hsp70 proteins . Coevolution of protein residues may emerge from different factors including phylogeny-driven preferences for compensatory substitutions and structural constraints imposed by protein stability , adaptability to binding partners and regulatory functions . In particular , coevolutionary signals of highly conserved residues in the protein core that undergo a small number of independent changes can be often overestimated , while correlated changes arising from molecular coevolution may be inadvertently suppressed [101–103] . To discriminate coevolutionary associations driven by functional constraints from those determined by common ancestry , the covariance metric based on MI calculations was adjusted by the average product correction ( APC ) [104–106] . Based on computations of coevolutionary residue matrices , we assembled a network of coevolutionary coupled residues in which the links between nodes ( residues ) represented mutual information shared by the respective residues . We then estimated cumulative mutual information ( cMI ) and ensemble-based proximity mutual information ( pMI ) residue profiles of the DnaK structures . A considerable spread in the cMI scores of functional residues was observed , particularly in the NBD subdomain IA and SBD-β regions , suggesting that cumulative coevolutionary score may not be a strong differentiating feature of the regulatory centers ( Fig 3B ) . Functional residues involved in DnaK allostery could be much better distinguished by their high pMI values ( Fig 3C and 3D ) , revealing strong propensities of regulatory sites to be surrounded by clusters of highly coevolving residues in the protein structure . Of particular interest were high pMI scores of key residues that communicate signal of the ATPase hydrolysis from the NBD region ( K70 , R71 , Y145 , F146 , R151 , R167 , I168 , E171 ) , to the inter-domain interface ( K414 , D481 ) , and the SBD-β allosteric hotspot ( L454 , L484 ) [36 , 53] . Conformational dynamics of the DnaK chaperone revealed several important trends and determined specific groups of residues that can be distinguished by their unique dynamic and coevolutionary signatures ( Fig 4 ) . Structurally stable regions in the ADP-DnaK and ATP-DnaK forms included residue segments 69–73 and 140–171 from the subdomain IA that featured catalytic residues K70 , E171 and conserved regulatory residues P143 , F146 , D148 , R151 , K155 , R167 , and I168 ( Fig 4A and 4B ) . To examine collective dynamics and identify hinge sites of global motions in the DnaK structures , we supplemented MD simulations by coarse-grained Gaussian network ( GNM ) analysis [107–109] . The GNM-based residue fluctuations averaged over the three lowest frequency modes were analyzed to characterize functional dynamics of the DnaK structures . The local maxima in these profiles correspond to flexible regions undergoing global structural changes , while the local minima are typically attributed to structurally rigid sites that serve as hinge points that coordinate collective dynamics . In the ADP-DnaK , the major hinge site ( D385 , V386 ) was located in the region connecting the subdomain IA of the NBD with the inter-domain linker ( Fig 4C ) . Among other local minima were residues 199–202 ( the inter-domain interface between IA and IIA subdomains ) and regulatory sites P419 and D481 ( SBD-β ) that bridge the inter-domain linker with the SBD-β subdomain . The slow mode profile of the ATP-DnaK structure showed that binding site residues ( K70 , R71 ) and regulatory sites ( P143 , Y145 , F146 , R151 , K155 , R167 , I168 , T221 , P419 , I438 , V440 , and L454 ) were largely immobilized in collective motions and could form hinge centers ( Fig 4D ) . These rigid residues were also evolutionary conserved and featured high pMI scores ( top 5% ) . Hence , an important category of functional sites involved in allosteric regulation of DnaK may be characterized by structural stability , proximity to global hinge centers and local structural environment that is enriched by highly coevolving flexible residues . We argue that these specific characteristics may be necessary for regulation of allosteric structural transitions and could distinguish regulatory sites from nonfunctional conserved residues . Structural mapping of high pMI sites showed that these residues occupy stable regions in the DnaK structures and could form local clusters near the nucleotide binding site and in the SBD-β subdomain ( Fig 5A and 5B ) . Several highly coevolving stable residues ( D194 , L195 , T199 , D201 ) were previously implicated in stimulation of the ATPase activity and form a local cluster in the subdomain IIA ( Fig 5C and 5D ) . Structural proximity of these residues to the catalytic center E171 and nucleotide binding site may impose limitations on conformational diversity of coevolving residues in this region , which may explain the observed stability . However , the vast majority of residues with high cMI scores tend to occupy flexible regions , primarily in the subdomain II of the NBD and the substrate binding region ( Fig 5C and 5D ) . These observations corroborated with the notion that coevolving flexible residues undergoing cooperative structural changes may constitute recognition elements of substrate binding sites [110 , 111] . Another group of regulatory Hsp70 residues ( D326 , K414 , N415 , Q442 and D481 ) revealed high cMI scores and intermediate mobility level ( Figs 4 and 5 ) . Several coevolving flexible sites in subdomain IIA ( D211 , E217 , V218 ) have been implicated in DnaJ recognition [112 , 113] . According to these studies , DnaJ domain interacts with DnaK loop ( residue 206–221 ) and mutations of residues D208 , E209 , D211 , 217 , and V218 on DnaK interfere with DnaK-DnaJ binding and stimulation of the ATPase activity . A strong coevolutionary signal of these residues was previously noticed in a related computational study [64] , suggesting that binding of the subdomain IIA to co-chaperones of the J-domain family may be meditated by a cluster of highly coevolving and structurally proximal residues . Notably , many single mutations and modifications of residue segments in the NBD regions can affect DnaJ binding or compromise ATPase stimulation ( residues Y146-D148 , R151 , D388 , D393 , R167 , N170 , T173 , E217-V218 , V388-L392 , L390-L391 ) [113] . We found that these residues exhibited a significant coevolutionary signal , confirming that several chaperone functions , including co-chaperone recruitment , regulation of the ATPase activity and allosteric control , may be mediated through allosteric coupling of coevolving residues . By consolidating conformational mobility profiles and coevolutionary residue propensities over all simulated Hsp70 proteins , we evaluated the extent of correlation between these parameters ( S2 Fig ) . An appreciable correlation was observed between residue mobility and pMI scores at the intermediate levels of conformational flexibility . Residues with small pMI values were associated with low sequence conservation and may be accompanied by the increased conformational mobility , but this trend does not take effect until an intermediate mobility level is reached . At the same time , no correlation was found between conformational mobility and cMI scores ( S2 Fig ) . While high pMI sites in the Hsp70 structures corresponded mostly to structurally stable residues , coevolving positions exhibited a wide range of conformational mobility with the peaks pointing to the middle part of the spectrum . These findings corroborated with evolutionary studies of protein dynamics [110 , 111] suggesting that highly coevolving residues may preferentially occupy regions of intermediate mobility . We argue that conformational variability of highly coevolving residues that surround rigid regulatory sites may enable concerted rearrangements of specific interactions associated with global allosteric changes . Although crystal structures of the ATP-DnaK and ATP-Sse1 chaperones are similar , their dynamics was somewhat different . MD simulations of the ATP-bound Sse1 structures produced fluctuation profiles that were exemplified by structural stability of the NBD residues , the inter-domain interface and the SBD-α region ( S3 Fig ) . At the same time , we noticed the increased conformational mobility of the substrate binding region in the SBD-β subdomain . These unique dynamic characteristics of the ATP-Sse1 structures resulted in the reduced number of local hinge sites that produce a smaller allosteric network with fewer mediating centers ( S3 Fig ) . In the Sse1 structures , only few high cMI residues occupied flexible regions in the SBD-β , which may also contribute to a limited allostery in this chaperone . We integrated coevolutionary analysis into construction and analysis of the residue interaction networks to test our hypothesis that dynamic and coevolutionary residue correlations may act as synchronizing forces to enable efficient and robust allosteric regulation . In this model , the network edges ( interactions ) are weighted based on both dynamic and coevolutionary residue correlations that determine the shortest communication paths between residue nodes . Residue centrality ( residue betweenness ) is a global network parameter that was computed to determine highly connected nodes in a global interaction network . A propensity of protein residues to serve as global mediating centers of allosteric interaction networks was evaluated by considering common peaks in the residue centrality and structure-based pMI profiles . We show that due to their unique networking and coevolutionary signatures these sites may control allosteric signaling and structural transformations during the Hsp70 functional cycle . A strong relationship was found between high centrality and functional significance of DnaK residues . Importantly , coevolutionary pMI scores ( Fig 3C and 3D ) and residue centrality profiles of the DnaK structures ( Fig 6A and 6B ) showed similar shapes , with regulatory sites mapped almost precisely onto the major peaks of these distributions . In the ADP-bound DnaK , three major broad peaks corresponded to a residue cluster in the subdomain IA ( residues 140–151 ) , the linker region , and residues 479–482 ( L6 , 7 loop ) ( Fig 6A ) . In the ATP-DnaK , the distribution peaks corresponded to the subdomain IA residues ( 140–154 , 161–175 ) , the SBD-β residues ( L454 , F476 , L484 ) and the inter-domain residues from loops L2 , 3 ( residues 412–420 ) , L4 , 5 loop ( 442-QGE-444 ) and L6 , 7 loop ( residues D481 , G482 ) ( Fig 6B ) . Among major peaks were the nucleotide binding site residues ( K70 , R71 ) , and functional residues of allosteric communication located at the inter-domain regions ( R151 , K155 , R167 , I168 , K414 , N415 ) . Structural mapping of functional sites showed that high centrality/high pMI sites R151 , K155 , R167 , I168 are interconnected and linked with a flexible SBD-β “arm” D481 at one side of the inter-domain interface ( Fig 6C and 6D ) . Another inter-domain juncture is formed through specific interactions between highly coevolving functional residues K414 , N415 ( L2 , 3 loop of the SBD-β ) and D326 from subdomain IIA that reside in structural proximity of high pMI residue T221 . Hence , the major inter-domain bridges may be established through coupling of coevolving functional residues that reside in local proximity of high centrality hinge centers . Global network centrality and local proximity to coevolving interfacial residues may facilitate cross-talk between functional hinge centers in coordination of allosteric changes . Mutational variants Y145A and D148A [42] , P143G and R151A [43] , K155D and R167D [44] , K414I [45 , 53] , D326V and N415G [47] may dramatically reduce or eliminate allosteric signaling in DnaK . In light of our results , the loss of regulatory function may be determined not only by disruption of specific inter-domain contacts , but also by global alterations in the network connectivity leading to the reduced efficiency of allosteric interactions . We argue that high network centrality and strong coevolutionary associations of regulatory sites may cause even minor mutations at these positions to be highly detrimental for allosteric regulation . By aggregating coevolutionary residue scores and residue centrality profiles from equilibrium ensembles of all simulated Hsp70 proteins , we evaluated the relationship between these parameters . There was only little correlation between residue centrality and cMI scores ( Fig 7A ) . However , an appreciable correlation was found between residue centrality and pMI scores . Furthermore , functional residues of DnaK regulation displayed consistently high coevolutionary and network centrality scores that were strongly correlated ( Fig 7B ) . We extended this analysis by considering experimentally known functional sites across all Hsp70 proteins . It appeared that the distributions of coevolutionary pMI scores and network centrality for functional sites were markedly shifted towards higher values of these parameters ( Fig 7C and 7D ) . The observed confluence of dynamics correlations and coevolutionary residue couplings with global networking features may determine modular organization of allosteric interactions and dictate localization of key mediating sites . We argue that coevolutionary and networking signatures of functional regions may be in harmony , acting as synchronizing forces that shape up the efficiency and robustness of allosteric regulatory mechanisms . These conclusions echoed recent revelations that coevolutionary relationships may be intimately linked with protein dynamics and determine conformational heterogeneity and functional landscapes of protein structures [114–117] . While many functional DnaK sites corresponded to high pMI residues , regulatory positions in the Sse1 structures featured markedly lower pMI values ( S4 Fig ) . Hence , structural environment of functional sites in the Sse1 structures may be deficient in highly coevolving residues . Moreover , the high centrality peaks in the Sse1 structures were less pronounced and not necessarily coincided with functional sites ( S4 Fig ) . Since properly positioned coevolving residues may provide a primary vehicle for executing cooperative structural changes , the lack of allosteric communication in Sse1 [38–41] may be associated with dislocation of mediating centers and insufficient coevolutionary coupling between functional regions . Notably , selection for specific functional constraints and lower substitution rates is a prominent feature of canonical Hsp70s such as DnaK [118] , whereas atypical Hsp70 chaperones , such as Sse1 , are characterized by the relaxed selection for functional constraints and higher substitution rates [119] . Our results corroborated with this evidence by revealing the reduced dynamic and coevolutionary coupling between functional regions in the Sse1 structures , which may be linked with deficient allosteric signaling observed in this Hsp70 chaperone . This may be contrasted with the observed convergence of dynamic and coevolutionary residue correlations in DnaK that may contribute to a highly cooperative allosteric mechanism with a broad network of mediating centers . To conclude , our results suggested that the interplay between residue coevolution and protein dynamics may be important in shaping up the nature of allosteric regulatory mechanisms that could range from a population-shift mechanism in DnaK to an entropy-driven mechanism adopted by Sse1 . To characterize topology and functional organization of allosteric interactions and communications in the Hsp70 structures , we performed a modular decomposition of protein structure networks using a community detection analysis [120] . Ordinarily , protein structure modularity approaches are based on the residue contact matrix [121 , 122] and do not include dynamic or evolutionary information . The network models of proteins based on residue contacts can often feature either the excessive modularity of rigid communities or produce overly flexible overlapping communities [123–126] . However , an appropriate balance between structural rigidity and flexibility is a cornerstone of protein functions and adaptability [125 , 126] . Using a community decomposition method , the residue interaction networks were divided into local modules in which residue nodes are strongly interconnected through both dynamic and coevolutionary correlations , whereas residues that belong to different communities may be sparsely connected and only weakly coupled . We show that this model can adequately describe a balance between structural rigidity and flexibility within local communities that enables efficient inter-modular connectivity and promotes allosteric signaling in the Hsp70 structures . An important question concerns functional significance of local communities and physical principles underlying modularity of the residue interaction networks . To clarify these issues , we first analyzed the nature and composition of conserved communities that are shared between DnaK structures and can be preserved during allosteric structural changes . An important finding was the emergence of conserved local communities that may be preserved to ensure structural stability and catalytic functions of the DnaK chaperone . In both ADP-DnaK and ATP-Dnak structures , a shared stable community ( K70-E171-P143 ) was detected that links catalytic residues K70 , E171 with the allosteric switch P143 ( Table 1 , Fig 8 ) . K70 and E171 are involved in catalysis of ATP hydrolysis , whereas P143 is a highly conserved residue that could act as a regulatory switch by assuming alternative conformations during ATP binding and hydrolysis [43] . Another conserved community in the subdomain IA ( V142-F146-T154 ) protects structural stability of a critical regulatory residue F146 . Mutation F146A can significantly reduce substrate release rate in the presence of ATP , thus pointing to the role of F146 in signal transmission from NB to SBD [53] . The two conserved communities centered around rigid functional sites P143 and F146 may be necessary for coupling ATP binding to the inter-lobe movements during allosteric transitions [53] . Several other conserved communities ( L324-F356-V331 ) and ( Q343-K270-M346 ) are shared by the DnaK structures forms and are responsible for structural stability of the subdomain IIA . According to single molecule optical tweezer experiments , nucleotide binding in DnaK is dependent on thermal stability of the subdomain IIA [127] . These studies also showed that stabilization and nucleotide-binding function of the lobe II in DnaK may be associated with structural preservation of residue cluster 330–345 in subdomain IIA . Our results may rationalize these experiments by showing that structural and evolutionary preservation of local interacting modules ( L324-F356-V331 ) and ( Q343-K270-M346 ) in the subdomain IIA may protect stability and nucleotide binding function of the NBD core . Intriguingly , the experimental data revealed that subdomain II regions could be mainly responsible for protein stability and nucleotide binding , while allosteric signaling may be primarily mediated by the regulatory residues in the subdomain I [127] . Our findings supported this assertion by revealing that functional centers with high network centrality may be consolidated in the subdomain IA . These residues could also form conserved and extremely stable interacting communities such as ( K70-E171-P143 ) that couples nucleotide binding residues K70 and E171 with the allosteric signaling switch P143 [43 , 53] . Several conserved communities ( I421-T420-I478 ) and ( L484-V440-L454 ) were also detected in the hydrophobic core of the SBD-β . According to NMR studies , these residues form a critical allosteric hotspot for communicating global dynamic changes from the NBD-SBD interface to the substrate binding site [36] . In the ADP-DnaK , this community links the SBD-β core with the inter-domain interface ( L484-V440-L454-I501 ) , while in the ATP-DnaK , the expanded module ( L454-V440-L399-L484 ) connects the SBD-β to the inter-domain linker . Finally , a conserved community in the SBD-β subdomain ( E444-S398-K414 ) links high pMI residues S398 and E444 with a functionally important inter-domain residue K414 ( Table 1 ) . This community may provide a stable bridge that transmits allosteric signal from the inter-domain residue K414 to the hydrophobic core of the SBD-β . Importantly , all conserved communities shared by the DnaK forms are formed by functional residues that display high network centrality and exhibit strong coevolutionary signals ( Table 1 , Fig 8 ) . We also examined another category of local communities that are associated with rearrangements of the inter-domain interactions responsible for global structural changes and a population shift in DnaK . In the ADP-bound DnaK , the inter-domain modules ( L397-E444-M515 ) , ( D481-K387-D385 ) , ( N415-D393-T395 ) , ( M515-N451-R445 ) , ( R467-D431-H544 ) , and ( E444-S398-K414 ) were centered on residues E444 and N451 ( top 5% pMI ) and included highly coevolving residues R445 , K414 , N415 ( top 5% cMI ) ( Fig 8A ) . These communities occupied three key regions of the inter-domain interface: a ) the inter-domain linker connected with the SBD functional sites D481 and N415; b ) the interface between SBD-β ( N451 , R445 ) and SBD-α ( M515 ) ; c ) the hinge interface between substrate binding loop ( D431 , R467 ) and SBD-α ( H544 ) ( Fig 8A ) . The disruption of these interaction communities during allosteric transition from ADP-bound to ATP-bound DnaK involves coordinated rearrangements in positions of the key SBD-β “handles” ( K414 , N415 and D481 ) that become tightly locked in the ATP-DnaK and are recruited into local modules ( D481-K155-R167-I168 ) and ( K414-N415-D326-T221 ) ( Fig 8B ) . One of these inter-domain communities ( D481-K155-R167-I168 ) strengthens a critical inter-domain juncture formed through specific interactions between D481 and I168 . Other communities ( V322-D326-K414 ) and ( T221-V218-V394-N415 ) link the NBD and SBD-β domains at another juncture of the interface ( Table 1 , Fig 8B ) . These stable modules couple functional residues K414 and N415 ( L2 , 3 loop ) with T221 and D326 from subdomain IIA . Importantly , local inter-domain communities are anchored by high pMI residues ( T221 , L454 ) and include highly coevolving residues ( D148 , D326 , K414 , N415 ) . Disruption of these communities through mutations K414I and N415G can affect substrate stimulation of the ATPase activity [53] . In the central inter-domain region , two communities ( Q442-D148-L454 ) and ( L454-V440-L399-L484 ) bridged the interfacial Q442-D148 pair with the key residues in the SBD-β core: L454 ( β5 strand ) , and L484 ( β7 strand ) . These communities are assembled around high pMI sites ( L454 , V440 ) and include highly coevolving functional residue D148 ( Table 1 ) . According to our findings , the reorganization of local communities during allosteric changes in DnaK may be determined by rearrangements of specific interactions formed by regulatory sites K414 and D481 . In the ATP-DnaK , these residues are involved in two critical inter-domain bottlenecks K414-D326 and D481-I168/D481-R151 that control transmission of allosteric signals ( Fig 8B ) . Moreover , the fidelity of allosteric signals navigating through these inter-domain passages may be protected by stability of local communities ( D481-K155-R167-I168 ) , ( V322-D326-K414 ) , and ( E444-S398-K414 ) . These findings may explain why mutations of D481 and K414 residues are the most detrimental for the intrinsic ATPase activity ( ~84 fold loss for D481A and D481K modifications and ~26 fold for K414I mutation ) [53] . We argue that the observed functional effects may result from significant alterations in the modular organization of allosteric interaction networks . To substantiate these arguments , we conducted alanine scanning of functional inter-domain residues F146 , R151 , I168 , D326 , K414 , and D481 . In these computations , we utilized the conformational ensemble obtained from MD simulations of the ATP-DnaK and engineered alanine mutations into 10 , 000 trajectory snapshots that were subsequently optimized by the 3Drefine method [128] . Using this “single trajectory” protocol to obtain conformational ensembles of mutational DnaK variants , we recalculated the dynamics and coevolutionary correlations between residues , reconstructed the residue interaction networks , and performed a community decomposition for each studied mutant ( S5 Fig ) . The results revealed an appreciable decline in the total number of local communities , confirming that mutations of functional inter-domain residues could disrupt not only interfacial communities but also lead to fragmentation of the global network , and thus reduce the efficiency of allosteric signaling . The high network centrality of F146 and D481 residues that are strategically positioned in the dense interfacial region of the ATP-DnaK , may explain the greater effect of mutations in these positions on modularity of allosteric interactions ( S5 Fig ) . To summarize , the performed community analysis addressed several important questions concerning modular organization of the residue interaction networks . First , we found that conserved communities may arise from requirements for structural stability and preservation of catalytic functions in DnaK . Second , it appeared that different communities in the ADP-DnaK and ATP-DnaK structures may be associated with rearrangements of specific interactions at the inter-domain regions that promote allosteric changes . Our results demonstrated that many regulatory sites in DnaK may be distinguished by their high centrality and integrating role in local interaction communities . The emergence of dynamic inter-domain modules that are anchored by high centrality sites and include coevolving flexible residues is a central result of this analysis . Dynamic and coevolutionary couplings between rigid and mobile residues within local communities may balance a strong intra-modular connectivity with weak inter-modular ties to propagate conformational changes . It may be suggested that coevolutionary dependencies of flexible residues in local communities may compensate the effects of some mutations and preserve modularity of the allosteric interaction network which may be required for efficient signaling . However , targeted mutations of high centrality mediating sites and residues involved in the inter-community connectivity may cause disruption of multiple interactions and significant rearrangements in modularity and efficiency of the allosteric interaction networks . We introduced a community-hopping model of allosteric communication pathways based on the notion that cooperative transitions may occur between local communities of tightly coupled interacting residues that could be more loosely coupled to one another . In this model , the interacting residues in local communities are typically spatially close in the protein structure and tend to switch their conformational states cooperatively . At the same time , each community could maintain only weak association with other communities . Collectively , these modules may form a weakly coupled assembly acting as a communication pathway in signal transmission . This model of allosteric pathways is rooted in the network formalism of protein structure and is motivated by a long-standing “weak-strong tie” hypothesis [129 , 130] . According to this theory , a tie ( or interaction strength ) may be determined by the underlying network topology , where “‘weak” ties ( interactions ) connect and transmit information between local communities consisting of “strongly” connected residues . A central assumption of this model is that the inter-community hopping between pairs of highly coevolving and dynamically correlated nodes may define “stepping stones” of optimal allosteric communication pathways . This model is based on allosteric communication propensities of protein residues that could be evaluated by considering fluctuations of the mean distance between a given residue and all other residues in the protein structure [131 , 132] . In this approach , a pair of residues would communicate with a high efficiency when their inter-residue distance fluctuates rather moderately . Alternatively , a pair of residues is expected to communicate poorly in the absence of correlated fluctuations leading to large variations in the inter-residue distance . We extended this model by relating CP values to average variations in the composite distance metric that measures residue distance fluctuations and variations in pMI score differences between a given residue and all other residues in the protein structure . A central assumption of this model is that the inter-community hopping between pairs of highly coevolving and dynamically correlated nodes may define “stepping stones” of optimal allosteric communication pathways . To address the experimentally detected dichotomy of DnaK allostery [53] , we performed a direct mapping of forward ( NBD-SBD ) and reverse ( SBD-NBD ) pathways in the DnaK structures ( Fig 9 ) . We selected K70 from the nucleotide binding site of the NBD as a starting point and residue D431 in the substrate binding site of the SBD as an end point . For simplicity , it was assumed that communication routes between these two residues could be representative of signal transmission pathways between the nucleotide and substrate binding sites . Modeling of the short communication pathways in the DnaK structures revealed an ensemble of efficient routes that navigated through a network of mediating residues with high network centrality . Despite the presence of multiple signaling routes , only several dominant forward and reverse pathways contributed 75%-90% of the population ( Table 2 ) . A certain divergence of forward and reverse pathways could be noticed in the ADP-DnaK structure ( Fig 9A ) . The most probable forward ( NBD-SBD ) pathway ( 55% occupancy in the ensemble ) connected the nucleotide binding site with R167 to reach the inter-domain linker and local community ( D481-K387-D385 ) centered around functional residue D481 . After reaching this critical juncture , the route moved through the SBD-β hydrophobic residues before locating residue F426 , which is a key allosteric hotspot in the SBD-β [36] . Upon reaching this point , the pathway was directed to the community ( R467-D431-H544 ) that links the SBD-β and SBD-α subdomain . Notably , the forward communication pathway traversed through major functional residues involved in allosteric regulation ( R167 , D481 and F426 ) . The most probable reverse ( SBD-NBD ) pathway in the ADP-DnaK ( 77% occupancy ) was somewhat different by proceeding through SBD communities ( L484-V440-L454-I501 ) , ( I412-T420-I478 ) before reaching regulatory sites P419 , D481 to cross the inter-domain interface and navigate to the binding site via I168 and R167 ( Fig 9A ) . At the same time , both forward and reverse pathways in the ADP-DnaK maneuvered through similar regulatory sites ( R167 , I168 , P419 , D481 , V440 , and F426 ) . A subtle yet functionally important dichotomy between forward and reverse pathways was also evident in the ATP-DnaK structure ( Fig 9B ) . The forward ( NBD-SBD ) communication pathways proceeded initially from K70 via community ( K70-E171-P143-D201 ) to P143 and then to functional site F146 through community ( I73-V142-F146-T154 ) . At this point , the first optimal forward path ( 48% occupancy ) crossed the inter-domain interface through D326-K414 bridge . The second most probable route ( 45% occupancy ) similarly connected K70 to F146 and then moved to I168 , R151 , and K155 via a critical community ( I168-K155-D481 ) to cross the inter-domain interface at another critical juncture I168-D481 ( Fig 9B , Table 2 ) . These two shortest pathways dominated the distribution of signaling routes and travelled through key mediating sites F146 , Y145 , I168 , R151 , K155 , and F426 . In contrast , a strong preference for a single reverse pathway ( 78% occupancy ) was found . Furthermore , the optimal SBD-NBD path was different from the forward route and navigated through different functional centers . In this case , the path moved from the substrate binding site by hopping between SBD-β communities ( L484-L399-V440-L454 ) and ( Q442-L454-L484-D148 ) to cross the SBD-NBD interface at a different juncture point ( L454 , D148 ) ( Fig 9B ) . This inter-domain connection appeared to be a preferential transition point for the reverse SBD-NBD pathway , but was not featured at all in the ensemble of forward NBD-SBD routes . The performed atomistic reconstruction of communication pathways in the DnaK structures is in excellent agreement with the recent functional studies [53] . These experiments dissected pathways of allosteric regulation by analyzing how mutations of functional residues could impede specific steps of signal transmission . Mutations Y145A , F146A , D481A , and D481K could abolish the forward ( NBD-SBD ) signaling and block inhibition of ATP hydrolysis in DnaK [53] . Of special interest , amino acid substitutions of F146 that could lead to deficient ATP-induced substrate release ( NBD-SBD direction ) , but produce only minor effects on substrate-induced stimulation of the ATPase activity ( SBD-NBD direction ) . Our results were fully consistent with these experiments , showing that forward pathways in the ATP-DnaK were obligated to proceed through F146 before reaching the inter-domain bridges D326-K414 ( path 1 ) and I168-D481 ( path 2 ) ( Fig 9B , Table 2 ) . Moreover , this communication hub was specific for the forward NBD-SBD pathways , but appeared to be far less important for the SBD-NBD signal transmission . On the other hand , alanine mutations of V440 , L440 and D148 residues strongly affected the SBD-NBD signaling and substrate stimulation of the ATP hydrolysis , but were less detrimental for signal transduction in the NBD-SBD direction [53] . According to our results , a single optimal SBD-NBD path navigated through local communities ( L484-L399-V440-L454 ) and ( Q442-L454-L484-D148 ) that were anchored by allosteric centers V440 , L454 , and L484 . This route is critically dependent on passing through a L454-D148 transition point that is specific for the reverse signaling , but was not observed in the ensemble of NBD-SBD pathways . In network terms , forward communication is critically dependent on conserved mediating centers of allosteric interactions whose mutations would be lethal for chaperone function . At the same time , reverse signaling invoked only few regulatory sites that are less critical for efficiency of allosteric interaction networks . These findings may rationalize the experimental evidence that efficient ATP-induced substrate release ( forward communication ) can be more critical for chaperone function than substrate stimulation of the ATPase activity ( reverse SBD-NBD signaling ) [53] . The goal of this study was to present a systematic computational analysis of the dynamic and evolutionary factors underlying allosteric structural transformations of the Hsp70 proteins . We investigated the relationship between sequence conservation , conformational dynamics , coevolutionary associations and organization of the residue interaction networks in the Hsp70 proteins . The central finding of this study is that functional centers of Hsp70 regulation could be distinguished by their specific dynamic , coevolutionary and networking signatures . We found that global features that differentiate functional residues include high network centrality and high pMI scores , indicating that local structural environment of key mediating centers may be enriched by coevolving residues . The key sites involved in allosteric signaling of DnaK corresponded to either invariant high pMI residues or coevolving residues with only conservative replacements in the Hsp70 family . A novel methodological aspect of this work was integration of three complementary factors that contribute to the modular organization of the residue interaction networks: the residue contact matrix , dynamic inter-residue correlation maps and structure-based coevolutionary residue correlations . We performed a community decomposition of the interaction networks in the Hsp70 structures and established functional significance and physical principles underlying modular organization of allosteric interactions . Conserved local communities may preserve structural stability and catalytic functions of the DnaK chaperone . Another category of local communities is involved in rearrangements of the inter-domain interactions responsible for global structural changes and a population shift in DnaK . The inter-domain communities in the Hsp70 structures harbor most of the functional residues implicated in allosteric regulation , suggesting that these sites could be integral for coordination of global structural changes . In network terms , mutations of these residues may give rise to global changes by simultaneously altering many interactions and triggering population shifts in the conformational equilibrium . Our results demonstrated that confluence of dynamics and coevolutionary associations between Hsp70 residues may determine efficiency of allosteric interaction networks and dictate the regulatory mechanism–from a highly cooperative population-shift in DnaK to a less cooperative entropy-driven allostery in Sse1 . By using allosteric residue propensities , we also developed a community-hopping model of allosteric communication pathways . Using this approach , we confirmed that efficient allosteric communications could be controlled by structurally stable functional centers that exploit coevolutionary coupled flexible residues in their local communities to propagate structural changes . We investigated a direction-specific nature of communication pathways in the DnaK chaperone and explained the role of specific residues mediating distinct steps of the Hsp70 cycle . This study reconciled a range of structural and functional experiments from a network-centric perspective , by showing that architecture and global properties of the residue interaction networks and communication pathways may be linked with specificity of allosteric regulatory mechanisms . All-atom MD simulations were performed for the following panel of full-length two-domain Hsp70 structures [133]: an ADP-bound DnaK ( pdb id 2KHO ) ; the crystal structure of an ATP-bound DnaK ( pdb id 4B9Q , 4JNE ) ; the crystal structure of the ATP-bound DnaK from multi-crystal single-wavelength anomalous diffraction ( SAD ) data set ( pdb id 4JN4 ) ; the crystal structure of DnaK in post-ATP hydrolysis state ( pdb id 2V7Y ) ; the crystal structure of the HSC bovine construct E213A/D214A mutant ( pdb id 1YUW ) ; the crystal structure of the of the HSC bovine construct E213A/D214A mutant ( pdb id 4FL9 ) ; the crystal structure of the native ATP-bound Sse1 ( pdb id 2QXL ) ; the crystal structure of ATP-Sse1 nucleotide exchange complex with the NBD of HSC70 bovine ( pdb id 3C7N ) ; the crystal structure of the selenomethionine-derivatized Sse1 construct in a complex with the NBD of hHsp70 ( pdb id 3D2E ) ; and the crystal structure of the native Sse1 in a complex with the NBD of hHsp70 ( pdb id 3D2F ) . These crystal structures included the apo states , the substrate-bound chaperone forms , and the nucleotide-bound Hsp70 structures . We have carried out two independent 500 ns and five independent 200 ns MD for each of the studied Hsp70 structures . ModLoop [134–136] and ArchPRED [137] homology modeling approaches were employed for reconstruction and optimization of missing loops in the Hsp70 structures . The chaperone structures were then optimized using the 3Drefine method [128] . All-atom MD simulations were performed with the aid of NAMD 2 . 6 package [138] . CHARMM22 force field [139 , 140] and the explicit TIP3P water model [141] were used in these simulations . The details of the MD protocol were previously reported and extensively discussed in our studies of Hsp70 chaperones [80 , 81] , Hsp90 chaperones [142 , 143] and protein kinases [144 , 145] . All MD simulations were done in the NPT ensemble at 1atm and 300K using Langevin piston coupling algorithm as described in our previous studies [80 , 81 , 142–145] . Collective motions and functional dynamics of the Hsp70 structures were modeled using the elastic network-based GNM approach [107–109] . The details of precise implementation of this approach were reported in a related study of Hsp70 chaperones [81] . Conformational mobility profiles in the essential space of low frequency modes were obtained using the oGNM [108] and ANM web servers [109] . Coevolutionary associations between residue pairs in the Hsp70 protein family were evaluated using MI analysis [99 , 100] . In this approach , multiple sequence alignment ( MSA ) profile of the Hsp70 protein family was obtained from Pfam database [146 , 147] . All sequences in the MSA within curated thresholds ( E-value = 10−2 and a column-inclusion threshold of 80% ) were included in the Hsp70 sequence alignment profile . A statistically significant and diverse number of sequences ( 16272 sequences ) in the Pfam database provided input for the MI computations . In MISTIC approach , sequence clustering is implemented to reduce sequence redundancy and sequence clusters are defined at a sequence identity threshold of 62% [148] . A lower bound of 400 sequences <62% identity is typically required in an MSA to yield statistically meaningful coevolutionary relationships . To discriminate coevolutionary associations driven by functional constraints from those determined by common ancestry , the covariance metric based on MI calculations was adjusted by the average product correction ( APC ) [104–106] . MI values of residue associations in the Hsp70 family corresponded to the Z-score normalized MI values that were adjusted through sequence clustering and APC correction . The Kullback-Leibler ( KL ) sequence conservation score KLConsScore was calculated using MSA profile of the Hsp70 protein family . The reference sequence in the alignment corresponds to DNAK_ECOLI ( residues 4–604 ) . The KL conservation score is computed as follows: KLConsScorei=∑i=1NlnP ( i ) Q ( i ) ( 1 ) P ( i ) is the frequency of amino acid i in a particular position and Q ( i ) is the background frequency of amino acid i obtained from the UniProt database [149] . A cumulative mutual information ( cMI ) score is a sequence-based parameter that measures the extent of mutual information shared by a given residue with all other protein residues . cMI is calculated as the sum of MI values above a threshold t = 6 . 5 for every pair in which a particular residue of interest appears [99 , 100]: cMIx=∑y , MI ( x , y ) >tMI ( x , y ) ( 2 ) A proximity mutual information ( pMI ) score estimates structural constraints imposed on coevolutionary dependencies . This parameter is defined as the average of cMI scores of all residues within a local structural proximity from a given residue in the protein structure [99 , 100] . The distance between each pair of residues was calculated as the shortest distance between any two heavy atoms that belong to each of these two positions . A threshold distance t = 5 Å is used to define structural residue proximity: pMIx=1N∑d ( x , y ) , tcMI ( x , y ) ( 3 ) For each residue , pMI score was computed using an ensemble-based definition of local residue environment . The amount of mutual information shared by a given residue with the spatially close neighboring nodes was obtained by averaging computations from 10 , 000 conformations along MD trajectories . A graph-based model of protein structure and topological residue connectivity are used to construct the residue interaction networks . In this network , residues are network nodes and edges represent residue interactions . The details of the graph construction using a particular interaction cut-off strength ( Imin ) were extensively discussed in the initial reports [70 , 71] and our previous studies [80 , 81 , 143–145] . The edges in the residue interaction network are then weighted based on dynamic residue correlations and coevolutionary couplings measured by the MI scores . In this model , weight wij is defined as the element of a matrix measuring the generalized correlation coefficient rMI ( xi , xj ) between residue fluctuations in structural and coevolutionary dimensions . The composite residue vector includes variables describing instantaneous residue positions in the three-dimensional space of protein structure and respective proximity-based MI score: wij=−log[rMI ( xi , xj ) ] ( 4 ) The edge lengths in the network are thus obtained using the generalized correlation coefficients rMI ( xi , xj ) associated with the dynamic correlation and mutual information shared by each pair of residues . The length ( i . e . weight ) of the edge that connects nodes i and j wij = −log[rMI ( xi , xj ) ] is calculated from the corresponding generalized correlation coefficient between these nodes . The matrix of communication distances is obtained using generalized correlation between composite variables describing both dynamic positions of residues and coevolutionary mutual information between residues . The ensemble of shortest paths is determined from matrix of communication distances by the Floyd-Warshall algorithm [150] that compares all possible paths between each pair of residue nodes . Using this protein structure network model , we computed the residue-based centrality parameter . The centrality of residue i is determined as its network betweenness computed as a fraction of the shortest paths between all pairs of residues that pass through residue i: Cb ( ni ) =∑j<kNgjk ( i ) gjk ( 5 ) where gjk denotes the number of shortest paths connecting j and k , and gjk ( i ) is the number of shortest paths between residues j and k that navigate through the node ni . Residues that populate a significant portion of the shortest paths connecting all residue pairs are characterized by high betweenness values ( high residue centrality ) . For each node n , the betweenness value can be normalized by the number of node pairs that exclude node n which is given as ( N - 1 ) ( N - 2 ) / 2 , where N is the total number of nodes in the connected component that node n belongs to . Network centrality analysis and community detection were done using CFinder program [151] . In this ensemble-based model , local interaction communities in the Hsp70 structures were evaluated using 10 , 000 conformations along MD trajectories . Local communities that remained stable and maintained their modular organization in more than 75% of the ensemble conformations were reported . The Girvan-Newmann algorithm [152 , 153] was used to maximize the modularity and optimize the quality of the community structure . This method utilizes the edge betweenness as a partitioning criterion and splits network into local communities , where the connections ( interactions ) within local communities are strong and dense , while the connections between communities are weaker and sparser . The implementation of this model is based on computation of allosteric communication propensities of protein residues . CP metric computes residue distance fluctuations and variations in pMI score differences between a given residue and all other residues in the protein structure over the course of MD simulations . For each residue , CP metric is evaluated as follows: CPi=3kBT〈w1 ( di−〈di〉 ) 2+w2 ( ΔpMi−〈ΔpMi〉 ) 2〉 ( 6 ) di=〈dij〉j* ( 7 ) ΔpMi=〈ΔpMij〉j* ( 8 ) where dij is the distance between residue i and residue j , ΔpMij is the difference between pMI scores of residues i and j; kB is the Boltzmann constant , T = 300K . di = ⟨dij⟩j* is the average distance from residue i to all other residues in the protein structure . ΔpMi is the average difference in pMi scores between residues i to all other residues j in the protein . In this expression , w1 and w2 are weighting factors adjusted to achieve optimal modularity of local communities . Based on optimization of network modularity and community partition by Girvan-Newmann algorithm [152 , 153] , each community consists of strongly connected and coupled residues , while different communities could maintain weak association with each other that are mediated by central network hubs . By evaluating communication propensities of residues in local communities , the candidate residues for the inter-community hopping are selected . A communication pathway in this model could be viewed as migration between strongly interacting residues within a local community that is coupled with the inter-community hopping event connecting a pair of coevolving and dynamically coupled residues from structurally proximal modules .
The diversity of allosteric mechanisms in the Hsp70 proteins could range from modulation of the inter-domain interactions and conformational dynamics to fine-tuning of the Hsp70 interactions with co-chaperones . The goal of this study is to present a systematic computational analysis of the dynamic and evolutionary factors underlying allosteric structural transformations of the Hsp70 proteins . We investigated the relationship between functional dynamics , residue coevolution , and network organization of residue interactions in the Hsp70 proteins . The results of this study revealed that conformational dynamics of the Hsp70 proteins may be linked with coevolutionary propensities and mutual information dependencies of the protein residues . Modularity and connectivity of allosteric interactions in the Hsp70 chaperones are coordinated by stable functional sites that feature unique coevolutionary signatures and high network centrality . The emergence of the inter-domain communities that are coordinated by functional centers and include highly coevolving residues could facilitate structural transitions through cooperative reorganization of the local interacting modules . We determined that the differences in the modularity of the residue interactions and organization of coevolutionary networks in DnaK may be associated with variations in their allosteric mechanisms . The network signatures of the DnaK structures are characteristic of a population-shift allostery that allows for coordinated structural rearrangements of local communities . A dislocation of mediating centers and insufficient coevolutionary coupling between functional regions may render a reduced cooperativity and promote a limited entropy-driven allostery in the Sse1 chaperone that occurs without structural changes . The results of this study showed that a network-centric framework and a community-hopping model of allosteric communication pathways may provide novel insights into molecular and evolutionary principles of allosteric regulation in the Hsp70 proteins .
[ "Abstract", "Introduction", "Results", "and", "Discussion", "Conclusions", "Materials", "and", "Methods" ]
[ "allosteric", "regulation", "crystal", "structure", "protein", "interaction", "networks", "condensed", "matter", "physics", "enzymology", "network", "analysis", "chaperone", "proteins", "protein", "structure", "dna", "dna", "structure", "crystallography", "enzyme", "chemistry", "computer", "and", "information", "sciences", "solid", "state", "physics", "proteins", "enzyme", "regulation", "proteomics", "molecular", "biology", "centrality", "physics", "biochemistry", "nucleic", "acids", "genetics", "biology", "and", "life", "sciences", "physical", "sciences", "macromolecular", "structure", "analysis" ]
2017
Computational Analysis of Residue Interaction Networks and Coevolutionary Relationships in the Hsp70 Chaperones: A Community-Hopping Model of Allosteric Regulation and Communication
A major challenge to developing a successful HIV vaccine is the vast diversity of viral sequences , yet it is generally assumed that an epitope conserved between different strains will be recognised by responding T-cells . We examined whether an invariant HLA-B8 restricted Nef90–97 epitope FL8 shared between five high titre viruses and eight recombinant vaccinia viruses expressing Nef from different viral isolates ( clades A–H ) could activate antiviral activity in FL8-specific cytotoxic T-lymphocytes ( CTL ) . Surprisingly , despite epitope conservation , we found that CTL antiviral efficacy is dependent on the infecting viral isolate . Only 23% of Nef proteins , expressed by HIV-1 isolates or as recombinant vaccinia-Nef , were optimally recognised by CTL . Recognition of the HIV-1 isolates by CTL was independent of clade-grouping but correlated with virus-specific polymorphisms in the epitope flanking region , which altered immunoproteasomal cleavage resulting in enhanced or impaired epitope generation . The finding that the majority of virus isolates failed to present this conserved epitope highlights the importance of viral variance in CTL epitope flanking regions on the efficiency of antigen processing , which has been considerably underestimated previously . This has important implications for future vaccine design strategies since efficient presentation of conserved viral epitopes is necessary to promote enhanced anti-viral immune responses . One of the greatest challenges in developing an effective T-cell based vaccine against HIV-1 is its high genetic variability [1] . Group M HIV-1 has expanded globally into 15 major clades , sub-clades and several interclade circulating recombinant forms . These continually evolving HIV-1 clades differ by over 30% in envelope amino acid sequences and viral isolates within the same clade may also differ by up to 15% [2] . The high rate of mutation from error-prone reverse transcription combined with replicative ability enables HIV-1 to adapt rapidly to immune and drug pressure , with the generation of multiple genetically distinct quasispecies within infected individuals . HIV-1 vaccines must overcome these obstacles to induce protective immunity against heterologous viral variants [3] . A critical component of HIV-1 control during the acute phase is the cytotoxic T-lymphocyte ( CTL ) response [4] , [5] . Therefore , many current vaccine strategies focus on identifying immunogens that elicit effective T-cell immunity against a diverse range of viral variants and characterising HIV-1 specific CTL responses in order to define the immune correlates of protection [5] . To counteract antigenic diversity , there is an increasing interest in developing HIV vaccines which elicit CTL responses to conserved epitopes , centralised sequences or immunogenic regions of high inter-clade homology [6] , [7] , [8] , [9] , [10] . Currently , the interferon gamma ( IFNγ ) producing ELISpot assay is frequently used to quantify the breadth and magnitude of CTL responses [11] , using peptides matched to consensus virus sequence or occasionally to autologous infecting virus sequence [12] [13] . Most data show that CTL recognition of epitope peptides is very sensitive to any change in the epitope peptide [13] , [14] . Thus the HLA type of the patient imprints changes on the sequence of the infecting virus , generally thought to be within the epitopes that have stimulated CTL responses [15] , [16] . However , while CTL may efficiently recognise exogenously loaded synthetic peptide matched to HIV-1 clade variants , it has been found that this does not necessarily correlate with CTL antiviral activity against HIV-1 infected cells displaying endogenously derived peptides [17]; for example , the artificial peptides may be added at non-physiological concentrations . Therefore , conventional peptide-based assays may over-estimate the ability of CTL to cross-recognise variant epitopes [18] . The use of exogenous synthetic peptides to quantify CTL responses may also fail to detect differences in the antiviral efficacy of CD8 T-cells that reflect variation in antigen processing efficiency within HIV-infected cells [19] , [20] . Remarkably , whilst much research has focused on recognition of exogenously added peptide epitopes , CTL recognition of virus-infected cells has been examined relatively rarely , and there has been no analysis of CTL recognition of invariant epitopes shared by diverse viral isolates and clades . The present study arose from the observation that CD8+ T cells specific for the highly conserved HLA B8 restricted Nef epitope FLKEKGGL ( FL8 ) failed to recognize HLA B8 positive cells infected with several HIV-1 isolates . Previously it has been shown that escape mutations can occur in the epitope flanking regions through impaired processing and presentation [21] , [22] , [23] , [24] , [25] , [26] , [27] , however , such studies have focused predominantly on a single viral isolate , mostly in circulating virus rather than selected viral strains in individual patients , or focused on classical escape mutation [28] . Therefore a range of HIV-1 isolates and vaccinia viruses expressing different Nef proteins , each of which shared this conserved epitope , were used to test responses from a set of CTL clones isolated from HLA B8+ patients . Overall , we evaluated CTL recognition and antiviral efficacy induced by a total of thirteen viral isolates containing the same conserved epitope . Surprisingly , only a small proportion ( 23% ) of these HIV-1 isolates induced optimal CTL recognition and antiviral efficacy . We found that variations in the flanking region had a profound effect on the presentation of this epitope , and viral isolates within the same HIV-1 clade were differentially recognized by FL8-specific CTL clones . Furthermore , we identified a phenylalanine motif in the FL8-epitope flanking regions of four HIV-1 isolates that led to an altered pattern of cleavage by the immunoproteasome that correlated with loss of CTL recognition . In conventional IFNγ ELISPOT and chromium release assays , we assessed three Nef FL8-specific CTL clones and four control Gag EI8-specific CTL clones for their recognition of HLA-B8+ matched C8166 target cells pulsed with peptides at different concentrations to measure functional avidity ( the peptide concentration that gives 50% maximum effect ) . The FL8-specific and EI8-specific CTL clones had comparable levels of functional avidity , measured by IFNγ release ( Figure 1A ) and in a lytic assay ( Figure 1B ) . In addition , there were no significant differences ( p>0 . 05 ) observed between FL8- and EI8- specific responses in the lysis assay at all peptide concentrations tested . This suggested that both peptides would be equally recognised in HLA-B8 target cells infected with HIV-1 . We then compared CTL recognition of synthetic Nef FL8 peptide pulsed exogenously onto the surface of uninfected C8166 targets with CTL antiviral efficacy against endogenously derived FL8 epitope presented on the cell surface of HIV-1HXB2 infected C8166 targets . Surprisingly , FL8-specific CTL did not recognise or mount an antiviral response against the virus-infected target cells . In a Viral Suppression Assay ( VSA ) , the panel of FL8- and EI8- specific CTL clones were co-cultured for four days with C8166 target cells infected with HIV-1HXB2 at five E∶T ratios , after which suppression of viral replication in the supernatant was quantified by measuring the level of Gag p24 by ELISA; and suppression of HIV-infected targets ( via CTL lysis or non-cytolytic inhibition ) was measured using intracellular p24 staining . Despite the similar functional avidity in peptide based assays , at a low infectious titre of virus , EI8-specific CTL exhibited superior suppression of viral replication compared to FL8-specific CTL when p24 was measured in supernatant by ELISA ( Figure 2A ) . At all E∶T ratios tested , co-culture with EI8-specific CTL resulted in undetectable levels of p24 Gag ( below the ELISA threshold of 10 pg/ml ) whilst viral suppression by FL8-specific CTL was negligible when compared to wells of virus-infected targets in the absence of CTL . Analysis of suppression at a high infectious titre showed similar results , with negligible suppression of virus-infected cells by FL8-specific T-cells at all E∶T ratios tested ( ranging from 37–89 ng/ml with a mean p24 of 62 ng/ml in the absence of CTL ) ( Figure 2B ) . Subsequent analysis of the p24 stained cells showed a similar pattern ( Figure 2C ) . At the 1∶1 and 1∶4 ratios , the control EI8 specific CTL clones demonstrated effective inhibition of viral replication , although they were unable completely to abrogate infection , whilst FL8-specific CTL did not differ from the controls in their suppression of viral infection at all three ratios tested . 2-way ANOVA with bonferroni post-test confirmed a statistically significant difference ( p<0 . 01 or p<0 . 001 ) between Nef-specific and Gag-specific clones at each E∶T ratio in the ELISA . We also developed a Live Virus Elispot ( LVE ) assay to assess IFNγ release by CTL exposed to HIV-infected target cells , as a marker of CTL antiviral activity over time . C8166 target cells were infected with HIV-1HXB2 and incubated for a period of 24 , 48 , 72 and 96 hours at 37°C . At each respective time point , the HIV-infected cells were co-cultured with HLA-matched FL8- and EI8-specific CTL clones on a pre-coated IFNγ ELISpot plate . Again , FL8-specific CTL did not mount an IFNγ response at any of the time points tested ( Figure 2D ) . In contrast , the EI8 epitope was recognised by the EI8-specific CTL clones , which generated a detectable IFNγ response that was significantly different ( p<0 . 001 ) between FL8- and EI8-specific clones at 48 , 72 and 96 hours post-infection in a 2-way ANOVA with bonferroni post-test . The three Nef FL8-specific clones were generated from three separate HLA-B8+ long-term non-progressors ( LTNP ) and control Gag EI8-specific clones were generated from one HLA-B8+ LTNP patient . The use of this pre-screened panel of T-cell clones removes the complexity of different TCR affinities , and thus variation in TCR/pMHC interactions , of polyclonal T-cell responses . To investigate whether CTL recognition and antiviral activity to a conserved epitope may depend on the infecting HIV-1 isolate , we chose five high titre HIV viruses ( three clade B laboratory strains and two clade A isolates ) that share the invariant Nef FL8 epitope for testing via in vitro Viral Suppression Assays ( VSA ) and Live Virus ELISPOTS ( LVE ) . FL8 is an immunodominant epitope that is highly conserved amongst HIV-1 Group M isolates in the Los Alamos National Laboratory ( LANL ) HIV sequence database . Proviral DNA for each virus was isolated from control wells containing infected C8166 targets in the absence of CTL , PCR amplified and sequenced to confirm the presence of the invariant FL8 epitope . The results from both assays and viral sequencing are summarised in Table 1 . Data from the viral suppression assay for each virus shows differing FL8-specific CTL antiviral activity for the different virus isolates , despite sharing the conserved FL8 epitope ( Figure 3 ) . FL8 specific CTL suppressed both clade A viral isolates HIV-192UG029 and HIV-193RW024 , as well as Clade B HIV-1MN , to below the threshold of detection ( 10 pg/ml ) in the p24 ELISA ( Figure 3C ) . However , clade B HIV-1HXB2 ( as characterised before in Figure 2A ) and clade B HIV-189 . 6 were not suppressed by FL8-specific CTL at any of the effector∶target ( E∶T ) ratios tested , and were not statistically different ( p>0 . 05 ) when compared to wells of virus-infected cells in the absence of CTL . Similar trends were obtained when analysing the corresponding infected co-cultures from the VSA ( Figure 3B ) . In HIV-192RW024 infected target cells , FL8 specific CTL reduced infection from 9% to <1% whilst a reduction was also observed in HIV-192UG029 infected cells at all three E∶T ratios tested , which were significantly different ( p<0 . 01 ) from control wells of infected cells in the absence of CTL . The clade B virus HIV-1MN was also efficiently suppressed by FL8 and EI8 CTL . However , FL8-specific CTL exerted no significant antiviral efficacy against HIV-1HXB2 or HIV-189 . 6 infected target cells , suggestive of impaired intracellular FL8 epitope processing and presentation . Similar result was observed while using HIV-1HXB2 and MN infected HLA B8+ PBMCs as target cells in VSA ( Figure S2 ) . Differential CTL antiviral activity to the FL8 epitope between viral isolates was also observed in the Live Virus Elispots ( LVE ) assay ( Figure 3C ) . FL8-specific CTL demonstrated a strong IFNγ response with an average magnitude of 46–101 SFUs at 96 hours post-infection against endogenously presented FL8 peptide by cells infected with clade A isolates HIV-192UG029 and HIV-192RW024 respectively . In contrast , a varied FL8-specific CTL response was observed against clade B infected cells , with a mean of 65 SFUs recorded for HIV-1MN , but 0 SFUs at 96 hrs post-infection for both HIV-1HXB2 and HIV-189 . 6 . Further LVE studies were conducted at 6 and 12 hours post-infection to assess ‘early’ CTL recognition of endogenous Nef and Gag epitopes but no IFNγ responses were observed . All high titre viruses were tested in parallel with the same panel of three highly avid FL8-specific clones and four control EI8-specific clones with the same results . All five viruses were CXCR4-tropic , able to replicate efficiently and were pathogenic , as confirmed by syncytia formation with 6–20% of C8166 cells ( p24+/CD4+ ) infected by Day 4 in the absence of CTL . A low MOI was chosen for physiologic relevance to an in vivo setting , with <20% infection observed , comparable to other SIV and HIV- based viral suppression assays [21] , [29] . However , similar results were also obtained at a higher MOI , so viral titre does not appear to impact upon the patterns observed ( as observed in figure 2B ) . Furthermore , since HIV-1HXB2 , MN and 93RW024 contained a conserved HLA-A24 restricted RW8 Nef epitope , we were also able to utilise CTL clones specific for RW8 to confirm that these viruses were able to present Nef derived peptides to CTL and were not simply Nef-deficient ( Figure S1 ) . Interestingly , although there is a general trend towards higher p24 at the lower E∶T ratios with the control gag clones , and with the Nef clones to a lesser extent , overall the E∶T ratio appears to have limited influence in the VSA . This is in accordance with the results from a similar assay [30] and may be attributable to the use of CTL clones that were pre-selected for their high avidity and ability to suppress viral replication at low E∶T ratios ( in comparison to polyclonal populations with differing avidity ) . It should also be noted that our control Gag EI8 clones specific to EIYKRWII were chosen as an internal control as they had similar functional avidity for the clade A intra-epitopic variant DI8 ( DIYKRWII ) when tested with peptides in the IFNγ ELISPOT and chromium release assays ( data not shown ) . However , these Gag-specific CTL were unable to respond to DI8 targets within the clade A viruses HIV-193RW024 and HIV-192UG029 in Figure 3 , due to more efficient immunoproteasomal processing for EI8 than DI8 ( data not shown ) ; however they still acted as good internal controls for clade B infections . Overall , our data from the viral suppression assays and live virus ELISPOTS clearly demonstrate that efficient CTL antiviral activity is heavily dependent on the infecting HIV-1 isolate , which cannot be accurately predicted from epitope sequence conservation alone or by clade-grouping . We also infected HLA-B8 B-cells with recombinant vaccinia virus constructs expressing HIV-1 Nef ( rVV-Nef ) from eight viral isolates , each from a different group M clade ( for simplicity abbreviated by their clade reference ) : HIV-192UG037 . 1 ( A ) , HIV-1MN ( B ) , HIV-196ZM651 ( C ) , HIV-194UG114 . 1 ( D ) , HIV-1CM235-32 ( AE ) , HIV-193BR020 ( F ) , HIV-192NG83 . 2 ( G ) and HIV-190CF056 ( H ) . The rVV-Nef expressing cells were co-cultured with FL8-specific CTL in two assays; a 51Cr-release assay and IFNγ ELISPOT . The results from both assays and the pre-determined rVV-Nef sequences are summarised in Table 1 . In the 51Cr-release assay , striking differences were observed in the level of CTL-induced lysis of cells infected with rVV-Nef from different viral isolates , which could be categorised into three distinct groups based upon their mean percentage lysis . Only in cells infected with rVV-Nef-D was the proportion of lysed cells ( 70% ) comparable to peptide-pulsed targets ( 75% ) indicative of optimal abundance of endogenously derived FL8 for efficient recognition and lysis ( Figure 4B ) . Subsequent analysis confirmed that there was no significant difference between rVVNef-D and the control in an ANOVA ( p>0 . 05 ) . In contrast , infection with viral isolates B , C , F and H resulted in reduced CTL recognition , with only 20–70% lysis compared to the peptide control , whilst no significant CTL killing ( >20% lysis ) was observed for isolates A , AE and G , suggestive of impaired or abolished FL8 processing and presentation . Concordant results were observed in the IFNγ ELISPOT ( Figure 4A ) . Only isolate D elicited levels of IFNγ release ( 132 SFCs ) similar to the peptide control ( 135 SFCs ) . In contrast , infection with B , C , F and H showed intermediate levels of 67 , 23 , 57 and 34 SFCs respectively and less than 10 SFCs for A , AE and G . In total , seven rVV-Nefs elicited CTL antiviral responses that were significantly different ( p<0 . 01 or p<0 . 001 ) from the CTL response to the FL8 peptide control . These data support the isolate-specific nature of CTL antiviral efficacy , despite epitope conservation between viruses . To ensure similar levels of Nef protein expression in these assays , rVV infection used the same plaque forming unit ( pfu ) titre and a CTL clone specific to a conserved Nef epitope HLA-A3 QK10 was utilised as an internal control . The strong CTL response to endogenous QK10 peptide presented by rVV-A , AE and G in the chromium release assay confirmed that these rVV efficiently infected the BCL targets and were not Nef-defective ( Figure S1 ) . To verify that the striking differences in CTL antiviral activity to a conserved epitope observed in Figure 3 and 4 were not attributable to differing Nef-induced down-regulation of surface MHC class I expression by the various virus isolates [31] , [32] , we infected target cells with virus and assessed HLA-B8 surface expression using a HLA-B8 specific monoclonal antibody . We compared HLA-B8 expression in HLA-B8+ targets infected with our five high titre HIV-1 isolates ( 48 hours post-infection ) and eight rVV-Nef constructs ( 4 hours post-infection ) , and also utilised three HLA-B8− negative targets as a negative control . Our results showed that although down-regulation of HLA-B8 surface expression was observed in the high titre virus infected cells when compared with uninfected or vaccinia-infected targets , there was no marked variation in HLA-B8 between the five high titre HIV-infected target cells ( Figure 5A ) . Furthermore , there was no marked variation in HLA-B8 expression between the eight rVV-infected target cells ( Figure 5B ) . However , we acknowledge that the direct measurement of HLA-B8 expression may vary between experiments . Whilst Lewis et al controlled for this with internal standards [33] , Nef-deleted viruses were not available for comparison with our wild-type viruses . Alternatively , we utilised B8-restricted Gag CTL controls in our viral suppression assays and live virus Elispots , since any effect of Nef mediated down-regulation of HLA-B8 ought to apply equally to this epitope and not be restricted to Nef epitopes . The CTL recognition of the Gag epitope was the same for the viruses tested , even when Nef recognition differed markedly . Together , these data therefore demonstrate that Nef-mediated down-regulation of HLA-B8 is not responsible for diminished CTL recognition of the conserved FL8 epitope on cells infected with HIV-1HXB2 and 89 . 6 isolates , or with rVV-Nef constructs A , AE and G . From the above results it seemed likely that differences in antigen processing might explain the failure of CTL to recognise cells infected with some of the virus isolates and rVV-Nef . Intracellular peptide processing is a multi-step pathway and amino acid variations within epitopes and their flanking regions can affect any step of this complex cascade [34] , which includes proteasomal or immunoproteasomal cleavage , TAP mediated transport to the ER lumen and N-terminal trimming by aminopeptidases . Alternatively , other proteases such as tripeptidyl peptidase II ( TPPII ) may act independently or in combination with the proteasome to generate epitope precursors . TPPII is known to play a pivotal role in the generation of the immunodominant HLA-A3 Nef QK10 epitope [35] , which is located 8 amino acid residues downstream from the HLA-B8 Nef FL8 epitope . To test the importance of different components of the antigen processing machinery on the generation and presentation of FL8 on the infected-cell surface , we chose two viruses that elicited a dominant FL8 response; high titre HIV-192UG029 and rVV-D ( HIV-194UG114 ) . Infected target cells were treated with inhibitors to block proteasome and immunoproteasome ( Epoxomicin ) , aminopeptidase activity ( Bestatin ) and tripeptidyl-peptidase II activity ( AAF-CMK ) at appropriate concentrations . Controls included inhibitor treated cells pulsed with peptide ( both HIV-infected and uninfected ) to ensure that the read out was not altered by inhibitor-induced cell death . The addition of epoxomicin to high titre HIV-192UG029 infected cells in a modified Live Virus ELISPOT intracellular antigen processing inhibition assay ( IAPIA ) completely abolished recognition of FL8 at 10 µM ( Figure 6A ) . In contrast , addition of bestatin and AAF-CMK , even at high concentrations of 10–100 µM , had little impact on CTL recognition . A similar pattern was observed when the same three inhibitors were added to BCL infected with rVV-D ( HIV-194UG114 ) in modified chromium release IAPIA ( Figure 6B ) . The addition of epoxomicin at 10 µM and 1 µM reduced CTL lysis by 82% and 53% respectively , whilst addition of Bestatin and AAF-CMK at 10 µM had minimal effect . The marked difference in CTL lysis at 100 µM is likely to represent partial inhibition of proteasomal activity when used at high concentrations . Overall , in contrast to the Nef QK10 epitope [35] , the result with AAF-CMK showed that TPPII is not involved in the FL8 processing pathway for the two viruses tested . Together , these inhibition assays clearly demonstrate that proteasomes and immunoproteasomes play a pivotal role in the endogenous processing of the HLA-B8 FL8 epitope which is required to initiate efficient CTL antiviral activity against HIV-192UG029 and HIV-194UG114 infected targets . Therefore , as a first key step in the processing pathway and in accordance with the antigen processing literature ( reviewed in [36] , [37] ) , the cleavage specificities of the immunoproteasomes and proteasomes in particular are likely to have a significant impact on epitope generation correlating with subsequent epitope abundance on the cell surface [38] . Since the FL8 epitope is conserved between all HIV-1 isolates tested in previous assays , we next investigated whether viral isolate-specific polymorphisms flanking FL8 could modulate the antigen processing efficiency of this epitope . We hypothesised that processing and production of the FL8 epitope may be more efficient in cells infected with clade A isolates HIV-192UG029 and HIV-193RW024 , and clade B HIV-1MN , whilst impaired antigen processing could account for poor CTL recognition of clade B HIV-1HXB2 and HIV-189 . 6 viruses in Figure 3 . Two 25-mer oligopeptides were therefore synthesised to span FL8 ( Nef90–97 ) and its flanking region ( Nef82–106 ) , one corresponding to HIV-1HXB2 and the other corresponding to HIV-192UG029 . The two oligopeptides differed in the flanking region by 3 amino acids including position 83 ( glycine/alanine ) , 85 ( valine/leucine ) , and 104 ( arginine/glutamine ) . Of these amino acid polymorphisms , the former two can be considered conservative variations , while replacement of arginine by glutamine leads to loss of a basic residue and replacement by an uncharged side chain . To test whether these residues affected antigen processing by the immunoproteasome , which is most commonly involved in CTL epitope generation , each oligopeptide was incubated with purified immuno-20S-proteasome ( i20S ) for 0 , 10 , 40 and 70 minutes . The peptide fragments resulting from immunoproteasomal digestion were identified using tandem mass spectrometry . Since immunoproteasomes and proteasomes rarely produce the exact 8–11mer peptide , but instead generate longer epitope pre-cursors that are often correctly cleaved at the C-terminus , we defined an ‘epitope precursor’ as a peptide fragment containing the intact FL8 epitope that is extended at the amino ( N ) terminus and carboxyl ( C ) terminus , and a ‘correct epitope pre-cursor’ as a peptide that is extended only at the N-terminus and correctly cleaved at the C-terminus . Both oligopeptides showed strikingly different digestion patterns ( Figure 7 ) when digested at these time points . Oligopeptide HIV-192UG029 produced three dominant FL8-containing precursor peptides , of which one was a correct C-terminally cleaved peptide ( KGAVDLSHFLKEKGGL ) ( Figure 8A ) . Although mass spectrometry is only semi-quantitative , UPLC-MSE analysis at each time point showed that this C-terminally cleaved peptide was present as early as 10 minutes post i20S digestion and increased in quantity at 40 and 70 minutes post digestion ( Figure 8B ) . In contrast , digestions of oligopeptide HIV-1HXB2 produced several FL8 containing precursor peptides , but these contained substantial N- and C-terminal extensions , and were only 3 amino acids shorter at most than the original 25-mer oligopeptide . None of the precursor peptides generated from digestion of oligopeptide HIV-1HXB2 contained the correct C-terminal cleavage for FL8 . Because of the large number of viral isolate-specific polymorphisms in the different viruses tested , it is difficult to define accurately which amino acid polymorphism ( s ) in the flanking regions present in different HIV-1 isolates critically impair immunoproteasomal cleavage of FL8 . We speculated that one or several amino acid polymorphisms may act in tandem to alter cleavage patterns . Since most of the flanking amino acid differences were in the N-terminal part of the FL8 epitope , we “swapped” the N-terminal sequence of the FL8 epitope region from the HIV-192UG029 virus that is recognized by CTL with the sequence derived from the rVV-Nef-A ( HIV-192UG037 ) virus that is not recognized ( see Table 1 and Figure 9A ) to create a “hybrid” FL8 epitope precursor . When digested by proteasomal proteolysis and analysed by UPLC-MSE , no correct C-terminal cleavage was observed , indicating that the N-terminal region of the FL8 epitope is critical for this cleavage step . We therefore further examined the polymorphisms observed in the PCR-derived sequence ‘tracked’ to either superior or abolished antiviral efficacy in our assays ( Table 1 ) . Interestingly , polymorphisms in our live viral sequences did not track to CTL antiviral efficacy . Conversely , we noted that the presence of a phenylalanine at position 89 immediately adjacent to the N-terminus of the FL8 epitope and an additional phenylalanine at position 85 , correlated with abolished CTL antiviral efficacy in three rVV isolates; A , AE and G . Variation of N- and C-terminal epitope flanking residues can influence the length and nature of epitope precursors that are generated [39] [40] [41] and the immunoproteasome is known to have a preference for cleaving after hydrophobic residues [42] . Thus , this motif is likely to have a pronounced impact on FL8 cleavage patterns . We therefore repeated the in vitro proteasomal digestion assay to test whether this phenylalanine motif may play a key role in modulating epitope processing . We chose two previously studied viruses that exhibited contrasting epitope generation in digests ( associated with contrasting CTL antiviral efficacy ) ; HIV-192UG029 and rVV-Nef-A ( HIV-192UG037 ) . Previous digestion of the HIV-192UG029 derived precursor peptide generated a correct C-terminally cleaved FL8 pre-cursor that was present in large quantities . In contrast , digestion of the HIV-192UG037 derived precursor peptide containing this phenylalanine motif generated no FL8-containing pre-cursors at any of the time points tested ( Figure 8 ) . Therefore , we took the HIV-192UG029 sequence ( KGAVDLSHFLKEKGGLDGLIYSRKR ) and designed two new 25-mer oligopeptides; the first in which we substituted the basic histidine residue with a large hydrophobic phenylalanine at position 89 immediately adjacent to FL8 ( HIV-192UG029+1: KGAVDLSFFLKEKGGLDGLIYSRKR ) and an additional substitution at position 85 ( HIV-192UG029+2: KGAFDLSFFLKEKGGLDGLIYSRKR ) in which hydrophobic valine was replaced with a hydrophobic phenylalanine . The original oligopeptide and two new oligopeptides were then digested with immunoproteasome for 0 , 10 , 40 , and 70 minutes and peptide fragments were identified using tandem mass spectrometry ( Figure 9 ) . The insertion of the phenylalanine motif markedly altered the pattern of immunoproteasomal cleavage . Whilst the digestion of the original HIV-192UG029 sequence generated the correct C-terminally cleaved FL8 pre-cursor in high quantity , the +1 and +2 oligopeptide sequences did not generate any C-terminally cleaved pre-cursor peptides . For the +1 oligopeptide , the breadth of intra-epitope cleavage was enhanced in comparison to the original peptide ( in which intra-epitope cleavage site was predominantly focused after the F ) . For the +2 oligopeptide , despite repeat analyses , only three fragments were identified in addition to the mother peptide . This is indicative that the motif markedly alters both cleavage patterns and the overall quantity of pre-cursors generated . Collectively , these data demonstrate that the phenylalanine motif in the N-terminal sequence can diminish epitope processing , and highlight that even a single virus-isolate polymorphism ( H89P ) in the flanking region can substantially alter epitope production . We have shown that striking differences exist in CTL antiviral efficacy to a conserved epitope shared between diverse viral isolates , and that from the panel of HIV-1 isolates only 23% were recognized by CTL . In experiments with five high titre HIV-laboratory strains in vitro , FL8-specific CTL demonstrated efficient viral suppression and a strong IFNγ response against endogenous FL8 peptide presented by cells infected with clade isolate HIV-193RW024 , and clade B HIV-1MN , and sub-optimal antiviral activity to clade A isolate HIV-192UG029 . However , no CTL antiviral activity was detected against a further two clade B strains , HIV-1HXB2 and HIV-189 . 6 . We also utilised a recombinant vaccinia virus system expressing Nef from eight viral isolates , each from a different group M clade ( A–H ) , to evaluate whether FL8 epitope-specific CTL recognition and antiviral activity differs between viral isolates . Again , we observed significant differences in CTL lysis and IFNγ secretion , with correct epitope processing and CTL recognition completely abolished in three viral isolates ( A , AE and G ) and impaired in a further four isolates ( B , C , F , H ) . Collectively , we demonstrate that a surprisingly large proportion ( 77% ) of Nef proteins with a conserved FL8 epitope , expressed by HIV-1 isolates or as recombinant vaccinia-Nef were suboptimally recognised by FL8- specific T-cells , despite the presence of the same epitope . Both endogenous expression systems clearly show that the antiviral efficacy of CD8+ T cells to an invariant epitope is heavily dependent on the infecting viral isolate , which occurs independently of clade-grouping . We have shown that inter-clade and intra-clade polymorphisms in the FL8 flanking region modulate epitope processing by the immunoproteasome , to enhance or impair epitope generation , which is associated with altered CTL recognition and antiviral activity in the infecting HIV-1 strains tested . Furthermore , we demonstrate that the ‘swapping’ of the flanking regions between viruses that are recognised and not recognised can appreciably modulate epitope processing , and we identify a N-terminal phenylalanine motif that can diminish epitope generation . When modified , this can help to optimize CTL responses elicited by appropriately designed vaccine vectors containing critical flanking sequences in addition to the epitope that enhance epitope antigenicity . Although it has been previously shown that flanking residues can impact epitope presentation by MHC , also for HIV-1 derived epitopes [19] , [23] such studies have focused on identifying amino acid variation predominantly within a single viral isolate . In contrast , our study evaluated the reason for variable CTL responses despite absolute conservation of a HIV Nef epitope shared by a panel of 13 HIV-1 whole viral isolates and recombinant vaccinia-Nef . Strikingly , our results strongly suggest that the impact of viral variation on efficient antigen processing has been seriously underestimated , and this has been reinforced by a reliance on peptide-based assays to measure T-cell responses in natural history and vaccine studies . Due to the current focus on developing HIV vaccines that elicit T-cells to conserved regions of the viral genome , this is an important finding that has implications for both vaccine design and evaluation of vaccine efficacy . Our results emphasise that the antiviral efficacy and cross-reactive potential of CD8+ T-cells should be assessed by their ability against cells infected with virus , and it cannot be accurately predicted solely on the basis of epitope conservation or based on the results of CTL assays using exogenously-loaded peptides . Since infecting viral isolates can exploit impaired antigen processing and presentation to hide from immune surveillance , epitope conservation between viruses may not accurately predict the cross-reactive potential and antiviral efficacy of CTL . The lack of epitope presentation in particular viral backbones may be one of the reasons that high levels of CTL elicited in people who become super-infected , either after stable HIV-1 infection [43] or after vaccination [44] , appear to be ineffective even when the infecting virus contains the identical epitope . Our data suggest that potential immunogens for cross-clade vaccine design should not be based solely upon invariant epitopes , but should focus upon conserved regions that include similar epitope flanking regions which have been tested via functional assays for endogenous presentation prior to use in vaccine constructs . Similarly , it will be important to measure the ability of vaccine elicited T cells to recognize a range of different virus isolates in whole-virus assays as well as the more conventional peptide-based studies . The use of whole-virus assays is not only informative on the efficacy of antigen processing and presentation , but can also be adapted to detect additional changes in CTL antiviral efficacy attributable to viral functions; which include differential protein expression kinetics , variation in protein expression levels by Tat , and the down-regulation of MHC class I by Nef [31] . For future vaccine design , it may also be necessary to reassess the relative value of utilising particular highly conserved and immunogenic regions of the virus , such as this FL8 epitope or the Gag p24 region [10] , [45] , since sequence conservation may mask a lack of presentation that undermines the efficacy of vaccine-induced responses . Interestingly , since Gag is much less variable between HIV-1 primary isolates and also during the natural course of infection due to fitness costs associated with changes in sequence [46] , the probability of gag epitopes being processed more efficiently among diverse viral isolates may be higher than for epitopes from other HIV proteins . This could potentially contribute to the efficacy of Gag-specific CTL observed in chronic HIV-infected patients [47] . Inter-clade and intra-clade virus specific polymorphims may also shape immunodominance as the targeting of HIV-specific CTL epitopes in a hierarchical pattern is sensitive to alterations in antigen processing and presentation [48] . FL8 is a well characterised immunodominant epitope during acute infection , with FL8-specific T-cells detected in over 70% of HIV-infected HLA-B8 participants tested [49] . Yet , immunodominance does not necessarily mean that the FL8 epitope is optimally processed in the majority of viral strains within each infected individual . The low antigen load generated by a viral isolate with sub-optimal or impaired epitope processing may be sufficient to prime T-cells , but insufficient to trigger cytotoxic killing in vivo , especially by low avidity CTL . In addition to the impact of viral isolate-specific polymorphisms , the high mutation rate of HIV may give rise to potential intra-epitopic and epitope-flanking escape mutations during acute infection [50] that subsequently diminish processing . Therefore , the T cells primed by the infecting virus might not be able to control mutated viruses that emerge during the course of infection . Also , strong immune pressure upon neighbouring epitopes such as B57-KF9 Nef in the first few weeks post-infection [13] or on epitopes overlapping with FL8 , such as B60-KL9 [51] , A2-FL11 , A3-AK9 , A3-DK9 , A24-HL9 , may alter the generation and consequent immunogenicity of FL8 . Furthermore , in the absence of sufficient antigen , where cells are infected by non-infectious virus or epitope processing is abolished , dendritic cells may acquire these HIV-infected targets and successfully cross-present HIV antigens to prime the expansion of HIV-specific T-cells [52] . Collectively , these mechanisms may explain why FL8-specific T-cells are immunodominant in acute infection , but unlike other responses , FL8 is an outlier that does not correlate with viral control at set point [53] . In our study , we have used rVV in addition to whole HIV to demonstrate our major conclusion that the anti-viral efficacy of B8 FL8 restricted responses are heavily dependent on viral isolates . Since the backbones for rVV and whole virus are different , it is possibile that rVV may not fully represent whole virus , and vaccinia expression levels and HIV expression levels are not equivalent . Ideally , direct HIV to HIV comparison would be very useful for the hypothesis . However , it is worth noting that the ELISPOT by using live virus ( MN ) infected cells or Vaccinia virus ( MN ) both give detectable FL8-specific responses which is indicative that the two assays are comparable . Processing data by using the sequences derived from rVV further confirmed the initial observation utilising rVV in IFNγ ELISPOT and lytic assays . Interestingly , analysis of viral amino acid sequence ( arising from PCR amplification and sequencing of our infected targets ) consistently differed from the expected Los Alamos National Laboratory ( LANL ) HIV Sequence Database for each of the five viruses , although the FL8 epitope remained unchanged . Whilst some of the viruses used in our assays were molecular clones , others were clinical isolate swarms , and therefore we could not exclude the possibility that there is heterogeneity in the FL8 epitope or flanking sequences . Yet consistent sequencing results were obtained when multiple PCR were performed , suggesting sequences showed in table 1 were dominant sequences . Considering that HIV undergoes an average of 1 mutation per genome per replicative cycle due to the error-prone nature of the reverse transcriptase , some mutations may become ‘fixed’ or continue to evolve in in vitro assays in which CD4+ T-cells are infected with HIV-1 laboratory strains , even under no selection pressure . Due to this high mutation rate , the LANL database represents a ‘snap shot’ of viral genomes and proteomes at a single time point , and therefore repeated sequencing is necessary . Our in vitro digests utilised immunoproteasomes as they are typically induced within the first day of viral infection by exposure to IFNγ or TNFα [54] [55] . Although immunoproteasomes can generate a different spectrum of epitopes from standard 20S proteasomes , the effects are thought to be relatively subtle , with more pronounced quantitative rather than qualitative differences [56] . Current estimates suggest that only 15% of peptides are the appropriate 8–11mer length for MHC-loading after proteasomal or immunoproteasomal digestion [57] , therefore the generation of extended FL8 precursors is not unusual . The N-terminal extensions are usually trimmed by aminopeptidases in the cytosol or after TAP-mediated transfer into the ER [58] , although successful CTL recognition of an endogenously derived 15-mer N-terminally-extended HIV-1 gp160 peptide was recently observed [59] . Since more than 99% of peptides are thought to be destroyed by cytosolic peptidases before encountering TAP [60] , a limitation to this technique is that the liberation of FL8-containing precursors in our in vitro assays cannot guarantee efficiency in subsequent processing and presentation steps of FL8 . However , the importance of the proteasome in liberating Nef peptides , and the strong correlation between the identities of extended epitopic precursors generated via in vitro proteasomal cleavage and naturally processed peptides acid-eluted from the surface of Nef-transfected cells , indicate that in vitro digests are a reliable tool [38] , [61] . These studies on Nef , together with our in vitro inhibition assays and immunoproteasomal digestions , strongly support the central role of the proteasome and immunoproteasome in altering peptide generation . In addition , since the specificity of TAP transport , ERAAP trimming and also HLA-binding affinity are dictated by internal epitope composition , particularly the C-terminal residue [37] , they are likely to be similar between viral strains with a conserved epitope ( post-proteasomal proteolysis ) . Therefore , these steps further along in the processing pathway are expected to have limited influence on the generation of the conserved FL8 epitope studied here , although this warrants further investigation . Interestingly , the overlapping HLA-B*60 KL9 epitope ( KEKGGLEGL ) is flanked by the amino acids FL at the N-terminus [49] , and the FL8 epitope ( FLKEKGGL ) is flanked by SH . These differences can be sufficient to determine a completely different outcome in their processing efficiency by the proteasome . In conclusion , our findings show that striking differences exist in the antiviral efficacy of CTL to an invariant epitope shared between viral isolates . Only a small proportion ( 23% ) of the HIV-1 Nef proteins , expressed by HIV-1 isolates or as recombinant vaccinia-Nef , elicited optimal FL8-specific CTL antiviral responses , whilst the majority demonstrated impaired or completely abolished epitope processing . We found that virus-specific polymorphisms in the flanking region to a conserved epitope substantially alters immunoproteasomal proteolysis , favouring , impairing or completely abolishing epitope generation , which is correlated with the efficiency of CTL antiviral activity in vitro and may play an important role in determining epitope immunogenicity and immunodominance required to prime T-cells in vivo . CTL antiviral efficacy is heavily dependent on the infecting viral strain; this occurs independently of clade-grouping . This is of major relevance for the design of future vaccines protective against genetically diverse strains of HIV-1 , and should be carefully evaluated when assessing the effectiveness of vaccine-induced T-cell responses . HIV-1 viral isolates 92UG029 , 93RW024 ( clade A ) were obtained from The UNAIDS Network for HIV Isolation and Characterization , and the DAIDS , NIAID . HIV-1 laboratory strains IIIB , MN , 89 . 6 ( clade B ) were obtained from the AIDS Research and Reference Reagent Program , Division of AIDS , NIAID , NIH , in addition to eight recombinant vaccinia viruses expressing Nef protein from different group M clades ( specified within brackets ) : HIV-192UG037 . 1 ( A ) , HIV-1MN ( B ) , HIV-196ZM651 ( C ) , HIV-194UG114 . 1 ( D ) , HIV-1CM235-32 ( AE ) , HIV-193BR020 ( F ) , HIV-192NG83 . 2 ( G ) and HIV-190CF056 ( H ) . Details and acknowledgment of each reagent are listed in supplemental material . CTL clones specific for HLA-B8 epitopes were generated by limiting dilution from the PBMCs of HIV-infected patients responding to the FL8 and EI8 epitope and maintained as described in Dong et al 2004 [62] . The Human T cell leukaemia C8166 line stably expressing CD4+ and HLA-B8 was maintained in R10 media ( RPMI 1640 with 100 U/ml of penicillin , 100 U/ml of streptomycin , 2 mM L-glutamine and 10% heat-inactivated fetal calf serum from Sigma-Aldrich ) . CTL lysis assays were performed using standard 51Chromium release assays as described elsewhere [63] . Slight modifications were made to test infection with recombinant vaccinia viruses ( rVV ) . In brief , a B-cell line ( BCL ) expressing HLA-B8 was incubated with 150 µCi of chromium for one hour at 37°C/5% CO2 , then washed extensively prior to infection with rVV encoding HIV-nef genes at 2×106 pfu/million cells for one hour in serum-free RPMI . Infected cells were allowed to recover for a further two hours in R10 . The labeled target cells were transferred to a round bottom 96-well plate and co-cultured with HLA-matched CTL clones in triplicate at an E∶T ratio of 40000∶5000 per well for a further 4 hours at 37°C . Supernatants were harvested and radioactivity was assessed using a beta-plate counter ( Wallac ) . Spontaneous chromium release was determined by analysing the release from target cells incubated in R10 media and maximum release of incorporated chromium was obtained from target cells treated with 5% Triton X-100 detergent . Specific lysis was calculated = 100 × ( experimental lysis − spontaneous lysis ) / ( maximum lysis − spontaneous lysis ) . Standard Human IFNγ ELISPOT assays were performed as described elsewhere [62] . Slight modifications were made when testing recombinant vaccinia viruses ( rVV ) . In brief , a HLA-B08 expressing BCL was infected with rVV encoding HIV-nef genes at 2×106 pfu/million cells for one hour in 1 ml RPMI and recovered as described above . Subsequently , the infected target cells were washed twice and added to pre-coated IFNγ ELISPOT plates with a HLA-matched CTL clone at an E∶T ratio of 400∶20 , 000 , in triplicate , in a final volume of 100 µ/well . Control wells included uninfected BCL ( negative ) and peptide pulsed uninfected BCL ( positive ) plus peptide pulsed rVV-infected BCL co-cultured with CTL . Assays were incubated for 6 hours at 37°C/5% CO2 and developed as normal . Differing HIV-1 strains ( 2-fold TCID50 ) were used to infect C8166 cells . Infected cells were washed twice , split with at least 1×106 cells/T25 flask cultured in a total volume of 2 ml R10 for a period of 6 , 24 , 48 , 72 and 96 hours post-infection at 37°C/5% CO2 . At each time point , cells were washed , counted and co-cultured with the panel of HLA-matched CTL clones in triplicate at one E∶T ratio of 400∶20000 on the pre-coated interferon gamma IFNγ ELISPOT plates at a final volume of 100 µl/well . Negative controls included the individual CTL clones co-cultured with uninfected target cell line in triplicate , and positive control included each CTL clone co-cultured with uninfected target cells pulsed with 2 µM of specific peptide . ELISPOT plates were incubated for 6 hrs at 37°C/5% CO2 and subsequently washed and developed as described previously . Spot forming units ( SFUs ) were counted using the ELISPOT reader system AID ELIspot 4 . 0 . Differing high titre HIV-1 strains ( TCID50 ) were used to infect the C8166 cell pellets during a 90 minute incubation at 37°C/5% CO2 and were subsequently washed ( 2× ) to remove free virus . 5×104 infected cells were co-cultured with HLA-matched HIV-1 specific CTL clones at differing E∶T ratios of 1∶1 , 1∶2 , 1∶4 , 1∶8 & 1∶16 in H10-IL2 ( 200 U/ml ) on a flat bottom 96-well plate , in a final volume of 200 µl per well , at 37°C for 4 days . Each condition was performed in triplicate , including one HLA-mismatched clone as a negative control and virus-infected cells in the absence of CTL as a positive control . Suppression of infected cells by CTL on Day 4 was directly monitored using an intracellular anti-p24 gag mAb ( KC57-RD1 , Beckman Coulter ) . The extracellular p24 content in the supernatant was also assayed by quantitative p24 antigen ELISA ( Immunodiagnostics ) in accordance with the manufacturer's protocol . Proviral DNA was isolated for each high titre virus from control wells containing only HIV-infected C8166 cells in the viral suppression assay using the PureGene DNA isolation Kit ( Gentra Systems ) as per the manufacturer's instructions . Nef was amplified from proviral DNA by nested PCR . The nested PCR amplification was carried out in a total volume of 50 µl using the Taq polymerase PCR reaction kit ( QIAGEN ) and AccuPrime Taq DNA Polymerase High Fidelity kit ( Invitrogen ) , according to the manufacturer's instructions . GTA GCT GAG GGG ACA GAT AG and TGC TAG AGA TTT TCC ACA C . Initial denaturing at 94°C for 120 seconds was followed by 30 cycles of denaturing at 94°C for 30 seconds , annealing at 52°C for 30 seconds and extension at 72°C for 90 seconds . A final extension at 72°C was run for 300 seconds . The internal pair of primers used was GAA GAA TAA GAC AGG GCT and AGG CTC AGA TCT GGT CTA A . Initial denaturing at 94°C for 120 seconds was followed by 30 cycles of denaturing at 94°C for 30 seconds , annealing at 56°C for 30 seconds and extension at 72°C for 90 seconds , with a final extension at 72°C was run for 300 seconds . The PCR products were checked for size on a 1% agarose gel and sequencing was performed in the MRC HIU Sequencing Facility , Weatherall Institute of Molecular Medicine . 1 million C8166 were infected separately with the five high titre viruses ( TCID50 ) for 48 hours whilst 1 million BCL were infected separately with the eight rVV ( 2×106 pfu/million targets ) for four hours . Both sets of virus-infected cells were stained with a biotin-conjugated anti-HLA-Class I B8 monoclonal antibody ( AB33716 , Abcam ) washed twice , and fixed with 1% paraformaldehyde . HLA-B8− negative cells were used to set the negative quadrants and to confirm negligible cross-reactivity of the antibody to other HLA-alleles expressed on target cells . Additional controls included the biotin-conjugated antibody in the presence and absence of streptavidin and uninfected cells . At least 105 live cells per infection were counted using a Cyan flow cytometer and analysed with FlowJo . The C8166 cells were pre-treated with proteasome inhibitor Epoxomicin , TPP II inhibitor AAF-CMK and aminopeptidase inhibitor Bestatin ( all Biomol ) at varying concentrations at 37°C for one hour prior . The cells were then infected with HIV-192UG029 at TCID50 for 90 minutes at 37°C/5% CO2 . Two controls comprising pre-treated cells not infected with HIV ( to determine the affect of the inhibitor on cell survival in the absence of HIV-induced cell death ) , and HIV infected untreated cells ( to determine the maximum response elicited in the absence of inhibitor treatment ) were used . All cells were washed after 90 minutes and transferred to a T25 flask containing R10 medium for 36 hrs before being used as targets in a modified Live Virus Elispot with a FL8-specific CTL clone at an E∶T ratio of 400∶20000 as described previously . Slight modifications were made to test infection with recombinant vaccinia virus ( rVV ) expressing Nef from HIV-1 clade D isolate HIV-194UG114 . In brief , BCL expressing HLA-B08 was incubated with 150 µCi of chromium for one hour at 37°C/5% CO2 and then washed extensively . The BCL were pre-treated with proteasome inhibitor Epoxomicin , TPPII inhibitor AAF-CMK and aminopeptidase inhibitor Bestatin ( all Biomol ) at varying concentrations in serum-free RPMI for one hour at 37°C . The cells were then infected with rVV-Nef-D at 2×106 pfu/million targets for one hour in serum-free RPMI , before recovery in R10 for a further one hour . The labelled target cells were then transferred to a round bottom 96-well plate and co-cultured with a HLA-matched CTL clone in triplicate at an E∶T ratio of 40000∶5000 per well for a further 4 hours at 37°C . Supernatants were harvested and radioactivity was assessed as described previously . Four 25-mer oligopeptides were designed to cover the FL8 epitope and flanking regions corresponding to the amino acid sequence as determined by sequencing for HIV-1HXB2IIIB , HIV-192UG029 and HIV-192UG037 and hybrid viral sequences . Extended peptides KAALDLSHFLKEKGGLDGLIYSQKR ( HIV-1HXB2IIIB ) , KGAVDLSHFLKEKGGLDGLIYSRKR ( HIV-192UG029 ) , KAAFDLGFFLKEKGGLDGLIYSKKR ( HIV-192UG037 ) , KAAFDLGFFLKEKGGLDGLIYSRKR ( HIV-192037N-029C ) , KGAVDLSFFLKEKGGLDGLIYSRKR ) ( HIV-192UG029+1 ) , KGAFDLSFFLKEKGGLDGLIYSRKR ) ( HIV-192UG029+2 ) : containing the HLA-B8 restricted FL8 epitope FLKEKGGL ( indicated in bold ) were synthesized by solid-phase F-moc chemistry on an automated peptide synthesizer ( Advanced ChemTech ) and purified to >98% by reversed-phase HPLC . In vitro proteasomal digestion of synthetic 25mer oligopeptides was conducted using immuno-20S ( i20S ) proteasome purchased from Biomol International , essentially as described [64] . For each oligopeptide , 1 µg of immuno-20S proteasome was incubated with 10 µg of peptide in a final volume of 100 µL of 20 mM Hepes , pH 7 . 8 , 2 mM magnesium acetate and 2 mM dithiothreitol and incubated at 37°C/5% CO2 . Sample aliquots were taken at several time points ( 0 , 10 , 40 , 70 minutes ) and reactions were terminated by the addition of 0 . 1 volume of formic acid ( FA ) . Control reactions without the i20S-proteasome were analysed to evaluate non-specific peptide degradation . The i20S-proteasome was pelleted by ultracentrifugation at 100000 g for 5 hours , and supernatants containing digested peptides were analysed using nano UPLC-high/low collision energy switching MS ( MSE ) as described [65] . In brief , the peptide digests were subjected to chromatographic separation by a NanoAcquity UPLC system coupled to a QTof premier tandem mass spectrometer ( Waters , Milford , MA , USA ) . For peptide precursor and fragment identification and simultaneous quantification , the instrument was run in MSE mode . Peptide peaks corresponding to the original undigested and shorter peptide fragments resulting from immunoproteasomal digestion were identified using a MASCOT search engine ( version 2 . 2 ) and a custom made database containing HIV-peptide sequences . Quantitative information was obtained from extracted ion chromatograms using MassLynx 4 . 1 software .
One of the greatest challenges to developing an effective HIV vaccine is the ability of HIV to rapidly alter its viral sequence . Such variation in viral sequence enables the virus to frequently evade recognition by the host immune system . To counteract this problem , there has been increasing interest in developing HIV vaccines that target T-cell responses to the regions of the virus that are highly conserved between strains of HIV . However , previous studies have focused on identifying amino acid variation predominantly within a single viral isolate , or have focused on classical within-epitope escape mutation . Our study assessed T-cell recognition of a conserved epitope shared by a total of 13 HIV strains . Strikingly , we show that only a small proportion of the viral strains were effectively recognised and targeted by the T-cells . In contrast , differences in amino acid sequence in the region flanking the epitope impaired the intracellular processing and presentation of epitope in the majority of HIV strains tested . Thus , our findings highlight that a large proportion of HIV strains may evade epitope-specific T-cell recognition despite absolute epitope conservation . This has important implications for both vaccine design and evaluation of vaccine efficacy .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "immunology/antigen", "processing", "and", "recognition", "infectious", "diseases/hiv", "infection", "and", "aids", "virology/immune", "evasion", "immunology/immunity", "to", "infections", "virology/host", "antiviral", "responses" ]
2011
The Antiviral Efficacy of HIV-Specific CD8+ T-Cells to a Conserved Epitope Is Heavily Dependent on the Infecting HIV-1 Isolate
Understanding the principles underlying the plasticity of signal transduction networks is fundamental to decipher the functioning of living cells . In Myxococcus xanthus , a particular chemosensory system ( Frz ) coordinates the activity of two separate motility systems ( the A- and S-motility systems ) , promoting multicellular development . This unusual structure asks how signal is transduced in a branched signal transduction pathway . Using combined evolution-guided and single cell approaches , we successfully uncoupled the regulations and showed that the A-motility regulation system branched-off an existing signaling system that initially only controlled S-motility . Pathway branching emerged in part following a gene duplication event and changes in the circuit structure increasing the signaling efficiency . In the evolved pathway , the Frz histidine kinase generates a steep biphasic response to increasing external stimulations , which is essential for signal partitioning to the motility systems . We further show that this behavior results from the action of two accessory response regulator proteins that act independently to filter and amplify signals from the upstream kinase . Thus , signal amplification loops may underlie the emergence of new connectivity in signal transduction pathways . In living cells , adaptation to rapid changes in environmental conditions requires coordinated rearrangements of basic cellular processes to adjust the cellular homeostasis to the new conditions . In general , receptor molecules sense environmental changes and translate them into a cellular response by phosphorylation of a downstream regulator . Because the cellular response must be integrated to various cellular processes , the phosphorylation cascade frequently involves a number of intermediates , allowing multiple regulation layers and branch points ( nodes ) in the regulatory circuit [1 , 2] . Thus , for a given pathway , identifying nodes and understanding how they participate in the regulation is of fundamental importance to elucidate how signals are integrated toward a cellular response . One possible approach to elucidate the underlying structure of a multi-component signaling pathway is to study its evolution because the diversification of signaling pathways is under strong selection pressure and signaling intermediates may have been selected in some organisms [3] . Consequently , some proteins that appear central to the regulation in a given genetic context may in fact be dispensable in a different context where their function is not required . For example , a protein that insulates a pathway from another will become dispensable if the secondary pathway is removed . Thus , tracking back the evolutionary history of signaling pathways can reveal core regulatory motifs and principles underlying the acquisition of additional regulations [4 , 5] . Bacteria are exceptional model systems for such studies because they are highly tractable experimentally and thousands of genome have been sequenced . In bacteria , signal transduction networks are frequently formed by so-called two-component systems . The core motif of a two-component system generally consists of a receptor , generally a membrane localized sensor histidine kinase ( HK ) and its cognate response regulator ( RR ) . Following activation by environmental signals , the HK uses ATP to autophosphorylate on a conserved histidine residue and the phosphoryl group is transferred to a conserved Asp residue of the RR protein , regulating a number of downstream processes , gene expression , secondary messenger synthesis or protein-protein interaction [6] . HK/RR pairs often form autonomous signal transduction systems [7] , but they are also modular and can be incorporated in more complex circuits , multiple phosphorelay systems and chemosensory-type systems [1 , 8] . In the enteric chemotaxis ( Che ) system , the HK ( CheA ) does not act directly as a sensor but resides in the cytosol where it is activated by a transmembrane Methyl-accepting protein ( Mcp ) via a coupling protein ( CheW ) . Following activation , CheA transfers a phosphoryl group to two RR domains; one of them , the CheY protein constitutes the system output and interacts with a protein of the flagellum ( FliM ) to switch the direction of its rotation . The second RR domain is carried by the CheB methyl esterase , and its phosphorylation activates the de-methylation of the Mcp to reset the system to a pre-signaling state ( adaptation , for a review , see [9] ) . In many bacteria , the core signaling apparatus of the enteric Che system has been coopted to the regulation of processes other than chemotaxis , such as surface motility , gene regulation and even cellular differentiation [8 , 10 , 11] . The genetic structure of these non-canonical chemosensory systems is quite diverse and their circuit architecture is generally poorly understood [8 , 10 , 11] . In this study , we investigate the evolution and genetic structure of a chemosensory-type system that controls two distinct motility machineries in Myxococcus xanthus . Myxococcus xanthus , a gram negative deltaproteobacterium , uses surface motility to form multicellular spore-filled fruiting bodies when nutrient sources become scarce . During this process , the Myxococcus cells can move as single cells or within large coordinated cell groups and reverse their direction of movement in a process where the cell poles rapidly exchange roles [12] . In the cell groups , a Type-IV pilus ( Tfp ) assembled at the cell pole ( the leading pole ) acts as grappling hooks and pulls the cell forward by retraction ( Fig 1A , [13] ) . Tfps retract when they are in contact with a cell surface-exposed exopolysaccharide ( fibrils ) , giving rise to a cooperative form of group movements called S ( Social ) -motility [14] . At the colony edges , the single cells are propelled by the recently characterized Agl-Glt apparatus , otherwise called the A ( Adventurous ) -motility system [15 , 16] . The Agl-Glt complex is assembled at the leading cell pole and moves directionally towards the lagging cell pole , promoting movement when it contacts the underlying surface ( Fig 1A , [12 , 16 , 17] ) . Thus , both motility systems are activated at the leading cell pole and their activation is switched coordinately to the opposite cell pole when cells reverse ( Fig 1A , [18–20] ) . The frequency of the reversal events is under the genetic control of a chemosensory-like system called Frz [21] . This control is essential for Myxococcus multicellular behaviors because frz mutants that reverse at low frequencies do not form fruiting bodies and form characteristic “frizzy” filament structures [22] . The molecular link between Frz and the motility systems requires three intermediate polarity proteins , MglA , MglB and RomR ( Fig 1A ) . MglA , a bacterial Ras-like G-protein , activates the motility systems at the leading cell pole ( Fig 1A , [23 , 24] ) . As all members of this family of molecular switches , MglA is active in association with GTP , a form that interacts with two motility system-specific proteins , AglZ ( A-motility ) and FrzS ( S-motility ) [19 , 20 , 25] . The polar localization of MglA results from combined regulations: ( i ) , by RomR , which recruits MglA-GTP to the cell poles [26 , 27]; and ( ii ) , by MglB , a MglA GTPase Activating Protein ( GAP ) , which prevents MglA access to the lagging cell pole ( Fig 1A , [23 , 24 , 28] ) . The polarity axis formed by MglA , MglB and RomR is stable unless it is contacted by upstream Frz signals , which provokes its dynamic switching and a reversal ( Fig 1A , [29] ) . In vivo , unknown signals are sensed by the Mcp homologue ( FrzCD ) and lead to the autophosphorylation of the kinase of the system ( FrzE ) from ATP via FrzA , the major CheW-like coupling protein ( Fig 1B , [30] . FrzB , another CheW-like homolog may also participate in the activation of the pathway , although contrarily to FrzA , it is not required for all Frz-dependent responses ( Fig 1B , [31] ) . Following activation , FrzE may then activate a handful of RR domains , including the C-terminal RR domain of FrzE ( FrzERR ) , the FrzZ protein and the N-terminal RR domain of RomR itself , to contact the polarity proteins and activate the polarity switch ( Fig 1B , [30 , 32–34] ) . Thus , the presence of multiple RR domains and a branch point at a circuit node formed by MglA asks how signaling from the FrzE kinase is channeled to the downstream motility systems ( Fig 1B ) . Overall , the motility regulation circuit is assembled from four interconnected modules with genetically separable functions: the control of the reversal frequency ( switch control module ) , the coordination of the A- and S-motility systems ( polarity control module ) and the function of A- and S-motility ( A- and S-motility modules , Fig 1B ) . By investigating the evolutionary origin of each module , we identified the core structure of the Frz pathway , genetically uncoupled A- and S-motility regulations and thus identified key regulations that led to the emergence of the evolved pathway . Remarkably , we find that adaptation of the ancestral circuit to A- and S-motility regulation required amplification of the signaling efficiency , suggesting that the evolution of signal reinforcement mechanisms may be linked to signal transduction pathway diversification in cells . From an evolutionary perspective , the co-regulation of the A- and S-motility complexes by the Frz system implied the connection of two machineries of different origins . While Tfp systems are found in all deltaproteobacterial genomes [26 , 35] , the A-motility Agl-Glt machinery is only present in Cystobacterineae , a Deltaproteobacteria family [15 , 16 , 36] . Thus , S-motility may be more ancient than A-motility and the emergence of A-motility in Cystobacterineae could have expanded the phenotypic repertoire and adaptive capabilities of this family of bacteria . Accordingly in Myxococcus , S-motility is required for fruiting body formation on soft and hard surfaces ( 0 . 5% vs 1 . 5% agar ) but A-motility is only required on hard surfaces ( Fig 2A and 2B ) . Importantly however , Frz regulation is required on both types of surfaces ( Fig 2A and 2B ) . To further understand how co-regulation of A- and S-motility emerged in Cystobacterineae , we performed in-depth phylogenetic analyses of the components of the switch , polarity and motility control modules . Measuring reversals of the S-motility system is difficult because they occur in large cell groups where single cells cannot be easily tracked . As mentioned in the introduction , S-motility results from the action of an extracellular EPS that provokes Tfp retraction [14] . In fact , the requirement for EPS can be bypassed , in single cell assays where the Myxococcus cells are allowed to move in a carboxymethylcellulose-coated microfluidic chamber ( See Experimental procedures ) . In this system , cells move by Tfp-dependent motility only with an average speed of 1 . 7 ± 0 . 8 μm . min-1 ( measured for 63 cells ) and Frz-dependent cellular reversals are observed and coincide with pole-to-pole oscillations of MglA-YFP and FrzS-YFP , similar to agar ( Figs 4A , 4B , and S8A ) [19 , 24] . A-motility is not active on the cellulose surface because the cell velocity and the reversal frequency of an A-motility A− ( aglQ ) mutant are unchanged compared to WT cells ( 1 . 6 ± 0 . 9 μm . min-1 , for 61 cells , S1 Movie and Fig 4B ) . Therefore , we used A+ strains for the rest of this study . We used the cellulose assay and high-throughput automated tracking ( S8B Fig and S2 Movie ) to test whether the core pathway defined by the phylogenetic analysis is sufficient to regulate Tfp-dependent reversals . Remarkably , Tfp-dependent reversals were still observed in the absence of each of the predicted accessory components , FrzZ , FrzB and AglZ ( Fig 4B ) . The frequency of Tfp-dependent reversals was equivalent to WT levels in the frzB and aglZ mutants , showing that these proteins are dispensable for the control of Tfp-dependent reversals ( Fig 4B ) . In the case of the frzZ mutant , the situation was intermediate , the reversal frequency was affected compared to WT cells but it was still significantly higher than in the frzE mutant ( Fig 4B ) . Because the reversal frequency of the frzZ mutant was intermediate , we decided to use a more precise reversal-scoring test to determine unambiguously if frzZ mutant cells can still reverse in a Frz-dependent way . Our interpretation could be biased by so-called Tfp-dependent “stick-slip” motions , a Frz-independent Tfp-driven movement that could be mistakenly counted as reversals ( S9A and S9B Fig , stick-slip motions are short range and generally appear distinct from bona fide reversals , [42] ) . Therefore , to score reversals with high accuracy , we monitored FrzS-YFP oscillations as a proxy for Frz-dependent reversals ( Figs 4A , S9A and S9B ) . As expected from the reversal measurements , pole-to-pole switching of FrzS-YFP coincident with directional changes was reduced on average in the frzZ mutant , but they were still observed and sometimes up to WT levels ( Fig 4C ) . Confirming this , pole-to-pole switching of FrzS-YFP was completely abolished in the frzE mutant and a frzE frzZ double mutant behaved like the frzE mutant ( Fig 4C ) . Therefore , although the reduction in the reversal frequency of the frzZ mutant is significant ( Fig 4B and 4C ) , Frz-dependent Tfp-reversals can occur in the absence of FrzZ ( but not in the absence of FrzE ) suggesting that FrzZ acts positively on Tfp-dependent reversals but is not strictly required for their activation . To test the properties of the predicted core Frz pathway , we further constructed a frzB , frzZ , aglZ triple mutant ( Δ3 ) . The Δ3 mutant still showed Frz-dependent reversals and its reversal frequency was again lower , similar to that of the single frzZ mutant ( Fig 4B ) . However , this lower reversal frequency did not translate into obvious S-motility developmental phenotypes because the Δ3 mutant formed fruiting bodies on soft agar , which is strictly S-motility dependent ( Fig 4D ) . This multicellular development was Frz dependent because a Δ3 frzE quadruple mutant did not form aggregates in similar conditions ( Fig 4D ) . On the contrary and as expected , the Δ3 mutant did not form aggregates on hard agar , a condition where A-motility is required ( Fig 4D ) . We conclude that although the predicted core Frz pathway has a lower activity than the evolved pathway containing FrzZ , this activity could be sufficient to allow strictly S-motility-dependent behaviors , ie the ability to make fruiting bodies on soft 0 . 5% agar surfaces . Thus , the acquisition of FrzB , FrzZ and AglZ could have further adapted this primary circuit to the regulation of two motility machineries , at least in part by boosting the signaling activity ( see below ) . The results above show that FrzE signaling is still efficient in the frzZ mutant ( albeit at lower efficiency ) , suggesting that another response regulator delivers FrzE signals to the polarity complex . RomR is a possible candidate because it contains a phosphorylatable Nt-RR domain and because it interacts directly with MglA . Thus , phosphorylation of RomR could link Frz signaling to the polarity switch [34] . The regulatory function of the RomR protein could not be tested in the past because RomR is essential for A-motility ( likely because MglA is delocalized in a romR mutant , [26 , 27 , 34] ) , and reversal frequencies are traditionally measured on A-motile cells . In the cellulose system , a romR deletion mutant showed WT Tfp-dependent motility ( 1 . 4 ± 0 . 8 μm . min-1 , for 65 cells , S3 Movie ) but it was dramatically affected in its ability to reverse in the FrzS-YFP oscillation assay ( Fig 4C ) . A frzZ romR double mutant also had abolished reversals showing that RomR acts downstream from FrzZ in the regulation ( Fig 4C ) . Importantly , while occasional reversals could be observed in frzE mutants , reversals were very rarely observed in romR and frzE romR mutants ( Fig 4C ) . We conclude that RomR acts downstream from FrzE and FrzZ in the reversal pathway and could thus be a central output protein of the Frz pathway . Both the frzZ and the frzB frzZ aglZ Δ3 mutants have similar and lower reversal frequencies than WT cells ( Fig 4B ) , suggesting that the presence of FrzZ increases the steady-state signaling activity . Because the Δ3 mutant still forms fruiting bodies on 0 . 5% agar ( Fig 4D ) , lower Frz activity may only translate in a developmental defect when A-motility is required . If so , differential Frz signaling activities may be required to regulate S-motility-dependent behaviors ( ie development on soft agar ) and A- and S-motility dependent behaviors ( ie development on hard agar ) . This hypothesis can be tested in a strain expressing the frz operon under the control of an IPTG-inducible promoter ( Fig 5A ) where Frz activity should be related to the level of Frz protein expression . In the absence of IPTG , such strain formed fruiting bodies on 0 . 5% agar but not on 1 . 5% agar , indicating that promoter leakage is even sufficient to restore S-motility-dependent aggregation ( Fig 5A ) . As expected , the addition of IPTG also restored fruiting body formation on 1 . 5% agar ( Fig 5A ) . Thus , developmental processes that require the S-motility system require lower Frz activity levels than developmental processes that require A- and S-motility . The FrzZ protein was thought to be central to Frz regulation because a frzZ mutant displays a typical frz phenotype on development hard agar [30] . However , if this phenotype is linked to lower Frz activity , it could be bypassed if Frz signaling is artificially increased in a frzZ mutant . To test this possibility , we took advantage of a chemical , Isoamyl alcohol ( IAA ) known to activate frz-dependent reversals . Although the exact target of IAA is not known , it appears to act on and de-methylate the FrzCD receptor ( directly or indirectly , [43] ) and its action is strictly Frz-dependent [31] . When added to hard developmental agar , IAA did not affect development of the WT strain up to concentrations of 0 . 075% , after which IAA blocked fruiting body formation ( Fig 5B ) . A frzE mutant showed the typical frz phenotype and as expected , this phenotype was neither rescued nor modified by IAA addition ( Fig 5B ) . Consistent with previous observations , the frzZ phenotype was indistinguishable from the frzE phenotype in absence of IAA ( Fig 5B ) . However , and contrarily to the frzE mutant , IAA rescued aggregation of the frzZ mutant up to 0 . 15% IAA , a high dose that disrupts aggregation in WT cells ( Fig 5B ) . Therefore , artificial activation of Frz signaling rescues the signaling defect of the frzZ mutant , suggesting that FrzZ acts to elevate Frz activity , allowing regulation of the two motility systems . To further investigate the function of FrzZ in Frz-signaling activity , we tested the contribution of FrzZ in a strain where the Frz receptor is hyper active . So-called frzon mutations map to the C-terminal domain of FrzCD and result in the expression of a truncated receptor protein ( FrzCDc ) [31] . Because the expression of FrzCDc is trans-dominant to the expression of FrzCD , frzon mutations have been suggested to hyper activate Frz signaling [44] . To first test this assumption , we purified FrzCD , FrzCDc , FrzA and the kinase domain of FrzE ( FrzEkinase , autophosphorylation of FrzE in vitro can only be detected if FrzERR is removed due to its phosphate sink activity , [32] ) and compared the capacity of FrzCD and FrzCDc to activate FrzEkinase autophosphorylation from ATP in vitro . While both FrzCD and FrzCDc were able to activate the autokinase activity of FrzEkinase in a dose-dependent manner , FrzCDc was a more potent activator ( Fig 6A and 6B ) . Thus , frzon mutations induce a hyper signaling state of the FrzE kinase . We then proceeded to test the contribution of FrzZ to Frz-signaling in a frzon background . In the cellulose chamber assay , frzon mutants reversed at high frequency , as expected ( Fig 6C , compare with the WT reversal frequency in Fig 4B ) . Remarkably , a frzon frzZ mutant still reversed but at a reversal frequency similar to the reversal frequency of the frzZ mutant ( Fig 6C , compare with Fig 4B ) . A frzon romR mutant did not reverse ( Fig 6C ) , confirming that Frz signaling is disrupted in absence of RomR . Thus , FrzZ acts downstream from the FrzCD receptor and exerts a positive effect on the transduction of Frz signals to the polarity switch . To investigate how FrzZ exerts its positive effect on Frz signaling activity , we took advantage of the cellulose system to develop a high-resolution single cell assay in which FrzS-YFP oscillations are measured directly as a function of stimulation levels; here , the addition of increasing doses of IAA . In this assay , we first established that reversals were induced by IAA in dose- and FrzE-dependent manners . In WT cells , IAA induced a sharp dose-dependent reversal response until a plateau was reached at an IAA concentration of 0 . 15% ( Fig 7A and 7B ) . As expected , a frzE mutant only reversed occasionally whatever the IAA concentration , showing that the observed IAA effects are strictly Frz-dependent ( Fig 7A and 7B ) . Consistent with previous results , a frzZ mutant still showed an IAA-dependent response but it was more gradual than the WT and showed lower amplitude at the higher IAA doses ( Fig 7A and 7B ) . We also used the IAA single cell assay to test the function of FrzERR , the FrzE receiver domain . The FrzERR domain is not absolutely essential for Frz signaling and has been suggested to inhibit signaling because FrzERR inhibits FrzE autophosphorylation in vitro ( Fig 1B , [32] ) . However , how such inhibition participates in Frz signaling is unclear . In the IAA assay , a frzERR mutant showed a behavior opposite of that of the frzZ mutant: this mutant reversed more than WT cells which was apparent IAA doses ranging between 0–0 . 075% ( Fig 7A and 7B ) . Thus , FrzERR prevents Frz-signaling at low stimulation levels . Remarkably , at concentrations higher than 0 . 075% the frzERR mutant stopped reversing and no longer responded to IAA ( Fig 7A and 7B ) . We hypothesize that this collapse is the result of an over-signaling state that disrupts Frz signaling function because ( i ) , a frzERR frzZ double mutant showed a composite phenotype: frzERR-type reversal frequencies at IAA doses ≤ 0 . 03% and frzZ-type reversal frequencies at the higher IAA concentrations ( Fig 7A and 7B ) ; and , ( ii ) , a frzon frzERR double mutant has a strongly reduced reversal frequency ( Fig 6C ) . In the enteric Che pathway , high chemoreceptor stimulation also inhibits signaling suggesting that similar mechanisms are at work in the Frz system [45 , 46] All together , the IAA experiments and the frzon mutant reversal frequencies suggest that FrzERR and FrzZ act independently in the pathway , FrzERR blocking activation at low signal levels and FrzZ amplifying signal transmission to allow a rapid switch-like response to stimulation ( which is required for the regulation of A- and S-motility ) . The Myxococcus motility apparatuses ( A- and S-motility ) allow this bacterium to perform an array of multicellular behaviors , which likely increases the competitiveness of this bacterium in the environment . Our results are consistent with an evolutionary scenario whereby these behaviors emerged following the stepwise assembly of four distinct functional modules , the Frz chemosensory apparatus , the polarity proteins MglAB , and the A- and S-motility systems , in a regulation pathway . Given that the studied genes likely form a minimal regulatory set and that not all players and interactions have been identified , a complete evolutionary scenario cannot be proposed . Nevertheless , we identify two major steps in the evolution of the pathway: In bacteria , the cooption of signaling modules formed by Mcp ( FrzCD ) , CheW ( FrzA ) , CheA ( FrzE ) , CheR ( FrzF ) and CheB ( FrzG ) homologues underlies the emergence of a large number of chemosensory-type pathways [11] . Therefore , it is likely that the primary deltaproteobacterial S-motility regulation apparatus first evolved by recruitment of MglAB to one such chemosensory system . RomR is a possible candidate to link Frz signaling to polarity regulation because this protein shares a similar evolutionary history ( Fig 3A and 3B ) , it interacts directly with MglA [26 , 27] and it is essential for reversals ( this work ) . However , we have not demonstrated that FrzE is the RomR kinase and thus formally , other intermediate proteins may relay FrzE signals to RomR . Nevertheless , our genetic analysis suggests that RomR functions as a core protein downstream from FrzE in the regulation pathway . Downstream from MglA , it is also possible that FrzS does not constitute the only link to the S-motility apparatus [12 , 18 , 47] , but because the interaction between MglA and FrzS is essential for S-motility [25] , the acquisition of FrzS was probably a key step for the emergence of the primary pathway . Diversification of the primary pathway to the regulation of A- and S-motility occurred in the Cystobacterinaea family of bacteria and coincided with profound modifications of the regulation system . Additions of AglZ and FrzZ were probably key to adapt the primary circuit to the emergence of a branch point downstream from the FrzE kinase . First , the duplication of a frzS ancestor gene and connection of AglZ to the A-motility apparatus might have created the branch point itself , connecting A-motility to MglA regulation . Aside from numerous amino acid substitutions , the main difference between AglZ and FrzS resides in the lengths of their coiled-coil domains . Thus , AglZ might have evolved a new interaction with the Agl-Glt machinery ( ie via the coiled-coil domain ) while retaining its ability to interact with MglA . Consistent with this , AglZ also interacts with MglA and it co-localizes with the Agl-Glt machinery [16 , 25 , 39 , 48] . Second , FrzZ was incorporated into the upstream regulatory circuit , allowing signal partitioning to the two motility systems ( see below ) . Other changes in the pathway not studied here may also participate in this regulation , including domain changes occurring in FrzCD and FrzG and the acquisition of FrzB ( S4 Table and Figs 3A , 3B , S3A and S3E ) . The function of FrzB does not appear redundant to that of FrzA , the major Frz CheW protein , because Bustamante et al . [31] showed that a frzA mutant is indistinguishable from a frzCD or a frzE mutant , while a frzB mutant still responds to IAA stimulation in a bulk agar motility assay [31] , which is consistent with our single cell experiments ( Fig 4B ) . This lack of redundancy may not be surprising because FrzB is not phylogenetically related to FrzA ( S3B and S3F Fig ) . It will be interesting to determine the exact function of FrzB and its potential connection to the branching of A-motility in the future . Using a high-resolution microfluidic single cell assay we were able to elucidate the individual contribution of the RR domain proteins of the pathway . The IAA stimulation Frz-dependent response curve showed a biphasic-type response with an overall sigmoidal shape ( Fig 8 ) . Because this response is entirely abolished in a frzERR frzZ mutant but not in the respective individual mutants ( Fig 7A and 7B ) , we conclude that FrzERR and FrzZ independently set distinct regimes of the signaling apparatus ( Fig 8 ) . More precisely , FrzERR inhibits signaling at low stimulation levels , blocking noisy activation of the polarity switch ( Fig 8 ) . The inhibition mechanism is likely that of a phosphate sink because hybrid kinases phosphotransfer to their covalently-attached receiver domain at very high efficiency and the half-life of the Aspartate-phosphate bond on FrzERR is very short lived , a property of response regulators with phosphate sink functions [32 , 49 , 50] . Signal inhibition by the receiver domain of hybrid kinases is also emerging in other systems [51] and could be a widespread regulation of hybrid kinases . At higher stimulation levels , the FrzERR inhibitory capacity becomes saturated and allows FrzE to phosphorylate the other receiver proteins of the pathway , FrzZ [30] , possibly RomR or any other unidentified receiver domain of the pathway . The combined action of these phosphorylation events results in a steep response ( Fig 8 ) . Phosphorylation events downstream from FrzE and independent of FrzZ would transduce the signal to the MglAB proteins , this could occur through RomR or other uncharacterized regulators . In parallel , the phosphorylation of FrzZ amplifies the signal , impacting both the slope and amplitude of the response ( Fig 8 ) . Remarkably , processes that require S-motility alone can accommodate a graded response of moderate amplitude , while processes that require A- and S-motility require FrzZ amplification ( Fig 8 ) . It will be essential to determine how Frz signals are processed by MglA and each motility system to understand these signaling intensity requirements . At the molecular level , FrzZ must act between the FrzE kinase and RomR because ( i ) , a hyper active FrzE kinase ( frzon mutation ) still requires FrzZ for maximal signal transmission ( Fig 6C ) , suggesting that FrzZ does not exert its action by feedback stimulation of the autokinase activity of FrzE; and ii ) , a frzZ romR mutant behaves like a romR mutant ( Fig 4C ) , showing clear epistatic relationships . FrzZ is a dual response regulator protein and it will be important to determine how the phosphorylation of each receiver domain contributes to signal amplification . The FrzZ phosphorylation sites appear partially redundant but only the phosphorylation of D52 and not D220 is important for the polar localization of FrzZ [33] . At the cell pole , the phosphorylated form of FrzZ could interact directly with RomR to facilitate the reversal switch . Two-component cascades frequently employ accessory response regulator domains to achieve a variety of functions in phosphorelays , signal inhibition or negative feedback loops [52 , 53] . To our knowledge , this is the first time that a signal amplifier function is identified for a RR protein and because FrzZ-like CheY-CheY fusion proteins are predicted in other complex two component systems ( Survey of the Microbial Signal Transduction database , [54] ) , characterizing the amplification mechanism may generally impact our understanding of bacterial signal transduction . In summary , this work reveals the modular structure of the Frz signal transduction pathway and suggests that the new pathway branch emerged at least in part by ( i ) , gene duplication followed by new functional specialization ( ie the emergence of AglZ and its connection to the A-motility system ) and ( ii ) , by the re-wiring of the signal flow , in this case an amplification system to partition signals at the branch point . These modifications are linked to the molecular structure of the Myxococcus regulation circuit where MglA acts as a regulation checkpoint , integrating upstream Frz signals into the coordinate regulation of the A- and S-motility machineries . In principle , amplification systems could operate in any signal transduction pathways that converge to the regulation of a checkpoint protein ( also called a master regulator ) . For example and conceptually similar to the Myxococcus system , eukaryotic Ras-like G-proteins must also partition their activity to several output proteins , ie during chemotaxis [55] . While different circuit designs could have evolved to achieve this function , the Myxococcus system could provide valuable lens to study the evolutionary and mechanistic processes that allowed one such diversification . Strains , plasmids and primers used for this study are listed in S1 , S2 , and S3 Tables . In general , M . xanthus strains were grown at 32°C in CYE rich media as previously described [31] . Plasmids were introduced in M . xanthus by electroporation . Mutants and transformants were obtained by homologous recombination based on a previously reported method [31] . E . coli cells were grown under standard laboratory conditions in Luria-Bertani broth supplemented with antibiotics , if necessary . Unless otherwise specified , soft and hard agar motility and development assays were performed as previously described [31] . In general , cell were grown up to an OD = 0 . 5 and concentrated ten times before they were spotted ( 10μL ) on CYE or CF [31] 0 . 5% ( soft ) agar or 1 . 5% ( hard ) agar plates for motility or for developmental assays . Colonies were photographed after 48 H or 72 H for motility or development , respectively . Developmental assays in the presence of Isoamyl alcohol ( IAA , Sigma Aldrich ) were performed similarly except that plates also contained IAA at appropriate concentrations . Single cell Tfp-dependent motility assays were initially developed by Sun et al . [13] in a system where Myxococcus cells are overlaid in methylcellulose . However , in this assay , reversals as observed on agar are not observed . In this media , the cells are loosely attached to the glass surface and they systematically detach and become tethered by one cell pole before a directional change is observed [13] . Many of these events may well result from actual reversal events , but other Tfp-dependent motions have been observed including stick-slip motions , sling-shot motions and walking up-right [42 , 56 , 57] . To avoid confusion linked to complex Tfp-dependent movements , we sought to optimize the methylcellulose assay . For this , homemade PDMS glass microfluidic chambers [58] were treated with 0 . 015% carboxymethylcellulose after extensive washing of the glass slide with water . For each experiment , 1mL of a CYE grown culture of OD = 0 . 5–1 was injected directly into the chamber and the cells were allowed to settle for 5 min . Motility was assayed after the chamber was washed with TPM 1mM CaCl2 buffer [58] . For IAA injections , IAA solutions made in TPM 1mM CaCl2 buffer at appropriate concentrations were injected directly into the channels and motility was assayed directly under the microscope . In TPM 1mM CaCl2 , we found that most WT motile cells left the field of view before reversing , making statistically reliable measurements of reversal frequencies difficult . This low reversal frequency is due to the absence of stimulating signals in these conditions . To increase the reversals counts and unless otherwise stated , Frz signaling was stimulated by adding 0 . 1–0 . 15% IAA to the TPM mix . Time-lapse experiments were performed as previously described ( Ducret et al . , 2013 ) using a Nikon TE2000-E-PFS inverted epifluorescence microscope . Image analysis was performed with a specific library of functions written in Python and adapted from available plugins in FIJI/ImageJ [59] . Cells were detected by thresholding the phase contrast images after stabilization . Cell tracking was obtained by calculating all objects distances between two consecutive frames , thus selecting the nearest objects . The computed trajectories were systematically verified manually and when errors were encountered , the trajectories were removed . The analysis of the trajectories is done automatically by a Python script that calculates the angle formed by the segments between the center of the cell at time t , the center of the cell at time t-1 and the center at time t+1 . Directional changes were scored as reversals when cells switched their direction of movement and the angle between segments was less than 90° . For non-reversing strains , the number of reversals for each cells was plotted against time using R software ( http://www . R-project . org/ ) . For strains that frequently reversed , the mean time between two reversals for each cells was plotted against time using R software . To further discriminate bona fide reversal events from stick-slip motions , the fluorescence intensity of FrzS-YFP was measured at cell poles over time . For each cell that was tracked , the fluorescence intensity and reversal profiles were correlated to distinguish bona fide reversals from stick-slip events with the R software . When a directional change was not correlated to a switch in fluorescence intensity , this change was discarded as a stick-slip event . The number of reversals was plotted against time using R software . Statistics were done using R software: Wilcox test was used when the number of cells was less than 40 in at least one of the two populations compared , and student test ( t-test ) was used for a number of cells higher than 40 . The genes encoding FrzEkinase , FrzA , FrzCD and FrzCDc were amplified by PCR using M . xanthus DZ2 chromosomal DNA as template and the forward and reverse primers listed in S3 Table . The amplified product was digested with the appropriate restriction enzymes , and ligated either into the pETPhos or pGEX plasmids generating pETPhos_frzEkinase , pETPhos_frzCD , pETPhos_frzCDc and pGEX_frzA which were used to transform E . coli BL21 ( DE3 ) Star cells in order to overexpress His-tagged or GST-tagged proteins . All constructs were verified by DNA sequencing . Recombinant strains harboring the different constructs were used to inoculate 400 ml of LB medium supplemented with glucose ( 1mg/mL ) and ampicillin ( 100μg/ml ) , and the resulting cultures were incubated at 25°C with shaking until the optical density of the culture reached an OD = 0 . 6 . IPTG ( 0 . 5 mM final ) was added to induce the overexpression , and growth was continued for 3 extra hours at 25°C . Purification of the His-tagged/GST-tagged recombinant proteins was performed as described by the manufacturer ( Clontech/GE Healthcare ) . In vitro phosphorylation assay was performed with E . coli purified recombinant proteins . 4 μg of FrzEkinase were incubated with 1μg of FrzA and increasing concentrations ( 0 . 5 to 7μg ) of either FrzCD or FrzCDc in 25 μl of buffer P ( 50 mM Tris-HCl , pH 7 . 5; 1 mM DTT; 5 mM MgCl2; 50mM KCl; 5 mM EDTA; 50μM ATP , 10% glycerol ) supplemented with 200 μCi ml-1 ( 65 nM ) of [γ-33P]ATP ( PerkinElmer , 3000 Ci mmol-1 ) for 10 minutes at room temperature in order to obtain the optimal FrzEkinase autophosphorylation activity . Each reaction mixture was stopped by addition of 5 × Laemmli and quickly loaded onto SDS-PAGE gel . After electrophoresis , proteins were revealed using Coomassie Brilliant Blue before gel drying . Radioactive proteins were visualized by autoradiography using direct exposure to film ( Carestream ) . 669 bp upstream from frzCD were amplified with primers CDind1 ( gaattcATGTCCCTGGACACCCCCAACGA ) and CDind2 ( actagtCATGGCCTGGATGAACTCGCCAAT ) and cloned into pGEM T-easy ( Promega ) to obtain plasmid pEM140 . pEM140 was digested with SpeI and EcoRI and the excised DNA fragment was cloned into pLacI ( a derivative of pAK20 [60] ) previously digested with the same enzymes . The resulting plasmid , pEM143 , is a derivative of pBBR1MCS carrying the lacI gene under its promoter and followed by the first 669 of frzCD and thus its integration by homologous recombination places the entire frz operon under IPTG control . Developmental plate assays were conducted in the presence ( 0 . 5 mM ) or absence of IPTG . For western blotting , strains were grown overnight with or without appropriate concentrations of IPTG . The cultures were concentrated to OD = 4 and western blotting was performed as previously described with 1/10 , 000 dilutions of anti-FrzCD ( Bustamante et al . , 2004 ) . A local protein database containing the 2 , 316 complete prokaryotic proteomes available in the NCBI ( http://www . ncbi . nlm . nih . gov/ ) as of May 23 , 2013 was built . This database was queried with the BlastP program ( default parameters , [61] ) using the full length sequences of the signaling proteins ( FrzF ( MXAN_4138 ) , FrzG ( MXAN_4139 ) , FrzE ( MXAN_4140 ) , FrzCD ( MXAN_4141 ) , FrzB ( MXAN_4142 ) , FrzA ( MXAN_4143 ) and FrzZ ( MXAN_4144 ) ) , the polarity control proteins ( MglA ( MXAN_1925 ) , MglB ( MXAN_1926 ) and RomR ( MXAN_4461 ) ) and the downstream proteins ( FrzS ( MXAN_4149 ) and AglZ ( MXAN_2991 ) ) of M . xanthus as a seed . The homology was assessed by visual inspection of each BlastP output ( no arbitrary cut-offs on the E-value or score ) . The retrieved sequences were aligned using MAFFT version 7 ( default parameters , [62] ) . Regions where the homology between amino acid positions was doubtful were removed using the BMGE software ( BLOSUM30 option; [63] ) . For each protein , preliminary phylogenetic analyses were performed using FastTree v . 2 using a gamma distribution with four categories [64] . Most of the studied proteins belong to very large protein families . Based on the resulting trees , the subfamilies containing the sequences from M . xanthus were identified and selected for further phylogenetic investigations . The corresponding sequences were realigned using MAFFT version 7 with the linsi option , which ensures accurate alignments . The resulting alignments were trimmed with BMGE as previously described . Maximum likelihood ( ML ) trees were computed using PHYML version 3 . 1 [65] with the Le and Gascuel ( LG ) model ( amino acid frequencies estimated from the dataset ) and a gamma distribution ( 4 discrete categories of sites and an estimated alpha parameter ) to take into account evolutionary rate variations across sites . Branch robustness was estimated by the non-parametric bootstrap procedure implemented in PhyML ( 100 replicates of the original dataset with the same parameters ) . Bayesian inferences ( BI ) were performed using Mrbayes 3 . 2 . 2 [66] with a mixed model of amino acid substitution including a gamma distribution ( 4 discrete categories ) . MrBayes was run with four chains for 1 million generations and trees were sampled every 100 generations . To construct the consensus tree , the first 2000 trees were discarded as “burn in” . The phylogenetic signal can be substantially increased by combining multiple sequence alignments of proteins involved in the same cellular function/biological process and sharing a common evolutionary history in a single large alignment ( also called supermatrix ) , [16 , 67–71] . Among the 12 studied genes , we showed that FrzF , FrzG , FrzCD and FrzE are always clustered together in genomes and share a similar evolutionary history . These genes were thus combined to build a supermatrix ( S4 Fig ) . For similar reasons , a second supermatrix was built by combining MglA and MglB ( S2 Fig ) . The ML and BI phylogenetic trees corresponding to these two supermatrices were inferred as previously described [16] . The 79 complete proteomes of Delta/Epsilonproteobacteria available at the NCBI in May 17 , 2013 were retrieved and assembled in a local database ( S5 Table ) . We used SILIX to build the protein families of homologous sequences present in these genomes ( default parameters; [72] ) . The homologous sequences corresponding to protein families present exactly in a single copy per genome ( 17 proteins ) were aligned using MAFFT Version 7 ( linsi option ) , trimmed with BMGE and combined to build a large supermatrix ( 6958 positions ) . The ML phylogenetic tree corresponding to this large supermatrix was inferred with PhyML , as described above .
Deciphering the circuit design of signal transduction networks is a fundamental question in cell biology . This task is challenging because many pathways are branched and control multiple cellular processes in response to one or several environmental signals . Studying pathway diversification in bacteria could be a powerful approach because these organisms contain so-called chemosensory systems , modular signaling units that have been adapted multiple times independently to regulate a large number of physiological processes . Here , we studied one such system , the Myxococcus xanthus chemosensory pathway ( Frz ) that controls the directionality of two distinct motility systems ( A- and S-motility ) . By experimentally uncoupling the regulations , we found that the Frz pathway evolved from a simpler ancestral system that only controlled S-motility originally . Two major pathway remodeling events allowed the recruitment of A-motility to the regulation , ( i ) the duplication of a connector protein which created the branch point and ( ii ) , the acquisition of a signal amplification mechanism to allow signal partitioning at the branch point . These results reveal the core structure of a complex chemosensory system and generally suggest that gene duplication and signal amplification underlie the diversification of signal transduction pathways .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[]
2015
Evolution and Design Governing Signal Precision and Amplification in a Bacterial Chemosensory Pathway
Tsetse flies serve as biological vectors for several species of African trypanosomes . In order to survive , proliferate and establish a midgut infection , trypanosomes must cross the tsetse fly peritrophic matrix ( PM ) , which is an acellular gut lining surrounding the blood meal . Crossing of this multi-layered structure occurs at least twice during parasite migration and development , but the mechanism of how trypanosomes do so is not understood . In order to better comprehend the molecular events surrounding trypanosome penetration of the tsetse PM , a mass spectrometry-based approach was applied to investigate the PM protein composition using Glossina morsitans morsitans as a model organism . PMs from male teneral ( young , unfed ) flies were dissected , solubilised in urea/SDS buffer and the proteins precipitated with cold acetone/TCA . The PM proteins were either subjected to an in-solution tryptic digestion or fractionated on 1D SDS-PAGE , and the resulting bands digested using trypsin . The tryptic fragments from both preparations were purified and analysed by LC-MS/MS . Overall , nearly 300 proteins were identified from both analyses , several of those containing signature Chitin Binding Domains ( CBD ) , including novel peritrophins and peritrophin-like glycoproteins , which are essential in maintaining PM architecture and may act as trypanosome adhesins . Furthermore , 27 proteins from the tsetse secondary endosymbiont , Sodalis glossinidius , were also identified , suggesting this bacterium is probably in close association with the tsetse PM . To our knowledge this is the first report on the protein composition of teneral G . m . morsitans , an important vector of African trypanosomes . Further functional analyses of these proteins will lead to a better understanding of the tsetse physiology and may help identify potential molecular targets to block trypanosome development within the tsetse . The concept of blocking trypanosome development within its tsetse host has been underexplored , primarily due to a lack of understanding the molecular events involved in the vector-parasite interactions and also difficulties in accessing an established colony of tsetse flies needed to implement such studies . Tsetse ( Diptera: Glossina spp ) are the sole cyclical vectors of trypanosomes in sub-Sahara Africa . Glossina morsitans morsitans belong to the morsitans ( savannah ) group that infest huge areas of sub-Saharan Africa and hinder the progression of cattle farming over ten million square kilometres [1] . They are also the vectors of the human pathogens Trypanosoma brucei gambiense and T . b . rhodesiense , which cause debilitating and ultimately fatal diseases if left untreated . Due to evidence of emerging parasite resistance to the current frontline therapeutics [2] , mammalian toxicity to treatment and no working vaccine , new disease transmission control ideas have shifted to investigating the vector-parasite interface rather than targeting parasite interactions within the mammalian host . For successful transmission to occur , salivarian trypanosomes must overcome many immunological and physical barriers to undergo a complex migration and development in the fly . Once ingested with a bloodmeal , the bloodstream form transforms in the midgut lumen into the procyclic stage within 1–2 days post-ingestion . After a successful differentiation into procyclics , the parasites then must avoid the proteolytic attack of tsetse digestive enzymes , reactive oxygen species [3] , immune peptides [4] and serum complement [5] . They do this by escaping to the ectoperitrophic space ( ES ) thereby crossing the peritrophic matrix ( PM ) , an acellular secretion that lines the midgut of many insects and could be present in more than one life stage [6] , [7] . After establishing an infection in the ES , the trypanosomes then colonise the proventriculus ( PV ) or cardia , where they continue to develop into long epimastigotes , which then cross the PM again en route to the salivary glands . In general , insect PMs are believed to be multi-functional and several roles have been proposed for this structure . Most functions depend on the selective permeability of the PM , but it is generally accepted that this tissue is analogous to the mucous secretions of mammalian digestive tracts [6] , [7] , [8] , in that it acts as a physical barrier to abrasive food particles and digestive enzymes . It has also been demonstrated that the PM acts as a biochemical barrier retaining ingested toxins [9] , [10] , [11] , thereby preventing cell damage and lethality to the insect . Perhaps more importantly , insect PMs impose physical barriers that prevent pathogens from reaching the midgut epithelium as demonstrated in mosquito studies [12] , [13] , [14] , and more recently , shown in two publications in Drosophila and Glossina [15] , [16] . There are two types of insect PMs described: type I and type II . Many heamatophagous adult diptera and important parasite vectors such as sand flies and mosquitoes possess a type I , which is secreted once from the midgut epithelial cells . Tsetse produce a type II PM , which is present prior to taking a bloodmeal and is continually secreted by a specialised group of cells in the PV . Electron microscopy , in combination with cytochemistry and lectin binding approaches , revealed that adult tsetse possess a highly organized , 3-layered PM ( ∼340 nm thick ) composed of glycosaminoglycans ( GAGs ) , glycoproteins of unidentified nature and chitin ( poly β- ( 1 , 4 ) -N-acetyl-D-glucosamine [GlcNAc] ) fibers [8] , [17] . In addition , very little is known on its overall protein composition and there is limited knowledge of the number of peritrophins that compose the tsetse PM . Until now , only Proventriculin 1 ( GmmPro1 ) and Proventriculin 2 ( GmmPro2 ) have been identified as putative components of the tsetse PM since these proteins are produced exclusively in the PM-secreting PV [18] . These putative peritrophins have barely been characterised , however , it is known that GmmPro2 is upregulated in susceptible tsetse lines ( salmon flies ) [19] . Peritrophins are structural PM proteins that are characterized by containing at least one chitin binding domain ( CBD ) that in turn have several conserved aromatic residues [20] . These CBDs interact with and bind chitin fibres present in the PM and other chitin containing proteins , which effectively influence PM tensile strength , elasticity and porosity , whilst the aromatic residues may bind carbohydrates . Peritrophins can also possess one or more mucin domains , reflecting the fact that they are believed to have evolved from mucins with the acquisition of CBDs [20] , [21] . These mucin domains possibly act as secretory compounds that aid water retention and resist enzymatic proteolysis . The teneral tsetse PM is the only partial physical barrier to trypanosome infection in the tsetse midgut and modifications to the PM as the fly ages may lead to a complete barrier to infection [22] . There is good evidence using electron microscopy that trypanosomes penetrate the tsetse PM [23] , [24] . However , this process must be dependent on the activity of PM-degrading enzymes since the pores in the tsetse PM are approximately 9 nm in size , which are too small for procyclic trypanosomes ( several microns long ) to pass through [25] , [26] . It is possible that proteins integral to the tsetse PM are important in infection establishment considering that parasites of other invertebrates secrete hydrolytic enzymes to degrade PM proteins in their respective hosts . To understand such strategies , a thorough revision of the composition and structure of the tsetse PM is required . This study provides the first insight into the overall protein content of the tsetse PM in an effort to understand , at the molecular level , the events involving trypanosome migration within the tsetse vector . Glossina morsitans morsitans ( Westwood ) were taken from an established colony at the Liverpool School of Tropical Medicine , which was maintained on sterile , defribinated horse blood ( TCS Biosciences ) at a relative humidity of 65–75% and an ambient temperature of 27°C±2°C . Experimental flies where collected at <24 hours post eclosion where they were briefly chilled at 4°C for initial sorting and kept in a 12 hour light and dark cycle in the same conditions as the colony until they were 72 hours old . All flies used in this study were teneral ( unfed ) male adults . PMs were dissected in sterile , chilled phosphate buffered saline solution ( PBS , 140 mM NaCl , 1 mM KCl , 6 mM phosphate buffer , pH 7 . 4 ) , transferred to 1 . 5 ml microcentrifuge tubes containing 200 µL of sterile PBS and centrifuged at 18 , 400×g for 5 minutes at 4°C . The supernatant was removed and the remaining PM pellet was washed three times in ice-cold distilled water for 10 minutes each at 18 , 400×g ( to remove excess salts , non-adhered bacteria and midgut contaminants ) then snap frozen and kept at −80°C until needed . PMs from ∼150 tsetse were thawed and re-suspended in 150 µL of 50 mM Tris-HCl ( pH 6 . 8 ) , containing 8 M urea , 3% SDS and 50 mM Dithiothreitol ( DTT ) . The sample was then sonicated in a sonicating ice-cold water bath 3 times for 5 minutes each and PM proteins precipitated with trichloroacetic acid ( TCA ) -acetone . Briefly , the PM suspension was mixed with 100% ice-cold acetone and 100% TCA ( 1∶8∶1 , V/V/V respectively ) and kept at −20°C for 1 hour [27] . After precipitation , the sample was centrifuged at 12 , 400× g for 15 minutes at 4°C , the supernatant discarded , and the protein pellet was washed twice with 1 ml ice-cold acetone . After the last wash , the remaining acetone was allowed to evaporate at room temperature , and the protein pellet was then re-dissolved in distilled water , mixed with Laemmli buffer [28] , and heated for 10 minutes at 95°C . In a separate experiment , 150 PMs were extracted and solubilized in urea buffer as described above , and then processed for in-solution tryptic digestion as described below . The PM protein preparation was fractionated on a NuPAGE ( Invitrogen ) precast 4–12% gel Tris-Bis gradient gel according to the manufacturer's recommendations . The gel was fixed overnight and the proteins were stained with colloidal Coomassie Blue G-250 ( Sigma ) as described by Neuhoff [29] , to allow sensitive visualization and destaining of proteins prior to mass spectrometry analysis . Approximately 10 µg/lane of a preparative PM protein urea extract were fractionated on a 12% SDS-PAGE and then transferred onto BioTrace polyvinylidene diflouride ( PVDF ) membrane at 90 V for 30 minutes . The membrane was then incubated overnight at 4°C in blocking buffer ( PBS/0 . 1% ( v/v ) Tween 20/5% ( w/v ) skimmed milk powder ) , containing 0 . 05% ( w/v ) sodium azide to prevent bacterial growth . After several washes in washing buffer ( PBS/0 . 1% ( w/v ) Tween 20 ) , separate membrane strips ( containing equal amounts of protein ) were probed for 1 hour at room temperature with either anti-tsetse or anti-bacterial primary antibodies: 1 ) mAb 4A2 ( mouse anti-Proventriculin 2 ( Pro2 ) ) 1∶25 dilution , 2 ) mAb TBRP/247 ( mouse anti-EP procyclin ) 1∶10 dilution , 3 ) mAb 3B2 ( mouse anti-lectin ) 1∶2 dilution 4 ) mAb 1H1 ( mouse anti-symbiont GroEL ) 1∶20 dilution , and 5 ) polyclonal rabbit anti-Sodalis glossinidus 1∶10 , 000 dilution . All antibodies were a generous gift from Prof . Terry Pearson ( University of Victoria , Canada ) . After several washes , the strips were incubated with a 1∶50 , 000 dilution of secondary antibody ( goat anti-mouse IgG ( antibodies 1–4 ) , or mouse anti-rabbit , Thermoscientific ( antibody 5 ) ( all conjugated to horse radish peroxidase ( HRPO ) ) at room temperature for 1 hour . After several washes , the strips were incubated with SuperSignal West Dura ( Pierce , UK ) peroxidase buffer and luminol/enhancer solution at a 1∶1 ratio , and developed by chemiluminescence , which continued for up to 3 hours . Excised gel plugs were destained in 50% acetonitrile/25 mM ammonium bicarbonate ( pH∼8 ) , reduced for 30 minutes at 37°C with 10 mM dithiothreitol ( Sigma ) in 50 mM ammonium bicarbonate and alkylated with 55 mM iodoacetamide ( Sigma ) in 50 mM ammonium bicarbonate for 30 minutes in the dark at room temperature . Gel plugs were washed for 15 minutes in 50 mM ammonium bicarbonate and dehydrated with 100% acetonitrile . Acetonitrile was removed and the gel plugs rehydrated with 0 . 01 µg/µL proteomic grade trypsin ( Sigma ) in 50 mM ammonium bicarbonate . Digestion was performed overnight at 37°C . Peptides were extracted from the gel plugs using successive 15 minute incubations of 2% ( v/v ) acetonitrile , 1% ( v/v ) formic acid . Peptide extracts were pooled and reduced to dryness using a centrifugal evaporator ( Jouan RC10-22 ) , and re-suspended in 3% ( v/v ) acetonitrile , 0 . 1% ( v/v ) TFA for analysis by mass spectrometry . For in-solution digestion , acetone precipitated PM material was solubilised with 0 . 1% ( v/v ) Rapigest ( Waters Corp . ) in 25 mM ammonium bicarbonate . The sample was heated at 80°C for 10 min , reduced with 3 mM DTT ( Sigma ) at 60°C for 10 min , and then alkylated with 9 mM iodoacetamide ( Sigma ) at room temperature for 30 min in the dark . Proteomic grade trypsin ( Sigma ) was added at a protein∶trypsin ratio of 50∶1 , and samples were incubated at 37°C overnight . Rapigest was removed by adding TFA to a final concentration of 1% ( v/v ) with incubation at 37°C for 2 hours . The peptide samples were then centrifuged at 12 , 000×g for 60 min at 4°C to remove precipitated Rapigest . Peptides were desalted using C18 Stage tips ( Thermo scientific ) , then reduced to dryness centrifugal evaporator ( Jouan RC10-22 ) , and re-suspended in 3% ( v/v ) acetonitrile , 0 . 1% ( v/v ) TFA for analysis by mass spectrometry . Peptide mixtures , generated by in-gel proteolysis of excised protein bands from polyacrylamide gels , were analysed by reverse-phase liquid chromatography ( RPLC ) using an UltiMate™ 3000 LC system ( DIONEX ) coupled to an LTQ ( Thermo Fisher Scientific ) mass spectrometer . Peptides ( 10 µl ) were injected onto a C18 column ( 2 µm particle size ( 100 ) , 75 µm diameter×150 mm long ) at nanoflow rate ( 300 nl/min ) and separated over a 50 minutes linear chromatographic gradient . The gradient consisted of the following phases: 0–30 min , 0–50% buffer B ( linear ) ; 30–30 . 1 min , 50–100% buffer B ( linear ) ; 30 . 1–35 min , 100% buffer B; 35 . 1–50 min , 0% buffer B . Full scan MS spectra ( m/z range , 400–2000 ) were acquired by the LTQ operating in triple-play acquisition mode . The top three most intense ions were selected for zoom scan and tandem MS by collision-induced dissociation ( CID ) . Peptide mixtures , generated by in-solution proteolysis , were analysed by on-line LC using the nanoACQUITY-nLC system ( Waters Corp . ) coupled to an LTQ-Orbitrap Velos ( Thermo Fisher Scientific ) mass spectrometer . Peptides ( ∼500 ng ) were injected onto the analytical column ( nanoACQUITY UPLC™ BEH130 C18 . 15 cm×75 µm , 1 . 7 µm capillary column ) at nanoflow rate ( 300 nl/min ) . The linear gradient consisted of 3–40% acetonitrile in 0 . 1% formic acid ( v/v ) over 120 min , followed by a ramp of 40–85% acetonitrile in 0 . 1% formic acid for 3 min . Full scan MS spectra ( m/z range . 300–2000 ) were acquired by the Orbitrap at a resolution of 30 , 000 . A data-dependent CID data acquisition method was used . The top 20 most intense ions from the MS1 scan ( full MS ) were selected for CID in the LTQ ion trap . Tandem MS data were searched against the Glossina morsitans morsitans database Glossina-morsitans-Yale_PEPTIDES_GmorY1 . 1 . fa . gz downloaded from VectorBase ( https://www . vectorbase . org/proteomes ) using the Mascot ( version 2 . 3 . 02 , Matrix Science , Liverpool ) search engine . Search parameters were a precursor mass tolerance of 10 ppm for the in-solution digest using the LTQ-Orbitrap Velos and 0 . 6 ppm for the lower resolution LTQ instrument . Fragment mass tolerance was 0 . 6 Da for both instruments . One missed cleavage was permitted , carbamidomethylation was set as a fixed modification and oxidation ( M ) was included as a variable modification . For in-solution data , the false discovery rate was filtered at 1% , and individual ion scores ≥30 were considered to indicate identity or extensive homology ( p<0 . 05 ) . Individual MS/MS spectra for single peptide hits with an ion score of 30 or above have been inspected manually and only included if a series of at least four continuous fragment ions were observed ( Supplemental Figure S1 and S2 ) . Tandem MS data were also searched against the Sodalis glossinidius peptide database generated from the latest re-annotated coding sequences [30] using the same search engine and parameters as described above . After colloidal Coomassie blue staining , many proteins with apparent molecular masses from ∼21 kDa up to >200 kDa were visualised , although a slight smeariness in a number of bands indicated the presence of highly modified proteins . Since many proteins do not stain with Coomassie blue ( e . g . mucins and peritrophins due to their high negative charge and acidity [33] ) , we decided to increase coverage by slicing the stained gel lane in 30 pieces from top to bottom ( Figure 1B ) . The individual bands were then excised , the proteins in-gel trypsinised and analysed by LC-MS/MS ( Figure 1 and Table 1 ) . This approach provided useful information regarding the relative abundance and masses of the different proteins ( Table 1 and S1 ) , whilst the in-solution analysis ( below ) increased the number of proteins identified . The most visually abundant proteins on the gel were a doublet migrating with relative molecular masses around 26 and 21 kDa ( bands 27 and 28 , respectively ) , which were identified as midgut trypsins . However , the most abundant and frequent hit in many of the bands analysed was a new type of peritrophin herein referred to as GmmPer66 ( discussed below in the peritrophins section ) . In addition , GmmPro2 , another known peritrophin-like protein that is produced in the PV [18] and the immunomodulatory TsetseEP protein [34] , [35] , were also detected in several bands ( Table 1 and S1 ) . The possible significance of the high occurrence of these proteins is discussed below . Furthermore , other peptidases , including GmmPro3 [18] , one serine peptidase and one putative metalloprotease , one chitinase , and several uncharacterised/conserved/hypothetical proteins were also found . Not surprisingly , abundant hits were also found for metabolic proteins , transporters and extracellular matrix proteins . The significance of the presence of these proteins is discussed below . In order to increase detection of PM proteins , a urea/SDS extract was also trypsinised in-solution and directly analysed by LC-MS/MS . A minimum of 195 G . m . morsitans proteins were identified . Only those with an ion score cut off of 30 or above were considered , with the majority of them having 2 or more identifying peptides and annotated on the VectorBase database ( version GmorY1 . 1 , 2013 ) and S . glossinidius genome . Proteins were classified and grouped by functional classifications ( Figure 2 ) , according to their GO terms and domain features as predicted by ExPASy Prosite , VectorBase and EMBL-EBI InterProScan . Hypothetical proteins were classified based on the presence of family domains . The majority of tsetse proteins ( 92% ) fit into 13 of the categories , whereas 15 proteins ( 8% ) could not be assigned to any category . However , all of these unknown proteins had orthologues in several insects and insect vectors , most of which had either no description or were described as conserved hypothetical proteins , suggestive of being ubiquitous among insects . Of the 195 proteins , 28 contained a predicted signal peptide ( SP ) , 26 were found to contain a transmembrane domain ( TM ) only , 16 had both a predicted signal peptide and at least one TM and the remainder ( 125 ) were predicted to be soluble ( i . e . neither SP nor TM domain ) . Interestingly , one of the most abundant hits corresponded to GmmPer66 . Two other novel peritrophins were also discovered: GmmPer12 ( GMOY011810 ) and GmmPer108 ( GMOY007191 ) . Western blot analysis was performed in order to validate some of the protein hits identified in both the in-gel and in-solution digested samples . Tsetse PMs were dissected , washed and solubilised with urea/SDS , processed for Western blotting and ∼10 PMs per lane probed separately with several anti-tsetse and two anti-Sodalis antibodies . As shown in Figure 3 , we were able to confirm by Western blotting the presence of one C-type lectin ( lane 1 ) , TsetseEP protein ( lane 2 ) and Pro2 ( lane 3 ) . In addition , the presence of symbiont proteins were confirmed using an anti-GroEL monoclonal antibody ( lane 4 ) , which cross-reacts with the GroEL of Wigglesworthia glossinidia and Sodalis glossinidius . To confirm that S . glossinidius , and not W . glossinidia , was isolated with the PM , an anti-Sodalis polyclonal antiserum was used ( lane 5 ) . This antiserum recognizes a suite of S . glossinidius proteins that produces a characteristic banding profile , including GroEL ( Mr∼60 kDa ) ( Haines , L . , unpublished ) ( Figure 3 ) . In total , five peritrophins were identified by mass spectrometry analysis from both in-solution and in-gel digestion and as such , has more than doubled the number of previously reported peritrophins from the Glossina PM ( Figure 4 ) . Both GmmPro1 ( GMOY011809 ) and GmmPro2 ( GMOY009587 ) are known to be synthesised in the tsetse PV and secreted during the formation of the PM [18] . The remaining three are novel , and this study is the first to positively identify them as being PM constituents . GmmPer12 ( GMOY011810 ) is a small peritrophin of 100 aa with a predicted molecular mass of ∼12 kDa and has a partial Peritrophin C Domain ( PCD ) . Originally , the PCD was thought to consist of 6 conserved cysteine residues [36] with the domain spanning 68–70 residues . Only recently has the PCD been shown to be composed of 120–121 residues and have a motif of 10 conserved cysteines [37] consisting of CX17CX9–10CX14CX9CX8–9CX19CX9–11CX14CX11C and those peritrophins thought to have a full PCD are now categorized as having partial domains . Partial domains may have come about through multiple duplication events or proteolytic degradation of full length proteins whilst retaining the ability to bind chitin . This proteolysis may occur before or after such CBD proteins have been incorporated into the matrix . Some partial CBDs have been shown to have trypsin and chymotrypsin cleavage sites embedded within the CBDs [38] suggesting these proteins are highly resistant to proteolysis owing to the folded nature of their structure through disulphide bond formation . GmmPer12 has a PCD of 4 conserved cysteine residues similar to that of GmmPro1 and GmmPro2 and is analogous to peritrophins found in other insects such as LcPer15 , a peritrophin found in the PM of the sheep blowfly Lucilia cuprina [39] ( Figure 5 ) . A predicted signal peptide between residues 19/20 suggests that GmmPer12 is secreted into the PM after synthesis . GmmPro1 , GmmPro2 and GmmPer12 are related to the peritrophin-15 family of proteins , integral proteins from the PMs of many insects [40] . This protein family is suggested to associate with the PM by binding to the ends of chitin fibrils giving structural support and preventing exochitinase action . The lack of N- and O-glycosylation on these 3 peritrophins supports this assumption . However , their intact forms appear to be absent in the PM , suggesting these three peritrophins are degraded and incorporated into the PM as partial fragments that have retained their ability to function as a chitin-binding domain . The updated Glossina VectorBase genome annotation has revealed that GmmPro2 is not 93 amino acids as previously reported [18] ( AAN52277 . 1 ) , but instead has an extension at its N-terminus making the protein 116 amino acids long . This is perhaps evidence that at least GmmPro2 ( and probably also GmmPro1 and GmmPer12 ) have evolved from a larger protein containing many CBDs . The majority of the proteins identified were hydrolytic enzymes including chitinases , amylases , exopeptidases and digestive enzymes such as trypsin . Although these may be midgut secreted proteins and only transiently associated with the PM , studies have shown these enzymes remain in the PM even after repeated washes and extraction with strong denaturants [14] . A tsetse Chitinase ( Cht1 ) was identified from both in-gel and in-solution analyses . Chitinases have been found associated in the PM of lepidopteran larvae where they are involved in the larvae moulting process [45] and are also found in the PM of adult mosquitoes where their role is less understood [46] . It has been suggested that during insect growth and development , chitin containing structures require the capacity to undergo remodelling and modification in order to allow for growth , maturation and repair [47] . This is especially true under certain conditions such as periods of moult or starvation where PM production can stop . In order for this to happen , tissue specific chitinolytic enzymes and chitin synthases are produced periodically . Chitinases are important in both the shedding of the cuticle during moults and growth and for the degradation and turnover of both the PM and trachea [47] . The fact that chitinases have now been identified in the tsetse PM suggests that PM chitinases in adult tsetse may be involved in degradation of the chitin fibrils thereby modifying the thickness , porosity and tensile strength of the PM during its extension along the length of the midgut . As expected , a large percentage of proteins ( 11% from the in-solution digestion ) were digestive enzymes such as trypsins , chymotrypsins , peptidases and serine proteases . Their identification may simply reflect transit across the PM to the endoperitrophic space in response to a blood meal , but given the fact that the flies used in this study did not receive a blood meal , it is possible that these enzymes are directly interacting with the PM in anticipation of feeding . These findings add to the reputable evidence that the PM improves digestion by concentrating the food bolus and filtering out indigestible components [48] . One serine protease , Proventriculin 3 or GmmPro3 , previously shown to be expressed in the tsetse PV , was also identified in this study suggesting that it might be physically associated to the PM . GmmPro3 is homologous to proteins of the serine protease S3 family and shares similarities with serine proteases from other haematophagus insects such as Stomoxys calcitrans and A . gambiae [18] . One serine protease inhibitor ( Serpin ) , GmmSpn4 , was also identified from the PM suggesting that serine proteases and serpins have a co-relationship involving blood meal digestion and may also modulate the PM structure until it is fully formed . Finally , proteases may also protect the passage of pathogens through the PM . In fact , the surface of procyclic trypanosomes gets “re-shaped” due to extensive proteolysis of the main surface glycoproteins , procyclins [49] , which partially may occur during PM crossing . From the in-solution analysis , 9 proteins ( ∼5% ) were identified as being involved in host-parasite interactions . These proteins were mainly C-type lectins ( CTLs ) , whose presence in the PM was corroborated by Western blotting ( Figure 3 ) . CTLs are Ca2+-dependent glycan binding proteins and play important roles in insect defence [50] . Carbohydrate binding events mediate a range of processes including cell/cell interactions , cell adhesion and are involved in cell apoptosis . They are also capable of recognizing pathogen-associated molecular patterns in a variety of microbes and in tsetse it has been suggested to be involved in the initial elimination of trypanosome burden by agglutinating parasites [51] , although so far no experimental evidence has proved this . Interestingly , from the in-gel analyses there were many hits for basement membrane-specific heparin sulfate proteoglycan core protein ( perlecan ) . Perlecan is a large proteoglycan with a multitude of diverse domains [52] . These domains bind to and cross-link numerous extracellular components and cell surface molecules . The N-terminal domain consists of ∼195 aa and contains three Ser-Gly-Asp attachment sites for large heparin sulfate chains or , occasionally , chondroitin sulfate . There is microscopy evidence showing that the G . m . morsitans PM contains glycosaminoglycan's ( GAGs ) in the layer facing the ectoperitrophic space ( epithelium side ) [53] , suggesting that this may be the location where perlecan may accumulate after secretion . Other domains include immunoglobulin , laminin and low-density lipoprotein ( LDL ) receptors that contain multiple cysteine residues able to form disulphide bridges . Perlecan also has an epidermal growth factor ( EGF ) domain , which is involved in ligand-recognition and protein-protein interactions . It is possible that identification of this protein is due to basement membrane contamination , however , if perlecan is a true PM protein , this may explain why proteins such as collagen , actin , lamin , laminin and fibronectin are found in a number of PM proteome studies [54] , [55] . It would be interesting to determine the exact place of perlecan synthesis . TsetseEP protein was also identified in both the in-gel and in-solution analyses . This is a unique tsetse protein of Mr∼36 kDa , which contains a characteristic extended glutamic acid-proline ( EP ) repeat domain at the C-terminus . Interestingly , its structure resembles that of the T . brucei EP-procyclins [31] , [34] , [56] . Studies have shown that TsetseEP probably acts as an antagonist to trypanosome infection [34] . TsetseEP is also highly upregulated in flies that have been challenged with gram-negative bacteria , which would suggest this protein may have an immunoprotective role [44] . The finding of TsetseEP in our analyses is intriguing . Although secretion of this molecule is enhanced by the presence of pathogenic microorganisms and it contains a lectin domain that may directly interact with pathogen's surface glycans , its elevated production during a midgut infection may also contribute to PM thickening , thus creating a stronger protective barrier . In Drosophila , there is genetic evidence showing that the PM structure changes in the presence of pathogenic bacteria [11] , [15] . In addition , an interesting recent work has shown that the Glossina PM becomes thinner in aposymbiotic flies ( i . e . lacking a midgut microbiome ) , which in turn increases PM permeability and allows an “easier” passage of trypanosomes through the PM [16] . Thus , although it remains to be determined how the structure of the tsetse PM changes in response to either pathogenic or non-pathogenic organisms , it may be possible that TsetseEP has a role in PM remodelling . Tsetse antigen 5 ( Tag5 ) was identified from the in-solution analysis . This protein of 259 amino acids is related to the large Crisp-Antigen 5 Plant pathogenesis protein families that are found in a huge diversity of organisms [57] . Mostly found in saliva of many insects , these proteins share a core sequence of approximately 200 amino acids that are responsible for their multiple functions . Antigen 5 has been proven as a potent venom allergen in hornets , wasps and fire ants and causes allergic reactions in humans [58] , [59] . Although primarily found in the salivary gland tissue of tsetse , it is reported to be expressed in the PV and midgut tissues [60] . A related protein in Drosophila , Antigen 5 related ( Agr ) , is also expressed in the PV of both larvae and adult flies [61] , [62] . Tag5 has also shown to be upregulated in a susceptible strain of tsetse ( salmon flies ) [19] . Tag5 may be a true constituent of the tsetse PM and as such may have a bearing on the digestion of the bloodmeal as studies have shown Tag5 prevents homeostasis [63] . As tsetse take up to 3 days to digest a bloodmeal , it is possible that the presence of Tag5 in the PM prevents the ingested bloodmeal from clotting quickly , thus aiding and facilitating digestion . Another protein identified and involved in immunity was glycoprotein CD36 , whose family members are conserved within mammals and have many representative orthologues in insects . They have a variety of functions including lipid transport , immune regulation , homoeostasis and adhesion . One function of CD36 is as a scavenger receptor , which recognizes molecular patterns presented by bacteria , pathogens and viruses and also pathogen infected cells [64] , [65] . An ortholog of CD36 in C . elegans , CO3F11 . 3 , is responsible for mediating host defences against fungal infection by stimulating the production of cytokines [66] . As a PM constituent , CD36 may have multiple roles from anti-homoeostasis to immune system mediation possibly involved in initial clearance of pathogens . In addition , this protein is highly resistant to proteolysis , which would be favourable given its putative location . Hemomucin , a 61 . 7 kDa protein containing extensive O-glycosylation at its C-terminus was also identified . It contains a domain showing strictosidine synthase , which is a key enzyme in alkaloid biosynthesis . Alkaloids are important in the immunity of plants and have been shown to be secreted in the venom of the fire ant where they act as potent inhibitors of bacteria [67] . Hemomucin from Drosophila proved likely to be involved in induction of antibacterial effector molecules after showing affinity for the snail lectin ( Helix pomatia hemagglutinin A ) . This protein was found to be expressed in the PV , suggesting that it may be incorporated into the PM after synthesis [68] . Proteins involved in stress response ( oxidation and reduction ) and protein folding ( heat shock and chaperones ) comprises a total 20% of the detected proteins . Some of these proteins may originate from the layer of epithelial cells that is in close proximity with the PM of teneral ( unfed ) flies . However , they may have a role in detoxification . Bloodmeal digestion leads to the rapid production of reactive oxygen species ( ROS ) , due to the breakdown of red blood cells , which causes the release of haem and iron . Accumulation of free haem leads to oxidative stress and these oxidation/reduction proteins are needed to detoxify the midgut environment [69] . It has been demonstrated in female Aedes aegypti that the PM of these insects are capable of binding haem during bloodmeal digestion as shown by histochemical studies [9] and a subsequent study has shown that at least one PM protein , the peritrophin AeIMUCI , is responsible for this interaction [10] . Haem-regulatory motifs ( HRM ) have also been found in peritrophins from 2 species of sandfly , Phlebotomus papatasi and Lutzomyia longipalpis [11] . A total of 27 S . glossinidius proteins were identified ( Table S2 and Table S3 ) . Given that Sodalis proteins have been identified within the PM and their presence verified by Western blotting ( Figure 3 ) suggests that secondary symbionts are intimately associated with the tsetse PM . Alternatively , these proteins may be secreted and incorporated into the PM , and thus they may have a functional role . The majority of these proteins were found to relate to metabolic activities within the bacteria . It has been well documented that Sodalis are important for many aspects of tsetse metabolism for example , cofactor and vitamin synthesis to compensate for the restricted diet of blood-meals [70] . Genes encoding biotin , lipoic acid , molybdenum cofactor , riboflavin and folic acid have all been found to be present in the genome of Sodalis glossinidius [71] . One interesting protein identified by mass spectrometry analysis was the Sodalis putative chitinase ( Accession No SG1474 ) . Studies have shown that when flies harbour a high density of Sodalis , they are more susceptible to trypanosome infection , thus it is entirely feasible to assume that these endosymbionts confer susceptibility to tsetse [72] , [73] , [74] . One possible explanation for this is that Sodalis may degrade chitin fibrils that comprise the tsetse PM , effectively remodelling it and providing an opportunity for trypanosomes to penetrate [72] , [75] , [76] . The primary carbon source during the growth of Sodalis is N-acetyl-β-D-glucosamine ( a monomer of chitin ) , which it produces from the breakdown of chitin using a secreted chitinase . Given that trypanosomes have no chitinase activity , it is reasonable to speculate that Sodalis breaks down PM chitin , leaving the PM vulnerable and unknowingly facilitating trypanosome crossing ( reviewed in [77] ) . In addition , the prevalence of trypanosome infection is highest when the fly is young and the PM is not yet fully formed . Proteins containing CBDs such as the peritrophins may have not yet been fully incorporated into the PM leaving the ends of PM chitin fibrils exposed . This may be the critical time point of chitinase activity , thereby degrading the PM and allowing trypanosomes to break through . Other parasites like Brugia malayi , Leishmania spp and Plasmodium spp secrete chitinases and proteases to degrade the proteins within the chitin meshwork and allow penetration of the PM [78] , [79] , [80] . Although the quantity of chitin has yet to be measured in the G . m . morsitans PM , the lack of chitinase expression in trypanosomes suggests that the chitin content of the tsetse PM may be probably low as reported in Lucilia cuprina larvae ( which also expresses a type II PM ) . Therefore , the tsetse PM chitin may not a real barrier to trypanosome infection [81] . Contrary to type I PMs , there is no molecular model representing the architecture of type II PMs . In the case of the Glossina PM , it is challenging to predict a model considering that it is composed of three layers [53] ( each one of different thickness and probably also in composition ) and because of the lack of EM localization of major peritrophins . However , based on its high abundance , number of CBDs and mucin domains , we hypothesise that GmmPer66 may play an essential role in interconnecting chitin fibres with other GmmPer66 monomers and/or other PM peritrophins , like GmmPer108 ( with 2 CBDs and 1 mucin domain ) or Pro2 ( with one CBD and several O-glycosylation sites ) . As suggested for other highly glycosylated molecules , the O-glycans from these peritrophins may serve to protect the PM from protease attack and retain water thus allowing the selective trafficking of molecules between the lumen and the ectoperitrophic space . It is conceivable that other peritrophins are also part of the tsetse PM , but their identification by MS was missed due to their resistance to trypsin . In fact , the G . m . morsitans genome contains a minimum of 41 peritrophins in the ( Attardo , G . et al . , in preparation ) . The study presented here has given a comprehensive overview of the main proteins that make up the tsetse PM identified using mass spectrometric techniques . Identification of at least 209 proteins from in-solution analysis and many more from in-gel analysis has provided a foundation of knowledge for which there is potential to expand . The identification of 3 novel peritrophins has expanded the list of known tsetse PM peritrophins from 2 to 5 . In addition , the unique banding pattern of one of these peritrophins , GmmPer66 , has provided us with useful insights into how their putative degree of crosslinking and how they are potentially incorporated into the PM . Although the quantity of chitin in the PM of G . m . morsitans has yet to be confirmed , the lack of chitinase activity in procyclic trypanosomes would suggest that the chitin component of the tsetse PM is extremely low and that chitin is not a real barrier to infection , proposing that the PM is composed mainly of glycoproteins rather than chitin . Therefore , a direct degradation of integral proteins may provide a pathway for trypanosome invasion through the PM . We are currently investigating candidate trypanosome proteases that may be participating in PM degradation . However , for proteases to act , glycosidases must first remove glycans ( i . e . chitin and GAGs ) . Therefore , it is intriguing that procyclic trypanosomes do not express the glycosidases to degrade any of these complex sugars . Alternatively , some of the PM-degrading glycosidases may be supplied by bacterial symbionts present in the tsetse midgut . With the completion of the Glossina genome project , a collaborative effort involving the VectorBase community and the Sanger Centre , there is great potential to reveal novel concepts about type II PMs . Insects with a type II PM are often more refractory to infection that those with a type I PM such as mosquitoes and sand flies . Whilst huge efforts have gone into researching larval type II PMs , this is the first study to concentrate on the protein composition of the adult type II PMs from an insect vector . The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium ( http://proteomecentral . proteomexchange . org ) via the PRIDE partner repository [82] with the dataset identifier PXD000594 and DOI 10 . 6019/PXD000594
African trypanosomes are transmitted by the haematophagous tsetse vector . For transmission to occur , bloodmeal ingested trypanosomes must overcome numerous barriers imposed by the fly . The first obstacle is the crossing of peritrophic matrix ( PM ) , a cell-free structure that protects the midgut epithelial cells from coming under attack by the hosts' digestive enzymes , aids in water retention and helps prevent harmful pathogens from establishing a systemic infection . Trypanosomes cross the tsetse PM at least twice in their development but how they do so remains to be elucidated . Despite being a recognised barrier to trypanosome infections , there is limited knowledge of the molecular components of the tsetse PM . In this study we identified nearly 300 PM proteins using two mass spectrometry approaches . Several of the identified components were peritrophins , which are a key group of glycoproteins essential for PM integrity . In addition , we detected proteins from Sodalis glossinidius , a commensal bacterium linked to increased susceptibility to trypanosome infection in tsetse . Our study provides the first comprehensive identification of proteins from the tsetse PM , which provides a starting point for research into potential targets for vector control .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results", "and", "Discussion" ]
[ "invertebrates", "medicine", "and", "health", "sciences", "pathology", "and", "laboratory", "medicine", "microbiology", "parasitic", "protozoans", "animals", "parasitology", "glossina", "protozoans", "tsetse", "fly", "epidemiology", "pathogenesis", "disease", "vectors", "insects", "arthropoda", "trypanosoma", "host-pathogen", "interactions", "vector", "biology", "protozoology", "biology", "and", "life", "sciences", "organisms" ]
2014
An Investigation into the Protein Composition of the Teneral Glossina morsitans morsitans Peritrophic Matrix
Dengue virus ( DENV ) is the leading cause of mosquito-borne viral illness and death in humans . Like many viruses , DENV has evolved potent mechanisms that abolish the antiviral response within infected cells . Nevertheless , several in vivo studies have demonstrated a key role of the innate immune response in controlling DENV infection and disease progression . Here , we report that sensing of DENV infected cells by plasmacytoid dendritic cells ( pDCs ) triggers a robust TLR7-dependent production of IFNα , concomitant with additional antiviral responses , including inflammatory cytokine secretion and pDC maturation . We demonstrate that unlike the efficient cell-free transmission of viral infectivity , pDC activation depends on cell-to-cell contact , a feature observed for various cell types and primary cells infected by DENV , as well as West Nile virus , another member of the Flavivirus genus . We show that the sensing of DENV infected cells by pDCs requires viral envelope protein-dependent secretion and transmission of viral RNA . Consistently with the cell-to-cell sensing-dependent pDC activation , we found that DENV structural components are clustered at the interface between pDCs and infected cells . The actin cytoskeleton is pivotal for both this clustering at the contacts and pDC activation , suggesting that this structural network likely contributes to the transmission of viral components to the pDCs . Due to an evolutionarily conserved suboptimal cleavage of the precursor membrane protein ( prM ) , DENV infected cells release uncleaved prM containing-immature particles , which are deficient for membrane fusion function . We demonstrate that cells releasing immature particles trigger pDC IFN response more potently than cells producing fusion-competent mature virus . Altogether , our results imply that immature particles , as a carrier to endolysosome-localized TLR7 sensor , may contribute to regulate the progression of dengue disease by eliciting a strong innate response . The innate immune system acts as the first line of defense for the sensing of viral infection . This involves rapid recognition of pathogen-associated molecular patterns ( PAMPs ) , including viral nucleic acids , by pattern recognition receptors ( PRRs ) . This recognition results in an antiviral response characterized by the production of type I interferons ( IFNs ) and expression of IFN-stimulated genes ( ISGs ) . This response suppresses viral spread by blocking the viral life cycle at multiple levels and also mediates immunomodulatory effects in surrounding tissues that impart the onset of the adaptive immune response [1] . The PRR can be cytoplasmic , e . g . , retinoic inducible gene-I ( RIG-I ) -like receptors ( RLRs ) and NOD-like receptors ( NLRs ) , or endosomal , e . g . , Toll-like receptors ( TLRs ) [1] . Thus , depending on their intracellular localization , virus-induced innate immune signaling typically occurs within cells that are either productively infected or that have internalized viral particles [1] , [2] . Recent studies illustrated the existence of alternative host sensing strategies by bystander plasmacytoid dendritic cells ( pDCs ) , which recognize infected cells [3] , [4] , [5] , [6] , [7] . pDCs are immune cells known to function as sentinels of viral infection and are a major type I IFN-producing cell type in vivo [8] , [9] . Using hepatitis C virus ( HCV ) as a model , we recently demonstrated that HCV infected cells can selectively package immunostimulatory viral RNA within exosomes that deliver their RNA cargo to pDCs , which , in turn , produce IFNα [3] . Exosomes also permit transfer to pDCs of distinct immunostimulatory viral RNAs , such as those of the negative strand lymphocytic choriomeningitis virus ( LCMV ) [4] . This sensing pathway is thought to assure recognition of infected cells and hence protects the host against viruses that defeat the pathogen-sensing machinery within the cells they infect . Virtually all viruses have evolved strategies that preclude antiviral signaling in the cell they infect [10] . For example , dengue virus ( DENV ) has evolved several evasion strategies that prevent IFN and ISG expression within infected cells [11] . Notably , the DENV NS2B-3 protease complex , by cleavage and degradation of an adapter of the cytoplasmic sensor-mediated signaling ( STING , also called MITA ) and by preventing phosphorylation and nuclear translocation of the downstream transcriptional factor , IFN regulatory 3 ( IRF3 ) , inhibits type I IFN production in DENV infected cells [12] , [13] , [14] , [15] . Despite these potent inhibitory mechanisms , expression of antiviral and inflammatory molecules is readily detected in DENV infected humans [16] , [17] . Their levels play a pivotal role in DENV infection clearance and pathogenicity [16] , [18] , [19] , thus highlighting the importance of elucidating the host sensing mechanisms leading to the IFN response during DENV infection . Here , we showed that pDCs are robust IFNα producer cells in response to DENV infected cells . In addition , we demonstrated that cell-to-cell contact- and TLR7-dependent pDC responsiveness leads to an antiviral state , inflammatory cytokine production as well as expression of co-stimulatory molecules by pDCs . Newly formed particles of DENV , like many viruses , undergo maturation by cleavage of the virus envelope protein , premembrane ( prM ) , in the secretory pathway that renders the virus infectious [20] . Yet , the prM cleavage site is suboptimal , leading to the secretion of about 30–40% immature , prM-bearing particles [21] , [22] , [23] , [24] , [25] , [26] . This evolutionarily conserved suboptimal site may be critical for the export of the infectious viral particles and/or may also positively contribute to viral infection by usurpation of humoral immune response , because anti-prM antibodies facilitate efficient binding and cell entry of prM-containing immature particles into Fc-receptor-expressing cells , a process called antibody dependent enhancement ( ADE ) [21] , [22] , [23] , [27] , [28] . Here , we report a previously unsuspected function of immature particles in innate immunity . Although the immature particles are not infectious , they are fully competent to trigger a robust type I IFN response by contacting non-permissive pDCs . Our results highlight the trade-off between efficient secretion of infectious viral particles and the production of a large amount of IFN-inducing immature particles . To investigate the mechanisms regulating the IFN response against DENV infection , primary human peripheral blood mononuclear cells ( PBMCs ) from healthy donors were exposed to supernatants containing DENV virions or DENV infected cells . We found that PBMCs specifically responded to co-cultivation with DENV infected cells but not to uninfected Huh7 . 5 . 1 cells , by a robust secretion of IFNα ( Figure 1A ) . In sharp contrast , supernatants from the DENV infected cells failed to trigger IFNα secretion by PBMCs ( Figure 1A ) . Plasmacytoid dendritic cells ( pDCs ) , which represent a rare PBMC population , i . e . 0 . 41% of PBMCs ( Figure 1B , upper panel ) , are known to produce IFNα [9] . Antibody-mediated pDC depletion from PBMCs ( Figure 1B , middle panel ) abolished IFNα secretion in response to co-culture with DENV infected cells ( Figure 1A ) . Similar results were also obtained using DENV infected BHK-21 cells ( Figures S1A and S1B ) . To rule out potential non-specific effects of the depletion procedure on innate cell responsiveness , we verified that IL-6 production triggered by lipopolysaccharide ( LPS ) exposure was maintained after pDC depletion ( Figures 1C and S1C ) . Consistent with the depletion results , the isolated pDC population ( Figure 1B , lower panel ) potently produced IFNα in response to co-culture with DENV infected cells , but not in the presence of their supernatants ( Figure 1A ) . A very limited number of pDCs ( i . e . , 2 , 000 pDCs ) was sufficient to produce a robust secretion of IFNα ( Figure 1A ) . Similar levels of IFNα production were detected after co-culture of infected cells with isolated pDCs as compared to total PBMCs , which contained a similar number of pDCs ( Figure 1D ) , further suggesting that pDCs are the main IFNα producer cells among PBMCs . We showed that the cells productively infected with DENV did not produce IFNα themselves ( Figure S2A ) . The pDC IFNα response increased as the duration of infection and , thus the replication levels , prior to co-culture increased ( Figure S2A ) . Remarkably , similar levels of IFNα secretion were reproducibly obtained with pDCs isolated from the blood of a cohort of 20 healthy donors ( Figure 1E ) . Together these results suggest that pDCs represent the main cell type in PBMC populations that produce IFN in response to co-cultivation with DENV infected cells and that this response was not induced by the addition of cell-free supernatants containing virus . To exclude the possibility that pDCs respond transiently to supernatants containing DENV , we quantified IFNα secretion in time course experiments . IFNα secretion was already detectable as early as 4 hours after co-cultivation of pDCs and DENV infected cells ( Figure 2A ) . IFNα levels concurrently increased over the time course of co-culture of DENV infected cells with either pDCs or PBMCs , and reached levels around 100 ng/mL after 16 hours of co-culture ( Figure 2A ) . In contrast , cell-free supernatants containing DENV did not trigger detectable IFNα production by pDCs or by PBMCs at any of time points analyzed ( Figure 2A ) . IFNα producer cells were markedly enriched in pDCs , characterized as a CD123-positive population , as compared to the CD123-negative population ( Figure 2B ) . For example , 12 hours after co-culture of DENV infected cells with PBMCs , ≈0 . 05% and ≈25–30% of CD123-negative and –positive cells , respectively , were IFNα positive ( Figure 2B ) . Consistently , the frequencies of IFNα producer cells ( i . e . , about 30% ) among pDCs ( i . e . , CD123-positive populations ) were comparable in co-cultures of DENV infected cells with PBMCs vs . isolated pDCs ( Figure 2B ) . Together these results suggested that IFNα is robustly produced only by pDCs that are co-cultured with DENV infected cells . Next , we showed that co-cultivation of DENV infected primary cells , i . e . , monocyte-derived macrophages ( mo-M ) and monocyte-derived dendritic cells ( mo-DC ) with pDCs ( isolated from the same donor ) , potently triggered pDC IFNα secretion ( Figures 3A and 3B ) . This stood in stark contrast to the corresponding cell-free supernatants containing virus or the parental uninfected cells did not , or very weakly , induced pDC IFNα production ( Figures 3A and 3B ) . Consistent with the previously reported inhibition of type I IFN production by the DENV NS2B-3 protease in infected cells [12] , [13] , [14] , [15] , DENV infected primary cells did not produced detectable levels of IFNα ( Figures 3A and 3B ) . Additionally , we determined if the production of IFNα by pDCs could be reproduced in response to co-culture with various cell types infected by DENV . Robust secretion of IFNα was triggered when pDCs were co-cultivated with DENV infected cell lines from different origins ( i . e . , human Huh7 . 5 . 1 , Hela and 293T cells or non-human BHK-21 and Vero cells ) , but not by the corresponding supernatants containing virus or the uninfected cells ( Figure 3C ) . DENV infected Vero cells were weaker IFNα inducers ( Figure 3C ) , consistent with lower levels of intracellular DENV RNA ( Figure 3D ) and infectious viral particle ( Figure 3E ) produced by these cells , suggesting that pDC IFNα induction is proportional , to some degree , with the level of viral replication . Remarkably , 293T cells infected by another member of the Flavivirus genus , West Nile virus ( WNV ) , but not the corresponding cell-free supernatants containing virus , also triggered robust IFNα production when co-cultured with pDCs ( Figure 3C ) . Similar to the results obtained using co-cultures with DENV infected cells , the pDC IFNα responses increased as the numbers of WNV infected cells increased ( Figures S2B and S2C ) . Together , these results demonstrated that the production of IFNα by pDCs in response to co-culture with DENV infected cells is not cell type specific and that pDCs similarly respond to another member of the Flavivirus genus . Cell-free supernatants containing virus from various infected cell types failed to trigger pDC IFNα production , even when added as crude non-filtered supernatants containing virus at concentrations as high as 20 infectious units per pDC ( Figure S3 ) , indicating that the transmission of the immunostimulatory signal to pDCs likely requires cell-to-cell contacts . To determine if contacts with DENV infected cells favors pDC sensing , we assessed IFNα production by pDCs cultured in transwell chambers with infected cells . Transwell cultures containing DENV infected monocyte-derived dendritic cells ( mo-DCs ) and pDCs separated by a 0 . 4 µm permeable membrane did not result in detectable levels of IFNα production by the pDCs ( Figure 4A ) . Similar results were obtained using DENV infected Huh7 . 5 . 1 , BHK-21 , Hela and Vero cells as well as WNV infected cells ( Figure 4B ) , confirming that this feature is not cell type specific or restricted to DENV . Similarly to IFNα , pDCs robustly produced IFNβ when in contact with DENV infected cells , but not when cells were physically separated by a transwell membrane ( Figure 4D ) . Consistent with these results , IFNβ production by pDCs was not triggered by supernatants from DENV infected cells and DENV infected cells did not themselves release detectable levels of IFNβ ( Figure 4D ) . In control experiments using identical transwell culture settings , an agonist of TLR7 , a viral RNA immune sensor [9] , triggered the production of both IFNα and IFNβ by the pDCs at levels similar to those obtained in the co-culture setting ( Figure 4C and 4E ) , thus ruling out potential non-specific effects of the experimental setting on pDC responsiveness . In agreement with previous reports [29] , [30] , vesicular stomatitis virus ( VSV ) or Influenza virus ( FluAV ) containing supernatants robustly triggered IFNα production by pDCs ( Figure 4F ) . Consistent with these results , VSV and FluAV infected cells in contact with pDCs ( Figure 4F , cocult ) or separated by a transwell membrane ( Figure 4F , TW ) , triggered IFNα production at similar levels . This suggested that contact with virus infected cells is not a universally employed mechanism to promote pDC activation by RNA viruses . Next , viral transmission across the transwell-membrane was assessed by quantifying infectious DENV ( Figure 4G ) and WNV ( Figure 4H ) on both sides of the membrane that separated infected cells from recipient cells . To evaluate the possible interference of recipient cells on the extracellular infectivity detection , we compared two types of recipient cells , i . e . , IFNα response-competent pDCs , which are non-permissive to infection ( Figure S4 ) and permissive cells ( Figure 4G and 4H , Naïve recipient cells ) . As expected from their size , infectious viral particles readily flowed across the 0 . 4 µm membrane ( Figures 4G and 4H ) , thereby permitting viral transmission from infected cells to naïve cells in the absence of direct contact ( Figures 4I and 4J ) . In sharp contrast , type I IFN production by the pDCs was induced exclusively under conditions where cell-to-cell contact was possible between infected cells and pDCs ( Figures 4A , 4B and 4D ) . Collectively , these results demonstrated that the exposure of pDCs to the DENV or WNV infected cell milieu either at defined time points ( Figure 3A–C ) or continuously ( Figures 4A , 4B and 4D ) failed to trigger a robust IFN response by pDCs , which were responsive to infected cells by cell-to-cell contact and/or in a short-range manner . pDCs typically respond to viral infection via endolysosome-localized TLR7- or TLR9 sensors that recognize RNA or DNA viral genomes , respectively [9] . Accordingly , we examined the transmission of DENV RNAs to co-cultured pDCs . The presence of DENV RNA in infected cells and co-cultured pDCs ( selectively labeled with DiI , a fluorescent membrane dye ) was assessed using a highly sensitive DENV RNA-specific fluorescence in situ hybridization ( FISH ) assay ( Figure 5A , upper panels ) . The analyses were performed after 5 hours of co-culture with DENV infected cells , at which time pDCs already produced IFNα ( Figure 2A ) . DENV RNA ( green ) was detected as discrete dots inside pDCs ( Figure 5A , lower panels ) . Inspection of consecutive Z-axis sections of co-cultures stained by combined DENV RNA FISH and anti-IFNα immuno-detections revealed that the frequency of DENV RNA-positive pDCs was elevated in both IFNα-positive ( i . e . , 85% ) and IFNα-negative pDCs ( i . e . , 74 . 5% ) ( Figure 5A , summary table ) . The specificity of these examinations was validated by the absence of DENV RNA-positive pDC when co-cultured with uninfected cells and when the FISH procedure was performed in the absence of the DENV RNA specific probe ( Figure 5A , summary table and Figure S5 ) . The presence of DENV RNA in IFNα-negative pDCs may reflect the time required for DENV RNA to trigger pDCs to produce enough IFNα to be detectable , which may not have occurred by 5 hours of co-cultivation . Alternatively , differential DENV RNA localization in intracellular compartments may modulate their recognition by innate sensors , and/or potential subsets of pDCs may be differentially responsive to the DENV RNA stimulus , in accordance with the maximal detection of about 30% IFNα-positive pDCs at plateau ( Figure 2B ) . Only a few DENV RNA dots were detected inside pDCs , suggesting that it is a rare event but sufficient to trigger pDC IFN production . Together , these results indicated that DENV RNA was readily transmitted from DENV infected cells to co-cultured pDCs , supporting the notion that DENV RNA might be recognized by pDC TLR7 . Accordingly , a TLR7 antagonist significantly inhibited pDC IFNα production induced by DENV infected cells ( Figure 5B ) . The specificity of this TLR7 antagonist was demonstrated by the inhibition of IFNα production induced by a TLR7 agonist ( R848 ) but not by a TLR9 agonist ( ODN2216 ) ( Figure 5B ) . Collectively , these results suggested that DENV infected cells transfer viral RNA to co-cultured pDCs and trigger TLR7-dependent IFNα production . Next , to further define the nature of the pDC-mediated antiviral state induced by contact with DENV infected cells , we examined the secretion of the inflammatory cytokines , IL-6 and tumor necrosis factor ( TNF ) -α , triggered by activation of the transcription factor NF-κβ , known to transduce antiviral signaling downstream of TLR7 [1] . TNF-α is known to play a pivotal role in the vascular leakage syndrome , a hallmark of dengue hemorrhagic fever [18] . Sensing of DENV infected cells , but not their supernatants , specifically triggered pDCs to produce IL-6 and TNF-α at levels comparable to those induced by treatment with a TLR7 agonist ( Figure 5C ) . In addition , ISGs ( i . e . , MxA and ISG56 ) were specifically up-regulated in co-cultures of DENV infected cells with pDCs or PBMCs ( Figure S6 ) , thus indicating the establishment of an antiviral state . Finally , we determined if DENV infected cells trigger pDC maturation as assessed by the up-regulation of the CD83 and CD86 markers at the cell surface . DENV infected cells , but not their supernatants , triggered a rapid increase in the surface expression of CD83 on co-cultivated pDCs ( i . e . , in CD123 marker-gated cells ) ( Figure 5D , left panel ) , accompanied by a slightly delayed augmentation of CD86 cell surface expression ( Figure 5D , left panel ) and by a concomitant increase in IFNα secretion ( Figure 5D , right panel ) . Collectively , these results demonstrated that sensing of DENV infected cells by TLR7 , a sensor of single stranded-RNA , triggers IFNα production by pDCs , along with the induction of the inflammatory response , an antiviral state and pDC maturation . To define how pDCs sense DENV infected cells , we analyzed the ability of cells harboring recombinant DENV genomes containing mutations specifying phenotypes deficient in various viral functions to trigger IFNα production by co-cultured pDCs . First , we tested cells containing DENV genomes encoding lethal mutations in the methyltransferase domain of the viral NS5 polymerase ( i . e . , Rep−/− ) [31] . As expected [31] , the triple mutation significantly reduced the intracellular level of DENV RNA at 48 hours post-transfection as compared to the wild type ( WT ) genome ( Figure 6A ) , reflecting a failure to amplify viral RNA ( Figure S7A ) . Consistently , this mutant did not express detectable amounts of intracellular viral proteins ( Figure S7B ) . Despite comparable intracellular viral RNA levels between the DENV WT and Rep−/− mutant genomes at the onset of co-culture i . e . , 24 hours post transfection , likely reflecting the input transfected RNA ( Figure 6A ) , Rep−/− DENV mutant genome harboring cells did not trigger IFNα production by co-cultured pDCs ( Figure 6D ) . Similarly , cells harboring DENV genomes encoding a four amino acid deletion in the capsid ( i . e . , amino acids V51-to-L54 ) , that significantly compromised both viral RNA replication ( Figures 6A and S7A ) and viral protein expression ( Figure S7B ) , failed to induce IFNα production by co-cultured pDCs ( Figure 6D ) . Together these results indicated that the pDC IFNα response requires active viral replication in neighboring DENV infected cells . Next , to address the requirement of viral genome release for pDC activation , we tested the effects of co-culture with cells harboring DENV genomes encoding point mutations in the envelope ( E ) glycoprotein , i . e . , the substitutions D215A , H244A or P217A , known to inhibit infectious viral production [32] , [33] . Consistent with previous reports [32] , [33] , the E glycoprotein mutations did not impair intracellular levels of either viral RNAs or proteins ( Figures 6A , S7A and S7B ) , but they all greatly compromised the production of infectious particles ( Figure 6B ) . Both the D215A and H244A mutations abrogated the release of viral RNA and structural proteins and the pDC IFNα response ( Figures 6C–D and S7C ) . Conversely , cells harboring DENV genomes encoding the P217A mutation triggered the IFNα response by pDCs ( i . e . , ≈36% relative to WT ) ( Figure 6D ) at various inducer cell concentrations ( Figure S7D ) and in proportion to the release of extracellular DENV RNA ( i . e . , ≈60% and 26% relative to WT at 24 and 48 hours post-transfection , respectively ) ( Figure 6C ) and viral structural proteins ( Figure S7C ) . Remarkably , the production of infectious virus ( Figure 6B ) was severely and disproportionally inhibited by the P217A mutation ( i . e . , 40-to-1 , 000 fold-reduction at 24-to-48 hours post-transfection ) as compared to the modest inhibition of the IFNα response by pDCs ( i . e . , ≈2 . 5 fold-reduction of IFNα response in the same time period ) ( Figure 6D and S7D ) . These results suggested that infectious virus production is not required and/or is not rate-limiting for pDC activation . Consistently , pDCs were not permissive to DENV infection ( Figure S4A ) , this latter observation is in line with the previous demonstration of pDCs as refractory to infection by other viruses [30] , [34] , [35] . Altogether , these results suggested that glycoprotein-dependent release of non-infectious viral components by DENV infected cells might trigger the IFNα response by contacting pDCs . To determine whether DENV surface proteins mediate the transmission of viral components to pDCs , we first assessed whether , similarly to DENV RNA ( Figure 5A ) , the DENV envelope proteins are transmitted into the pDCs , by inspection of consecutive Z-axis sections of DiD-labeled pDCs in co-culture with cells harboring the WT and DENV genomes encoding E protein mutations ( Figure S8 ) . Similar to DENV genome , we observed the E glycoproteins ( E GP ) in dot-like structures inside the pDCs . The frequencies of E GP dot-positive pDCs were elevated when in the co-cultures with either cells harboring the WT genome ( i . e . , around 90% ) or the P217A mutation ( i . e . , above 65% ) ( Figures 6E and S8 ) , which was in proportion to the release of extracellular DENV RNA ( i . e . , 60% relative to WT particles at 24 hours post-transfection ) ( Figures 6C ) . In contrast , cells harboring the DENV genome encoding the H244A mutation in E , which do not release viral particles and fail to trigger the IFN response by pDCs ( Figures 6B , 6C , 6D , S7C and S7D ) , demonstrated little to no transmission of the E GP into the pDCs ( Figures 6E and S8 ) . Because the intracellular levels of viral components ( i . e . , viral RNA , E and capsid proteins ) were equivalent for cells harboring DENV genomes encoding the H244A point mutant , as compared to WT genome ( Figures 6A and S7B ) , the results suggested that pDC IFNα production is activated by the glycoprotein-mediated transmission of viral components from DENV infected cells into contacting pDCs . Next , we tested the impact of expressing the DENV surface proteins alone ( Figure S9A ) on pDC IFN induction . Expression of the envelope proteins alone is known to result , in absence of nucleocapsid , in the release of viral envelope containing-membrane vesicles , the sub-viral particles ( SVPs ) ( Figure S9B ) [36] . Although the glycoproteins were readily transmitted from cells expressing only the DENV surface proteins to the co-cultured pDCs ( Figure S9D–F ) , IFNα production was not triggered ( Figure S9C ) . These observations are in agreement with the transmission of DENV RNA to pDCs and activation by the TLR7 RNA sensor ( Figure 5A and 5B ) . To corroborate these results , we determined whether pDC activation by contact with DENV infected cells requires an internalization-dependent mechanism by testing inhibitors of dynamin ( Dynasore ) [37] , of clathrin-mediated endocytosis ( Chlorpromazine [38] ) and of macropinocytosis ( Gö6983-PKC inhibitor [39] , [40] ) . Inhibitors of both dynamin and clathrin-mediated endocytosis , but not macropinocytosis , abrogated pDC IFNα production triggered by DENV infected cells ( Figure S10A ) , without any effect on the ongoing DENV replication and viral production ( Figures S10B–C ) . In addition , these inhibitors did not markedly impair pDC IFNα production induced by a TLR7 agonist ( Figure S10A ) , a cell-permeable imidazoquinoline , which passively diffuses inside the pDCs [41] , thus ruling out potential side-effect downstream of TLR7 recognition . These results , in agreement with the requirement of the endolysosome localized-sensor , TLR7 ( Figure 5B ) , suggested that pDC IFNα production triggered by DENV infected cells requires glycoprotein-mediated secretion of non-infectious viral components , which are subsequently internalized by co-cultured pDCs . These results demonstrated that pDC activation triggered by DENV infected cells is distinct from that induced by cells infected by other viruses , such as HCV , LCMV and classical swine fever virus ( CSFV ) , which does not require viral structural protein expression [4] , [5] , [7] . Next , we sought to study the regulation by cell contacts of DENV surface protein-dependent transfer and activation of pDCs . First , the cytoskeleton organization at the cell interface between pDCs and DENV infected cells was determined by confocal microscopy analysis . We observed an accumulation of the actin network at the cell contacts ( Figure 7A–E ) , while the microtubule network was not markedly modified at this location ( Figure S11A , left panel ) . In agreement with the importance of secreted structural components for pDC activation ( Figure 6 ) , specific immunostaining of non-permeabilized cells revealed that envelope proteins ( i . e . , E GP and prM ) were both present as clusters at the interface between pDCs and infected cells ( Figures 7F–Q and S12 ) . These observations prompted us to define the impact of the cytoskeleton network on cell contact-dependent pDC IFNα production . We showed that two inhibitors of the cytoskeleton network , Latrunculin B and Nocodazole ( i . e . , actin and microtubule depolymerizing drugs , respectively ) disrupted the actin network in co-cultures of pDC/DENV infected cell ( Figure 8A ) , consistent with previous reports [42] , [43] . As expected , the microtubule network was only perturbed by Nocodazole treatment ( Figure S11A ) [44] . By imaging flow cytometry analysis of GFP expressing-DENV infected cells co-cultured with pDCs ( stained by pDC marker CD123 ) ( Figures S13 ) , we showed that the frequency of conjugates between pDCs and DENV infected cells was greatly decreased by inhibitors of the cytoskeleton network ( Figure 8B and S13 ) . Both these inhibitors , in conjunction with the loss of actin accumulation at the contacts ( Figure 8A ) , impaired E glycoprotein clustering ( Figure 8C ) . Indeed , quantifications performed in a “double-blind” set-up revealed that , while E GP clustering was readily observed at the cell interface in untreated co-cultures ( i . e . , ≈60% of the pDCs at close proximity with DENV infected cells harboring E GP clustering ) , these frequencies were reduced to 15% for co-cultures treated with either inhibitor ( Figure 8D ) . Importantly , similar treatments inhibited IFNα production by the pDCs ( Figure 8E ) . Neither compound inhibited DENV RNA replication in the infected cells and infectious viral production ( Figures 8F–G ) , nor did they prevent the internalization ability of pDC , as assessed by membrane dye uptake ( Figure S11B ) . In addition , they did not inhibit pDC IFNα production triggered by a TLR7 agonist ( Figure 8E ) , thus ruling out potential nonspecific effects of these compounds on pDC responsiveness . Altogether , these results suggested that the cytoskeleton-dependent regulation of cell contacts and apposed GP clustering likely favors the subsequent activation of IFNα production by the pDCs . The phenotypic analysis of a virus production defective mutant ( i . e . , P217A ) ( Figure 6 ) revealed that infectious virus production is not required and/or is not rate-limiting for pDC activation . Like many viruses , DENV infected cells release immature non-infectious particles harboring uncleaved precursor membrane proteins ( prM ) , that are generated by inefficient cleavage of prM by the resident trans-Golgi protease furin [21] , [22] , [23] , [24] , [25] , [26] . To determine whether immature particles can serve as vehicles from DENV infected cells to contacting pDCs , we first determined the presence of prM protein dots inside co-cultured pDCs by using an antibody recognizing the pr peptide [45] and by examining consecutive Z-sections by confocal microscopy analysis . Dots of prM were observed inside pDCs co-cultured with DENV infected cells ( Figures S14A and S14C ) with very little background staining in pDCs co-cultured with uninfected control cells ( Figures S14B and S14C ) , suggesting that prM ( and/or pr peptide ) , along with the E GP ( Figures 6E , S8 and S9 ) are readily transferred to the pDCs . Next , to determine the ability of immature particles to convey immunostimulatory RNAs to pDCs , we tested the effects of DENV genomes encoding mutations in the furin cleavage site of the prM protein ( i . e . , the substitutions R88A , K90A and R91A ) , which , as expected from previous reports with single mutations [26] , failed to produce infectious virus ( Figure 9C ) . By contrast , RNA replication , intracellular viral protein expression ( Figures 9A , S7A and S7B ) , release of viral components ( Figures 9B and S7C ) , and transmission of viral components to the pDCs ( Figures 9D and S8D ) were maintained at levels comparable to the WT counterparts . Remarkably , the pDCs produced similar levels of IFNα when comparing contacting cells producing non-infectious immature virions vs WT DENV ( Figure 9E ) . Similar results were obtained when using various concentrations of cells harboring WT/mutant DENV genome ( Figure S7D ) . Therefore , these results suggested that cells producing immature particles potently trigger IFNα production by contacting pDCs . Next , to define the specific function of uncleaved prM-containing particles in pDC activation by DENV infected cells , we designed experiments aiming at modulating the levels of prM maturation . Firstly , we assessed the impact of an inhibitor of furin . As expected , this inhibitor markedly decreased the maturation of DENV particles , as shown by an increased prM∶E ratio measured by ELISA ( Figure 9F ) . The production of extracellular infectious DENV was also reduced in a dose-dependent manner upon furin treatment ( Figure 9H ) , while the levels of intracellular DENV RNA were unchanged ( Figure 9G ) . Remarkably , inhibition of prM cleavage enhanced IFNα productions by co-cultured pDCs in a dose-dependent manner ( Figure 9I ) . Increased pDC activation was observed despite a reduction in the release of physical particles , as shown by extracellular DENV RNA measurement ( Figure 9H ) . Altogether these results suggested that the activation of pDCs triggered by contacting infected cells inversely correlates with the levels of prM maturation . To further confirm these results , we studied the impact of furin up-regulation . As expected , cells overexpressing furin produced viral particles containing reduced prM∶E ratios ( i . e . , ≈10-fold reduction ) ( Figure 9J ) . The specific infectivity of DENV particles was increased upon furin overexpression ( i . e . , ≈3-fold increase in the ratios of infectivity to extracellular DENV RNA , comparing furin-overexpressing cells to counterpart control cells ) . Thus , cells overexpressing furin were compared to counterpart cells that produced either similar levels of intracellular and/or extracellular DENV RNA , or alternatively , similar production of infectious virus , by using different MOIs ( Figure 9K–L ) . Our results indicated that cells producing more mature particles were clearly impaired at triggering IFNα production by co-cultured pDCs ( Figure 9M ) . Altogether these results demonstrated that cells producing immature DENV particles are very potent at inducing IFNα production by pDCs , as compared to cells releasing mature virions . DENV has rapidly emerged in recent years as the most significant arboviral disease of humans , with greater than half of the world population at risk of infection [46] . Despite many years of research , the virus–host interactions that determine dengue pathogenesis are still incompletely understood [47] . Nonetheless , the self-limiting febrile symptoms observed in most DENV-contracted cases and the short course of illness suggest a key role for innate immune defenses in controlling DENV infection at early stages [18] . Accordingly , in vivo studies have demonstrated a critical role for type I IFNs in the host defense against DENV [16] , [18] , [19] . Furthermore , the activation of pDCs strongly correlates with the disease outcome of DENV infected patients [48] . Importantly , a study of children with DENV infections across a broad range of illness severities suggested that a blunted blood pDC response to systemic infection was associated with higher viremia levels and was a key step in the pathogenic cascade toward severe disease [49] . Although the activation mechanism and exact function are still elusive , altogether , these findings highlight the critical roles played by pDCs and the IFN response on disease progression in DENV infected individuals . Here , we revealed that DENV infected cells potently trigger IFNα secretion by non-permissive pDCs , a host response that bypasses the evasion from the innate response within infected cells . Furthermore , we demonstrated that TLR7-dependent IFNα production by pDCs in response to infected cells is concurrent with other hallmarks of innate immunity , such as inflammatory cytokine secretion , ISG up-regulation and pDC maturation . In agreement with our results , Rodriguez at al . showed that DENV-containing supernatants failed to trigger pDC IFNα production [50] , while other reports suggested that they triggered pDC activation [48] , [51] . This discordance may be explained by the preparation and concentration of supernatants and large number of pDCs that were used in the latter reports . Remarkably , the results of our study demonstrated that , despite continuous exposure to the infected cell milieu , physical separation from infected cells precludes the IFN response by pDCs . Consistently , strong pDC IFNα secretion was induced by co-cultured DENV infected cells ( i . e . , up to 0 . 5 µg/ml ) , indicating that cell-to-cell contact is a key feature of pDC activation . Interestingly , cell-to-cell transmission of immunostimulatory signals appears to be a common characteristic of pDC induction , as shown in this report for two members of the Flavivirus genus , DENV and WNV and as previously reported for other viruses , i . e . HCV , HIV , LCMV and CSFV [3] , [4] , [5] , [6] , [7] . Specifically , our previous results obtained in the context of HCV indicated that pDC stimulation occurs via viral RNA-containing exosomes . In this context , we suggested that the concentration of immunostimulatory exosomes in the supernatants was below an activating threshold for pDC stimulation , while this threshold might be reached in the intercellular space when cells are in contact [3] . Importantly , we showed here that viral structural components are detected in clusters at the interface between pDCs and infected cells . This finding suggests that cellular surface molecules and/or structures might concentrate the PAMP-carrier at the cell contacts , thereby enhancing transmission to pDCs . We further revealed that the actin network is pivotal for both this clustering of viral components at the pDC-infected cell interface , likely by regulating cell-to-cell contacts , and for pDC activation . Based on this observation , it is conceivable that the cytoskeleton structure serves as a platform contributing to the cell-to-cell transmission of viral components to the pDCs . Additional experiments will be required to test these hypotheses and to determine whether , for the various viruses that trigger the pDC IFN response in a cell-to-cell contact dependent manner , the mechanism of activation involves either common or distinct cellular factors and/or structures at the contacts . The mechanism we have identified is distinct from the conventional induction of the innate response , which typically occurs by the recognition of viral nucleic acids within infected cells [1] , [2] . Moreover , in contrast with the previously characterized induction of pDC IFNα production through contact with infected cells [3] , [4] , [5] , [7] , here we have defined a sensing pathway , which requires an E glycoprotein-dependent secretion of viral components , notably viral RNA , to trigger the pDCs . As such , it is different from the mechanism of induction by cells infected by other viruses , which does not require viral structural proteins [3] , [4] , [5] , [7] . Indeed , our results illustrate the crucial role of DENV envelope proteins in the induction of the innate response by neighboring IFN producer pDCs that are not permissive to infection . Importantly , our results revealed that cells producing uncleaved prM-containing immature particles triggered IFNα by pDCs more potently than cells efficiently producing fully mature virions . These immature particles are known to be deficient for the membrane fusion step , which occurs in the endo-lysosomal compartment during cell entry [52] , [53] , [54] . Interestingly , recognition of viral RNA by TLR7 sensor also takes place in this cellular compartment [1] , [55] . Therefore , based on these findings , we suggest a working model in which an extended retention within the endosomal compartment of fusogenic-deficient immature particles may favor the exposure of their viral genome for TLR7 recognition . In contrast , mature virions , which are fusion-competent , could escape from this compartment by membrane fusion . Additional experiments will be required to firmly validate and generalize this new concept . Several reports have demonstrated that a large proportion of uncleaved prM-containing immature particles are released from DENV infected cells , i . e . , 30-to-40% of viral particles [21] , [22] , [23] , [24] , [25] , [26] , on which prM content is variable on a per-particle basis [56] , [57] . Consistently , we showed that furin overexpression reduced the levels of immature virus , otherwise , produced by DENV infected cells , and concomitantly with reduced pDC IFN response . Although direct proof is still required , current evidence supports the in vivo existence of uncleaved prM-containing virus . Previous studies have demonstrated that a proportion of B cells isolated from DENV infected individuals produces monoclonal antibodies against prM [58] , [59] . In addition , the characterization of these anti-prM antibodies indicated that they are a major component of the serological response to DENV infection , leading to increased replication in Fc receptor-bearing cells via antibody-dependent enhancement ( ADE ) [56] , [58] , [60] . Importantly , our results illustrate a previously unsuspected function of these immature particles in innate immunity in mediating an IFN response by non-permissive bystander pDC . Indeed , the results of our study imply that the suboptimal furin-cleavage sequence , likely evolutionarily conserved to favor efficient export of infectious virus by preventing premature membrane fusion in the secretory pathway and cell entry of immature virus into Fc-receptor-expressing cells by ADE [21] , [22] , [23] , [27] , [28] , might also , by producing an IFN-inducer , contribute to regulate dengue pathogenesis . It is possible that pDC activation by infected cells elicits a strong local innate response that may lead to viral replication suppression or , alternatively , to the possible subsequent recruitment of DENV permissive cells and systemic viral spread . It is also conceivable that the interplay between pDCs and other cells regulating the innate responses , in turn , modulates this newly identified innate sensing mechanism of infected cells and/or the homing of pDCs to the infection site . Productive infection of cells with a wide range of enveloped viruses depends critically on the processing of the viral surface glycoproteins by cellular proteases [20] . Yet , depending on viral variants/strains , such cleavages might be limited by the differential requirement for certain host proteases , as their expression can be tissue-restricted . These selective requirements may contribute to their virulence , as proposed for influenza virus [61] . Additionally , suboptimal cleavage sites are evolutionarily maintained by sequence features , such as , e . g . , the presence of acidic residues or glycosylation sites adjacent to the cleavage site [23] , [62] . These events lead to the release of viral particles with uncleaved glycoproteins , as shown for viruses such as , e . g . measles virus [63] , influenza virus [61] , [64] , DENV and WNV [53] , [56] . Therefore our results , by uncovering a functional role of immature viral particles in innate immunity , may have broad implications for our understanding of the host-virus relationship . Huh-7 . 5 . 1 [65] , Vero E6 ( ATCC CRL-1586 ) , Hela ( ATCC CCL-2 ) and HEK-293T ( ATCC CRL-1573 ) cells were maintained in Dulbecco's modified Eagle medium ( DMEM ) ( Life Technologies ) supplemented with 10% FBS , 100 units ( U ) /ml penicillin , 100 mg/ml streptomycin , 2 mM L-glutamine and non-essential amino acids ( Life Technologies ) at 37°C/5% CO2 . BHK-21 cells ( ATCC CCL-10 ) were maintained in Eagle's MEM ( Life Technologies ) with the same supplements . pDCs were isolated from 450 ml of blood from healthy adult human volunteers which was obtained according to procedures approved by the “Etablissement Français du sang” ( EFS ) Committee . PBMCs were isolated using Ficoll-Hypaque density centrifugation . pDCs were positively selected from PBMCs using BDCA-4-magnetic beads ( MACS Miltenyi Biotec ) and cultured as previously described [3] . Monocytes were positively selected from pDC-depleted PBMCs using CD14-magnetic beads ( MACS Miltenyi Biotec ) according to the manufacturer's instructions , with a typical purity of 95% of CD11c-positive cells . CD14+ cells were then differentiated to monocyte-derived DCs ( mo-DCs ) by incubation for 6 days in RPMI 1640 medium supplemented with 10% FBS , 100 U/ml penicillin , 100 mg/ml streptomycin , 2 mM L-glutamine , non-essential amino acids , 1 mM sodium pyruvate and 0 . 05 mM βmercaptoethanol ( Sigma-Aldrich ) with 500 U/ml human granulocyte-macrophage colony-stimulated factor ( GM-CSF ) and 2 , 000 U/ml human interleukin 4 ( IL-4 ) ( MACS Miltenyi Biotec ) , as previously described [12] . To generate monocyte-derived macrophages , monocytes were cultured in the same medium as for the mo-DCs with 500 U/ml GM-CSF for 6 days . The antibodies used for immunoblotting were mouse anti-E glycoprotein ( 4G2 and 3H5 ) kindly provided by P . Despres ( Pasteur Institut , Paris , France ) ; mouse anti-capsid ( 6F3 ) kindly provided by J . Aaskov ( Queensland University of Technology , Brisbane , Australia ) ; mouse anti-actin ( AC74 , Sigma Aldrich ) . The antibodies used for immunostaining were mouse PE-conjugated anti-CD123 , mouse APC-conjugated anti-BDCA-2 , mouse APC-conjugated anti-IFNα ( MACS Miltenyi Biotec ) , mouse PerCP-conjugated anti-CD83 ( eBioscience ) , and mouse APC-conjugated anti-CD86 ( BD Bioscience ) , mouse anti-DENV prM ( clone DM-1 , Abcam ) , mouse anti-alpha tubulin ( DM1A , Sigma Aldrich ) ; Ficoll-Hypaque ( GE Healthcare Life Sciences ) ; LPS , TLR7 agonist ( R848 ) and TLR9 agonist ( ODN2216 ) ( Invivogen ) ; TLR7 antagonist , IRS661 ( 5′-TGCTTGCAAGCTTGCAAGCA-3′ ) synthesized on a phosphorothionate backbone ( MWG Biotech ) ; Fc Blocking solution ( MACS Miltenyi Biotec ) ; Golgi-Plug and permeabilization-wash solution ( BD Bioscience ) ; IFNα and IFNβ ELISA kit ( PBL Interferon Source ) ; IL-6 and TNFα ELISA kit ( Affymetrix , eBioscience ) ; Lipofectamine 2000 ( Life Technologies ) ; 96-well format transwell chambers ( Corning ) ; LabTek II Chamber Slide System , 96-Well Optical-Bottom Plates and Nunc UpCell 96F Microwell Plate ( Thermo Fisher Scientific ) ; CF488A-conjugated phalloidin ( Biotium ) ; Vibrant cell-labeling solution ( CM-DiI , Life Technologies ) ; Hoescht and Alexas-conjugated secondary antibodies ( Life Technologies ) ; iScript cDNA synthesis kit ( Biorad ) , qPCR kit ( Life Technologies ) . Latrunculin B , nocodazole , chlorpromazine , dynasore and Gö6983-PKC were purchased from Sigma-Aldrich . Viral stocks of the prototypic DENV-2 strain New Guinea C ( NGC ) ( AF038403 ) were produced using in vitro RNA transcripts prepared from DENV-2 infectious plasmid clone pDVWS601 plasmid [66] linearized with XbaI and using mMESSAGE mMACHINE T7 Kit ( Ambion ) . In vitro transcribed RNA was introduced into BHK-21 cells by electroporation as previously described [31] . Briefly , 5 µg of in vitro transcribed RNA was used to transfect 4×106 cells by electroporation . Six hours post-transfection , the culture medium was refreshed . Virus containing supernatants collected at 3 days post-electroporation were clarified through a 0 . 45 µm filter ( Corning ) . Viral stocks of WNV ( lineage II , strain 956 D117 3B ) [67] were produced by transfection of 2×106 HEK-293T cells with 4 µg of the plasmid pWNII-GFP [67] using the Lipofectamine 2000 transfection reagent ( Invitrogen ) in optiMEM . Six hours post-transfection , the medium was refreshed with HEK-293T culture medium . Virus containing supernatants collected 48 h post-transfection were clarified through a 0 . 45 µm filter ( Corning ) . Viral stocks of vesicular stomatitis virus ( VSV-GFP , infectious titer of ≈109 Tissue Culture Infectious Dose ( TCID50 ) /ml ) were produced as previously described [68] and kindly provided by Dr J . Perrault ( Department of Biology , Center for Microbial Sciences , San Diego State University , CA , US ) . Viral stocks of Influenza A Virus ( FluAV , A/WSN/33 strain , delta NS1 , i . e . , infectious titer of ≈106 plaque forming unit ( PFU ) /ml ) were produced as previously described [69] and kindly provided by Dr R . Le Goffic ( Unite de Virologie et Immunologie Moleculaires , Jouy-en-Josas , France ) . Sixteen hours prior to co-culture with pDCs , Huh7 . 5 cells , which are known to be deficient for the RIG-I signaling pathway [70] , were used to rule out the confounding contribution of IFNα production by the infected cells themselves and infected at MOI of 0 . 1 and 0 . 5 for VSV-GFP and FluAV , respectively . Introduction of mutations into the genomic length DENV-2 strain NGC cDNA clone pDVWS601 , encoding amino acid substitutions into the E glycoprotein ( i . e . , H244A , D215A , P217A ) and NS5 ( i . e . , Rep−/− , containing the multiple amino acid substitutions G81A , G83A and G85A ) have been described previously [31] , [32] . Mutations encoding amino acid substitutions in prM ( amino acids R88A/K90A/R91A ) and an in frame four amino acid deletion in the capsid ( amino acids V51-to-L54 ) were first introduced into DENV-2 subgenomic cDNA fragments by overlap-PCR ( OL-PCR ) using mutagenic primers . The sequences of the primers are described in the Table S1 . The OL-PCR fragments were purified and cleaved with BsrGI and SphI and then transferred into the pDVWS601 plasmid that had been cleaved with the corresponding restriction enzymes . The presence of the mutations and sequence of the PCR derived regions were confirmed by sequencing . In vitro RNA transcripts were prepared from the parental and mutated pDVWS601 plasmids as described above and transfected into Huh7 . 5 . 1 cells using the Lipofectamine 2000 transfection reagent ( Life Technologies ) , according to the manufacturer's instruction . One µg of RNA was used to transfect a 60% confluent cell monolayer contained in a single well of a 6-well plate following the manufacturer's protocol . Six hours post-transfection , the cells were either harvested for the quantification of viral RNA ( 6 hour time point ) or washed 3 times with PBS and fresh culture medium added to the cells for additional incubation times . At 24 and 48 hours post-transfection , the cells were harvested for the determination of RNA and protein levels and their supernatants collected for the quantification of viral RNA and infectious titer or concentrated by ultracentrifugation for the determination of protein levels by Western blot . At 24 hours post-transfection , the cells were harvested and co-cultured with isolated pDCs for 18–20 hours . Forty-eight hours prior to co-culture , cells were infected at a MOI of 3 using a viral stock of WNV or DENV . Unless otherwise indicated , 2×104 pDCs were co-cultured with 105 infected cells , transfected cells or uninfected parental cells , or treated with 100 µl of supernatant from the latter cells in a 200 µl final volume in 96-well round-bottom plates incubated at 37°C/5% CO2 . Eighteen to twenty hours later , cell-culture supernatants were collected and the levels of IFNα , IFNβ , TNFα and IL-6 were measured using a commercially available ELISA kits specific for IFNα and IFNβ ( PBL Interferon Source ) , TNFα and IL-6 ( Affymetrix ) , following the manufacturer's instructions . When indicated , 105 infected cells or uninfected cells were co-cultured with 3×104 pDCs or with 105 naïve recipient cells , as indicated , in 96-well format transwell chambers separated by a 0 . 4 µm membrane ( Corning ) . At the indicated times , cells were harvested and resuspended using 0 . 48 mM EDTA-PBS solution ( Life Technologies ) . After incubation with Fc receptor blocking reagent ( MACS Miltenyi Biotec ) for 10 minutes at 4°C , surface staining of pDC markers , CD123 and BDCA-2 and/or the cell differentiation markers CD83 and CD86 were detected by a 40 minute incubation at 4°C with 5 µg/mL of the indicated combinations of PE-conjugated mouse anti-CD123 , APC-conjugated anti-BDCA-2 , PerCP-conjugated anti-CD83 , and APC conjugated anti-CD86 , respectively , diluted in staining buffer ( PBS without calcium and magnesium , with 2% FBS ) , followed by PBS washes . Cells were then fixed by incubation for 20 minutes at room temperature with 4% paraformaldehyde , followed by 20 minutes incubation with 0 . 1 M glycin-PBS at room temperature and two PBS washes . For intracellular-immunostaining of IFNα , cocultivated cells were treated with 1 µl/ml GolgiPlug solution ( BD Bioscience ) before collection . After fixation and CD123-staining steps , IFNα was detected by a 40 minute incubation with APC-conjugated mouse anti- IFNα ( MACS Miltenyi Biotec ) diluted at 1∶10 in permeabilization buffer ( BD Bioscience ) . Cells were then washed twice with permeabilization buffer and resuspended in staining buffer . Flow cytometric analysis was performed using a Digital LSR II , and the data were analyzed with Flow Jo software ( Tree Star ) . The corresponding control isotypes served to define the specific signal . After isolation , 5×104 pDCs were stained by using 0 . 5 µM Vibrant cell-labeling solution ( CM-DiI , Life Technologies ) by successive incubations for 10 and 15 minutes at 37°C and 4°C respectively . Labeled pDCs were washed twice with PBS and then co-cultured with pre-plated DENV infected cells for 5 hours at 37°C in glass bottom 96 well-plate ( Fisher Scientific ) , pretreated with poly-L-lysine at 8 µg/mL . After 4% PFA fixation at room temperature and PBS washing , DENV plus strand RNA was detected using a probe set that targets a region between nucleotide positions 8437-to-9685 in the DENV-2 NGC genome ( Panomics/Affymetrix ) according to the manufacturer's instructions . For IFNα immunostaining , the cells were permeabilized by incubation for 7 minutes in PBS containing 0 . 3% ( v/v ) Triton - and 3% ( w/v ) BSA , then incubated with mouse anti-IFNα antibody ( MACS Miltenyi Biotec ) at 2 µg/ml in PBS containing 3% BSA for 40 minutes at room temperature , followed by an incubation with Alexa 647-conjugated anti-mouse antibody ( Life Technologies ) and Hoechst dye for 40 minutes at room temperature . As controls , FISH detection of DENV RNA were performed in co-cultures of pDCs with non-infected cells and in co-cultures of pDCs with DENV infected cells by omitting DENV-specific probe and by following the same procedure of hybridization and immunostaining . Images were acquired with a Zeiss LSM 710 laser scanning confocal microscope and analyzed with Image J ( http://rsb . info . nih . gov/ij ) and IMARIS ( Bitplane Inc . ) software packages . After immune-isolation , pDCs were stained with 0 . 5 µM Vibrant cell-labeling solution ( CM-DiI ) as above-described . 4-to-5×104 DiI-labeled pDCs ( DiI-pDCs ) were co-cultured with 4-to-5×104 DENV infected Huh7 . 5 . 1 cells for 8 hours at 37°C . For analysis of DENV E and prM transfer into pDCs and cell contacts , co-cultures were performed in LabTek II Chamber Slide System ( Nunc ) . After 4% PFA fixation and three PBS washes , cells were permeabilized 7 min with 0 . 1% Triton in PBS prior immunostaining . For analysis of DENV surface protein clustering at the cell contacts , co-cultures were incubated in a 96-Well Optical-Bottom Plates . After 4% PFA fixation and three PBS washes , immunostainings were performed without permeabilization step , as previously described [71] . After blocking step ( PBS 3% BSA ) actin filaments were stained with CF488A-conjugated phalloidin ( Biotium ) at 1 . 25 U/mL , α-tubulin was stained with mouse anti-α tubulin ( DM1A clone , from Sigma ) at 1∶2000-dilution , DENV E glycoproteins were detected using anti-E antibody ( 3H5 clone ) at 1∶500-dilution and anti-PrM antibody ( DM-1 clone , Abcam ) at 1∶50-dilution and IFNα was detected by a mouse anti-IFNα ( Miltenyi ) at 1∶20-dilution . Antibodies were diluted in 3% BSA-PBS and added to the cell for 1 hour incubation at room temperature . After three PBS-washes with PBS , cells were incubated with an Alexa 647-conjugated-anti-mouse antibody ( for detection of anti-α-tubulin and anti-E antibodies ) or Alexa 488-conjugated-anti-mouse ( for detection of anti-PrM antibody ) at 1∶1000-dilution in 3% BSA-PBS , added to the cells along with Hoechst diluted at 1∶500 ( Molecular Probes ) for 1 hour incubation at room temperature . After three washes with PBS , cells in 96 wells plate were directly observed and cells in Labtek were mounted with mowiol prior observation . Images were acquired with a Zeiss LSM 710 laser scanning confocal microscope and analyzed with Image J ( http://rsb . info . nih . gov/ij ) and IMARIS ( Bitplane Inc . ) software packages . RNAs were isolated from cells or supernatants harvested in guanidinium thiocyanate citrate buffer ( GTC ) by phenol/chloroform extraction procedure as previously [3] . The efficiency of RNA extraction and reverse transcription-real-time quantitative PCR ( RT-qPCR ) was controlled by the addition of carrier RNAs encoding Xef1α ( xenopus transcription factor 1α ) in vitro transcripts in supernatants diluted in GTC buffer . DENV RNA and Xef1α and glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ) mRNA levels were determined by RT-qPCR using an iScript RT kit ( Biorad ) and a One-Step PCR Master Mix kit for qPCR and analyzed using StepOnePlus Real-Time PCR system ( Life Technologies ) . The sequences of the primers used for the RT-qPCR are described in Table S1 . Extracellular and intracellular DENV RNA levels were normalized for Xef1α and GAPDH RNA levels , respectively . Infectivity titers in supernatants were determined by end-point dilution using Huh 7 . 5 . 1 cells . Foci forming unit ( ffu ) were detected 72 hours after infection by GFP expression for WNV and anti-E glycoprotein specific immunofluorescence for DENV . Briefly , Huh 7 . 5 . 1 cells were fixed with 4% PFA and permeabilization by incubation for 7 minutes in PBS containing 0 . 1% Triton . Cells were then blocked in PBS containing 3% BSA for 15 minutes and incubated for 1 hour with mouse anti-E glycoprotein ( clone 3H5 ) hybridoma supernatant diluted at 1∶200 in PBS containing 1% BSA . After 3 washes with PBS , cells were incubated 1 hour with secondary Alexa 555-conjugated anti-mouse antibody ( 1∶1'000-dilution ) and Hoechst dye ( 1∶1'000-dilution ) in PBS containing 1% BSA . Percentage of E-positive cells and GFP expressing cells was determined using a Zeiss Axiovert 135 microscope . Viral supernatant were filtrated through a 0 . 45 µm filter ( Corning ) and concentrated prior to Western blot analysis by ultracentrifugation at 110 , 000× g for 2 hours at 4°C using a SW41 rotor . The pellets were re-suspended in PBS . Viral pellets and cell lysates were extracted using lysis buffer ( 150 mM NaCl 50 mM Tris HCl pH 8 , 1% NP40 , 0 . 5% Deoxycholate , 0 . 1% Sodium dodecyl sulfate ) and analyzed by Western blotting using hybridoma supernatant-containing anti-E ( 4G2 ) and anti-capsid ( 6F3 ) at the dilution of 1∶500 and actin at 1 µg/ml followed by secondary horse radish peroxidase-coupled antibodies and chemiluminescence . Huh 7 . 5 . 1 cells were transduced with retroviral based vector pseudotyped with VSV glycoprotein to stably express GFP , as previously described [72] . Forty-eight hours prior co-culture with pDCs , GFP-expressing Huh 7 . 5 . 1 cells were infected at a MOI of 3 using a viral stock of DENV . 105 GFP-expressing DENV infected cells were co-cultured with 3×104 pDCs in low-adherence micro-plate designed for cell harvesting by temperature reduction ( Nunc UpCell 96F Microwell Plate from Thermo Scientific ) for 5 hours at 37°C in presence , or not , of Latrunculin B and Nocodazole ( 1 µM ) , as indicated . After 4% PFA fixation , co-cultured cells were harvested by equivalent multi-pipetting at room temperature and washed three times with staining buffer ( PBS without calcium and magnesium with 2% FBS ) . After incubation with Fc receptor blocking reagent ( MACS Miltenyi Biotec ) for 10 minutes at 4°C , surface staining of a pDC marker , CD123 , was detected by a 40 minute incubation at 4°C with 5 µg/mL of APC-conjugated mouse anti-CD123 , diluted in staining buffer , followed by washes with staining buffer . Co-cultured cells were analyzed by Image Stream X technology ( Amnis ) at magnification ×60 using IDEAS software . The cell population defined as pDC/DENV cell conjugates comprises conjugates of at least one CD123+ cell and at least one cell solely GFP+ cell among the total of APC+ cells , GFP+ cells and conjugates . The cell populations were sorted by using masks ( IDEAS software ) to eliminate i/the non-specific signals i . e . , double positive single cells and ii/cells with background levels for APC signal . Post-cell sorting , the accuracy of the gated cell population in regards to the defined criteria was controlled by a visual inspection of the individual pictures in the gated cells population ( i . e . , assessment with 90 randomly picked pictures of the population defined as conjugates ) . The percentages of gated single cells or conjugates with an accurate phenotype according to the defined criteria among the total of examined pictures per category of cell population were: 97% for GFP+ gated population , 99% for APC-CD123+ gated population and 89% for conjugates . 293T cells , which stably express furin , were generated by transfection using Polyethylenimine and selected using hygromycin ( at 5 µg/ml ) . The decRRVKR-CMK inhibitor ( Calbiochem ) was used to inhibit the Furin activity in Huh7 . 5 . 1 co-cultured with the pDCs , as the indicated concentrations . The levels of prM maturation were analyzed by detection of E and prM by ELISA , as previously described [58] . Briefly serial dilutions of viral supernatants were incubated on anti-E ( 4G2 ) antibody coated 96-well plate . Then , E and prM were detected by using a humanized version of 3H5 mAb ( hu3H5 ) and anti-prM , respectively . The prM∶E ratios were calculated using the viral supernatant dilution with E detection in the linear range . DENV-2 NGC prM and E genes were cloned under the control of CMV promoter , by amplification from pSVprME [73] using primers ADVprME_Fwd ( GATCCCCGGGACCGCCACCATGGTGAA ) and ADVprME_REV ( GATCCCCGGGAGCTTGATATCAGGCCTGC ) and cloned into the Sma I site of the adenovirus shuttle vector pDC104 under the control of the CMV promoter to produce pAdvprME . AdvprME was transfected into cells using the Xtreme-GENE HP DNA Transfection Reagent , follow the manufacturer's instructions . Six hours post-transfection , the cells were washed with PBS and fresh culture medium added to the cells for additional incubation times . At 48 hours post-transfection , the cells were harvested and co-cultured with isolated pDCs for 18–20 hours . Parallel determination of intracellular protein levels by Western blot in harvested cells and their supernatants concentrated by using vivaspin concentrator with centrifugation at 3000 g for 30 min ( cut-off 100 KDa , Sartorius ) . Paired Student's t-test was used to analyze data . Data considered significant demonstrated p-values less than 0 . 05 . Data were also analyzed using a two ways non-parametrical analysis of variance ( ANOVA ) , followed by comparison with Levene Test , analyzed with xlstat software . Triangles indicate the experimental conditions that belong to a separated group statistically different from the others .
Viral recognition by the host often triggers an antiviral state , which suppresses viral spread and imparts adaptive immunity . Like many viruses , dengue virus ( DENV ) defeats the host-sensing pathway within infected cells . However , in vivo studies have demonstrated a key role of innate immunity in controlling DENV infection . Here we report that sensing of DENV-infected cells by non-permissive innate immune cells , the plasmacytoid dendritic cells ( pDCs ) , triggers a cell-contact- and TLR7-dependent activation of a strong antiviral IFN response . This cell-to-cell sensing involves transmission of viral elements that are clustered at the interface between pDCs and infected cells and is regulated by the actin network . Importantly , we revealed that uncleaved prM surface protein-containing immature particles play a key function in stimulating the innate immune response . These non-infectious immature particles are released by infected cells as a consequence of a suboptimal cleavage site , which is an evolutionarily conserved viral feature that likely favors the export of infectious virus by prevention of premature membrane fusion in the secretory pathway . Therefore our results highlight a conceptually novel trade-off between efficient infectious virus release and the production of IFN-inducing particles . This concept may have broad importance for the many viruses that , like DENV , can disable the pathogen-sensing machinery within infected cells and can release uncleaved glycoprotein-containing non-infectious particles .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "innate", "immune", "system", "immunity", "virology", "emerging", "viral", "diseases", "biology", "and", "life", "sciences", "immunology", "microbiology", "immune", "system" ]
2014
Sensing of Immature Particles Produced by Dengue Virus Infected Cells Induces an Antiviral Response by Plasmacytoid Dendritic Cells
Pathogens that traffic in blood , lymphatics , or interstitial fluids must adopt strategies to evade innate immune defenses , notably the complement system . Through recruitment of host regulators of complement to their surface , many pathogens are able to escape complement-mediated attack . The Lyme disease spirochete , Borrelia burgdorferi , produces a number of surface proteins that bind to factor H related molecules , which function as the dominant negative regulator of the alternative pathway of complement . Relatively less is known about how B . burgdorferi evades the classical pathway of complement despite the observation that some sensu lato strains are sensitive to classical pathway activation . Here we report that the borrelial lipoprotein BBK32 potently and specifically inhibits the classical pathway by binding with high affinity to the initiating C1 complex of complement . In addition , B . burgdorferi cells that produce BBK32 on their surface bind to both C1 and C1r and a serum sensitive derivative of B . burgdorferi is protected from killing via the classical pathway in a BBK32-dependent manner . Subsequent biochemical and biophysical approaches localized the anti-complement activity of BBK32 to its globular C-terminal domain . Mechanistic studies reveal that BBK32 acts by entrapping C1 in its zymogen form by binding and inhibiting the C1 subcomponent , C1r , which serves as the initiating serine protease of the classical pathway . To our knowledge this is the first report of a spirochetal protein acting as a direct inhibitor of the classical pathway and is the only example of a biomolecule capable of specifically and noncovalently inhibiting C1/C1r . By identifying a unique mode of complement evasion this study greatly enhances our understanding of how pathogens subvert and potentially manipulate host innate immune systems . Complement is an interconnected system of serum and cell surface proteins that comprises a primary arm of innate immunity . Complement acts in the recognition and clearance of microbial invaders and terminal activation of the complement cascade can result in direct lysis of pathogens . In addition to its function as a ‘first-line-of-defense’ , complement plays an extensive role in mediating inflammatory responses , maintaining cell homeostasis , and in coordinating tissue development and repair [1] . Complement also acts as a crucial bridge between the innate and adaptive immune systems by promoting B-cell differentiation and regulating T-cell immunity [2 , 3] . The complement system is initiated by one of three pathways , termed the classical pathway ( CP ) , lectin pathway ( LP ) , and alternative pathway ( AP ) , which are defined by their mode of pattern recognition . The CP has traditionally been called the ‘antibody-dependent’ pathway of complement owing to its activation by antigen-bound antibodies of the IgG or IgM types . However , the CP is also activated by binding of the complement protein C1q to a diverse set of microbial and apoptotic cell surface structures [4] . C1q serves as the pattern recognition molecule of the CP and is itself an integral subunit of the CP initiator zymogen known as the first component of human complement , C1 . C1 is a large ( ~790 kDa ) multiprotein complex formed by the interaction of one C1q molecule , two C1r protease molecules and two C1s protease molecules . In the absence of calcium , C1 quickly dissociates , as calcium ions stabilize the formation of the heteromeric C1r2C1s2 complex as well as its interaction with C1q . Binding of C1q to pathogens or altered-self surfaces triggers the autocatalysis of C1r which in turn cleaves C1s proenzyme to form fully activated C1 [5] . In contrast , the LP is initiated via recognition of pathogen-associated molecular patterns ( PAMPs ) by mannan-binding lectin ( MBL ) or ficolins , which in turn bind and autoactivate mannan-binding lectin associated serine proteases ( MASPs ) [6] . Finally , the AP is continuously activated via constitutive low-level hydrolysis of complement protein C3 [7] . Independent of initiation mode , all pathways converge at the level of complement component C3 , which is cleaved by multi-component enzymes called C3 convertases . The CP C3 convertase arises when activated C1 cleaves complement component C4 and C2 forming C4bC2a . The LP intersects with the CP at this step as activated MASP-2 also cleaves C4 and C2 . On the other hand , the predominant AP C3 convertase ( C3bBb ) is formed by interaction of C3b , complement factor B , and the protease complement factor D . As their name implies , C3 convertases bind and cleave C3 resulting in the proteolytic release of the small soluble anaphylotoxin C3a , which acts to stimulate recruitment of neutrophils , monocytes and macrophages to the sites of activation . The second byproduct of C3 cleavage , C3b , covalently attaches to surfaces near the activation site where it serves as the central opsonin of the cascade eventually leading to the downstream effector functions of complement , including direct killing by formation of the terminal complement complex which has also been termed the membrane attack complex . Host cells express several soluble and membrane associated proteins collectively referred to as regulators of complement activation , which are critical in preventing complement attack on healthy host tissue . For example , the CP is down regulated on host cells by the action of C4b-binding protein which interferes with the formation as well as accelerates the decay of the CP/LP C3 convertase while simultaneously serving as a cofactor for the complement factor I mediated degradation of C4b [8] . Complement factor H serves an analogous role in the regulation of the AP C3 convertase and in doing so acts as the dominant negative regulator of complement amplification [9] . Blood-borne pathogens , or those that traffic in interstitial fluid or lymphatics , which lack endogenous regulators of complement activity , must adopt strategies to successfully evade complement attack . One such pathogen , Borrelia burgdorferi , is the etiologic agent of Lyme disease and is the leading cause of vector-borne illness in the United States according to the Centers for Disease Control and Prevention ( CDC ) . Lyme disease is often accompanied by a local erythema migrans lesion and can lead to severe clinical outcomes such as carditis , neurological dysfunction , and arthritis [10 , 11] . B . burgdorferi is transmitted to humans via the bite of infected hard ticks . During the ticks blood meal spirochetes enter the mammalian host and subsequently disseminate to remote tissues [10 , 11] . If therapeutic intervention is not sought , B . burgdorferi is able to persistently colonize a large number of tissues including joint , skin , heart , and the central nervous system [10 , 11] . B . burgdorferi appears to avoid complement-mediated killing from the AP by expressing a group of virulence factors known as Csp proteins ( CspA and CspZ ) and those from the OspE/F family [12–18] . These proteins are also referred to as complement regulator-acquiring surface proteins ( CRASPs ) [19 , 20] . These bacterial surface proteins recruit human factor H , factor H-like protein 1 , and factor H-related proteins , which serve as the major endogenous negative regulators of the AP [12 , 13 , 20–23] . In addition , human factor H is also recruited to the surface of relapsing fever Borrelia spp . where similar AP inhibition would occur [24 , 25] . By hijacking these key host complement regulatory molecules , B . burgdorferi , as well as and other B . burgdorferi sensu lato isolates , subverts the deleterious effects of AP activation . Activation of the CP has previously been shown for Lyme disease spirochetes [26 , 27] and studies employing mouse models deficient in factor H , factor B , or C3 have shown that the CP and/or LP play significant roles in controlling early stages of borrelial infection [28] . Indeed , the importance of spirochetal strategies to subvert CP activation are underscored by the ability of B . burgdorferi as well as the relapsing fever spirochetes B . recurrentis and B . duttonii to recruit the host CP regulators C4b-binding protein and/or C1 esterase inhibitor ( C1-INH ) to their surface via interactions with specific borrelial lipoproteins [29–31] . Herein we report the identification of the borrelial lipoprotein BBK32 as a potent and specific inhibitor of the CP capable of forming high-affinity interaction with C1 . We go on to localize the anti-complement activity of BBK32 to the C-terminal region and demonstrate a molecular mechanism by which BBK32 noncovalently inactivates the central CP initiating serine protease C1r . To our knowledge , BBK32 represents the first example of a C1r specific inhibitor of biomolecular origin and is the first noncovalent protein inhibitor of the C1 complex to be described . Thus , this work significantly expands our knowledge of how pathogens recognize and evade human innate immunity by defining a new mechanism by which the pathogen B . burgdorferi prevents activation of the classical pathway of complement . In light of the apparent ability of B . burgdorferi to suppress the CP ( discussed above ) , we hypothesized that novel interactions exist between B . burgdorferi surface proteins and the CP initiating enzyme complex , complement C1 . To explore this hypothesis we adopted a Far Western approach designed to probe for interaction of B . burgdorferi B31 lysate proteins with C1 . The initial profile showed that biotinylated C1 specifically recognizes borrelial proteins with apparent molecular masses of 17 , 28 , and 48 kDa ( Fig 1A ) . We also tested the reactivity of lysates harvested from cultures grown under conditions that require the Rrp2-RpoN-RpoS regulatory system for their induction [32–35] ( e . g . , pH 6 . 8 relative to conventional growth conditions of pH 7 . 6 ) and thus mimic , in part , some aspects of the mammalian host environment [32–37] that stimulate bbk32 expression . Given that the 48-kDa band was near the known SDS-PAGE migration position of the borrelial lipoprotein BBK32 [38] , we also tested lysates originating from cells that lacked an intact bbk32 locus ( bbk32::StrR; [39] ) ( Fig 1A ) . These experiments show that the 48-kDa band is inducible under mammalian host-like conditions and is also completely absent from bbk32::StrR lysates . The two lower molecular weight bands did not change in intensity in response to altered growth conditions and remained present in the B . burgdorferi lacking intact bbk32 , indicating that they were not proteolytic fragments of BBK32 and instead are distinct protein species ( Fig 1A ) . Immunoblot analysis also confirmed that the 48-kDa C1-reactive species co-migrated with BBK32 in the parent strain and was missing in the bbk32::StrR mutant ( Fig 1B ) . Next , we were interested in determining if the C1-binding proteins were surface exposed in B . burgdorferi . To this end , we used a proteinase K accessibility assay that serves as a readout for the surface localization of borrelial proteins [40 , 41] . The results indicated that all three C1 reactive proteins were eliminated following protease treatment suggesting that they are all on the outer surface of B . burgdorferi ( Fig 1C ) . To address the integrity of the B . burgdorferi cells , we tested whether the subsurface endoflagellar structural protein FlaB was affected by the addition of proteinase K . The levels of FlaB between mock and protease treated cells were not different ( Fig 1D ) , indicating that the B . burgdorferi cells used in this experiment were structurally intact . The identity of the 17 and 28-kDa proteins targeted by biotinylated C1 in the Far Western assay remain unclear at this time . However , in addition to gel migratory position , three lines of evidence suggested that the identity of the 48-kDa band was indeed BBK32; ( i ) it is induced under mammalian-like conditions ( Fig 1A ) , ( ii ) it is surface exposed ( Fig 1C ) , and ( iii ) it is not detected in samples containing the bbk32::StrR allele ( Fig 1A and 1B ) . To further investigate a potential C1/BBK32 interaction we next produced a recombinant form of BBK32 ( residues 21 to 354 ) which lacks only the 20 residue signal peptide [38] , and immobilized this “full-length” BBK32 ( referred to as BBK32-FL hereafter ) protein on the surface of a Biacore sensor chip . Surface plasmon resonance ( SPR ) was used to quantitatively measure the interaction of purified C1 with immobilized BBK32-FL in a running buffer of HBS-T-Ca2+ . Strong and dose-dependent binding was observed ( Fig 2A ) and kinetic evaluation of the resulting sensorgrams indicates that BBK32-FL binds C1 with high affinity , as a dissociation constant ( KD ) of 3 . 9 nM ( Table 1 ) was calculated . BBK32 has long been known for its ability to bind the extracellular matrix glycoprotein fibronectin ( Fn ) [38 , 42 , 43] . The Fn/BBK32 interaction is primarily mediated by anti-parallel β-strand addition of residues 126–190 originating from the intrinsically disordered N-terminal region of BBK32 with Fn type I domains along an extended region of Fn [42 , 44] . In addition to Fn , glycosaminoglycans ( GAGs ) also act as a mammalian host ligand for BBK32 and are recognized by a distinct N-terminal binding site ( BBK32 residues 45–68 ) [45] . In contrast , the C-terminal region of BBK32 ( residues 206–354 and hereafter referred to as BBK32-C ) is a highly basic globular domain rich in α-helical secondary structure [46] , which to date has not been ascribed a specific function . To determine if the binding site for C1 on BBK32 could be localized to the N-terminal region of BBK32 ( residues 21–205 , and hereafter referred to as BBK32-N ) or the globular C-terminal region , we next immobilized recombinant versions of BBK32-N and BBK32-C on the surface of an SPR sensor chip . Intriguingly , no detectable response was measured when C1 was injected over the BBK32-N fragment ( Fig 2C ) ; however , the C-terminal domain of BBK32 retained a high-affinity for C1 ( KD = 2 . 3 nM ) ( Fig 2B , Table 1 ) . Furthermore , BBK32-C has a very similar kinetic profile ( ka = 8 . 4 x 104 M-1s-1 , kd = 1 . 9 x 10−4 s-1 ) to that measured for C1/BBK32-FL ( ka = 7 . 2 x 104 M-1s-1 , kd = 2 . 8 x 10−4 s-1 ) . The interaction of both BBK32-FL and BBK32-C with C1 was strongly dependent on calcium as binding was nearly abolished when C1 was injected in a buffer containing the calcium chelator ethylene glycol tetraacetic acid ( EGTA ) ( Fig 2D–2F ) . Taken together , these results demonstrate that the C-terminal region of BBK32 forms a tight calcium-dependent interaction with human complement C1 . Given the apparent high-affinity interaction between BBK32 and C1 we sought to determine if BBK32 modulates complement activation . To this end , we evaluated the effect of various concentrations of BBK32-FL , BBK32-N and BBK32-C in an ELISA-based assay of complement function . When conditions specific for CP activation were used , both BBK32-FL ( IC50 , C3b = 34 nM , IC50 , C4b = 34 nM ) and BBK32-C ( IC50 , C3b = 4 . 7 nM , IC50 , C4b = 5 . 6 nM ) potently inhibited the generation of the downstream complement activation products C3b and C4b ( Fig 3A and 3B , Table 2 ) . In contrast , BBK32-N failed to inhibit CP activation at any concentration used . Importantly , when conditions were used to selectively activate the AP or LP , no significant effect could be measured at concentrations up to 1 μM of BBK32-FL or BBK32-C ( Fig 3C ) . To further investigate their complement inhibitory activities we next measured residual complement-mediated hemolysis in the presence of 1 μM BBK32 proteins . Strikingly , 1 μM BBK32-FL or BBK32-C conferred nearly 100% protection to sensitized sheep red blood cells in the presence of complement ( i . e . , normal human serum ) using a standard assay of CP-mediated hemolysis ( CP50 ) ( Fig 4A ) . Furthermore , BBK32 inhibited the CP in a dose-dependent manner with calculated IC50 values of 110 nM and 60 nM for BBK32-FL and BBK32-C , respectively ( Fig 4B and Table 2 ) . Similar inhibition of the CP could not be detected for the N-terminal BBK32 fragment ( Fig 4A and 4B ) . Consistent with the ELISA based complement assay ( Fig 3C ) , BBK32 proteins did not have an effect on AP-mediated hemolysis of rabbit blood cells ( Fig 4C ) . Taken together , these results demonstrate that BBK32 acts as a potent and specific inhibitor of the classical pathway of human complement and this inhibitory activity locates to the C-terminal globular region of BBK32 . The data presented above indicate that BBK32 can form a high-affinity interaction with C1 and that this interaction results in a specific inhibition of the classical pathway of human complement . The calcium-dependence of the C1/BBK32 interaction ( Fig 2D–2F ) raised the distinct possibility that BBK32 was capable of recognizing only the fully formed calcium-mediated C1 complex rather than binding to an individual C1 subunit with high-affinity . To determine if this was the case we used SPR and monitored the response generated by individual 50 nM injections of C1q , C1r enzyme , C1s enzyme , and C1s proenzyme over the surface of a BBK32-FL or BBK32-C SPR biosensor ( Fig 5A and 5B ) . Surprisingly , we found that BBK32 binds specifically to C1r , but not to other C1 components ( Fig 5A and 5B ) . In agreement with C1/BBK32 binding activity , BBK32-C binds C1r with similar affinity to that of full-length BBK32 ( KD , C1r/BBK32-FL = 15 nM vs . KD , C1r/BBK32-C = 32 nM ) ( Fig 5C and 5D and Table 1 ) . The ability of BBK32 to interact with C1r could also be observed using a Far Western approach where B . burgdorferi lysates were probed with biotinylated C1r ( Fig 5E ) . A band corresponding to the 47-kDa BBK32 protein is detected in the parent lysate but is absent in the lysate from the bbk32 mutant strain . Intriguingly , two additional “non-BBK32” bands are also capable of interacting with C1r using this technique . While the identities of these proteins are currently unknown , we note the presence of a 28-kDa-reactive band present in the parent strain ( denoted with an asterisk ) that matches the migration position of a C1 reactive band ( Fig 1 ) . Kinetic evaluation of the SPR data suggest that both the BBK32-FL and the BBK32-C proteins recognize C1 preferentially over C1r with an increased affinity being attributed primarily to an increase in the association rate ( Table 1 ) . Nonetheless , these binding data strongly suggest that the major BBK32 binding site on the C1 complex is mediated by C1r . Interestingly , we found that the C1r/BBK32 interaction was also dependent on calcium ( S1A and S1B Fig ) . C1r itself possesses multiple calcium binding sites and the structural and functional consequences of C1r/calcium binding are complex [47–50] . For example , calcium binding of C1r results in large-scale conformational changes of the C1r CUB2 domain from a disordered structure in the absence of calcium to a fully folded structure in its presence [49] . The dependence of the BBK32/C1r interaction on calcium suggests that BBK32 recognizes a calcium-dependent C1r conformation , which further implicates the calcium-dependent C1 complex as the physiologically relevant BBK32 ligand . The ability of BBK32 to bind both C1 and C1r and to specifically inhibit the classical pathway of complement suggested that BBK32 interfered with the enzymatic activity of the C1r protease . To determine if this was the case we first measured the effect of various concentrations of BBK32-FL , BBK32-C , and BBK32-N on the in vitro conversion of C1s proenzyme by previously activated C1r enzyme ( Fig 6A ) . Indeed , both BBK32-FL and BBK32-C inhibited C1r in dose-dependent fashion , while BBK32-N failed to inhibit up to a 10 μM final concentration . Quantification of these data by densitometry was performed by evaluating the peak intensity of the C1s proenzyme band ( Fig 6B ) and calculated IC50 values are reported in Table 2 . BBK32 proteins do not affect the activity of previously activated C1s ( C1s enzyme ) as C4 is cleaved by C1s in the presence of 1 μM BBK32 proteins ( Fig 6C ) . These results indicate that BBK32 specifically inhibits C1r enzyme by preventing the processing of its natural substrate , the C1s proenzyme . When C1 is incubated at 37°C , it becomes activated by the autocatalysis of C1r proenzyme and subsequent C1r enzyme cleavage of the C1s proenzyme [51] . To determine if BBK32 affects C1r and C1s zymogens within the C1 complex , we incubated BBK32 proteins ( 5 μM ) with C1 ( 40 nM ) for two hours at 37°C and co-immunoprecipitated ( co-IP ) C1 using a C1q monoclonal antibody . Bound fractions were subjected to SDS-PAGE , followed by Western immunoblot analysis , and the presence of C1r , C1s , C1q , or BBK32 was then assessed . When reactions were probed with C1r antibody ( Fig 7A ) , a processed form of C1r was observed for the buffer only control as indicated by the presence of C1r chain 1 ( 57 kDa ) and C1r chain 2 ( 35 kDa ) . Interestingly , reactions containing BBK32-FL or BBK32-C contain only the proenzyme form of C1r ( 92 kDa ) . In contrast , reactions incubated with BBK32-N or an independent negative control ( recombinant borrelial lipoprotein OspC ) were indistinguishable from the buffer only reaction . These results demonstrate that BBK32 inhibits the autocatalysis of the C1r proenzyme . In agreement with the presence of inactivated C1r within the C1 complex , we found essentially only C1s proenzyme ( 86 kDa ) when we probed the reactions containing BBK32-FL or BBK32-C with the C1s polyclonal antibody ( Fig 7B ) . As would be expected by the presence of activated C1r , we found cleaved C1s in the buffer only , BBK32-N , and OspC reactions . C1q is detected at identical levels in all reactions , confirming the validity of the co-IP approach ( Fig 7C ) , and importantly , BBK32-FL or BBK32-C but not BBK32-N were pulled down with the C1 complex ( Fig 7D ) . Taken together , these results show that BBK32 bound C1 is trapped in an inactive form whereby the autocatalytic activation of C1r is inhibited and subsequent C1r cleavage of C1s proenzyme is blocked and that the BBK32 inhibitory effect is specific for the C-terminal half of the protein . We were interested in determining if similar activities were also observed with native BBK32 produced in B . burgdorferi . To test whether surface exposed BBK32 could mediate binding to components of the classical complement pathway , we incubated infectious B . burgdorferi and a bbk32::StrR derivative ( strain JS315 ) with immobilized C1 ( grown under conditions that induce bbk32 ) . Interestingly , both the parent and the bbk32 isogenic mutant readily bound C1 and no significant difference in binding could be detected ( S2 Fig ) . One plausible explanation for this result is a likely layer of functional redundancy provided by the presence of additional B . burgdorferi proteins capable of C1 recognition as evidenced in the overlay analysis ( Fig 1A ) . To provide a less complex outer surface environment for B . burgdorferi , we used strain B314 , which is missing linear plasmids ( lp ) and thus does not synthesize many lp-encoded borrelial proteins associated with both mammalian and tick infectivity , specifically OspAB , DbpBA , and BBK32 [52 , 53] . A shuttle vector containing bbk32 with its native promoter ( pCD100 ) was transformed into strain B314 , along with a vector-only control ( pBBE22luc ) . An equivalent amount of protein was loaded from each strain ( Fig 8A ) and subsequent immunoblot and biotin-labeled C1 overlay analysis showed that , as expected , B314/pCD100 made detectable levels of BBK32 capable of binding C1 whereas B314/pBBE22luc did not ( Fig 8B and 8C ) . To determine whether natively produced BBK32 mirrored the C1/C1r-binding profile of purified recombinant BBK32 , we exposed these strains to immobilized C1 and C1r . Unlike the wild-type strain , the B314 control strain ( B314/pBBE22luc ) exhibited very little binding to C1 or C1r ( Fig 8D ) . In contrast , the presence of BBK32 in B314/pCD100 transformed organisms promoted a significant enhancement of spirochete binding to both of these targets ( Fig 8D ) . We were next interested if the production of BBK32 could promote resistance to serum . For this purpose , we again utilized B . burgdorferi strain B314 , which in addition to its avirulent phenotype , is rendered serum sensitive , presumably due to its aforementioned loss of lp content [53] . We hypothesized that the selective addition of BBK32 would make strain B314 resistant to complement dependent killing and , to assess this , tested B314/pBBE22luc and B314/pCD100 for their relative resistance to serum . Consistent with an anti-complement activity , the presence of natively expressed BBK32 in B314/pCD100 provided significant protection against serum based immobilization relative to strain B314/pBBE22luc that lacked BBK32 ( Fig 9 ) . As expected , heat inactivation resulted in limited killing with either strain tested ( Fig 9 ) . When conditions were used that selectively block the CP/LP but keep the AP intact ( NHS + Mg-EGTA ) [27] both strains exhibited equivalent resistance/sensitivity independent of BBK32 levels ( Fig 9 ) . These data imply that a majority of the inactivation observed for serum-sensitive B314/pBBE22luc is due to the adverse effects of the CP/LP . Furthermore , these data strongly suggest that the BBK32-dependent protection observed with untreated serum is due to the ability of BBK32 to neutralize either the CP or LP . Coupled with the C1/C1r-specific binding profile ( Figs 1 , 2 and 5 ) , CP-specific inhibition ( Figs 3 and 4 ) , and C1r inhibitory activity ( Figs 6 and 7 ) of recombinant BBK32 proteins along with the CP-specific function of C1r [54] , the serum resistance mediated by BBK32 in this assay is best explained by BBK32-dependent inactivation of the CP . In order to survive the destructive forces of the human complement cascade , successful microbial pathogens have evolved a number of sophisticated evasion strategies . Although specific modes of complement recognition and inactivation are widely varied , these inhibitory mechanisms can be conceptually grouped into three forms: direct recruitment or mimicry of host regulators of complement activity such as complement factor H , enzymatic degradation of complement components by direct or indirect means , or inhibition through direct interaction with complement proteins [55] . In addition to employing each of these evasion strategies , B . burgdorferi , takes advantage of its complex mammalian-tick lifestyle by exploiting anti-complement molecules produced by the tick salivary glands during tick feeding [20] . Recruitment of factor H related molecules by the borrelial CspA , CspZ , and the OspE/F proteins has been studied extensively [12–23 , 56] and C4b-binding protein has been reported to interact with B . burgdorferi , B . afzelii , and B . garinii [30] , while the relapsing fever spirochetes have been shown to bind factor H [24 , 25] C4b-binding protein and C1-INH [29 , 31] . In addition to their direct recruitment , an example of host complement regulator mimicry has also previously been reported for B . burgdorferi [57] . To our knowledge , the only previously reported example of a borrelial factor possessing direct and novel complement inhibition activity is the recently described terminal complement complex inhibition function of CspA [20 , 23] . CspA along with its ability to bind factor H/factor H like-1 and plasminogen also acts a direct inhibitor of the terminal complement complex by binding complement components C7 and C9 and interfering with C5b-9 complex assembly [23] . All three pathways of complement can be activated by B . burgdorferi and all result in direct complement-mediated killing of the spirochete [20 , 23] . In addition to becoming activated by immune complexes , the CP has also been shown to kill B . burgdorferi in the absence of specific antibodies [27] . The requirement for protection from complement attack for B . burgdorferi is evidenced by the production of a number of virulence factors ( now to include BBK32 ) that specifically target and inactivate complement [20 , 23] . In this study we determined that the C-terminal globular region of the borrelial lipoprotein BBK32 exhibits a potent CP-specific inhibitory activity that confers serum resistance to a normally serum-sensitive B . burgdorferi strain . While the focus of this study was on elucidating the anti-complement activity and mechanism of BBK32 , our probe of lysates with C1 and C1r suggest a robust interaction between B . burgdorferi and the CP may exist . Despite the linkage of B . burgdorferi to complement resistance , the small number of experimental infectivity studies employing mice deficient in key components of complement has shown a surprisingly limited role for complement in controlling B . burgdorferi burden [28 , 58 , 59] . Thus , it is of interest to consider potential roles for the inhibition of the CP beyond protection from complement-mediated attack . For example , upon colonizing lymph tissue B . burgdorferi disrupts the normal formation of germinal centers ( GC ) [60 , 61] . Lack of normal GC development ultimately results in reduced antibody titers against B . burgdorferi in experimental infection [60] . Local complement C4 deposition on follicular dendritic cells ( FDC ) is significantly reduced in B . burgdorferi infected lymph nodes and this is speculated to be responsible for the premature collapse of GC responses due to diminished antigen presentation by FDCs [60] . In this regard , it is of interest to determine if BBK32 mediates this lymphoid specific effect , resulting in the observed reduction in the humoral immune response to borrelial antigens . It would be of additional interest to understand the potential role of BBK32 anti-complement activity in the context of spirochetal persistence in a natural reservoir animal such as Peromyscus leucopus [62] . Studies to address some of these possibilities are currently underway . In the current study we show that BBK32 acts directly by binding to and inhibiting C1 via a novel mechanism involving the noncovalent inhibition of C1r enzymatic activity . A model for the inhibition of C1 by BBK32 is depicted in Fig 10 . When C1 engages a surface via C1q binding , this information is transmitted by coordinated conformational changes within the C1 complex ultimately triggering autocatalysis of C1r and subsequent activation of C1s [63] . Under physiological conditions this activation is extremely rapid [64] and controlled on the surface of host cells by the only known endogenous inhibitor of C1 , C1-INH . C1-INH covalently modifies the C1r and C1s active sites and promotes their release from ligand-bound C1q [65] . C1-INH is a member of the serpin family of protease inhibitors and along with inhibiting both C1r and C1s it also inactivates a number of blood proteases involved in the complement , contact , fibrinolytic , and coagulation systems [65] . The broad protease-binding specificity and covalent inhibitory mechanism of C1-INH stands in stark contrast to that of BBK32 , which specifically binds and inhibits C1r and had no detectable effect on the homologous C1s protease . It has been shown that serum deficient in C1r is unable to undergo CP activation while retaining a fully functional LP [54] . In our studies BBK32 did not block the LP or AP suggesting it is unable to inhibit the primary serine proteases of these pathways ( i . e . , MASPs and factor D , respectively ) . Instead , BBK32 specifically inactivates the CP by preventing the autocatalysis of C1r proenzyme and subsequent cleavage of C1s proenzyme ultimately rendering C1 entrapped in a zymogen form . Since its discovery as a fibronectin and glycosoaminoglycan-binding protein expressed on the surface of B . burgdorferi [38 , 43] , BBK32 has been the subject of intense study [39 , 44–46 , 66–68] , including a recent report indicating a role in bloodstream survival [69] . In this regard , BBK32 has been shown to be critical for borrelial pathogenesis [39 , 68] , capable of exploiting host fibronectin function [44 , 46 , 70–73] , shown to be involved in borrelial vascular adhesion mechanisms [69 , 72 , 73] , and in promoting joint colonization [45] . Surprisingly then , infectivity of a B . burgdorferi bbk32 mutant shows a limited infectivity phenotype when experimental infection is done at a high inoculum dose ( i . e . , 105 ) [39 , 68 , 74] . However , this effect is restricted to high doses , as lower inoculum doses ( e . g . , 103 ) exhibit a significant reduction in colonization [39] that is most apparent using in vivo imaging for detection [68] . Furthermore , bbk32 mutant strains exhibit a delay in the ability to disseminate [68] . Nonetheless , one potential explanation for the relatively mild phenotype of the bbk32 mutant is a layer of functional redundancy present in the spirochete . For example , several borrelial proteins besides BBK32 are now known to bind directly to fibronectin [75–77] , and although much remains to be known about their specific functional activities , it is possible that these proteins could overlap with BBK32’s fibronectin related functions . It seems a similar situation may exist for the BBK32 anti-complement activities described here . The presence of multiple bands in the Far Western lysate probes ( Figs 1 and 5E ) suggest that B . burgdorferi may express one or more “non-BBK32” proteins capable of interacting with both C1 and C1r . The existence of additional borrelial proteins able to compensate a loss of BBK32 may further explain the muted bbk32 mutant phenotype . Determining the identity of these proteins along with evaluating their potential to interfere with the CP remains an important next step . Orthologs of BBK32 are found not only in Lyme disease spirochetes , but also in relapsing fever spirochetes where they have been divided into three groups based on phylogenetic relationships [78] . Interestingly , the Lyme disease spirochete B . valaisiana strain ZWU3 Ny3 was shown to possess a novel mechanism of complement inhibition independent of host complement regulator recruitment/mimicry [79] . Although a molecular mechanism has not been elucidated , the authors hypothesize that the inhibitory mechanism present in this strain of B . valaisiana relies on direct interaction with complement components [79] . We note that a BLASTP search of BBK32 from B . burgdorferi B31 reveals a gene located on linear plasmid 28 of B . valaisiana strain VS116 that encodes a 113 residue hypothetical protein that contains 60% identity and 72% similarity with the C-terminus of BBK32 ( residues 181 to 303 ) . While it is not known if this is a functional gene product possessing BBK32 anti-complement activity , it is intriguing that the hypothetical protein lacks the GAG and Fn-binding sequences that are the hallmark of BBK32 . Nonetheless , future studies are required to understand if the complement inhibitory activity of the BBK32 C-terminal domain is conserved across borrelial species . In the same light , it remains to be seen if BBK32 anti-complement mechanisms are found in other human pathogens that are known to have evolved similar molecular mechanisms of host interaction . For instance , BBK32 itself shares nearly identical modes of fibronectin interaction with a group of fibronectin-binding proteins from staphylococcal and streptococcal bacterial species [44 , 80–83] . While these Gram-positive encoded proteins lack sequence conservation with C-terminal BBK32 sequences , it may be important to assess their potential role in complement interaction . The possibility is raised that an underlying benefit to the microorganism exists in producing proteins capable of simultaneously interacting with host extracellular matrix molecules like fibronectin and components of the complement system . Such an example of a synergistic mode of interaction exists for the extracellular fibrinogen-binding ( Efb ) protein expressed by Staphylococcus aureus . A disordered N-terminal region of Efb binds directly to human fibrinogen , while a highly basic globular domain originating from the C-terminal region of the protein binds with high affinity to complement component C3 [84–86] . Although each molecular interaction individually contributes to virulence , a ternary fibrinogen-Efb-C3b complex can form , encapsulating the bacteria in a ‘fibrinogen-shield’ , which results in direct inhibition of phagocytosis [87] . While the similarity of the BBK32 molecular architecture to that of staphylococcal Efb is striking , it is currently unknown how BBK32/fibronectin binding affects BBK32 anti-complement activities or if an analogous functional synergism exists . The work presented here provides the conceptual framework to explore this and related questions on complex host-pathogen interactions involving extracellular matrix and complement proteins . B . burgdorferi B31 strains ML23 [88] , JS315 ( ML23 bbk32::StrR;[39] ) , and B314 were grown in BSK-II media supplemented with 6% normal rabbit serum ( Pel-Freez Biologicals , Rogers , AR ) under microaerobic conditions at 37°C , 5% CO2 atmosphere , at either a pH of 6 . 8 or 7 . 6 . The serum sensitive strain B314 ( kindly provided by Tom Schwan ) is a non-infectious variant of strain B31 that lacks most linear plasmids [43 , 45 , 52 , 53] . All B . burgdorferi cells were enumerated by dark field microscopy . Native bbk32 was cloned into the shuttle vector pBBE22luc [68] using the following approach . The oligonucleotide primers BBK32Comp-BamHI-F ( 5’-ACGCGGATCCGTACTTTGTTCACCCTCTTGATAGC-3’; BamHI site is in bold ) and BBK32Comp-SalI-R ( 5’-ACGCGTCGACATATTATGTAGCCTGTTTTAAATT-3’; SalI site is underlined ) were used to PCR amplify bbk32 from strain B31 genomic DNA . The PCR-amplified product contains 213 bp of upstream sequence and 37 bp downstream from the translational start site and stop codon of the 1 , 062 bp bbk32 gene , respectively . The resulting 1 . 3 kb fragment was cloned into plasmid pCR-Blunt-II-TOPO and transformed in Mach1-T1R Escherichia coli cells ( F– ϕ80lacZΔM15 ΔlacX74 hsdR ( rK– , mK+ ) ΔrecA1398 endA1 tonA; ThermoFisher ) . The resulting construct was digested with BamHI and SalI and the 1 . 3 kb fragment subsequently cloned into BamHI and SalI cut pBBE22luc . The resulting construct , which contained bbk32 expressed from its native promoter , was designated pCD100 . Transformation of strain B314 with pCD100 and pBBE22luc was done as previously described [89] . Transformants were selected for resistance to kanamycin and screened by PCR to confirm the presence of pCD100 using primers BBK32Comp-BamHI-F and BBK32Comp-SalI-R . B . burgdorferi strain ML23 was grown under inducing conditions ( 37°C , 5% CO2 , pH 6 . 8 ) and harvested by centrifugation at 5 , 800 x g , and washed twice with PBS . The cell pellet was resuspended in 0 . 5 ml of either PBS alone , or PBS with proteinase K ( to a final concentration of 200 μg ml-1 ) . All samples were incubated at 20°C for 40 min . Reactions were terminated by the addition of phenylmethylsulfonyl fluoride ( PMSF ) to a final concentration of 1 mM . Cells were again pelleted by centrifugation ( 9 , 000 x g for 10 min at 4°C ) , washed twice with PBS containing 1 mM PMSF , and resuspended in Laemmli sample buffer [90] . Samples corresponding to 5 x 107 whole cell equivalents were run on SDS-PAGE gel , transferred to PVDF membranes and probed with biotinylated C1 in a Far Western analysis or immunoblotted with a monoclonal antibody to borrelial FlaB ( Affinity BioReagents , Inc . ) , as described below . DNA fragments encoding residues 21 to 205 or 206 to 354 of BBK32 ( B . burgdorferi B31 strain ) were PCR-amplified from pQE30-BBK32-FL plasmid DNA [46] using oligonucleotide primers that appended BamHI and NotI sites at the 5’ and 3’ ends , respectively . Restriction digested DNA fragments were then sub-cloned into the pT7HMT vector [91] . Sequence confirmed plasmids were transformed into E . coli strain BL21 ( DE3 ) for protein production . Expression and purification of recombinant BBK32-FL was performed as previously described [46] . Recombinant BBK32-N and BBK32-C were overexpressed by inoculating 1 liter of Terrific Broth with 10 ml of an overnight BL21 ( DE3 ) at 37°C , induced with 1 mM isopropyl-D-thiogalactopyranoside upon reaching an optical density at 600 nm of 0 . 6–0 . 8 , shifted to 18°C shaking , and allowed to express overnight . Overnight cultures were harvested by centrifugation at 5 , 000 x g for 10 min and resuspended in Ni-NTA-binding buffer ( 20 mM Tris ( pH 8 . 0 ) , 500 mM NaCl , 10 mM Imidazole ( pH 8 . 0 ) ) and lysed by microfluidization . A clarified cell extract was obtained by centrifugation at 25 , 000 x g for 35 min and the supernatant was applied to a nickel agarose ( Gold Bio ) column previously equilibrated in Ni-NTA buffer , washed with 5 column volumes ( CV ) of Ni-NTA-binding buffer , and eluted with 2 CV’s of Ni-NTA elution buffer ( 20 mM Tris ( pH 8 . 0 ) , 500 mM NaCl , 500 mM imidazole ( pH 8 . 0 ) ) . The HIS-myc affinity tag was removed by enzymatic digestion with tobacco etch virus ( TEV ) protease in the presence of 5 mM β-mercaptoethanol at room temperature for 2 h . The digestion reaction was allowed to continue overnight at 4°C while being dialyzed against Ni-NTA-binding buffer . The TEV digested sample was then incubated with a nickel agarose column and the flow through fraction was collected and subjected to gel filtration chromatography using a HiLoad Superdex 200 PG column ( GE Healthcare ) equilibrated in 20 mM Tris ( 8 . 0 ) , 200 mM NaCl . Peaks were analyzed by SDS-PAGE and fractions corresponding to BBK32 proteins were pooled and concentrated using Amicon centrifugal filters ( EMD Millipore ) , aliquoted , and stored at -80°C until use . Purified C1 , C1r enzyme , C1q , C1s enzyme , C1s proenzyme , and C4 were obtained from Complement Technology ( Tyler , TX ) . Human C1 or C1r proteins were biotinylated using EZ-link Sulfo-NHS-LC-Biotin ( Thermo Fisher Scientific ) at the molar ratio suggested by the manufacturer . The labeling reaction was quenched by the addition of Tris HCl pH 7 . 6 to a final concentration of 20 mM . The sample was diluted in protein-binding buffer ( 100 mM NaCl , 20 mM Tris pH 7 . 6 , 1 mM EDTA , 10% glycerol , 0 . 2% Tween-20 , 2% non-fat milk ) [92] and used below in the Far-Western assay . For Far Western analyses , B . burgdorferi protein lysates were resolved by SDS–PAGE [90] and gels were transferred to PVDF membranes as described [93] . The membrane was blocked in 10% non-fat milk , washed with PBS , 0 . 2% Tween-20 , and then 20 μg biotinylated C1 or C1r ( both from CompTech ) , at 1 μg/ml in protein binding buffer [92] , was incubated with the membrane overnight at 4°C . The membrane was washed extensively in PBS , 0 . 2% Tween-20 , and then incubated with Vectastain solution ( Vectastain Elite ABC Kit , Vector Laboratories ) , as instructed by the manufacturer , to enhance the signal . The membrane was then washed and developed using the Western Lightning Chemiluminescent Reagent plus system ( Perkin Elmer , Waltham , MA , USA ) . Conventional immunoblotting was done as previously described [94] . Production of BBK32 in B314/pCD100 was evaluated using either a monoclonal antibody to BBK32 ( a generous gift from Seppo Meri , University of Helsinki ) diluted to 1:4000 or , in the case of the co-immunoprecipitations ( see below ) , a polyclonal antibody against BBK32 diluted 1:1500 . Immunoblots to detect the endoflagellar antigen FlaB were done using a monoclonal to B . burgdorferi FlaB ( Affinity BioReagents ) diluted 1:20 , 000 . Appropriate anti-rabbit Ig or anti-mouse Ig HRP conjugates ( Life Technologies ) were diluted 1:5000 and used to detect primary antibodies on the PVDF membranes . Immune complexes were detected using the Western Lightning Chemiluminescent Reagent plus system ( Perkin Elmer , Waltham , MA , USA ) . Direct binding of C1 and subunits of the C1 complex to BBK32 was assessed by SPR using a Biacore 3000 instrument ( GE Healthcare ) at 25°C . HBS-T-Ca2+ ( 20 mM HEPES ( pH 7 . 3 ) , 140 mM NaCl , 0 . 005% ( v/v ) Tween 20 , 5 mM CaCl2 ) was used as the running buffer and a flowrate of 10 μl min-1 were used in all experiments . A BBK32 biosensor was created by immobilizing recombinant BBK32 proteins on separate flowcells of a C1 sensor chip ( GE Healthcare ) . In all cases immobilization was achieved using standard amine coupling chemistry by activating the flowcell surface for 7 min at 5 μl min-1 with an equal volume mixture of 0 . 1 M N-hydroxysuccinimide and 0 . 4 M ethyl ( dimethylaminopropyl ) carbodiimide . Next , BBK32-FL at 20 μg ml-1 in 10 mM sodium acetate ( pH 4 . 0 ) , BBK32-N at 25 μg ml-1 in 10 mM sodium acetate ( pH 4 . 5 ) , or BBK32-C at 5 μg ml-1 in 10 mM sodium acetate ( pH 5 . 5 ) were injected and allowed to react until the desired surface density was reached . Finally , 1 M ethanolamine ( pH 8 . 5 ) was injected for 7 min at 5 μl min-1 to quench remaining reactive groups . A reference flowcell was generated by activating the surface followed by immediate quenching . Final immobilization densities reported in resonance units ( RU ) were as follows: BBK32-FL ( 165 RU ) , BBK32-N ( 150 RU ) , and BBK32-C ( 240 RU ) . All solution phase analytes were exchanged into running buffer just prior to injection . Purified C1 was injected as a twofold concentration series in triplicate consisting of 0 . 20 , 0 . 39 , 0 . 78 , 1 . 6 , 3 . 1 , 6 . 3 , 13 , 25 , 50 , and 100 nM for 3 min followed by 5 min of dissociation . The surface was regenerated to baseline by injecting HBS-T-EGTA ( 20 mM HEPES ( pH 7 . 3 ) , 140 mM NaCl , 0 . 005% Tween-20 , and 10 mM EGTA ) for 1 min followed by a 30 s injection of a solution containing 0 . 1 M glycine ( pH 2 . 2 ) and 2 . 5 M NaCl . Direct interaction of C1r enzyme with BBK32 proteins was assessed using an identical protocol to C1 except for a concentration series consisting of 0 . 39 , 0 . 78 , 1 . 6 , 3 . 1 , 6 . 3 , 13 , 25 , 50 , 100 , and 200 nM was used and the injection time was increased to 5 min followed by 10 min of dissociation . Binding of C1r enzyme to active BBK32 proteins was assessed on a second C1 biosensor with different coupling densities BBK32-FL ( 80 RU ) and BBK32-C ( 330 RU ) . A separate concentration series consisting of 2 . 0 , 3 . 9 , 7 . 8 , 16 , 31 , 63 , 125 , 250 , 500 , and 1000 nM was injected in duplicate for 5 min followed by 15 min of dissociation . Kinetic analysis was performed for each set of sensorgrams resulting from C1 or C1r enzyme injections using BIAevaluation software 4 . 1 . 1 ( GE Healthcare ) using a 1:1 ( Langmuir ) binding model and fitting Rmax locally . To determine the effect of calcium on the interaction of BBK32 , a single concentration of C1 ( 50 nM ) or C1r enzyme ( 200 nM ) was injected in triplicate in a running buffer of HBS-T-Ca2+ or HBS-T-EGTA . To evaluate relative binding of individual C1 complex components to BBK32 , triplicate injections were performed at a fixed 50 nM concentration in HBS-T- Ca2+ running buffer . Response was corrected for the molecular weight of each component ( C1 = 790 kDa , C1q = 410 kDa , C1r enzyme = 92 kDa , C1s proenzyme and C1s enzyme = 86 kDa ) . To delineate the effect of BBK32 proteins on the CP , LP , and AP , we adopted an ELISA based assay previously described [95] . Costar EIA/RIA plates ( Fisher Scientific ) were incubated overnight at room temperature with 3 μg ml-1 human IgM ( CP initiator ) ( Athens Research & Technology ) , 25 μg ml-1 Salmonella enteriditis LPS ( AP initiator ) ( Sigma Aldrich ) , or 20 μg ml-1 of mannan from Saccharomyces cerevisiae ( LP initiator ) ( Sigma Aldrich ) in a coating buffer consisting of 100 mM Na2CO3/NaHCO3 ( pH 9 . 6 ) . All subsequent steps were preceded by three consecutive washes with TBS-T buffer ( 50 mM Tris ( pH 8 . 0 ) , 150 mM NaCl , 0 . 05% ( v/v ) Triton X-100 ) and all reaction volumes were 100 μl . Plates were blocked in PBS-T-BSA ( 137 mM NaCl , 2 . 7 mM KCl , 10 mM Na2HPO4 , 1 . 8 mM KH2PO4 , 1% ( w/v ) bovine serum albumin , and 0 . 05% ( v/v ) Tween-20 ) for 1 h . Reactions consisted of pooled complement human serum ( Innovative Research ) , at 1% ( CP/LP ) or 20% ( AP ) final concentration , various concentrations of BBK32 proteins , and CP/LP buffer ( 20 mM HEPES ( pH 7 . 3 ) , 0 . 1% ( w/v ) gelatin , 140 mM NaCl , 2 mM CaCl2 , 0 . 5 mM MgCl2 ) or AP buffer ( 20 mM HEPES ( pH 7 . 5 ) , 0 . 1% ( w/v ) gelatin , 140 mM NaCl , 5 mM MgCl2 , 10 mM EGTA ) . Serum/BBK32 mixtures were then added to wells and incubated at 37°C for 1 h . Downstream complement activation was measured by detecting C3b deposition using a 1:300 dilution of an anti-C3d monoclonal antibody ( 030–08 , Santa Cruz Biotechnology ) or 1:300 dilution of an anti-C4d monoclonal antibody ( C4-1 , Cell Sciences ) incubated at room temperature for 1 h , and subsequent 1 h room temperature incubation with a 1:5000 dilution of goat anti-mouse HRP secondary antibody ( Thermo Scientific ) . HRP-labeled antibody was detected using 1-step Ultra TMB ( Thermo Scientific ) for 10 min . The reaction was stopped by addition of 2 M sulfuric acid and the absorbance was measured at 450 nm using a VersaMax microplate reader ( Molecular Devices ) . Wells containing serum only or where serum was replaced with buffer were treated as 100% and 0% signal , respectively . All experiments were performed a minimum of three times . IC50 values were evaluated by variable slope four-parameter nonlinear regression analysis performed using GraphPad Prism 5 . 0 . Inhibition of CP-mediated hemolysis by BBK32 proteins was assessed using a modified classical pathway hemolytic assay ( CP50 ) . Sheep erythrocytes ( 5 x 108 cells ml-1 ) sensitized with human IgM ( Complement Tech ) were centrifuged at 500 x g at 4°C for 3 min and resuspended in GHB++ buffer ( 20 mM HEPES ( pH 7 . 3 ) , 140 mM NaCl , 0 . 1% gelatin ( w/v ) , 0 . 15 mM CaCl2 , and 0 . 5 mM MgCl2 ) . Final volumes were 100 μl and reactions began by mixing 35 μl GHB++ with 20 μl of BBK32 proteins previously diluted into GHB++ at various concentrations , followed by 20 μl NHS ( 1% v/v ) final concentration , and 25 μl sheep erythrocytes sensitized with human IgM . Reactions were incubated at 37°C for 1 h with intermittent shaking and clarified by centrifugation at 1000 x g for 3 min . 50 μl of each reaction were transferred to a 96-well flat-bottom half-area microplate and absorbance was measured at 541 nm using a VersaMax microplate reader ( Molecular Devices ) . A well containing no BBK32 was considered as 100% lysis and background absorbance was measured by replacing NHS with buffer . Percent lysis was calculated by subtracting background readings from each well and comparing each reading to 100% controls . Inhibition of AP-mediated hemolysis by BBK32 proteins was assessed using a modified alternative pathway hemolytic assay ( APH50 ) . Rabbit erythrocytes ( Complement Tech ) at 5 x 108 cells ml-1 were washed by centrifugation at 500 x g at 4°C for 3 min and resuspension in GHBS° buffer ( 20 mM HEPES ( pH 7 . 5 ) , 140 mM NaCl and 0 . 1% gelatin ( w/v ) ) . Reactions began by diluting 5 μl of 0 . 1M MgCl2-EGTA into 30 μl GHBS° , followed by 20 μl of BBK32 proteins , followed by 20 μl of NHS , and finally 25 μl rabbit erythrocytes . Reactions were allowed to incubate at 37°C for 30 min with intermittent agitation , clarified , and diluted 1:10 in a 96-well plate . Absorbance was measured at 412 nm and % lysis was computed as described for the CP50 assay . All experiments were repeated between two and four times . IC50 values were evaluated by variable slope four-parameter nonlinear regression analysis performed using GraphPad Prism 5 . 0 . The ability of BBK32 proteins to inhibit the enzymatic cleavage of C1s proenzyme by C1r enzyme in vitro was performed as follows . Reactions were carried out in HBS-Ca2+ ( 20 mM HEPES ( pH 7 . 3 ) , 140 mM NaCl , 5 mM CaCl2 ) . Reaction volumes were 10 μl and consisted of 5 μl BBK32 protein previously diluted into water , 1 μl C1s proenzyme ( 1 μg μL-1 ) , 1 . 5 μl C1r enzyme ( 333 nM ) , and 2 . 5 μl of 4x HBS-Ca2+ . Following overnight incubation at 37°C , each reaction was mixed with 5 μl reducing SDS-PAGE Laemmli sample buffer , boiled for 5 min , and 7 . 5 μl of each reaction were separated on a 10% Tris-tricine SDS-PAGE gel . Following Coomassie blue staining , digital images of destained gels were captured using a FluorChem M imaging system ( ProteinSimple ) . Densitometry was performed using AlphaView SA 3 . 4 . 0 software ( ProteinSimple ) and the normalized peak height of the band corresponding to C1s proenzyme was plotted against the concentration of BBK32 present in each reaction . IC50 values were evaluated with GraphPad Prism 5 . 0 using four parameter variable slope nonlinear regression analysis and constraining the top and bottom values to 100 and 0 respectively . All experiments were repeated between two and four times . The ability of BBK32 proteins to inhibit the enzymatic cleavage of C4 by C1s enzyme in vitro was performed as follows . Reaction volumes were 10 μl and consisted of 2 . 5 μl C4 ( 1 mg ml-1 ) , 1 μl C1s enzyme ( 1 mg ml-1 ) , 1 . 5 μl PBS , and 5 μl of BBK32 proteins previously diluted into PBS . Reactions were incubated at 37°C for 30 min and stopped by the addition of 5 μl SDS-PAGE reducing buffer , boiled for 5 min , and evaluated by SDS-PAGE . Monoclonal antibodies to human C1q were captured by Protein G beads ( Thermo Scientific ) at room-temperature for 2 h as instructed by the manufacturer . Purified C1 complex ( 40 nM ) was incubated at room temperature for 2 h with either 5 μM BBK32-FL , BBK32-N , BBK32-C , or full-length recombinant OspC . All reactions were performed in HEPES++ buffer ( 20 mM HEPES , 140 mM NaCl , 0 . 15 mM CaCl2 and 0 . 5 mM MgCl2 , pH 7 . 3 ) . Subsequently , C1 , pre-incubated with borrelial recombinant proteins , was added to Protein G beads pre-loaded with C1q monoclonal antibodies and incubated at 37°C for 2 hr . The beads were then washed 5 times with HEPES++ buffer containing 0 . 2% Tween-20 , followed by the addition of 50 μl Laemmli sample buffer . Samples were then boiled for 10 min , separated by SDS-PAGE , transferred to PVDF membranes , and subjected to Western immunoblot analysis . The components of the C1 complex or BBK32 were detected with polyclonal goat anti-human C1q ( diluted 1:5000 ) , goat anti-human C1r ( diluted 1:3000 ) , sheep anti-human C1s ( diluted 1:3000 ) , or rabbit polyclonal antibody to BBK32 ( diluted 1:1500 ) followed by incubation with either HRP-conjugated TrueBlot ( Rockland Antibodies ) directed against goat or sheep immunoglobulin ( diluted 1:2000 ) ( for the C1 proteins ) or goat anti-rabbit Ig/HRP ( Life Technologies ) diluted 1:5000 ( for BBK32 ) . The membranes were washed extensively in PBS , 0 . 2% Tween-20 , and developed using the Western Lightning Chemiluminescent Reagent plus system ( Perkin Elmer , Waltham , MA , USA ) . B . burgdorferi adherence assay was done as previously described with slight modifications [94] . Briefly , poly-D-lysine pre-coated coverslips ( Corning Biocoat ) were coated with 1 μg human C1 ( Comptech ) , C1r ( Comptech ) , or BSA respectively and incubated at 4°C overnight . The coverslips were washed thoroughly in PBS to remove excess unbound proteins . The coverslips were then blocked with 3% BSA at room temperature for 1 hr . B . burgdorferi strains ML23/pBBE22luc and JS315/pBBE22luc were grown to mid-logarithmic phase at 37°C , 5% CO2 , pH 6 . 8 to induce expression of bbk32 . Strains B314/pBBE22luc and B314/pCD100 were grown to mid-logarithmic phase at 32°C , 1% CO2 , pH 7 . 6 since the production of BBK32 in B314/pCD100 was high independent of growth condition . All B . burgdorferi strains were subsequently diluted to 107 organisms/ml in BSK-II medium without serum . The resulting B . burgdorferi samples , in 0 . 1 ml volumes , were applied onto the coverslips and incubated for 2 hr at 32°C . Unbound bacteria were removed from the coverslips by gentle washing with PBS; this wash step was repeated 7 times . The coverslips were applied to a glass slide and attached spirochetes were counted by dark field microscopy . Binding of spirochetes to their respective targets was scored by dark field microscopy . Complement sensitivity assays were performed as previously described [27 , 96] . Briefly , B . burgdorferi strains were grown to exponential phase at 32°C , 1% CO2 , pH 7 . 6 , and 106 cells suspension in 80 μl of BSK-II medium was added to 20 μl of normal human serum ( NHS ) or 10 mM Mg/EGTA treated NHS to give a final volume of 100 μl with 20% serum . The samples were placed in microtiter plates and the suspensions were sealed and incubated at 32°C for 2 h . Heat-inactivated normal human serum ( hiNHS ) was used as a control . After incubation , B . burgdorferi suspensions were scored under dark field microscope and the percentage of viable B . burgdorferi cells were calculated from randomly chosen fields and based on immobilization , loss of cell envelope integrity , and overt lysis .
The human complement system is a connected network of blood proteins capable of recognizing and eliminating microbial intruders . To avoid the destructive force of complement activation , many microorganisms that enter the bloodstream express molecules that disrupt key steps of the complement cascade by interacting with specific complement components . In this study we show that the causative agent of Lyme disease , Borrelia burgdorferi , expresses a surface-protein termed BBK32 that targets and inhibits the first component of complement , designated C1 . Upon binding to human C1 , BBK32 traps this initiating protease complex of the classical pathway of complement in an inactive state , and prevents the downstream proteolytic events of the pathway . Our study defines a new mechanism by which microbes are able to escape the human innate immune system and identifies complement protease C1r as a previously unknown target of bacterial anti-complement molecules . Thus , discovery of the complement inhibitory activity of the borrelial protein BBK32 significantly advances our understanding of how disease-causing bacteria survive in immune competent hosts .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[ "complement", "system", "medicine", "and", "health", "sciences", "immune", "physiology", "pathology", "and", "laboratory", "medicine", "zymogens", "enzymes", "pathogens", "immunology", "microbiology", "enzymology", "enzyme", "inhibitors", "bacteria", "bacterial", "pathogens", "immune", "system", "proteins", "complement", "inhibitors", "borrelia", "burgdorferi", "proteins", "medical", "microbiology", "microbial", "pathogens", "borrelia", "enzyme", "precursors", "immune", "system", "complement", "activation", "biochemistry", "physiology", "biology", "and", "life", "sciences", "proteases", "organisms" ]
2016
Borrelia burgdorferi BBK32 Inhibits the Classical Pathway by Blocking Activation of the C1 Complement Complex
Neurons that respond selectively but in an invariant manner to a given feature of natural stimuli have been observed across species and systems . Such responses emerge in higher brain areas , thereby suggesting that they occur by integrating afferent input . However , the mechanisms by which such integration occurs are poorly understood . Here we show that midbrain electrosensory neurons can respond selectively and in an invariant manner to heterogeneity in behaviorally relevant stimulus waveforms . Such invariant responses were not seen in hindbrain electrosensory neurons providing afferent input to these midbrain neurons , suggesting that response invariance results from nonlinear integration of such input . To test this hypothesis , we built a model based on the Hodgkin-Huxley formalism that received realistic afferent input . We found that multiple combinations of parameter values could give rise to invariant responses matching those seen experimentally . Our model thus shows that there are multiple solutions towards achieving invariant responses and reveals how subthreshold membrane conductances help promote robust and invariant firing in response to heterogeneous stimulus waveforms associated with behaviorally relevant stimuli . We discuss the implications of our findings for the electrosensory and other systems . Efficient processing of incoming sensory information is essential to an organism’s survival . Thus , understanding the strategies used by the brain to process such information ( the neural code ) remains an important problem in neuroscience . There is growing experimental evidence that the representation of sensory information changes from a dense code , in which neurons respond differentially to most if not all behaviorally relevant stimuli , to a sparse code , in which neurons instead respond selectively but in an invariant manner to a given stimulus ( e . g . a human face ) [1–12] . This is thought to reduce energy consumption , enhance the ability to recognize a particular feature , and facilitate readout by higher brain structures [13] . Despite their seemingly contrary nature , both response selectivity and invariance to identity preserving transforms of the target stimuli ( e . g . the same object seen from different viewpoints or the same sound heard at different intensities ) have been shown to progressively increase as information propagates to higher brain regions [8 , 10–12 , 14] . Most strikingly , neurons in the medio-temporal cortex can respond similarly to stimulus patterns that are only abstractly related ( a picture of a given person vs . the name of this person written on a piece of paper ) [6] . The mechanisms by which sparse coding and feature invariance emerge in the vertebrate brain remain largely unknown but are critical for understanding brain disorders as well as improving artificial intelligence [15] . Here we studied the emergence of feature invariant representations of natural stimuli in the electrosensory system of the gymnotiform weakly electric fish Apteronotus leptorhynchus . These animals continuously emit a quasi-sinusoidal electric field through an electric organ discharge ( EOD ) and rely on perturbations of this field caused by objects whose conductivity is different than that of the surrounding water ( e . g . prey , conspecifics ) to obtain information about their surroundings ( see [16–21] for review ) . Peripheral electroreceptor neurons respond to increases in EOD amplitude through increases in firing rate and relay this information to pyramidal neurons within the hindbrain electrosensory lateral line lobe ( ELL ) that in turn project to the midbrain torus semicircularis ( TS ) , which is at a similar processing stage as the inferior colliculus in mammals [17] . Natural stimuli are comprised in part by those caused by conspecifics . In particular , the interference between the EODs of two fish that come into contact ( <1 m distance ) will give rise to a sinusoidal amplitude modulation stimulus , a beat , whose frequency is equal to the difference between the two EOD frequencies . Due to a sexual dimorphism in EOD frequency , interactions between same-sex and opposite-sex conspecifics give rise to low ( <30 Hz ) and high ( >30 Hz ) frequency beats , respectively [22] . These fish moreover will display electrocommunication stimuli called “chirps” that consist of brief ( <40 ms ) increases in EOD frequency , which occur on top of the beat and can give rise to very different waveforms [22–26] . Previous electrophysiological studies have found that both peripheral electroreceptors [27–29] as well as hindbrain pyramidal neurons [30–32] tend to use dense neural codes as they give strong responses to both beat and chirp stimuli that vary based on the waveform used . In contrast , TS neurons tend to use a sparse neural code as they display much more selectivity . In particular , some TS neurons can respond selectively but differentially to chirps [32] . Here we report that some TS neurons can instead respond selectively but in an invariant fashion to different chirp waveforms , thereby providing a potential neural correlate of robust perception of natural electrocommunication signals by permitting reliable signal detection and information extraction in higher brain areas . Since hindbrain ELL neurons providing afferent input to TS neurons did not display such invariant responses , we concluded that feature invariance emerges at the level of the TS presumably by nonlinear integration of ELL afferent input . In order to investigate the underlying mechanisms , we built a model of a TS neuron receiving realistic input based on the Hodgkin-Huxley formalism . Our results show that multiple combinations of parameter values could give rise to feature invariant responses to different chirp waveforms mimicking those seen experimentally . Importantly , while a spiking nonlinearity was sufficient to observe such responses , the addition of subthreshold h- and T-type currents increased the regions in parameter space that gave rise to feature invariance . Our results have important implications for understanding the emergence of feature invariant responses in sensory systems . While it is well known that natural electrocommunication stimuli display heterogeneous waveforms [25 , 27] , a systematic investigation and quantification of these heterogeneities has not been done to date . Thus , we first investigated and quantified heterogeneities in naturally occurring electrocommunication stimuli . We used a well-established behavioral assay [34] in which the restrained animal receives a stimulus mimicking another fish’s EOD and recorded the animal’s behavioral responses ( see Methods ) . On average the EOD frequency increases associated with small and big chirps were 50 . 1±0 . 5 Hz ( min: 30 . 1 Hz; max: 76 . 5 Hz; n = 486 ) and 261 . 2±2 . 2 Hz ( min: 155 . 4 Hz; max: 459 . 3 Hz; n = 395 ) , respectively . Overall , we found large differences between the waveforms associated with small chirps as evidenced from the four example traces shown in the top panel of Fig 1C . In contrast , we found smaller differences amongst the waveforms associated with big chirps that were all relatively similar to each other as they were mainly characterized by a pronounced decrease in the beat amplitude ( Fig 1C , bottom panel ) . We quantified the difference between the observed small chirp and big chirp waveforms using a similarity index ( see Methods ) and found overall larger values for big chirps than for small chirps ( two-sample K-S test , p = 0 . 0013 ) ( Fig 1D ) . In order to investigate the source of the observed heterogeneities in the recorded small chirp but not big chirp waveforms , we quantified each by two attributes: namely duration and beat phase ( see Methods ) . Our results show that both small and big chirp duration were distributed over different but relatively narrow ranges of values ( Fig 1E ) . In contrast , both small and big chirps occurred at all phases of the beat with uniform probability ( Fig 1E ) . Thus , our results suggest that the phase of the beat at which the chirp occurs but not its duration is an important source of heterogeneity in the resulting stimulus waveform for small but not big chirps . This result can be understood intuitively as follows . First , we note that small and big chirps tend to occur on top of lower and higher frequency beats , respectively . Thus , the beat period is then longer relative to the chirp duration for the former and is thus expected to have more of an effect in creating a heterogeneous set of waveforms . We nevertheless note that the fact that there is a concomitant decrease in EOD amplitude for big but not small chirps also likely contributes to the fact that there are less heterogeneities in the waveforms resulting from the former . We next investigated whether electrosensory midbrain neurons responded to big and small chirps in a feature invariant manner . To do so , we recorded the responses of TS neurons ( N = 137 ) ( Fig 2A ) to both small and big chirp stimuli that captured the relatively heterogeneous waveforms seen for the former and the relatively homogeneous waveforms seen for the latter ( Fig 2B ) . Previous studies have found that TS neurons either do not respond selectively or respond selectively but differentially to different chirp stimulus waveforms [32] . Here we focused on neurons that responded selectively to the chirp stimulus and not the beat and whose responses to different chirp waveforms were similar ( see Methods ) . We found that some TS neurons in our dataset ( N = 9 ) responded selectively to both small and big chirp stimuli but in a similar manner through silence during the beat and by the firing of 1–2 action potentials at a short latency ( ~15 ms ) after the chirp onset ( Fig 2B ) . We quantified whether the response was selective to the chirp waveform using the chirp selectivity index ( CSI ) as done previously [32] ( see Methods ) and we quantified differences between spiking responses to different chirp waveforms using the Victor-Purpura distance metric ( VPD ) [35] ( see Methods ) . We obtained CSI = 1 and VPD = 1 . 19 for the example neuron shown in Fig 2 . We also computed a feature invariance index ( FI ) score that captured a neuron’s ability to respond selectively but invariantly to chirps . We obtained FI = 0 . 99 for the example neuron shown in Fig 2 . Perhaps the simplest potential explanation for the experimentally observed invariant responses of TS neurons described above is that they are simply inherited from their afferent ELL pyramidal neurons . Previous studies have found two types of ELL pyramidal neurons [36]: ON-type neurons respond with excitation while OFF-type neurons instead respond with inhibition to increases in EOD amplitude , respectively . We thus recorded ELL pyramidal neuron responses to the same stimuli presented to TS neurons ( Fig 3A ) . In contrast to TS neurons and consistent with previous results [37] , ON ( n = 25 ) and OFF-type ( n = 20 ) ELL pyramidal cells displayed pronounced responses to the beat in the form of phase locking ( Fig 3B and 3C ) . Although ON and OFF-type ELL pyramidal cells also responded to all chirp waveforms , they did not do so in an invariant manner as they were excited by some chirp waveforms but inhibited by others ( Fig 3B and 3C ) . We quantified the responses of ON and OFF-type ELL pyramidal cells to chirps using CSI , VPD , and FI . Overall , both ON and OFF-type ELL neurons displayed significantly smaller CSI values than TS neurons ( Fig 3D , left ) indicating that they tended to respond to both the beat and the chirp , consistent with previous findings [30–32] . Moreover , there were significantly greater differences between the responses of ON and OFF-type ELL neurons to different chirp waveforms as compared to that of TS neurons ( Fig 3D , middle ) . Therefore , the responses of TS neurons to chirps were significantly more invariant than those of ELL neurons ( ON: 0 . 000±0 . 000 , OFF: 0 . 004±0 . 004 , TS: 0 . 54±0 . 09 , one-way ANOVA with Tukey-Kramer correction , p<0 . 05 ) ( Fig 3D , right ) . We conclude that the feature invariant responses observed in TS are not simply inherited from ELL neurons and must rather result from TS neurons integrating such input . We next tested whether pooling the activities of ON and OFF-type pyramidal cells might give rise to feature invariant responses . This is important because previous results have found strong heterogeneities in the responses of ON and OFF-type pyramidal cells to chirp stimuli [30 , 31] . Overall , we found that pooling the responses across either ON-type , OFF-type , or both types did not give rise to feature invariant responses to natural electrocommunication stimuli ( Fig 3E ) . While such pooling gave rise to slightly higher FI values than seen for single neurons , the FI values were still much lower than those observed in TS neurons ( Fig 3E ) . We thus conclude that nonlinear integration of ELL input by TS neurons is necessary to give rise to the experimentally observed feature invariant responses in the midbrain . To investigate whether , and if so how , nonlinear integration of ELL input by TS neurons is sufficient to observe feature invariant responses to chirp stimuli , we built a model TS neuron based on the Hodgkin-Huxley formalism that included different membrane conductances seen experimentally in TS neurons ( Fig 4 ) . Importantly , the afferent ELL input to the model was the weighted sum of the population-averaged experimentally observed responses of ON and OFF-type pyramidal neurons to chirp stimuli convolved with an alpha function to mimic synaptic input ( see Methods ) . Model parameters were similar to those used in previous modeling studies of TS neurons [38–40] or varied systematically . We then used a constrained differential evolution algorithm to identify combinations of parameter values that gave rise to the highest feature invariant responses as quantified by FI ( see Methods ) . This algorithm identified multiple sets of physiologically realistic parameter values that all gave rise to feature invariant responses that matched those seen experimentally . Fig 5 shows five such examples . For each set of parameter values , the model neuron responded to each chirp with 1–4 action potentials , as seen experimentally , in a similar fashion as quantified by similar FI scores despite having different parameter values . Importantly , we had both T- and h-type conductances set to zero ( i . e . gT = gh = 0 μS ) for model 2 indicating that these subthreshold membrane conductances are not necessary to observe feature invariance . Thus , our model predicts that the spiking nonlinearity is sufficient to produce feature invariant responses to natural electrocommunication stimuli . We next investigated why different combinations of parameter values all gave rise to feature invariant responses to natural electrocommunication stimuli . To do so , we systematically varied model parameters . Specifically , we varied the bias current Ibias , the T-type calcium conductance gT , the h-type conductance gh , the maximum synaptic conductance gsyn , the noise intensity σnoise , and the fraction of ON-type input σB . We observed negatively sloped bands when varying any two parameters except σB within this set ( Fig 6A and 6B and 6C as well as the left panels of Figs 7A and 8A and 8C ) , indicating that increases/decreases in one parameter could be compensated for by decreases/increases in the other parameter . In contrast , we did not observe such negatively sloped bands when varying both σB and any of Ibias , gT , gh , gsyn , or σnoise ( Fig 6D and left panels of Figs 7B and 7C and 8B ) , indicating that proper tuning of σB , which gives the relative balance of ON vs . OFF-type input received by the model neuron is necessary to achieve feature invariance as a change in this parameter cannot be compensated for by changing other parameters . We note that , by definition , parameter regions with high feature invariance correspond to regions with high CSI and low VPD values ( S1 and S2 Figs ) . The implications of these findings are discussed below . While our modeling results have shown that the subthreshold membrane conductances gh and gT were not necessary to observe feature invariance , we nevertheless investigated whether these might increase the regions in parameter space for which feature invariance was observed . To do so , we compared the robustness of feature invariance to varying model parameters as quantified by the % of pixels for which we had FI ≥ 0 . 7 when varying parameters pairwise in models with and without subthreshold membrane conductances . Figs 7 and 8 show the FI values obtained when varying parameters pairwise in model neurons with ( left panels ) and without ( right panels ) subthreshold conductances . It is seen that high FI values were obtained for larger regions in parameter space with subthreshold conductances ( Figs 7 and 8 , compare left and right panels ) . On average , the robustness with subthreshold conductances was higher ( ∼16% ) than without ( ∼3% ) , indicating that the subthreshold membrane conductances gh and gT , while not necessary to obtain feature invariant responses , do significantly increase the regions in parameter space for which feature invariance is observed . Our model made important predictions that: 1 ) a spiking nonlinearity was sufficient to give rise to feature invariance and that 2 ) maximum feature invariance was obtained when our model neuron received inputs from both ON and OFF-type sources . To test 1 ) , consider that , if a spiking nonlinearity is sufficient to give rise to feature invariance , which first requires that the neural response be selective to the chirp stimulus , then we expect that the membrane potential response would display less selectivity than the spiking response . We tested this hypothesis by comparing CSI values obtained from the membrane potential to those obtained from spikes across chirp stimuli for two neurons that were recorded from intracellularly . Confirming our hypothesis , we found that the CSI values obtained from the membrane potential ( 0 . 40±0 . 06 ) were significantly less than that obtained from the spiking responses ( 0 . 85±0 . 06 ) ( p = 0 . 002 , signrank test , N = 10 ) . To test 2 ) , we investigated the membrane potential responses of feature invariant TS neurons to sinusoidal input . We note that ON and OFF-type ELL pyramidal cells respond only during the rising and falling phases of such stimuli , respectively [36 , 37] . Thus , if feature invariant TS neurons receive excitatory input from both ON and OFF-type ELL pyramidal cells , then we would expect to see membrane potential depolarizations during both the rising and falling phases of the sinusoidal stimulus . Fig 9A shows the membrane potential response of an example TS neuron to sinusoidal stimulation . Confirming our hypothesis , we observed depolarizations during both the rising and falling phases of the stimulus ( Fig 9A , dashed red lines ) . The average membrane potential response to one stimulus cycle ( Fig 9B ) was clearly bimodal . This was confirmed by computing the power spectral density of the membrane potential that showed maximum power at twice the stimulus frequency ( Fig 9C , red arrow ) and much less power at the stimulus frequency ( Fig 9C , black arrow ) . We computed a bimodality index whose value is 1 if the neuron responds equally at two phases π radians apart and zero if the neuron only responds at one phase ( see Methods ) . We found values of 0 . 67 and 0 . 52 for both TS neurons ( Fig 9C , inset ) . The implications of these results are discussed below . We investigated the coding of natural electrocommunication stimuli in both hindbrain and midbrain electrosensory neurons . Our characterization of natural electrocommunication stimuli revealed strong heterogeneities in waveforms that gave rise to differential response patterns in hindbrain pyramidal neurons . Surprisingly , we found that some TS neurons displayed feature invariant responses to natural electrocommunication stimuli through nonlinear integration of balanced hindbrain ELL neuron afferent input . In order to understand the underlying mechanisms , we built a model based on the Hodgkin-Huxley formalism . Systematically varying model parameters and finding parameter values giving rise to the highest levels of feature invariance through an evolutionary algorithm revealed that very different combinations of parameter values could give rise to approximately similar degrees of feature invariance . This was because the effects of changing a given parameter could be compensated for by changing another parameter . This compensation was seen for all parameters tested with the notable exception of the balance between ON and OFF-type inputs . Although a spiking nonlinearity was sufficient to observe feature invariance , our results nevertheless showed that subthreshold membrane conductances enhanced the robustness of feature invariance overall . We then verified our modeling predictions experimentally and found that: 1 ) the membrane potential response was less selective to chirps than the spiking response; 2 ) feature invariant TS neurons responded with membrane potential depolarizations during both the rising and the falling phase of sinusoidal stimuli suggesting that they do indeed receive input from both ON and OFF-type ELL pyramidal neurons . Our model made the important novel prediction that multiple combinations of parameter values can give rise to feature invariant responses to natural electrocommunication stimuli . The occurrence of similar neural network output despite considerable differences in underlying cellular properties has been previously observed in mathematical models [41–46] . However , such phenomena have mostly been observed experimentally in invertebrate model systems where a given neuron type can be reliably identified across individuals [47–51] . The invariance in output pattern is thought to promote robust function despite perturbations or variability during development [49 , 52 , 53] by permitting homeostasis through compensatory mechanisms [54–56] as well as genetic alterations [49 , 57] . Our results have shown that feature invariant responses to natural electrocommunication stimuli could be obtained in a realistic model through multiple combinations of parameter values . This was because changes in some parameters could be compensated for through changes in others , thereby leading to correlations between parameters as observed elsewhere [58] . Previous studies have suggested that such optimization can be achieved through coupled control of ion channels [48 , 53 , 58] , which could be applicable for h- and T-type currents in this case . Such studies are needed to uncover whether the electrosensory system uses most if not all of the solutions available to give rise to feature invariance . In general experimental studies have found the variability in neuronal and circuit properties to be less than that found in mathematical models: this is likely due to an incomplete understanding of the more complex biological constraints imposed on the molecular , cellular , and network levels [51 , 52 , 59] . Interestingly , we found that not all changes in parameter values could be compensated for as the best feature invariant responses were observed when the model TS neuron received nearly balanced inputs from ON and OFF-type ELL pyramidal neurons , which was confirmed experimentally . This suggests that these TS neurons correspond to the “ON-OFF” ( or type 3 ) neurons that have been previously observed in TS [60 , 61] and that were found to respond selectively to the second order features of electrosensory stimuli [40] . The fact that the neurons responding selectively to the second order features of electrosensory stimuli considered previously tended to spike in response to sinusoidal stimuli [40] whereas the neurons displaying selective but invariant responses to chirps considered here did not suggest that these are not the same neuron type . Since only a small percentage ( ∼5% ) of TS neurons in our dataset displayed selective and invariant responses to chirps , we hypothesize that these must correspond to 1 of the 50 previously anatomically identified cell types within TS [62] . An experimental verification of these predictions is at best challenging as it would require identification of which neural class ( es ) in TS show selective and invariant responses to chirps as well as selective responses to second order features of electrosensory stimuli . Such experiments should also directly verify whether these neurons do indeed receive inputs from both ON and OFF-type ELL pyramidal neurons through direct stimulation of afferent synaptic connections in vitro . These studies are beyond the scope of this paper . Moreover , while previous studies have found that TS neurons receive large amounts of neuromodulatory inputs [63 , 64] , the effects of these on sensory processing have only been studied in hindbrain pyramidal neurons [34 , 65–67] . Thus , further studies are needed to understand how neuromodulators affect sensory processing within TS and whether the feature invariant responses seen in this study are robust to such neuromodulators as seen elsewhere [68] . We also note that previous anatomical studies have found that TS neurons receive inhibitory input exclusively from other neurons located within TS [62] . While our modeling results show that such inhibition is not necessary in order to observe selective but feature invariant responses to chirps , it is conceivable that such inhibition could be used to enhance response selectivity as well as similarity . Further studies are needed to understand the role played by inhibition on sensory processing by TS neurons . Finally , we note that future studies should test whether TS neurons display feature invariant responses to more chirp waveforms . In particular , previous studies have reported that small chirps can occur on top of high frequency beats [24 , 29] . The fact that TS neurons responded similarly to both big chirps occurring on top of high frequency beats and small chirps occurring on top of low frequency beats would suggest that they would respond similarly to small chirps occurring on top of high frequency beats but further studies are required to test this prediction . While previous studies have reported that both peripheral electroreceptor [27–29] as well hindbrain pyramidal [30–32 , 69] neurons can respond to natural electrocommunication stimuli , these have all shown differential responses to different chirp waveforms . In contrast , behavioral studies have shown that weakly electric fish display robust and similar behavioral responses to small chirps despite such heterogeneous waveforms [23 , 24 , 70] , thereby suggesting that the animal perceives these waveforms as similar . If this hypothesis is true , then our results showing that some TS neurons respond in a feature invariant manner would provide a neural correlate of such invariant behavioral responses implying that perception of natural electrocommunication stimuli is largely independent of their characteristics . However , previous studies have shown large heterogeneities in the responses of TS neurons to stimuli: some neurons respond to moving objects in a directionally selective manner [39 , 71 , 72] , others to second order stimulus features [40] , and others to natural electrocommunication stimuli [32] . Such diversity in responses is likely to be a signature of several different parallel processing pathways for natural electrosensory stimuli . The feature invariant responses of some TS neurons to natural electrocommunication stimuli are likely to serve as a reliable detection signal for their occurrence in time [32] . Such a signal would be advantageous as small and big chirp stimuli respectively occur during aggressive and courtship behavior [23 , 24 , 26] , which might help the animal better prepare for each context . It is nevertheless possible that previously described other TS neurons that respond selectively but differentially to different chirp waveforms convey information about chirp attributes in parallel [32] . This is an attractive hypothesis as these TS neurons tended to respond either to small or to big chirps , which would provide a neural correlate of behavioral results suggesting that the animal can actually distinguish between small and big chirp waveforms , which is consistent with the fact that big chirps instead constitute an attractive signal for a potential mate ( see [26] for review ) . Behavioral studies in which the behavioral responses to different small and big chirp waveforms are explicitly considered and compared are needed to test whether the animal can actually distinguish between different waveforms and to test whether and , if so , how perception depends on natural electrocommunication stimulus attributes . We furthermore note that previous studies have found that electrosensory pyramidal neurons display large heterogeneities and can be classified into different classes that are associated with differential expression of ionic conductances [73 , 74] , amount of descending input [37 , 75] , and dendritic morphology [76] . Anatomical studies have shown that the electrosensory lateral line lobe is organized into columns each containing one member from each pyramidal cell class . While all classes project to the midbrain TS [75] , the specific pattern of innervation is not known . Further studies are needed to uncover this pattern . The emergence of feature invariant neuronal responses appears to be a ubiquitous strategy for sensory processing across species and systems . Indeed , such neurons have been observed in both the visual [6 , 12 , 14 , 77 , 78] and auditory [9–11 , 79] pathways . In general , previous studies have noted an increase in both response selectivity and invariance as information propagates to higher brain regions [10 , 12 , 14] . Thus , feature invariance appears to be linked to sparse coding [1] . Moreover , it is generally agreed that feature invariance develops in stages with neurons in higher brain areas displaying progressively more invariant responses [10 , 12 , 14] . Hence , our results showing that some TS neurons display feature invariant responses to natural electrocommunication stimuli , together with previous results showing the emergence of sparse coding in TS [32] , are consistent with those obtained elsewhere . In particular , we found that nonlinear integration of balanced input from ON and OFF-type neurons was sufficient to give rise to feature invariant responses . This mechanism is likely to be applicable to other systems: this is because neurons in more peripheral brain areas also tend to respond to identity preserving transformations of a stimulus through different patterns of excitation and inhibition [79] . Further , ON and OFF-type neurons have been observed in the more peripheral brain areas ( e . g . retina ) of other systems [80 , 81] and studies have shown that higher order neurons likely receive mixed input from both ON and OFF-type neurons [82 , 83] . Additionally , the membrane conductances used in our model are generic and found ubiquitously in the central nervous system [84–86] . These observations , together with many anatomical and physiological similarities between the electrosensory and other systems ( see [17–19] for review ) , suggest that the results obtained in this study are applicable to other systems . We have provided the first experimental evidence that some midbrain electrosensory neurons respond to heterogeneous natural electrocommunication stimuli selectively but in an invariant manner . We have further shown that a simple model receiving input from ON and OFF-type hindbrain neurons that is consistent with known anatomy could reproduce experimental findings for multiple parameter combinations . While a spiking nonlinearity was sufficient to give rise to feature invariance in our model , the addition of subthreshold membrane conductances increased the set of parameter values leading to such invariance . It is likely that the mechanisms giving rise to feature invariance in the electrosensory system will also be found elsewhere . McGill University’s animal care committee approved all procedures . McGill University holds a certificate of ‘Good Animal Practice’ from the Canadian Council on Animal Care and is also certified by the US National Institutes of Health Public Health Service under the 'Policy on Humane Care and Use of Laboratory Animals' with Assurance number A-5006-01 . The weakly electric fish Apteronotus leptorhynchus was used exclusively in this study . Fish were acquired from tropical fish suppliers , acclimated to the laboratory as per published guidelines [87 , 88] . Surgical procedures were explained in detail previously [32 , 34 , 39 , 40 , 89 , 90] . The animal was immobilized with 0 . 1–0 . 5 mg injection of tubocurarine ( Sigma ) intramuscularly . The fish was then transferred to a recording tank and respirated via a mouth tube with flow rate of 10 mL/min . We then glued a metal post rostral to the exposed area of the skull after topical application of lidocaine ( 2% ) to ensure stability during recording . We then drilled a small hole of ∼2 mm2 over the cerebellum and the ELL area in the case of ELL recordings [34 , 91–94] , and over the midbrain optic tectum in the case of TS recordings [32 , 38–40] . Fish were restrained by placing them in a “chirp chamber” as previously described [34] . Chirps were identified as increases in the animal’s own EOD frequency that exceeded 30 Hz and were segregated into small ( type II ) and big ( type I ) as done previously [33] . We recorded chirp responses to sinusoidal waveforms mimicking another fish’s EOD whose frequency was set 10 and 80 Hz above the animal’s own EOD frequency for type II and type I chirps , respectively . The fish’s EOD was recorded and the instantaneous EOD frequency was computed from the inverse of the timing difference between successive zero crossings . The chirp duration was defined as the full width at half-maximum of the frequency increase . The time of occurrence of the chirp was defined as the time at which the EOD frequency is maximal . The beat phase at which the chirp occurred was obtained by expressing the time at which the chirp occurred at relative to the nearest time occurrence of a local maximum of the beat in the past , dividing by the beat period , and multiplying the result by 360° . We computed a similarity metric to capture the variability for small and big chirps that was defined by: SM=1−RMSE/σ , where RMSE is the root-mean squared error between two chirp waveforms computed as: RMSE=⟨[Si−⟨Si⟩−Sj+⟨Sj⟩]2⟩ and σ is the maximum error given by: σ=max[max ( Si ) −min ( Si ) 2 , max ( Sj ) −min ( Sj ) 2] Here <…> denotes the average over time which was computed over a time window of 37 . 5 ms centered on the chirp onset . Density plots were constructed using a binwidth of 0 . 55 ms for duration , 14 degrees for the phase distribution and a binwidth of 0 . 038 for the similarity metric . The number of events occurring per bin were counted and normalized to the maximum number of events found . To assess whether the phase at which a chirp ( small chirp or big chirp ) occurred was homogeneously distributed , we generated 1000 surrogate phase values . These were obtained by randomly permuting a set of phase values between 0 and 2π with the same number of elements as the actual dataset for small and big chirps . The confidence interval was set to be 3 times the standard deviation obtained from the surrogate phase values . Extracellular recordings were made from pyramidal cells within the lateral segment of the ELL because these are most sensitive to the stimuli used in this study [30 , 31] and from TS neurons using metal filled micropipettes [95] as described previously [91] . Intracellular recordings from TS neurons were made using patch pipettes as described previously [71 , 72 , 96] . The recorded signals were sampled at 10 kHz and were digitized by a Power1401 with Spike2 software . ON and OFF-type pyramidal cells can easily be distinguished based on their responses to sinusoidal stimuli as their responses are then in and out of phase , respectively [36] . We note that the electric organ of A . leptorhynchus is neurogenic and is thus not affected by injection of curare-like drugs . All stimuli consisted of amplitude modulations ( AMs ) of the animal’s own EOD . They were produced by first generating a train of sinusoidal waveforms that were triggered by the zero crossing of each EOD cycle with a frequency slightly greater than the fish’s own EOD frequency . The train thus remains synchronized to the animal’s EOD and will either add or subtract depending on its polarity . The modulation waveform ( i . e . the stimulus ) is then multiplied ( MT3 multiplier , Tucker Davis Technologies ) with the train and the resulting signal was applied to the experimental tank after being isolated from ground ( A395 linear stimulus isolator , World Precision Instruments ) via two chloridized silver wire electrodes located ~15 cm on each side of the animal [37] . Stimulus intensities were similar to those used in previous studies [34 , 91] . Chirp stimuli were generated as previously described [32] and were each presented at least 20 times to each neuron . All analysis was performed using custom-built routines in Matlab ( The Mathworks , Natick , MA ) . Action potential times were defined as the times at which the signal crossed a suitably chosen threshold value . From the spike time sequence we created a binary sequence X ( t ) with binwidth dt = 0 . 5 ms and set the content of each bin to equal the number of spikes the time of which fell within that bin . Peri-stimulus time histograms ( PSTHs ) were obtained by averaging the neural responses across repeated presentations of a given stimulus with binwidth Δt = 0 . 1 ms and were smoothed with a 10 . 8 ms long boxcar filter for the small chirps and 5 ms for the big chirp . To quantify the selectivity of a given neuron for the chirp stimulus , the chirp selectivity index ( CSI ) was computed as follows: CSI=RC−RBRC+RB where RC and RB represent the maximum firing rates of the PSTH during the chirp and beat , respectively , similar to what was done previously [32] . The window used to define the chirp time was 100 ms in length , starting at the onset of the chirp in the stimulus . The CSI ranges between -1 and 1 , representing perfect selectivity for the beat at -1 and the chirp at 1 . To measure the selectivity of a model across multiple chirp stimuli , the average CSI was used: CSIavg=1N∑i=1NCSIi where N is the number of chirp stimuli tested and CSIi is the CSI to chirp stimuli i . We also computed the CSI from the membrane potential minus its minimum value in the same way as described above . We first computed the average membrane potential response to sinusoidal stimulation . The bimodality index was computed from the average membrane potential response minus its minimum value in the same way as described previously [40] . First , we performed a circular permutation such that the maximum signal value is now located at 0 . The bimodality index was then obtained by dividing the signal value at half the stimulus period by the signal value at 0 . The invariance of a neuronal response across different stimuli was quantified by comparing the similarity of the spike trains across chirp types and trials . The Victor-Purpura distance ( VPD ) , a metric-space measure of the distance between two spike trains , was used to quantify this [35] . Briefly , the VPD computes the total cost of transforming one spike train into another via an optimal series of elementary operations and used q = 100 s−1 as done previously [32] . Thus , the feature invariance of a neuron was computed by the average VPD across all pairs of trials as follows: VPDavg=1M∑i=1N∑j=iN∑k=1NT∑l=uijNTVPD ( Ci ( k ) , Cj ( l ) ) where NT is the number of trials for each chirp stimulus , M is the number of combinations of pairs of trials ( i . e . the number of times the VPD is computed ) , uij is 1 if i ≠ j and is k + 1 otherwise , and VPD ( Ci ( k ) , Cj ( l ) ) represents the VPD between the spike trains in response to the kth presentation of chirp i and the lth presentation of chirp j , respectively . For computing the invariance of neural responses of population averages , individual spike trains were not accessible for calculating the VPD . Therefore , the root-mean squared error ( RMSE ) between the PSTHs of the responses of the populations to the various chirps was computed instead . Average PSTHs were computed by averaging the PSTHs across a given population of cells for each chirp type , and these were used to compute the average pairwise RMSE between all pairs of PSTHs as follows . First , in order to prevent small differences in timing from biasing the result ( because an identical response shifted slightly in time will have a high RMSE ) , a cross-correlogram was computed between the two PSTHs , followed by a circular shift of one PSTH by the time shift at the maximum value of the cross-correlogram ( i . e . leading to maximum overlap between the PSTHs ) . These PSTHs were then used to compute the final average RMSE as: RMSEavg=1P∑i=1N∑j=iN1L∑k=1L ( PSTHi ( kΔt ) −PSTHj ( kΔt ) ) 2 where P is the number of pairs of PSTHs considered , L is the number of bins for the PSTH and Δt is the binwidth . The measures for selectivity and spike train invariance were used to compute a feature invariance ( FI ) score as follows: FI=H0 ( CSIavg−αVPDavg ) where H0 is the Heaviside step function , and α >0 is a constant . We note that the choice of α>0 is arbitrary and does not affect the qualitative nature of our results as long as the value chosen is not too large . We chose α = 0 . 01 in order to emphasize the fact that the responses of some TS neurons to different chirp waveforms were far more similar and selective than those of ELL neurons . We only included TS neurons for which we obtained FI>0 . 2 . Importantly , we obtained FI<0 . 2 for all ELL neurons in our dataset . For computing FI scores of population averages via their PSTHs , an FI score based on the RMSE was instead used: FIRMSE=H0 ( CSIavg−γRMSEavg ) where γ = 0 . 0041 is a constant whose value was chosen by requiring that FI = FIRMSE for a representative , feature invariant TS neuron . We also compared VPDavg and RMSEavg computed from single ELL and TS neuron spike train and PSTH responses , respectively . Overall , we found a strong positive correlation between both quantities ( R = 0 . 90 , N = 54 , p<<10−3 ) , indicating that using either measure will not affect our results qualitatively . We furthermore found similar values of FI and FIRMSE for ELL ON-cells ( 0 . 000±0 . 000 vs . 0 . 0055±0 . 0036 ) , OFF-cells ( 0 . 004±0 . 004 vs . 0 . 0232±0 . 0180 ) , and TS neurons ( 0 . 54±0 . 09 vs . 0 . 577±0 . 082 ) . Statistical significance was assessed through one-way analysis of variance ( ANOVA ) with the Tukey-Kramer method of correcting for multiple comparisons at the p = 0 . 05 level . Values are reported as mean ± standard error throughout the text . Our model is based on the Hodgkin-Huxley formalism and considers a single TS neuron receiving excitatory synaptic input from ON and OFF-type ELL pyramidal cell populations based on known anatomical data [62] . We obtained the synaptic input by pooling the recorded activities of ON and OFF-type pyramidal cells in response to the chirp stimuli used in this study . The population-averaged PSTH was then converted to a time varying conductance as follows: gON ( t ) =ζgmax ( PSTHE ( t ) *[tτsynexp ( 1−tτsyn ) H0[t]] ) gOFF ( t ) =ζgmax ( PSTHI ( t ) *[tτsynexp ( 1−tτsyn ) H0[t]] ) where “*” is the convolution operator and PSTHE , PSTHI are the population-averaged PSTHs for ON and OFF-type pyramidal cells , respectively . We used gmax = 0 . 13 μS , ζ = 0 . 0005 , and τsyn = 20 ms . The time varying synaptic current applied to the model TS neuron was then equal to: Isyn ( t ) =−2Ws ( σBgON ( t ) + ( 1−σB ) gOFF ( t ) ) ( V−Esyn ) =−2gsyn ( σBgON ( t ) /gmax+ ( 1−σB ) gOFF ( t ) /gmax ) ( V−Esyn ) where Ws is a constant that was varied systematically , the parameter σB controls the relative proportion of input from ON-type pyramidal cells , which is important as previous studies have shown that this proportion can vary substantially across TS neurons [40] , Esyn = 0 mV is the reversal potential of the excitatory synapse . Note that we report the value of gsyn = Ws gmax in the figures . TS neurons were modeled with the Hodgkin-Huxley formalism [97] , via the following system of stochastic differential equations: dVdt=1C ( INa+IKDR+Ih+IT+Ileak+Isyn+Ibias+σnoiseξ ( t ) ) dηdt=Φη∞ ( V ) −ητηdhdt=h∞ ( V ) −hτhdndt=n∞ ( V ) −nτn where C = 1 μF is the cell membrane capacitance , V is the transmembrane voltage , INa is the spiking sodium current , IKDR is the delayed rectifier potassium current , Ih is the hyperpolarization activated current mediated by HCN channels , IT is the low-threshold T-type calcium channel current , Isyn is the synaptic current from the ELL defined above , Ileak is the leak current , Ibias is a constant bias current , and η , h , and n represent the voltage-dependent channel activation variables with Φ = 2 and τη = 30 ms . ξ ( t ) is a time-varying stochastic Gaussian white noise process with mean zero and standard deviation 0 . 8 . We used σnoise = 1 nA unless otherwise noted . The equations governing T-type calcium , sodium , and delayed rectifier potassium channel activation/inactivation were as follows: m∞ ( V ) =αm ( V ) αm ( V ) +βm ( V ) αm ( V ) =0 . 1 ( V+40 . 7 ) 1−exp ( −0 . 1 ( V+40 . 7 ) ) βm ( V ) =4exp ( −0 . 05 ( V+49 . 7 ) ) η∞ ( V ) =10 . 5+0 . 25+exp ( V+826 . 3 ) s∞ ( V ) =11+exp ( −V+637 . 8 ) n∞ ( V ) =αn ( V ) αn ( V ) +βn ( V ) αn ( V ) =0 . 01 ( V+40 . 7 ) 1−exp ( −0 . 1 ( V+40 . 7 ) ) βn ( V ) =0 . 125exp ( −0 . 0125 ( V+50 . 7 ) ) τn ( V ) =0 . 05αn ( V ) +βn ( V ) h∞ ( V ) =11+exp ( 0 . 151 ( V+73 ) ) τh=exp ( 0 . 033 ( V+75 ) ) 0 . 011[1+exp ( 0 . 083 ( V+75 ) ) ] and the ionic currents are given by: INa=−gNam∞3 ( V ) ( 0 . 85−n ) ( V−ENa ) IKDR=−gkn4 ( V−EK ) Ih=−ghh ( V−Eh ) IT=−gTs∞3 ( V ) η ( V−ECa ) Ileak=−gleak ( V−Eleak ) where the reversal potentials of the ionic conductances are given by: ENa = 60 mV , EK = -85 mV , Eleak = -65 mV , Eh = -30 mV , ECa = 120 mV . The maximal conductances , unless otherwise noted , are given by gNa = 30 μS , gK = 10 μS , gleak = 0 . 18 μS , gh = 7 μS , and gT = 0 . 32 μS . Parameters for Ih were the same as those used previously [84 , 98] . We concentrated on the subthreshold currents IT and Ih because these can promote spiking following hyperpolarization and have furthermore been shown to be present in midbrain neurons [71 , 99] . The above system of equations was numerically simulated with the Euler-Maruyama algorithm [100] , using an integration time-step of 0 . 025 ms . The model TS neuron’s output was analyzed in the same fashion as the experimental data . The robustness of feature invariance to changes in model parameter values was computed as the % of pixels for which we had FI ≥ 0 . 7 . We used a differential evolution ( DE ) algorithm [101 , 102] in order to find sets of parameter values for which the model neuron’s output to different input chirp waveforms matched that seen experimentally in TS neurons . Specifically , we implemented a variant of a previously proposed method for global parameter estimation of Hodgkin-Huxley models [102] . DE evolves a set ( or population ) of parameter vectors ( i . e . “individuals” ) by minimizing a fitness function Ffit over a series of iterations ( i . e . “generations” ) . In keeping with the notation used in previous studies [101 , 102] , we denote Xkr ( i ) as parameter i for individual r of generation k . First , the population of K individuals is randomly initialized with values for each of the D parameters , uniformly distributed within the boundary constraints for each . Here , the parameter vector is made up of σB , WS , Ibias , gh , gT ( i . e . D = 5 ) . For each individual at every generation , a new individual is constructed by two operations consisting of "differentiation" and "recombination" . In differentiation , the rth new parameter vector Xk , trialr is built by combining three other individuals Xkr1 , Xkr2 , and Xkr3 , where r1≠r2≠r3: Xk , trialr=Xkr1+ ( Xkr2−Xkr3 ) F∀r=1 , … , N where the differential weight F = 0 . 5 , and the three individuals are chosen based on a probability distribution that is preferentially weighted for more fit ( i . e . lower fitness score ) individuals: pkri=λexp ( −Ffit ( Xkri ) max∀j ( Ffit ( Xkj ) ) ) ∀ri=1 , … , N where λ is a normalization constant such that the sum of probability values is equal to one . Recombination is then performed as follows: Xmutr ( i ) ={Xk , trialr ( i ) ifu<CRXkr ( i ) otherwise∀r=1 , … , N;i=1 , … , D where u is a random variable generated from a uniform distribution U ( 0 , 1 ) and with crossover probability CR = 0 . 9 . Selection is finally performed to produce the next generation via: Xk+1r={XmutrifFfit ( Xmutr ) ≤Ffit ( Xkr ) Xkrotherwise∀r=1 , … , N In this study , the fitness function for a given individual was defined as: Ffit ( Xkr ) =exp ( −FIXkr ) where FIXkr is the FI score computed from simulating a model using the parameters encoded in individual Xkr on the five chirp stimuli . To encode boundary constraints of the parameter values within physiologically realistic ranges , the “resampling” approach was adopted , based on previous empirical results [103] . Whenever a parameter ( gene ) value violates its constraints , resampling enforces the boundary conditions by rerunning the differentiation step with three new individuals chosen uniformly randomly .
We provide the first experimental evidence showing that midbrain electrosensory neurons in the weakly electric fish species Apteronotus leptorhynchus can respond in an invariant manner to the heterogeneous stimulus waveforms associated with natural electrocommunication stimuli . Interestingly , hindbrain neuron populations providing afferent input did not display feature invariant responses . In order to understand the mechanisms that mediate the emergence of feature invariance in midbrain neurons , we built a model based on the Hodgkin-Huxley formalism . An evolutionary algorithm identified multiple combinations of parameter values that all gave rise to responses that matched those seen experimentally . In particular , balanced input from ON and OFF-type cells was necessary to observe feature invariance . Moreover , while a spiking nonlinearity was sufficient to observe invariant responses , the addition of subthreshold membrane conductances to the model enhanced the regions in parameter space for which feature invariant responses were observed . Further analysis of membrane potential responses confirmed our modeling predictions . Our model thus makes the important prediction that multiple mechanisms will lead to feature invariance and we discuss the implications for sensory processing in the electrosensory and other systems .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[]
2015
Electrosensory Midbrain Neurons Display Feature Invariant Responses to Natural Communication Stimuli
Translational errors occur at high rates , and they influence organism viability and the onset of genetic diseases . To investigate how organisms mitigate the deleterious effects of protein synthesis errors during evolution , a mutant yeast strain was engineered to translate a codon ambiguously ( mistranslation ) . It thereby overloads the protein quality-control pathways and disrupts cellular protein homeostasis . This strain was used to study the capacity of the yeast genome to compensate the deleterious effects of protein mistranslation . Laboratory evolutionary experiments revealed that fitness loss due to mistranslation can rapidly be mitigated . Genomic analysis demonstrated that adaptation was primarily mediated by large-scale chromosomal duplication and deletion events , suggesting that errors during protein synthesis promote the evolution of genome architecture . By altering the dosages of numerous , functionally related proteins simultaneously , these genetic changes introduced large phenotypic leaps that enabled rapid adaptation to mistranslation . Evolution increased the level of tolerance to mistranslation through acceleration of ubiquitin-proteasome–mediated protein degradation and protein synthesis . As a consequence of rapid elimination of erroneous protein products , evolution reduced the extent of toxic protein aggregation in mistranslating cells . However , there was a strong evolutionary trade-off between adaptation to mistranslation and survival upon starvation: the evolved lines showed fitness defects and impaired capacity to degrade mature ribosomes upon nutrient limitation . Moreover , as a response to an enhanced energy demand of accelerated protein turnover , the evolved lines exhibited increased glucose uptake by selective duplication of hexose transporter genes . We conclude that adjustment of proteome homeostasis to mistranslation evolves rapidly , but this adaptation has several side effects on cellular physiology . Our work also indicates that translational fidelity and the ubiquitin-proteasome system are functionally linked to each other and may , therefore , co-evolve in nature . Fidelity of protein synthesis has a substantial impact on cellular survival [1 , 2] . There might be incorporation of one amino acid for another ( missense errors ) , premature termination of protein synthesis , frameshift errors and read-through of stop codons ( nonsense errors ) . Even if the cell has the right protein sequence at the right time , errors can occur during folding , synthesis of posttranslational modifications , or translocation of proteins across membranes . Recent works indicate that many steps in protein production are strikingly error-prone . For example , the average missense error rate during translation is between 10−3 and 10−4 [3] . While such estimates are rough and are based on a variety of methodologies , they clearly indicate that error rates during protein production are three to five orders of magnitude higher than DNA-replication errors . Perfectly replicated unicellular genomes are commonplace , but perfectly synthesized proteomes never occur . Although the exact amount is currently debated [4 , 5] , it appears that a large fraction of the proteins in eukaryotic cells is defective . Most of these faulty proteins are degraded by the proteasome or aggregate [1] . Indeed , disruption of translational fidelity with aminoglycoside antibiotics kills bacteria [6] , cells with impaired proofreading have abnormal morphologies [7 , 8] , and enhanced translational error rates in mammals cause disease [9] . For example , a single mutation in a tRNA synthetase yields widespread translation errors and protein misfolding and consequent death of Purkinje cells in the mouse brain [9] . Moreover , decreased accuracy of protein synthesis due to altered codon–anticodon interactions leads to protein aggregation and saturation of the protein quality-control networks [10] . Similarly , mutants conferring a hyper-accurate translation phenotype grow more slowly than wild-type cells [8 , 11] as excessively accurate ribosomes are kinetically less efficient than wild-type ribosomes . Thus , accuracy of translation has an optimal value reflecting a trade-off between costs of kinetic proofreading and cellular side consequences of erroneous protein accumulation . A variety of quality-control mechanisms exist that reduce the rate at which errors occur ( error prevention ) or limit harmful effects if an error has already been made ( error mitigation ) [12] . Error reduction may be achieved by an improved proofreading machinery or usage of codons with corresponding highly abundant tRNAs [12] . Furthermore , as protein synthesis errors frequently lead to protein misfolding and aggregation , cells may achieve enhanced tolerance to errors through modification of the cellular chaperone networks , protein degradation pathways , or by the evolution of robust protein folding mechanisms [13–15] . The relative contribution of these pathways to safeguarding the integrity of biological information has remained largely a terra incognita , not least due to the shortage of laboratory evolution studies . To investigate this problem systematically , we took advantage of the availability of a previously engineered Saccharomyces cerevisiae strain that mistranslates proteins constitutively at high level [16] . The engineered construct ( tRNACAGSer ) misincorporates serine ( Ser ) at leucine ( Leu ) sites encoded by the CUG codon . As CUG codons in the yeast genome are distributed over 89% of its protein coding genes , this system is ideal to study the impact of protein mistranslation events on a proteome scale . Since more than 60% of the CUG encoded residues are buried in S . cerevisiae proteins [17] , and serine and leucine are chemically very different from each other , mistranslation events generate protein misfolding and proteotoxic stress [10] . To shed light on the evolution of safeguarding mechanisms , we initiated laboratory evolutionary experiments with a strain with initially high mistranslation rate . Fitness rapidly increased during the course of laboratory evolution and was mediated by large-scale chromosomal duplication and deletion events . Robustness to translational errors was achieved through accelerated protein turnover , albeit at a considerable energetic cost . Using a synthetic tRNACAGSer construct on a plasmid , we have previously engineered codon misreading in a diploid yeast strain that misincorporates serine at the leucine CUG codon [16 , 18] . A previous quantitative mass-spectrometry study indicates that the tRNACAGSer misincorporates 1 . 4% of serine at the CUG sites [19] . As the background mistranslation error in yeast is around 0 . 001% [20] , the construct generates a 1 , 400-fold increase in mistranslation at CUG sites on a proteome-wide scale . We used established protocols specifically designed to measure fitness in yeast populations [21 , 22] . Growth was assayed by monitoring the optical density at 600 nm wavelength ( OD600 ) of liquid cultures of each strain on synthetic complete ( SC ) leucine dropout medium . The tRNACAGSer construct reduced growth rate by 40% , suggesting that mistranslation has a substantial fitness cost in an otherwise stress-free environmental condition ( Fig 1A ) . Next , we initiated laboratory evolutionary experiments , starting from a single clone , that carried the tRNACAGSer plasmid . Eleven replicate populations were cultivated in a standard laboratory medium and 1% of each parallel evolving culture was diluted into a fresh medium every 48 h . Populations were propagated for approximately 250 generations ( 70 d ) . To control for potential adaptation to the medium , we also established 11 populations starting from the isogenic wild-type strain that carried the empty vector ( instead of tRNACAGSer ) . Next , we isolated a single colony from each independently evolved population after 70 d of evolution . All starting and evolved lines were subjected to high-throughput fitness measurements by monitoring growth rates in liquid cultures . We found that the evolving wild-type control lines showed a marginal 3 . 7% fitness improvement ( S1 Fig ) . By contrast , the fitness of the lines with tRNACAGSer improved by 13 . 3%–51 . 7% , and many of them approximated wild-type fitness ( Fig 1A ) . Plasmid sequence analysis in the evolved lines confirmed that the rapid fitness gain was not due to defective mutations in tRNACAGSer ( S2 Fig ) . Next , tRNACAGSer was swapped for the corresponding empty vector in all starting and evolved lines . Most evolved lines showed 3%–11% growth deficit compared to their ancestor ( Fig 1B ) . Thus , the mutations that have accumulated during the course of laboratory evolution are favorable when mistranslation rate is high but generally reduce fitness otherwise . To investigate the molecular changes underlying adaptive evolution in the laboratory , we re-sequenced the complete genomes of 11 independently evolved clones , all of which showed substantial fitness improvements under high mistranslation rate . In total , we identified 431 independent mutational events ( SNPs and large genomic rearrangements ) . On average , we detected 39 . 18 point mutations , 1 deletion , and 2 duplications per line . A strong pattern of parallel evolution emerged at the level of genes and functional modules . Of the mutated genes , 26% with non-synonymous mutations were shared by at least two lines , and some were shared extensively ( S1 Table ) . Several functional units were repeatedly mutated . The examples include rRNA maturation ( MRD1 [S000006316] ) , transcription initiation and elongation ( BDF1 [S000004391] , TFG2 [S000003237] ) , and the ubiquitin-proteasome system ( SSM4 [S000001292] , HSM3 [S000000476] ) . Large-scale duplications ( including segmental or whole-chromosome duplication ) and genomic deletions were also prevalent ( Table 1 ) . In total , 33 large-scale duplication and partial chromosome deletion events were observed . Many of the evolved lines accumulated 4–5 gross chromosomal rearrangements . Strikingly , however , only seven major different types of such events were observed , most of which occurred repeatedly . For example , over 90% of the evolved lines lost a small part of one copy of chromosome V , which spanned over 127 kb . The evolved lines also frequently carried duplications of large segments of chromosomes IV ( 540 kb ) and VII ( 238 kb ) . Borders of these duplicated genomic regions were flanked by transposon of yeast ( Ty ) elements , indicating that the observed genomic changes were mediated by homologous recombination events between Ty elements ( S2 Table ) . The repeated evolution of chromosomal rearrangements at the same breakpoint suggests that these mutational events confer adaptive advantage to the carrying cells . To investigate this issue further , we grouped the mutated genes into several major functional categories based on database information . The observed large-scale duplication events were enriched in genes involved in ribosomal biogenesis ( e . g . , DRS1 [S000003931] , SOF1 [S000003934] , RIX7 [S000003957] ) , rRNA processing ( e . g . , GRC3 [S000003958] , NOC3 [S000003992] , SDO1 [S000004012] ) , and proteasome core complex assembly ( PRE9 [S000003367] ) . Significantly , the small genomic region repeatedly deleted on chromosome V carries two genes encoding deubiquitinating enzymes ( UBP3 [S000000953] , UBP5 [S000000946] ) . Last , we studied the impact of the accumulated mutations on genomic expression . The analysis focused on two independently evolved strains ( lines 1 and 4 ) , both of which carried typical large-scale genomic rearrangements ( Table 1 ) . We grew these strains to mid-log phase in standard laboratory medium and measured genomic expression relative to the ancestor strain ( DNA microarrays were used for this purpose ) . Only genes that showed at least a 2-fold change in expression were considered further . As expected , copy number variation had a significant impact on gene expression in the evolved lines ( Fig 2 , see S3 Table for the full dataset ) . To reveal the functional enrichment of genes with changed expression unrelated to chromosomal duplication or deletion , we focused on genomic segments that have not undergone copy number changes during the course of evolution . The two strains showed correlated changes in genomic expression ( Spearman correlation , r = 0 . 54 p < 10−6 ) . Given this correlation , we focused on the set of genes with expression changes in the same direction in the two strains . A total of 425 genes were affected in both strains , of which 168 were induced and 257 repressed , relative to the ancestor . Functional enrichment analysis revealed that genes involved in rRNA processing , ribosome biogenesis and amino acid biosynthesis were up-regulated ( Table 2 ) . The precise molecular mechanisms underlying up-regulation of ribosomal genes in the evolved strains are unknown . It is worth noting , however , that genes encoding ribosomal proteins form the tightest cluster of coordinately regulated genes , and one of the most prominent transcription factors controlling the ribosomal regulon is IFH1 ( S000004213 ) [24] . Significantly , IFH1 is located on chromosome XII , and this chromosome has undergone a complete duplication during the course of laboratory evolution ( Table 1 ) . Taken together , the genomic analyses demonstrate that evolution preferentially targeted genes involved in translation and the ubiquitin-proteasome system . Based on results of the previous section , we next asked whether the observed mutations influence protein synthesis rate . We used an established biochemical assay; protein pulse labeling with [14C ( U ) ]-L-Amino Acid Mixture [25 , 26] . To ensure that changes in protein synthesis rate reflect the impact of the accumulated mutations in the evolved lines ( rather than growth rate mediated effects of mistranslation ) , tRNACAGSer was swapped for the corresponding empty vector in the ancestor and the evolved lines . Ancestor and evolved cells were incubated with a mixture of carbon-14 labeled L-amino acids . The rate of carbon-14 labeled amino acids incorporation in newly synthesized proteins reflects directly the rate of mRNA translation in vitro . Protein synthesis was quantified according to carbon-14 labeled L-amino acids incorporation detected by a scintillation counter . We found that the evolved lines show higher protein synthesis rate than the corresponding ancestor ( Fig 3 ) . Several mutated genes were involved in the regulation of tRNA transcription ( RNA polymerase III ) , tRNA export ( SOL1 [S000005317] ) , tRNA surveillance and degradation ( TRF5 [S000005243] ) . A serine tRNA was also repeatedly mutated , mostly in the variable arm of this molecule ( Fig 4A , S4 Table ) , which is recognized by the seryl-tRNA synthetase ( SerRS ) . Indeed , the yeast SerRS recognizes the three G-C base-pairs of the variable arm of serine tRNAs and the discriminator base at position 73 ( G73 ) . Based on these observations , we assumed that evolution has acted to alter tRNA stability and cellular abundance . Changes in the tRNA pool could subsequently reduce the rate of mistranslation during the evolution period . Misreading activity of tRNACAGSer was then tested by using a previously developed β-galactosidase assay [10] . Briefly , the Escherichia coli LacZ gene contains 54 CTG codons , and misincorporation of serine at these leucine codons generates a combinatorial array of mutant β-galactosidase molecules . The altered stability of these statistical protein ensembles can be quantified using thermal denaturation assays [10] . The high number of CTG codons present in the LacZ gene combined with the different chemical properties of serine ( polar amino acid ) and leucine ( hydrophobic amino acid ) make β-galactosidase a highly sensitive reporter , allowing for monitoring misreading activity by tRNACAGSer . Prior to evolution , presence of tRNACAGSer caused a 62% reduction in β-galactosidase activity ( Fig 4B ) . Next , we compared β-galactosidase activities of the ancestor and the evolved tRNACAGSer carrying lines . A relatively small but significant increase in β-galactosidase activity was found in the evolved lines , indicative of a slight reduction of misreading by tRNACAGSer . When tRNACAGSer was swapped for the corresponding empty vector , all evolved lines had high β-galactosidase activities , similar to that of the ancestor ( Fig 4C ) . This result confirms that the presence of tRNACAGSer is crucial for serine misincorporation . Taken together , these results indicate that evolution acted to reduce mistranslation rate , probably by influencing tRNA aminoacylation , tRNA pool distribution , or usage during translation . Indeed , we note that quantification of tRNA expression by northern blot revealed a small but significant reduction in the cellular abundance of mistranslating tRNA in three evolved lines ( S3 Fig ) . Even if a translational error is not prevented , the fitness consequences that ensue could still be reduced [1 , 12] . Fitness cost of mistranslation could be partly due to protein misfolding , protein aggregation , and consequent induction of cellular toxicity [1] . Indeed , a prior work showed that tRNACAGSer in yeast initiates a proteotoxic stress response and activates the unfolded protein response pathway [10] . Therefore , we hypothesized that changes in the ubiquitin-proteasome system mitigate the harmful consequences of mistranslation by reducing the extent of protein aggregation . To shed light on a possible link between mistranslation and protein aggregation , we measured the level of cellular aggregation in cells subjected to mistranslation and compared it with that of the wild-type control . An established method based on aggregation of a fluorescently tagged human protein ( VHL , von Hippel–Lindau ) was applied [27] . Active quality-control machinery in wild-type yeast prevents the aggregation of VHL [28] . However , upon overload of the quality-control machinery ( including Hsp70 and Hsp90 chaperone complexes ) , misfolded VHL molecules form aggregates that are seen as foci in the cells [27] . As the fluorescent tag ( mCherry ) remains functional , localization of aggregation spots in the cells are detectable by fluorescence microscopy . We found that tRNACAGSer initiated aggregation of VHL in the ancestor ( Fig 5A ) . This result demonstrates that targeting of VHL-mCherry for proteasomal degradation is compromised in these cells . In sharp contrast , the evolved strains showed substantially decreased VHL focus formation , indicative of a reduced aggregation propensity in these lineages ( Fig 5B ) . One may argue that aneuploidy , such as partial chromosomal duplications found in the evolved lines , are expected to have an opposite effect on protein aggregation , since these genomic changes can overload the protein quality-control machinery [28] . Nevertheless , this is no longer so when aneuploidy-tolerating mutations in a deubiquitinating enzyme ( such as Ubp3p ) are also present [29] . We previously showed evidence for the loss of a single genomic copy of UBP3 in most of the evolved lines . In addition , we have also demonstrated that partial chromosomal duplications are beneficial for speeding up translation ( due to the overrepresentation of ribosome-associated genes in these genomic regions ) . Therefore , these genomic rearrangements are more likely to be favorable , without any negative effects on aggregation . As a net outcome , we found that the level of protein aggregation is reduced , rather than increased , in the evolved lines ( Fig 5B ) . Based on the above findings , we next asked whether selection acted to increase proteasome-mediated degradation of misfolded proteins . Indeed , short-lived proteins display higher average aggregation propensity [30] , indicating that efficient degradation of high-turnover proteins is maintained to minimize the danger of aggregation . The genomic analysis is consistent with the hypothesis , as genes related to the ubiquitin-proteasome system were repeatedly mutated ( S1 Table ) . For example , three independently evolved lines carried mutations in the HSM3 gene , which encodes a proteasome-interacting protein involved in the assembly of the 19S proteasomal regulatory particle . Perhaps most significantly , a small segment of chromosome V carrying genes of deubiquitinating enzymes ( UBP3 , UBP5 ) was repeatedly lost in over 90% of the evolved lines ( Table 1 ) . Tagging of a target protein by ubiquitin specifies cellular location and frequently directs it to the 26S proteasome for degradation [31] . Importantly , the ubiquitin-proteasome system is actively involved in eliminating misfolded proteins . Proteasome-associated deubiquitinating enzymes , such as Ubp3p , cleave ubiquitin–protein bonds , and thereby enhance the free intracellular level of ubiquitin [32] . By reversal of ubiquitination , Ubp3p diverts proteins away from the proteasome system . Accordingly , reduced Ubp3p level decreases the intracellular level of free ubiquitin and simultaneously increases the fraction of proteins destined for destruction [32] . Ubiquitin is critical for the survival of yeast cells in the presence of protein synthesis inhibitor cycloheximide [33] . Therefore , the evolved lines are expected to show increased sensitivities to this agent . This was indeed so ( Fig 6A ) . Along with the observed mutations , a biochemical assay supports the notion that proteasome activity changed during the course of laboratory evolution . Proteasome activity of the ancestor and the evolved lines was quantified as described previously [34] , using a fluorogenic peptide as substrate . Briefly , total protein was extracted from exponentially growing cells and activity was determined for 100 μg of protein extract in assay buffer with s-LLVY-MCA . The measured fluorescence emission showed that proteasome chymotrypsin-like activity increased 1 . 4–2 . 8-fold in the laboratory evolved lines compared to the ancestor ( Fig 6B ) . Both translation- and proteasome-mediated degradation of proteins are exceptionally energy-consuming cellular processes . Therefore , the accelerated proteome turnover is expected to incur substantial energy costs , leading to fitness deficit when external nutrients are limited . Indeed , in agreement with expectations , the evolved lines showed reduced growth rate in glucose- and amino-acid–limited medium compared to their ancestor ( S4A and S4B Fig ) . To investigate whether the enhanced energetic costs of accelerated protein turnover and the associated fitness deficit invokes a selection pressure to alter cellular physiology , we measured glucose uptake using established protocols [35] . The comparison revealed that evolved cells generally internalized more glucose molecules than the ancestor cells did ( Fig 7 ) . Several mutations could be responsible for the observed increase in glucose uptake , two of which are especially noteworthy . Nine out of 11 strains carried a segmental duplication of chromosome 4 , which spans over 540 kb ( Table 1 ) . This small region contains three genes involved in hexose transport ( HXT3 [S000002753] , HXT6 [S000002751] , and HXT7 [S000002750] ) . Significantly , prior work showed that when yeast is evolving in a glucose-limited environment , populations amplify an overlapping region of chromosome IV , which includes these high-affinity hexose transporters [36 , 37] . Moreover , MTL1 ( S000003255 ) , a membrane sensor of stress during glucose starvation , was also mutated in seven lines independently . Taken together , the genetic and biochemical data suggest that the evolved lines demand more glucose for cellular proliferation . Last , we tested the response of the evolved lines to complete nutrient depletion . Using standard lifespan assays [38] , we found that 8 out of the 11 studied evolved lines rapidly lost viability upon prolonged culturing in starvation ( Fig 8A ) . We suspected that changes in the ubiquitin-proteasome system could partly be responsible for these changes . Due to the exceptionally high energetic costs of translation , mature ribosomes are selectively degraded by autophagy upon nutrient starvation [39] . This pathway , termed ribophagy [39 , 40] , has an important role in adjusting the number of ribosomes to match the cellular needs . Importantly , ribophagy crucially depends on the expression of UBP3 [39] . Therefore , the evolved lines with loss of a single genomic copy of UBP3 may exhibit impaired capacity to degrade mature ribosomes . To analyze quantitative changes in ribosome destruction , an established protocol was used [27] . The ribosomal protein ( Rpl25p [S000005487] ) was tagged with green fluorescent protein ( GFP ) . Whereas the fused protein is distributed evenly in the cytoplasm under nutrient-rich conditions , the protein accumulates in the vacuole during starvation [27] . As previously demonstrated , the level of cellular relocalization of this construct is a reliable indicator of ribosome turnover [27] . The evolved lines carried a single genomic copy of the UBP3 gene , the only exception being line 7 , where both copies remained intact ( Table 1 ) . Remarkably , all lines with reduced UBP3 dosage showed markedly reduced Rpl25p–GFP levels in the vacuole during starvation compared to both line 7 and the corresponding ancestor ( Fig 8B and 8C ) . We note also that Rpl25p–GFP level was still somewhat lower in line 7 than that in the ancestor , indicating the existence of other mutations influencing ribophagy . Although we cannot exclude the possibility that some degradation occurs from unrelated processes , these results clearly indicate that vacuolar processing of Rpl25p–GFP proteins is impaired in the evolved lines . Finally , mortality rate upon starvation is expected to be a complex process also dependent on the rate of using the energy resources of the cell . Future works are needed to elucidate the molecular details of these processes . The cellular damage of mistranslation can have many different sources [1] , including ( i ) loss of protein function , ( ii ) extra clean-up costs by overloading the quality-control systems involved in degradation or refolding of misfolded proteins , and ( iii ) induction of toxicity by protein aggregation . To investigate how organisms mitigate the deleterious effects of mistranslation during evolution , a mutant tRNA was expressed in S . cerevisiae . The construct induced over a 1 , 000-fold increase in mistranslation at the CUG codon and exposed the global detrimental effects of codon-specific ambiguity on the proteome . By integrating evolutionary experiments and genomic and functional analyses of the evolved lines , the following main conclusions were reached . First , the fitness defect due to mistranslation was rapidly mitigated during laboratory evolution . Evolution reduced the rate at which errors occur ( error prevention ) and mitigated the harmful effects of errors ( error mitigation ) as well . Interestingly , we failed to find evidence for mutations at individual CTG sites . This is not completely unexpected , as mistranslation at CUG codons affects thousands of sites simultaneously , each of which probably have a relatively small contribution to fitness individually . Therefore , the selection pressure for reduction of mistranslation rate locally ( i . e . , at individual genic sites ) may be too small to be detected in the laboratory . Second , our work demonstrates that mistranslation initiates rapid evolution of genomic architecture ( see also [41] ) . Convergent evolution of nearly identical chromosomal duplications indicates that these mutational events confer specific fitness advantages [42–44] , most likely by simultaneous dosage increment of protein complexes and cellular subsystems involved in error mitigation . Indeed , chromosomal duplications are known to promote microbial evolution under environmental stress [45–47] . Interestingly , we found repeated occurrence of loss of function mutations in a repressor of the transcription of histone gene ( HIR2 [S000005564] ) . Inactivation of HIR2 decreases heterochromatin-mediated gene silencing and increases chromosomal instability due to defective kinetochore formation [48 , 49] . Future work should elucidate whether HIR2-mediated chromosomal instability promotes the rise of adaptive duplicated chromosome segments during evolution . Third , reduction of protein aggregation was a main target of evolution . Indeed , maintenance of proteome homeostasis is crucial for cell survival [50] . Proteins must reach their native conformation , refolded when necessary , and damaged proteins must be degraded . Folding and degradation of misfolded proteins are assisted by the concerted action of molecular chaperones and the proteasome [51] . In times of proteotoxic stress caused by mistranslation , these protein quality-control mechanisms are overwhelmed , leading to the accumulation of misfolded proteins and protein aggregates [52] . Fourth , the evolutionary adjustment of proteome homeostasis to mistranslation is achieved through acceleration of protein turnover [53] , a process that is determined by the combined rates of protein synthesis and ubiquitin-proteasome system mediated degradation . This is in agreement with prior works suggesting that proteins with high turnover rate and , thus , with short lifetime will have lower risk of misfolding than long-lived proteins [30] . As newly synthesized polypeptides compete for the protein folding machinery , a large fraction of error-free proteins are degraded shortly after translation [1] . Deubiquitinating enzymes , such as Ubp3p , have a central role in rescuing misfolded but partly active proteins from proteasomal degradation [32] . Fifth , there is a strong evolutionary trade-off between survival under starvation and adaptive mechanisms underlying tolerance to mistranslation . The evolved lines showed fitness defects and impaired capacity to degrade mature ribosomes upon nutrient limitation . Moreover , as a response to the energy demands of accelerated protein turnover , the evolved lines exhibited increased glucose uptake partly achieved by selective duplication of hexose transporter genes . To summarize , accelerated protein turnover is an effective first line of defense against protein mistranslation , but due to its exceptionally high energetic demand , this strategy is only feasible in nutrient-rich environments . This leads to the prediction that translation fidelity should vary across environments: microbes living in nutrient-poor conditions are expected to show much less tolerance to mistranslation . Due to the shortage of comparative data on translational fidelity across related microbial species , testing this prediction is not yet possible . Robustness to mistranslation in our study was achieved by disposing of proteins prone to aggregation rather than by improving the efficiency of protein folding of individual proteins . An unexpected aspect of our work is that there were no mutations found in known chaperone-encoding genes . Moreover , functional enrichment analysis ( Table 2 ) revealed that the accumulated mutations had no major effect on the expression level of chaperone-encoding genes . Probably , such mutations would be of little or no benefit , as a large fraction of mistranslated proteins in our study is damaged beyond repair . Indeed , CUG-encoded residues in yeast proteins are mostly buried and located in functionally conserved positions [17] . Moreover , given the large differences in Ser and Leu in chemical properties ( including hydrophobicity ) , this amino acid change is expected to cause abrupt changes in protein structure . Our work has several other implications for future studies . It supports the idea that partial or complete chromosome duplications fuel rapid evolutionary adaptation [42–44] . By altering the dosages of numerous , functionally related proteins simultaneously , these genetic changes introduce large phenotypic leaps that enable adaptation even in relatively small populations [44] . However , a prior study suggested that due to their deleterious side effects , chromosomal duplications are only transient solutions that are later augmented or replaced by more refined mechanisms through individual point mutations [47] . Protein turnover decreases with age in unicellular and multicellular organisms alike , resulting in an increase in the amount of intracellular damaged proteins [53] . Decreased protein turnover rate with age is the result of cumulative damage to the various components of the protein synthesis and degradation machinery [53] . Our work indicates that translational fidelity and the ubiquitin-proteasome system are functionally linked to each other and they could have a strong influence on cellular longevity . Our work focused on the detrimental effects of mistranslation . However , as mistranslation triggers an unfolded protein stress response , it also facilitates survival under acute stress [16 , 54] . It would be important to see whether evolution of tolerance to mistranslation interferes with such adaptation . There might be a strong negative trade-off between cellular robustness to mistranslation and mistranslation-mediated preadaptation to stressful conditions . This possibility can readily be tested in the laboratory . Finally , our work also sheds light on the problem of genetic code evolution [55] . Natural alterations to the standard genetic code have been discovered in many species . Interestingly , Candida albicans maintained variable serine and leucine incorporation levels at CUG sites [56] . The “ambiguous intermediate” theory states that tRNA mutations expand the decoding capacity of tRNA , leading to a transient state of ambiguous decoding of a single codon by both its cognate tRNA and the mutant tRNA . Proponents of the theory argued that such codon ambiguity is an important first step for gradual codon reassignments [57] . However , this initial step of ambiguous decoding is expected to lead to a serious decline in fitness due to the synthesis of non-functional or toxic proteins . Our work suggests that this problem is not fatal for the ambiguous intermediate theory . Selection pressure against genotypes with ambiguous decoding may only be temporary , as organisms can readily evolve tolerance to mistranslation . S . cerevisiae strains used in this study were self-diploids based on BY4742 background ( MATα; his3Δ1; leu2Δ0; lys2Δ0; ura3Δ0 ) . CUG ambiguous cells were obtained by transformation with the single-copy LEU2 plasmid containing the C . albicans G33-tRNACAGSer gene ( pUKC715 ) [18] . Control cells were transformed with the single-copy vector pRS315 ( vector alone ) . Yeast transformations were carried out using the lithium-acetate method [58] . Transformants were selected on leucine dropout synthetic complete medium ( SC-leucine: 5g/L ammonium sulfate , 1 . 7 g/L Yeast Nitrogen Base , supplemented with 2% glucose and with amino-acid mix , without leucine ) . Laboratory evolution experiment was conducted using leucine dropout synthetic complete liquid medium ( SC-leucine ) . Starting from a single clone carrying tRNACAGSer , 11 populations were inoculated into 100 μl liquid medium in a 96-well format plate ( Greiner ) and were incubated at 30°C in a shaking incubator . Using the same protocol , we also established 11 control wild-type evolving lineages carrying the corresponding empty vector . One percent of the stationary phase cultures were transferred to fresh medium every second day using handheld pintools ( VP407 , V&P Scientific , Inc . ) . The laboratory evolution was carried out for 72 d . We used established protocols specifically designed to measure fitness in yeast populations [21 , 22] . The fitness of the ancestor and the evolved lines were measured in 10–20 replicates each . The growth curves were monitored over a 48 h incubation period in a Biotek Powerwave XS2 automated plate reader in 384 density plates filled with SC-leucine liquid medium . During the kinetic run , the optical density of each well was recorded at 600 nm ( OD600 ) every 4 . 5 min . Between the optical readings , the cultures were incubated at 30°C , with alternating shaking speed ( 1 , 000–1 , 200 rpm ) . Growth rate was used as proxy for fitness and was estimated as previously [59] . To test the sensitivity to protein synthesis inhibitors , the culture medium was supplemented with subinhibitory concentration of cycloheximide ( 0 . 06 μg/ml , Biochemica ) . Fitness was estimated as above . To test the sensitivity under nutrient limitations , culture medium with limited carbon source ( SC-leucine supplemented with 1% glucose ) and culture medium with limited amino acid source ( SC-leucine supplemented with 0 . 25% amino acid dropout mix ) was used . Fitness was estimated as above . To test whether the detected fitness improvements in the evolved lines were not due to loss of function mutations in the mistranslating tRNACAGSer plasmid , which could potentially arise during the course of laboratory evolution , the tRNACAGSer plasmid was swapped with the original tRNACAGSer plasmid in the evolved lines . To investigate the effects of genomic mutations in the absence of high mistranslation rate , the tRNACAGSer carrying plasmid was eliminated , and the empty vector was introduced into the evolved lines . For each strain , we confirmed tRNACAGSer loss and subsequent plasmid gain using the appropriate plasmid marker . The quantification of translation rate was done by using protein pulse labeling method with [14C ( U ) ]-L-Amino Acid Mixture [25 , 26] . Amino acid incorporation was performed for the wild type and all evolved lines without the plasmid containing the tRNACAGSer in 15 replicates each . Briefly , 2 × 107 cells were collected , re-suspended into 2 ml of pre-warmed minimal medium and the suspension was incubated 20 min at 30°C with agitation . 20 μl of cold [14C ( U ) ]- L-Amino Acid Mixture was added ( Perkin Elmer , 0 . 1 mCi/ml ) and the mixture was incubated 10 min at 30°C with agitation . Amino acid incorporation was stopped by the addition of 60 μl of cycloheximide ( 20 mg/ml ) and ice incubation . Cells were washed once with cold water and frozen at -80°C . Protein was then extracted by re-suspending cell pellets in 300 μl Lysis buffer ( 50 mM potassium phosphate buffer pH 7 , 1 mM EDTA , 5% ( V/V% ) glycerol , 1 mM phenylmethylsulfonyl fluoride , complete mini protease inhibitor cocktail ( Roche ) and 100 μl of glass beads . Cells were disrupted using a Precellys ( Bertin Technologies , Montigny-le-Bretonneux , France ) disrupter ( five cycles of 10 s at 5 , 000 rpm and 1 min on ice between cycles ) and centrifuged at 5 , 000 g for 10 min . A total of 50 μl of supernatant was applied on 1 cm2 square paper microfiber filter ( GF/C , Whatman , Maidstone , United Kingdom ) . Amino acid incorporation was measured using a scintillation counter ( Beckman ) and protein extracts were quantified using the bicinchoninic acid ( BCA ) protein quantification Kit ( Pierce . Rockford , IL , United States ) . [14C ( U ) ]-L-Amino acid incorporation was normalized against the total amount of protein for each . In order to check for a possible reduction in mistranslation rate , the misreading activity of the tRNACAGSer in the ancestor and evolved lines was tested using the β-galactosidase method [10] . Nine replicates of ancestor and evolved lines were used per assay . The E . coli LacZ gene contains 54 CTG codons and misincorporation of Ser at these Leu codons generates a combinatorial array of mutant β-galactosidase molecules ( statistical β-gal ) whose altered stability can be quantified using thermal denaturation assays . Yeast cells containing the empty vector and the tRNACAGSer plasmid , respectively , were co-transformed with the pUKC815 plasmid , which contains the promoter of yeast phosphoglycerate kinase ( PGK1 [S000000605] ) gene , the N-terminal 33 amino acids of PKG1 gene fused in frame to the E . coli lacZ gene , encoding β-galactosidase [20] . Yeast cells expressing both plasmids were selected in leucine and uracil dropout SC medium . Approximately 2 . 5 × 106 ancestor and evolved cells from the exponential phase were harvested by centrifugation , respectively . Cells were washed and resuspended in 800 μl of Z-buffer ( 60 mM Na2HPO4 , 40 mM NaH2PO4·2H2O , 10 mM KCl , 1 mM MgSO4·7H2O , 50 mM 2-mercaptoethanol , pH 7 . 0 ) , 20 μl of 0 . 1% SDS , and 50 μl of chloroform . Cell suspensions were mixed for 30 s and incubated in triplicate at 47°C in a water bath for 10 min . This β-galactosidase unfolding step was followed by a refolding step , which was carried out by incubating samples on ice for 30 min . Residual β-galactosidase activity was then quantified at 37°C . For this , the assay tubes ( 200 μl ) were incubated for 5 min at 37°C and then 4 mg/mL of the ο-nitrophenyl-β-D galactopyranoside ( ONPG , Sigma-Aldrich ) substrate were added to each tube . After 5 min , the reactions were stopped by the addition of 400 μl of 1M Na2CO3 . β-galactosidase activity was determined by monitoring ο-nitrophenol synthesis at 420 nm and normalized for the total β-galactosidase produced by each line . To control for potential differences in total β-galactosidase expression across strains , we quantified β–galactosidase expression level using western blot . Total protein fractions were analyzed under reducing conditions using 10% SDS-PAGE and blotted onto nitrocellulose membranes ( Hybond ECL , Amersham ) according to standard procedures . β-galactosidase was detected using anti-β-gal rabbit IgG primary antibody ( Invitrogen ) at 1:5 , 000 dilution . Bound antibody was visualized by incubating membranes with a IRDye680 goat anti-rabbit secondary antibody ( Li-cor Biosciences , Lincoln , NE , US ) at 1:10 , 000 dilution . Detection was carried out using an Odyssey Infrared Imaging system ( Li-cor Biosciences ) . The amount of β-galactosidase was normalized to the amount of ADH1 ( S000005446 ) present in the total protein fraction . An established method was applied to estimate cellular propensity for protein aggregation [27] . It is based on the aggregation of a fluorescently tagged human protein ( von-Hippel-Lindau [VHL-mCherry] ) . This human protein is prone to misfolding in the absence of its cofactor elongin BC , a protein absent from the S . cerevisiae genome . However , the active quality-control machinery of S . cerevisiae can prevent the aggregation of this fusion protein . It results disperse cytosolic localization of the mCherry . Upon accumulation of erroneous proteins , the control machinery becomes overloaded , and the human protein becomes aggregated , while the fluorescent tag remains functional . In this case , the red fluorescence will appear as an aggregated focus point . The expression of the VHL-mCherry was driven by a galactose inducible promoter , from a plasmid ( pGAL-VHL-mCherry [CHFP] ) . Six replicates of ancestor and wild-type strains and 3 replicates of evolved lines were grown on leucine and uracil dropout SC medium , containing 2% raffinose as carbon source . Saturated cultures were diluted into the same selection medium , which contains 2% raffinose as carbon source and 2% galactose , to induce the expression of the fusion protein . Micrographs were taken from live cells using PerkinElmer Operetta High Content Imaging System . The ratio of cells with aggregated foci was calculated by dividing the number of cells with aggregated foci with the total number of cells expressing VHL-mCherry . The level of the mistranslating tRNA ( expressed from single-copy plasmids ) was monitored by northern blot analysis . C . albicans SN148 ancestor empty vector cells were used as positive control . Fractionation of tRNAs was carried out on 12%–15% polyacrylamide ( 40% Acril:Bis ) gels containing 8 M urea ( 0 . 8 mm thick , 30 cm long ) . In each gel slot , 50 μg of total RNA sample was loaded and gels were electrophoresed at 500 V for 16 h . Fractioned tRNAs were localized by UV shadowing , the portion of the gel containing tRNAs was cut and transferred onto a nitrocellulose membrane ( Hybond N , Amersham ) using a Semy-dry Trans Blot ( Bio-Rad ) . For hybridization , probes were prepared using 10 pmol of dephosphorylated oligonucleotide and 4 μl of ɣ-32P-ATP ( 5000Ci/mmol ) ( Perkin Elmer ) in 1x T4 kinase buffer , 10 mM spermidine and 16 units of T4 kinase ( Takara ) . Labelling reactions were incubated at 37°C for 1 h and then probes were extracted using phenol:chloroform:isoamyl alcohol ( PCIA ) . The hybridization protocol was performed as described by Jacques Heitzler and colleagues [60] . Membranes were pre-hybridized for 30 min at 55°C in a hybridization solution [5x Denhardt’s solution ( 1% Ficol , 1% Polyvinylpyrrolidone and 1% Bovine serum albumin ) , 6x SSPE ( 3 M NaCl , 173 mM NaH2PO4 , 25 mM EDTA ) and 0 . 05% SDS] . Membrane hybridizations were performed overnight in the above buffer using probes GCGACACGAGCAGGGTTC for detection of tRNASerCAG and GCGGAAGCCGGGAATCGAAC for detection of the control tRNAGlyCCC . Membranes were then washed four times ( 3 min each time ) in 2x SSPE , 0 . 5% SDS at 53°C and exposed overnight with intensifying screens and developed using a Molecular Imager FX ( Bio-Rad ) . In ancestor and evolved lines , proteasome activity was quantified as described previously [19] in three replicates each . From the middle of the exponential phase , 2 × 108 cells were collected , washed and frozen at -80°C . Cell pellets were resuspended in 350 μl of lysis buffer ( 10 mM Hepes , 10 mM KCl , 1 . 5 mM MgCl ) , two-thirds volume of glass beads and were disrupted using a Precellys disrupter ( three cycles of 10 s at 5 , 000 rpm followed by 2 min on ice between cycles ) . Pellets were centrifuged for 5 min at 3 , 000 g followed by 10 min at 15 , 000 g . Protein extracts were quantified using the BCA protein quantification kit ( Pierce ) . 100 μg of protein extracts were resuspended in assay buffer ( 10 mM Tris pH 8 , 20 mM KCl , 5 mM MgCl ) to a final volume of 100 μl and were incubated at 37°C for 15 min . The proteasome substrate N-SLLVY-MCA ( Sigma ) was added to a final concentration of 50 μM , and cells were incubated at 37°C during 60 min with agitation . Activity was measured using a Perkin Elmer Luminescence Spectrometer ( LS 50B ) at 365 nm ( excitation ) and 435 nm ( emission ) . Glucose uptake of the cells were measured as described previously [35] . Briefly , 3–3 replicates of wild type and evolved lines ( in absence of the mistranslating tRNACAGSer ) were inoculated into 20 ml of leucine dropout SC liquid medium to an initial optical density of 0 . 01 . The cultures were incubated at 30°C , shaking at 200 rpm . Optical densities were recorded every 3 h . Glucose content of the media was measured using Glucose assay HK kit ( Sigma ) in parallel with optical density measurements . To assess glucose uptake rate , the measured glucose content in the medium was normalized to the cell number ( calculated from the OD ) of the given line . Ancestor and evolved strains expressing the tRNACAGSer were transformed with plasmid pRS316-RPL25-GFP [38] and were grown in minimal medium without leucine and uracil to an OD600 = 0 . 2–0 . 8 . Nine to 11 replicates were used for each strain . Cells were pelleted at 3 , 000 g for 5 min . The supernatant was removed and after a washing step , the cells were incubated for 24–48 h in starvation medium ( 0 . 17% yeast nitrogen base without amino acids and without ammonium sulfate , 2% glucose ) . Cells were then poured onto a microscope slide previously coated with a bed of 1% of agarose . The localization of GFP-tagged proteins was scored after 24 h using a Zeiss MC80 Axioplan 2 light microscope , equipped for epifluorescence microscopy with the filter set HE38 . Microscope fields were randomly chosen and at least 100 cells were analyzed per sample . Photographs were taken using an AxioCamHRc camera and the number of cells with vacuolar fluorescence was counted and normalized for the total number of cells [61] . To measure chronological lifespan , a previously described method was used [38] . Briefly , the ancestor and evolved lines ( in absence of tRNACAGSer ) were inoculated into 10 ml leucine dropout SC medium , three replicates each . The medium was supplemented with a 2-fold excess of the histidine , methionine and uracil to avoid possible artifacts due to auxotrophic deficiencies of the strains . Initial densities of the cultures were set to OD600 = 0 . 05 . After 3 d of growth , when the cultures reached saturation , aliquots from the culture were collected and were plated onto YPD plates in serial dilutions . The plates were incubated in 30°C for 2–3 d , and viability was assessed by counting colony forming units . The viability at this time point was considered as initial viability ( 100% ) . Cells from saturated cultures were washed and resuspended in sterile water , to remove excess nutrients , originating from autolysed cells . The cultures were kept on incubating at 30°C with shaking , and in every 3 d , the same volume of aliquots were collected and subjected to viability assessment as mentioned above . Total RNA was extracted using the hot-phenol method [62] with few modifications . Briefly , 50 ml of exponentially growing cells were harvested and frozen overnight at −80°C . Cell pellets were resuspended in 0 . 5 ml of lysis buffer ( 10 mM Tris pH 7 . 5 , 10 mM EDTA , 0 . 5% SDS ) and 0 . 5 ml of acid phenol chloroform ( 5:1 pH 4 . 7 , Sigma ) . The samples were vigorously mixed and heat incubated at 65°C for 1 h . The aqueous phase was separated from the phenolic phase by centrifugation at 4°C . The aqueous phase was re-extracted with the same volume of chloroform:isoamyl alcohol ( 24:1 , Fluka ) . RNA was precipitated overnight at −80°C with ethanol 100% and 3 M sodium acetate pH 5 . 2 . RNA was pelleted by centrifugation and resuspended in sterile MilliQ water . Total RNA samples were treated with DNaseI ( Amersham Biosciences ) according to the commercial enzyme protocol and quantification and quality control was performed using the Agilent 2100 Bioanalyzer system . Gene expression profiling was performed using the Agilent protocol for One-Color MicroarrayBased Gene Expression Analysis Quick Amp Labeling v5 . 7 ( Agilent Technologies ) . Briefly , cDNA was synthesized from 600 ng of total RNA using Agilent T7 Promoter Primer and T7 RNA Polymerase Blend and labeled with Cyanine 3-CTP . Labeled cDNA was purified with RNeasy mini spin columns ( QIAGEN ) to remove residual Cyanine 3-CTP . Dye incorporation and quantification was monitored using the Nanodrop 1000 Spectrophotometer . To prepare hybridization , 1 . 65 μg of Cy3-labeled cRNA were mixed with the fragmentation mix ( Blocking Agent and Fragmentation Buffer ) and incubated for 30 min at 60°C . Finally , GEx Hybridization Buffer HI-RPM was added and the preparation was assembled in the custom made Agilent arrays ( yeast G4813A ) . Slides were prepared using Agilent gasket slides according to the manufacturer instructions . Each hybridization was carried out for 17 h at 65°C , in an Agilent hybridization oven . After washing and drying , the microarrays were scanned using the Agilent G2565AA microarray scanner ( Agilent ) . Probes signal values were extracted from microarray scan data using Agilent Feature Extraction Software ( Agilent ) . The microarray raw data was submitted to the GEO database and has been given the following accession number: GSE65718 . Data were normalized using median centering of signal distribution with Biometric Research Branch BRB-Array tools v3 . 4 . 0 software . Microarray data analysis was carried out with MEV software ( TM4 Microarray Software Suite ) [63] . Student’s t test was applied to identify genes that showed statistically significant ( p ˂ 0 . 01 ) differences in expression between control ( ancestor with empty vector ) and evolved lineages . Genomic DNA extraction was carried out using the Genomic-tip 100/G kit ( Qiagen ) according to the manufacturer’s protocol . Quantification and quality assessment were performed using the Picogreen fluorescence based quantification assay . For Illumina sequencing , genomic DNA was prepared and sequenced using the manufacturer-supplied protocols and reagents , as follows . One library per sample was constructed using Illumina DNA Sample Prep standard protocol and with an insert size of 400–500 bp . Briefly , 5 μg of high molecular weight genomic DNA ( gDNA ) was fragmented by Covaris sonication device . Following sonication , DNA fragments were end-repaired and A-tailed . Adapters were then ligated via a 3′ thymine overhang . Finally , ligated fragments were amplified by PCR . The library was applied to an Illumina Flowcell for cluster generation . Sequencing was performed on a Genome Analyzer IIx instrument using ~150 bp paired-end reads . Raw sequence data , 146 bp paired end reads with expected insert size of 400–500 bp , from each sample were trimmed by removing consecutive bases on both 5′- and 3′ flanks with base quality less than 20 . Trimmed reads that did not pass filtering criteria for ambiguity ( N content < 5% ) , complexity ( score ≥ 10 ) , length ( 50 bases or longer ) , and average base quality ≥ 20 were removed using Bamtools [64] . Remaining reads were mapped to the reference genome of Saccharomyces cerevisiae S288C , obtained from the Saccharomyces Genome Database [65] using BWA [66] . Processing and filtering of mapped reads were done using Samtools [67] . After removal of duplicates , read pairs aligning to opposite strands , or those where predicted insert size did not match actual size , were removed . Additionally , read pairs were removed where one or both reads had low mapping quality ( MQ < 20 ) or had less than 95% sequence identity to the reference . Mapped reads were analyzed using Samtools to produce read pileups [67] , detect single nucleotide variants and call genotypes . Small insertions and deletions ( indels ) were not called . Bases with low base quality or with read depth less than three or higher than twice the sample average coverage were called as unknown genotype . Sequencing data were archived in the European Nucleotide Archive under accession number PRJEB8951 . Funspec ( acronym for "Functional Specification" ) , a web-based cluster interpreter for yeast [23] was used for functional enrichment analysis . The gene set of interest was uploaded to the web interface of Funspec and clustering was done using automated algorithms , based on various knowledge resources . Intersections of the input list were sought with any given functional category . A GO category was termed as enriched significantly , if the genes annotated to a particular GO term are significantly overrepresented ( p < 0 . 05 ) in the given gene set , using the whole genome as the background set of genes . The accession numbers for genes mentioned in this paper are from the Saccharomyces Genome Database ( http://www . yeastgenome . org ) .
Although fidelity of information transfer has a substantial impact on cellular survival , many steps in protein production are strikingly error-prone . Such errors during protein synthesis can have a substantial influence on viability and the onset of genetic diseases . These considerations raise the question as to how organisms can tolerate errors during protein synthesis . In this paper , for the first time , we study organisms’ capacity to evolve robustness against mistranslation and explore the underlying cellular mechanisms . A mutant yeast strain was engineered to translate a codon ambiguously ( mistranslation ) . This thereby overloads the protein quality-control pathways and disrupts cellular protein homeostasis . This strain was used to study the capacity of the yeast genome to compensate for the deleterious effects of protein mistranslation . We found that mistranslation led to rapid evolution of genomic rearrangements , including chromosomal duplications and deletions . By altering the dosages of numerous , functionally related proteins simultaneously , these genetic changes introduce large phenotypic leaps that enable adaptation to mistranslation . Robustness against mistranslation during laboratory evolution was achieved through acceleration of protein turnover—a process that was determined by the combined rates of protein synthesis and ubiquitin-proteasome system-mediated degradation . However , as both translation and active degradation of proteins are exceptionally energy-consuming cellular processes , accelerated proteome turnover has substantial energy costs .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[]
2015
Evolution of Robustness to Protein Mistranslation by Accelerated Protein Turnover
Arthropod-borne flaviviruses cause life-threatening diseases associated with endothelial hyperpermeability and vascular leak . We recently found that vascular leak can be triggered by dengue virus ( DENV ) non-structural protein 1 ( NS1 ) via the disruption of the endothelial glycocalyx-like layer ( EGL ) . However , the molecular determinants of NS1 required to trigger EGL disruption and the cellular pathway ( s ) involved remain unknown . Here we report that mutation of a single glycosylated residue of NS1 ( N207Q ) abolishes the ability of NS1 to trigger EGL disruption and induce endothelial hyperpermeability . Intriguingly , while this mutant bound to the surface of endothelial cells comparably to wild-type NS1 , it was no longer internalized , suggesting that NS1 binding and internalization are distinct steps . Using endocytic pathway inhibitors and gene-specific siRNAs , we determined that NS1 was endocytosed into endothelial cells in a dynamin- and clathrin-dependent manner , which was required to trigger endothelial dysfunction in vitro and vascular leak in vivo . Finally , we found that the N207 glycosylation site is highly conserved among flaviviruses and is also essential for West Nile and Zika virus NS1 to trigger endothelial hyperpermeability via clathrin-mediated endocytosis . These data provide critical mechanistic insight into flavivirus NS1-induced pathogenesis , presenting novel therapeutic and vaccine targets for flaviviral diseases . Dengue virus ( DENV ) is a mosquito-borne flavivirus , and infection with any of its four serotypes ( DENV1-4 ) can result in inapparent infection , classic dengue fever , or dengue hemorrhagic fever/dengue shock syndrome–severe manifestations characterized by vascular leak that can lead to shock and death [1] . West Nile virus ( WNV ) is a related flavivirus that causes encephalitis [2] , and Zika virus ( ZIKV ) , which recently emerged and generated large epidemics across the Americas , can cause Guillain-Barré syndrome in adults and microcephaly and other congenital birth defects in babies born to women infected during pregnancy [3–7] . The 10 . 7-kb RNA genome of flaviviruses encodes three structural and seven non-structural ( NS ) proteins , including NS1 . We recently described a novel cell-intrinsic role for DENV NS1 in increasing permeability of human endothelial cell monolayers in vitro and systemic vascular leak in vivo via disruption of components of the endothelial glycocalyx-like layer ( EGL ) [8–10] . This EGL pathway is distinct from the cytokine-mediated pathway involving activation of peripheral blood mononuclear cells ( PBMCs ) previously described [10 , 11] . The NS1-mediated degradation of the EGL involves activation of the protease cathepsin L , heparanase , and sialidases , which in turn disrupt the EGL . However , the upstream pathway initiated by NS1 , resulting in cathepsin L activation , EGL degradation , and therefore vascular leak , as well as the molecular determinants of NS1 required for inducing endothelial hyperpermeability , have not yet been identified . All flavivirus NS1 genes are highly homologous and encode a 352-amino acid polypeptide with slight variation in glycosylation status that results in a molecular weight of 46–55 kDa [12] . High-resolution structures of the C-terminal half [13] and full-length [14] DENV NS1 , as well as the closely related WNV NS1 , revealed dimers containing distinct domains for membrane association and interactions with the immune system and provided structural insights into flavivirus NS1 assembly and antibody recognition [14] . In flavivirus-infected cells , NS1 dimers are found intracellularly and also on the cell surface , while secreted forms are comprised of atypical barrel-shaped hexamers containing lipid cargo [15] . DENV NS1 contains two conserved N-linked glycans ( N-glycans ) at Asn-130 ( N130 ) and Asn-207 ( N207 ) , and NS1 proteins derived from different cells/species exhibit distinct types of N-glycosylation [16] . Previous studies investigating the importance of the N-glycans on NS1 found that DENV2 deglycosylated at either site exhibited significant attenuation of neurovirulence in mice compared to the wild-type ( WT ) virus [17] . However , the mechanism of this attenuated virulence has never been elucidated , due in part to the fact that NS1 N-glycosylation mutant viruses are unstable , with significant defects in viral replication and proliferation that prevent further conclusive mechanistic study [17] . This has been a major roadblock in efforts to study the dual role of NS1 in viral replication and pathogenesis . In mammalian cell-derived DENV NS1 , complex glycans are attached at Asn-130 and high-mannose glycans are attached at Asn-207 [18]; these N-glycans are involved in NS1 dimer stability and secretion [18 , 19] . In this study , to investigate the role of these N-glycosylation sites directly on NS1-induced vascular leak independently from their roles in viral replication , we generated recombinant NS1 in a mammalian expression system , which incorporates human glycosylation patterns . We sought to investigate whether the glycosylation status of NS1 may contribute to its ability to induce endothelial hyperpermeability . Using both transendothelial electrical resistance ( TEER ) and an immunofluorescence assay ( IFA ) with confocal microscopy , we show that the NS1-N207Q mutant does not induce endothelial barrier dysfunction or degradation of the EGL of human pulmonary microvascular endothelial cells ( HPMEC ) . Further , we demonstrate that there are no differences in cell binding per se between DENV WT NS1 and the NS1-N207Q mutant , but the latter is not efficiently internalized by endothelial cells , indicating that internalization of NS1 is required for endothelial cell-intrinsic pathogenesis . Furthermore , we show that endothelial cells internalize NS1 via clathrin-mediated endocytosis and that its inhibition by either small molecule inhibitors or siRNA knock-down prevents NS1-induced hyperpermeability and EGL degradation in vitro and vascular leak in vivo . Finally , we identify N207 as a conserved residue of NS1 among multiple flaviviruses ( such as DENV , ZIKV , and WNV ) that is critical for NS1 internalization and endothelial hyperpermeability in biologically relevant human cells . Taken together , our results indicate that the N207 glycosylation site is required for NS1 internalization by endothelial cells and that clathrin-mediated endocytosis is a key step in the pathway leading to flavivirus NS1-induced endothelial hyperpermeability and barrier dysfunction . To examine the role of NS1 glycosylation status on endothelial barrier dysfunction , we generated several NS1 constructs targeting glycosylation sites in DENV2 NS1 ( henceforth referred to as DENV NS1 ) , including WT , the single N-glycosylation mutants N130Q and N207Q , and the double mutant N130Q+N207Q ( Fig 1A and S1 Table ) . DENV NS1-WT and DENV NS1-N207Q were successfully secreted and purified as oligomers ( MW: >250 kDa ) ( Fig 1B and 1C and S1A Fig ) ; the DENV NS1-N207Q mutant ( monomer ~43 kDa ) was slightly smaller than the NS1-WT monomer ( ~48 kDa ) due to loss of the N207 glycosylation site ( Fig 1B ) , consistent with previous publications [19] . In contrast , the N130Q mutant was not efficiently secreted , nor was the double mutant N130Q+N207Q ( S1B Fig ) ; thus , these two constructs were not further characterized . 6xHis-tagged recombinant DENV NS1 proteins were purified using nickel-nitrilotriacetic acid resin agarose beads and subjected to dialysis; the resulting purified DENV NS1 proteins ( WT and N207Q ) were visualized by silver stain and found to exhibit purity similar to commercially available NS1 ( >95% ) , as previously described [8] ( Fig 1D ) . To confirm the successful removal of an N-glycan at position 207 , we digested WT NS1 ( both commercial and in-house produced ) and the NS1-N207Q mutant with endoglycosidase H ( Endo H; removes high mannose glycans ) or PNGase F ( removes complex glycans and high mannose glycans ) , and measured their digestion sensitivity through observation of a gel shift by Western blot . The commercial and in-house produced WT NS1 proteins were sensitive to digestion by both Endo H and PNGase F , suggesting they possessed both the high mannose glycan at position 207 as well as the complex glycan at position 130 . In contrast , the NS1-N207Q mutant was sensitive to digestion with PNGase F but not Endo H , suggesting that it retained the complex glycan at position 130 but no longer possessed the high mannose N-glycan at position 207 , confirming its removal by the N207Q mutation ( S1C Fig ) . We next compared the size and stability of the NS1-WT and NS1-N207Q proteins . When subjected to size-exclusion chromatography , both our in-house produced NS1-WT and NS1-N207Q eluted at similar fractions , suggesting they are of comparable size and conformation ( S2A and S2B Fig ) . We then utilized an NS1-specific monoclonal antibody ( 9NS1 ) previously reported to detect NS1 in a native but not in a denatured conformation [20] , to further confirm that NS1-WT and NS1-N207Q were in a comparable conformation . Indeed , we found that 9NS1 recognized WT NS1 ( both commercial and in-house produced ) comparably to NS1-N207Q in native conditions but did not detect any of the three under denaturing conditions . This confirms both that 9NS1 functions as a conformational antibody and that the conformation of NS1-WT and NS1-N207Q is comparable ( S3A Fig ) . As expected , several additional NS1-specific monoclonal antibodies that are not conformation-dependent , as well as an antibody targeting the 6xHis-tag , recognized all NS1 proteins comparably in native and denaturing conditions ( S3A Fig ) . Next , to confirm that our purified and dialyzed NS1 proteins were stable at 37°C over the time-frame relevant for our in vitro cell culture-based assays , we diluted the NS1 proteins in our standard cell culture medium and incubated these proteins at 37°C for varying times . We found that both NS1-WT and NS1-N207Q are stable under these conditions both at time-points relevant for our in vitro assays ( 7 hours ) and even at longer time-points ( 48 hours ) ( S3B–S3D Fig ) . As expected , in these experiments , the addition of SDS or proteinase K degraded NS1 ( S3B–S3D Fig ) . Taken together , these data indicate that our purified NS1 proteins ( both NS1-WT and NS1-N207Q ) are comparable in size , conformation , and stability . To investigate the functional role of NS1 glycosylation on NS1-mediated pathogenesis , we first studied the effects of the proteins on endothelial barrier function by measuring the TEER of HPMEC , since a major site of vascular leak in severe dengue occurs in the lungs [8] . Our DENV NS1-WT triggered a similar decrease in TEER values as commercial DENV2 NS1 ( DENV NS1+ ) ( Fig 1E ) , indicating disruption of endothelial barrier integrity . In contrast , the relative TEER of the DENV NS1-N207Q mutant was similar to that of mock-treated cells , demonstrating that mutation of the N207 glycosylation site in DENV NS1 completely prevents induction of endothelial hyperpermeability in HPMEC ( Fig 1E and S4A Fig ) . Similar results were observed in human brain microvascular endothelial cells ( HBMEC ) ( Fig 1F and S4B Fig ) ; these cells model barrier function of the blood-brain barrier , which is relevant due to cases of dengue encephalitis and reports of DENV detected in the brain in autopsies of dengue shock cases [21 , 22] . Together , these data suggest that the N207 glycosylation site is a critical molecular determinant of DENV NS1-mediated pathogenesis . We previously demonstrated that DENV NS1 binds to endothelial cells and activates cathepsin L , which then activates heparanase , leading to cleavage of heparan sulfate from proteoglycans on the cell surface [9] . NS1 also activates sialidases , which cleave sialic acid on the cell surface [9] . Together , these events contribute to NS1-induced degradation of the EGL , leading to endothelial barrier dysfunction [9 , 10] . To mechanistically determine why the DENV NS1-N207Q mutant no longer triggers hyperpermeability of HPMEC , we first asked whether the DENV NS1-N207Q mutant was still able to bind to the surface of endothelial cells . We found that both the DENV NS1-N207Q mutant and WT DENV NS1 ( both commercial and in-house-produced ) bound to the surface of HPMEC as visualized by IFA ( Fig 2A and 2B and S5A and S5B Fig ) . However , the DENV NS1-N207Q mutant did not trigger EGL disruption , as determined by disruption of sialic acid , activation of cathepsin L , and cleavage of heparan sulfate , visualized by IFA at 1 and 6 hours post-treatment ( hpt ) with NS1 ( Fig 2A–2E and S5A and S5C Fig ) . These results suggest that the N-glycosylation of DENV NS1 at residue 207 plays a role in the activation of the endothelial cell-intrinsic mechanisms leading to disruption of the EGL . Though the DENV NS1-N207Q mutant was still able to bind to the surface of endothelial cells without leading to disruption of the EGL , it surprisingly appeared to bind to the surface of HPMEC at substantially higher levels than WT DENV NS1 ( both commercial and in-house-produced ) after 6 hours at 37°C ( Fig 2A and 2B ) . To further investigate this finding , we evaluated NS1 levels on the surface of HPMEC at 4°C , a temperature at which internalization via endocytosis should be greatly reduced [23 , 24] . This examines whether the apparent enhanced binding of DENV NS1-N207Q at 37°C was due to an intrinsically enhanced ability of the DENV NS1-N207Q mutant to bind to endothelial cells or through some other mechanism , such as deficient endocytosis . We found that the DENV NS1-N207Q mutant bound to HPMEC at similar levels as both commercial and in-house-produced WT DENV NS1 when incubated at 4°C for 1 hour , suggesting the DENV NS1-N207Q does not intrinsically bind more to HPMEC compared to WT DENV NS1 ( Fig 3A , top row , and Fig 3B , left panel ) . Instead , when NS1 proteins were incubated with HPMEC at 37°C ( allowing for protein internalization ) for 1 hour , the DENV NS1-N207Q mutant again bound at higher levels compared to WT NS1 , suggesting that it remained on the cell surface , while both commercial and in-house-produced WT NS1 proteins were internalized ( Fig 3A , bottom row , and Fig 3B , right panel ) . NS1 is known to bind to heparan sulfate moieties on the surface of endothelial cells [25] . To determine whether the NS1-N207Q mutant still required heparan sulfate moieties to bind to the surface of endothelial cells , we pretreated cells with recombinant heparanase to selectively remove heparan sulfate from the cell surface , and then compared the ability of NS1-WT and the NS1-N207Q mutant to bind to these cells . As expected , the in-house produced NS1-WT exhibited a significant decrease in binding to heparanase-treated cells compared to untreated cells . Similarly , the NS1-N207Q mutant also exhibited a significant decrease in binding to heparanase-treated cells compared to the untreated cells , suggesting that the NS1-N207Q mutant still requires heparan sulfate to bind to the endothelial cell surface ( S6A and S6B Fig ) . As expected , levels of heparan sulfate were dramatically decreased in endothelial cells pretreated with heparanase compared to untreated control cells ( S6C Fig ) . Our previous observations , and the observations of others , support the concept of NS1 internalization by endothelial cells [9 , 26] . To further support the notion that the DENV NS1-N207Q mutant has a defect in internalization , we incubated WT DENV NS1 and DENV NS1-N207Q with HPMEC at 37°C and compared levels of NS1 after stripping surface-bound NS1 from HPMEC; the remaining NS1 is presumed to have been internalized . We found that WT DENV NS1 ( both commercial and in-house-produced ) appeared at higher levels than the DENV NS1-N207Q mutant in the HPMEC cell lysate at 1 hpt following trypsin-mediated removal of surface-bound NS1 , as measured by Western blot , supporting our hypothesis that the DENV NS1-N207Q mutant has an internalization defect ( Fig 3C and 3D ) . Non-trypsin-treated cells served as a control for the initial levels of NS1 added to cells ( S7A Fig ) . To visualize intracellular WT NS1 , we incubated NS1 with HPMEC at 37°C for 1 hour and examined permeabilized cells by IFA for colocalization of WT DENV NS1-WT with Rab5 , a regulatory GTPase associated with early endosomes [27] . Both WT NS1 and NS1-N207Q proteins were detected; however , only WT DENV NS1 , but not DENV NS1-N207Q , colocalized with Rab5 ( Fig 3E and 3F and S7B Fig ) . This pattern can also be visualized in an animated single 2D projection of NS1 internalization in HPMEC created using multiple confocal Z-stack images ( S1 and S2 Movies ) . After binding to the cell surface , the WT NS1 protein ( in red , indicated by white arrowheads ) moves inside the cell , where it encounters Rab5 ( in green ) , which results in a coalescence signal ( in yellow ) , suggesting their spatial colocalization in the cell . Interestingly , as the Z stack images move into a deeper Z axial space , this colocalizing signal disappears , and only the red signal for NS1 persists , suggesting its mobilization into a different intracellular compartment ( S1 Movie ) . This process does not occur with the NS1-N207Q mutant protein , where most of the NS1 ( in red ) seems to remain on the cell surface , as little to no signal appears to colocalize with Rab5 , suggesting a lack of internalized protein ( S2 Movie ) . To further demonstrate the localization of NS1 in the early endosome , we show that NS1 also colocalizes with overexpressed Rab5-GFP as well as an additional early endosome marker ( EEA1 ) ( S7C and S7D Fig ) Collectively , these data demonstrate that while both WT DENV NS1 and DENV NS1-N207Q bind comparably to endothelial cells , only WT DENV NS1 is internalized , suggesting that endothelial cell binding and internalization may be distinct stages of NS1-induced cell-intrinsic endothelial hyperpermeability . Based on these results , we hypothesized that DENV NS1 internalization by endothelial cells is a critical step in the activation of EGL disruption and endothelial hyperpermeability . Endocytosis is a major route for transporting molecules into cells to regulate many cellular signaling processes and is commonly facilitated by two main endocytic molecules: clathrin and caveolin [28] . To interrogate the mechanism by which DENV NS1 is internalized by human endothelial cells , we incubated DENV NS1 with HPMEC and examined localization of internalized NS1 with clathrin and caveolin . We detected colocalization of WT DENV NS1 ( both commercial and in-house-produced ) with clathrin ( Fig 4A and 4C ) but not with caveolin in both HPMEC ( Fig 4B and 4D ) and HBMEC ( S8A–S8D Fig ) , suggesting that NS1 is internalized into endothelial cells in a clathrin-dependent manner . Interestingly , although DENV NS1-N207Q colocalized with clathrin comparably to DENV WT NS1 in HPMEC ( Fig 4A and 4C ) , only WT DENV NS1 colocalized with the early endosome marker Rab5 ( Fig 3E–3F ) . These data suggest that while binding of the DENV NS1-N207Q mutant to endothelial cells is sufficient to recruit clathrin to the NS1 binding site , it is not sufficient to promote internalization [29 , 30] . To determine whether internalization of DENV NS1 is dependent on clathrin , we utilized a common chemical inhibitor of clathrin-mediated endocytosis . Pitstop 2 is a selective inhibitor of clathrin-mediated endocytosis that acts by blocking ligand access to the clathrin N-terminal domain [31] . Pretreating HPMEC with Pitstop 2 abrogated NS1 colocalization with Rab5 , suggesting inhibition of internalization ( Fig 4E and 4F ) . These data suggest that DENV NS1 is internalized in a clathrin-dependent manner . Cathepsin L and downstream cellular enzymes are activated in endothelial cells treated with NS1 [9]; as such , we hypothesized that internalized NS1 would colocalize with cathepsin L . Since the peak of NS1-induced cathepsin L activation occurs as early as 30 minutes post-treatment [9] , we conducted an internalization time-course experiment to monitor the lifespan of NS1 within cells to determine whether internalization kinetics correlate with cathepsin L activation . WT NS1 ( DENV NS1+ ) was incubated with HPMEC for 45 minutes at 4°C to allow NS1 to bind to cells . Afterwards , cells were washed with cold PBS to remove any unbound NS1 and were then allowed to incubate at 37°C for varying times . Upon temperature shift to 37°C , NS1 accumulated rapidly in HPMEC , reaching a maximum at 15 minutes post-temperature shift , and then rapidly decreased until dropping to near undetectable levels at 3 hours post-temperature shift ( S9A–S9C Fig ) . These kinetics correlate well with the previously observed kinetics of NS1-triggered activation of cathepsin L [9] . Intriguingly , NS1 colocalized with the early endosome marker Rab5 , the lysosomal marker LAMP1 , and cathepsin L , suggesting that the presence of NS1 in these intracellular compartments may activate cathepsin L ( Fig 3E , S9D and S9E Fig ) . These data indicate that NS1 rapidly and transiently internalizes into cells and localizes to compartments where cathepsin L is present , suggesting that internalization of NS1 may be required for activation of cathepsin L , and thus downstream disruption of the EGL . To test whether DENV NS1 internalization is required for EGL disruption , we evaluated the effect of Pitstop 2 treatment on cathepsin L activation and heparan sulfate degradation ( Fig 5A and 5B ) . Pretreatment of HPMEC with Pitstop 2 prevented the activation of cathepsin L and subsequent EGL degradation by WT DENV NS1 , as demonstrated by decreased cathepsin L activity and reduced heparan sulfate on the surface of HPMEC , respectively ( Fig 5A and 5B ) . We next asked whether DENV NS1 internalization is required to trigger endothelial hyperpermeability in vitro . Dynasore , a common inhibitor of endocytosis , antagonizes the small GTPase dynamin , which is required for clathrin-mediated endocytosis [32 , 33] . Pretreatment of HPMEC with either Dynasore or Pitstop 2 abolished DENV NS1-triggered endothelial hyperpermeability as measured by TEER ( Fig 5C and S10 Fig ) . To confirm that these inhibitors blocked clathrin-mediated endocytosis in our system , we measured the capacity of both Pitstop 2 and Dynasore to block internalization of transferrin , a protein well appreciated to be taken up via clathrin-mediated endocytosis [34] . As expected , Pitstop2 and Dynasore , but not the DMSO vehicle control , blocked internalization of transferrin into HPMEC ( S11A and S11B Fig ) . Because chemical inhibitors can have off-target effects , we genetically knocked down clathrin , dynamin , and caveolin to confirm the role of clathrin-mediated endocytosis in DENV NS1 EGL disruption . Utilization of small interfering RNAs ( siRNA ) targeting either the clathrin heavy chain , dynamin I and II ( I/II ) , or caveolin-1 mRNA enabled us to knock down expression of the proteins encoded by these transcripts . Knocking down clathrin and dynamin , but not caveolin , abolished NS1-mediated EGL disruption , as determined by heparan sulfate surface expression ( Fig 6A and 6B , S12 Fig ) . To confirm that our siRNA knock-down experiments successfully blocked clathrin-mediated and caveolin-mediated endocytosis in our system , we conducted internalization assays in these cells using transferrin as well as human serum albumin , which is known to be internalized via caveolin-mediated endocytosis [35] . As expected , transferrin internalization was significantly inhibited by knocking down clathrin and dynamin expression , while internalization of human serum albumin was only inhibited by knocking down caveolin expression in cells ( S11C , S11D and S12 Figs ) , confirming the specificity of our siRNA system . These results indicate that NS1 is internalized via clathrin-mediated endocytosis , which is critical for NS1-induced hyperpermeability and EGL degradation . To determine whether clathrin-mediated endocytosis is required for NS1 to trigger vascular leak in vivo , we utilized a mouse model of localized leak in the dermis [10] . In this assay , NS1 or control proteins are injected intradermally ( ID ) into the depilated backs of mice; up to four different conditions can be tested per mouse . Fluorescently labeled soluble dextran ( Alexa Fluor 680-dextran ) is then administered intravenously , and accumulation of Alexa Fluor 680-dextran in the backs of mice ( where ID injections of NS1 or controls occurred ) is monitored at 2 hours post-dextran injection . We injected mice with NS1 or PBS mixed with a vehicle control or with a cocktail of Pitstop 2 and Dynasore to antagonize clathrin-mediated endocytosis locally ( 4 conditions total ) . Injection with NS1 mixed with a vehicle control showed significantly greater fluorescent signal ( i . e . , leakage ) compared to the PBS conditions or NS1 mixed with the clathrin-mediated endocytosis inhibitor cocktail , suggesting that this cocktail blocked NS1-mediated leak in the dermis . These data suggest that NS1-triggered leakage in vivo , in the mouse dermis , requires clathrin-mediated endocytosis ( Fig 7A and 7B ) . Given the importance of the N207 glycosylation site for DENV NS1-mediated endothelial hyperpermeability , we investigated its prevalence among related viruses and found a high level of conservation of this site in multiple flaviviruses ( Fig 8A ) . We then expanded our study to examine the role of the N207 glycosylation site of NS1 on the endothelial permeability of HBMEC in two closely related neurotropic flaviviruses–WNV and ZIKV . We first generated the WNV and ZIKV NS1-WT and NS1-N207Q mutant recombinant proteins ( S1 Table ) . All proteins were secreted equivalently and appeared stable throughout the purification/dialysis procedure; further , all proteins were determined to be pure by silver staining and in a hexameric/oligomeric structure by native gels ( S13A–S13F Fig ) . We found that WT WNV and ZIKV NS1 ( commercial or produced in-house ) triggered endothelial hyperpermeability as measured by TEER , indicating that WT NS1 proteins are capable of inducing endothelial hyperpermeability in HBMEC [36 , 37] ( Fig 8B and 8C and S14A and 14B Fig ) . In contrast to the WT NS1 proteins , both the WNV NS1-N207Q and ZIKV NS1-N207Q mutants failed to trigger endothelial hyperpermeability of HBMEC , suggesting that the N207 glycosylation site is a critical determinant for endothelial hyperpermeability induced by NS1 from multiple flaviviruses ( Fig 8B and 8C and S14A and S14B Fig ) . To determine whether the N207 residue is also required for internalization of NS1 from related flavivirus such as WNV and ZIKV , as it is for DENV NS1 , we conducted a binding/internalization assay of NS1-WT and NS1-N207Q from DENV , WNV , and ZIKV using HPMEC and HBMEC . In this internalization assay , we measured colocalization of NS1 with Rab5 and interpreted Rab5-positive NS1 puncta to represent binding and internalization while Rab5-negative puncta represent binding only . As observed previously [37] , DENV NS1-WT bound well to both HPMEC and HBMEC while WNV and ZIKV NS1 only efficiently bound to HBMEC , and this was also the case for the NS1-N207Q mutants ( S15A–S15C Fig ) . Further , within HBMEC , to which ZIKV and WNV NS1 efficiently bound , the NS1-WT proteins but not the NS1-N207Q mutants of DENV , WNV , and ZIKV were efficiently internalized ( S15A–S15C Fig ) . These data suggest that like DENV NS1 , the N207-glycosylation site is critical for internalization of WNV and ZIKV NS1 proteins . We next investigated whether clathrin-mediated endocytosis was a common pathway for flavivirus NS1-mediated pathogenesis . We found that Pitstop 2 blocked heparan sulfate shedding on HBMEC induced by WT WNV NS1 and WT ZIKV NS1 ( Fig 8D–8E ) . Taken together , these results indicate that internalization of NS1 via a clathrin-mediated endocytosis pathway is a conserved mechanism for NS1-induced endothelial hyperpermeability for multiple flaviviruses . In this study , we demonstrate that the N207 glycosylation site of DENV NS1 is essential for inducing EGL degradation and hyperpermeability of human endothelial cells . We found that the DENV NS1-N207Q mutant binds to cells at similar levels as WT DENV NS1 but is retained on the cell surface , in contrast to WT DENV NS1 , which is rapidly internalized . Additionally , internalization of WT DENV NS1 was dynamin- and clathrin-dependent , but independent of caveolin . Further , clathrin-mediated endocytosis of WT DENV , WNV , and ZIKV NS1 was required for endothelial hyperpermeability and EGL disruption . Taken together , these results indicate that the N207 glycosylation site is required for endothelial cell internalization via clathrin-mediated endocytosis and endosomal trafficking of NS1 , which is necessary for the activation of enzymes such as cathepsin L that lead to EGL degradation and increased endothelial permeability ( Fig 9A and 9B ) . Using a transgenic expression system , it was previously shown that the N130 glycan of DENV NS1 was required for stabilization of the secreted NS1 hexamer whereas the N207 glycan facilitated secretion and extracellular protein stability [19] . In our system , the N-glycan at position 130 was required for protein secretion while the N-glycan at position 207 was dispensable for protein secretion and stability , in contrast to Somnuke et al . These differences could be explained by the methods used to produce and purify NS1 proteins between the two studies ( e . g . , a different signal sequence mediating secretion and a different strategy and methodology utilized for protein purification ) . In our current mammalian protein expression system utilizing 293F suspension cells , the DENV NS1-N207Q mutant was efficiently secreted; however , neither the DENV NS1-N130Q mutant nor the double mutant N130Q+N207Q was efficiently secreted , suggesting that the N130 N-glycan site is essential for NS1 secretion . These observations are supported by a previous study that demonstrated that ablation of the homologous glycosylation sites in yellow fever virus NS1 resulted in viral mutants with impaired NS1 secretion [12] . Due to this technical limitation , the N130Q and N130Q+N207Q constructs were not further characterized in this study . However , based on the significant homology at NS1-N130 across all flaviviruses , it is possible that NS1-N130 may also carry out important pathological functions , which urge further investigation . We have previously shown that NS1 activates key endothelial cell-intrinsic pathways , including the sialidase and cathepsin L/heparanase pathways , and that these contribute to NS1-induced hyperpermeability of human endothelial cells [9]; however , how these enzymes become activated remain unknown [9] . Our results here provide evidence that internalization of NS1 via clathrin-mediated endocytosis is required for the activation of these pathogenic processes and demonstrate that NS1 localization in endosomes and lysosomes , where cathepsin L is localized , correlates with cathepsin L activation . Further studies defining the biochemical interactions required for this activation are needed and are an area of active investigation . Our data suggest that NS1 binding and internalization are distinct steps in triggering endothelial hyperpermeability . Because both WT DENV NS1 and DENV NS1-N207Q are able to bind to HPMEC , but only WT DENV NS1 is internalized , we hypothesize that NS1 initially binds to heparan sulfate moieties on endothelial cells [25] , independently of the N207 N-glycan . DENV NS1 may then interact with an additional binding partner ( s ) in a manner dependent on the N207 N-glycan , which leads to internalization , endosomal trafficking , and activation of cathepsin L and/or sialidases . Why is the N207 N-glycosylation site critical for internalization of DENV NS1 ? First , the N207 N-glycan may be required for engaging a specific receptor on endothelial cells , either directly through a physical interaction or indirectly through a conformational change . Based on the available crystal structure of NS1 , amino acid 207 is not predicted to be surface-exposed [14] . This observation favors a model where a conformational change of NS1 allows this N-glycan to be exposed and mediate an interaction required for internalization or where the N-glycan itself mediates a conformational change of NS1 that allows another domain to mediate an interaction required for internalization . Second , glycosylation patterns have been shown to be important for lateral diffusion of proteins on the plasma membrane [38 , 39] . Thus , an alternate possibility is that the N207 N-glycan is required for NS1 to migrate to the proper microdomain to interact with functional components on the lipid membrane required for internalization . Furthermore , NS1 is secreted from DENV-infected cells predominantly as a soluble hexameric barrel-shaped , high-density lipoprotein with a hydrophobic core containing lipid cargos such as triglycerides , cholesteryl esters , and phospholipids [15 , 40 , 41] . The functional importance , if any , of these lipid cargos is yet to be determined . A third possibility for the role of the N207 N-glycan in NS1 internalization is that this sugar moiety may be important for acquisition of specific lipid cargos that play a role during NS1 binding to endothelial cells or even internalization; thus , further investigation of the lipid contents of the mutant NS1 is needed to uncover a potential role of the N207 N-glycan in lipid cargo acquisition as well as a possible role of these lipid cargos in NS1-mediated pathogenesis [42] . Intriguingly , the NS1-N207Q mutant binds to cells and even recruits clathrin to the cell surface , but fails to be internalized and traffic into early endosomes . Because of these observations , an important remaining question is at what stage of the internalization process is the NS1-N207Q mutant blocked ? Cell binding of NS1-N207Q alone may be sufficient to trigger the proper signaling pathway ( s ) to recruit clathrin to the cell surface but may not be sufficient to trigger the next stage of internalization , perhaps mediated through the engagement of the N207 N-glycan to a receptor . Previous reports of stalled clathrin-coated pits suggest that recruitment of additional factors , such as the clathrin-interacting protein epsin , are required for full maturation and budding of the endosome into the cell [30 , 43] . Thus , another potential explanation may be that the N207 N-glycan is required to recruit additional factors to the clathrin-coated pits that promote internalization . An understanding of the role that other cellular factors , like epsins , play in the maturation of NS1-containing clathrin-coated pits and subsequent internalization is needed to fully elucidate the internalization defect of NS1-N207Q . Clathrin-mediated endocytosis of NS1 may have implications beyond NS1-triggered endothelial hyperpermeability . It is known that DENV invades cells via endocytosis [44] , and pharmacologically blocking clathrin-mediated endocytosis retards DENV replication [45] . Further , a recent study demonstrated that DENV NS1 interacts with DENV structural proteins , which modulates infectious viral particle production [46] . In light of our finding that NS1 is internalized via clathrin-mediated endocytosis , NS1 could potentially play a role in mediating clathrin-mediated endocytosis of flavivirus virions . Further studies are needed to elucidate , and potentially differentiate , clathrin-mediated endocytosis of virions via the E protein and/or NS1 . Regarding overall pathogenic mechanisms , our investigation specifically focuses on a cell-intrinsic pathway by which NS1 mediates the disruption of the EGL , but flavivirus-induced vascular dysfunction is a complex and multifactorial process . It can be mediated through the NS1 endothelial cell-intrinsic pathway [9 , 10 , 36 , 37] , through production of proinflammatory cytokines by immune cells induced by either the virus or NS1 [11 , 47] , through NS1-activated platelets [48] , or even potentially through direct viral infection leading to apoptosis of endothelial cells [49 , 50] . Though endothelial cells can be infected in vitro by DENV , findings from autopsy studies suggest that this does not occur in vivo [22 , 51 , 52] . ZIKV infects human endothelial cells in vitro and fetal endothelial cells in a mouse model [4 , 53 , 54] , suggesting another pathway by which ZIKV may trigger endothelial dysfunction . Dissecting the relative contributions of each of these pathways in flavivirus-induced pathology in vivo is critical for designing appropriate therapeutics that could be used to treat flavivirus-induced vascular dysfunction . All flaviviruses must disseminate from the blood into specific tissues , where they replicate to high levels . They can do this by infecting migrating immune cells , direct infection of endothelial cells [55] , or through dysfunctional endothelial cell barriers [36 , 56] . Our data indicate that the cell-intrinsic endothelial dysfunction pathway is triggered not only by DENV NS1 , but also by neurotropic WNV and ZIKV NS1 in HBMEC . Our previous study demonstrates that NS1 from multiple flaviviruses ( including DENV , WNV , and ZIKV ) mediates endothelial dysfunction and vascular leak in a tissue-specific manner reflecting viral tropism [37] . We speculate that this conserved NS1-mediated cell-intrinsic endothelial dysfunction pathway mediates virus dissemination of different flaviviruses into an appropriate tissue , promoting viral replication . In summary , our current study not only identifies the mechanism of internalization of secreted NS1 by human endothelial cells via a dynamin- and clathrin-mediated , caveolin-independent endocytosis pathway , but also pinpoints a single amino acid , the N207 glycosylation site of NS1 , as a key determinant of NS1-mediated endothelial hyperpermeability . Mutation of N207 prevents NS1 internalization by human endothelial cells and inhibits disruption of the EGL and endothelial barrier function in vitro and in vivo . Further , our work identifies N207 of NS1 as a conserved functional residue among multiple flaviviruses ( ZIKV and WNV , in addition to DENV ) that is essential for flavivirus NS1-mediated endothelial dysfunction via clathrin-mediated endocytosis in human cells . This new molecular insight into the mechanism of flavivirus NS1-triggered endothelial hyperpermeability is crucial for understanding pan-flaviviral pathogenesis , as well as for developing antiviral therapies and NS1-based vaccine approaches . A panel of recombinant DENV2 ( strain 16681 ) , WNV ( NY99 ) , and ZIKV ( Nica1-16 ) NS1 proteins was generated , including wild-type ( WT ) and site-specific mutant constructs , and the primers utilized are listed in S1 Table . Amplicons consisting of the WT DENV or WNV NS1 gene preceded by the CD33 signal sequence and followed by a 6xHis-tag were a gift from M . S . Diamond ( Washington University in St . Louis ) , along with the pmab vector . The WT ZIKV NS1 gene fragment was designed according to the ZIKV Nica1-16 strain sequence [4] and synthesized by Bio Basic Inc . To introduce mutations , the Phusion High-Fidelity DNA Polymerase ( New England BioLabs; NEB ) was used to perform overlap-extension PCR according to the manufacturer’s instructions . XbaI and MluI-HF restriction endonucleases ( NEB ) were used to digest the PCR amplicons and the pmab vector . A Quick Ligation ( NEB ) was performed to ligate the inserts with the vector , and 5-alpha Competent E . coli cells ( NEB ) were used for transformation . Colonies were selected based on expected XbaI and MluI-HF double digestion profiles and correct Sanger sequencing results . FreeStyle 293F suspension cells ( Thermo Fisher Scientific ) derived from human embryonic kidney cells were cultured in FreeStyle 293 Expression medium with 1% penicillin/streptomycin at 37°C and 8% CO2 and were maintained at 0 . 15–1 . 2 x 106 cells/ml on a cell shaker at 135 rpm . The HPMEC-ST1 . 6r line was kindly donated by Dr . J . C . Kirkpatrick at Johannes Gutenberg University , Germany , and was grown using the EGM-2 bullet kit ( Clonetics , Lonza ) and maintained as previously described [8] . HBMEC were donated by Dr . Ana Rodriguez at New York University and maintained using endothelial cell medium with growth supplement ( ScienCell Research Labs ) . Transfections were performed using FreeStyle 293F suspension cells and FreeStyle MAX transfection reagent ( Thermo Fisher Scientific ) according to the manufacturer’s protocol , and supernatant containing NS1 were collected 48 hours post-transfection . To prevent protein degradation , a tablet of EDTA-free protease inhibitor cocktail ( Roche Life Sciences ) was added to every 150 ml of supernatant collected . Cell supernatants were stored at -80°C prior to purification . NS1-containing supernatants were then thawed and concentrated in an Amicon filter with a 100-KDa size cutoff ( UFC9100 , Millipore ) . The 6xHis-tagged recombinant DENV NS1 proteins were purified by batch method using nickel nitrilotriacetic acid ( Ni-NTA ) resin agarose beads ( Thermo Fisher Scientific ) . In brief , Ni-NTA resin agarose beads were first equilibrated in binding buffer ( 20 mM sodium phosphate , 500 mM sodium chloride , 20 mM imidazole , pH 7 . 4 ) , and then supernatants containing NS1 were diluted 1:2 in 2X binding buffer and added to the previously equilibrated Ni-NTA resin agarose beads . NS1-containing supernatants were allowed to bind to the Ni-NTA resin for 30 minutes at 4°C rocking end-over-end . Once binding was completed , the resin was washed 4X in wash buffer ( 20 mM sodium phosphate , 500 mM sodium chloride , 30 mM imidazole , pH 7 . 4 ) . Concentrated NS1 was eluted from the Ni-NTA resin using an imidazole-containing elution buffer ( 20 mM sodium phosphate , 500 mM sodium chloride , and 200 mM imidazole , pH 7 . 4 ) . This purified NS1 stock was then subjected to dialysis against 1X PBS for 48 hours at 4°C . Typical concentrations of NS1 stocks obtained ranged from 0 . 1–0 . 5 mg/ml . The Pierce BCA Protein Assay Kit ( Thermo Fisher Scientific ) was used to quantify the purified recombinant proteins . For the Endo H/PNGase F digestion Western blots , protein stability assays Western blots , NS1 internalization kinetic Western blots , siRNA knockdown Western blots , and ZIKV/WNV native-PAGE Western blots total cell lysates or recombinant proteins were collected in protein sample buffer ( 0 . 1 M Tris [pH 6 . 8] , 4% SDS , 4 mM EDTA , 286 mM 2-mercaptoethanol , 3 . 2 M glycerol , 0 . 05% bromophenol blue ) and then resolved by SDS-PAGE . The same protocol and buffer were used for native-PAGE but without SDS . The resolved proteins were then transferred onto nitrocellulose membranes and probed with primary antibodies diluted in TBS/0 . 1% Tween 20 ( TBST ) containing 5% skim milk , rocking overnight at 4°C . Secondary antibodies were diluted in TBST with 5% skim milk for 1-hour rocking at room temperature . After each antibody incubation , membranes were washed three times for 5 minutes each with TBST . Membranes were visualized using ECL reagents on a ChemiDoc system with Image Lab software from Bio-Rad . All antibodies were probed by the above method except rabbit anti-Caveolin1 and rabbit anti-DynaminI/II which were probed in TBST with 5% bovine serum albumin according to the manufacturer’s instructions . The following antibodies were used: mouse anti-6xHis ( MA1-21315 , Thermo Scientific ) , anti-NS1 monoclonal antibody ( 7E11 ) , rabbit anti-clathrin heavy chain ( ab172958 , Abcam ) , rabbit anti-Dynamin I/II ( 2342S , Cell Signaling ) , rabbit anti-caveolin 1 ( 3267 , Cell Signaling ) , mouse anti-βactin ( sc-47778 , Santa Cruz Biotechnology ) , mouse anti- αtubulin ( ab4074 , Abcam ) , goat anti-mouse HRP ( 405306 , Biolegend ) , and donkey anti-rabbit HRP ( 406401 , Biolegend ) . For the N-glycan digestion Western blots , 1 μg of denatured NS1s were digested with 750 units Endo H ( NEB ) or 500 units PNGase F ( NEB ) treatment at 37°C for 1 hour according to manufacturer’s protocol . For the NS1 stability assays 100 ng of protein were mixed with EGM-2 media ( Lonza ) and incubated in a tissue-culture incubator for the times indicated in the figures . For all other Western blots , SDS-PAGE gels were transferred semi-dry onto a polyvinylidene difluoride ( PVDF ) membrane for 7 minutes [57] at 25V . The membrane was blocked and incubated with primary anti-6xHis-tag antibody overnight . After washing , a goat anti-mouse IgG secondary antibody conjugated to Alexa Fluor 680 ( Thermo Fisher Scientific ) was added at a dilution of 1:5 , 000 for 1 hour , and blots were imaged using the LI-COR Odyssey imaging system . Anti-NS1 mAb 7E11 ( generously donated by R . Putnak at the Walter Reed Army Institute of Research ) was used for detection of DENV NS1 . 0 . 25 milligrams ( 500 ul of a 0 . 5 mg/ml stock ) of purified and dialyzed NS1-WT and NS1-N207Q were injected into a Superose 6 Increase 10/300 GL column ( GE Life Sciences ) previously equilibrated with 1X PBS . Validation of peaks , detected by UV light , were done by evaluating 2 . 5 ul of a given fraction by SDS-PAGE and probing with a NS1-speficic monoclonal antibody ( 7E11 ) . NS1 capture ELISA: MaxiSorp ELISA plates ( Thermo Scientific Nunc ) were coated with 50 μl of an anti-NS1 specific monoclonal antibody ( 7E11 ) ( 5 μg/ml ) and incubated overnight at room temperature . The next day the plate was blocked with 1% BSA in PBS for 1 hour then washed twice with 1X PBS . Then 50 ul of NS1 containing supernatant was added to wells and incubated for 1 hour at room temperature . Plates were then washed 3 times in PBS + 0 . 1% Tween20 and incubated for 1 hour at room temperature with another NS1-specific monoclonal antibody ( 2B7 ) . After an additional three washes in PBS + 0 . 1% Tween20 a peroxidase-labeled goat anti-mouse secondary antibody ( Jackson ImmunoResearch ) at 0 . 5 μg/ml in 1% BSA-PBS was added to the wells and incubated for an additional hour . After 3 additional washing steps , the assay was developed using an ABTS-HRP substrate kit ( KPL ) , according to the manufacturer’s specifications . NS1 direct ELISA: To examine the conformational status of the recombinant NS1 proteins used in this study , we performed a NS1 direct ELISA using a panel of recombinant NS1 proteins ( NS1+ , NS1-WT , and NS1-N207Q ) and anti-NS1 monoclonal antibodies ( mAb ) that interacts with the soluble NS1 protein in a conformational-dependent manner . Briefly , ELISA plates ( Nunc , Thermo Scientific ) were coated overnight at 4°C with a single concentration of NS1 ( 200 ng/mL ) diluted in sterile DPBS ( 1X , Gibco ) under two conditions: native ( no SDS ) or denaturing ( SDS 0 . 1% , boiling at 99°C for 5 minutes ) . The next day , plates were washed twice using wash buffer ( PBS + 0 . 05% Tween-20 ) and blocked using blocking buffer ( 3% bovine albumin , BSA in washing buffer ) for 45 minutes at room temperature ( RT ) . Then , plates were probed individually with anti-NS1 mAbs ( 0 . 25 μg/mL ) which bind NS1 proteins under denaturing conditions ( e . g . 7E11 and 2B7 ) or native conditions ( e . g . 9NS1 ) . All recombinant NS1 proteins in this study contain a 6xHis-tag . The anti-6xHis-tag mAb ( 0 . 25 μg/mL , Thermo Scientific ) was used as positive control for NS1 detection . A peroxidase conjugated goat-anti mouse IgG ( H+L ) antibody ( Jackson Immunoresearch Inc . , ) ( 1:5000 ) and TMB ( 3 , 3′ , 5 , 5′-Tetramethylbenzidine , Sigma ) were used to develop the assay . Plates were allowed to develop for 30 minutes at room temperature and were then stopped with a solution of sulfuric acid ( H2SO4 , 2N ) . Plates were read at a 450 nm wavelength using a microplate reader . To measure the functional effects of the recombinant NS1 proteins , HPMEC and HBMEC were cultured in duplicate wells ( 80 , 000 cells in 0 . 3 ml ) in the apical chamber of 24-well Transwell polycarbonate membrane inserts ( Transwell permeable support , 0 . 4 μm , 6 . 5 mm insert; Corning Inc . ) and incubated with 5 μg/ml NS1 protein ( WT or mutants , 1 . 5 μg total protein ) . TEER was measured as previously described [8 , 9] . Resistance values were measured in Ohms ( Ω ) at sequential 2-hour time-points following the addition of test proteins using an Epithelial Volt Ohm Meter ( EVOM ) with “chopstick” electrodes ( World Precision Instruments ) . Untreated endothelial cells grown on Transwell inserts were used as negative untreated controls , and inserts with medium alone were used for blank resistance measurements . Relative TEER represents a ratio of resistance values ( Ω ) as follows: ( Ω experimental condition-Ω medium alone ) / ( Ω non-treated endothelial cells-Ω medium alone ) . Confluent HPMEC monolayers grown on gelatin-coated coverslips ( 0 . 2% , Sigma ) were treated with 5 μg/ml of DENV2 NS1+ from the Native Antigen Company ( Oxfordshire , UK ) as a positive control , WT DENV NS1 produced in-house , or the DENV NS1-N207Q mutant and incubated for 1 hour at 37°C/4°C or for 6 hours at 37°C , depending on the experiment . To remove heparan sulfate from the cell surface , HPMEC were treated for 24 hours , prior to NS1 treatment with 0 . 5 units of recombinant heparanase ( H3917-50 UN , Sigma ) . An anti-6xHis-tag antibody conjugated to DyLight 647 ( Thermo Fisher Scientific ) was used to detect NS1 protein bound to the cell surface . Untreated cells were used as a negative control . Images were acquired using a Zeiss LSM 710 AxioObserver-34-channel spectral detector confocal microscope and processed using ImageJ software . For MFI quantification , the threshold for each individual channel ( RGB ) was adjusted and converted to grayscale . Then , mean grayscale values and integrated density from selected areas were obtained , along with adjacent background readings , and plotted as mean fluorescence intensity ( MFI ) . HPMEC were grown and imaged as described above . To assess the effect of flavivirus NS1 on the integrity of EGL components such as sialic acid and heparan sulfate , HPMEC monolayers were treated with DENV WT ( Native Antigen or produced in-house ) or NS1-N207Q mutant proteins ( 5 μg/ml ) and fixed with 4% paraformaldehyde ( PFA ) at different time-points ( 1 , 6 hpt ) . A primary antibody against heparan sulfate ( clone F58-10E4 , Amsbio ) was incubated overnight at 4°C , and detection was performed using a secondary goat anti-mouse IgM antibody conjugated to Alexa 488 ( Thermo Fisher Scientific ) . The cell surface expression of sialic acid was visualized using the lectin wheat germ agglutinin ( WGA ) conjugated to Alexa 647 ( Thermo Fisher Scientific ) [9] . The proteolytic activity of cathepsin L was evaluated using the Magic Red assay cathepsin L detection kit ( ImmunoChemistry Technologies ) [9] . Nuclei were stained with Hoechst ( blue ) . Untreated cells were used as a control for basal sialic acid and heparan sulfate expression and cathepsin L activity . Images were acquired using the Zen 2010 software ( Zeiss ) and analyzed with ImageJ software . For representative pictures , an area of ~1 . 8 μm2 ( 1 . 25x1 . 40 μm ) containing ~28–30 cells was used . MFI quantification was performed as described above . To evaluate the internalization of DENV NS1 proteins in HPMEC , we used Western blot and IFA . Briefly , confluent monolayers ( pre-chilled for 10 minutes at 4°C ) grown on culture plates and/or glass coverslips were exposed to 10 μg/ml of different NS1 proteins ( as indicated above ) and incubated for 45 minutes at 4°C to facilitate NS1 protein adsorption but not internalization . Cultures were then transferred to 37°C to allow protein internalization . For the internalization kinetic experiments , cells were washed twice with PBS , before transfer to 37°C , to remove any unbound NS1 . One hour later ( or at the indicated times for kinetic analysis ) , cell supernatants were removed , and cell monolayers were rinsed 3X with PBS and detached with trypsin-EDTA , which also removed any surface-bound , non-internalized NS1 proteins from the outer cell membrane . Cells were pelleted and lysed using SDS gel loading buffer , then the cell lysate was separated using an SDS-PAGE gradient gel ( 4–20% ) and visualized by Western blot with primary anti-NS1 mAbs ( 1G5 . 3 and 1H7 . 4 , obtained from Paul Young , University of Queensland , Australia ) [58] and secondary anti-mouse antibody conjugated to Alexa 750 ( ab175740 , Abcam ) . α-Tubulin was used as a protein loading control ( anti-α-tubulin ab4074 , Abcam ) , and Rab5 was used as an early endosome marker ( ab18211 , Abcam ) . Anti-rabbit IgG conjugated to Alexa 680 was used as secondary antibody ( ab175773 , Abcam ) . Images were acquired and processed using a LI-COR Odyssey system and ImageJ software , respectively . For the internalization Western blot , at the indicated time post temperature shift , cell monolayers were rinsed 3X with PBS and then 2X with pre-chilled acid wash buffer ( glycine 100 mM , and 150mM NaCl , pH 2 . 5 ) , which removed any surface-bound , non-internalized NS1 proteins from the outer cell membrane . Cells were then collected in western blot sample buffer . Additionally , internalized NS1 ( 10 μg/ml ) was visualized by IFA by co-staining NS1 and Rab5 . Briefly , after 45 minutes at 4°C ( to normalize protein adsorption ) , cultures were transferred to 37°C for 1 hour ( or the time indicated in the figure for the kinetic analysis ) to facilitate protein internalization . For the internalization kinetic experiments , cells were washed twice with PBS before transfer to 37°C , to remove any unbound NS1 . Before fixation of samples within the kinetic analysis , cell supernatants were removed , and cell monolayers were rinsed 3X with PBS and then 2X with pre-chilled acid wash buffer ( glycine 100 mM , and 150mM NaCl , pH 2 . 5 ) , removing surface-bound NS1 . After fixation with 4% PFA , samples were processed by indirect IFA and confocal microscopy imaging . For experiments using inhibitors , compounds were added to wells 30 minutes before the addition of DENV NS1 protein . Initially , a co-localization plugin analysis in ImageJ software was used to define the co-localizing spots between NS1 and Rab5 in each experimental condition . The amount of spatial overlap between the two signals ( NS1 “red” and Rab5 “green” ) was obtained using four different frames from the maximum projections of two RGB images based on the object-based approach ( JACoP ) and defined by the Manders’ Coefficient , as previously described [59] . As an additional analysis to quantify the amount of internalized NS1 protein in HBMEC and HPMEC monolayers by IFA , the number of NS1-positive ( NS1+ ) puncta per cell ( n = 200 cells ) and those co-stained with the early endosome marker Rab5 ( NS1+Rab5+ ) were manually counted using the multi-point tool in ImageJ . An animated single 2D projection of NS1 internalization in HPMEC was created by collapsing multiple image-sections ( maximum projections ) obtained from confocal Z-stack acquisition analyses . Briefly , HPMEC monolayers grown on glass coverslips were treated with DENV NS1 ( WT ) , and the glycosylation mutant ( DENV NS1-N207Q ) for 1 hour at 37°C . Cell monolayers were processed for immunofluorescence assay as described above co-staining for NS1 protein and the early endosome marker Rab5 . Z-stacks were acquired on a Zeiss LSM 710 inverted confocal microscope ( CRL Molecular Imaging Center , UC Berkeley ) using a Plan-Apochromat 40X dry objective . Projections of multiple Z-stack sections ( 20 μm range: 20 slices; 1 μm interval ) were performed and animated using Z stack reconstruction plugin in Image J , then saved as Audio Video Interleave ( AVI ) format . Scale bar , 1 μm . IFA was used to determine whether clathrin-mediated endocytosis or caveolin-mediated uptake contribute to internalization of NS1 by endothelial cells . HPMEC or HBMEC were grown and imaged as described above . Briefly , after 45 minutes at 4°C ( to normalize protein adsorption ) , plates were transferred to 37°C for 30 minutes to facilitate protein internalization . After fixation with 4% PFA , samples were processed by IFA and confocal microscopy imaging . A colocalization plugin analysis in ImageJ software was used to define the colocalizing spots between NS1 and clathrin or caveolin in each experimental condition . The amount of spatial overlap between the two signals ( NS1 , red , and Clathrin , Caveolin , Rab5 , green ) was obtained using four different frames from the maximum projections of two RGB images based on the object-based approach ( JACoP ) and defined by the Manders’ Coefficient as previously described [59] . Primary antibodies against clathrin heavy chain ( Clathrin Heavy Chain [D3C6] XP Rabbit mAb [4796S] , Cell Signaling ) , caveolin ( Caveolin-1 ( D46G3 ) XP Rabbit mAb #3267 , Cell Signaling and Anti-Caveolin-2 antibody EPR5471 [ab133484] , Abcam ) , Rab5 ( Anti-Rab5 antibody—Early Endosome Marker [ab18211] , Abcam ) , EEA1-early endosome marker ( AB2900 , Abcam ) , cathepsin L ( BMS166 , eBioscience ) , LAMP1 ( ab25630 , Abcam ) , mouse anti-6xHis ( MA1-21315 , Thermo Scientific ) , and rabbit anti-6xHis ( ab232492 , Abcam ) , together with the secondary antibodies goat anti-mouse IgG conjugated to Alexa 647 ( ab150115 , Abcam ) , goat anti-mouse IgG conjugated to Alexa 488 ( ab150117 , Abcam ) , donkey anti-rabbit IgG conjugated to Alexa 568 ( ab175470 , Abcam ) were used in IFA experiments . An anti-6xHis-tag mAb ( HIS . H8 ) conjugated to DyLight 647 ( Thermo Fisher Scientific ) was used for NS1 protein binding assays . Additionally , cells were infected with a baculovirus Rab5-GFP overexpression system ( CellLight Early Endosomes-GFP , BacMam 2 . 0 , C10586 , Thermo Fisher Scientific ) , according to the manufacturer’s instructions , to visualize colocalization of NS1 with Rab5 . Selective inhibitors of clathrin-mediated endocytosis ( Pitstop 2 , Abcam ) and dynamin ( Dynasore , Sigma ) were used in IFA experiments and/or TEER assays at concentrations that do not affect cell viability . Cell viability was determined using the Promega CellTox Green Cytotoxicity Assay following the manufacturer’s instructions . The internalization of human transferrin , an iron-binding protein well-known to be internalized via clathrin-mediated endocytosis , was used as a control to examine the effectiveness of clathrin-mediated endocytosis inhibitors . Briefly , 30 minutes post-treatment with Pitstop 2 ( 25 μM ) and Dynasore ( 50 μM ) at 37°C , transferrin conjugated to Alexa 568 ( 20 μg/ml ) ( T23365 , Thermo Fisher Scientific ) was added to HPMEC monolayers cultured on cover slips and incubated for 10 minutes at 37°C . Then , plates were immediately transferred to an ice-bed; the transferrin inoculum was removed and cell monolayers were washed twice with cold PBS , and then washed twice with pre-chilled acid wash buffer ( glycine 100 mM , and 150mM NaCl , pH 2 . 5 ) to remove cell-surface associated but non-internalized transferrin [33] . After this , coverslips were processed as described above for imaging after fixation with 4% PFA . For transient knock-down analysis of clathrin heavy chain ( HC ) , dynamin I/II , and caveolin-1 , the siRNA-mediated knock-down system from Santa Cruz Biotechnology was used . In brief , Control siRNA-A ( sc-37007 ) , Clathrin HC siRNA ( h ) ( sc-35067 ) , Dynamin I/II ( sc-43736 ) , and caveolin-1 siRNA ( h ) ( sc-29241 ) were resuspended in siRNA Dilution Buffer ( sc-29527 ) according to the manufacturer’s instructions . HPMEC were then transfected with siRNAs using siRNA Transfection Reagent ( sc-29528 ) in siRNA Transfection Medium ( sc-36868 ) according to the manufacturer’s instructions . Seventy-two hours post-transfection , cells were assayed for knock-down efficiency by Western blot analysis and used for further experiments . Primary antibodies used for Western blot were as follows: clathrin HC ( ab172958 , Abcam ) , Dynamin I/II ( 2342S , Cell Signaling ) , caveolin-1 ( D46G3 ) XP Rabbit mAb #3267 ( Cell Signaling ) , and α-tubulin ( ab4074 , Abcam ) . Additionally , HPMEC monolayers transiently transfected with distinct siRNAs were analyzed for internalization of fluorescently labeled-transferrin-A568 ( 20 μg/ml ) ( T23365 , Thermo Scientific ) and albumin-FITC ( 100 μg/ml ) ( GTX28030 , GeneTex ) after incubation at 37°C for 10 minutes and 1 . 5 hours , respectively . The amount of internalized proteins was evaluated by IFA analyses using the same acid wash buffer method described above . In vivo NS1-induced endothelial hyperpermeability was measured using a rodent model of localized vascular leak , as previously described [10] . Briefly , the dorsal area of 6-week old female WT C57BL/6 mice ( Jackson Labs ) was depilated 3–4 days prior to each experiment . Hair was initially trimmed using hair clippers and then removed using Nair ( Church & Dwight ) . Excess Nair was thoroughly wiped off using 70% ethanol and water . On the day of the experiment , DENV2 NS1 ( 15 μg ) was incubated for 15 minutes at room temperature in the presence of 0 . 25 mg/ml Pitstop 2 ( abcam ) and 0 . 5 mg/ml Dynasore ( abcam ) . NS1-inhibitor mixtures and controls were then were injected intradermally ( ID ) into distinct sites in the depilated dorsal skin of mice . Then , 150 μl of 10-kDa dextran conjugated with Alexa Fluor 680 ( 1 mg/ml; Sigma ) was delivered by retro-orbital ( RO ) injection . Two hours after injection , mice were euthanized using isoflurane , and the dorsal skin was removed and placed in Petri dishes . Tissues were then visualized and scanned using a fluorescent detection system ( LI-COR Odyssey CLx Imaging System ) at a wavelength of 700 nm , and leakage surrounding the injection sites was quantified using Image Studio software ( LI-COR Biosciences ) . Statistical analyses were performed using GraphPad Prism 6 software , and all graphs were generated using Prism 6 . For quantification of MFI in in Figs 2B–2E , 3B , 3D and 3F and S5B and S5C Fig , an ordinary one-way ANOVA using Dunnett’s multiple comparison test with multiple comparisons was used . For MFI quantification of IFA images in Figs 4F , 5B , 6B and 8E and S6B Fig , unpaired parametric t-tests were used to determine the significance of treatment with inhibitors or siRNA knockdown of various proteins . For dermal leak experiments in Fig 7 , a nonparametric Mann-Whitney U test was used to determine significance of inhibitor treatment . For MFI quantification of IFA images in S11B and S11D Fig , ordinary one-way ANOVA analyses using Dunnett’s multiple comparison test with multiple comparisons to either the DMSO group ( S11B Fig ) or the untreated group ( S11D Fig ) were used to determine the significance of inhibitor treatment or siRNA knockdown on internalization of transferrin or albumin by HPMEC .
Vascular leak is a hallmark of severe dengue disease . Recently , our group revealed a critical role for NS1 in induction of endothelial hyperpermeability and vascular leakage in an endothelial cell-intrinsic manner . However , the upstream pathway triggered by NS1 , as well as the molecular determinants of NS1 required for this phenomenon , remain obscure . Gaining insight into this endothelial cell-intrinsic pathway is critical for understanding dengue pathogenesis , developing novel antiviral therapies , and developing NS1-based vaccine approaches that pose a minimal risk of antibody-dependent enhancement . Our current study expands our knowledge of this novel pathway not only by identifying the requirement of internalization of secreted NS1 via clathrin-mediated endocytosis , but also by pinpointing the NS1 molecular determinant ( N207 ) required to trigger vascular leak . Further , our work identifies N207 as a residue conserved among multiple flaviviruses ( Zika virus and West Nile virus , in addition to DENV ) , which is critical for NS1-mediated vascular leak in biologically relevant human endothelial cells modeling interstitial compartments in the lung or the blood-brain barrier . Thus , our study identifies endocytosis and a single amino acid ( N207 ) of the NS1 viral toxin as critical for pan-flavivirus pathogenesis , representing a novel target for anti-flaviviral therapy and vaccine development .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "methods" ]
[ "cell", "binding", "cell", "physiology", "medicine", "and", "health", "sciences", "pathology", "and", "laboratory", "medicine", "chemical", "compounds", "gene", "regulation", "endothelial", "cells", "pathogens", "cell", "processes", "salts", "microbiology", "epithelial", "cells", "viruses", "rna", "viruses", "glycosylation", "small", "interfering", "rnas", "animal", "cells", "proteins", "medical", "microbiology", "sulfates", "microbial", "pathogens", "gene", "expression", "biological", "tissue", "chemistry", "recombinant", "proteins", "endocytosis", "biochemistry", "rna", "west", "nile", "virus", "cell", "biology", "secretory", "pathway", "anatomy", "flaviviruses", "post-translational", "modification", "viral", "pathogens", "nucleic", "acids", "genetics", "epithelium", "biology", "and", "life", "sciences", "cellular", "types", "physical", "sciences", "non-coding", "rna", "glycobiology", "organisms" ]
2019
Endocytosis of flavivirus NS1 is required for NS1-mediated endothelial hyperpermeability and is abolished by a single N-glycosylation site mutation
Blood flukes of the genus Schistosoma are platyhelminth parasites that infect 200 million people worldwide . Digestion of nutrients from the host bloodstream is essential for parasite development and reproduction . A network of proteolytic enzymes ( proteases ) facilitates hydrolysis of host hemoglobin and serum proteins . We identified a new cathepsin L termed SmCL3 using PCR strategies based on S . mansoni EST sequence data . An ortholog is present in Schistosoma japonicum . SmCL3 was heterologously expressed as an active enzyme in the yeast , Pichia pastoris . Recombinant SmCL3 has a broad pH activity range against peptidyl substrates and is inhibited by Clan CA protease inhibitors . Consistent with a function in degrading host proteins , SmCL3 hydrolyzes serum albumin and hemoglobin , is localized to the adult gastrodermis , and is expressed mainly in those life stages infecting the mammalian host . The predominant form of SmCL3 in the parasite exists as a zymogen , which is unusual for proteases . This zymogen includes an unusually long prodomain with alpha helical secondary structure motifs . The striking specificity of SmCL3 for amino acids with large aromatic side chains ( Trp and Tyr ) at the P2 substrate position , as determined with positional scanning-synthetic combinatorial library , is consistent with a molecular model that shows a large and deep S2 pocket . A sequence similarity network ( SSN ) view clusters SmCL3 and other cathepsins L in accordance with previous large-scale phylogenetic analyses that identify six super kingdoms . SmCL3 is a gut-associated cathepsin L that may contribute to the network of proteases involved in degrading host blood proteins as nutrients . Furthermore , this enzyme exhibits some unusual sequence and biophysical features that may result in additional functions . The visualization of network inter-relationships among cathepsins L suggests that these enzymes are suitable ‘marker sequences’ for inclusion in future phylogenetic analyses . Proteases ( proteolytic enzymes , peptidases ) provide essential functions in all life forms [1] . Proteases function as key elements of parasitism including hatching , excystment , tissue/cell invasion , nutrient acquisition and immune evasion [2] , [3] . For trematode parasites causing diseases of medical and veterinary importance , proteases operate at the host-parasite interface facilitating migration , digestion of host proteins and probably immune evasion [3] , [4] . Within the family Schistosomatidae , three major species infect more than 200 million people worldwide [5] . After penetration of human skin by aquatic larvae ( cercariae ) , immature parasites ( schistosomula ) migrate within the vascular system to the final predilection site where females produce eggs upon maturation . Parasite development and fecundity rely on nutrients ingested from the host bloodstream . A network of proteases with differing catalytic mechanisms “Clans” as described in the MEROPS database ( http://merops . sanger . ac . uk/ ) has been identified in the schistosome gut and facilitates digestion of proteins to absorbable peptides and amino acids [6]–[8] . For Schistosoma mansoni , the component proteases thus far characterized include Clan CA S . mansoni cathepsin B1 ( SmCB1 ) , SmCL1 ( SmCF ) and SmCL2 , SmCC , a Clan CD asparaginyl endopeptidase ( SmAE ) , a Clan AA aspartic protease SmCD and a Clan MF leucine metallo-aminopeptidase [7] , [9] . Proteolytic networks associated with host protein degradation and comprising the same protease clans have been described for other parasitic platyhelminths [4] and are conserved across phylogenetically diverse organisms such as Plasmodium [10] , nematodes [11] and arthropods [12] . Given their central importance in the biology of the parasite , gut proteases have been tested as vaccine candidates for disease prophylaxis [13] , [14] and are potential chemotherapeutic targets [15] , [16] . As immunodominant antigens , some schistosome gut proteases have been experimentally proven as serodiagnostic antigens [17] . In this study , we have identified and characterized a new cathepsin L in S . mansoni , SmCL3 . From the original expressed sequence tag ( EST ) [18] we have cloned and sequenced the full-length open reading frame ( ORF ) , and heterogeneously expressed the enzyme in the yeast , Pichia pastoris . The hydrolytic activity and specificity of the recombinant protease were characterized using active site-directed affinity probes , peptidyl substrates and a positional scanning-synthetic combinatorial library ( PS-SCL ) . Monospecific antibodies localized SmCL3 to the gut . Distinct from SmCL1 and SmCL2 , the N-terminus of the SmCL3 zymogen is extended by approximately 30 amino acids , and the enzyme exists primarily as a zymogen in the parasite rather than as a fully processed mature enzyme . Sequence similarity clustering and visualization using Cytoscape [19] places SmCL3 in the metazoan cathepsin L cluster along with SmCL2 and cathepsins L from the liver fluke , Fasciola spp . . This cluster is distinct from a second group of cathepsins F that includes SmCL1 and those from other trematode parasites such as Opisthorchis , Paragonimus and Clonorchis . S . mansoni ( a Puerto Rican isolate ) is maintained in the laboratory by cycling between the freshwater snail , Biomphalaria glabrata , and the golden hamster , Mesocricetus auratus . Hamsters are maintained in barrier facilities as approved by the Institutional Animal Care and Use Committee of the University of California San Francisco ( IACUC ) . All animal experiments were carried out in accordance with the same protocols approved by the IACUC . Infections with S . mansoni are initiated by subcutaneous injections of 500–1000 cercariae . At 6–7 weeks post-infection , hamsters are euthanized with intra-peritoneal injections of sodium pentobarbital ( 50 mg/kg ) , and adult worms harvested by reverse perfusion of the hepatic portal system [20] in RPMI 1640 medium ( Invitrogen ) . Complete Medium 169 containing 5% fetal calf serum and 1% ABAM ( Antibiotics/Antimycotics: Sigma-Aldrich ) , was used to maintain immature ( schistosomula ) and adult worms in vitro [21] . For preparation of schistosomula , cercariae were harvested from the infected snails by light induction for 1 h , and chilled on ice in a 50 ml falcon tube . The water was poured off and replaced with chilled incomplete Medium 169 ( without serum ) in preparation for shearing of tails , a method modified from Colley and Wikel [22] . Cercariae were then passed back and forth 15 times between two 10 ml syringes connected by a double-headed 22 gauge needle . Upon deposition into a 5 cm Petri dish , the lighter tails were separated from heads by swirling and aspiration with a Pasteur pipet . The nascent schistosomula were then collected and washed three times in Incomplete Medium 169 . After recovery from hamsters , adult worms were washed 5 times in incomplete Medium 169 . Both schistosomula and adults were maintained in complete Medium 169 under a 5% CO2 atmosphere at 37°C . Miracidia ( the stage infective to the snail ) were prepared from eggs trypsinized from infected liver tissue and hatched in freshwater . A partial sequence encoding the cathepsin L3 was obtained from the S . mansoni EST database [18] . Gene-specific primers were used to verify the cathepsin L3 gene sequence . Briefly , S . mansoni mRNA was isolated from adult worms using the FastTrack 2 . 0 isolation kit ( Invitrogen ) , and single strand cDNA was prepared using Superscript III Reverse Transcriptase ( Invitrogen ) with an oligo-dT18 primer . Purified cDNA was then used as template for PCR using Taq Platinum polymerase ( Invitrogen ) and gene-specific primers , SmCL3frd1 ( 5′-GCCTGGCTCTGTAAATGTTGAG -3′ ) and SmCL3rev1 ( 5′- CATATGGATAGGAAATCTCAGAATC -3′ ) . A 350 bp product was amplified and subsequently cloned into pCR 2 . 1-TOPO cloning vector ( Invitrogen ) for propagation in E . coli . Five positive clones were analyzed for sequence verification . Full-length cathepsin L gene was retrieved by rapid amplification of cDNA ends ( RACE ) using the GeneRacer Kit ( Invitrogen ) according to the manufacturer's instructions . Gene specific primers for 3′ RACE were SmCL3 3′RACE frd1 ( 5′- GTTGCGTGGATATAAAGTCACTAG -3′ ) and SmCL3 3′RACE frd2 ( 5′- GCTATCAGACATAAAGGGTCGAC -3′ ) . For 5′RACE the primers were SmCL3 5′RACE rev1 ( 5′- GTCGACCCTTTATGTCTGATAGC -3′ ) and SmCL3 5′RACE rev2 ( 5′- CTAGTGACTTTATATCCACGCAAC -3′ ) . Final amplicons were cloned into pCR 2 . 1-TOPO cloning vector and sequenced . To verify the entire ORF sequence , PCR incorporated Platinum Taq polymerase , cDNA from adult worms and primers directed to the 5′ and 3′ ends of the SmCL3 gene . The resulting amplicons were cloned into pCR 2 . 1-TOPO cloning vector and 10 randomly selected positive E . coli clones were sequenced . Total RNA was extracted from S . mansoni eggs , daughter sporocysts extracted from hepatopancreases of snails patent for infection , cercariae , newly transformed schistosomula ( incubated in vitro for 24 h ) , and adult worms using Trizol reagent according to the manufacturer's instructions ( Invitrogen ) . The precipitation step was omitted and RNA from the aqueous phase was purified using the RNA Isolation Kit ( Stratagene ) according to the manufacturer's instructions . The concentration of RNA was determined by absorbance at 260 nm using a ND-1000 Spectrophotometer ( NanoDrop ) . Single-stranded cDNA was synthesized from 1 µg of total RNA using SuperScript III reverse transcriptase ( Invitrogen ) and an oligo d ( T ) 18 reverse primer according to the manufacturer's protocol , and the resulting cDNA was purified . Quantitative PCR ( qPCR ) was carried out using the SYBR-green MasterMix Plus Kit ( Eurogentech ) with 1 µl of purified cDNA and each of 2 sets of forward and reverse primers ( 0 . 1 µl; 2 . 4 µM each; Table S1 ) that had been designed using the Primer 3 software ( http://frodo . wi . mit . edu/cgi-bin/primer3/primer3_www . cgi , [23] ) and designed to amplify 150–250 bp fragments . Triplicate reactions were carried out in a final volume of 25 µl in 96 well plates in a MX 3005P Real-Time PCR cycler ( Stratagene ) . The amplification profile consisted of an initial hot start ( 95°C for 10 min ) followed by 40 cycles comprising 95°C for 30 s , 55°C for 1 min and 72°C for 30 s . The ROX dye and S . mansoni cytochrome C oxidase I ( SmCyCOx ) ( GenBank AF216698 , [24] ) were always used as a reference dye and reference gene , respectively . Upon completion of the amplification , the dissociation curve was examined for potential primer dimerization . The cycle threshold ( CT ) values were averaged and the standard deviation was determined . The relative expression levels were calculated using the formula 2 − ( SmCyCOx CT – Gene of interest CT ) [25] . The primary amino acid sequence coding the SmCL3 gene was analyzed by SignalP ( http://www . cbs . dtu . dk/services/SignalP/; [26] ) to identify the predicted starting position of the proenzyme which was then amplified with Pfx DNA polymerase ( Invitrogen ) using the cloning primers SmCL3picZB frd , 5′-GATACTGCAGATTCTGGTTTCAGAAAGTGGTC-3′ ( Pst I restriction site underlined; note: the Kex 2 yeast protease processing site is placed upstream in the expression pPICZ αB vector ) and SmCL3picZB rev , 5′-TAAGCGGCCGCTCATACTAGAGGGTATGAAGCCGCACTGGCA-3′ ( Not I restriction site underlined , termination codon in italic ) . Alternatively , a histidine-tagged reverse cloning primer , SmCL3picZB revHis , 5′-TAAGCGGCCGCTCACATCATCATCATCATCATTACTAGAGGGTATGAAGCCGCACTGGCA-3′ ( Not I restriction site underlined , termination codon in italic , 6×His-tag in bold ) , was used to amplify a C-terminal histidine- tagged form ( SmCL3-his ) to facilitate subsequent purification and concentration of the SmCL3 expression product . The resulting PCR products were sub-cloned into the expression vector pPICZ αB ( Invitrogen ) , as previously described [27] and sequences verified . Transformation of P . pastoris and protein expression were carried out as described previously [27] , [28] . The induction yeast medium containing recombinant enzyme was filtered ( 0 . 45 µm ) , lyophilized and stored at −20°C until use . The powder was resuspended to 10% of the induction volume , and desalted using PD-10 columns ( GE-Healthcare ) and eluted in 50 mM sodium phosphate ( pH 6 . 0 ) for non his-tagged enzyme , or 50 mM sodium phosphate , pH 7 . 5 , 500 mM NaCl for SmCL3-his SmCL3-his was purified further on a HisTrap 5 ml column ( GE-Healthcare ) . The column was equilibrated with 50 mM sodium phosphate , 500 mM NaCl , pH 8 . 0 . Enzyme sample was loaded and the column washed with 25 ml of 50 mM sodium phosphate , 500 mM NaCl , 20 mM imidazole , pH 6 . 0 and eluted in same buffer containing 500 mM imidazole . Salt and imidazole were then removed by buffer exchange on an Amicon Ultra 10 kDa spinning column by 5 wash and centrifugation steps into 50 mM citrate , 100 mM sodium phosphate buffer , pH 6 . 0 . The presence of active recombinant enzyme was verified by protease activity assay ( see Protease activity assays ) and by SDS-PAGE gels with protein visualized either with iodinated clan CA affinity label 125I-DCG-04 ( see Active site labeling ) or with SafeStain protein dye ( Invitrogen ) . The cleavage sites used to generate the active recombinant enzyme were identified by N-terminal protein sequencing ( Protein and Nucleic Acid Facility , Stanford University ) . The recombinant enzyme was stored at −20°C . To determine glycosylation status , recombinant SmCL3 activity was inhibited for 30 min at RT with 10 µM K11777 and deglycosylated using endoglycosidase H ( Endo-H , Roche ) according to the manufacturer's protocol . Samples were then resolved by 15% SDS-PAGE . The specific irreversible affinity probe for Clan CA cysteine proteases , 125I-DCG-04 [29] was used to label the active site of recombinant SmCL3 at pH 6 . 0 , as previously described [30] . Prior to radiolabeling , control samples were incubated for 20 min in the presence of 10 µM of the Clan CA cysteine protease inhibitor E-64 ( L-trans-epoxysuccinyl-leucylamide- ( 4-guanido ) -butane; Sigma ) or preheated at 70°C . Labeled SmCL3 samples were resolved by SDS-PAGE ( 15% Tris-HCl Criterion gel; Biorad ) and visualized by autoradiography using a Typhoon Trio 8600 Variable Mode Imager ( GE Healthcare ) . Proteolytic activity was measured with the synthetic fluorogenic dipeptidyl substrate Z-Phe-Arg-AMC ( benzyloxycarbonyl-phenylalanylarginine-7-amido-4-methylcoumarin; Bachem ) . Assays were performed in black 96-well plates as described previously [28] . Briefly , recombinant SmCL3 enzyme was pre-incubated for 10 min at RT ( room temperature ) in 50 mM citrate , 100 mM sodium phosphate , pH 3 . 0–8 . 0 or 100 mM glycine , pH 7 . 0–11 . 0 . All buffers contained 100 mM NaCl and 2 mM dithiothreitol ( DTT ) in a final volume of 100 µl . The reactions were started by adding 100 µl of the same buffer solution containing 40 µM Z-Phe-Arg-AMC . Release of free AMC was measured at excitation and emission wavelengths of 355 and 460 nm , respectively , in a Labsystems Fluoroskan II fluorescent plate reader ( Thermo Electron Corporation ) . For pH stability assays , recombinant SmCL3 samples were incubated in 50 mM citrate , 100 mM sodium phosphate , 2 mM DTT , pH 3 . 0–8 . 0 at 37°C for 1 h . Enzyme activities were analyzed at pH 6 . 0 using fluorescent dipeptidyl substrate Z-Phe-Arg-AMC and active site labeling with 125I-DCG-04 . The Km value and kcat ( turnover rate ) for SmCL3 with Z-Phe-Arg-AMC were determined by nonlinear regression analysis Prism 4 ( GraphPad ) . Rates were obtained from substrate concentrations ( 0 . 2–150 µM ) with a fixed enzyme concentration of 3 nM . Assays were performed in black 96-well plates in 50 mM citrate , 100 mM sodium phosphate , pH 6 . 0 at a final volume of 200 µl . Release of free AMC was measured at 25°C in a Flex Station fluorescent plate reader ( Molecular Devices ) . Kinetic analyses with irreversible cysteine protease inhibitors were performed as previously described [31] . Enzyme ( ∼3 nM ) in 100 µL 50 mM citrate , 100 mM sodium phosphate , pH 6 . 0 ( see above ) , was added to inhibitor dilutions in 100 µL of the same assay buffer containing 25 µM Z-Phe-Arg-AMC . Progress curves were recorded for 5 min in the Flex Station fluorescent plate reader at 25°C ( less than 5% of substrate consumed ) over a range of dilutions ( 0 . 5 , 0 . 4 , 0 . 3 , 0 . 2 , 0 . 1 , 0 . 05 , and 0 µM ) of inhibitors the cysteine protease inhibitors E-64 or K11777 ( N-methyl piperazine-ureaphenylalanyl-homophenylalanyl-vinylsulfone-benzene [32] , [33] dissolved in DMSO ( final DMSO in assay was 0 . 5% ) . Inhibitor dilutions giving simple exponential progress curves over a wide range of kobs ( first order observed inhibition constant ) with r∧2 values ≥ to 0 . 9 were used to determine kinetic parameters . The value of kobs , the rate constant for loss of enzyme activity , was determined from an equation for pseudo first order dynamics using Prism4 software ( GraphPad ) . When kobs varied linearly with inhibitor concentration , kass ( complex formation constant ) was determined by linear regression analysis [34] . If the variation was hyperbolic , indicating saturation inhibition kinetics , kinact ( maximal inactivation rate constant ) and Ki ( inhibition constant ) were determined from an equation describing a two step irreversible inhibitor mechanism ( kobs = kinact [I]o/ ( [I]o+Ki* ( 1+[S]o/Km ) ) ) and nonlinear regression analysis using Prism . 4 . Recombinant SmCL3 ( ∼100 nM ) was incubated overnight at 37°C with bovine albumin or bovine hemoglobin ( 1 mg/ml; Sigma ) in 50 mM citrate , 100 mM sodium phosphate , 2 mM DTT , pH 3 . 0–10 . 0 . After incubation , a 20 µl sample was resolved by 10% Bis-Tris NuPage Novex gel with MES buffer running buffer ( Invitrogen ) . PS-SCL were employed as previously described [35] . All 20 amino acids were incorporated in tetrapeptides where cysteine was omitted and norleucine included . Assays involving either SmCL3 or SmCL3-his were carried out in black 96-well microtiter plates at pH 6 . 0 , as described previously [35] , [36] . Release of 7-amino-4-carbamoylmethylcoumarin ( ACC ) was measured in a Perkin-Elmer LS50B luminescence spectrometer with excitation and emission wavelengths set to 380 and 460 nm , respectively . One mg of purified recombinant SmCL3-his was resolved by SDS-PAGE ( 12% Tris-HCl Criterion gel; Biorad ) . Gels were briefly stained in SimplyBlue Safe Stain to visualize the SmCL3 protein band and then washed with water . The protein band was excised and homogenized in sterile saline using a glass homogenizer . Five Swiss-Webster mice were injected with a 100 µl mixture of antigen and adjuvant 4 times at 14 day intervals . The first injection was administered intraperitoneally in Freunds Complete Adjuvant ( Sigma ) in a ratio 3∶1 . Three subsequent subcutaneous injections contained antigen in TiterMax Gold adjuvant ( Sigma ) at a 2∶1 ratio . For control sera , blood samples were withdrawn from mice receiving acrylamide samples alone . Seven days after the last injection , mice were euthanized and exsanguinated . After clotting , serum was separated from blood cells and then the IgG fraction isolated using a HiTrap Protein G column ( GE-Healthcare ) , according to the manufacturer's protocol . For immunoblotting , S . mansoni soluble protein extracts were prepared by sonication in 50 mM citrate , 100 mM phosphate , pH 5 . 0 over an ice bath in the presence of Protease Inhibitor Cocktail ( Sigma ) . After brief centrifugation at 8 000 g for 5 min at 4°C , supernatants containing soluble proteins were collected . Extracts ( 20 µg per well ) and recombinant SmCL3 were resolved by SDS-PAGE ( 15% Tris-HCl Criterion gels ) and transferred onto a PVDF membrane ( Biorad ) . Membranes were blocked overnight at 4°C in 5% non-fat dry milk in Tris-buffered saline containing 0 . 1% Tween 20 ( TBS-T ) and washed 3×5 min in TBS-T . After washing , membranes were incubated for 1 h with anti-SmCL3 or control purified polyclonal IgG ( 1∶1000 ) in TBS-T . Membranes were then washed 3×15 min in TBS-T and incubated for 1 h with anti-mouse IgG-HRP conjugate ( GE Healthcare ) at a dilution of 1∶2000 . After washing in TBS-T 3×15 min , followed by a single wash in TBS for 5 min , membranes were developed using an enhanced chemiluminescent kit ( ECL Western Blotting Detection Reagents , GE Healthcare ) according to the manufacturer's instructions . Immunoreactivity was visualized by exposure to the SuperRX Medical X-Ray Film ( Fuji ) . Perfused adult S . mansoni worms were fixed in 0 . 1% glutaraldehyde in PBS , pH 7 . 4 at RT for 2 h , washed 3×15 in PBS , pH 7 . 4 and stored at 4°C prior to use . Samples were then embedded in JB-4 ( Polyscience ) , sectioned at 2 . 5 µm , placed on glass slides and dried at 60°C for 5 min . Incubation with mouse control or anti-SmCL1 IgG antibodies at 1∶200 dilutions in TBS-T and secondary Alexa Fluor 594 anti-mouse IgG ( Invitrogen ) was carried out as described [37] . Localization was observed using a laser scanning microscope LSM 510 META ( Carl Zeiss ) . Soluble extract from adult worms was prepared as described above . The extract was size fractioned using pre-equilibrated column Superdex 200 ( GE- Healthcare ) according to manufacture's instructions . Eluted fractions were resolved by SDS-PAGE ( 15% Tris-HCl Criterion gel , Biorad ) and transferred onto a PVDF membrane ( Biorad ) and SmCL3 was detected by Western blot analysis . The SmCL3 protein sequence was used as a query in a web-based blastp at http://blast . ncbi . nlm . nih . gov [38] search of the Protein Data Bank ( PDB; http://www . rcsb . org/pdb ) using the default setting of filtering low-complexity regions . The fourth best hit was used as the template for modeling because this hit had a good E-value and also included an inhibitor complexed with the protein , which improves modeling results . The template was human cathepsin V complexed with vinyl sulfone inhibitor K11777 [32] , [33] , pdb ID 1FH0 , ( with identical chains A and B , solved to 1 . 6 Å resolution ) . The BLAST alignment of SmCL3 and 1FH0 had 59% sequence identity ( 135/227 residues ) , E-value = 2e-74 . The alignment from the BLAST search was used with the homology modeling program PLOP [39] . In order to show the active site as a substrate would likely bind , views of the model were generated in Chimera [40] as follows: The template 1FH0 , chain A , was aligned to the SmCL3 model in Chimera using the Matchmaker tool . The template was then hidden except for the inhibitor . The catalytic Cys172 and His317 were colored yellow and blue , respectively . The residues in the S2 binding pocket that are ≤5 Å from the inhibitor are shown in ball-and-stick format . The predicted residues in this pocket are identical to those in the 1FH0 template ( cathepsin V ) except for one residue which is Leu216 in SmCL3 ( colored light green ) and Phe69 in 1FH0 . Other important active site residues Gln166 and Asn337 aligned closely with the same corresponding residues in 1FH0 ( not highlighted in the model ) . To analyze SmCL3 prodomain structure , protein sequence was also imported into the protein modeling program interface Maestro ( Maestro 8 . 5207 , Schrodinger , LLC ) , and the secondary structure prediction program PSIpred [41] run on the sequence through the Maestro interface using the Prime application ( Prime 2 . 0208 , Schrodinger , LLC ) . Secondary structure prediction programs such as PSIpred are about 75% accurate ( http://cubic . bioc . columbia . edu/eva/sec/res_sec . html ) . SmCL3 was queried against the UniRef100 database ( http://www . ebi . ac . uk/uniref/ ) [42] of non-redundant protein sequences using the program blastp [38] . A perl script was then used to select 1025 sequence hits scoring at E-value≤1e−30 and where the alignment length was at least 80% of the query length . The sequences were filtered to a set of 297 sequences ≤60% identical to each other using the program CD-HIT [43] . An all vs . all blastp search of these representative sequences was then performed to find sequence similarity relationships between all 297 proteins . Perl scripts were used to parse the species names from the UniRef IDs and to key species to class using data from NCBI Taxonomy ( http://www . ncbi . nlm . nih . gov/Taxonomy ) . The resulting data of sequence similarity relationships and node labels were formatted , colored by class and visualized using sequence similarity networks ( SSNs ) for visualization of relationships across diverse protein superfamilies [44] in the ‘organic’ layout with Cytoscape v2 . 4 . 1 [19] . An E-value cutoff threshold of 1e−60 was used for drawing edges between sequences . Cytoscape is an open source bioinformatics software platform for visualizing many types of biological networks ( http://www . cytoscape . org/index . php ) . In the ‘organic’ view , each representative sequence is displayed as a colored “node” with lines connecting nodes signifying a BLAST E-value relationship better than the cutoff value . The 247 nodes that formed clusters are shown; more highly interconnected nodes have shorter edges than less well-connected nodes . To aid interpretation of the output , the nodes were also colored to correspond to a super kingdom classification proposed by Simpson and Roger [45] . For details about included gene sequences see supporting Cytoscape data ( Fig . S1; Note: you have to download Cytoscape v2 . 4 . 1 program at http://www . cytoscape . org/index . php ) . PCR strategies based on EST information [18] led to the amplification , sequencing and characterization of a novel cathepsin L gene in S . mansoni that we term SmCL3 in accordance with the previously used nomenclature [7] , [46] . PCR screening did not identify other gene isoforms . The open reading frame ( ORF ) consists of 1113 bp ( 370 amino acids; GenBank accession EU022371 ) that encodes a pre-proenzyme ( Fig . 1 ) . The signal leader sequence was predicted to have a length of 16 amino acid residues . The 130 residue pro-peptide sequence was predicted from a multiple sequence alignment using BLASTP 2 . 2 . 18 ( http://www . ncbi . nlm . nih . gov/blast/ ) [38] . Mw/pI values , calculated by the Compute pI/Mw program ( http://www . expasy . org ) [47] , are 41 . 2/6 . 5 , 39 . 4/6 . 5 , and 24 . 1/4 . 9 kDa for the full length , zymogen and mature proteins , respectively . Cys172 , His317 , Asn337 form the protease's catalytic triad that is essential for peptidolytic activity . Gln168 , a residue expected to be involved in the formation of the oxyanion hole , is present . The mature ( catalytic ) domain has 3 putative disulfide bonds typical of other cathepsin L enzymes [48] . Potential N-linked glycosylation sites are at positions 194 and 252 ( Fig . 1 ) . Compared to typical cathepsins L , the pro-peptide of SmCL3 is unusually long with an N-terminal extension of approximately 30 amino acids , similar to the S . japonicum ortholog ( SjCL3; GenBank AAW27185; [49] ) and two more distant Clonorchis sinensis cathepsins L ( Genbank ABK91809 , ABJ89815; Hu et al , unpublished ) . Also , an asparagine residue , present in the pro-peptide of previously characterized S . mansoni proteases and a site of trans-activation by asparaginyl endopeptidases [30] , [50] is absent in SmCL3 . Like other cathepsins L , the pro-peptide contains a type I-29 protease inhibitor motif [51] , [52] ( Fig . 1 ) . A variant of the ERFNIN motif , found in other cathepsin L family pro-peptides [53] , is present as ERFNMN . A second motif , GNFD , which is involved in intramolecular processing of other Clan CA proteases [54] , is also present in the pro-peptide ( Fig . 1 ) . The elongated prodomain is not random coil but is predicted to be alpha helix by protein modeling using Maestro . SmCL3 was successfully expressed in the yeast P . pastoris , fully processed and activated; i . e . , without the presence of the pro-peptide . Typical yields of recombinant SmCL3 were 30–50 mg/l of expression media . Peptidolytic activity was evident with or without the C-terminal 6×His-tag using the dipeptidyl substrate Z-Phe-Arg-AMC . As judged by kinetic analyses and assays with the positional scanning synthetic combinatorial library ( see SmCL3 positional scanning below ) , the presence of this 6×His-tag had no effect on catalysis and this expression variant was , therefore , used for subsequent studies . The Clan CA specific inhibitor , E-64 , eliminated peptidolytic activity , thus verifying the catalytic mechanism as consistent with cysteine proteases . Using SDS-PAGE , the estimated molecular mass was ∼32–34 kDa which decreased to 28–30 kDa after enzymatic deglycosylation ( Fig . 2 ) consistent with the use of at least one of the two potential glycosylation sites by Pichia . Purified SmCL3 was labeled with the cysteine protease affinity probe , 125I-DCG-04 ( Fig . 3 ) and cleaved gelatin on zymogram gels ( not shown ) . Though the enzyme was expressed as fully active , some processing heterogeneity was noted by N-terminal protein sequencing of the purified expression product . The most abundant cleavage site ( as predicted above ) was after the Lys153 ( HTK↓LPS , Fig . 1 ) . A less abundant and slower migrating protein species was also produced by Pichia ( Fig . 3 ) . 125I-DCG-04-labeling confirmed the band as a variant form of SmCL3 ( Fig . 3 ) . We attempted to sequence this minor band but without success . SmCL3 is catalytically active over a broad pH range . Hydrolysis of Z-Phe-Arg-AMC displayed a bell-shaped pH profile from pH 3 . 0–11 . 0 with optimal activity around pH 6 . 5 ( Fig . 4A ) . Bovine albumin and bovine hemoglobin were degraded: albumin was partially hydrolyzed with a pH optimum around 6 . 0 ( Fig . 5A ) ; hydrolysis of hemoglobin was complete at pH 4 . 0–6 . 0 with partial hydrolysis at lower and higher pH values ( Fig . 5B ) . These pH dependencies for activity correlated with the enzymatic stability of SmCL3 between pH 4 . 0–6 . 0 as measured with both Z-Phe-Arg-AMC and 125I-DCG-04 ( Fig . 4B ) . No loss of activity was recorded after incubation of enzyme for 30 min , 1 and 3 h . However , at other pH values , a time dependent decrease in activity was measured . Therefore , the difference in profiles between the activity and stability experiments is possibly due to the instability of the enzyme at pH values equal to or greater than 7 . 0 . Kinetic constants obtained for SmCL3 with Z-Phe-Arg-AMC were: Km = 20 . 2 µM and kcat/Km = 410 mM−1 s−1 . Inhibition constants ( kobs at 1 nM of inhibitor ) measured for SmCL3 with E-64 and K11777 were 26 . 5 and 140 nM−1 s−1 , respectively . Consistent with other Clan CA proteases [35] , SmCL3 prefers the basic amino acids lysine and arginine at the P1 subsite position ( Fig . 6 ) . At P2 , the enzyme prefers hydrophobic amino acids , especially bulky aromatic residues . Upon a search of the literature involving PS-SCL assays , the P2 preferences of SmCL3 was found to closely resemble those of human cathepsin V [35] , a cathepsin L-like protease . In particular , there is an overriding preference for tryptophan and equal preference for phenylalanine and leucine in the P2 sites of both enzymes . Screening at P3 and P4 revealed greater promiscuity . Notably , SmCL3 is able to accept aspartic acid at P2 and P3 positions , which is similar to human cathepsin F [35] . For the three-dimensional model of SmCL3 , the X-ray crystallographic structure of human cathepsin V complexed with a peptidyl vinyl sulfone inhibitor , K11777 , was used as template . We used this template because of its high percentage identity ( 59% ) to SmCL3 and because the structure was solved with an inhibitor in the active site thereby likely making the modeling of the active site more accurate . From the structure-based alignment , the predicted interaction of the modeled structure with K11777 is depicted in Fig . 7 . The predicted residues in the S2 binding pocket of SmCL3 are identical to those in cathepsin V with the exception of a Leu residue ( Leu216 ) which is phenylalanine ( Phe69 ) in cathepsin V . This substitution enlarges what is already a deep and wide pocket , and consistent with the results from the PS-SCL , appears well adapted to accept the side chains of bulky aromatic residues , such as tryptophan , tyrosine and phenylalanine . The prodomain region was lacking in the template and so is not included in the model . However , secondary structure prediction indicates that five helices are likely to form in the SmCL3 prodomain ( not shown ) . Quantitative PCR demonstrated that SmCL3 is predominantly expressed in those developmental stages infecting the mammalian host ( Fig . 8A ) , a result that is in accord with the protein expression profile as shown by immunoblots with specific polyclonal anti-SmCL3 IgG ( see below ) . Most mRNA for SmCL3 was detected in transformed schistosomula in vitro , and adult male and female worms . Expression profiling by qPCR indicated that SmCL3 mRNA is 50 to 1000 fold less abundant relative to previously described gut-associated proteases in S . mansoni adults [7] , [8] ( Fig . 8B ) . SmCL3 mRNA is also less abundant than that of the tegumental/parenchymal SmCB2 [27] ( more than 100-fold ) , but is found in greater quantities than the endoplasmatic reticulum protease , SmER-60 [55] ( more than 10-fold ) . By immunobloting with specific polyclonal anti-SmCL3 IgG ( Fig . 9 ) , native SmCL3 was detected in extracts of both adults and newly-transformed schistosomula 1 h after in vitro transformation . Weaker reactivity was detected in extracts of eggs and no reaction was found in extracts of miracidia and cercariae . Control mouse IgG antibodies were non-reactive throughout ( not shown ) . Unlike the immuno-reactivity observed at approximately 30 kDa for the recombinant enzyme ( Fig . 9 , lanes 1 and 2 ) , the major reactive band in schistosome extracts migrated with a mass of approximately 40 kDa ( Fig . 9 , lanes 3 , 4 and 7 ) , a mass that corresponds to that of the SmCL3 zymogen . Attempts to process in trans pro-SmCL3 within extracts using other recombinant proteases such as SmCB1 [30] and a asparaginyl endopeptidases from tick [36] or S . mansoni [30] failed , as did incubation of worms extracts overnight at 37°C in an effort to endogenously process the zymogen ( not shown ) . The data suggest , therefore , that SmCL3 is present in its major form as a zymogen rather than as mature catalytically active enzyme . As judged by immunoblotting , size exclusion chromatography of S . mansoni adult soluble protein extracts separated the putative SmCL3 zymogen ( Fig . 10 , fractions 24–27 ) from the immunoreactive protein species of 30 kDa – the possible mature enzyme ( fractions 30–32 ) , and of 13 and 11 kDa – possible SmCL3 fragments ( fractions 35–38 ) . SmCL3 was not detected by specific polyclonal IgG in excretory/secretory ( E/S ) products of adult worms maintained in culture medium . Nevertheless , SmCL3 was detected by antibody in the regurgitant when adult worms were induced to regurgitate in water ( data not shown ) . SmCL3 was localized to the gastrodermis of both adult sexes with some reaction in the female vitellaria using confocal microscopy with mouse anti-SmCL3 IgG and Alexa Fluor 594 secondary antibodies ( Fig . 11A , 11C , and 11D ) . No reaction was observed in the tegument and parenchyma . No staining was observed with control mouse polyclonal IgG ( Fig . 11B ) . A network view of primary protein sequence similarity relationships among cathepsin L type enzymes was generated using the software Cytoscape [19] . Each sequence is represented as a square node , except for cathepsin L sequences from platyhelminths which are indicated by circular nodes and those representing Trematoda are enlarged circular nodes ( Fig . 12 ) . Of immediate interest is that the clustering of cathepsin L sequences agrees closely with the taxonomic organization of the kingdoms of life into six supergroups [45]: Opisthokonta , Plantae , Chromalveolata , Amoebozoa , Rhizaria and Excavata ( Fig . 12 ) . SmCL3 ( white circular node ) is found within a large cluster of closely related invertebrate ( light blue squares ) and vertebrate metazoan ( dark blue squares ) cathepsins L that together make up the super kingdom Opisthokonta . This large cluster also includes the S . japonicum ortholog , SjCL3 , two C . sinensis cathepsin L genes , the SmCL2 gene and related cathepsins L from Fasciola gigantica and F . hepatica . The cluster is distinct from a cluster of cathepsins L that is restricted to the Plantae super kingdom , an organizational level of primary plastid endosymbionts comprising plants , green and red algae . More disparate clusters of cathepsin L sequences are found in the super kingdom Chromalveolata ( secondary symbionts; contains apicomplexan parasites such as Toxoplasma and Cryptosporidium ) , the Amoebozoa ( includes the parasite Entamoeba histolytica ) . Another compact cluster displayed in Fig . 12 is entirely composed of baculovirus cathepsin L-like genes ( encircled black ) and is least connected to the other clusters . The Cytoscape view also resolves a cluster of sequences that is enriched in cathepsins F and W , which are subtypes of cathepsin L ( encircled in orange ) . This cluster includes sequences of greater phylogenetic diversity including SmCLl ( aka SmCF ) [7] , cathepsins F from Opisthorchis viverrini , Clonorhis sinensis , Paragonimus westermani and Metagonimus yokogawai , and Excavata parasitic kintetoplastid cathepsins . Finally , a small cathepsin H cluster , another subtype of cathepsin L ( encircled in green ) is resolved that from the clusters containing cathepsins L and F/W . For sequence details see supporting Cytoscape data ( Fig . S1; Note: after downloading Cytoscape ) . Growth , maturation and fecundity of the schistosome parasite in the mammalian host rely on nutrients ingested from the host bloodstream . A number of proteases are expressed in the gut of S . mansoni and are involved in the degradation of hemoglobin and serum proteins [7] , [8] . This multienzyme network includes two cathepsins L , SmCL1 ( aka SmCF ) and SmCL2 [7] , [46] , [56] . Although sequences for other cathepsins L exist in the EST datasets [18] and in first pass assembly of the genome ( Mashiyama , Caffrey , Sajid , unpublished ) , nothing is known about their contribution to schistosome metabolism . Here , we identified , heterologously expressed and characterized a novel gut-associated cathepsin L that we term SmCL3 . A sequence for an ortholog in S . japonicum ( SjCL3 ) also exists ( GenBank AAW27185; [49] ) . SmCL3 possesses sequence characteristics consistent with those of other cathepsins L . These include six Cys residues forming three disulphide bonds [48] , an active site catalytic triad of Cys , His and Asn [57] , the residue Gln168 involved in the formation of the oxyanion hole , a pro-peptide that contains an I29 inhibitor family sub-domain and a variation of the ERFNIN motif ( ERFNMN ) that is typical for cathepsins L [53] . This motif , together with the motif GNFD [54] , is probably involved in intra-cellular trafficking and processing . Other sequence features of SmCL3 are more unusual , especially when compared to other helminth cysteine proteases associated with the gut . First , an Asn residue , found between the pro-peptide and mature domain of other gut cathepsins in Schistosoma [30] , [50] and fasciolids [50] , [58] , and demonstrated to be a processing site for pro-cathepsin activation by an asparaginyl endopeptidase ( AE ) [30] , is absent . Unlike recombinant S . mansoni pro-cathepsin B1 expressed in Pichia that requires trans-processing by an endogenous AE for full activity [30] , SmCL3 is already fully processed in Pichia induction medium at the predicted cleavage site , as judged by Edman N-terminal sequencing and proteolytic activity . This suggests that recombinant pro-SmCL3 is capable of auto-catalytic activation and maturation . Secondly , the SmCL3 zymogen has an unusually long pro-peptide comprising 130 residues . Approximately the first 30 amino acids of the pro-peptide share some homology with the SmCL3 ortholog in S . japonicum and two C . sinensis cathepsins L . For the SmCL3 prodomain , five helical structures were predicted which imply some regulatory or supplementary structural role . Prodomains that are extended N-terminally , though different in sequence , are also found in the gut-associated cathepsins L of the animal parasitic nematode Gnathostoma spinigerum [59] and the plant parasitic nematode Meloidogyne incognita [60] . These extensions may confer additional functionality to the zymogen , perhaps in protein trafficking or as binding sites for other proteins . It is also possible that this extension may be associated with the fact that the major form of the enzyme in the parasite apparently exists as a zymogen and/or the enzyme seems not to be secreted into the gut lumen ( see discussion below ) . SmCL3 cleaves albumin and hemoglobin most efficiently at pH values between 4 . 0 and 6 . 0 . The pH dependency of hydrolysis of the Z-Phe-Arg-AMC synthetic substrate results in a bell-shaped curve from pH 3 . 5 to 11 . 0 with an optimum at 6 . 5 . At least 40% of total activity can be detected between pH 4 . 0 and 10 . 0 . A similar bell-shaped pH profile was measured for SmCL1 [46] , which , unlike SmCL2 , was able to cleave peptidyl substrate at basic pH . The acidic pH optima measured against both protein and peptidyl substrates correlates with the pH of the gut lumen ( ∼pH 6 . 5 ) [30] , [61] and with the lower pH ( ∼4 . 0 ) micro-environments thought to form upon fusion of gut lamellae and where it is hypothesized that the bulk of gastrodermal proteolysis by cysteine and aspartic proteases takes place [8] . Recombinant SmCL3 possesses peptidolytic characteristics consistent with its classification as a Clan CA Family C1 protease: it is effectively inhibited by the Clan CA-specific inhibitors E-64 and K11777 [32] , [33] , and labeled by the affinity probe DCG-04 [29] . Positional scanning using diverse synthetic substrate libraries revealed a typical Clan CA preference profile: no single amino acid preference in S4 and S3 but a strong preference for lysine and arginine in the S1 subsite [35] . However , in S2 ( the subsite driving specificity in Clan CA proteases ) , hydrophobic amino acids ( Trp>Tyr>Phe/Val>Leu ) are preferred . These preferences at P2 are similar to those of human cathepsin V [35] but differ , for example , from F . hepatica cathepsins L1 and L2 that exhibit a singular preference for Leu and Pro in the case of FhCL2 [62] , [63] . In support of the S2 preferences demonstrated biochemically with the PS-SCL , the 3D structural model of SmCL3 , using K11777 as the bound ligand , visualizes a large and deep S2 pocket . The amino residues forming the S2 pocket are identical to those of cathepsin V with the exception of one substitution of a Leu ( residue 216 ) instead of Phe . SmCL3 is developmentally regulated at both the mRNA and protein levels being mainly expressed in those stages ( schistosomula and adult ) infecting the definitive mammalian host and thus suggesting a function ( s ) particular to these developmental stages . That one of these functions is associated with the digestion of host blood proteins is supported by confocal microscopy using polyclonal IgG that localizes SmCL3 to gastrodermis of adult worms . The hypothesis is consistent with the ability of the enzyme to degrade biologically relevant protein substrates , i . e . , hemoglobin and bovine serum albumin , as discussed above . Given that the transcription of SmCL3 is 50 to1000 fold less than other gut-associated proteases , the actual proteolytic contribution by SmCL3 to total proteolysis in the gut remains to be determined . Because of its localization , it is conceivable that SmCL3 operates with the other gut proteases to complete the degradation of host proteins as nutrients [7] , [8] . However , unlike proteases such as SmCB1 , SmCL1 and SmCL2 [30] , [44] , [56] , SmCL3 was not detected in worm E/S products when maintained in isotonic culture medium . However , when adult worms were induced to regurgitate in water , i . e . , exposed to hypotonicity , potentially causing damage to gut cells , SmCL3 could then be detected by specific antibody in the regurgitant ( data not shown ) . This suggests that SmCL3 is normally retained within the gastrodermal epithelium and is not secreted . Of note is that the G . spinigerum cathepsin L is likewise not detected in E/S products even though it is found in the gastrodermis [59] . Apart from the gut , some immuno-reaction for SmCL3 was also noted in the vitellaria of female S . mansoni , a finding consistent with the presence of a small amount of SmCL3 in eggs by immunoblotting . Therefore , SmCL3 may also function in egg and/or miracidial metabolism . By immunoblotting , SmCL3 was detected in parasite extracts at a molecular mass of approximately 40 kDa , i . e . , consistent with that of the zymogen . A similar situation was noted recently for the G . spinigerum cathepsin L – the major species of that enzyme also migrated at approximately 40 kDa in worm extracts [59] . Attempts to trans-process the SmCL3 zymogen in worm extracts using other recombinant proteases failed , as did incubating worm extracts overnight at 37°C . After size-exclusion chromatography of adult worm extracts , in addition to the resolution of a major 40 kDa protein , minor immuno-reactive protein species were detected at approximately 28 , 15 and 13 kDa . These may represent the mature deglycosylated enzyme , and two degradation products , respectively . It seems , therefore , that the major form of SmCL3 in S . mansoni is a zymogen . The retention of the pro-peptide with the mature domain opens the possibility of a distinct function ( s ) for the zymogen , including the possibility of a limited or discrete processing activity against protein and peptide substrates . Precedents for protease zymogens that exhibit peptidolytic activity exist [30] , [64] . Often , the strength of the association between the pro-peptide and mature domain is pH dependent [65] , allowing access to and cleavage of small peptide substrates . To investigate evolutionary relationships of the full SmCL3 protein sequence and its cathepsin L neighbors , we examined the top 1 , 000 hits in a BLAST search of the SmCL3 protein sequence . We found only 16 other trematode sequences for which the N-terminus extended far enough to overlap at least 75% of the prodomain of SmCL3 . Most of these sequences were highly similar to each other and after filtering to 90% identity , there were only 5 sequences including the SmCL3 sequence . As expected , a multiple sequence alignment of these sequences showed a highly conserved catalytic domain , and a more variable prodomain region ( data not shown ) . In order to visualize more distant relationships between SmCL3 and cathepsin L-like sequences , we constructed a SSN using the program Cytoscape [19] . We have recently established that SSNs show good agreement with information provided by phylogenetic trees and allow a clear view of all of the represented proteins in a dataset together with easy associations to functional and other types of information [44] . Based on full-length sequences , and as visualized with the software Cytoscape ( Fig . 12 ) , the sequence clustering of cathepsin L proteins recapitulates the recent proposed partition of Eukarya into six “super kingdoms” based on multivariate phylogenetic analyses [45] . SmCL3 is found within the Opisthokonta super kingdom ( that includes animals and fungi ) . The Opisthokonta cathepsins L also includes the SjCL3 ortholog from S . japonicum , SmCL2 [66] , and two C . sinensis cathepsins L ( Hu et al , unpublished ) , as well as a collection of sequences from F . hepatica and F . gigantica [58] . The network clearly resolves a cluster that is enriched in cathepsins F/W ( encircled in orange ) and another enriched in cathepsin H ( encircled in green ) . This confirms the previous distinction of the subgroup cathepsins F/W from the main body of cathepsins L , which arose as a result of gene fusion between an ancestral cathepsin L and a cystatin ( cysteine protease inhibitor ) [67] , and from another cathepsin L subgroup , cathepsins H , which contain a specific mini-chain formation to function as aminopeptidases [68] . The inclusion within the cathepsin F/W cluster of SmCL1 ( SmCF ) , other trematodal cathepsins F and the kinetoplastid cathepsins L supports and extends previous phylogenetic data [48] , [69]–[71] . Overall , given the close agreement of the distance relationships observed here between cathepsins L and the taxonomic separation of the tree of life proposed previously [45] , we would suggest that cathepsins L are useful ‘marker genes’ for inclusion in future phylogenetic analyses . More in-depth studies of this and other issues in a global analysis of the members of this super-family may be enlightening future work . To conclude , SmL3 is a gut-associated protease with some unusual sequence and biophysical features . The enzyme may function as part of the network of proteases [8] that facilitates the digestion of host proteins by the schistosome parasite . As inhibitors of Clan CA proteases are therapeutic in animal models of schistosomiasis [15] , [16] , it is possible that the inhibition of SmCL3 , either alone or in concert with other cysteine proteases , may prove clinically beneficial .
Parasitic infection caused by blood flukes of the genus Schistosoma is a major global health problem . More than 200 million people are infected . Identifying and characterizing the constituent enzymes of the parasite's biochemical pathways should reveal opportunities for developing new therapies ( i . e . , vaccines , drugs ) . Schistosomes feed on host blood , and a number of proteolytic enzymes ( proteases ) contribute to this process . We have identified and characterized a new protease , SmCL3 ( for Schistosoma mansoni cathepsin L3 ) , that is found within the gut tissue of the parasite . We have employed various biochemical and molecular biological methods and sequence similarity analyses to characterize SmCL3 and obtain insights into its possible functions in the parasite , as well as its evolutionary position among cathepsin L proteases in general . SmCL3 hydrolyzes major host blood proteins ( serum albumin and hemoglobin ) and is expressed in parasite life stages infecting the mammalian host . Enzyme substrate specificity detected by positional scanning-synthetic combinatorial library was confirmed by molecular modeling . A sequence analysis placed SmCL3 to the cluster of other cathepsins L in accordance with previous phylogenetic analyses .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results", "Discussion" ]
[ "biochemistry", "infectious", "diseases", "infectious", "diseases/neglected", "tropical", "diseases", "computational", "biology/comparative", "sequence", "analysis", "biochemistry/protein", "chemistry", "infectious", "diseases/helminth", "infections", "biochemistry/bioinformatics", "computational", "biology", "infectious", "diseases/tropical", "and", "travel-associated", "diseases" ]
2009
SmCL3, a Gastrodermal Cysteine Protease of the Human Blood Fluke Schistosoma mansoni
The establishment of precise neuronal connectivity during development is critical for sensing the external environment and informing appropriate behavioral responses . In the visual system , many connections are organized topographically , which preserves the spatial order of the visual scene . The superior colliculus ( SC ) is a midbrain nucleus that integrates visual inputs from the retina and primary visual cortex ( V1 ) to regulate goal-directed eye movements . In the SC , topographically organized inputs from the retina and V1 must be aligned to facilitate integration . Previously , we showed that retinal input instructs the alignment of V1 inputs in the SC in a manner dependent on spontaneous neuronal activity; however , the mechanism of activity-dependent instruction remains unclear . To begin to address this gap , we developed two novel computational models of visual map alignment in the SC that incorporate distinct activity-dependent components . First , a Correlational Model assumes that V1 inputs achieve alignment with established retinal inputs through simple correlative firing mechanisms . A second Integrational Model assumes that V1 inputs contribute to the firing of SC neurons during alignment . Both models accurately replicate in vivo findings in wild type , transgenic and combination mutant mouse models , suggesting either activity-dependent mechanism is plausible . In silico experiments reveal distinct behaviors in response to weakening retinal drive , providing insight into the nature of the system governing map alignment depending on the activity-dependent strategy utilized . Overall , we describe novel computational frameworks of visual map alignment that accurately model many aspects of the in vivo process and propose experiments to test them . Processing sensory information is a critical task of the central nervous system , requiring the establishment of precisely ordered synaptic connectivity during development . In the visual system , image-forming regions are organized into topographics maps , such that neighboring neurons monitor adjacent regions of visual space [1 , 2] . The development of topographic connections in the visual system has been the focus of intense study , both experimentally and theoretically , elucidating general principles underlying neural circuit wiring [3 , 4] . However , these studies have focused primarily on the mechanisms by which topographic connectivity is established for a single projection . In regions that integrate visual information , multiple converging inputs must establish topography and be aligned with one another to facilitate integration [5] . Yet , little is known about the mechanisms by which topographic maps of space are aligned in these regions , in part due to a lack of computational frameworks that model this process . The superior colliculus ( SC ) is a critical multisensory integration center that receives visual , somatosensory , and auditory inputs that inform goal-directed head and eye movements [6–8] . The SC receives visual inputs from retinal ganglion cells ( RGCs ) and Layer 5 pyramidal neurons in the primary visual cortex ( V1 ) [9] . Each of these inputs projects to distinct , but overlapping , sublaminae of the superficial SC , where they are organized topographically and in alignment with one another [10] . The mapping of retinocollicular projections occurs during the first postnatal week in mice , and a combination of molecular cues [11–17] , correlated neuronal activity [18–20] and competition [21 , 22] have been demonstrated to regulate the establishment of precise retinocollicular topography . The mechanisms by which V1 inputs establish topography and alignment with retinal inputs are less clear . Mapping of V1 corticocollicular inputs occurs during the second postnatal week in mice , after retinocollicular topography has been established . Previously , we demonstrated that retinal input instructs the alignment of V1 axons in a manner dependent on the normal pattern of spontaneous activity [23] . Subsequent studies confirmed that correlated spontaneous activity originating in the retina propagates throughout V1 and the SC in vivo [24] , supporting its possible role as the instructive cue for alignment . Further , the timing of spiking acitivty in V1 and the SC is consistent with activity-dependent visual map alignment [25] . However , the underlying mechanisms of activity-dependent alignment remain unclear . Theoretical modeling of neural circuit development is a powerful tool to both better describe complicated processes and generate novel hypotheses regarding circuit wiring [26] . Indeed , several mathematical models have been developed to describe topographic mapping of retinocollicular projections [27–32] . However , each of the current models has weaknesses and cannot replicate the full complement of empirical data obtained from in vivo studies of mutant mice [4] . Further , no theoretical models of visual map alignment have been developed , hindering our ability to probe potential mechanisms of this critical developmental event . Here , we describe two novel models of visual map alignment in the SC , each of which utilizes a different activity-dependent mechanism for visual map alignment , providing an in silico platform to investigate strategies used in vivo . First , a Correlational Model assumes that SC neuron firing is driven only by RGC inputs . In this case , alignment of V1 inputs is guided by simple correlation between V1 axon activity and RGC-driven SC activity . Second , an Integrational Model assumes that V1 inputs can drive firing of SC neurons in addition to RGCs . Under these conditions , alignment is driven by weighted integrated activity of both RGCs and V1 inputs . Importantly , both models replicated with high fidelity visual map alignment as observed in wild type ( WT ) conditions , as well as that observed in transgenic and knockout mouse models . Interestingly , the models could be differentiated in silico , as they predicted different behaviors when the retinal drive component was weakened under transgenic model conditions . Based on these findings , we conclude that either correlational or integrational mechanisms may be utilized to achieve visual map alignment , suggest in vivo experiments that may be able to distinguish between the two , and speculate on the potential biological advantages of each . In the present study we develop two novel models of visual map alignment in the SC , specifically focusing on the projection from V1 to the SC , which develops during the second postnatal week in mice [23] . We assume that retinocollicular and retino-geniculo-cortical connections have been established during the first postnatal week [33–35] , i . e . topography has been established by RGCs in the SC , and V1 neurons are projecting axons from a topographically ordered region ( Fig 1 ) . Without a loss of generality , we utilize a common coordinate system based on retinal space to describe the topographic organization in the SC and V1 , which allows us to avoid ambiguity when map alignment is not one-to-one , for example in cases of mutant mice . That is , any location in the SC or V1 is a vector r → in two-dimensional Ω-space , normalized to unit size , associated with the corresponding retinal location ( see Fig 1 ) . More specifically , two two-dimensional maps Φ : r → R → r → S C and Ψ : r → R → r → V 1 , are representations of retinal inputs from location r → R into the SC ( r → S C ) and into V1 ( r → V 1 ) , correspondingly . In order to make direct comparisons to in vivo anatomical data , retinal space is projected onto the appropriate axes in both the SC and V1 . Specifically , the nasal-temporal ( N-T ) axis of the SC projects along the posterior-anterior ( P-A ) axis of the SC and is represented along the medial-lateral ( M-L ) axis of V1 , whereas the dorsal-ventral ( D-V ) axis of the retina projects along the L-M axis of the SC and is represented along the A-P axis of V1 . It is important to note that Ω-space is designed for WT mice , and all projection distortions caused by genetic manipulation are part of the models . In simulations , we study Ω-space in a regular grid where the retina , SC and V1 are represented as two-dimensional layers of 100x100 neurons . The models present new corticocollicular inputs in SC as a number of connections/synapses ( n ) between axons originating from a given cortical location r → l ∈ Ω , Ψ - 1 : r → V 1 → r → l and dendrites of SC neurons located in r → s ∈ Ω , Φ - 1 : r → S C → r → s . This number is a vector function n ( r → s , r → l ) , which is simulated as a four-dimensional array . To model the development of corticocollicular connections , we extended a stochastical model [22] that was developed to model the establishment of retinocollicular topography and showed best qualitative assessment against experimental data [4] . As in the original approach , the model minimizes total energy E in the V1-SC system , which is a function of connectivity . For both models we consider total energy as a sum of chemoaffinity energy ( Ea ) , axonal competition energy ( Ec ) and activity-dependent energy ( Eu ) : E = E a + E c + E u ( 1 ) The minimum of total energy E represents the most stable configuration of corticocollicular connections . We used a modified simulated annealing algorithm , described in [22] , to find the minimum of total energy ( see Methods section ) . Both models share the same representations for the chemoaffinity and competition energies , as described in [22 , 36] with minor modifications , but differ in the representation of activity-dependent energy . However , it is important to note that in both cases , the activity-dependent energy function in our model is different from those used for modeling retinocollicular development . Two descriptions for activity-dependent energy reflect different assumptions in model definitions . The first model is based on the assumption that new synapses of V1 axons onto SC neurons are significantly weaker than established synapses with RGCs; therefore this model considers only correlation between activity of SC neurons driven by retinal inputs and V1 neurons . We refer to this model as the “Correlational” model . In the second model , we assume that SC neurons integrate activity of both RGC and V1 inputs and that the effect of V1 inputs is not negligible , which we refer to as the “Integrational” model . All components for each model are described below . The major distinction between our Correlational and Integrational models is the ability of V1 inputs to drive SC neurons during the process of visual map alignment . In both models , retinal inputs have strong drive , which instructs V1 inputs to align with the retinal map . However , we noted that during some simulations of the Integrational model under any condition , transient clusters of V1 terminals could be observed in the SC . This anecdotal observation suggested that the Correlational and Integrational models might behave differently under conditions in which retinal drive were reduced during visual map alignment . To test this , we performed an in silico experiment in which we simulated visual map alignment under Isl2EphA3/EphA3 conditions , but with weakened ability of retinal input to drive SC neuron firing . In the Correlational model , weakening retinal drive is equal to a gradual decreasing of Eu , which we model by scaling down the γu parameter . For this analysis , we simulated the termination patterns of V1 axons projecting from the center of the L-M axis ( rxl = 0 . 5 ) under Isl2EphA3/EphA3 conditions . As expected , simulations in which retinal drive is similar to previous simulations ( e . g . γu = 10 ) , projections from V1 are bifurcated into two termination zones along the A-P axis ( Fig 5A ) . And , not surprisingly , when retinal drive is dramatically reduced ( e . g . γu = 0 . 1 ) , V1 axons terminate broadly along the A-P axis and only in a single termination zone ( Fig 5 ) . Interestingly , the transition was gradual between a single broad termination zone when retinal drive is weak to duplicated termination zones when retinal drive is high . This pattern of change is reminiscent of supercritical pitchfork bifurcation observed in dynamical systems [42] , and implies that two termination zones of cortical axons may be a result of bi-stability for individual axons . In simulations with the Integrational model , we modeled the weakening retinal drive by decreasing the factor ξu . Similar to observations from the Correlational model , simulations with high retinal drive ( e . g . ξu = 4 ) resulted in a bifurcation of V1 projections , while in those with weak retinal drive ( e . g . ξu = 0 . 04 ) a single termination zone was observed . Interestingly , the width of connection densities were not as wide under the latter conditions compared with simulations in the Correlational model , due to the ability of local V1 inputs to drive correlated activity in the Integrational model . Further , we observed that the transition from projections to a single termination zone when retinal drive is weak to duplicated termination zones when retinal drive is strong was much sharper for the Integrational model compared to the Correlational model . Taken together , these in silico experiments suggest that the models can be differentiated . Importantly , the diagrams generated by these experiments are not strictly classical bifurcation diagrams for dynamical systems [42] . Despite this , they reveal characteristic features of the total energy function , which affects the dynamics of connectivity patterns during development . Previously , it was noted that energy functions for competition [22] and for activity-dependence [36] have stable fix points which is an attribute of dynamical systems . Although our simulations should be considered only as optimization procedures , the minimum of energy function and corresponding peaks of connection density , are stable fix points of a dynamical system . The development of visual inputs in the SC occurs as a two step process: first , retinocollicular inputs establish topographic order in a manner dependent on molecular cues , correlated neuronal activity and competition during the first postnatal week; second , V1 inputs are instructed by RGCs to terminate in alignment with the retinocollicular map in a manner dependent on spontaneous activity . Thus , both of our computational models of visual map alignment focus on the establishment of topography by V1 neurons and are based on previous stochastic models that describe the development of retioncollicular topography [22 , 43] . Based on our previous work demonstrating the importance of correlated spontaneous activity during visual map alignment [23] , the most critical component of each model is the activity-dependent energy . For the Correlational model , activity-dependent alignment is achieved via Hebbian “fire together , wire together” rules [37] , wherein simple correlations between firing patterns of V1 axons and SC neurons are used . In contrast , the Integrational model considers the possibility that V1 inputs can drive SC neuron firing during visual map alignment . Although both of the models presented here are based on the stochastic models previously decribed to model retinocollicular development , it is critical to note that the process , and thus the performance , of the models is fundamentally different . When modeling retinocollicular development , the landscape of activity-dependent energy in any given region of the SC is essentially flat prior to simulation , due to the random connectivity . As such , during modeling , newly established connections form the energy profile , progressively developing energy wells in each region as dictated by the local density of RGC inputs , until a stable configuration is achieved . In contrast , when modeling the alignment of V1 inputs in the SC , the landscape of activity-dependent energy is in a pre-defined state by RGC inputs ( Eqs 4–7 , Supplementary S2 Fig ) . Indeed , these differences revealed themselves in the behavior of our models during in silico experiments performed in which we weakened retinal drive . Under these conditions , the Integrational model performs similar to modeling retinocollicular development , in that activity-dependent energy progressively decreases . Alternativley , in simulations with the Correlational model , which most closely resembles the retinocollicular mapping models on which our alignment models are based , activity-dependent energy can only decrease when retinal drive is sufficient . Understanding the nature of interactions between retinal and V1 inputs during visual map alignment is critical for developing a more robust model of this process . Importantly , both models replicate in vivo findings from WT and mutant animals , though with subtle differing degrees of fidelity . For example , while both models predict that V1 projections will bifurcate to align with a duplicated retinal map under Isl2EphA3/EphA3 conditions , neither predicts that the termination zone area of posterior-projecting V1 axons will be larger than anterior-projecting V1 axons , as we previously found [23] . This limitation may derive from innacurate estimation of the distance over which activity is properly correlated in the SC of Isl2EphA3/EphA3 mice . On one hand , since an entire azimuth representation is compressed into approximately half the SC , the relevant correlation distance may need to be halved as well . On the other hand , correlations between V1 and SC activities may actually be correlated over larger distances in Isl2EphA3/EphA3 mice , since two locations separated by a significant distance will fire with similar timing . Further , it remains unclear why only one of the two termination zones of V1 axons in Isl2EphA3/EphA3 does not refine as well as those observed in WT animals . It may be related to the differences in subtypes of RGCs that project to each domain [44] , given that distinct subtypes may participate differently during spontaneous retinal waves [45] . Elucidation and incorporation of these parameters of spontaneous activity into future models is necessary to overcome the limitations of our current models . Another limitation of these models is the underlying assumption that the representation of visual space in each region is symmetrical , which does not accurately reflect anatomical and functional data . Indeed , in several species , portions of the visual field are over-represented in the retina , V1 and the SC . In the mouse visual system , which these computational frameworks are meant to model , RGC density is highest centrally with a slight ventral bias ( i . e . upper visual field ) and decreases with eccentricity [46] . Similarly in both V1 and the SC , the central visual field is over-represented [20 , 47] . However , in other species , the asymmetric representation of visual space can differ between regions . For instance , in the macaque , lower visual field is over-represented in V1 [48] , while upper visual field is over-represented in the SC . How might alignment be achieved in such a situation and could our models account for this ? While we did not model this directly , possible distortions of symmetry , such as expansions and contractions , are included in the Φ and Ψ functions and , thus , are implicit to the model . However , the pliability of such distortions are limited by the competition energy component of our models , and , therefore , these models may not be ideal for investigating more drastic “sign reversals . ” Application of our models in these contexts may have to incorporate changes in competition energy . Another caveat is that our models deal strictly with development , where we model the pattern of spontaneous activity driving alignment to influence all regions of the retinotopic map uniformly . However , if non-uniform , experience-dependent changes drive differences in asymmetry between regions , then distinct mechanisms , and thus models , may be needed to describe this process . It is also critical to note that these models focus solely on the alignment of excitatory inputs from the retina and V1 onto excitatory principal cells of the superficial SC , ignoring putative connections with inhibitory populations . Indeed , the SC is densely packed with inhibitory neurons that modulate both the response to visual stimuli and the sensorimotor transformation to saccadic eye movements [49 , 50] . However , while GABAergic synapses are present in the SC during the period of retinocollicular map formation and visual map alignment [51] , they are weak and their role in either process is not clear . Regardless , inclusion of the development of connections between V1 neurons and inhibitory inputs in the SC , as well as lateral connections within the SC , would make for a more robust model . In order to distinguish the Correlational and Integrational models from one another , we leveraged the duplicated map of azimuth in Isl2EphA3/EphA3 to perform a modified bifurcation analysis . To do so , we performed simulations with both models in which we varied the parameter relating to the strength of retinal drive ( γu for Correlational and ξu for Integrational ) . For the Correlational model , we found that increasing retinal drive led to a gradual transition from a single , broad map to a sharply tuned duplicated map . The shape of this curve was strikingly similar to that of the supercritical pitchfork bifurcation associated with dynamical systems , albeit a static version rooted in a spatial domain . The behavior of the Integrational model to increasing retinal drive under Isl2EphA3/EphA3 conditions was strikingly different . Here , the transition from single to duplicated map was sharp , and suggestive of multistability within the system . Interestingly , we previously found that the retinocollicular map in heterozygous Isl2EphA3/+ mice can be organized in one of three possible ways [43] , reminiscent of the either/or prediction of the Integrational model observed here . Together , these findings suggest the possibility that the development of topography in general may observe the rules of multistable systems . In general the Integrational model is more robust to variation of retinal input strength . It shows smaller variance in alignment accuracy to a broader range of retinal input strengths ( Fig 5B ) , which may be considered as potential biological advantage . In contrast weakening retinal inputs below some threshold gradually distorts topographic map alignment in the Correlational model ( Fig 5A ) . Therefore , the in silico tests performed here on simplified computational models of a complex biological process are severely limited in their predicitive powers . Further , our data do not favor conclusively either the Correlational or Integrational model and more data are needed to determine if either is a valid representation of in vivo processes . Given that both models are able to replicate the limited in vivo data from mutant animals , the question of which is utilized remains unresolved . An exploration of the biological advantages of each may point towards which mechanism might be utilized . On one hand , the Correlational model might be energetically favorable compared to the Integrational model , since developing V1 inputs do not need to invest in expressing the full complement of pre-synaptic machinery at each transient early contact . Additionally , one might imagine that use of a correlational mechanism might lead to faster refinement , again making it more energetically favorable . However , our in silico modeling does not indicate that the Correlational model resolves to a steady state faster than the Integrational model ( S4 Fig ) , though this is not necessarily representative of the speed of refinement in vivo . On the other hand , the energy investment required to execute the Integrational model may confer other advantages to the development of visual circuitry in the SC . For instance , multiple subtypes of visual neurons are found in the SC [52] , and the ability of V1 inputs to contribute to SC neuron firing during development may help to ensure that they integrate into the appropriate sub-circuit . In support of this possibility , recent evidence suggests that fine-grain topography in the SC may be sacrificed to allow for the establishment of microdomains of neurons tuned to the same aspect of the visual scene [53] . However , critical aspects of the nature of the developing circuitry in the SC remain unknown , preventing us from favoring one model over the other . One key piece of evidence that might distinguish these models relates to the distinction between the two formulations: namely , whether V1 inputs can drive SC neuron firing during development . Electron microscopy studies indicate that V1 inputs form synaptic contacts onto SC cells during development , which mature over time [54] . However , to our knowledge no study has explored the physiological characteristics of corticocollicular inputs throughout development , perhaps due to the circuitous route from V1 to the SC preventing the isolation of the preserved tract in a slice . A potential alternative may be to leverage the power of optogenetics to expresses light-excitable channels in V1 during development . Slices could then be made of the SC and the termials of corticocollicular afferents stimulated while recoding from SC neurons . Understanding the potency of V1 inputs over the course of visual map alignment would provide substantial insight to the mechanisms underlying this critical event , as well as inform the development of more accurate models of visual map alignment . Here we have described two novel computational models of the development of alignment between retinal inputs and those from V1 in the SC . The major difference between the models relates to the mechanism of activity-dependent refinement . The models behave differently in in silico experiments in which retinal drive the SC is weakened during simulations , suggesting differences in the nature of map alignment depending on the mechanism of activity-dependent refinement . Overall , the Correlational and Integrational frameworks presented here accurately model known aspects of visual map alignment , but further experimentation is needed to determine which of the activity-dependent mechanisms is utilized in vivo . A modified simulated annealing algorithm [22] was used to find the minimum of energy function ( Eq 1 ) . For each neuron in the 100x100 grid , the algorithm produces 15 , 000 steps , 150 , 000 , 000 steps in total . At each step , the algorithm adds one connection and removes another one randomly . The probability to accept or reject addition or removal of a connection is modeled by the sigmoid function from changing in in total energy ( ΔE ) as followed: P = 1 1 + e 4 Δ E ( 8 ) Initially connections are randomly distributed such that each neuron receives 50 connections on average . We also tested our models under two extreme initial conditions: totally disconnected and all-to-all connected networks . No variation in results were found under either condition . We confirmed that 150 , 000 , 000 steps are enough by performing a simulation when number of steps was doubled . Neither model , under any parameter set , achieved better convergence with double the number of steps ( S4 Fig ) . Therefore , we conclude that 150 , 000 , 000 steps allow our algorithms to reach steady-state energy minimums . The modified simulated annealing algorithm was implemented in Cython computer language with Python wrapper . We used the Python numerical library ( numpy ) and GNU scientific library ( gsl ) for random number generation , matrix manipulations and operation vectorization . One optimization procedure for the Correlational model requires on average 10 hours of single processor time , while the Integrational model needs approximately 16 hours of single processor time . Source code on the model and required scripts will be made public available via ModelDB website after publication ( https://senselab . med . yale . edu/ModelDB/showModel . cshtml ? model=195658 ) . We studied the robustness of parameters to variation , as well as general model behavior , in a wide range of parameter space , which was estimated to require around 1 . 3 years of simulation time on four cores of a desktop computer . In this study , we exploited embarrassingly parallel computing on 1344 cores of a high performance Cray XE6/XK7 cluster to speed up computations to one week . A connectivity four-dimensional array ( n ) was sampled to verify one dimension mapping . To obtain connectivity density , standard Silverman method [55] implemented in the Python scientific library ( scipy ) was used .
In order to process sensory stimuli , precise connections must be established between sensory neurons during development . In the visual system , many connections are organized topographically , such that neighboring neurons monitor adjacent regions of space . In the superior colliculus ( SC ) , converging topographic inputs must be aligned with one another to facilitate integration and preserve the spatial order of the visual scene . In this paper , we propose two novel computational models to describe the alignment of visual inputs in the SC . We demonstrate that both models are able to replicate experimental data obtained from wild type and mutant animals . Interestingly , each model performed differently in response to hypothetical experiments , suggesting they could be differentiated empirically . Thus , we put forth testable models of visual map alignment in the SC and propose experiments to determine which may be used during development .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[ "medicine", "and", "health", "sciences", "topographic", "maps", "ocular", "anatomy", "social", "sciences", "neuroscience", "simulation", "and", "modeling", "ganglion", "cells", "nerve", "fibers", "vision", "neuronal", "dendrites", "research", "and", "analysis", "methods", "geography", "animal", "cells", "axons", "cartography", "cellular", "neuroscience", "psychology", "retina", "cell", "biology", "anatomy", "earth", "sciences", "neurons", "biology", "and", "life", "sciences", "cellular", "types", "sensory", "perception", "ocular", "system", "afferent", "neurons", "retinal", "ganglion", "cells" ]
2016
Novel Models of Visual Topographic Map Alignment in the Superior Colliculus
Germ line specification is a crucial step in the life cycle of all organisms . For sexual plant reproduction , the megaspore mother cell ( MMC ) is of crucial importance: it marks the first cell of the plant “germline” lineage that gets committed to undergo meiosis . One of the meiotic products , the functional megaspore , subsequently gives rise to the haploid , multicellular female gametophyte that harbours the female gametes . The MMC is formed by selection and differentiation of a single somatic , sub-epidermal cell in the ovule . The transcriptional network underlying MMC specification and differentiation is largely unknown . We provide the first transcriptome analysis of an MMC using the model plant Arabidopsis thaliana with a combination of laser-assisted microdissection and microarray hybridizations . Statistical analyses identified an over-representation of translational regulation control pathways and a significant enrichment of DEAD/DEAH-box helicases in the MMC transcriptome , paralleling important features of the animal germline . Analysis of two independent T-DNA insertion lines suggests an important role of an enriched helicase , MNEME ( MEM ) , in MMC differentiation and the restriction of the germline fate to only one cell per ovule primordium . In heterozygous mem mutants , additional enlarged MMC-like cells , which sometimes initiate female gametophyte development , were observed at higher frequencies than in the wild type . This closely resembles the phenotype of mutants affected in the small RNA and DNA-methylation pathways important for epigenetic regulation . Importantly , the mem phenotype shows features of apospory , as female gametophytes initiate from two non-sister cells in these mutants . Moreover , in mem gametophytic nuclei , both higher order chromatin structure and the distribution of LIKE HETEROCHROMATIN PROTEIN1 were affected , indicating epigenetic perturbations . In summary , the MMC transcriptome sets the stage for future functional characterization as illustrated by the identification of MEM , a novel gene involved in the restriction of germline fate . The life cycle of flowering plants alternates between a diploid sporophytic and a haploid gametophytic phase . In contrast to animals , which form gametes directly by meitotic division from a diploid germline , plants form female and male spores by meiotic division during megasporogenesis and microsporogenesis , respectively . Subsequently , during mega- and microgametogenesis , the spores develop by mitotic division and cell differentiation into the female and male gametophytes , respectively . The two morphologically distinct gametophytes develop within specialized reproductive structures in the female and male organs of the flower , the ovules and the anthers . The multicellular , haploid gametophytes ultimately give rise to the gametes . In the ovule , the archespore , which arises from a sub-epidermal cell , is the first cell of the reproductive lineage ( “germline” ) [1] . In the model plant Arabidopsis thaliana , the archespore differentiates directly into the megaspore mother cell ( MMC ) , which is committed to undergo meiosis and gives rise to a tetrad of haploid megaspores . In Arabidopsis , as in most species , only one of these , the functional megaspore ( FMS ) , survives while the others degenerate . The FMS occupies a defined position within the ovule suggesting that position is important for its determination and survival [2] . The importance of signaling from sporophytic ovule tissues for differentiation of the MMC and selection of the FMS has been discussed [3] , [4] . The FMS develops into the haploid embryo sac ( female gametophyte ) through three rounds of mitosis followed by cellularization , typically forming a seven-celled embryo sac , including two gametes ( the haploid egg cell and the homo-diploid central cell ) , two synergids , and three antipodals [2] , [5] , [6] . Double fertilization of the female gametes by one sperm cell each initiates seed development with the fertilized egg cell giving rise to the diploid embryo and the fertilized central cell to the triploid endosperm . Some plant species can produce asexual seeds through a process known as apomixis . To initiate apomictic reproduction , an unreduced embryo sac is formed either from a sporophytic ( “somatic” ) cell of the ovule ( apospory ) or from a MMC that omits or aborts meiosis ( diplospory ) [4] . The egg cell subsequently develops into an embryo without fertilization ( parthenogenesis ) . So far , little is known about the genes underlying the developmental program of MMC specification and differentiation . Analysis of the molecular basis underlying early reproductive development is particularly difficult due to the low abundance and inaccessibility of the relevant cells . Expression in the Arabidopsis MMC has so far only been shown for a few meiotic genes [7]–[11] and SPOROCYTLESS/NOZZLE ( SPL/NZZ ) . SPL/NZZ is a plant-specific protein related to MADS-domain transcription factors , which plays an important role for the initiation of sporogenesis [12]–[15] . In spl/nzz mutants the nucellus is reduced and the archespore usually fails to undergo differentiation to form a MMC [12] , [14] . Apart from the MMC , SPL/NZZ is expressed in sporophytic tissues during early stages of ovule development and in flowers , leaves , seedlings , and stems [12] , [14] , indicating broader functions in plant development . Interestingly , SPL/NZZ modulates the expression of YUCCA2 and YUCCA6 , genes that function in auxin biosynthesis , to regulate lateral organ development [16] . Auxin has been proposed to play an important role for gametophyte development in Arabidopsis [17] . It was suggested that an auxin gradient established in the developing embryo sac influences cell type specification [17] . Recently , small RNAs were shown to be involved in regulating cell fate determination by introducing epigenetic modifications at the DNA or chromatin level . ARGONAUTE ( AGO ) proteins are involved in this mechanism by regulating mRNAs during miRNA- or siRNA-guided post-transcriptional gene silencing . It has been demonstrated that Arabidopsis AGO9 is required to restrict the differentiation of sub-epidermal cells into MMCs in pre-meiotic ovules [18] . In contrast to wild-type plants , more than one enlarged sub-epidermal cell was frequently observed in ago9 mutants . Such a phenotype has so far been observed only in a small number of mutants in maize and rice [19]–[21] . In ago9 mutants female gametophyte development from the MMC and a second sub-epidermal , sporophytic cell was observed , resembling features of apospory [18] . Enriched expression of AGO9 , as well AGO1 , AGO2 , AGO5 , and AGO8 , was also found in the Arabidopsis egg cell , supporting a role of the RNA-based silencing mechanism in the female gametophyte [22] . However , a possible role of these genes in the differentiation of female reproductive structures remains to be unveiled . To obtain more insights into the genetic and molecular bases underlying megasporogenesis in Arabidopsis , we used a combination of laser-assisted microdissection ( LAM ) and Affymetrix ATH1 GeneCHIP profiling [22] to analyze the transcriptome of the MMC and the surrounding sporophytic tissue . Statistical analyses of gene expression identified genes and functions significantly enriched in the MMC as compared to other cell types and tissues . In particular , translational regulation was identified as an important feature . Also the molecular function of ATP-dependent helicase activity was enriched . We found that mutations in a RNA-helicase gene , named MNEME ( MEM ) after the Greek muse of memory [23] , lead to defects during megasporogenesis and mega-gametogenesis and arrest at early embryonic stages . In particular , in ovules of heterozygous mem/MEM plants , more than one enlarged sup-epidermal cell developed instead of the single MMC . Furthermore , two gametophytic cells instead of one were frequently observed , a phenotype similar to that described in Arabidopsis ago9 mutants and mutants in the DNA-methylation pathway in maize , which affect epigenetic regulation [18] , [19] . Interestingly , we observed altered epigenetic modifications in gametophytic nuclei of mem mutants . In summary , this study describes the transcriptional network of the Arabidopsis MMC suggesting a role for RNA processing and translational control at early stages of sexual reproduction and revealing an important function of a novel RNA-helicase , MEM , in the restriction of the germline lineage to only one cell per ovule . To investigate the transcriptome of the Arabidopsis MMC and the surrounding sporophytic tissue of the ovule ( sporo_nucellus ) , we combined LAM with microarrays ( Figure 1 ) . MMCs and the surrounding nucellar tissue were isolated separately by LAM ( Figure 1A–E ) . Because of the small size of the ovules at this young developmental stage and the structural limitations of dried sections required for LAM , limited cross-contamination of the samples could not completely be avoided . Between 560 and 930 sections were pooled per sample . The extracted total RNA was subjected to two rounds of linear amplifications , labeled , and hybridized to Affymetrix ATH1 arrays . As the default algorithm for the generation of present and absent calls performs poorly on data from amplified samples [22] , [24] , an alternative algorithm , AtPANP , was adapted and applied to calculate present/absent p values [22] . This algorithm has been shown to outperform the default algorithm for the generation of present and absent calls in terms of accuracy and precision on data from cell-type-specific LAM samples [22] . However , the AtPANP algorithm was based on non-matching probes on the ATH1-array to determine the background signal in accordance with Arabidopsis TAIR7 genome annotation [22] . Therefore , we updated the array annotation and the negative probe selection based on the TAIR9 genome release ( http://www . arabidopsis . org; http://brainarray . mbni . med . umich . edu/Brainarray/default . asp ) . The re-annotated array targets 21 , 504 genes , which is 75% of all genes annotated in the genome ( 64% including putative pseudogenes and transposable element genes ) . A total of 6 , 650 genes were found to be expressed in the MMC ( i . e . , significantly detected above background , hereafter referred to as present/P ) in three out of four arrays ( Figure 1G ) , and an additional 2 , 465 genes were detected as present in two MMC replicates ( referred to as marginal/M; Table S1 ) . Together , 9 , 115 genes showed putative expression in the MMC , which is a bit more than the 8 , 850 genes identified to be expressed in the mature female gametophyte [22] . For the sporophytic nucellar tissue , 10 , 081 genes were detected as P , while an additional 1 , 442 genes were M ( Figure 1H ) . This is in agreement with the expectation that more genes are expressed in the heterogeneous nucellus tissue as compared to a single cell type . We validated our dataset using different independent approaches: ( I ) expression analysis using RNA in situ hybridization , ( II ) analysis of reporter gene expression in transgenic plant lines carrying putative cis-regulatory elements driving the E . coli uidA gene encoding ß-glucuronidase ( GUS ) , ( III ) investigation of enhancer trap lines , and ( IV ) comparison to the literature . For 12 genes we could confirm predominant or exclusive expression in the MMC within developing ovules ( Figure 2A–L , Table S2 ) . In addition , PUMILIO12 ( PUM12 ) and ATP-BINDING CASSETTE B19 ( ABCB19 ) were confirmed to be expressed in the nucellus tissue , using lines expressing GUS and GFP as reporters , respectively ( Figure 2M and 2N ) [22] , [25] . To date only five genes have been described to be expressed in the MMC: the meiotic genes DISRUPTION OF MEIOTIC CONTROL1 ( DMC1 ) , SOLO DANCERS ( SDS ) , DYAD/SWITCH1 ( SWI1 ) , MULTIPOLAR SPINDLE1 ( MPS1 ) , and SPL/NZZ [7]–[12] , [14] . DMC1 , SDS , and DYAD/SWI are important for homologous recombination , sister chromatid cohesion , synapsis , and bivalent formation during meiotic prophase I , and MPS1 is a gene involved in spindle organization in meiocytes [7]–[11] , [26] . Except for MPS1 these genes were present in our MMC dataset; SWI1 could not be analyzed as it is not represented on the ATH1 array . In addition , NZZ/SPL , WUSCHEL ( WUS ) , WINDHOSE1 ( WIH1 ) , WIH2 , and AGO9 were present in the dataset from surrounding nucellus tissue , consistent with the literature [12]–[15] , [18] , [27] , [28] . These independent validations provide strong evidence for the accuracy of the expression datasets . The MMC undergoes meiosis and eventually gives rise to the haploid embryo sac . Consequently , genes encoding proteins important for meiosis are expected to be expressed in this cell type . However , although mainly developmental stages starting just before meiosis and ending at prophase of meiosis I were sampled , it should be noted that development from archesporial cell to mature MMC is a continuous process , such that the samples contain multiple developmental stages . We found that five additional genes active during female meiosis without known functions in somatic tissues were expressed in the MMC ( P or M ) . MUTL-HOMOLOGUE1 ( MLH1 ) , MUTS HOMOLOG4 ( MSH4 ) , RECOMBINATION8/SYNAPTIC1 ( REC8/SYN1 ) , and PARTING DANCERS1 ( PTD1 ) function during prophase of meiosis I , and TARDY ASYNCHRONOUS MEIOSIS ( TAM ) controls the transitions between prophase and the first meiotic division as well as between meiosis I and meiosis II [29]–[35] . The identification of genes with important roles during prophase of meiosis I as expressed in our MMC dataset is consistent with the developmental stages covered in our sampling . To identify new genes with a role in MMC specification and differentiation and to obtain novel insights into the transcriptional basis and molecular mechanisms underlying megasporogenesis , we analyzed our transcriptome datasets of the MMC and the surrounding sporophytic nucellus tissues by hierarchical clustering and an analysis of gene enrichment . In particular we compared the transcriptome of the MMC with the transcriptomes of ( I ) the surrounding nucellar tissue , ( II ) the cells of the mature female gametophyte , and ( III ) an additional 70 gametophytic and sporophytic cell types and tissues from a tissue atlas ( as described in [22] plus additional samples , see Methods ) . In addition , we compared ( IV ) the expression in the nucellar tissue with the tissue atlas . The MMC develops from the selected archespore , which is closely related in cell lineage to the surrounding tissue . It can thus be assumed that they share , to a certain extent , similar gene expression patterns . Nevertheless , the MMC is morphologically and functionally distinct from the surrounding cells . The determination of the MMC can be viewed as the delineation of a committed cell lineage that corresponds to the animal germline . Thus , the MMC and the egg cell of the mature embryo sac are the first and the last stage of the plant germline lineage . To relate the transcriptome of the MMC and surrounding tissue to the recently investigated transcriptomes of cell types of the mature female gametophyte ( egg cell , central cell , and synergids ) and to the male gametophyte ( pollen ) , we applied hierarchical agglomerative sample clustering . Cell-type- and tissue-specific datasets cluster together , indicating good reproducibility of the data ( Figure 1F ) . All datasets from the female germline lineage and the sporophytic nucellus tissue cluster closer together and group separately from pollen . In addition , the MMC shares more characteristics with the sporophytic nucellar tissue than with gametophytic cells , in agreement with their close relationship with respect to cell lineage . The mature female gametophyte is separated from the MMC by only a few cell cycles . Potentially , they share expression of a subset of genes important for the identity of the germline lineage . However , other genes will be important either for differentiation of the female gametophyte and the gametes or for MMC specification and megasporogenesis , the transition from the sporophytic to the gametophytic phase . Thus , a comparison of transcriptional profiles from these two developmental stages can provide important insights into the molecular basis of cell specification and cell fate acquisition . We found 2 , 451 genes differentially expressed in the MMC and the cells of the mature female gametophyte ( egg cell , central cell , and synergid cells [22] ) at a false discovery rate below 0 . 05 ( Figure 3 ) [36] . We now focused on 796 genes with significantly enriched expression in the MMC in all three contrasts . A functional classification of these genes identified translational regulation control pathways and functions related to ribosome biogenesis and structure as highly over-represented ( p value <0 . 01 , Table 1 , Table S3 ) . In addition , mainly different metabolic functions and transport processes , particularly for the transport of different ions , were significantly enriched ( p value <0 . 01 , Table 1 , Table S3 ) , but also the molecular functions “structural constituent of chromatin” and “ATP-dependent helicase activity” ( Table S3 ) . Interestingly , genes annotated in the gene ontology ( GO ) term “embryonic development” also were identified as near significantly enriched ( p value = 0 . 011 , Table 1 , Figure S1 ) . To obtain more insight into the molecular mechanisms underlying the development of the MMC in contrast to the mature gametophyte , we analyzed our dataset for enrichment of protein family ( PFAM ) domains and gene families . Three gene families , the “cytoplasmic ribosomal gene family” , the “eukaryotic initiation factor family” , and the “proton pump interactor ( PPI ) ” gene family , as well as 34 PFAM domains , including 10 ribosomal protein domains , were identified as significantly enriched ( Table S4 , Fisher's exact test , p value <0 . 01 ) . In addition , the “HMG ( high mobility group ) box” , the “eIF-6 family” , and the “DEAD/DEAH-box helicases” belonged to the protein domains significantly over-represented ( Table S4 ) . Together , this analysis suggests that translational regulation is a major feature underlying MMC specification , paralleling an important feature of the animal germline ( reviewed in [37] ) . In addition , specific RNA-helicases play crucial roles in germline development in animals ( reviewed in [37] ) . Interestingly , DEAD/DEAH-box helicases were specifically enriched in the MMC as compared to the cells of the embryo sac . With few exceptions , these genes were also more highly expressed in the MMC than in pollen or sperm ( Figure S2 ) , supporting their importance for megasporogenesis as compared to gamete differentiation . The comprehensive tissue atlas allowed us to identify genes with preferential expression in the MMC and the surrounding nucellus tissue . In the nucellus tissue 134 genes were significantly enriched as compared to the tissue atlas not including the MMC ( adjusted p value <0 . 01 [38] , Table S5 ) . Functional gene classification identified the molecular functions “acid phosphatase activity” , “protein serine/threonine phosphatase activity” , “structural constituent of ribosome” , “RNA binding” , and the biological process “oligopeptide transport” as upregulated in nucellus tissue ( Table S6 ) . One of the oligopeptide transporters significantly enriched in the nucellus , OLIGOPEPTIDE TRANSPORTER9 ( OPT9 ) , was previously described as highly expressed in microspores and bicellular pollen [39] , suggesting a role during reproductive development . Including the MMC in the analysis , 49 genes were significantly enriched in nucellus tissue as compared to the tissue atlas ( Figure S3 , adjusted p value <0 . 01 [38] ) . Analysis of this set of genes revealed the gene families “cytochrome P450” and “monolignol biosynthesis” as significantly enriched ( Fisher's exact test , p-value <0 . 01 ) . In the MMC , 82 genes were significantly enriched as compared to the tissue atlas ( excluding sporo_nucellus , Figure S4 , Table S7 , adjusted p value <0 . 01 [38] ) . Based on these genes , functional gene classification suggests roles for “tyrosine biosynthesis” , “translation” , “acid phosphatase activity” , and “ATP-dependent RNA-helicase activity” for MMC differentiation ( Table S8 ) . When including the nucellus samples in the tissue atlas , still 13 genes were significantly enriched in the MMC , suggesting that those genes might play specific roles during MMC specification and differentiation ( Figure 4 ) . Among these 13 genes is SDS , involved in homologous chromosome pairing during meiotic prophase I [8] . AT2G20390 and AT4G38390 encode unknown proteins . AT3G07140 encodes a GPI-transamidase GPI16 subunit protein , with a putative function in adding GPI anchors to proteins linked to the cell surface . AT2G30940 ( Figure 2B ) , encoding a protein tyrosine kinase , and AT1G11270 ( Figure 2G ) , coding for a Cyclin-like F-box protein , are enriched in the MMC , potentially with functions in inter- or intra-cellular signaling and cell cycle regulation , respectively . AT2G39240 encodes an RNA polymerase I transcription factor , and AT1G61990 encodes a protein related to mitochondrial transcription factors . Arabidopsis PUMILIO23 ( AtPUM23 ) is an RNA-binding protein located in the nucleus [40] . AT1G15710 is a prehenate dehydrogenase potentially involved in tyrosin biosynthesis . Also YUCCA2 , a gene involved in auxin biosynthesis and AT1G29440 , encoding an auxin-responsive gene related to SMALL AUXIN UPREGULATED68 ( SAUR68 ) , are predominantly expressed in the MMC , supporting the importance of auxin signaling for early stages of reproductive development . In addition , an ATP-dependent RNA-helicase , AT5G39840 , which we named MNEME ( MEM ) after the Greek muse of memory [23] , is amongst these 13 genes specifically enriched in the MMC . Although only two of the genes are annotated as unknown proteins , none of these genes have been functionally characterized in detail so far , except for YUCCA2 and SDS . This might be due to their rather specific expression in a rare cell type . Interestingly , we discovered the expression of DEAD/DEAH-box helicases as well as genes with functions related to translation also in the comparison of the MMC transcriptome against the tissue atlas , supporting the evidence that these are dominant features of the MMC . Our transcriptional dataset suggests the importance of DEAD/DEAH-box helicases during early developmental stages of the female reproductive lineage . One of the helicases , MEM , is encoded by one of the genes preferentially expressed in the MMC ( Figure 4 , Figure S5 , Table S9 ) , suggesting for a role in MMC specification and differentiation . To study the potential function of MEM during reproductive development , we analyzed two independent T-DNA insertion lines , mem-1 and mem-2 , inserted in the first exon and in the 3′UTR 50 bp downstream of the stop codon , respectively . Indeed , heterozygous mem-1 and mem-2 plants showed fertility defects with 40% ( N = 563 ) and 33% ( N = 627 ) of arrested ovules or aborted seeds , respectively . Transmission efficiency of the mutant alleles was analyzed in reciprocal crosses of heterozygous mem-1/MEM or mem-2/MEM plants with the wild type and showed a reduced transmission through the female but not the male gametophyte ( Table 2 , [41] ) . This indicates that mem is a female gametophytic mutant . Indeed , only 4% ( N = 214 ) of seeds were arrested or developmentally delayed after pollinating wild-type flowers with pollen of mem-2/MEM plants , in contrast to 26% ( N = 142 ) of arrested seeds observed in siliques of heterozygous mem-2/MEM plants pollinated with wild-type pollen . Although mem is transmitted through both male and female gametophytes , homozygous plants have not been identified , indicating that they are either not viable or only survive at very low frequency , implying embryo lethality . As MEM is predominantly expressed during early stages of reproduction , we first studied megasporogenesis in plants carrying a mutant mem-1 or mem-2 allele in more detail . In wild-type plants , one archespore becomes selected in the sub-epidermal layer of the ovule and differentiates into a MMC . However , in 6% ( N = 141 ) of wild-type ovules , we observed initiation of two MMCs before meiosis , in agreement with the 5%–6% reported previously [2] , [18] . In ovules of mem-1/MEM and mem-2/MEM plants , however , 18% ( N = 275 ) and 22% ( N = 171 ) form either more than one enlarged sub-epidermal cell with characteristics of the MMC or an MMC with adjacent abnormal cells ( Figure 5E–I ) . In addition , at the onset of megagametogenesis , instead of one FMS and the remnants of the three degenerated megaspores , a second gametophytic cell was often observed , or the FMS was flanked by abnormal cells ( Figure 5J–L ) . To analyze whether these cells are differentiated gametophytic cells , we used the ANTIKEVORKIAN ( AKV ) cell-identity reporter previously shown to mark nuclei during megagametogenesis prior to cellularization [42] , [43] . In wild-type plants , this reporter was expressed in the FMS , but not in the degenerated megaspores ( Figure 5N ) . Occasionally , a very weak staining was observed in the degenerated megaspore adjacent to the differentiating functional megaspore ( ∼10% , N = 87 ) . In heterozygous mem-1 or mem-2 mutant plants , however , nuclei of two adjacent cells were often marked as gametophytic ( Figure 5O ) , as observed in ∼34% ( N = 134 ) and ∼33% ( N = 123 ) of analyzed ovules , respectively . An increased number of gametophytic nuclei were often observed during early stages of megagametogenesis , likely derived from two female gametophytes ( Figure 5M , Figure S6I–K ) . In addition , the shape of the gametophyte ( Figure S6A and S6B ) and the positioning of gametophytic cells in the ovule , or nuclei in the gametophyte , were affected ( Figure S6B–D , H , L ) . Therefore , a second gametophytic cell likely initiated gametophyte development resulting in an unusually positioned developing embryo sac ( Figure S6C and S6D ) . To determine whether the two FMS-like cells give rise to two mature embryo sacs in one ovule and whether megagametogenesis in mutant ovules could give rise to normally developed mature gametopyhtes , flowers of heterozygous mem-1/MEM and mem-2/MEM mutant plants were analyzed 3 days after emasculation . Although a second normal mature embryo sac was never observed , in mem-1/MEM at least 44% ( N = 177 ) showed mutant phenotypes in the mature female gametophyte . An additional 8% of all ovules could not clearly be classified . In the most abundant mutant class , the female gametophyte harbored a normal structure with all cell types except that the polar nuclei in the central cell did not fuse ( 23% of total ovules analyzed , Figure 6D ) . In the second most abundant mutant class , gametophytes were abnormally narrow with fused polar nuclei ( 13% ) ( Figure 6B ) . Other phenotypes included untypical positioning of the putative egg cell or other cells ( ∼3% ) and absent gametophytes ( ∼5% , Figure 6C ) . Similar phenotypes were observed in ovules of mem-2/MEM plants ( Figure S6E–G ) . In summary , heterozygous plants carrying a mutant mem-1 or mem-2 allele ( I ) are affected during megasporogenesis , particularly in the selection of the MMC and FMS , indicating haplo-insuffiency of the MEM gene , and ( II ) have a gametophytically controlled defect in the development of the embryo sac and seed . Double fertilization initiated seed development but , in comparison to the wild type , developmental progression was delayed in mem-1 and mem-2 derived seeds , which finally arrested at different early embryonic stages ( from one-cell to mid-globular stage , Figure 6E–G , Figure S7 ) . To gain more insights into the embryonic function of MEM we studied embryogenesis in mem-1/MEM mutant plants in more detail . At 2 days after pollination ( DAP ) , when the majority of wild-type embryos had undergone two or three cell divisions ( two- to four-cell embryo proper ) , the majority of mutant embryos had divided only once or not at all ( Figure S7A and S7C–E ) . Endosperm development was delayed in comparison to the wild type ( Figure S7C–E ) . At 3 DAP , a proportion of unfertilized ovules and seeds ( likely arrested around the zygote stage ) had started degeneration and collapsed ( Figure 6E , Figure S7B and S7F ) . Only about 10% of embryos with a developmental delay developed into a two- or four-cell embryo , while the majority of wild-type embryos had reached the octant or early globular stage ( Figure S7B ) . At 4 DAP , the majority of arrested seeds had collapsed and only infrequently , in about 1% of all ovules and seeds ( N = 149 ) , arrest at the mid-globular stage was observed ( Figure 6G ) . As the MEM gene was identified as significantly enriched in the MMC , this finding suggests that either ( I ) carryover of stable transcripts present in the MMC of a heterozygous plant is enough to sustain later stages until early seed development , ( II ) transcripts present in the selected MMC prevent early arrest , but de novo transcription is required at later developmental stages , or ( III ) during early stages of reproduction MEM determines the developmental fate of the gametophyte ( e . g . , by setting an epigenetic state that is interpreted only later in development ) . Alternatively , ( IV ) other ATP-dependent RNA-helicases enriched in the MMC might act redundantly during megasporogenesis . As a first approach to investigate these possibilities , we studied the transcript abundance and expression during reproductive development in more detail . By array analysis of the cells of the mature female gametophyte [22] , expression of MEM was neither observed in the gametes ( egg , central cell , sperm ) or the synergids ( Figure 4 , Table S9 ) , nor in the transcriptomes of embryo and endosperm , except for marginal expression in one globular embryo sample ( embryo_proper_globEmb; P in one of two replicates; Table S9 ) . To confirm the transcriptome data we analyzed the expression of MEM by RNA in situ hybridization during megasporogenesis on buds harboring mature female gametophytes , and during early seed development . Highest expression was detected during megasporogenesis , in the archespore ( Figure S5A ) , the MMC ( Figure 5A , Figure S5B ) , and the FMS ( Figure 5B and 5C ) . Weaker signals were detected in the sporophytic tissues of the developing ovule , while no specific signals were detected in the mature gametophyte or the sense controls ( unpublished data and , Figure S5C ) . These data independently confirm the accuracy of our transcriptome dataset and show that MEM is highly expressed in the MMC and FMS , while it is either absent or strongly down-regulated in the mature female gametophyte . During early stages of seed development , a weak signal was detected in the endosperm , while in embryos signals were rarely observed and hardly distinguishable from background , likely due to very low transcript levels at the detection limit ( Figure S5D–F ) . The specific enrichment of MEM expression during megasporogenesis together with the developmental arrests of the embryo sac or early embryo suggested that MEM might either directly or indirectly determine molecular responses that occur later in development . In plants as well as in animals , epigenetic modifications based on histone modifications and DNA-methylation play important roles in regulating gene expression . Such epigenetic marks determine the chromatin structure and , thus , the transcriptional state of a cell ( reviewed by [44] , [45] ) . The LIKE HETEROCHROMATIN PROTEIN1/TERMINAL FLOWER2 ( LHP1/TFL2 ) protein has previously been shown to associate with euchromatic repressive marks [42] , [46] . It binds to H3K27me3 methylation marks established by Polycomb group proteins in euchromatic regions and is of functional importance for the interpretation of these marks [47]–[49] . In the Arabidopsis mature female gametophyte , LHP1 binds repressive chromatin marks in the nuclei of the egg cell and the synergids and , to a much lower extent , the central cell [42] . In the mature gametophyte before fusion of the polar nuclei it is equally expressed in the two unfused polar nuclei and the egg cell and synergid nuclei [42] . To study potential changes in the establishment of this epigenetic mark in gametophytes of heterozygous mem-1 and mem-2 mutant plants , we analyzed the distribution of LHP1 in mature embryo sacs . We crossed plants carrying a LHP1/TFL2 construct in translational fusion to GFP ( pTFL2:TFL2-GFP; [46] ) to heterozygous mem-1/MEM and mem-2/MEM plants . Plants of the F2 generation from these crosses were selected for the TFL2-GFP marker and the presence of the mem-1 or mem-2 allele . Wild-type ovules showed strong signals in the nuclei of synergids and the egg cell , and a weaker signal in the central cell nucleus as recently described ( [42]; Figure 7A ) . We focused on the most abundant mutant class with two unfused polar nuclei , as such embryo sacs were well distinguishable from the wild type while being morphologically closest to the wild type . No GFP signal was detected in mem-1 and mem-2 mutant female gametophytes with unfused polar nuclei ( N = 20 and N = 10 , respectively ) and only rarely in other mutant classes ( Figure 7B , C , G ) . This finding suggests that MEM is directly or indirectly involved either in the proper establishment of euchromatic repressive marks in the germline or their interpretation by regulation of LHP1 . Changes in the epigenetic setup of a cell might also involve changes in chromatin structure . The H2B-YFP marker under the control of the AKV promoter reflects some aspects of chromatin structure during megagametogenesis . In ovules of plants carrying a mutant mem-1 or mem-2 allele , we frequently observed a different distribution of H2B-YFP as compared to the wild type , as shown for developing gametophytes with four gametophytic nuclei , which might be derived from two MMC-like cells ( Figure 7D–F; Figure S8 ) . Mutant and wild-type gametophytes could be distinguished by the unusual position and size of gametopyhtic nuclei in mem gametophytes . In particular , more than five heterochromatic foci indicated by spots of high signal intensity ( chromocenters ) were often observed in mutant gametophytic nuclei of mem-1/MEM or mem-2/MEM plants , changes in chromatin structure we did not observe in gametophytes of wild-type plants expressing this marker ( unpublished data ) , which show a more equal distribution of H2B-YFP not exceeding five chromocenters . These results indicate changes in chromatin structure of gametophytic nuclei and possibly higher ploidy , as more chromocenters were detectable than expected for haploid nuclei . Intriguingly , higher H2B-YFP fluorescence was observed by quantification of the signal intensity in the nuclei of additional FMSs or developing female gametophytes in mem mutants as compared to the wild type . In addition , similar differences were observed within one ovule between the FMSs or developing gametophytes in the normal position and the additional FMSs or gametophytes in more micropylar positions ( Figure S8 ) , suggesting a higher ploidy of the latter . In summary , MEM plays a role for key steps of plant reproduction , including megasporogenesis , megagametogenesis , and embryogenesis . Importantly , heterozygous plants carrying a mutant mem-1 or mem-2 allele were affected in restriction of the germline lineage to one cell per ovule primordium , a phenotype resembling mutants in the small RNA and DNA-methylation pathways , both important for epigenetic regulation [18] , [19] . Interestingly , changes in the epigenetic setup of mem gametophytic nuclei were observed , providing an explanation for the defects found at later developmental stages when MEM expression is not detectable anymore . Nonetheless , a function of the extremely low MEM expression levels during early seed development cannot be excluded . The characterization of MEM illustrates the usefulness of our MMC transcriptome dataset for the identification of genes and functions important for megasporogenesis and early development of the plant female reproductive lineage . Formation and specification of the MMC is a key step in plant reproduction , marking the developmental switch from the sporophytic ( “somatic” ) fate to the reproductive or “germline” lineage . To our knowledge , we present the first transcriptome analysis of the MMC and the surrounding nucellus tissue in the sexual model plant Arabidopsis . Hierarchical sample clustering revealed that the MMC transcriptome is clearly distinct from that of the surrounding nucellus or the cells of the mature gametophyte . Our data indicate that translational control , ribosome biogenesis , and the expression of DEAD/DEAH-box helicases are major features of MMC specification in plants . This resembles an important feature of the animal germline , where translational regulation is a fundamental and highly conserved mechanism for restricting gene activity ( reviewed in [37] ) . While transcription is active early in gametogenesis in animals , differentiation into sperm cells and oocytes is under translational control as the reproductive cells enter meiosis ( reviewed in [50] ) . Particularly , specific RNA-helicases like Vasa and the eukaryotic translation initiation factor eIF4A , but also the RNA-binding proteins Boule , Bruno , and Pumilio , together with the zinc finger protein Nanos , are determinants of the Drosophila germline . They are involved in regulating the maintenance of stem cell fate and the differentiation to the gametes [37] . In flowering plants , similar molecular mechanisms might be required for the transition from undifferentiated meristematic cells of the nucellus to the reproductive fate . Interestingly , Arabidopsis homologues or proteins harboring similar functional domains as those described in animals were identified among the genes with specifically enriched expression in the Arabidopsis MMC , including the three DEAD/DEAH-box helicases MEM , eIF4A ( AT1G72730 ) , and AT3G16840; two PUMILIO ( PUM ) proteins , PUM7 and PUM23; as well as different RNA-binding and zinc finger proteins ( Table S7 ) . These findings suggest that similar regulatory pathways may be involved in germline specification and the development of female gametes in plants and animals . This is in line with the recent finding of shared features of epigenetic regulation through the small RNA pathway in plant and human gametes [22] . Genes involved in different small RNA pathways were found to be expressed during Arabidopsis megasporogenesis ( Figure S9 , Figure S10 ) and , as previously reported , during male gametogenesis ( Figure S9 , Figure S10 [51] ) . Thus , genes involved in small RNA pathways , including members of the AGO gene family , appear to play important roles in the regulation of germline development and the maintenance of germline integrity not only in animals but also in plants [52] , [53] . However , it remains to be elucidated whether these similarities represent the consequence of convergent evolution or ancestral features . Interestingly , a functional divergence of small RNA pathways has been found between these kingdoms—for example , in their requirements for target recognition [54] and the absence of the PIWI-clade of AGO proteins , whose members are abundantly expressed in the animal germline [55] , in plants . In animals , functional interactions between the RNA-based silencing pathway and the germline-specific Vasa family of RNA-helicases have been described ( reviewed in [56] ) . However , while Vasa and Vasa-like DEAD-box RNA helicases are widely conserved in the animal kingdom , no Vasa proteins have been discovered in plants . Nevertheless , in plants other DEAD/DEAH-box helicases may have similar functions . We analyzed two independent mutant lines with T-DNAs disrupting the MEM gene , encoding an RNA helicase with highly specific expression in the MMC . Future studies will be required to elucidate whether MEM may be functionally interacting with the small RNA pathway . Interestingly , mem mutants affect archespore selection and MMC specification leading to the initiation of two gametophytes in one ovule . These abnormalities resemble recently described Arabidopsis mutants involved in the small RNA pathway ( ago9 , sgs3 , rdr6 ) [18] and maize mutants in the DNA-methylation pathway [19] . Similar to these mutants , an additional enlarged cell in mem/MEM ovules may proceed to form a gametophyte without undergoing meiosis , as it occurs in aposporous apomicts [4] . This is consistent with the finding that additional developing FMSs or gametophytes have a higher ploidy than those in the wild-type position . Identification of the molecular players controlling apospory and other components of apomixis is a long-standing goal in plant research , as apomixis leads to the production of clonal offspring , a feature that has important agricultural applications [57] . However , unlike AGO9 , which has been detected in the somatic cells that form additional MMC-like cells [18] , MEM shows enriched expression in the MMC , suggesting that non-cell autonomous components regulate germline fate . It has been postulated for a while that the MMC suppresses the development of additional MMCs in a non-cell-autonomous fashion [2] , but the molecular components were not known . Nevertheless , though at significantly lower levels , MEM expression was detected in the nucellus cells such that cell-autonomous effects in the cells neighboring the MMC cannot be fully excluded . Although additional embryo sacs are formed , two mature gametophytes within one ovule have neither been observed in mem mutants , nor reported for Arabidopsis mutants affecting small RNA pathways or maize mutants defective in DNA-methylation [18] , [19] . It remains to be elucidated whether the additional gametophytes in mem mutants can occasionally give rise to viable offspring . With respect to defects in seed development the above mutants vary , too: while ago9 and one of the maize mutants ( dmt102::Mu/dmt102::Mu ) are nearly fully fertile , the other maize mutant ( D103 RNAi lines against Dmt103 ) shows seed abortion as we observed it for mem . Apart from MEM , other DEAD/DEAH-box helicases are enriched in the MMC . These helicases might have distinct or redundant functions . In Arabidopsis ATP-dependent RNA-helicases are a large protein family with 78 annotated members , generally involved in unwinding stable RNA ( or DNA ) duplexes using ATP as a source of energy . RNA-helicases in general are involved in multiple processes of RNA metabolism and play a role in developmental processes including pollen tube guidance , megagametogenesis , and seed development , as already demonstrated for MAGATAMA ( MAA ) , Arabidopsis thaliana RNA HELICASE36/SLOW WALKER3 ( RH36/SWA3 ) , and FREYA ( FEY ) [58]–[61] . In addition , embryo sac development arrest15 ( eda15 ) mutant plants , carrying a mutant allele of AtSUV3 , a gene with homology to MEM , develop abnormal numbers of nuclei during gametophyte development [62] . SUV3 genes are evolutionary highly conserved from purple bacteria to higher eukaryotes including plants and humans [63] . They are involved in unwinding dsDNA , dsRNA , and RNA-DNA heteroduplexes [64] . While SUV3 proteins studied so far are localized predominantly in the mitochondria , the human SUV3 ortholog is partially present in the nucleus and is probably involved in chromatin maintenance , cell-cycle regulation , and the regulation of apoptosis [65] . In the mem/MEM heterozygous mutants instead of one , two cells with FMS characteristics were often observed . It is also possible that an apoptosis defect in one of the three degenerating megaspores might result in a second surviving FMS-like cell; however , surviving FMS-like cells should have reduced ploidy unlike what we observed . In summary , the functional analysis of MEM revealed structural abnormalities from the onset of megasporogenesis to embryo development , suggesting that MEM function is required at several stages of reproductive development . The enriched abundance of MEM transcript in the MMC , together with the observed changes in LHP1 binding and chromatin structure in mem female gametophytic nuclei , suggests an involvement of MEM in establishment of the epigenetic landscape in the female gametophyte . In this way , MEM expression during megasporogenesis might be relevant for the regulation of transcriptional control at later stages of reproductive development . The importance of the epigenetic state of the mature gametes for the transition from gametophyte to seed development has recently been demonstrated [42] . Interestingly , the observed changes in chromatin structure in mem mutant gametophytic nuclei are in agreement with the functions of the human SUV3 ortholog in chromatin maintenance [65] . Apart from this , the enrichment of genes regulating chromatin structure in the MMC as compared to the mature gametophyte suggests a more general role of epigenetic regulation in the acquisition of germline fate in the female reproductive lineage . Recent studies also provide evidence for an involvement of epigenetic regulation in the differentiation between sexual gametophyte formation and apospory [18] , [19] . However , it remains to be seen whether the modifications in epigenetic marks and chromatin structure observed in mem mutant gametophytes play a role in this respect . Apart from MEM , a number of genes enriched during megasporogenesis as compared to the mature gametophyte play important roles during embryo development , such as MATERNAL EFFECT EMBRYO ARREST63 ( MEE63 ) and several EMBRYO DEFECTIVE ( EMB ) genes [62] , [66]–[70] . In these cases , gene function might be masked by haplo-sufficiency or redundancy during megasporogenesis and become apparent only during embryonic development . Alternatively , a subset of genes expressed during early stages of reproduction might determine the developmental fate of later stages—for example , by establishing epigenetic marks required for activation or repression of gene expression later in development . However , given the evidence for the importance of translational control during gametophyte development , transcripts present in the MMC might encode proteins whose activities are only required at later stages of reproductive development . A total of 13 genes were significantly enriched as compared to the tissue atlas including the sporophytic nucellus . This specificity of expression suggests an importance of the gene function for the developing MMC , as demonstrated by the characterization of one of those genes—MEM—for MMC specification and gametophyte development . Notably , we found YUCCA2 and AT1G29440 , genes involved in auxin synthesis and signaling , enriched in the MMC . An auxin gradient established during megagametogenesis has recently been proposed to be important for cell specification [17] . However , to date no role for auxin has been ascribed for megasporogenesis . As YUCCA2 expression is modulated by SPL/NZZ during lateral organ development [16] , it might link homeotic gene function underlying reproductive organ development with gametogenesis . In conclusion , our study indicates that similar molecular mechanisms are acting upon germline specification and differentiation in animals and in plants . Control of translational regulation is a dominant feature in the transcriptome dataset and RNA processing involving RNA-helicases plays an important role for early stages of female gametophyte development . MEM , a gene encoding a helicase with significantly enriched expression in the MMC , plays important roles for restriction of the reproductive fate to only one cell per ovule primordium and for gametophyte development . Thus , this transcriptome analysis of the Arabidopsis MMC provides insights into the molecular basis of a key step of plant reproduction . A detailed understanding of the mechanisms underlying megasporogenesis is not only interesting from a fundamental point-of-view , but also the precondition for the manipulation of this pathway towards apomixis , which is of great importance for plant breeding and seed production . Arabidopsis thaliana ( L . ) Heynh . , accession Landsberg erecta , was used for LAM sample preparation , as specimen for in situ hybridizations , and for plant transformation throughout this study . Arabidopsis thaliana Col-0 plants were used as wild-type plants in the context of the mutant analysis . Seedlings were grown on MS plates for 7–12 d before transfer to soil ( ED73 , Universalerde , Germany ) and grown in a growth chamber at 16 h light / 8 h darkness at 21°C and 18°C , respectively . Plants were treated with a 10% milk suspension and nematodes against powdery mildew and black flies , respectively . Enhancer trap lines and T-DNA insertions were ordered from the Cold Spring Harbor Trapper Collection ( http://genetrap . cshl . edu/ ) or NASC ( http://arabidopsis . info ) and grown as described above . pTFL2:TFL2-GFP and pAKV:H2B-YFP marker lines were kindly provided by K . Goto and W . -C . Yang , respectively . The PUM12-GUS reporter line was described previously [22] . The pABCB19:ABCB19-GFP ( pPGP19:PGP19-GFP ) marker line was kindly provided by M . Geisler [25] . To prepare material for LAM , inflorescences were fixed on ice in farmers' fixative ( ethanol:acetic acid 3∶1 ) , vacuum infiltrated two times for 15 min , and stored on ice overnight . The fixative was replaced by 70% ethanol and young buds were selected under the dissecting scope . Subsequently , tips of the ovaries were dissected using injection needles , cleared in chloralhydrate:glycerol:water ( 8∶1∶2; w:v:v ) , and subjected to microscopic analysis . Buds with ovules harboring MMCs ( predominantly before meiosis or at meiosis I ) were embedded in Paraplast X-tra in an ASP200 embedding machine ( Leica Microsystems , Wetzlar , Germany ) as described [22] . Paraplast embedded samples were stored at 4°C until further use . Thin sections of 6–7 µm were prepared from the samples using a RM2145 Leica microtome and mounted on PET metal frame slides ( Molecular Machines and Industries ( MMI ) , Glattbrugg , Switzerland ) using methanol . Slides were dried overnight on a heating table at 42°C and subsequently dewaxed two times for 10 min in Xylol ( Merck , Darmstadt , Germany ) . LAM was performed with a SL µCut and a CellCut Plus Instrument ( MMI ) . MMCs and surrounding sporophytic nucellus tissue were subsequently isolated and collected separately on MMI isolation caps . On average , ∼65 MMC sections were collected per day on one isolation cap ( estimated to be equivalent to ∼50–55 MMCs ) . In addition , one or two ovary sections were isolated per slide to control for RNA quality . LAM samples were stored at −80°C until extraction . RNA was isolated using the PicoPure RNA isolation kit ( Arcturus Engineering , Mountain View , USA ) following the manufacturer's instructions with modifications . For extraction , caps were covered with 10–11 µl of Extraction buffer from the kit , incubated at 42°C for 30 min , and pooled for binding on the column . RNA integrity was tested on a Bioanalyzer ( Agilent , Santa Clara , USA ) , using control sections dissected after collection of MMCs and surrounding nucellar tissue from the same slides . After optimization , RNA integrity was good and reproducible at ∼RIN7 . Isolated RNA from ∼560 to 930 pooled sections of MMCs or the surrounding nucellar tissue were subjected to two rounds of linear amplification with the MessageAmpII Kit ( Ambion , Foster City , USA ) , following the manufacturer's instructions . During the second round of amplification , biotin-11-UTP ( Ambion ) was incorporated in the amplified aRNA for array analysis . Quantity and fragment size distribution of the amplified product was analyzed using a Nanodrop and the Bioanalyzer . Samples with amplification yields between 16 µg and 40 µg were used for samples MMC1 to MMC3; for MMC4 , three samples with suboptimal amplification yields between 2 . 6 and 7 . 1 µg were pooled . Amplification yields from the sporophytic nucellar samples ranged between 38 µg and 65 µg . 15 µg labeled aaRNA was fragmented and hybridized onto the Arabidopsis ATH1 GeneChip ( Affymetrix ) for 16 h at 45°C as described in the technical manual . The hybridization , staining , washing , and subsequent array scanning was performed as described previously [22] . Original data-files ( . CEL ) are deposited on ArrayExpress ( http://www . ebi . ac . uk/arrayexpress/ ) under accession E-MEXP-3137 ( megaspore_mothercell , including the four MMC samples ) and under E-MEXP-3138 ( sporo_nucellus , including the four datasets of the surrounding nucellus ) . Total RNA was isolated from Arabidopsis Col-0 inflorescences using the RNeasy Plant Mini Kit ( Qiagen , Hilden , Germany ) and treated on column with DNaseI . The RNA was subsequently reverse transcribed to cDNA using SuperscriptII Reverse Transcriptase ( Invitrogen , Carlsbad , USA ) . Fragments for cloning of in situ probes were PCR amplified with Taq ( Sigma , St Luis , USA ) ; for primer sequences , see Table S10 . Fragment cloning and in situ hybridizations were performed as previously described [22] with modifications: In situ hybridizations were performed on 7 or 8 µm thin sections of inflorescences or buds . Pictures were captured on a Leica DMR microscope ( Leica Microsystems , Bensheim , Germany ) , cropped , and processed in Adobe Photoshop Version 8 . 0 . 1 ( Adobe Systems Inc . , San Jose , CA , USA ) . 5′ upstream sequences of genes of interest were PCR-amplified from Arabidopsis thaliana Col-0 genomic DNA , using primers containing the 5′attB sites ( for primer sequences , see Table S10 ) . The PCR products were cloned into pDONR207 ( Gateway Cloning , Invitrogen ) , using site-directed recombination according to the manufacturer's recommendations . The resulting entry clones were recombined with the destination vector pMDC162 ( the At2g24500 promoter ) or pSS240 ( other entry clones ) [71] , [72] , producing the final binary vectors containing the uidA reporter gene encoding β-glucuronidase ( GUS ) . Buds were opened and transferred to the GUS reaction buffer for 24–72 h at 37°C ( 4 mM 5-Bromo-4-chloro-3-indoxyl-beta-D-glucuronic acid cyclohexylammonium salt ( Biosynth AG , Staad , Switzerland ) , 10 mM ETDA , 0 . 1% Triton X-100 , 2 mM potassium ferrocyanide , 2 mM potassium ferricyanide , 100 mM phosphate buffer ( pH = 7 . 2 ) ) , dissected , and mounted in clearing solution ( 1×PBS , 20% lactic acid , 20% glycerol ) . Wild-type plants were transformed using the floral dip method [73] . At least three independent F1 lines were analyzed per construct . Enhancer trap lines were GUS-stained following the same protocol . Two independent T-DNA insertion lines disrupting At5g39840 ( mem-1 and mem-2; SAIL_182_A07 and SALK_11370 , respectively ) were analyzed for phenotypes during reproductive development . Developmental arrest during early seed development was counted after opening the silique with injection needles . For histological analysis , ovules and developing seeds were cleared as described above and subjected to microscopic analysis . For analysis of developmental defects during embryo and endosperm development , mem1/MEM and wild-type plants were pollinated 2 d after emasculation and siliques were fixed as described after 2 , 3 , or 4 DAP . Lines mem-1 and mem-2 were genotyped with primers 5′-GAATTTCATAACCAATCTCGATACAC-3′ , 5′-TACTGCAGACCTCACGAAACC-3′ , and 5′-GTCGAGTCTGCAGTGTTTTCC-3′ , and with primers 5′-CTTTGACGTTGGAGTCCAC-3′ [74] , 5′-AATCGAGTGTTTGCAACAACC-3′ , and 5′-GCTAACGAGAGTTCAACACCG-3′ , respectively . Position of the T-DNA left border was analyzed by sequencing . For analysis of transmission efficiency at least three heterozygous mutant plants per insertion line were crossed as female or male to the wild-type [41] . The progeny from these crosses were genotyped . pTFL2:TFL2-GFP and pAKV:H2B-YFP marker lines were crossed to heterozygous mem-1 and mem-2 mutant plants as female . The F1 and F2 generations of these crosses were used for the analyses . F2 progeny from the cross of the heterozygous mutants with the TFL2 marker line were analyzed for expression of the marker . From the 42 progenies analyzed from the cross with the mem-1 mutant line , GFP signal was observed in all plants; among the progeny from the cross with mem-2 , 18 plants out of 23 were clearly GFP positive , one plant was negative , and in the other plants background fluorescence could not be discriminated from signal . This high frequency suggested more than one copy of the pTFL2:TFL2-GFP marker in the genome . Thus , a complete or close linkage of the marker to the mem-1 and mem-2 alleles is unlikely . In the heterozygous mem mutants carrying the TFL2:TFL2-GFP marker an unusually high percentage of ovules arrested at early stages of reproductive development were observed ( 57% and 52% in heterozygous mem-1 and mem-2 mutants , respectively ) . However , phenotypes observed in mature gametophytes resembled the phenotypes of mem-1/MEM and mem-2/MEM plants . For clearing , GUS-staining , and in situ hybridizations , the slides were viewed under a Leica DMR microscope ( Leica Microsystems , Bensheim , Germany ) and pictures were taken with a digital camera for microscopes ( Magnafire model S99802 , Optronics , USA ) . Confocal images were acquired using a Confocal Laser Scanning Microscope ( Leica SP2 , Leica ) . GFP or YFP signal and chlorophyll auto-fluorescence were simultaneously acquired with laser excitation 488 nm and emissions of 500–530 nm for GFP and 590–720 nm for chlorophyll . For quantification of fluorescent signals fluorescence intensity of the nuclei expressing H2B-YFP was measured on 3-dimensional reconstructions of confocal series using IMARIS ( Bitplane , CH ) . Contour surfaces were generated for individual nuclei and the intensity sum was used to calculate the relative intensity . The nucleus with the lowest intensity within one ovule was set to 1 . Quality control was performed as previously described [22] . Robust-Multiarray Analysis ( RMA , [75] ) was performed using the Bioconductor software ( Version 2 . 6 , http://www . bioconductor . org ) implemented in the statistical software “R” Version 2 . 10 . 1 ( http://www . r-project . org ) . Reannotation of an array can significantly alter the interpretation of a microarray dataset [76] . Therefore , we used reannotation information where probes were mapped to predicted gene sequences of the TAIR9 Arabidopsis genome release ( downloadable at http://brainarray . mbni . med . umich . edu/ ) . The reannotated array targets 21 , 504 genes , representing around 64% of the Arabidopsis genome utilizing 219 , 079 single probes . From the latter , 1 , 732 single probes match multiple genes in the genome perfectly and were removed from the mappings for the analysis using the dChip software ( see below ) , which removed 251 probesets from the analysis . The Bioconductor package affxparser [77] was used to generate a new chip description file ( . cdf-file ) where the multiple mappings had been removed . Base-level annotations were downloaded from the Bioconductor homepage ( Version 2 . 6 ) , which includes the Gene Ontology ( GO ) mappings . Protein family ( PFAM ) and gene family ( FAM ) information for Arabidopsis were downloaded from TAIR9 ( http://www . arabidopsis . org ) . We made use of extensive microarray datasets from Arabidopsis for comparing the molecular profile of MMCs and gametes in contrast to tissues of the rest of the organism/body . We processed an Arabidopsis atlas consisting of mixed tissue and single-cell tissues as described previously [22] and added the following two datasets: ( 1 ) laser-microdissected early embryo and endosperm stages ( Harada-Goldberg dataset provided by Ryan Christopher Kirkbride: GSE12404 record in GEO ( http://www . ncbi . nlm . nih . gov/gds ) , as used in [76] ) and ( 2 ) cell-sorted subdomains of the shoot apical meristem [78] . For finding single genes that show enrichment in MMCs or the sporophytic nucellus , log2-transformed dChip expression indexes were imported into R [79] . A linear model was fitted on the data and modified t tests , implemented in the limma-package [36] , were used to test every contrast of a given cell type against all other tissues/cell types . Genes with an adjusted p value smaller than 0 . 01 in all contrasts ( Benjamini-Hochberg adjustment; [38] ) were considered significant . For finding genes that show enrichment in the MMC as compared to cells of the mature female gametophyte [22] , the same method as described for the tissue atlas was used , only that RMA was used for processing the data to generate log2-scale expression indexes and genes were identified at a false discovery rate below 0 . 05 [36] . We firstly applied a pre-filtering step and restricted the analysis to probesets with evidence of expression for at least three out of 13 arrays ( four replicates MMC and three replicates each for egg cell , central cell , and the synergids ) as analyzed by AtPANP . After fitting the linear model and identifying differentially expressed genes using the moderated F-statistic ( at a false discovery rate below 0 . 05 ) [36] , each contrast of the MMC against egg cell , central cell , and synergids was examined separately: genes significantly upregulated in all three contrasts were selected as “MMC enriched . ” Heatmaps were generated using the Bioconductor package gplots [80] , using hierarchical agglomerative clustering ( complete linkage ) and euclidean distance . Heatmaps were based on log2-transformed mean expression values generated by dChip [81] , except for the genes differentially expressed in the MMC and the cell types of the mature female gametophyte ( Figure 3 ) , where the heatmap was based on log2-scale expression values generated by RMA [75] . In order to calculate present/absent p values we applied a previously described method called AtPANP [22] , which is a modified version of the original PANP method [82] . The method makes use of internal negative control for the ATH1 GeneChip that consists of probes that do not match sequences from the latest Arabidopsis genome release anymore . These negative probes were determined via BLAST [83] . For this , probes present on the ATH1 GeneChip but not used in the probeset annotation were queried against the TAIR9 cDNA and BAC databases ( downloaded from www . arabidopsis . org ) , using the standalone BLAST executable function “blastall” Version 2 . 2 . 23 ( ftp://ftp . ncbi . nih . gov/blast/executables/ ) . Probes that matched either genomic or cDNA sequences with more than two mismatches only were considered reliable measures for background ( a total of 1 , 574 probes ) . Single negative probes were randomly assembled into sets of 11 , thus constituting negative probe sets . We generated a total of 2 , 000 negative probesets by resampling randomly from the pool of negative probes . Negative probe set signals were then calculated using the RMA algorithm [75] , an algorithm that has been shown to be robust for the analysis of data from amplified RNA [84] . An empirical signal background distribution for each individual array was used to determine the probeset signal threshold for a given false-positive rate—as implemented in the pa . calls-function from the Bioconductor package PANP [82] . p value calculations on resampled negative probesets were repeated 20 times and averaged in order to get more robust results . A p value threshold of 0 . 02 was considered significant ( referred to as “present” ) and a transcript considered expressed when called “present” in at least three out of four replicates and marginally expressed when called “present” in at least two out of four replicates . Venn diagrams of present call overlaps were drawn using the software VENNY [85] . For Gene Ontology ( GO ) analysis we used the Bioconductor package topGO [86] . We used a Fisher's exact test to test for overrepresented GO terms in combination with the function “weight . ” We also used a two-sided Fisher's exact test and comparison against the whole array-genome to test for misrepresentation of protein and gene families .
Germline specification is a key step in sexual reproduction . In plants , the reproductive lineage or “germline” doesn't arise early in development , as it does in animals; rather , the germline is specified during flower development . In the female reproductive organs of the flower , a single sporophytic cell in each ovule is selected and differentiates into a megaspore mother cell ( MMC ) , which will undergo meiosis . Despite the importance of the specification of the MMC as the first committed cell of the germline lineage , little is known about the genetic and molecular bases of this process . We performed a cell-type-specific transcriptome analysis of Arabidopsis MMCs using laser-assisted microdissection and microarrays . Statistical data analysis comparing these results with the transcriptomes of 71 other types of cells and tissues revealed the importance of translational control pathways and RNA helicases for plant germline development , a feature reminiscent of the animal germline . We further characterized a novel MMC-enriched RNA helicase , called MNEME , and showed that it plays important roles in MMC differentiation and the restriction of the plant germline to only one cell per ovule . This example illustrates the usefulness of our transcriptome dataset for the identification of novel candidate genes involved in this crucial step of plant reproduction .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "developmental", "biology", "plant", "science", "model", "organisms", "plant", "biology", "genetics", "biology", "genomics", "molecular", "cell", "biology", "genetics", "and", "genomics" ]
2011
Transcriptome Analysis of the Arabidopsis Megaspore Mother Cell Uncovers the Importance of RNA Helicases for Plant Germline Development
LRRK2 plays an important role in Parkinson's disease ( PD ) , but its biological functions are largely unknown . Here , we cloned the homolog of human LRRK2 , characterized its expression , and investigated its biological functions in zebrafish . The blockage of zebrafish LRRK2 ( zLRRK2 ) protein by morpholinos caused embryonic lethality and severe developmental defects such as growth retardation and loss of neurons . In contrast , the deletion of the WD40 domain of zLRRK2 by morpholinos targeting splicing did not induce severe embryonic developmental defects; rather it caused Parkinsonism-like phenotypes , including loss of dopaminergic neurons in diencephalon and locomotion defects . These neurodegenerative and locomotion defects could be rescued by over-expressing zLRRK2 or hLRRK2 mRNA . The administration of L-dopa could also rescue the locomotion defects , but not the neurodegeneration . Taken together , our results demonstrate that zLRRK2 is an ortholog of hLRRK2 and that the deletion of WD40 domain of zLRRK2 provides a disease model for PD . Parkinson's disease ( PD ) is a common neurodegenerative disorder characterized by the selective loss of dopaminergic neurons of the substantia nigra pars compacta ( SNpc ) and movement symptoms , including resting tremor , rigidity and postural instability [1] . The vast majority of PD patients are idiopathic , but a small number of patients show a familial inheritance where mutations in α-synuclein , Parkin , DJ-1 , ubiquitin-C-hydrolase-L1 ( UCHL1 ) or Leucine-rich repeat kinase 2 ( LRRK2 ) play an important role [2] , [3] . Of the identified disease genes for PD , mutations in LRRK2 are the most prevalent in both familial and sporadic PD patients [4]–[6] , and show an interesting diversity in terms of population distribution as well as functional impact . For example , G2019S variant within the kinase domain was found to be a high-penetrate gain-of-function mutation ( associated with enhanced kinase activity ) ; it appears to be the most common mutation in the majority of the populations studied except Asian ones [7]–[9] . In contrast , G2385R variant within the WD40 domain results in a loss-of-function mutation ( associated with the reduced kinase activity ) and is a common susceptibility allele in Asian populations , but absent in Caucasians [10] , [11] . The diverse spectrum of pathogenic mutations within the multiple domains of LRRK2 protein ( see below ) and the complex mechanisms by which these mutations influence the development of PD suggest that LRRK2 may be a master regulator of the disease development [12] . LRRK2 is therefore not only clinically important to link the familial and sporadic forms of PD , but also biologically significant for understanding the etiology of the disease [13] . The biological function of LRRK2 is , however , largely unknown . Human LRRK2 ( hLRRK2 ) encodes an unusually large protein composed of multiple functional domains , including armadillo repeats , ankyrin repeats , two enzymatic S/T kinase and Roc GTPase domains , COR and WD40 domains ( as the dimerization motif ) , indicating that LRRK2 is a complex multifunctional protein [14] , [15] . The functional studies of LRRK2 were largely carried out by over-expressing either the wild-type or mutant allele of the MAPKKK and ROC domains of hLRRK2 in in vitro and in vivo model systems [16] . These transgenic studies suggested that the hyperactive kinase activity of LRRK2 may be cytotoxic and cause neurodegeneration . However , over-expression of the wild-type human LRRK2 protein does not always achieve the same mutant effect , as demonstrated in rat studies [17] , raising a question on how much can be inferred from such ‘gain-of-function’ analysis of mutant alleles in understanding the normal function of LRRK2 . When measuring the kinase activity of LRRK2 mutations using moesin as substrate [18] , the most frequent mutation , G2019S , is the only one showing stimulated kinase activity . Hence , the mechanism by which LRRK2 mutation induces PD is more complex than previously imagined and is not only due to an increase in LRRK2 kinase activity . Analysis of loss-of-function mutations in Drosophila ( dLRRK2 ) revealed conflicting findings; and it was not clear whether the disruption of dLRRK's function caused parkinsonism-like neurodegeneration and locomotive defects in this model system [19] , [20] . So far , there are no reports on the function of WD40 or other domains in vertebrate models , such as the mouse or rat and the decreased kinase model for LRRK2 is very limited . In this study , we performed the first in vivo loss-of-function study of LRRK2 in zebrafish . We cloned the zebrafish homolog of human LRRK2 and performed a series of molecular and genetic analyses to characterize its expression and biological functions , particularly the role of the WD40 domain , in embryonic and neuronal development . Through a TBLASTN analysis of hLRRK2 protein sequence against zebrafish cDNA sequences and a subsequent TBLASTX analysis of the identified zebrafish cDNA sequences against human cDNA sequences , we identified XM_682700 as the zebrafish homolog of hLRRK2 . To clone the full length transcript of zLRRK2 , we performed further RACE analysis using mRNAs isolated from the brain of adult fish and identified a 9168 bp transcript carrying both start and stop codons . The size of this transcript matched the zlrrk2 mRNA detected by Northern analysis ( Figure 1B ) . This 9168 bp transcript consists of 51 exons , spanning 118 kb genomic sequences ( chr25:37299901–37361611 , UCSC Genome Browser , Dec 2008 ) and encodes a protein of 2533 amino acid residues ( Text S1 ) . This experimentally cloned full length transcript is different from the ensembl predicted cDNA of 7410 bp that consists of 59 exons and encodes a protein with 2470 amino acid residues . The zLRRK2 protein contains all the functional domains of the hLRRK2 protein . There is a high degree conservation of amino acid sequences between the zLRRK2 and hLRRK2 proteins , with the highest conservation within the kinase domain ( 71% ) ( Figure 2 ) . A phylogenetic analysis ( Figure S1 ) revealed that the zLRRK2 protein was clustered together with the hLRRK2 protein as well as the LRRK2 proteins of other animal species . A temporal expression analysis by quantitative RT-PCR ( qRT-PCR ) ( Figure 1A ) indicated that the maternal mRNA of zlrrk2 could be detected at the pre-MBT ( mid blastula transition: from one cell to sphere ) stages and was then degraded by the beginning of the gastrula stage . The zygotic expression of zlrrk2 was first detectable at the tail bud stage ( the last stage of gastrulation ) and increased gradually during the segmentation and pharyngula stages , reaching a peak around 24 hours post fertilization ( hpf ) . After a short period of reduction , the expression of zlrrk2 increased again through the hatching and larval stages up to , at least , 10 days post fertilization ( dpf ) . At both 24 hpf and 6 dpf , a strong expression of zlrrk2 could be detected in the brain by whole mount in situ hybridization ( WISH ) analysis , and zlrrk2's expression in the brain is ubiquitous ( Figure 1E ) . In adult fish ( older than three months ) , zlrrk2 mRNA was detected in the brain , muscle , ovary and gut by Northern blot and qRT-PCR analyses ( Figure 1B and 1C ) , but full-length zLRRK2 protein was predominantly detected in brain by Western blot ( Figure 1D ) . Microinjection of morpholinos targeting the ATG start site into embryos effectively abolished the expression of zLRRK2 protein , as determined by Western blot analysis ( Figure S2 and Figure S3A ) . Knockdown of zLRRK2 expression resulted in severe embryonic lethality ( ∼90% of 64 embryos examined ) within 3 dpf . The surviving morphants showed developmental retardation , such as slow growth , reduced brain size and heart edema compared to the wild type fish ( Figure S3 ) . WISH analysis showed a loss of TH+ neurons in the diencephalon of the surviving morphants ( Figure S3B ) , which was consistent with the reduced level of tyrosine hydroxylase detected by the Western blot analysis ( Figure S3A ) . Both heart edema and TH+ neuron loss phenotypes are morpholino concentration-dependent and can be partially rescued by over-expression of human LRRK2 ( Figure S3 and Figure S4 ) ( zLRRK2 was not used for the rescue , because its expression will be blocked by ATG morpholinos ) . However , due to the developmental retardation , it is not clear whether the loss of TH+ neurons in the diencephalon is an indication for a specific role of zLRRK2 . The severe embryonic defect of the zLRRK2 knockdown also prevented us from studying its impact on locomotive movement . It has been shown that G2385R variant within the WD40 domain was associated with a very moderate risk for PD development [21] . We therefore hypothesize that the deletion of WD40 domain may lead to a weaker phenotype than the translational block of zLRRK2 expression , allowing us to study the specific role of zLRRK2 in neurodevelopment and locomotive movement . To delete the WD40 domain , we designed morpholinos that specifically interrupted the splicing of the 45th exon of zlrrk2 and consequently introduced a pre-mature stop codon just upstream of the WD40 domain ( Figure S5A ) . Delivery of this splicing-blocking morpholinos into embryos caused a production of truncated zlrrk2 mRNA without the WD40 domain ( zLRRK2-ΔWD40 ) , as confirmed by RT-PCR and sequencing analyses ( Figure S5B and S5C ) . As hypothesized , zLRRK2-ΔWD40 morphants showed a largely normal embryonic development , at least up to 7 dpf , without any distinguishable morphological defects , except a mild blood accumulation between the swimbladder and yolksac ( Figure S6 ) . Western blot analysis of whole fish lysate of zLRRK2-ΔWD40 morphants ( 3 dpf ) ( Figure 3A and Figure S2 ) showed significant loss of full-length ( 280 KD ) zLRRK2 protein and TH protein expression . The reduction of TH expression was also confirmed by qRT-PCR analysis ( Figure S5D ) . Consistently , the WISH analysis ( at 3 dpf ) showed a loss of TH+/DAT+ DA neurons in the diencephalon of the zLRRK2-ΔWD40 morphants ( Figure 3B ) . As expected , the phenotypes of the zLRRK2-ΔWD40 morphants are morpholino concentration-dependent ( Figure S4 ) . To further investigate the impact of zLRRK2-ΔWD40 on neurodevelopment , we microinjected the zLRRK2-ΔWD40 morpholinos into the embryos of the Tg ( DeltaD∶GAL4/UAS∶Kaede ) line [22] , where neurons are labeled by Kaede expression ( driven by detlaD promoter ) . At 18 somite stage of embryonic development , no obvious neuron cell loss could be observed in the ΔWD40 morphants ( compared to the wild-type fish ) ( Figure 3C ) . At 6 dpf , the forebrain and hindbrain of the morphants appeared to be normal and indistinguishable from the control siblings , but the midbrain , particularly the optic tectum of the morphants contained far fewer neurons than the control siblings ( Figure 3C ) . Using the TUNEL assay , we found an enhanced apoptosis throughout the zLRRK2-ΔWD40 morphants ( Figure 3D ) . We also stained axonal microtubules using an acetylated-tubulin antibody and found a reduction and disorganization of axon tracts , most prominently in the optic tectum of the zLRRK2-ΔWD40 morphants ( Figure 4 ) . These results indicated that the deletion of the WD40 domain causes the loss of neurons and the reduction and disorganization of axon tracts in the brain , including the DA loss in the diencephalon of the morphants . Over-expression of either wild-type zLRRK2 or hLRRK2 ( Figure S7 ) could rescue both the DA neuron loss ( Figure 3B , Figure 5A and 5B ) and axon tract disorganization ( Figure 4 ) of the zLRRK2-ΔWD40 morphants , confirming that the neurodegenerative phenotypes of the zLRRK2-ΔWD40 morphant was a specific effect of the WD40 domain deletion due to splicing-blocking morpholinos , instead of off-target effect or unspecific toxicity of morpholinos . Furthermore , the successful rescue of the neurodegenerative phenotype by wild-type hLRRK2 confirmed zLRRK2 to be the functional ortholog of hLRRK2 . To investigate the locomotion behavior of the zLRRK2-ΔWD40 morphants , we measured the swimming distance of larval fish within time windows of 30 seconds . As shown in Figure 6 , the zLRRK2-ΔWD40 morphants moved much smaller distances than the wild-type fish . Like the neurodegenerative defects , this reduced swimming activity could be rescued by over-expressing either zLRRK2 or hLRRK2 ( Figure 6 and Figure S8 ) . Intriguingly , this reduced swimming activity of zLRRK2-ΔWD40 morphants could also be rescued by the administration of Levo-dopa ( L-dopa ) , a compound that is widely used to treat PD ( Figure 6 and Figure S8 ) . The administration of L-dopa , however , did not rescue the neurodegenerative phenotype of the morphants , as demonstrated by TH labeling ( Figure 3B , Figure 5A and 5B ) . In addition to the investigation of the WD40 deletion , we also investigated the impact of the over-expression of human LRRK2 G2019S and G2385R mutant alleles in zebrafish . The over-expression of both the mutant alleles could induce a similar blood accumulation between the swim bladder and yolk sac as the zLRRK2-ΔWD40 deletion and a mild loss of TH+ cell compared to wild-type ( Figure S4 and Figure S9 ) . Furthermore , unlike the wild-type zLRRK2 and hLRRK2 , both hG2019S and hG2385R alleles could only partially rescue the loss of TH+ neurons in the zLRRK2-ΔWD40 morphants ( Figure 5C ) . In this study , we provide strong evidence that zLRRK2 is an ortholog of hLRRK2 . The proteins of zLRRK2 and hLRRK2 show a conservation of amino acid sequence and share an identical domain structure . Phylogenetically , zLRRK2 is clustered together with hLRRK2 , instead of hLRRK1 , as well as the LRRK2 proteins of other animal species . Finally , the successful rescue of the defects of zLRRK2 morphants , in terms of both neurodegeneration and swimming abnormality , by over-expressing hLRRK2 mRNA , provides the most convincing evidence for a functional conservation of LRRK2 between zebrafish and human . zLRRK2 shows a dynamic expression profile in zebrafish . During embryonic development , zLRRK2 transcript is mainly restricted to the brain , but demonstrating ubiquitous expression within brain , as observed in mouse , rat and human brains [23] , [24] . In adult fish , zlrrk2 mRNA was expressed in multiple tissues or organs . Western blot analysis ( using an antibody against the WD40 domain ) , however , showed a rather restricted expression of zLRRK2 protein in the brain . Together with the previous finding that the LRRK2 protein isolated from transgenic mouse brain showed a higher kinase activity than from transgenic mouse lung or transfected cultured cells [25] and the suggestion that hLRRK2 has several splicing forms ( AceView [26] ) ; the differential expression patterns of zLRRK2 mRNA and protein in various tissues may suggest a complex mechanism for regulating zLRRK2 splicing and expression . zLRRK2 plays an important role in neuronal development . The involvement of zLRRK2 in neurodevelopment is first suggested by the retarded brain development and the loss of TH+ neurons in the zLRRK2 ATG morphants and further evidenced by the neurodegenerative phenotypes of the zLRRK2-ΔWD40 deletion . The zLRRK2-ΔWD40 deletion caused a significant loss of DA neurons in the diencephalon , and other types of neurons are also likely affected . Interestingly , our preliminary study showed that the zLRRK2-ΔWD40 deletion had a rather limited impact on the development of neurons during early embryonic development . Considering that , 1 ) zLRRK2 shows a ubiquitous expression in the brain , 2 ) the zLRRK2-ΔWD40 deletion leads to an increased apoptosis activity across the brain , and 3 ) the midbrain , particularly the optic tectum of zebrafish is very stress-sensitive , the loss of neurons in the zLRRK2-ΔWD40 morphans is more likely due to a neurodegeneration process instead of the interruption of normal neuronal development . The loss of DA neurons in zLRRK2-ΔWD40 deletion likely happens as a result of a rather broad neurodegeneration within several regions of the brain . We speculate that LRRK2 may be important for neuron survival and thus play a more prominent role in neural maintenance , rather than development of neurons . The interruption of normal LRRK2 function may cause neurons to become more sensitive to factors that might trigger cell death . The zLRRK2-ΔWD40 deletion also caused a significant reduction and disorganization of axon tracts , more prominently in the midbrain . This is consistent with the previous finding from transgenic mouse study that LRRK2 is involved in neurite growth [27] . Since LRRK2 interacts with microtubule through Roc domain [12] and the WD40 domain can bind to the Roc domain [28] , the reduced and disorganized axon tracts in the midbrain of the zLRRK2-ΔWD40 morphants may be due to an interruption of the microtubule cytoskeleton [29] . This would be consistent with the well established requirement for microtubules in axon outgrowth . Our result has also supported the recent hypothesis that Parkinsonism may be due to a disorganized ‘microtubule railroad’ system , which can be the consequence of a faulty of motor [30] or perhaps microtubule . However , further study will be needed to elucidate whether the reduced and disorganized axon tracts truly reflects the interrupted ‘microtubule railroad’ system and thus causes neurodegeneration . This is the first demonstration of the role of the WD40 domain of LRRK2 in neural development and/or neural maintenance . The WD40 domain is known to mediate protein-protein interaction in many contexts , such as signal transduction , transcription regulation , cell cycle control , apoptosis and cytoskeleton assembly [31] . The WD40 domain has been suggested to play a crucial role in LRRK2 self-interaction and autophosphorylation , which regulates the kinase activity of LRRK2 [28] . The deletion of the WD40 domain causes a partial reduction in kinase activity in vitro , which could be restored to a normal level by the over-expression of the gain-of-function mutation R1441C [32] . We have previously shown that the G2385R risk variant in the WD40 domain increases neuronal apoptosis under cellular stress [33] , providing further support for the functional role of WD40 domain . However , the different phenotypic impact of blocking the kinase activity of LRRK2 ( by knocking-down the protein expression ) and deleting the WD40 domain suggests that LRRK2 may influence the neurodevelopment through other mechanisms beyond the modification of its kinase activity . We have demonstrated a locomotion defect in the zLRRK2-ΔWD40 morphant . More importantly , we confirmed that the locomotion defect likely happens as a direct result of the dopamine insufficiency ( due to the loss of DA neurons ) , since the defect can be rescued by supplementing dopamine through the administration of L-dopa . The administration of L-dopa did not rescue the loss of DA neurons in the diencephalon of zLRRK2-ΔWD40 morphant . This is consistent with the therapeutic effect of L-dopa in treating human PD condition where the treatment can only offer a temporary relieve of clinical symptoms , but cannot restore the degeneration of DA neurons [34] . The morphant phenotypes of the zLRRK2-ΔWD40 deletion seem to closely mimic the human condition of PD at both molecular and physiological levels . In addition , our preliminary study has shown that the over-expression of human point mutations , such as G2019S and G2385R , shows a similar impact on neural development as the WD40 deletion in zebrafish . This is consistent with the dominant effect of human point mutations , such as G2019S , to induce PD-like phenotypes in other animal models . Furthermore , MPTP treatment was shown previously to trigger the similar degeneration of DA neurons and locomotion behavior defects in zebrafish as in human [35] , [36] . Taken together , these studies have demonstrated that zebrafish can be used for studying PD-related neurodegeneration , and the deletion of WD40 domain in zebrafish provides a potential disease model for PD . It is noteworthy to point out that the neuron loss of zLRRK2-ΔWD40 can be observed as early as in the late stage of embryonic development . As a limitation of this model , the early-onset phenotype of zLRRK2-ΔWD40 does not fully recapitulate the late-onset PD phenotype in human . This difference may , at least partially , due to the fact that the knock-down effect of splice-blocking morpholino could be up to 90% in zebrafish , whereas PD patients usually carry heterozygous point mutations of LRRK2 . Consequently , the early-onset phenotype of zLRRK2-ΔWD40 may be due to more severe mutational effect of WD40 deletion than heterozygous point mutations in human . Furthermore , although the molecular function of LRRK2 is conserved between zebrafish and human , the neuronal system , including dopaminergic one , may not be fully conserved between two species . As a consequence , mutations of functionally conserved LRRK2 may show partially different phenotypes . It is not truly unexpected because it is rather uncommon for animal models to recapitulate the full phenotype of human disease . Several animal models for PD were developed in recent years . DJ-1 knockout mice show decreased motor functions , increased striatal dopamine level without the loss of DA neurons [37] , [38] and increased sensitivity to MPTP and oxidative stress [39] . Consistently , the knockdown of DJ-1 expression in zebrafish did not result in a loss of DA neurons , unless under the exposure to pro-oxidant hydrogen peroxide and the proteasome inhibitor MG132 [40] . In PINK1 knockdown [41] or knockout mice [42] , there were no changes in striatal dopamine level , nigral DA neurons numbers and motor activity . In Parkin knockout mice , mutants with the deletion of exon2 [42] showed no abnormalities compared to wild-type mice in terms of the nigral DA neurons numbers and motor activity , while the mutants with the deletion of exon3 [43] showed behavioral deficits , but without DA neurons loss . The conditional LRRK2 G2019S model [27] was reported to have no obvious neuropathological or motor abnormalities at 12 months of age . Over-expression of UCH-L1 in zebrafish did not result in a discernible phenotypic effect [44] . Therefore , the previous vertebrate models did not show , in a consistent fashion , the progressive loss of nigrostriatal dopaminergic neurons and motor defects . In Drosophila , the expression of wild-type and mutant forms of human α-synuclein lead to a progressive DA neuron loss [45] , and the loss could be suppressed by the over-expression of parkin [46] . Drosophila parkin-null mutants also showed motor deficits [47] and DA neuron degeneration [48] , [49] . Recently , Lee et al . found that the loss of Lrrk2 in Drosophila lead to impaired locomotive activity and degeneration of DA [20] . However , Wang et al's study did not confirm this observation and instead found an increased sensitivity to oxidative stress [19] . The over-expression of hLRRK2 wild-type or G2019S mutant allele in Drosophila resulted in loss of DA neurons , locomotor dysfunction and early mortality , which could be rescued by the administration of levodopa [50] . Although this invertebrate model recapitulates several features of human PD , a recent study showed that dLRRK2 is not an ortholog of hLRRK2 [51] , dampening the relevance and importance of this Dorsophila LRRK2 model for PD . In summary , we have demonstrated that zLRRK2 is an ortholog of hLRRK2 . As a vertebrate model , the zLRRK2-ΔWD40 morphant recapitulates some key molecular , physiological and behavioral hall-marks of PD . Together with other animal models , this potential vertebrate model provides opportunities to investigate the biological mechanisms underlying the development of PD . The fact that the locomotion defect of the zLRRK2-ΔWD40 morphant can be ‘treated’ by L-dopa also raises the possibility that this zebrafish model may be used for screening new drugs to treat PD . A TBLASTN analysis of the human LRRK2 protein against zebrafish cDNA sequences yielded two hits , XM_682700 and XM_682192 . Through reciprocal TBLASTX , we found that XM_682700 was the possible homolog of human LRRK2 ( hLRRK2 ) , whereas XM_682192 was the possible homolog of human LRRK1 . As indicated , XM_682700 is a predicted cDNA of zebrafish LRRK2 ( zLRRK2 ) with about 6 kb sequences . To verify whether it is a true coding gene , we blasted this sequence against the genome of zebrafish and identified 6 EST sequences ( BI884532 , EB935015 , BI882500 , AL918398 , BQ258400 and CD758533 ) . To identify the full length transcript of zLRRK2 , we performed RACE analysis by using mRNA isolated from the brain of adult fish and the sequence of BI884532 ( most 5′ end ) for designing the primer of 5′-RACE and AL918398 ( most 3′ end ) for designing the primer of 3′-RACE . Likewise , using the sequences of the same two ESTs , a pair of primers was designed to amplify the middle part of the zLRRK2 transcript . After cloning and sequencing , we identified a 9168 bp transcript carrying both start and stop codons . 5′ RACE and 3′ RACE were performed by using the GeneRacer Kit ( Invitrogen , USA ) according to manufacturer instructions . Gene Specific Primer for 5′ end is 5′ CTGCATTTCAGCAACACAGG 3′ and Gene Specific Primer for 3′ end is 5′ AAGTCCAGCGTGTAGCTGAGCGTGGAAATG 3′ . qRT-PCR was performed with HIGH CAPACITY CDNA REVERSE TRANSCRIPTION KIT and SYBR Green 1 PCR Master Mix . Gene specific primers are 5′GACTCCGAGGCGATACAG 3′ and 5′ CAAGGGCACTCAGACAGG 3′ . Internal control beta-actin primers are 5′ TGGCAAAGGGAGGTAGTTG 3′ and 5′GTGAGGAGGGCAAAGTGG 3′ . Wild type AB line and Tg ( DeltaD∶GAL4/UAS∶Kaede ) line zebrafish were maintained according to methods described in The Zebrafish Book ( Westerfield , 1995 ) . Details are provided in Protocol S1 . Procedure was followed by the method described in The Zebrafish Book ( Westerfield , 1995 ) . Gene specific primers are: 5′ TGCAAACGGAGGTAAAAACC 3′ and 5′AGATGATCCTGGTCCCACAG 3′ for zlrrk2; 5′AAGGATGGCTTGGAGGAC3′ and 5′CTCGGAGGGTGGAGTAGA3′ for th . PCR product was cloned into pGEMT vector for probe synthesis . For dat , 5′GGGGTTCAGTTCACCTCCTC3′ and 5′CATTAACCCTCACTAAAGGGAAGACTCCATCCCTCCCATAGC3′ ( with T3 promoter ) were used for PCR and probe synthesis . Three different morpholino antisense oligonucleotides ( translation start site of zlrrk2 , ATG–ACAACTCCTCTATTTCTGCCATGAT; intron 45 splice donor junction; EI–CACAAGCAGATTTATTAACCTGTGC; intron 44 splice acceptor junction , IE–GCTCCTGAAACACAGCATTAGGAAC ) were obtained from Gene Tools ( Philomath , OR ) and injected at the one- to two-cell stage . Details of splicing interfering mopholino design are provided in Protocol S1 . Efficacy of morpholinos directed against splice sites was evaluated using RT-PCR ( Figure S5 ) with forward primers F1-5′TGCAAACGGAGGTAAAAACC 3′ and in conjunction with reverse primer R1- 5′AGATGATCCTGGTCCCACAG 3′ . Dosage-dependent effects of morpholinos were observed ( Figure S4 ) . Gene specific primers for probe synthesis are 5′GTTGGCGTTCTGCCGGGT CC 3′ and 5′ AAAGCGGCCGCATTAAGCAGCGTTTCTCTCATTCTGCGG 3′ . Details are provided in Protocol S1 . The anti-zLRRK2 antibody used in this study is developed from the C-terminal ( within WD40 domain , CSTRKPKVHSEDQSR ) regions of LRRK2 . Western analysis was conducted using standard techniques . Details are provided in Protocol S1 and Figure S10 . The truncated protein of zLRRK2-ΔWD40 cannot be recognized by this antibody due to its absence of the WD40 domain . For plasmid rescue , pCI-neo vector ( Promega , USA ) harboring zlrrk2 or hLRRK2 cDNAs tagged by Flag and hLRRK2 G2019S or G2385R cDNAs tagged by myc were used . At 2 to 3 days after the microinjection of the linearized plasmids by SfiΙ , total fish homogenate was subjected to anti-Flag Western blot analysis , clearly showing that both the zlrrk2 and hLRRK2 cDNAs could be expressed in zebrafish ( Figure S7 ) . Upon confirming the expression of zlrrk2 and hLRRK2 cDNAs in zebrafish , the rescue experiment was performed by the co-microinjection of the plasmids and the ATG or splicing-blocking morpholinos into embryos . Dosage-dependent effects of plasmid over-expression were observed ( Figure S4 ) . For the rescue of the ATG morphants , only the hLRRK2 cDNAs was used , because the ATG morpholinos will block the protein expression of the zlrrk2 cDNA . For L-dopa rescue , L-dopa ( 1 mM ) ( Sigma ) was applied at 5 dpf stage and behavior test was performed at 6 dpf . Apoptosis assay was carried out by using In Situ Cell Death Detection Kit , TMR red ( Roche ) . Procedures are referenced from Michael Hendricks and Suresh Jesuthasan's work [52] . Details are provided in Protocol S1 . On a Zeiss LSM510 META confocal microscope , live embryos' brains were imaged using an Achroplan 10X/0 . 30 water immersion objective and alpha tubulin stained brains were imaged using a EC Plan-Neofluar 10x/0 . 30 objective . Projection of confocal z-stacks was done using Zeiss software . Video recording for the behavior analysis of 6 dpf larva were taken with a Sony HDR-SR12E . Behavior analysis was done using NIH ImageJ . Protein domains of zebrafish LRRK2 were predicted by SMART ( Simple Modular Architecture Research Tool , http://smart . embl-heidelberg . de ) . Each domains and whole protein of zLRRK2 and hLRRK2 were aligned by ClustalW2 ( http://www . ebi . ac . uk/Tools/clustalw2/index . html ) . The animal protein sequences of the COR and Kinase domains are obtained from Marin's study [14] . Details are provided in Protocol S1 .
Parkinson's disease ( PD ) is a degenerative disease of the brain ( central nervous system ) that often impairs motor skills , speech , and other functions . PD was long thought to be caused by environmental factors , but the discovery of several gene mutations in the patients ( mostly with familial form of PD ) clearly demonstrated the involvement of genetic factors in the development of PD . Among the identified genes , LRRK2 was discovered to be one of the most important genetic causes of PD . The biological function of LRRK2 was , however , largely unknown . In this study , we studied the function of LRRK2 in zebrafish by blocking the normal function of LRRK2 . The zebrafish showed features of neurodegeneration and locomotion defects , similar to those of PD patients . The defects of the fish could be rescued by expressing the normal protein of LRRK2 , and the locomotion defect could also be rescued by the administration of L-dopa that is commonly used for treating PD patients . We have therefore developed a zebrafish model for PD that can be used for understanding the mechanism underlying the development of PD and will be helpful for future screening of new drugs to treat PD .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "genetics", "and", "genomics/disease", "models", "developmental", "biology/neurodevelopment", "neurological", "disorders/movement", "disorders" ]
2010
Deletion of the WD40 Domain of LRRK2 in Zebrafish Causes Parkinsonism-Like Loss of Neurons and Locomotive Defect
Cell autonomous cancer dependencies are now routinely identified using CRISPR loss-of-function viability screens . However , a bias exists that makes it difficult to assess the true essentiality of genes located in amplicons , since the entire amplified region can exhibit lethal scores . These false-positive hits can either be discarded from further analysis , which in cancer models can represent a significant number of hits , or methods can be developed to rescue the true-positives within amplified regions . We propose two methods to rescue true positive hits in amplified regions by correcting for this copy number artefact . The Local Drop Out ( LDO ) method uses the relative lethality scores within genomic regions to assess true essentiality and does not require additional orthogonal data ( e . g . copy number value ) . LDO is meant to be used in screens covering a dense region of the genome ( e . g . a whole chromosome or the whole genome ) . The General Additive Model ( GAM ) method models the screening data as a function of the known copy number values and removes the systematic effect from the measured lethality . GAM does not require the same density as LDO , but does require prior knowledge of the copy number values . Both methods have been developed with single sample experiments in mind so that the correction can be applied even in smaller screens . Here we demonstrate the efficacy of both methods at removing the copy number effect and rescuing hits from some of the amplified regions . We estimate a 70–80% decrease of false positive hits with either method in regions of high copy number compared to no correction . CRISPR based loss-of-function screens have emerged as a powerful tool to interrogate multiple species and models [1] . The technology has been quickly adopted to identify essential genes in cancer , including several cancer cell line screens [2–4] . However , as reported in two studies [5 , 6] and further discussed by others [7] , genes in regions of copy number amplification display strong lethal phenotypes by CRISPR-Cas9 cutting ( as opposed to CRISPRi [8] ) , regardless of the true biological essentiality of the targeted gene . This results in a significant number of false positive hits in samples with large copy number alterations as is often the case in cancer models . One way of mitigating this problem of false positives would be to simply discard any hits found in amplified regions . This is a viable strategy when considering aggregate profiles [9] , but runs the risk of yielding many false negatives when looking at individual hits . Especially when copy number events are an important oncogenic driver and identifying the essential gene in the amplicon is of interest to target discovery [10] . Therefore , to fully leverage CRISPR based screens , it is important to understand and correct for the observed copy number bias . Here , we propose methods to correct for the copy number artefact , while rescuing the true positives within the amplicons . The corresponding R scripts are also provided ( https://doi . org/10 . 6084/m9 . figshare . 5140057 . v3 ) . To the best of our knowledge two other methods have recently been proposed in [11 , 12] . In this study wee used the data published by Munoz et al . [5] , where the copy number artefact has been observed ( Fig 1A ) , i . e . a negative correlation of sensitivity ( calculated as Log FC ) with copy number . To illustrate the methods , we focused on the astrocytoma cell line SF268 and the gastric cancer cell line MKN45 , as these two cell lines have amplicons where the driver has been well characterized , YAP1 and MET , respectively [13–16] . The sgRNA library used targeted 2722 human genes with an average coverage of 20 reagents per gene . In addition , a second screen performed on MKN45 , using a different library of genes with a coverage of 10 reagents per gene , was used to evaluate the methods described herein . We then evaluated our methods on the Avana dataset [11] . To account for the copy number artefact , we propose the Local Drop Out ( LDO ) method . LDO aims to correct phenotype scores for each guide by taking into account guide scores targeting the other genes in its direct genomic neighbourhood . It assumes that most genes display little or no phenotype upon knock-out in such screens ( ~2 weeks or less ) and does not rely on copy number measurements . If multiple neighbouring genes show similarly strong drop out values by exhibiting a significant reduction of viability score , it is assumed that the observed phenotype is due to a copy number effect rather than a true dependence of the cell line . This assumption is corroborated by observations made in large RNAi screens [17 , 18] where only a single or few genes are identified as drivers of focal copy number events . The density of the screen influences the size of the copy number events that can be detected: the higher the density of the genes selected to be included in the screen , the more focal the detected copy number events can be . The LDO method uses a two-step process: 1 . A list of potential hits is defined; 2 . The remaining “neutral” genes are used to estimate the copy number effect on viability and the viability are corrected based on the estimate . In step one , a list of potential gene hits is defined that minimizes false negatives , and in step two , allows the estimation of copy number effect on viability to be based on “neutral” genes . The potential hit list can be defined in several ways . Prior knowledge can be used , e . g . lists of pan-lethal genes available in the public domain , to determine an initial list of essential guides for consideration . Alternatively , we propose to identify cell line specific genes that are either essential or growth enhancing by calculating each guide’s vulnerability score compared to a weighted mean sensitivity of neighbouring guides not in that gene , i . e . assessing the difference between the dependence score of one guide against the average vulnerability observed in the guides targeting different genes on the same locus . The weighted mean sensitivity is calculated as follows . Let g be a guide in the set G of all guides in a specific chromosome or chromosomal arm , with gih the ith guide targeting gene h and Gh be the set of guides targeting gene h . Let E1 be the set of guides targeting known essential genes . An exponential distribution with parameter ω = 100’000 bp is used . Additionally , let the genomic position of guide g be xg and the viability score induced by guide g be Sg , then the weighted mean sensitivity , excluding essential guides and guides targeting the same gene , mS ( gih ) for guide gih can be written as: mS ( gih ) =1N∑g∈G\{Gh∪E1}Sge−ω|xgih−xg| With N=∑g∈G\{Gh∪E1}e−ω|xgih−xg| For each guide , the first iteration of the corrected sensitivity S1 value is obtained from subtracting the weighted mean sensitivity for that guide to the original sensitivity value without correction ( Sgih1=Sgih−mS ( gih ) ) . Using this measure , the guides with absolute values above the μth percentile across the entire genome are considered as guides displaying potential true phenotypes ( hits ) and are not used in the second iteration of the method ( by default μ = 85 ) . In this analysis , we have used a list of a priori essential genes compiled from [4] ( see Mat & Met for more details ) . Removing the known pan-essential genes is not a requirement of the method , but can improve the accuracy of the resulting CN correction . In particular in the case of successively located pan-essential genes which could otherwise be confused for a CN event . We have used μ = 85 since it represents a prior belief that we can expect about 15% of genes ( and therefore 15% of guides in our design ) to display a true phenotype in the screen independent of the copy number . The parameter can be modified , e . g . one might expect a larger percentage of genes displaying a phenotype in longer screens . This procedure is equivalent to increasing the set of essential guides in set E2 which is then sample specific and contains the set E1 and all the guides identified above the μth percentile . In the second step of the LDO correction , all guides below the μth percentile are used to fit a regression tree to estimate the copy number effect . These guides are highly enriched in guides showing no phenotypes or “neutral” gene guides , i . e . the set of guides g ∈ G\E2 . Here , a one dimensional regression tree is used to estimate the sensitivity of the guides as a function of the genomic location alone . The copy number effect identified by the regression is then removed from the original sensitivity score Sg to obtain the LDO corrected sensitivity score SgLDO . Specifically , the regression tree T formulates the copy number induced sensitivity SCN at position x as follows: SCN ( x|T ) =SCN ( x|{Sm , Rm}1M ) =∑m=1MSm1 ( x∈Rm ) Where {Rm}1M are subregions of the genome , and x is a genomic position . Sm are the estimated copy number induced sensitivity values in region Rm . Using only the guides g ∈ G\E2 , we try to find the regression tree T which minimizes the error: e ( T ) =∑g∈G\E2[Sg−SCNT ( xg|Sm , Rm ) ]2 with respect to Sm and Rm . In practice , a regularization term is added to avoid overfitting . Thus , the objective is to identify the tree T which minimizes the following term: minT∈T ( β ) [e ( T ) +α|T|] where |T| is the number of terminal nodes of the tree and the complexity parameter α measures the “cost” of adding another region Rm to the model . The higher the cost , the shallower the tree . Also additional constraints can be set on the universe T ( β ) of potential trees T . In particular , one can consider the universe T ( β ) with a minimum number β of guides per region Rm . The regression is performed iteratively with increasing values of the cost parameter α and constant value of β . By default the parameter α is initialised to 10−k with k = 3 and β is set to twice the mean number of guides per gene . The value k is iteratively decreased in increments ik set by default to 0 . 1 . This process decreases the complexity of the regression tree until all regions RF∈{Rm}1M contain at least 3 genes . The last resulting tree TLDO is used to calculate the LDO corrected sensitivity scores SgLDO defined by: SgLDO=Sg−SCN ( xg|TLDO ) In Fig 1B , the essentiality score before and after LDO correction is shown for the YAP1 amplicon in SF268 . In Fig 1B ( left ) , ANGPTL5 , KIAA1377 , C11orf70 , BIRC3 , BIRC2 , TMEM123 , MMP7 , and MMP20 display equivalently significant phenotypes believed to be entirely due to the copy number artefact . On the other hand , YAP1 shows a stronger phenotype relative to its neighbouring genes . In Fig 1B ( right ) , the resulting corrected sensitivity scores are shown in the YAP1 amplicon in SF268 . The copy number effect has been successfully removed and YAP1 still scores as significantly lethal , thereby being identified as the amplicon’s driver , as is expected from existing shRNA screens and reported elsewhere [13 , 14] . Overall , LDO removes the copy number effect beyond the YAP1 amplicon in SF268 and in MKN45 cell lines , as shown in Fig 1C . The number of guides with log2 ( CNA ) larger than 2 and LogFC below -0 . 5 is decreased from 98 to 37 guides in MKN45 and 267 to 38 guides in SF268 . Finally , and although this is not the main motivation for the development of this method , the LDO strategy can be used to predict copy number alterations in the screened samples ( see S3 Fig ) . Although the method applied in this screen was able to successfully recover the driver in the YAP1 amplicon , this is not always the case as shown in Fig 2A . From shRNA screens and other reports [15 , 16] , MET is expected to be the driver of this amplicon . Therefore , one could expect the MET guides to display a stronger relative drop out compared to the rest of the genes in the amplicon . However this was not the case and thus applying the LDO correction did not enable the recovery of MET as the driver of the amplicon . The degree of amplification does not appear to explain the lack of differential MET effect in MKN45 considering that the amplification in SF268:YAP1 is equivalent to what is seen in MKN45:MET . One potential reason for this lack of relative drop out is the quality of the guides used . The screen was rerun with different guide designs . The result for the MET amplicon in sample MKN45 is shown in Fig 2B and in this case , MET does display a stronger phenotype than the rest of the amplicon . This highlights the need for careful library design ( S1 Fig ) . To verify the generalizability of this method we applied LDO to the Avana data set of 342 cell lines screened with a genome-wide CRISPR library [11] . We used the guide level dependency scores as well as the CCLE [19] copy number provided by Meyers & al . The guide scores were further scaled so that the mean guide scores targeting nonessential and essential genes are equal to 0 and -1 respectively in each sample . The reference set of nonessential and essential genes established in [20] was used for the scaling . In Fig 3A we show that in the Avana data set LDO again markedly decreases the correlation between gene dependency and copy number compared to the uncorrected data . To assess the risk of over correction by LDO , we considered the recall of the guides targeting essential , nonessential and CN amplified regions ( i . e . regions with copy number > 5 ) . A similar strategy was used in [12] . In Fig 3B the recall curves for cell line DAN-G is presented before and after LDO correction . We notice that the recall for the essential genes is barely affected by the correction , indicating that the LDO method does not markedly impact the sensitivity to detect these genes . The recall curve for the nonessential group of genes is also not strongly affected by the correction as opposed to the recall of the amplified genes which is strongly reduced in DAN-G . To assess these recall curves across the whole dataset the area under the recall curve ( AURC ) was used . In Fig 3C the AURC before and after LDO correction is presented for all samples . The AURC for the essential genes remains unaffected by the correction while the AURC for the nonessential genes is slightly increased . The median AURC across samples is shifted from 87 . 3 to 87 . 1 and from 41 . 8 to 44 . 6 for the essential and nonessential genes respectively . This is in contrast to the shifts in median AURC observed for the amplified genes from 62 . 5 to 50 . 5 when using all samples or from 77 . 5 to 52 . 6 when considering only samples with at least 25 genes in amplified regions . In contrast to the LDO method , the GAM method is a supervised strategy requiring orthogonal data , such as copy number for the correction . To do this , we used a generalized additive model ( GAM , [21] ) framework and modelled the sensitivity to CRISPR-mediated gene knock-out as a function of copy number to yield an adjusted CRISPR-mediated gene knock-out estimate . In addition to its ability to leverage the copy number values when available , the potential benefit of this framework compared to LDO is that it can be extended to consider any arbitrary number of additional features ( both linked to artefactual or true effects ) potentially relevant for the purpose of modelling the phenotype ( e . g . gene expression , multi-alignment of guides , etc . ) . In this analysis , only the copy number measurements were used . Once the model has been fitted , the effect of the artefactual components of the sensitivity can be removed from the observed phenotype in order to keep the “biologically-relevant” component ( in this example only the artefactual copy-number effect is considered and removed ) . Unlike the LDO method , the GAM method is insensitive to the screen density and would be preferred should a sparse coverage of the genome be considered in the screen . Additionally the GAM method does not require a prior list of known essential genes to be performed . The GAM structure can be written as follows: E ( Sg ) =α+s1 ( x1g ) +⋯+sp ( xpg ) Where E ( Sg ) is the expected sensitivity of guide g; x1g , … , xpg are the predictor variables for g and s1 ( x1g ) , … , sp ( xpg ) denote the smoothing functions estimated by non-parametric means from the data . Finally , α is the intercept . Note the lack of a linker function in the above equation compared to the canonical GAM framework , since we consistently use the identity function . For the purpose of fitting the GAM to the data , we use the R implementation from the mgcv package [22] with default parameters , so that penalized thin plate regression spline models are used for the smoothing . This framework enables us to take into account an arbitrary number of predictor variables to model both linear and non-linear dependencies of the data . The aim is to remove from the measured sensitivity Sg the components of the model , which are deemed to come from artefactual predictor variables ( e . g . copy number ) but keep those coming from variables which are considered true predictors of biological sensitivity ( e . g . gene expression ) . Instead of x1g , … , xpg let us further differentiate the predictor variables into the artefactual variables x^1g , … , x^lg and the explanatory variables xl+1g , … , xpg so that the GAM corrected sensitivity can be written as SgGAM=Sg−∑i=1lsi ( x^ig ) In this presentation a single artefactual predictor x^1 is used which represents the copy number value at the position of guide g . The GAM corrected sensitivity score SgGAM can then be used in lieu of the original sensitivity score with the same hit-defining thresholds and interpretation . The correction of the copy number artefact in SF268 and MKN45 using GAM is shown in S2 Fig . The use of high-throughput CRISPR screens to identify cell autonomous cancer dependencies has become routine . However , as shown in previous studies , these screens display high rates of false positive hits in regions of high copy number amplifications . In this report , we describe two methods , Local Drop Out ( LDO ) and General Additive Model ( GAM ) , to correct for this copy number bias , thereby enabling the identification of true positive hits while reducing false positives substantially . In both cases the methods were developed with experimental setups in mind utilizing only a few number of cell lines , including single model experiments . Thus , making these methods appropriate for a broad range of experiments . As a result the CN artefact corrections proposed are performed at the level of single samples . We applied both methods to previously published screening data of 2722 genes performed in the SF268 and MKN45 cell lines . The utility of the methods were shown by way of two examples: first , the YAP1 dependency in SF268 was recovered , while removing 8 false positive genes from the hit list ( ANGPTL5 , KIAA1377 , C11orf70 , BIRC3 , BIRC2 , TMEM123 , MMP7 , and MMP20 ) ; second , the MET dependency in MKN45 was recovered in one of the two screens , while removing 3 false positive hits ( CAV1 , ST7 , and ING3 ) . Overall , the number of guides with log2 ( CNA ) larger than 2 and LogFC below -0 . 5 is decreased from 98 to 37 guides in MKN45 and 267 to 38 guides in SF268 when using LDO; with GAM the number of guides are reduced to 28 and 37 guides for MKN45 and SF268 respectively . Additionally the LDO method was applied to an external data set of 342 viability screens . There the method again markedly lowered the copy number effect on cell viability while retaining the sensitivity to known essential genes thus demonstrating the generalizability of the method to a larger data set . These methods , however , do have limitations . We observed that rescuing true positives within amplicons is only possible if the driver mutation in the amplicon of interest is displaying a stronger drop out relative to the neighbouring genes . Depending on the guides used , this is not always the case as demonstrated with MET in MKN45 in our first screen . Despite this caveat , both methods are still able to remove false-positives , although the true positive is not rescued in this case . We would argue that in a typical screening effort , the loss of a few true positives is preferred to a large amount of false positives , as would be the case if one does not correct for the copy number effect . Indeed , a lot of effort and resources can be spent chasing an elusive false positive . The second obvious alternative is to remove any amplified region from the subsequent analysis , which means a large amount of false negative hits , since those would not even be considered for further analysis , but also relies on prior available copy number measurements which is not always the case . Another limitation is that these methods are highly dependent on guide scores obtained in the screen which can be variable . In our MKN45: MET example , it is unclear what the reason for the difference in drop out of the driver is . Guide design could be an explanation , however if CRISPR genome editing does indeed generate two cellular responses in cancer cells as suggested in [6]: an early anti-proliferative DNA damage response and a later gene dependant effect , the number of doublings before harvesting could also be an explanation . Whereby cell lines with long doubling times would only undergo enough doublings to sustain the DNA damage response but not enough to signal a differential effect from the driver genes . This hypothesis however does not seem to fit with the doubling times of 29h and 44h for MKN45 and SF268 respectively as reported in the Cancer Cell Line Encyclopedia ( CCLE ) [19] . Outside of these limitations , each of the two methods presented offer different advantages to the correction for the copy number induced false positives in loss-of-function CRISPR screens . The LDO method can correct for the copy number artefact even when copy number is not known beforehand as long as the density of the CRISPR screen is high enough to capture the copy number events with confidence . As opposed to LDO which is unsupervised , the GAM correction is a supervised method requiring copy number measurements . It is however not dependent on high density screens and can additionally incorporate an arbitrary number of predictor variables in its model . As a supervised method GAM remains the appropriate method should the copy number information be available . The fact that LDO does not need any copy number information also enables the user to infer copy number alterations based on CRISPR screens by exploring the magnitude of the correction that was applied to the different genomic regions ( S3 Fig ) . To collect an initial list of essential genes the results from [23] was used . In particular the essentiality of each gene was established in [23] using a genome-wide single guide CRISPR screen in 4 cancer cell lines . The strength of the essentiality is reported as an adjusted p-value in the accompanying data . Here , the genes with a maximum adjusted p-value of 0 . 05 across all 4 cell lines are used as de facto essential genes if and only if the accompanying CRISPR score is also smaller than -1 . This results in a list of 814 potential essential genes ( S1 Table ) . The choice of the exponential decay function in the weighted mean calculation is arbitrary ( and any weighing function can easily be used instead in the provided scripts ) . Any monotonously decreasing function could be used , or , for example , a simple sliding window . The size of the window , or the value picked for ω in the exponential decay case , should be chosen so as to borrow the information from as many genes as possible while still remaining within the bounds of the expected copy number event sizes that are expected to be observed . The exponential decay function has the advantage of putting more weight to the genes in the direct neighbourhood of the gene of interest and thus even if the size of the window considered is relatively large the estimate remains relatively robust . Similarly the values for α1 , β1 , the minimum number of three genes per short CN event and the choice of only considering events with correction values larger than the 1 . 5 times the median average deviation of the background noise were chosen arbitrarily based on a priori expectation of the effects we wish to correct for .
Cancer vulnerabilities have been identified by systematically disrupting individual genes in cancer cells and observing the resulting effect on cell proliferation . In recent years , a new gene editing technique called CRISPR has made it easier and cheaper to disrupt genes by precisely and completely suppressing the function of individual genes by cutting through its DNA . However , an artefact of the approach yields false positives: using CRISPR to target genes in regions of the genome which are abnormally repeated , called copy number alterations ( CNA ) , has been shown to kill the investigated cells irrespective of the true essentiality of the amplified genes . This artefact is a particular issue when studying tumours , since CNAs are common in cancer . Additionally cancer-specific genes are known to selectively drive amplification , making the ability to assess the essentiality of genes in these regions even more important . Here we describe and provide the code to computationally correct for this artefact and recover the true essentiality of CNA genes .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[ "biotechnology", "genome", "engineering", "cancer", "detection", "and", "diagnosis", "medicine", "and", "health", "sciences", "engineering", "and", "technology", "synthetic", "biology", "genomic", "library", "screening", "health", "care", "crispr", "synthetic", "bioengineering", "oncology", "decision", "analysis", "management", "engineering", "screening", "guidelines", "molecular", "biology", "techniques", "synthetic", "genomics", "research", "and", "analysis", "methods", "bioengineering", "synthetic", "genome", "editing", "decision", "trees", "cancer", "screening", "molecular", "biology", "molecular", "biology", "assays", "and", "analysis", "techniques", "diagnostic", "medicine", "genetic", "screens", "library", "screening", "gene", "identification", "and", "analysis", "genetics", "biology", "and", "life", "sciences", "gene", "amplification", "health", "care", "policy" ]
2018
Correction of copy number induced false positives in CRISPR screens
Birds have a unique bone physiology , due to the demands placed on them through egg production . In particular their medullary bone serves as a source of calcium for eggshell production during lay and undergoes continuous and rapid remodelling . We take advantage of the fact that bone traits have diverged massively during chicken domestication to map the genetic basis of bone metabolism in the chicken . We performed a quantitative trait locus ( QTL ) and expression QTL ( eQTL ) mapping study in an advanced intercross based on Red Junglefowl ( the wild progenitor of the modern domestic chicken ) and White Leghorn chickens . We measured femoral bone traits in 456 chickens by peripheral computerised tomography and femoral gene expression in a subset of 125 females from the cross with microarrays . This resulted in 25 loci for female bone traits , 26 loci for male bone traits and 6318 local eQTL loci . We then overlapped bone and gene expression loci , before checking for an association between gene expression and trait values to identify candidate quantitative trait genes for bone traits . A handful of our candidates have been previously associated with bone traits in mice , but our results also implicate unexpected and largely unknown genes in bone metabolism . In summary , by utilising the unique bone metabolism of an avian species , we have identified a number of candidate genes affecting bone allocation and metabolism . These findings can have ramifications not only for the understanding of bone metabolism genetics in general , but could also be used as a potential model for osteoporosis as well as revealing new aspects of vertebrate bone regulation or features that distinguish avian and mammalian bone . The genetic basis of bone mineral density in the chicken is of both theoretical and practical importance . On the theoretical side , domestication is an example of intense directional selection causing wide-ranging differences in the phenotype of domestic animals compared to their wild progenitors . In the case of the domestic chicken , descended from the wild Red Junglefowl , bone metabolism is intrinsically coupled with the production of eggs . Hence , bone and fecundity phenotypes form a suite of correlated traits that have likely been tied with fitness during domestication . In addition , the comb , a sexual ornament in both male and female chickens [1 , 2] , appears to be tied with bone mineral density . In terms of practical applications , bone mineral density is a predictor of osteoporosis , a debilitating condition in both chickens [3 , 4] and humans [5] . Describing quantitative trait genes for bone mineral density can give new clues about the biological processes underlying osteoporosis . Also , to the extent that the variants are still segregating in commercial chicken flocks , they may provide targets for selection . Bone is made up of a mineral matrix , mostly hydroxyapatite , in an organic matrix dominated by collagen , lipids , proteoglycans and other bone structural proteins . Bone is continually being remodelled by osteoblasts , which produce bone components , and osteoclasts , which break them down . Bone metabolism in female birds is special in that they produce a medullary bone , which serves as a reservoir for calcium used in production of the eggshell [6 , 7] . Avian bone remodelling is quicker than mammalian and coordinated with the laying cycle [8 , 9] . Osteoporosis , loss of bone density , and bone fractures are a major health issue for layer chickens in production , likely exacerbated by the strains that high egg production and quick growth put on domestic chickens [10] . There is substantial genetic variation in bone traits in domestic layer chickens [11–13] . Studies involving dietary treatments suggest that the cause of osteoporosis in hens is not merely calcium deficiency [13 , 14] , and that there is a genetic basis to bone strength independent of calcium metabolism . With osteoporosis also being a serious health problem in humans , genetic mapping studies in humans and animal models have been performed to search for genes and processes involved in bone density . As is the case with heritable quantitative traits in general , the bone mineral density loci that have been identified so far leave most of the genetic variance unexplained ( see review by [15] ) . Linkage mapping rarely has the resolution to isolate single molecular genes , but linkage mapping in humans has found variants affecting bone density in genes such as LRP5 [16–18] and BMP2 [19] . Genome-wide association studies , on the other hand , bring associated regions down to few or sometimes a single gene . In recent years , several such association studies have been published , implicating common genetic variants in human bone density variation . The hits include known bone-related genes such as osteoblast regulator osterix , osteoclast regulators RANKL and osteoprotegerin , and Wnt pathway genes such as β-catenin and WNT16 [20–31] . However , even with the high resolution of a GWAS , it is not necessarily the gene closest to the variant that is the causal gene . Authors have successfully employed eQTL mapping or genetical genomics and systems genetics approaches such as coexpression networks , to isolate quantitative trait genes for bone density in mice , including Alox5 [32] , Alox15 [33] , Bicc1 [34] , Asxl2 [35] and Darc [36] . Our aim is to study the genetics of bone traits in the chicken using modern genomics methods . Due to the divergence of domestic chickens from Red Junglefowl during domestication , we can use a wild by domestic chicken cross as a study system for bone genetics . Quantitative trait locus ( QTL ) mapping as a top down mapping approach has been widely successful when it comes to finding chromosomal regions associated with various traits . However , going from the quantitative trait locus , usually a broad chromosomal region harbouring many genes , to causative genes or even variants , is difficult . The main obstacle to quantitative trait gene identification is the long linkage blocks in the pedigrees employed for mapping experiments . Making an advanced intercross line is one strategy to increase informative meioses and mapping resolution [37] . The chicken has a high recombination rate and a relatively compact genome compared other vertebrate models such as the mouse [38] . To aid quantitative trait gene identification , we turn to genetical genomics or expression QTL ( eQTL ) mapping to female femoral gene expression measured by microarrays . While gene regulation is complex , regulatory variation in molecular intermediates can be assumed to have simpler genetic architecture than the phenotypic trait , which is affected by the combined output of many such molecular pathways . In cases where the causative variant acts by means of gene expression , expression QTL mapping helps by reducing the number of candidate genes to the ones whose expression map to the QTL region . Beyond overlapping bone QTL and expression QTL , we consider the regression between bone trait and gene expression level of these positional candidate genes . In this way , we only retain candidates whose gene expression is associated with the bone trait . Beginning with the medullary traits , we find two candidate causal genes , each for a different QTL . These genes are drebrin ( DBN1; NM_205499 ) for diaphyseal medullary area on chromosome 13 , and collagen type XI alpha 1 ( COL11A1; represented by the EST sequence 603372847F1 ) for diaphyseal medullary content on chromosome 8 . Additionally , a QTL for diaphyseal medullary area on chromosome had three candidates: EST probesets 603847051F1 , 603961442F1 and 603846396F1 . These correspond to two unknown cDNA clones . There are three cortical trait QTL with a single candidate each . Heat shock transcription factor family member 5 ( HSF5; ENSGALG00000001031 ) is a candidate for diaphyseal cortical thickness on chromosome 19 . Also , tissue specific transplantation antigen P35B ( TST3A; ENSGALG00000016129 ) is a candidate for diaphyseal cortical area on chromosome 2 . This QTL colocalises with another for diaphyseal cortical thickness , which has both TST3A and KH domain-containing , RNA-binding , signal transduction-associated protein 3 ( KHDRBS3; ENSGALG00000016203 ) as candidates . Mitochondrial ribosomal protein S18A ( MRPS18A; ENSGALG00000010296 ) and an uncharacterized gene from the Ensembl database ( ENSGALG00000010303 ) are candidates for diaphyseal cortical thickness on chromosome 3 . Immunoglobulin superfamily , member 9B ( IGSF9B; ENSGALG00000001450 ) and POU class 2 associating factor 1 ( POU2AF1; NM_204175 ) are also candidates for a diaphyseal cortical thickness QTL on chromosome 24 . Finally , on chromosome 1 we find an uncharacterized Ensembl gene ENSGALG00000021181 and EST probeset 603234378F1 as candidates for diaphyseal cortical density . However , both these probesets correspond to the predicted gene LOC771935 ( XM_001235144 . 2 ) , which has since been removed from GenBank . We also detect gene expression candidates for total bone and bone circumference . Glucosamine ( N-acetyl ) -6-sulfatase ( GNS; NM_001199559 ) is the sole candidate for a diaphyseal total bone density and a diaphyseal endosteal circumference QTL on chromosome 1 . Also , centromere protein O ( CENPO ) , represented by EST 603256840F1 , is a candidate for a diaphyseal endosteal circumference QTL on chromosome 3 . On chromosome 13 , secreted protein , acidic , cysteine-rich ( osteonecin ) represented by the EST probeset 603470410F1 and xylosylprotein beta 1 , 4-galactosyltransferase , polypeptide 7 ( B4GALT7; NM_001039911 ) are candidates for a metaphyseal total content QTL . This QTL colocalises with the metaphyseal medullary area QTL for which drebrin is a candidate . B4GALT7 and osteonectin are also among the candidates for a third , colocalising medullary content QTL . These genes may affect both total bone content and medullary content . In a similar fashion , TSTA3 , KHDRBS3 are candidates for diaphyseal total bone content , as well as the cortical traits mentioned above . However , in the case of diaphyseal total content , the QTL also has a third candidate: protein tyrosine kinase 2 ( PTK2 ) , represented by EST probeset 603469577F1 . We also detect Ras-related protein Rab-24 ( RAB24 ) and SUMO-interacting motifs containing 1 ( SIMC1 ) as candidates for diaphyseal total area on chromosome 13 . Finally , we compared our candidate quantitative trait genes with results from genome-wide association studies ( GWAS ) in mice and humans . 30 out of 76 candidate probesets could be mapped to a mouse ortholog . Three of them were associated with bone mineral density in the GWAS by Farber et al . [35] . These genes were somatostatin receptor 5 ( SSTR5; NM_001024834 ) , calcium channel , voltage-dependent , T type , alpha 1H subunit ( CACNA1H; ENSGALG00000005215 ) , mitochondrial ribosomal protein S18A ( MRPS18A ) , and glycoprotein 1b , alpha polypeptide ( GP1BA; ENSGALG00000021693 ) . These genes are located in QTL that have other associated candidates , but these previous associations increase our confidence in them as potential causative genes . We also searched for our candidates among the top associated genes for bone mineral density in the dbGAP Association Results Browser , and found no overlap . Hotspots of trans-eQTL occurring in the same region could reflect regulatory variants with many downstream effects . We searched for such hotspots by comparing the number of overlapping trans-eQTL with the number of trans-eQTL observed with simulated uniformly distributed eQTL intervals . In this way we found 12 clusters of trans-eQTL , making up putative hotspots on chromosomes 1 , 2 , 3 , 4 , 5 and 25 ( see S4 Table ) . Four of these hotspots overlap female bone QTL . In particular , the hotspot on chromosome 3 overlaps the cortical thickness QTL for which we detect candidate genes MRPS18A and ENSGALG00000010303 . Further dissection of these loci would be necessary to know whether the clusters are genuine trans-regulatory hotspots involved in bone function . Drebrin ( DBN1 ) is the sole candidate for one of our medullary QTL . Drebrin binds actin filaments and is known for its role in the nervous system . However , it is also expressed in lymphocytes and involved in connecting chemokine receptor CXCR4 to the cytoskeleton [39] . This receptor , in turn , is involved in differentiation of osteoclasts from hematopoietic cells [40] . While drebrin has not been studied in connection with bone metabolism , we hypothesise that eQTL effects on drebrin expression could affect medullary bone through osteoclasts . We also detect collagen XI alpha 1 as a candidate for medullary content on chromosome 13 with multiple sources indicating effects on bone . Mutations in this gene cause the rare human chondrodysplasias fibrochondrogenesis [41] , Marshall syndrome and Stickler syndrome [42 , 43] , and may predispose to osteoarthritis [44] . Cell culture experiments suggest that it inhibits osteoblast differentiation [45] . We find HSF5 to be associated with cortical thickness . Heat shock factors ( HSF ) regulate heat shock protein ( HSP ) genes and activate them in response to cellular stressors . Furthermore , there is some evidence that heat shock responses may be involved in bone formation . Bone growth can be stimulated by heat , and heat treatment increases mineralisation and heat shock protein expression in cell culture [46] . Molecular studies of the RANKL promoter and HSF2 knock-down suggest that this osteoclast regulator is regulated by heat shock factors [47 , 48] . However , there is little known about HSF5 specifically . TSTA3 , our candidate for a diaphyseal cortical area QTL , is involved in protein glycosylation by contributing to the synthesis of GDP-fucose , which is used by fucosyltransferases in glycosylation of cell adhesion proteins . Knockout studies of TSTA3 in mice suggest that fucosylation is needed to regulate myeloid cell differentiation by Notch signalling [49] . Notch also regulates both osteoblast and osteoclast differentiation [50 , 51] . There are two additional candidates for the colocalising cortical thickness QTL: KHDRBS3 , which encodes an RNA binding protein , and PTK2 , protein tyrosine kinase . PTK2 is involved in osteoblast differentiation [52 , 53] . Along with the unknown Ensembl gene , ENSGALG00000010303 , MRPS18A is a candidate for diaphyseal cortical thickness , whilst MPRS18A encodes a mitochondrial ribosomal protein . These ribosomes translate proteins required for mitochondrial function . Hence mutations in mitochondrial ribosomal proteins can cause oxidative phosphorylation deficiencies ( such as OMIM phenotypes 614582 , 611719 , 610498 ) [54] . Such defects can affect diverse organ systems , since mitochondrial function is crucial for cellular metabolism . To our knowledge , no such disorders with specific bone-related symptoms are known , although there is still the possibility of mitochondrial effects on bone . Active osteoclasts , for instance , are rich in mitochondria [55] . Quantitative changes in mitochondrial activity could influence the balance of bone remodelling . Finally , we find two candidates for a cortical thickness locus on chromosome 24: IGSF9B and POU2AF1 . There is little known about IGSF9B , except that it is involved in cell adhesion at synapses [56] . POU2AF1 , however , is known as a transcription factor in B cells of the immune system . It regulates B cell maturation , and is mutated in some forms of leukaemia [57 , 58] . While no direct connection between POU2AF1 expression and bone is known , the immune system affects bone regulation . For example , B cells produce cytokines with effects on bone , including osteoclast-stimulating RANKL [59–61] . GNS is our single candidate for a couple of colocalising QTL for diaphyseal total bone and endosteal circumference on chromosome 1 . It encodes a glucosamine sulfatase that breaks down heparane sulphate . Heparan sulphates are polysaccharides that make up part of the extracellular matrix and are part of proteoglycan glycoproteins . As such , heparan sulphates have regulatory functions in development , including regulation of osteoblasts [62 , 63] . Mutations in the GNS gene cause mucopolysaccharidosis type IIID ( Sanfilippo syndrome D; OMIM phenotype 252940 ) [64 , 65] . It is a lysosomal enzyme , and loss of function causes accumulation of heparan sulphate in the organelle with deleterious effects on several organ systems . Symptoms includes effects on bone , though these may be secondary effects [66] . Taken together , this leads to the speculation that the GNS eQTL may affect bone density through either a signaling effect of heparan sulphate or through a lysosome-dependent mechanism . CENPO , a candidate for an endosteal circumference QTL , encodes a centromere protein and is involved in chromosomal segregation during mitosis [67] . While proliferation of different bone cell populations is important for bone metabolism , its involvement in bone is unknown . Another locus for total bone content has two candidates: osteonectin and B4GALT7 . Osteonectin is a glycoprotein produced by osteoblasts . It contributes to mineralisation by binding mineral crystals and linking to collagen , and it functions in bone remodelling [68 , 69] . B4GALT7 is another candidate involved in protein glycosylation , a glycosyltransferase that participates in synthesis of proteoglycans , and in particular heparane sulphate . Mutations in human cause Ehler—Danlos syndrome ( OMIM phenotype 130070 ) , affecting connective tissue and causing skeletal deformations [70–72] . Finally , a locus for diaphyseal total area has RAB24 and SMIC1 as candidates . RAB24 is one of the small GTPases of Ras-related proteins that are usually involved in intracellular trafficking . However , RAB24 seems to participate in cell division and chromosome segregation [73] . Potential functions in bone are unknown , but RAB24 was down-regulated in a gene expression study of stimulated mineralisation in the osteoblast-like mouse cell line MCT3-E1 [74] . The other candidate at the same QTL , SIMC1 , regulates the protease calpain 3 in skeletal muscle [75] . The number of mouse GWAS candidates for bone density that were also found as candidates in our study was restricted to four shared associations . One of them is MRPS18A , which has been discussed above . The others are from QTL that contain several gene expression candidates . SSTR5 encodes one of the receptors for the hormone somatostatin . It regulates growth hormone secretion , and mutations in the gene are associated with acromegaly [76 , 77] . The site of action of growth hormone regulation should be the pituitary , but there is also evidence somatostatin and its receptors affect bone precursor cells [78 , 79] . CACNA1H encodes , Cav3 . 2 , a voltage gated calcium channel subunit . It is expressed in osteoblasts and chondrocytes during bone development in mice [80 , 81] . GP1BA encodes a von Willebrand protein receptor expressed on the surface of platelets . Mutations cause bleeding diseases Bernard-Soulier syndrome ( OMIM phenotype number 231200 and 153670 ) and platelet-type von Willebrand disease ( OMIM phenotype number 177820 ) . von Willebrand factor can interact with osteoprotegrin to regulate osteoclast differentiation [82] , and a transgenic GP1BA von Willebrand disease model has increased bone mass and reduced osteoclast activity [83] . We found no overlap between our candidate quantitative trait genes and the top associations for human bone traits in the dbGAP Association Results Browser . However , the Framingham osteoporosis study found a region around KHDRBS3 as a potentially pleiotropic association with lumbar spine and femoral neck bone mineral density [84] . As mentioned previously in connection with the candidates , several of the genes highlighted here cause rare Mendelian diseases in humans , some of them with known skeletal phenotypes . Local cis-eQTL effects need not be as drastic as the loss-of-function mutations that cause Mendelian diseases , and could be restricted to the bone by tissue-specific regulatory elements . Based on the functional literature and the above associations , many of the genes found are plausible candidates for bone traits . However , we find a number of unexpected or unknown candidates . This is reasonable given the dearth of knowledge about the genetic machinery of complex traits in general . However , it could also be due to peculiarities of chicken bone metabolism as compared to mammals such as humans and mice . Previously unknown or poorly understood genes , coupled with the relative lack of human and mouse GWAS overlaps in our data , potentially indicate the novel aspect of chicken bone metabolism . In the female chicken femoral bone , calcium is continually removed from bone to supply the forming eggshell on a 24-hour basis throughout laying . This might require genetic mechanisms different from those in mammals . Understanding this mechanism will offer novel insights into bone metabolism . We also find that several known genes involved in bone , including Matrix Gla protein , TRAF2 , PDZRN3 and PDZRN4 display cis-eQTL effects . While they are not candidates for the phenotypic QTL detected in this study , this regulatory variation in bone genes suggest that there is more subtle variation in bone metabolism between wild and domestic chickens . Among the most highly expressed genes that also have an eQTL , we note BF1 and BF2 genes on chromosome 16 , involved in immunity in the chicken , the MHC class II beta chain gene BLB1 , the MHC B-G antigen , and cathepsin S involved in degrading protein antigens for MHC II presention . While evolution of the immune system under domestication is not the focus of this work , we note that this suggests changes to adaptive immunity . Genetic variation affecting bone traits between wild and domestic genotypes is of interest to the evolutionary study of chicken domestication , and potentially provides insights into the molecular genetic regulation of bone mineral density . However , variation present within domestic breeds is most useful to breeders . Whether the QTL we detect segregate in contemporary domestic chickens can only be found by studies of these populations . We do at least find some replicability of the results of the F2 generation of the intercross used in this work [85 , 86] . When comparing the advanced intercross bone results to the previous F2 results , about half of the F2 QTL overlapped at least one of the F8 QTL . Possible reasons for differences between F2 and advanced intercross mapping include thee limited power to detect small-effect QTL and the larger linkage blocks causing the F2 study to detect the aggregate effect of several linked QTL rather than single QTL . The latter phenomenon should be more pronounced for polygenic traits with epistatic interactions , and bone traits display both a large number of loci and considerable epistasis . Expression QTL mapping is an established genomics approach that has had several successes in quantitative trait gene identification in animals . The localisation of a QTL from a linkage study is always rough , due both to the inherently limited resolution of experimental crosses , and the statistical uncertainties of QTL mapping . Our candidate gene prioritisation builds on the overlap of 1 . 8 LOD drop intervals , corresponding to approximate 95% confidence intervals , between QTL and eQTL . In the next step , we test for an association between the gene expression level and the bone phenotype . We will suffer false negatives when variants act by other means than gene expression changes in adult bone tissue . Conversely , a candidate gene found might not be causative if the gene expression is correlated with the trait for some other reason , such as the gene being downstream of the causative gene in a regulatory pathway . For instance , if the bone QTL causes the proliferation of some population of bone cells , causing increases in mRNA of cell division related genes such as CENPO or RAB24 and mitochondrial protein genes such as MRPS18A . Some of the QTL intervals have many candidate genes that are highlighted by our approach ( see S2 Table ) . While it is possible for a QTL region to harbour multiple causative genes , false positives are likely among these candidates . In the end , conclusive evidence for a quantitative trait gene can come through experimental manipulations in cell culture or transgenic animals . Our QTL and eQTL mapping results provide compelling candidates for such experiments . In conclusion , we find several potential quantitative trait genes for bone traits in the chicken , using to our knowledge the first eQTL analysis of bone tissue in the chicken . They are supported by genetical genomics evidence in the form of QTL-eQTL overlap and trait-expression associations . While the genes are not particularly obvious candidates , some of them have connections to bone metabolism in the literature , and a handful have been associated with bone mineral density in mice or humans . Further investigation of the molecular mechanisms of these potential quantitative trait genes could reveal unexplored pathways of bone density regulation . The study was approved by the Regional Committee for Ethical Approval of Animal Experiments ( Jordbruks verket DNR# 122–10 ) . Birds were sacrificed by cervical dislocation and decapitation , as per the guidelines of the permit . The intercross was established from one Red Junglefowl male originating from Thailand and three White Leghorn females of the L13 line . QTL mapping has previously been performed in the F2 generation ( see [85 , 86] ) . The intercross was maintained with approximately a hundred individuals per generation at Linköping University and expanded for QTL mapping in generation eight . Five batches of eighth generation intercross chickens were hatched and kept at the Linköping University chicken facility . They originated from 107 pairings of 122 F7 individuals . The chickens lived in 3 x 3 meter pens with three levels , access to perches , and food and water ad libitum . Chickens were culled when they were 212 days old . The study was approved by the Regional Committee for Ethical Approval of Animal Experiments . Femurs were dissected out after sacrifice , the right stored frozen at -20°C for bone phenotyping while a piece cut from the left was frozen in liquid nitrogen and stored at -80°C for RNA isolation . We used peripheral quantitative computer tomography ( Stratec XCT—Research SA , Stratec Medizintechnik , Germany ) to measure a metaphyseal ( 6% of femur length ) and a diaphysial ( 50% of femur length ) cross section from each femur . We used the CORTMODE1 setting with a density threshold of >1000 mg/cm3 to measure cortical bone , PEELMODE2 with a threshold of 1000 mg/cm3 to measure endosteal area , density and bone content , PEELMODE2 with thresholds 1000 mg/cm3 and 150 mg/cm3 to measure medullary area , density and bone content , and PEELMODE2 with a threshold of 150 mg/cm3 to measure total area . In total , we gathered bone phenotypes from 227 male and 229 female chickens . Female chickens were also tested in egg-laying trials , as detailed in [87] . There were two fecundity trials of two weeks each . In the first , eggs were collected daily , and in the second , females were given two dummy eggs initially , and kept laid eggs . Since the second trial was closest to the time of sacrifice , we considered the number of eggs produced during this trial as a possible covariate for mapping of bone phenotypes . Females that produced no eggs at all were excluded . Bone and fecundity phenotypes from this subset of the intercross that was assayed for gene expression have previously been used to test for associations between phenotypes and gene expression of comb candidate genes; see [87 , 88] . Summary statistics for traits are given in S5 Table . The chickens were genotyped for 652 SNP markers spread across the sequenced part of the chicken genome . DNA was isolated from blood samples using a standard TRIS extraction . Genotyping was performed with the Illumina Golden Gate platform at Uppsala Seq and SNP platform . We performed QTL mapping with R/qtl [89] and Haley-Knott regression using forward-selection for multiple-QTL models and pairwise scans to search for epistasis . All analyses included sex , batch and body weight as covariates . For female traits we also considered egg laying in the second fecundity trial as a possible covariate . If there was a significant association between the bone phenotype in question and the total weight of eggs produced we included the egg covariate in QTL mapping . To adjust for cryptic relatedness structure , which can be a confounder in advanced intercrosses , we applied principal component analysis on the genotype matrix , and included principal components as covariates in QTL mapping . Empirical significance thresholds were established by means of permutation tests , shuffling the phenotypes while preserving the correlation structure of the genotypes . Genomic support intervals around the QTL were formed with the 1 . 8 LOD drop method [90] . See S1 Data for marker informativeness values , and S2 Data for phenotype and genotype data in R/qtl format . At the time of dissection , a piece was cut from the middle of the left femur and stored in liquid nitrogen . Bone samples were disrupted with a hammer while frozen and then homogenised on a FastPrep 24 instrument using ceramic beads ( Lysing matrix D , MP Biomedicals ) and Tri reagent ( Ambion ) . RNA was isolated using Tri reagent according to the manufacturer's protocol . RNA was further purified with the Fermentas GeneJet RNA purification kit ( Thermo Scientific ) . After treatment with DNAse I , double-stranded cDNA was synthesised with a combination of Fermentas RevertAid Premium First-strand cDNA synthesis kit , DNA polymerase I , RNase H and T4 DNA polymerase ( Thermo Scientific ) according to protocols provided with the kit . The cDNA was labelled with the NimbleGen One Colour labelling kit ( Roche ) and hybridised to NimbleGen 12x135k custom gene expression microarrays . Scanning was performed with a NimbleGen microarray scanner ( Roche ) . Subarrays were discarded due to low fluorescence intensity or visual uneven fluorescence resulting in 125 individual samples . Micoarrays were designed to cover all RefSeq and Ensembl genes as well as a database of ESTs . EST probesets were annotated by alignment to the chicken genome ( version 2 . 1/galGal3 ) with BLAT [91] ( see S3 Dataset ) . The name of each probeset used in tables contains the database accession ( from RefSeq , Ensembl or the NCBI EST database ) of the gene model or EST sequence used to design it . Selected gene expression data with bone and fecundity phenotypes from this subset of the intercross has previously been used for targeted genetical genomics of QTL regions for comb size; see [87 , 88] . Microarrays have been uploaded to ArrayExpress under accession number E-MTAB-3141 . Expression QTL mapping was performed with Haley-Knott regression as implemented in R/qtl [89] . Local , putative cis-acting , eQTL were mapped within an interval of 100 cM around the position of the probeset . The interval was expanded to the closest flanking markers spanning at least 50 cM in each direction . Global trans-acting eQTL were mapped using the entire genetic map . A probeset was assigned a cis-eQTL when the logarithm of odds passed the cis-threshold on any markers in the window around the location of the probeset , and a trans-eQTL if it passed the ( higher ) trans-threshold somewhere else in the genome . 1 . 8 LOD drop intervals , expanded to the closest markers , were used to form confidence intervals around eQTL . Thresholds were generated by permutation separately for cis and trans associations either using the whole-genome or the 100 cM regions around probeset locations . Each iteration , the individual identities were resampled , a 100 probesets were subsampled , and the maximum LOD score from an eQTL scan of this permuted dataset was saved . The process was repeated 10000 times , and the 95th percentile LOD score was used to generate the significance threshold . We used a two-phase method to search for quantitative trait genes for bone phenotypes: 1 ) we overlapped 1 . 8 LOD drop intervals from bone QTL and expression QTL; 2 ) tested for an association between probeset expression levels and bone traits with a regression model including body mass as covariate . Also , the same egg fecundity phenotype was included as covariate , in cases where the QTL in question was detected with a fecundity covariate . The regression between probeset expression level and bone trait included the body weight as a covariate . p-values from the t-test of the regression coefficient were adjusted by Bonferroni correction for the number of uncorrelated cis-eQTL ( those where p-value for pairwise correlation test > 0 . 05 ) in the interval . To investigate clustering of diastal trans-eQTL , we simulated uniformly placed eQTL on an interval the size of the sequenced chicken genome and counted the maximum coverage of simulated eQTL intervals in 1000 iterations . We regard any region with coverage above the 95th percentile of the simulated maximum coverage , which was 22 , as an eQTL cluster . We compared our candidate quantitative trait genes with previously published genome-wide association studies ( GWAS ) in human and mouse . We used the mouse GWAS data from [35] and the human top associations catalogued in the dbGAP Association Results Browser ( http://www . ncbi . nlm . nih . gov/projects/gapplusprev/sgap_plus . htm ) . Using Ensembl version 70 , we mapped Ensembl and RefSeq probesets to Ensembl gene identifiers and to orthologous human and mouse genes . In the mouse dataset we extracted the p-values for SNPs in 200 kb windows around the start of the Ensembl gene model . The threshold for genome-wide significance used in the original study was 4 * 10–6 . For the human associations , we exported all associations for bone mineral density in humans from dbGAP Association Result Browser with a p-value less than 10–5 . We mapped the pair of flanking genes listed in the Association Results Browser to chicken Ensembl gene identifiers and overlapped them with the list derived from our results .
In this work we seek to further the understanding of bone genetics by mapping bone traits and gene expression in the chicken . Bone in female birds is special due to egg production . In this study , we combine the genetic mapping of bone traits with bone gene expression to find candidate quantitative trait genes that explain the differences between wild and domestic chickens in terms of bone production . The concept of combining genetic mapping and gene expression mapping is not new , and has already been successful in isolating bone-related genes in mammals , however this is the first time it has been applied to an avian system with such unique bone modelling processes . We aim to reveal new molecular mechanisms of bone regulation , and many of the candidates we find are new , highlighting the potential this technique has to identify the potential differences between avian and mammalian bone biology .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[]
2015
Genetic Regulation of Bone Metabolism in the Chicken: Similarities and Differences to Mammalian Systems
Genome-wide association studies ( GWAS ) have discovered thousands loci associated with disease risk and quantitative traits , yet most of the variants responsible for risk remain uncharacterized . The majority of GWAS-identified loci are enriched for non-coding single-nucleotide polymorphisms ( SNPs ) and defining the molecular mechanism of risk is challenging . Many non-coding causal SNPs are hypothesized to alter transcription factor ( TF ) binding sites as the mechanism by which they affect organismal phenotypes . We employed an integrative genomics approach to identify candidate TF binding motifs that confer breast cancer-specific phenotypes identified by GWAS . We performed de novo motif analysis of regulatory elements , analyzed evolutionary conservation of identified motifs , and assayed TF footprinting data to identify sequence elements that recruit TFs and maintain chromatin landscape in breast cancer-relevant tissue and cell lines . We identified candidate causal SNPs that are predicted to alter TF binding within breast cancer-relevant regulatory regions that are in strong linkage disequilibrium with significantly associated GWAS SNPs . We confirm that the TFs bind with predicted allele-specific preferences using CTCF ChIP-seq data . We used The Cancer Genome Atlas breast cancer patient data to identify ANKLE1 and ZNF404 as the target genes of candidate TF binding site SNPs in the 19p13 . 11 and 19q13 . 31 GWAS-identified loci . These SNPs are associated with the expression of ZNF404 and ANKLE1 in breast tissue . This integrative analysis pipeline is a general framework to identify candidate causal variants within regulatory regions and TF binding sites that confer phenotypic variation and disease risk . Genome-wide association studies ( GWAS ) have identified more than 90 genomic loci and common genetic variants associated with breast cancer [1–5] . The single nucleotide polymorphisms ( SNPs ) associated with breast cancer have been shown to be enriched in DNA regulatory regions [6 , 7] , with few residing in coding regions of genes . The mechanisms by which most of these variants contribute to breast cancer biology remain unknown [8–11] . The effects of putative causal non-coding SNPs are challenging to interpret as they may alter transcription factor ( TF ) binding sites [12] , lncRNA structure [13] , splicing [14] , transcription start or termination signals , or DNA shape [15] . Non-coding SNPs that alter TF binding sites are the most easily interpreted because they have the potential to modulate gene expression to mediate their effects on disease risk [16] . Therefore , it is possible to identify putative causal SNPs by focusing on those that alter TF binding sites in breast tissue . TF dysregulation is a hallmark of many cancers [17 , 18] . Genes encoding TFs in tumor cells are often amplified , deleted , rearranged via chromosomal translocation , or subjected to point mutations that result in a gain- or loss-of-function [18] . For example , transcriptional amplification of c-Myc reduces rate-limiting constraints for tumor cell growth and proliferation; high c-Myc expression correlates with tumor aggression and poor clinical outcome [19] . Estrogen receptor ( ER ) is a TF that regulates cell proliferation , which is the defining feature of luminal breast cancers [20] . Identifying the full set of TFs that function within a cell type remains a challenge . Enzymatic accessibility assays identify open chromatin in the genome , which is an indirect measure of regulatory element activity and TF binding events [21–25] . TFs that directly or indirectly recruit cofactors , such as histone modifiers and nucleosome remodelers , recognize sequence motifs that are enriched in regions of open chromatin characterized by active histone marks and enzymatic hypersensitivity peaks [26 , 27] . These cofactors maintain the chromatin structure at regulatory elements . Therefore , one strategy for identifying the functional TFs in a cell is to query the sequence underlying accessible regions for over-represented motifs [28–31] . Alternatively , depletion of signal within hypersensitive regions ( footprints ) [30 , 31] coupled to motif analysis can be used to infer TF binding . The reliance on hypersensitivity footprinting to define a near-comprehensive set of TF motifs is limiting because many TFs do not have footprints for biological reasons [31] and the enzymes exhibit sequence-specificity that can be misinterpreted as footprint signatures [31–33] . These methods strictly identify motifs that are over-represented within regions of open chromatin , but families of TFs often contain paralogous DNA binding domains and thus recognize indistinguishable sequence motifs . One can directly measure the expression of TFs to identify putative functional TFs among related TFs [34]; however , expression of a TF is an imperfect proxy for TF function . For example , many nuclear receptors are not transcriptionally functional in the absence of ligand regardless of expression levels . Here we propose an approach that integrates open chromatin genomic data and gene expression data to identify candidate TFs that are functional in breast cancer-relevant cells . Expression quantitative trait loci ( eQTL ) analysis identifies genetic variants that correlate with gene expression differences in a population . eQTL analysis complements genetic association data by predicting causal genes whose expression differences dictate organismal phenotypes [35 , 36] . The gene that is responsible for a trait may be located relatively far from the GWAS associated SNPs , as the causal SNPs may modulate TF binding and TFs can act distally to regulate gene transcription . In these cases , preferential binding of a TF to one allele causes differential regulation of gene expression to confer the phenotype [37] . Several studies have provided evidence of causal relationships for gene expression mediating the association between GWAS SNPs and traits [16 , 38] . Phenotype-associated SNPs are enriched for eQTLs , suggesting that eQTL analysis can enhance discovery of causal genes associated with complex phenotypes [39 , 40] . The goal of this study is to gain mechanistic insight into how breast cancer susceptibility alleles confer disease risk . First we identify sequence motifs in breast-relevant tissue and cell lines that are enriched within enzymatic hypersensitive sites and support their functional role with evolutionary conservation analysis and gene expression data . Next we identify breast cancer risk alleles that are predicted to modulate TF binding and we perform eQTL analysis to find candidate causal risk genes . Regulatory regions are bound by a milieu of protein complexes and regulatory nucleic acids , such as transcription factors , histone modifiers , nucleosome remodelers , and lncRNA . Sequence-specific TFs are directly or indirectly responsible for the recruitment of downstream transcriptional modifiers in a sequence-dependent manner . Therefore , we sought to identify all TFs that bind within open chromatin in breast cancer-relevant cells and tissues . We performed ATAC-seq in a mammary epithelial cell line ( MCF10A ) to complement publicly available DNase-seq data from Encyclopedia of DNA Elements ( ENCODE ) [41] and the Roadmap Epigenomics project [42] . We quantified open chromatin and identified regulatory elements genome-wide in five breast cancer-relevant cell lines and breast tissue: MCF7 cells , MCF10A cells , T47D cells , cultured human mammary epithelial cells ( HMEC ) , and primary breast variant HMECs ( vHMEC ) . Many regions of chromatin accessibility are shared between the cell lines , although the degree of accessibility for a region can vary between cell types ( Fig 1 ) . Enzymatic accessibility coverage ( peaks ) [28 , 29] or depletions of signal in hypersensitive region ( footprints ) [30 , 31] coupled to motif analysis are routinely used to identify TFs that maintain open chromatin structure . We performed iterative rounds of de novo motif analysis using the sequence underlying enzymatic hypersensitivity peaks to identify overrepresented TF recognition sites in each data set ( S1 File ) . All but one of the motifs we identified were previously characterized and described in databases [43–47] . These motifs found in hypersensitive peaks are , on average , evolutionarily conserved ( Fig 2 ) . Additionally , we identified a potential TF recognition sequence that has no known cognate TF binding partner; we refer to this sequence element as an orphan motif ( Fig 2B ) . This orphan motif is evolutionarily conserved , as measured by phastCons [48] and phyloP [49] scores , in hypersensitivity peaks . Hypersensitivity footprints result from protection of the DNA by a bound TF [50]; however , approximately half of all TF-bound motifs do not exhibit composite footprints [31 , 51] because the TF dissociates during the nuclei isolation procedure . We corrected the DNase data for intrinsic sequence bias [33] , but we do not observe a composite footprint for this orphan motif ( Fig 3B ) . However , the hypersensitivity pattern surrounding the motif is not uniform—the region downstream of this orphan motif is more hypersensitive than upstream . This directional pattern of enzyme accessibility is common with many TFs [29] , including CTCF ( Fig 3A ) . We hypothesize that this orphan motif is functional and directs chromatin accessibility by serving as a recognition site for an uncharacterized TF . In addition to this orphan motif , we identified hundreds of position-specific weight matrices ( PSWM ) from our exhaustive analyses , many were redundant between cell lines and found in multiple rounds of motif analysis in the same cells . To reduce the complexity of these data , we mapped the comprehensive set of motifs found in MCF7 , MCF10A , T47D , HMEC and vHMEC into distinct non-redundant motif families . This operation resulted in identification of 37 sequence motif classes across the five breast cancer-relevant cell lines and tissue ( S1 Fig ) . Many TFs share paralogous DNA binding domains and these TFs often recognize the same sequence motifs . We defined the full set of TFs that recognize each motif by using known TF/sequence interaction data from ChIP-seq [43 , 45 , 52] , protein binding microarrays [53 , 54] , and SELEX [46 , 47] data . We identified 23 TF families in at least two cell lines/breast tissue ( Table 1 ) . Fourteen TF families were uniquely found in one cell line or tissue ( Table 1 ) . These 37 TF families represent the binding motifs for 235 distinct TFs ( Table 1 ) . However , these TFs are not all expressed in breast tissue . We identified the TFs that are most likely candidates for maintaining open chromatin and the gene regulatory expression profiles in breast-relevant cells by examining the relative expression of all of the TFs in each family using TCGA expression data ( Fig 4 and S2 Fig ) . For example , ESR1 , ESR2 , and PPAR-γ contain paralogous DNA binding domains and they recognize indistinguishable sequence elements ( S1 Fig ) . We find that ESR1 is the most highly expressed TF that recognizes the motif ( Fig 4 ) . This result is consistent with the biological role of ESR1 in the etiology of breast cancer and breast biology compared to ESR2 and PPAR-γ . Similarly , FOXA1 is the most well-charactered TF within the Forkhead Box family of TFs in terms of estrogen signaling and interplay with ER [55] . As expected , we find that FOXA1 is the most highly expressed TF in the family of 22 Forkhead Box TFs ( Fig 4 ) . We find that many of these highly expressed TFs correlate with breast cancer survival time in a subtype-specific manner ( S3 Fig ) . Silencing of IRF7 pathways in breast cancer cells promotes breast cancer metastasis , and high expression of the IRF7-regulated genes with breast cancer is associated with prolonged survival [56] . Similarly , we find that high expression of IRF7 is correlated ( P = 0 . 029 ) with positive breast cancer patient outcome in Luminal A subtype ( S3 Fig ) . We find that high expression of BATF ( P = 0 . 0035 ) and TP73 ( P = 0 . 0077 ) is correlated with breast cancer patient survival in HER2+ and Basal-like subtypes , respectively ( S3 Fig ) . Taken together , these data support the notion that TF expression levels may serve as biomarkers of patient outcome . A major goal of this study was to identify a set of plausible causal SNPs that modulate TF binding from a list of SNPs associated with breast cancer in GWAS-defined loci . We identified a total of 463 SNPs in strong LD ( r2 ≥ 0 . 8 ) with the most associated breast cancer GWAS SNPs defined in 93 distinct genomic loci; these SNPs are predicted to affect the binding of TFs belonging to at least 30 TF families . Six examples of candidate causal breast cancer-associated SNPs are shown in Table 2 . Transcription factors from the following TF families are predicted to have their binding affected: CTCF , GABPA , RUNX , GRHL2 , USF1 , ZBTB33 , and ZNF143 . For example , SNP rs11540855 ( 3′ UTR of ABHD8 on chromosome 19p13 . 11 ) is within a DNase-defined regulatory element in human mammary epithelial cells and should affect the binding of CTCF ( Fig 5 ) . rs11540855 is in strong LD with rs8170 ( r2 = 0 . 98 ) , which is associated with breast cancer risk [58–60] . Likewise , rs3760982 ( 1 . 1kb 5′ of KCNN4 on chromosome 19q13 . 31 ) is associated with breast cancer susceptibility [3] and we find that its A allele is predicted to enhance RUNX binding ( Fig 5 ) . More examples of candidate causal breast cancer-associated SNPs disrupting TF binding sites within breast cancer GWAS loci are shown in the supporting information ( S4 Fig ) . Two SNPs ( rs4414128 and rs8103622 ) are predicted to strongly affect CTCF binding ( Fig 6 and S5 Fig ) ; therefore , we tested our predictions of how the alleles would affect CTCF binding by analyzing ENCODE ChIP-seq data . The SNPs rs4414128 and rs8103622 showed allele-specific binding that was consistent with the direction predicted based on tolerated degeneracy from the consensus binding site ( Fig 6 and S5 Fig ) . For rs4414128 ( Fig 6 ) , 34 of 44 cell types/replicates show the expected C preference with a range between 52% and 75% . Eight cell lines/replicates show modest allelic imbalance favoring the T allele ( 31–49% of the reads spanning the SNP ) ; two experiments are balanced . Importantly , both replicates of human mammary epithelial cells ( HMEC ) show an allelic imbalance favoring C in 58% and 75% of the reads . Across 37 cell types ( and replicates ) that are heterozygous and normal karyotype , rs8103622 ( S5 Fig ) shows allele-specific preference of C as expected in 34 cell types/replicates with the range between 53% and 88% . The other three instances exhibit allelic balance , with allele frequencies between 48% and 52% . We identified rs11540855 as the most significant ( P-value: 2 . 75E-10 ) eQTL SNP in the GWAS locus that is in strong LD ( r2 = 0 . 90 ) with the most associated breast cancer GWAS SNP rs8170 [58] ( Fig 7C ) . The G/G genotype at rs11540855 is predicted to compromise CTCF binding ( Fig 5 ) and G/G individuals have , on average , higher expression levels of ANKLE1 ( Fig 7A and 7B ) . The most significantly associated GWAS SNP in this locus ( rs8170-T ) increases the risk of ER-negative breast cancer with an odds ratio of 1 . 10 [60] and the T allele is in LD with the G allele of rs11540855 . Therefore , higher expression of ANKLE1 ( Fig 7A and 7B ) is associated with increased breast cancer risk . We also identified rs3760982 as one of the most associated eQTL SNPs ( P-value: 3 . 16E-07 ) and rs3760982 is correlated with ZNF404 expression ( Fig 8C ) . The A/A genotype at rs3760982 is predicted to increase RUNX binding ( Fig 5 ) and is correlated with higher expression of ZNF404 in breast cancer tumor samples ( Fig 8A ) and breast tissue ( Fig 8B ) . The rs3760982-A allele is associated with an increased risk ( odds ratio of 1 . 06 ) of breast cancer [3] , thus higher expression of ZNF404 correlates with increased breast cancer risk . Therefore , we prioritized SNPs within GWAS loci that are predicted to affect transcription factor binding and module expression of ANKLE1 and ZNF404 to confer breast cancer risk . GWAS have discovered more than 90 genetic loci and common genetic variants associated with breast cancer susceptibility [1–5] , and the majority of SNPs in these loci are enriched in non-coding regions . Non-coding genetic variants can contribute to complex traits and diseases through many molecular mechanisms [12–14 , 61 , 62] including having an effect on TF binding affinities , which can result in differential gene expression . Herein , we describe an integrative genomics methodology to identify a near-comprehensive set of TFs that are actively maintaining open chromatin in a cell type . We identify which GWAS-relevant SNPs are predicted to modulate TF binding intensity and use ChIP-seq data to confirm our predictions . Lastly , we use eQTL data to identify the likely target genes of SNPs that affect TF binding affinity . Taken together , this approach can identify likely causal SNPs associated with breast cancer risk . We found that rs3760982 variants are predicted to modulate RUNX binding ( Fig 5 ) and we confirm previous work showing that rs3760982 is an eQTL for ZNF404 ( Fig 8 ) [3 , 63] . The A allele of rs3760982 conforms more stringently to the RUNX consensus sequence and this allele is predicted to enhance RUNX binding; allele-specific ChIP-seq in breast tissue , would test whether RUNX family TFs prefer binding the A allele in vivo . The RUNX family of TFs are canonical transcriptional activators [64–67] , so we hypothesize that increased RUNX binding is a mechanism by which the ZNF404 is regulated . Testing this hypothesis would necessitate specific gene editing of rs3760982 and subsequent measurements of ZNF404 expression . While CRISPR-mediated [68] deletions of genetic elements is routine , precise changes of specific alleles remains a challenge . Deletion of the rs3760982 variant by CRISPR , followed by measuring ZNF404 expression would confirm or refute the role of rs3760982 variants in ZNF404 expression . One could also test allele-specific expression of alleles within transcription units that are phased with rs3760982 variants . We propose using precision global run-on sequencing [69] to measure nascent RNA expression to capture informative intronic SNPs . While genomic approaches are a means to develop novel hypotheses , the advent of genetic editing approaches permits hypothesis testing to define mechanisms by which genes and genetic variants contribute to disease risk . We found that rs11540855 is an eQTL for ANKLE1 ( Fig 7 ) and rs11540855 variants are predicted to affect CTCF binding ( Fig 5 ) . The rs11540855 SNP is in high LD ( r2 = 0 . 90 ) with the breast cancer-associated GWAS SNP rs8170 , which was first found as a modifier of breast cancer risk in BRCA1 mutation carriers [58] . Subsequently , this SNP was found to be associated with breast cancer susceptibility in ER-negative breast cancer [58–60] . ANKLE1 is an evolutionarily conserved non-membrane-bound LEM protein that harbors endonuclease activity , but its cellular functions remain uncharacterized [70 , 71] . Future work will need to determine the allele-specificity of CTCF binding at rs11540855 and test the role that this site has upon ANKLE1 expression . These approaches will be able to define the relationship between TF binding and gene expression , but it is challenging to develop a physiologically relevant model of breast cancer risk that incorporates human genetic variation . GWAS-identified SNPs are common and typically confer relatively small differences in risk . Further , the cumulative affects of differential gene expression over the lifetime of an individual cannot be easily recapitulated in a controlled environment . Our research identified a previously uncharacterized DNA sequence motif that is enriched in open chromatin , evolutionarily conserved , and is associated with directional hypersensitivity profiles ( Figs 2 and 3 ) . We hypothesize that this orphan motif is the recognition site for a previously uncharacterized transcription factor . Future work , such as DNA affinity chromatography [72] , will be needed to identify this candidate TF . Genomic approaches are ideally suited to address fundamental biological questions in a relatively unbiased manner . Integrative genomic approaches and analyses can clarify the null-hypothesis and permit the development of novel hypotheses that were previously inconceivable . These approaches serve as a first-step in understanding the biology of breast cancer risk and targeted experimental follow-up is necessary to define the mechanistic roles of genes and genetic variants in breast cancer susceptibility and disease progression . We cultured MCF10A cells in Dulbecco’s modified Eagle’s medium ( Invitrogen ) with 5% horse serum ( Invitrogen ) , 1% penicillin/streptomycin ( Invitrogen ) , 20 ng/ml EGF ( Peprotech ) , 0 . 5 μg/ml hydrocortisone ( Sigma ) , 100 ng/ml cholera toxin ( Sigma ) and 10 μg/ml insulin ( Sigma ) in a humidified incubator at 37°C with 5% CO2 . The ATAC-seq library was prepared as previously described [73] with several modifications: 1 ) IGEPAL CA-630 was omitted from the lysis buffer; 2 ) we performed two additional wash steps with lysis buffer; and 3 ) we performed PCR-clean up using AMPure XP beads to select DNA <600 bp . The MCF10A ATAC-seq data were deposited in the Gene Expression Omnibus ( GEO ) database , with accession number GSE89013 . We mapped reads to the hg38 human reference genome using Bowtie2 [74] and merged replicate aligned files . We used the merged data for all subsequent analysis; refer to S1 File for ATAC-seq data analysis details . We performed iterative rounds of de novo motif analysis using a 120-base pair window centered on the summit of hypersensitivity as defined by ATAC-seq or DNase-seq ( S1 File ) . In each cell type we found hundreds of over-represented position specific weight matrices ( PSWMs ) . We identified all instances of each PSWM within breast-specific regulatory elements , while accounting for the possibility that the reference genome contains variants that will conform more or less strictly to the PSWM . This approach allowed us to identify potential binding sites that contain SNPs , even if the reference allele does not match the queried PSWM . Although we identified hundreds of distinct PSWMs , many PSWMs are similar to one another and are likely to represent redundant specificity of a single TF or TF family . To consolidate similar PSWMs into known TF families , we systematically classified several public PSWM repositories [43–47] into families with distinct features . PSWMs were first divided into clusters based on connectivity; connectivity between motif nodes was measured by negative log10 E-value as calculated by TOMTOM [75] . An edge was inferred between two motif nodes if their similarity exceeded a negative log10 E-value of 10 . We defined a motif cluster as a connected set of nodes; connectivity is defined by the existence of a path between every pair of nodes . A fast greedy modularity algorithm [76] further divided each motif cluster into families . We downloaded a curated list of breast cancer associated SNPs from the GWAS catalog [77] . The SNP that exhibits the most statistically significant association with breast cancer in any locus may not be causal due to linkage disequilibrium ( LD ) and the sampling variation that interrogated the specific SNP . To better define the list of likely causal variants for each locus , we identified all SNPs satisfying the following three criteria: 1 ) SNPs that are in strong LD ( r2 ≥ 0 . 8 ) with the most significant reported GWAS SNP; 2 ) SNPs that are located within putative TF binding sites identified by hypersensitivity assays; and 3 ) SNPs that are within critically important positions that affect TF binding affinity ( Information Content ≥ 0 . 5 ) . We analyzed ENCODE CTCF ChIP-seq data for allele-specific preference of SNPs that are predicted to modulate CTCF binding affinity . All CTCF ChIP-seq data are provided within S1 File . We analyzed the highest intensity CTCF sites to ensure that sequencing reads would span the query SNP . We exclusively queried normal karyotype cell lines that had SNPs that were heterozygous within each locus to reduce the chances that copy number variations ( i . e . , aneuploidy ) of alleles would skew our analyses . To identify putative causal genes whose expression may be affected by polymorphisms , we performed eQTL analysis using TCGA breast cancer data [57] with fastQTL [78] . We imputed the genotypes from TCGA SNP6 arrays that were hybridized with DNA extracted from the blood of patients with breast cancer . We retrieved genotype data from dbGaP ( phs000178 . v9 . p8 ) and imputed genotypes using the Michigan Imputation Server [79] with the following parameters: 1000G Phase 1 v3 Shapeit2 Reference Panel , ShapeIT Phasing , Mixed Population , and Quality Control/Imputation Mode . Following imputation , we removed SNPs with the following features: Hardy-Weinberg Equilibrium p < 1 × 10−6 and minor allele frequency ( MAF ) < 5% . We used UCSC-curated TCGA RNA-seq data [57] from breast cancer patient solid tumor samples as the gene expression data to identify eQTLs . TCGA clinical data were incorporated as the covariates for eQTL analysis such as sample RNA concentration , RIN value , sex , and ethnicity . We performed Principal Component Analysis ( PCA ) on the quantitative variables from clinical data and used the first three principal components as covariates . We retained other qualitative variables as categorical covariates .
The promise of effective personalized medicine is dependent upon the ability to identify genetic variants in the population that influence disease risk and then use this information to accurately predict the likelihood of disease incidence for individual patients . High-risk individuals may be entered into clinical trails , pre-clinical intervention strategies , or increased frequency of screening to detect early disease onset . However , the contribution of any one genetic variant to increase disease susceptibility is typically small , with many potential causal variants in the genomic region associated with risk . Therefore , it is important to understand the biological mechanisms by which the variants within a genetic region influence disease susceptibility by refining the set of all variants to those that are highly plausible to be causal . Herein , we describe a method to integrate molecular genomics data with genetic epidemiological data to inform on the underlying molecular mechanisms that influence breast cancer risk . This approach identifies the important transcription factors that directly regulate gene expression to modulate disease susceptibility .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "methods" ]
[ "cell", "binding", "genome-wide", "association", "studies", "cell", "physiology", "medicine", "and", "health", "sciences", "breast", "tumors", "chemical", "characterization", "cancers", "and", "neoplasms", "oncology", "genome", "analysis", "sequence", "motif", "analysis", "epigenetics", "chromatin", "research", "and", "analysis", "methods", "sequence", "analysis", "bioinformatics", "chromosome", "biology", "gene", "expression", "breast", "cancer", "binding", "analysis", "genetic", "loci", "cell", "biology", "database", "and", "informatics", "methods", "genetics", "biology", "and", "life", "sciences", "genomics", "computational", "biology", "human", "genetics" ]
2017
Identification of breast cancer associated variants that modulate transcription factor binding
Curation and interpretation of copy number variants identified by genome-wide testing is challenged by the large number of events harbored in each personal genome . Conventional determination of phenotypic relevance relies on patterns of higher frequency in affected individuals versus controls; however , an increasing amount of ascertained variation is rare or private to clans . Consequently , frequency data have less utility to resolve pathogenic from benign . One solution is disease-specific algorithms that leverage gene knowledge together with variant frequency to aid prioritization . We used large-scale resources including Gene Ontology , protein-protein interactions and other annotation systems together with a broad set of 83 genes with known associations to epilepsy to construct a pathogenicity score for the phenotype . We evaluated the score for all annotated human genes and applied Bayesian methods to combine the derived pathogenicity score with frequency information from our diagnostic laboratory . Analysis determined Bayes factors and posterior distributions for each gene . We applied our method to subjects with abnormal chromosomal microarray results and confirmed epilepsy diagnoses gathered by electronic medical record review . Genes deleted in our subjects with epilepsy had significantly higher pathogenicity scores and Bayes factors compared to subjects referred for non-neurologic indications . We also applied our scores to identify a recently validated epilepsy gene in a complex genomic region and to reveal candidate genes for epilepsy . We propose a potential use in clinical decision support for our results in the context of genome-wide screening . Our approach demonstrates the utility of integrative data in medical genomics . Interpretation of high-resolution array comparative genomic hybridization ( aCGH ) data is made challenging by the large number of copy number variation ( CNV ) events identified in each individual . Analogous problems arise in interpretation of deep sequencing data where the number of variants rapidly outstrips the capacity for manual curation . Moreover , because of the recent expansion of human populations , most variation in an individual genome is rare and restricted among family lineages , making distinction between rare and pathogenic variation difficult [1] . Given the scale of variation and the challenge of profile interpretation , a number of groups have developed and utilized computational and machine learning tools to prioritize genetic data [2] , [3] . Huang and colleagues analyzed the characteristics of a group of genes and their protein products known to cause phenotypes in the haploinsufficient state and compared them to those that were repeatedly deleted in a control population of apparently healthy individuals ( i . e . those haplosufficient ) [4] . The differences between these groups of genes were used to develop a general quantitative model to predict whether a gene deletion is likely to be deleterious . While there is broad applicability to such a prioritization scheme , it provides little guidance to help a clinician determine whether a given deletion has a role in a specific phenotype in an individual patient . Other studies assigned genes to networks based upon particular disease phenotypes , and while useful for directing further studies , these approaches did not attempt to quantify the likelihood of a gene's appropriate assignment to a given disease trait or its propensity for actually contributing to disease [5]–[7] . Other researchers have developed a computational model that takes into account genomic structure and functional elements to predict whether a given CNV is associated with intellectual disability ( ID ) [8] . This algorithm represents an excellent tool , however , it does not specifically predict or rank which gene ( s ) within the CNV are most dosage-sensitive or likely to be relevant to the phenotype . Such specific predictions are necessary to inform clinical interpretation and to aid the development of disease-centered diagnostics . Moreover , none of these tools use comparative variant frequency information between affected phenotypes in a large database to inform the scoring and prioritization schemes . Epilepsy is a common neurological disorder for which improved computational tools could be extremely beneficial . With over 50 million individuals affected , the prevalence of epilepsy ranges from 0 . 2 to 2% depending on the population studied [9] . In the United States , the overall prevalence is approximately 0 . 5% , with a disproportionate number of cases in infants , children , and the elderly [10] . The epilepsies are currently grouped into genetic , structural/metabolic , or unknown etiologies [11] . To date , only a fraction of patients with suspected genetic forms of epilepsy have an etiological diagnosis , meaning accurate recurrence risk , prognosis , and disease-specific surveillance and treatment information are rarely available . The lack of specific diagnoses is at least partly due to the complex inheritance , variable expressivity , and incomplete penetrance of many forms of epilepsy; although some examples of Mendelian segregation are recognized [12] . The role of CNVs in common neurological diseases has become increasingly clear , and there are well-studied CNVs that cause isolated or syndromic disorders including ID [13] , autism spectrum disorders ( ASDs ) [14] , [15] , and schizophrenia [16]–[18] . Although each CNV itself is rare among individuals with a given disease , when considered as a group , structural variation of the genome is a common cause of such phenotypes . A number of well-described syndromic disorders with epilepsy are caused by CNVs , including chromosome 1p36 deletion syndrome , Angelman syndrome , and MECP2 duplication syndrome . A number of studies testing large cohorts of individuals have demonstrated that various CNVs are associated with a wide range of epilepsy phenotypes including non-syndromic idiopathic epilepsy [19]–[22] . We hypothesized that information about individual genes gleaned from large-scale knowledge sources could be integrated into an epilepsy-specific pathogenicity score . We further hypothesized that these scores could be combined with frequency information of gene disruption among individuals with epilepsy to prioritize candidate genes and interpret variants identified in personal genomes . We used a fixed set of training genes previously published as variant in Mendelian epilepsies to determine training patterns for epilepsy genes in these high dimensional data types and subsequently developed a score matching the training set for each available gene in each knowledge source . To utilize variant frequency information together with our pathogenicity scores , we took advantage of a Bayesian approach in which the gene pathogenicity scores were used to develop informative prior probabilities for the expected increase in the frequency of variants in an epilepsy population as compared to non-neurologic controls . This statistical analysis determined Bayes factors as further scores we used to rank and prioritize genes . We then applied these gene-level scores to characterize CNVs harbored by individuals in a well-defined cohort of subjects with epilepsy identified by electronic medical record ( EMR ) review , and used this analysis of CNVs to assess our pathogenicity score . We also evaluated the Bayes factors comparing the results of our epilepsy cohort to a matched cohort with non-neurologic indications , and used this method to explore a possible role of multiple genes disrupted within the genome of a single individual with epilepsy . Finally , we examined the possible utility of our scheme as a clinical decision support tool for patients undergoing genome-wide testing . Genes involved in the same disease are often similarly annotated in knowledge databases , are expressed in similar tissues or have gene products that physically interact [2] . We hypothesized that genes involved in epilepsy would show such characteristics . Indeed , analysis of a set of 83 genes with known epilepsy associations reveals that epilepsy genes form highly connected networks in multiple datasets ( Figure 1 ) . Thus , we concluded that it would be reasonable to interrogate these knowledge sources to identify as-of-yet unknown genes associated with the phenotype by correlating the features of the training genes with those of all other RefSeq genes . To improve our prioritization , we sought to include information about how often a given gene was mutated among individuals with epilepsy compared with a background population ( Figure 2 ) . These complementary strategies and their integration are described below . We hypothesized that a bioinformatic approach could consolidate information from multiple biological fields into an integrated score of pathogenicity on a genome-wide scale . To this end , we validated six “features” ( see methods ) using biological information from large-scale knowledge sources and comparing known epilepsy genes , including 20 recognized as causative by the International League Against Epilepsy [23] ( Table S1 ) , to all annotated RefSeq genes . We considered Gene Ontology ( GO ) and Mouse Genome Informatics ( MGI ) phenotype annotation , protein-protein interaction ( PPI ) data , human tissue expression patterns , micro RNA ( miRNA ) targeting , and the Kyoto Encyclopedia of Genes and Genomes ( KEGG ) pathway data to develop an omnibus epilepsy pathogenicity score to predict whether loss of function of a given gene might be relevant to the phenotype . To determine the efficacy of our scoring mechanism in an unbiased way , we cross-validated our approach . We removed each of the training genes and recalculated genome-wide pathogenicity scores based on the remaining training genes; we then evaluated the procedure on the gene excluded from training . The composite result is presented in Figure S1 . The cross-validation demonstrates that individual feature scores as well as the composite mean pathogenicity score detect known epilepsy genes more efficiently than random chance . The PPI score is the most efficient feature , with an area under the curve ( AUC ) of 0 . 84 . The AUC of the composite score is 0 . 86 . The individual feature scores and composite mean pathogenicity scores of some well-known genes are presented in Table 1 . Table S2 provides scores for all RefSeq genes . Having concluded that our pathogenicity score is capable of differentiating known epilepsy genes from other genes throughout the genome , we hypothesized that pathogenicity scores among genes deleted in individuals would correlate with an epilepsy phenotype . We theorized that patients with non-neurologic phenotypes would be less likely to harbor CNVs containing genes with high pathogenicity scores , while patients with epilepsy would have CNVs harboring genes with higher scores . To increase the likelihood of identifying an effect , we investigated the CNVs of a well-described “analysis cohort” of subjects with epilepsy identified from unbiased independent review of EMRs ( see methods ) . We compared their CNVs to those of a matching analysis cohort of subjects referred to our diagnostic center for non-neurologic indications . We computed the maximum pathogenicity score among genes varying in copy number for each subject , filtering erroneous calls ( see methods ) . Considering all CNVs – both gains and losses – across the two analysis cohorts , the total pathogenicity burden is not significantly different between subjects with epilepsy and subjects without neurological disease ( data not shown ) . However , considering only genes harbored within genomic deletions reveals that the maximum scoring gene of deletion CNVs is significantly higher ( Wilcox Signed Rank Test , p<5 . 2×10−4 ) in patients with epilepsy ( Figure 3A ) . The maximum scoring genes of genomic duplications are not significantly different between the two patient groups ( Figure 3B ) . To exclude the possibility that the variation observed in patient-wide pathogenicity scores was due to known epilepsy genes , we elected to remove the training epilepsy genes ( Table S2 ) from our calculations . In support of our findings , the maximum score of genes disrupted among patients with epilepsy remains statistically greater than those with non-neurologic indications ( Wilcox Signed Rank Test , p<0 . 011 , Figure 4A ) . Having evidence that our pathogenicity score is correlated with epilepsy phenotype at the patient level , we sought to further improve our gene-based prioritization by including gene deletion frequency among a large cohort . This step is important because the pathogenicity score is uninformed about mutation or variant frequency of genes . To accomplish this integration , we took advantage of a Bayesian approach coupled with CNV data collected among 23 , 578 individuals referred to our diagnostic center for genome wide CNV testing because of a variety of phenotypes . We computationally identified individuals with indications for diagnostic test consistent with epilepsy ( n = 1616 ) and those consistent with disease , but of non-neurologic etiology ( n = 2940 , see methods ) . This resulted in two matched “frequency cohorts” of individuals distinct from our well-phenotyped “analysis cohorts . ” We determined the observed deletion frequency for each gene as described above . We then parameterized a family of gamma prior distributions that modeled the baseline deletion frequency for each gene by setting the mean of the distribution equal to the observed deletion rate in the non-neurologic frequency cohort . Figure 4 demonstrates this processes for KCTD15 , with the baseline distribution shown in grey . We then allowed the prior mean of the gamma distribution of each gene to increase based on its epilepsy-specific pathogenicity score , while the variance was constrained to be a constant multiple of the mean . This approach resulted in a second family of prior rate distributions informed not only by gene knowledge but also background deletion frequency ( Figure 4 , pink distribution ) . Next , we computed the total probability of the observed rate of deletions for each gene among the 1 , 616 subjects in the epilepsy frequency cohort ( Figure 4 , red line ) under the background model ( grey distribution ) and the pathogenicity informed prior ( pink distribution ) . We calculated the ratio of these probabilities , called the Bayes factor , for each gene to allow us to further prioritize genes associated with the epilepsy phenotype . Table 2 lists the 10 RefSeq genes with the highest Bayes factors , presenting only the maximum scoring gene for recurrent deletions with multiple high scoring genes . Table 3 lists 10 additional genes with high Bayes factors but without known associations with epilepsy . Finally , we computed posterior rate distributions for each gene , taking into account gene knowledge from the pathogenicity score , background deletion frequency from the non-neurologic cohort as well as the observed rate in the epilepsy frequency cohort ( Figure 4 , blue distribution ) . Table S2 lists the frequencies , Bayes factors and posterior rate distribution parameters for each RefSeq gene . Having designed Bayes factors to more robustly prioritize candidate epilepsy genes , we revisited our analysis cohort of subjects with epilepsy that we previously analyzed with the pathogenicity score alone . Using a method analogous to our previous analysis performed at the level of subjects , we calculated the maximum Bayes factor among genes deleted in each individual . Because the deletion of a gene in a given subject with epilepsy necessarily influences the Bayes factor of that gene , we used a cross-validation approach , and recalculated the genome wide Bayes factors for each subject leaving out their contribution to the frequency data . We discovered that the cross-validated maximum Bayes factors were significantly higher among subjects with epilepsy than those referred for non-neurologic indications ( Wilcox signed rank test p<1 . 1×10−6 , p<9 . 4×10−6 without training genes Figure 5A ) . Given that many of our features preferentially identify genes that are more highly expressed in the brain ( not the least of which being the gene expression feature , data not shown ) , we were concerned that our Bayes factors might be identifying genes associated with neurologic phenotypes rather than epilepsy in particular . To examine this , we generated a cohort of individuals referred for ASDs and not epilepsy who also had abnormal aCGH studies . In keeping with the Bayes factor score as specific for epilepsy , the maximum Bayes factor is higher among subjects with epilepsy than those referred for ASDs ( Wilcox signed rank test p<6 . 5×10−5 , p<2 . 1×10−9 without training genes Figure 5A ) . We were also interested to explore whether the Bayes factors of more than one gene deleted in each individual subject might be correlated with phenotype , thus suggesting a digenic or oligogenic effect . To this end , we performed the same cross-validation calculation , but excluded the contribution of the genes with the single highest Bayes factor from each patient . Notably , the genes with the second highest Bayes factors are significantly higher among subjects with epilepsy compared with individuals referred for non-neurologic indications ( Wilcox signed rank test , p<2 . 1×10−5 , p<5 . 1×10−5 without training genes , Figure 5B ) . Such a comparison considering the third highest scoring gene and excluding the two highest scoring genes also results in a significant difference ( Wilcox signed rank test , p<5 . 4×10−6 , p<1 . 8×10−6 without training genes , Figure 5B ) . As additional assessment of the utility of our scoring method , we calculated the scores of genes published as potentially related to epilepsy by Lemke et al [24] . Because the training genes are by definition higher scoring , we elected to exclude them from this analysis . Genes in the published list but not included in the epilepsy training genes ( n = 263 ) have significantly higher pathogenicity scores than the genome wide average ( Wilcox signed rank test , p<0 . 014 ) . The same genes also have significantly higher Bayes factors ( Wilcox signed rank test , p<7 . 6×10−12 ) . Another potential use of our Bayes factor scoring metric is in the identification of candidate epilepsy genes at regions with known associations but no known causative gene [25] . As a proof of principle , we analyzed the 34 RefSeq genes harbored in the recurrent , low-copy repeat mediated 16p11 . 2 deletion . No training gene was identified from this region because no definitive association has been made between a gene and epilepsy , although approximately 24% of patients with 16p11 . 2 deletions experience seizures [26] . If we use frequency data among subjects with epilepsy and those referred for non-neurologic indications alone , little information can be gained because of the recurrent nature of the deletion ( Figure 6 ) . In fact , because of recent advances in oligonucleotide aCGH probe design , over time additional probes have been placed in regions closer to the flanking LCRs that mediate the CNV formation . Because of this technical artifact , more subjects with epilepsy were calculated to have deletions of SLC7A5P1 than those subjects with non-neurologic indications . However , if we instead use Bayes factors , taking into account both frequency and gene knowledge , the highest scoring gene is identified as KCTD13 ( Figure 6 ) . Dosage of this gene has recently been shown to correlate reciprocally with the phenotype of head size in a Zebrafish model , a hallmark of the 16p11 . 2 deletion and duplication syndromes [27] . The same authors report a patient with a complex rearrangement of KCTD13 with many of the features of 16p11 . 2 deletion syndrome . Nonetheless , given the high scores of both DOC2A and TAOK2 , it is not unreasonable to hypothesize that one or more other genes in the region might also contribute to the epilepsy seen in these subjects . Figure S2 presents similar analyses of other loci with recurrent deletions associated with epilepsy . Given that our data are derived from diagnostic testing , we were interested to explore our composite Bayes factor result as a possible clinical decision support tool to aid in discrimination of an epilepsy phenotype . Figure 7 shows the sensitivity and specificity of the maximum Bayes factor among deleted genes in an individual subject when used as a binary decision rule to discriminate between epilepsy and non-neurologic phenotypes across a range of Bayes factor cut-off values . These parameters are relevant for subjects with abnormal aCGH tests and no neurologic indications other than epilepsy . As an example , having a deleted gene with a Bayes factor of greater than 1 discriminates with a sensitivity and specificity of 0 . 62 and 0 . 60 , respectively . These statistics are highly influenced by recurrent deletions at the Velocardiofacial locus with a maximum Bayes factor of 2 . 87 . Choosing a cut off of 2 . 88 ( thus excluding the effects of the Velocardiofacial region ) results in a sensitivity and specificity of 0 . 37 and 0 . 84 , respectively . Given the imperfection of our indication and coding data , such a cut off rule would suggest 16% of patients with non-neurological indications should receive increased suspicion based on their CNV data . We also attempted to construct a decision rule based upon the contribution of multiple genes deleted in a given patient , but concluded the single highest scoring gene produced the best results ( data not shown ) . Rapid expansion of the human population [28] together with relaxed negative selective pressures secondary to increased food supplies and improved medical care [1] as well as the possible influence of higher mutation rates [29] have skewed much of the allele frequency spectrum of human genomic variation toward rare or private variants . Purifying selection is expected to eliminate highly deleterious alleles from a population over time [30] , yet it is precisely the new and rare variations that contribute to human disease . We should expect that novel rare and private variants will continuously be discovered , and there are a nearly infinite number of possible variants and combinations of variants that can occur . Thus , a fundamental shift in the approach to variant interpretation must occur from simple cataloging of variants at a locus to prediction of the possible effects of highly rare or newly identified variants by integrating the state of knowledge about genes and disease processes . We contend that effective diagnostics must ultimately incorporate some aspects of discovery , inferring the relevance of new and arcane genomic variants for patient phenotypes by leveraging known information and multiple sources of evidence . In essence , our approach seeks to automate aspects of expert interpretation processes that are currently undertaken by clinical molecular geneticists and diagnostic laboratories on a daily basis . These professionals consider what is known about mutated genes—for example whether they are expressed in effected tissues or if their protein products are involved in applicable pathways . They then consider the frequency of mutations both among normal individuals and patients with similar and related phenotypes . Together with other information and years of experience , the geneticist combines these data into an assessment of variant relevance . Although our computational method cannot be as effective as an experienced human at interpretation of an individual variant in a single patient , it does have the advantage of scalability to many variants and to large cohorts of individuals with different phenotypes . Moreover , this approach and others like it can help to facilitate the interpretation of an expert by providing additional triage of large-scale variant data . Our method comprises two integrated steps: phenotype specific pathogenicity scoring and Bayesian analysis using frequency data . The pathogenicity scoring approach provides a quantitative method to evaluate genes with respect to a fixed phenotype using known phenotype specific disease genes as a target , leveraging many sources of knowledge . However , since the model depends highly on the “epilepsy genes” ( Table S1 ) , the choice of the training genes themselves inherently introduces the bias of past knowledge . Moreover , the training genes were not otherwise sub-structured to consider their distinct functions or roles in epilepsies with diverse etiology; this simplification was mirrored in the EMR review , where we made binary decisions about the appropriateness of the epilepsy assignment without consideration of natural history data that might otherwise inform or refine the interpretation of genetic data . Our simplified initial approach might be improved by future methods better informed by sub-classifying the training genes and refined consideration of the phenotype data . Another facet of past knowledge bias is that the computation relies on available gene data from the literature and public databases . Thus , the pathogenicity score is only as effective as the a priori knowledge for each individual gene . If little or no information is known about a gene , or worse yet if a gene is not annotated in the RefSeq , the algorithm cannot accurately calculate a score . To overcome this limitation we attempted to incorporate less biased information such as gene expression data and other types of genome wide scores , such as miRNA target prediction . In cases where genes were missing features–such as lack of MGI phenotype data—we attempted to impute missing values using linear regression and other methods . Ultimately , we concluded that restricting analysis to the available reported data provided better results than statistical imputation ( data not shown ) . More work in this area is warranted . Likewise , the knowledge sources we utilized are themselves imperfect . In ontological systems , the failure to annotate a gene to a category can represent an unobserved value in the annotation system , such as a phenotypic assay that was not performed , and not evidence of a negative annotation . This property is often summed up as: “the absence of evidence is not evidence of absence . ” While this issue is an important limitation that requires further study , we believe data will improve over time , making ontological systems progressively more informative as annotations become more comprehensive genome-wide . A key advantage of our method is incorporation of observed variant frequency data from over 20 , 000 genome-wide assays performed by high-resolution aCGH at out diagnostic lab in addition to the computational gene scoring approach . The epilepsy cohorts and comparator non-neurologic cohorts were comprised of phenotypically affected individuals with segmental findings . Our approach was to model the differential frequency of CNVs affecting each gene between these two groups using our pathogenicity score to inform the rate distribution . Subsequently , we are able to use the machinery of Bayesian model comparison to compute those genes where the epilepsy scoring improved the fit from what would be expected without this phenotype-based knowledge . The Bayes factor summaries allowed us to rank individual genes using the computational score , but the real frequency data—which are driven by molecular mutation events in actual human populations—necessarily incorporate structural and genomic feature information that are not part of the pathogenicity score . By exploiting variant frequency in actual subjects , our approach utilizes this genomic information without requiring us to explicitly model or otherwise include the complex biological processes underlying mutation . Using this approach at the genome and cohort-wide level , our analysis was able to highlight a number of potentially novel genes as relevant to epilepsy . The highest scoring candidate gene identified by our method is ERBB4 , encoding a member of the ErbB subfamily of tyrosine kinases that functions as a neuregulin receptor [31] . Rare variants of ERBB4 have been associated with increased risk for schizophrenia [32]; an intronic deletion between exons 7 and 8 was also identified in a patient with an ASD [33]; and a patient with a de novo reciprocal translocation t ( 2;6 ) ( q34;p25 . 3 ) , apparently disrupting the ERBB4 gene , was identified with early myoclonic encephalopathy [34] . Recent experiments also showed that Erbb4−/− mice exhibit increased susceptibility to chemically induced seizures [35] . This evidence taken together with our analysis suggests that mutations of ERBB4 may be associated with a number of epileptic phenotypes . Given the sizable mutation frequency difference of ERBB4 between the epilepsy and non-neurologic cohorts , identification of the gene would have likely been possible using frequency information alone . However , our method is also able to call attention to genes that , although rarely mutated , are highly similar to the training genes . A number of the genes listed in Table 3 exemplify this principle . For example , although GRM5 , encoding the metabatropic glutamate receptor 5 , was identified as deleted in only one subject , it's Bayes factor is in the 98th percentile among genes deleted at least once in any cohort . This gene is interesting because Grm5−/− mice have increased susceptibility to pharmacologically induced seizures and the human protein product is highly connected to the epilepsy training genes . Such information could easily be overlooked given the subject's ( 3 . 6 Mb ) deletion also includes 17 other RefSeq genes . Prioritization of rarely mutated genes that are novel to the epilepsy cohort is an important aspect of our approach . Notably the exclusion of an individual's mutations from the frequency data for the Bayes factor calculation prevents contribution of such genes to our analysis cohort assessment , lowering our statistical power . Given the observation of rare variation in human disease , it is likely that some of these variants contribute to patients' seizures . Additional studies will be required to validate the associations of these genes with epilepsy . In addition to the prioritization of individual genes , our method also naturally lends itself to calculation of multi-variant genetic load , and we were able to evaluate evidence for digenic and oligogenic effects in our analysis cohort . We saw that not only was the maximum Bayes factor among deleted genes significantly higher in the epilepsy cohort versus non-neurologic comparators , but we also observed significant differences in the second and third highest scoring genes . Previous analysis of a large set of CNV data has suggested that copy number changes of multiple genes at distant loci but in similar networks may compound at the molecular level to contribute to phenotypic variation seen with well-known recurrent genomic disorders [36] . This previous study relied on relatively more common recurrent CNVs together with second site mutations . In contrast , our method collapses different deletion alleles at the gene level and then more globally for the phenotype by scoring variants . This allows us to identify differences at the cohort level generally rather than considering individual pairs of variants for which there is very little statistical power given their rarity . Our data suggest that , at least in some patients , the deleterious effects of mutations in two or more genes involved in similar processes may interact on the molecular , cellular , or organism level to results in seizures . In other patients , genomic structural abnormalities may have little influence . Previous studies of sequence variant load in epilepsy failed to identify differences between subjects with epilepsy and control individuals [37]; however , this analysis focused entirely on channel genes whereas our method was intentionally designed to be broad and to include many gene families known to be involved in epilepsy . Because of the strong correlation of Bayes factors with epilepsy phenotype , we investigated sensitivity and specificity of our score as a clinical decision support tool as a natural extension of our integrated analysis . Our approach was to discriminate individuals with epilepsy from those with non-neurologic indications based on their maximum Bayes factors . The difficulty experienced by physicians making use of genome wide tests represents a major limitation in clinical practice [38] . However , if genome wide results can be condensed into a quantitative score and studied epidemiologically , as are other quantitative test results ( i . e . serum Troponin-I ) , the results may be made more accessible for physicians to interpret and guide management . For example , if a patient had a high epilepsy-specific Bayes factor but was not known to experience seizures , it might be reasonable to modify patient care . Such information could alert the clinician to provide counseling about seizure signs , symptoms , and first aid . In some instances , the prescription of an emergency abortive medication might be indicated . Neurologists are frequently asked to discern whether a patient's spells are epileptic or non-epileptic in nature; when the pre-test probability for seizures is high due to known genetic risk factors , then clinical decisions such as ordering an electroencephalogram ( EEG ) or longer-term video EEG monitoring or making the decision to start a medication to treat epilepsy could be impacted . While the Bayes factor score certainly does not capture the subtle differences in phenotypes caused by different alleles at various loci , this approach and extensions of it might in the future be helpful for front-line providers to identify which colleagues can provide useful insight into a particular patient's treatment or augment the clinical decision making process . Because of its modular nature , our scoring mechanism can be recalculated at any time as more confirmed epilepsy genes are discovered , other loci throughout the genome are annotated and more fully characterized , and as additional variant frequency data are constantly recruited into our clinical genomic database . Although our model was designed to prioritize genes varying in dosage among samples tested by aCGH , the data and computational framework are not specifically tied to haploinsufficiency . Therefore , the pathogenicity scores and Bayes factors presented in Table S2 may be applicable to a wide variety of data sets . With the decreasing cost of genome-wide sequencing strategies , analysis of patient genomes in epilepsy will likely result in an explosion of sequence variants of uncertain significance [39] . We hypothesize that together with improved algorithms designed to predict the effect of nucleotide changes on protein function , such pathogenicity scores and Bayesian methods should facilitate prioritization of sequence variants . Overall , this study represents a step towards a quantitative framework of phenotype-specific variant interpretation . We suggest our method could be utilized for any other phenotype for which a sufficiently broad set of training genes can be generated; and although we restricted our analysis to copy number variants , this approach could be fruitfully applied to sequence variants as well . Our results have highlighted novel candidates in epilepsy and have provided further evidence of oligogenic inheritance in human disease . We believe that integrative approaches such as ours will become more accurate and useful with improved knowledge about genes and the molecular basis of disease as well as with the increased availability of genome wide profiles . We developed a training set of clinically accepted genes that , when altered , result in epilepsy ( Table S2 ) , by searching the Online Mendelian Inheritance in Man ( OMIM ) database for all entries containing the words “epilepsy” or “seizure . ” All genes matching these criteria were then manually curated based on published evidence that alterations in a given gene cause or increase susceptibility to epilepsy . Gene alterations that cause syndromes in which epilepsy is poorly penetrant , for which suspected pathogenicity is based on only a few poorly characterized patients , or for which statistically significant association was not demonstrated in at least one study were excluded . Manual review resulted in the identification of 83 epilepsy training genes and included the twenty epilepsy genes recognized by the ILAE [23] . We chose not to exclude or differentially weight genes associated with autosomal recessive or x-linked epilepsy as to have a more broadly defined pathogenicity score . This aspect is an area for further research . Individual features were designed by comparing the epilepsy training gene set to every RefSeq gene based on Gene Ontology [40] , MGI phenotypes [41] , miRNA targeting [42] , KEGG molecular interaction network data [43] , GeneAtlas expression distribution [44] , and PPI networks [45] . For the GO , MGI phenotypes , KEGG , and microRNA targeting data , gene level feature scores were determined by a four-step process . First , we identified annotations in each ontological system that are enriched among the set of epilepsy training genes . We used p-values for enrichment , odds ratios and number of training genes annotated to select categories; the cut-off values used to define enrichment for each annotation system are listed in Table S3 . Second , we used a novel scoring method to quantify each gene's annotation match to the training genes . For detailed descriptions and formulae , see Text S1 . We used these scores to determine a pair of empirical distributions , one for the training genes and one for the background genes; for each gene we computed the fraction of genes ( training and all other RefSeq separately ) found to have scores equal to or higher than the index gene's score . The logarithm of the ratio of these probabilities served as the metric for each gene . The log-ratio was transformed by subtracting the mean and dividing by the standard deviation and served as a standardized score for a given gene in a given annotation system . Table S2 lists all annotation system scores for each RefSeq gene . For gene expression we used the GeneAtlas data to identify those tissues where the epilepsy training genes were most highly expressed and where they were least expressed . Then , for each gene , we determined a score motivated by the T-statistic to measure the difference between high vs . low tissue expression . We again used the probability ratio and transformation approach to determine a score for each gene ( Table S2 ) . For PPI data , we used a measure based on network communicability between each gene's protein product and those of the epilepsy training genes as our feature . Communicability measures the total number of paths that connect pairs of nodes in a network scaled by the length of each path [46] . This concept takes into account the principal that the existence or non-existence of a direct interaction between proteins does not capture fully how connected two gene products are . For example , if a pair of proteins shares a number of interacting partners , they should be considered closer than two proteins that are only connected via a single sequence of interactors of the same length . Given the scale and connectedness of the PPI network , we chose to consider all paths of length six or less ( see Text S1 ) . As before , we used the probability ratio and transformation approach to determine a score for each gene ( Table S2 ) . Composite pathogenicity scores were generated from the mean of all 6 features . When data were missing because feature information was not present for a given gene in a given annotation system , we calculated the mean of the available features ( Table S2 ) . Genes for which no data was available in any feature received a zero score . To establish the performance of our scoring procedure , we performed leave-one-out cross validation . We sequentially dropped one gene out and established the pathogenicity scores for all sources of information ( GO , MGI , KEGG , miRNA , expression , and PPI ) using the remaining training genes and the RefSeq complement of the full training set . Then , for the omitted training genes , we determined the individual ratios and the mean of ratios across the features . Finally , we compiled the cross-validated results and determined the percentages of epilepsy training genes that met or exceeded each percentile cutoff . Approval to conduct retrospective analyses of clinical laboratory data and clinical records from the Baylor College of Medicine ( BCM ) Molecular Genetics Laboratory ( MGL ) databases and the Texas Children's Hospital ( TCH ) electronic medical record ( EMR ) using protocol H-27825 was obtained from the institutional review board for BCM and affiliated hospitals . Our goal was to identify all local patients with an epilepsy diagnosis who had a clinical aCGH study performed at our institution . We searched the TCH EMR using ICD-9-CM codes 333 . 2 , 345 , 649 . 4 , 780 . 39 and their respective subordinate four and five digit codes , where applicable , in order to identify patients with an epilepsy diagnosis who were seen at TCH between February 2004 and April 2011 . We then identified the subset of these patients who had a clinical aCGH study in the MGL database . In an attempt to detect additional patients with epilepsy who had aCGH performed but for whom the clinician may not have properly coded the epilepsy diagnosis in the EMR , we also searched the MGL database referral indications using the search terms “epilepsy” , “seizure” , “infantile spasm” , “convulsion” , and variations thereof . In total , we identified 1 , 641 local patients with evidence of an epilepsy diagnosis who also had aCGH performed during the queried period of time . Medical records for patients with abnormal aCGH test interpretations were reviewed by a diplomate of the American Board of Psychiatry and Neurology with Special Qualification in Child Neurology . Patients for whom a diagnosis of epilepsy ( defined as recurrent unprovoked seizures ) could not be confirmed were excluded . This resulted in a set of 295 patients with abnormal aCGH results and confirmed epilepsy . 84 of patients were tested by BAC arrays or with other manufacturers' array platforms that were incompatible with assessment across cohorts and thus were excluded , leaving 211 patient with EMR confirmed epilepsy and genome-wide results . For comparison purposes , we also generated an analysis cohort of patients with an abnormal aCGH report from a comparable array version but no neurological referral indications or ICD-9-CM codes consistent with neurological disease . Because patients without any neurologic indication or diagnosis comprise a considerable minority of our aCGH database , we were forced to expand our search to patients referred from outside hospitals . A list of included and excluded indication and code classes are listed in Table S4 . We also generated a comparator cohort of subjects with indications consistent with autism but not epilepsy by searching indications for the word “autism” and removing any subjects in the epilepsy cohort . Array CGH was performed on genomic DNA extracted from the patients' peripheral blood lymphocytes using various versions of oligonucleotide microarrays depending on the date of submission . Each oligonucleotide array was custom designed by the MGL at BCM ( Houston , TX , USA ) and manufactured by Agilent Technologies ( Santa Clara , CA , USA ) . Data were analyzed utilizing a custom designed BCM statistical analysis package implemented in the R programming language ( R Core Development Team ) . Segments of each analysis cohort individual's genome potentially varying in copy number were determined from oligonucleotide log2 patient vs . control intensities as previously described [11] , [47] . To select calculated intervals representing true positives , we limited our analysis to intervals smaller than 15 Mb , with at least 4 variant probes , and a mean log2 ratio <−0 . 3 or >0 . 21 . Gene lists were calculated by selecting RefSeq genes for which any part of the annotated sequence is contained within the minimum interval defined by the first and last deleted or duplicated oligonucleotide probe for each interval . The maximum composite pathogenicity score was then computed for each patient . To inform our analysis using frequency information , we generated gene deletion frequency data from our clinical database . Because observed frequency rates may be influenced by changes to the microarray design , we elected to limit our analysis to 23 , 578 patients tested by the BCM version 8 array series . We selected individuals based on their indication for procedure as described in the analysis cohort section . This process resulted in 1 , 616 individuals with indications consistent with epilepsy and 2 , 940 individuals with indications consistent with disease but of non-neurologic etiology . These frequency cohorts fundamentally differ from the analysis cohorts in that they also contain CNVs from individuals that did not have clinically abnormal array CGH findings . Because 73 individuals in the epilepsy analysis cohort were tested by version 8 arrays , they occur in both the analysis and frequency cohorts . For each RefSeq gene , we parameterized a prior gamma distribution with a mean equal to the observed rate of deletion CNVs among subjects referred to our diagnostic center for non-neurologic indications and variance constrained to be a constant multiple of the mean . If no deletion CNV was observed , we supplied a rate of 0 . 085 deletions per thousand subjects , equal to 1/4th the lowest observed rate . We then parameterized a second gamma distribution by allowing the mean parameter to increase for genes with positive mean pathogenicity scores by scaling the mean as an increasing function of the pathogenicity score . To also take into account the rarity of the variant in question , we developed a scaling function that adjusted the influence of the pathogenicity score inversely with variant frequency in the control population ( see Text S1 and Figure S3 ) . Variants commonly seen in the population would be expected to have less rate change among subjects with epilepsy . Conversely , rarely variant genes in the control population with high pathogenicity scores would be modeled to be more highly variant in epilepsy . This process resulted in two prior gamma distributions for each gene , one informed only by frequency in a non-neurologic cohort and one informed both by frequency and gene knowledge as encoded in our pathogenicity score . For genes where the pathogenicity score was negative , we retained the same prior distribution for the non-neurologic cohort . Next we calculated the probability of the observed rate of deletion CNVs among the epilepsy frequency cohort ( see Text S1 ) under the two gamma distributions . The ratio of these probabilities , the Bayes factor , served as an overall score for each gene . Finally , we calculated the posterior rate parameters for each gene taking into account background frequency , pathogenicity , and the observed rate among subjects with epilepsy . Table S2 lists the Bayes factors and posterior rate parameters for each RefSeq gene . We calculated lists of genes disrupted by deletions in each individual for all members of the autism , epilepsy and non-neurologic cohorts as described above . Because 73 members of the analysis cohort are part of the frequency cohort , their personal contribution to frequency data artificially inflates the Bayes factor score by definition . To avoid this inflation , we recalculated the genome-wide Bayes factors again , but removing their contribution to the frequency data . We then selected genes with Bayes factors greater than 1 and sorted them in decreasing order . We performed statistical analysis on the first , second , and third highest scoring genes for each individual . Table S5 lists the top three genes and their Bayes factors for each individual .
Improvements in sequencing and microarray technologies have increased the resolution and scope of genetic testing . As a result , millions of variations are identified in each personal genome of unrelated individuals . In the context of testing for genetic diseases , identifying the variant or variants contributing to illness among such a large number of candidates is difficult . Conventional studies to identify causative variants have relied on patterns of higher frequency in affected patients compared with individuals that are well . However , it is often the rarest variations that cause human disease , making frequency information alone less useful . Many groups have turned to computational analysis to aid in interpretation of genetic variants . Epilepsy is a disease where such tools would be useful , as only a fraction of patients with suspected genetic epilepsy have a specific genetic diagnosis . To help improve variant interpretation in epilepsy , we used computational analysis to combine knowledge about genes from large cloud information sources with mutation frequency from our diagnostic laboratory to score all genes as to how likely they are to be associated with epilepsy . We use these scores to identify possible candidate genes in epilepsy , and explore other downstream applications .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[]
2013
Fusion of Large-Scale Genomic Knowledge and Frequency Data Computationally Prioritizes Variants in Epilepsy
The Sec1/munc18 protein family is essential for vesicle fusion in eukaryotic cells via binding to SNARE proteins . Protein kinase C modulates these interactions by phosphorylating munc18a thereby reducing its affinity to one of the central SNARE members , syntaxin-1a . The established hypothesis is that the reduced affinity of the phosphorylated munc18a to syntaxin-1a is a result of local electrostatic repulsion between the two proteins , which interferes with their compatibility . The current study challenges this paradigm and offers a novel mechanistic explanation by revealing a syntaxin-non-binding conformation of munc18a that is induced by the phosphomimetic mutations . In the present study , using molecular dynamics simulations , we explored the dynamics of the wild-type munc18a versus phosphomimetic mutant munc18a . We focused on the structural changes that occur in the cavity between domains 3a and 1 , which serves as the main syntaxin-binding site . The results of the simulations suggest that the free wild-type munc18a exhibits a dynamic equilibrium between several conformations differing in the size of its cavity ( the main syntaxin-binding site ) . The flexibility of the cavity's size might facilitate the binding or unbinding of syntaxin . In silico insertion of phosphomimetic mutations into the munc18a structure induces the formation of a conformation where the syntaxin-binding area is rigid and blocked as a result of interactions between residues located on both sides of the cavity . Therefore , we suggest that the reduced affinity of the phosphomimetic mutant/phosphorylated munc18a is a result of the closed-cavity conformation , which makes syntaxin binding energetically and sterically unfavorable . The current study demonstrates the potential of phosphoryalation , an essential biological process , to serve as a driving force for dramatic conformational changes of proteins modulating their affinity to target proteins . Intracellular membrane fusion in eukaryotes is mediated by a well-conserved fusion machinery composed of SNARE ( soluble N-ethylmaleimide-sensitive factor attachment protein receptor ) and SM ( Sec1/munc18-like ) proteins [1] . In the early studies , munc18a was shown to bind syntaxin , one of the central SNARE members and block ternary SNARE complex formation , suggesting that it plays a negative regulatory role [2] , [3] . However , genetic and biochemical studies indicated that SM proteins play a positive essential role as demonstrated by their null mutants; studies with mutated worms , flies and mice lacking munc18a , revealed a dramatic decrease in secretory granule fusion , docking and priming [4] , [5] , [6] . Therefore , the central hypothesis , to date , is that SM proteins play several roles depending on their mode of binding to the SNARE members [7] , [8] . The first mode of interaction that was discovered [9] relates to the binding of munc18a to a stable closed-conformation of syntaxin . This mode of interaction allows the specific transfer of syntaxin through the endoplasmic reticulum and the Golgi apparatus to the plasma membrane , keeping syntaxin from engaging to ectopic intracellular SNARE complexes [10] , [11] . Recent studies demonstrate that SM proteins bind only or additionally to a short peptide present at the N-terminus of syntaxin , designated as the N-peptide [1] , [10] , [12] . This mode of interaction was intensively investigated in the last few years and its importance is under a strong debate . One of the main hypotheses for the role of the interaction of munc18a with the N-terminal of syntaxin is that this interaction allows munc18a to bind the SNARE ternary complex suggesting a stimulatory role for munc18a in the last stages of SNARE-mediated fusion [13] . The rat munc18a , which was structurally resolved as part of the complex with syntaxin-1a [9] , [12] , is an arched-shaped three-domain protein ( Figure 1A ) that embraces syntaxin in a cavity located between domains 3a and 1 ( Figure 1 , A and B ) . Phosphorylation by protein kinase C ( PKC ) or phosphomimetic mutations in residues 306 and 313 ( S306D , S313D ) of munc18a modulate this interaction by reducing the affinity to form a complex [14] , [15] . Previous studies have suggested that replacement of the polar serine moieties in domain 3a of munc18a by phosphate groups or negatively charged glutamates disrupts the complex due to electrostatic repulsion between munc18a and the adjacent area of syntaxin ( Figure 1B ) , which contains acidic residues [14] . In the present study , munc18a dynamics was studied , for the first time , using molecular dynamics ( MD ) simulations under different conditions for several hundred nanoseconds . We show that in the absence of syntaxin , wild-type munc18a exhibits a dynamic equilibrium between several states , differing in the size of the syntaxin-binding site ( the cavity between domains 3a and 1 ) . In the next step , we examined the dynamic behavior of the phosphomimetic munc18aS306D , S313D and we show that following in-silico insertion of the mutations into the wild-type structure , munc18a adopted a rigid closed-cavity conformation which makes syntaxin binding less probable . The closed-cavity conformation is induced specifically by the PKC phosphomimetic mutations and reversible upon dephosphorylation of the protein back to the wild-type form . In the present study , munc18a dynamics was studied using molecular dynamics ( MD ) simulations , a powerful method in which the dynamics and conformational changes in proteins can be followed in a virtual fashion . We performed three MD simulations of the wild-type munc18a ( termed 1 , 2 and 3 , Table 1 ) as described in details in the Methods section . The simulations were performed under the same conditions; accept for applying two distinct informatics tools; Swiss-Pdb [16] or Rosetta [17] , [18] , [19] for in silico reconstruction and structural modeling prediction of regions in the protein that were not resolved in the crystal structure [9] . The high resemblance of the basic dynamics characteristics of munc18a in the three simulations ( Text S1 and Figures S1 and S2 ) allowed us to evaluate the general relative inter-domain motions of the protein and attribute them to the activity and function of the protein . We monitored specifically the changes in the main syntaxin-binding site of munc18a , which is the area of the cavity between domains 3a and 1 [9] . We first measured the change in the distance between the centers of mass of domains 3a and 1 ( Figure 1C ) and of the distance between specific residues ( Gly 26 in domain 1 and Glu273 in domain 3a ) on both sides of the cavity ( Figure 1 , D–E ) during the simulations . The measurements showed that the distances frequently change , indicating structural fluctuations of these domains and dynamic changes in the size of the cavity , becoming wider or narrower ( Figure 1 , C–E ) . During the simulations , the main motions of the protein were isolated from its overall movement using an essential dynamics ( ED ) analysis . ED analysis is a method for isolating the various modes of motion of a protein during the simulation by yielding a set of eigenvectors corresponding to its internal motions namely the amplitudes and the directions of the motions [20] . The vectors are scaled according to the time scale of the motion from the slowest undulations which generally correspond with motions of large regions in the protein , and up to the fast and high-frequency local fluctuations . The ED analysis of the wild-type munc18a simulations clearly illustrated that the main motion vectors exhibit opening and closure of the cavity between domains 3a and 1 ( Figure 1 , F–G ) dramatically changing its size . The high flexibility in the size of the munc18a cavity probably assists in binding or unbinding of syntaxin or other target proteins that bind munc18a in other regions as well ( such as CDK5 for example ) [9] , [21] . Squid munc18 ( sSec1 ) , a homolog of the rat protein ( munc18a ) , has been crystallized as a free protein , i . e . unbound to the squid syntaxin [22] , and three variations of the structure are available . The following section examines the similarity between the dynamics behavior of the wild-type munc18a during the simulations and the resolved crystal structures of its squid homolog , sSec1 . Figure 2 presents a superposition of the three available sSec1 crystal structures ( 1EPU . pdb , 1FVF . pdb and 1FVH . pdb ) and the munc18a crystal structure ( 3C98 . pdb ) . In the three simulations , domain 3a , and particularly the β-hairpin ( residues 263–280 ) exhibited high structural variability , sampling manifold structures ( Figure 2B ) . Similarly , the three resolved sSec1 structures exhibit high variability among them in the structure of domain 3a . In the simulations , domain 1 remarkably preserved its secondary structure and we observed a clear rotational motion of this domain . Similarly , superposition of the three crystal structures of the squid protein shows that they share the same secondary structure for domain 1 , but domain 1 is positioned in a slightly different angle reflecting a rotation motion of this domain ( Figure 2C ) . The β-hairpin in domain 3a of munc18a ( residues 261–280 ) plays a prominent role in the interaction of munc18a with syntaxin-1a . Eight amino-acid residues out of the 19 that compose the β-hairpin are engaged in interactions with the H3 domain of syntaxin , making the hairpin an essential element in the binding of syntaxin , and in keeping syntaxin in its closed ( inactive ) structure . Therefore , any fluctuations in the position of the β-hairpin might influence the affinity of syntaxin to munc18a and might cause syntaxin to alternate to its open structure . Comparison of the munc18a structure to the crystal structure of Sly1p , the yeast Golgi homolog of munc18a ( 1mqs . pdb , downloaded from PDB [23] ) indicates that the hairpin of the later , although partially unstructured , resides in a much higher position than in the munc18a crystal structure ( Figure 3A ) . In this structure , the Golgi resident syntaxin ( sed5p ) is absent from the cavity area and the crystal structure only includes its N-terminal peptide which is bound to the area of domain 1 . Interestingly , in simulation 2 of munc18a ( Table 1 ) , we traced a prominent motion of the β-hairpin of the protein , moving during the simulation from its original location in the crystal structure outwards and upwards , protruding from the rest of the protein ( Figure 3 , B and C ) . This motion , resulting in a position similar to that seen in the crystal structure of the Sly1p , confirms the possibility raised before that the motion of the β-hairpin might serve as a mechanism for the release of syntaxin from the cavity area [22] . To further confirm this notion , we reconstructed the full structure of Sly1p ( see Methods ) and performed a simulation of its dynamics under the same conditions of the munc18a simulations . We focused on the movement of the β-hairpin in domain 3a ( residues 298 to 327 ) during the simulation . Indeed , during the 20 ns of the simulation , we observed an extensive rotation-translation movement of this region downwards approaching domain 3a ( Figure 3D ) and consequently narrowing the width of the cavity . Thus , the β-hairpin might serve as a gate for the cavity , opening and closing the cavity when needed . In the current study , we show that this motion can occur spontaneously with no interference from any additional factor ( s ) ; although we cannot determine the probability of this type of motion in the free or syntaxin-bound munc18a , these data support the hypothesis that the β-hairpin serves as a switch for syntaxin-binding or unbinding . After characterizing , in detail , the dynamics of the wild-type munc18a , the next step of our study was to examine the dynamic behavior of the phosphomimetic munc18aS306D , S313D and determine the differences compared to the wild-type dynamics . Characterizing the differences in the dynamics of munc18aS306D , S313D will assist to determine a mechanism that might explain the reduced affinity of syntaxin to munc18aS306D , S313D/phosphorylated munc18a . Following in-silico insertion of the mutations ( See Methods ) into the wild-type structure , the mutant was simulated under the same conditions as the wild-type ( ∼35 ns , simulation M1 , Table 1 ) . Strikingly , analysis of the fluctuations in the distance between the centers of mass of domains 3a and 1 indicated a marked decrease in the distance between the centers of mass of domains 3a and 1 ( Figure 4A ) . Already in the first ∼3 ns of the simulation , the distance decreased from 3 . 9 nm to 3 . 3 nm , and during the rest of the simulation , the distance stabilized ( ∼3 . 5 nm ) exhibiting only further minor fluctuations ( Figure 4A ) . Observation of the mutant dynamics showed that the decrease in the distance between the centers of mass of domains 3a and 1 represents a process of closure of the cavity between these domains . The closure leads to the preferential stabilization of a distinct closed-cavity conformation of munc18aS306D , S313D , a conformation that probably cannot bind syntaxin via the cavity . Further examination of the closure process shows that the conformational change in the protein includes a structural disruption in the area of the mutations , which is located in domain 3a , on the side opposite to the cavity ( Figure 4 , B–D ) . Calculation of the average local RMSF ( Root Mean Square Fluctuations ) in the area of the mutations ( residues 306–313 ) showed a large increment of 28% to 120% in the specific RMSF values of these residues with respect to the wild-type values , indicating substantial movements of this region ( Figure 4E ) . This structural disruption on one side of domain 3a might destabilize its overall structure , allowing the area adjacent to the cavity to move towards domain 1 , located on the other side . ED analysis performed both using Dynatraj and by GROMACS for the most dominant motions of the munc18aS306D , S313D simulation ( Methods ) , demonstrated that the closing motion of the cavity ( Figure 5A ) occurs as part of the most dominant motion in the simulation . The GROMACS-based ED analysis shows that the most dominant motion in the munc18aS306D , S313D simulation is twice the size of the main motion of the wild-type munc18a simulation and encompasses 37% of the total movement of the protein during the simulation ( Figure 5B ) . To examine whether the phosphomimetic mutations can induce the closure of munc18a cavity also when the structure was already well-relaxed , In-silico phosphomimetic mutations were inserted into the well-relaxed structure of munc18a ( the structure obtained after 35 ns simulation of the wild-type , see Methods and Table 1 ) , and the structure was simulated from that point for another 35 ns ( Simulation M2 , Table 1 ) . The phenomenon of cavity closure was clearly reproduced , but the time course of the process was different ( data not shown ) . Comparison of the structures in the first and last frames of this simulation clearly illustrates two distinct conformations: the initial open-cavity conformation and the final closed-cavity conformation ( Figure 5 , C and D ) . For comparison to the wild-type simulation ( Figure 1E ) , the change in the distance between residues Gly 26 in domain 1 and Glu 273 in domain 3a , during this munc18aS306D , S313D simulation , is presented . The distance between these residues decreased during the last 20 ns of the simulation ( Figure S3 ) demonstrating again the closure of the cavity . In addition , a video of the trajectory of this simulation , demonstrating the closure process is presented ( Video S1 ) . To determine the relative stability of the structures that munc18aS306D , S313D samples during the simulation , and to identify the most stable and dominant structure in the mutant simulations , we had used another quantitative analysis tool for the simulations termed , cluster analysis ( [24] , Methods ) . Briefly , Cluster analysis segments the structures that the protein samples during the simulation into sub-groups ( termed , clusters ) . The structures are divided to clusters according to an adjustable RMSD ( Root Mean Square Deviations ) cut-off value that defines the variance between structures that populate the same cluster ( [24] , Methods ) . Comparison of the cluster analyses performed for the phosphomimetic munc18aS306D , S313D and the wild-type munc18a simulations , using the same RMSD cut-off value , showed that munc18aS306D , S313D samples fewer conformations than the wild-type during the simulations , having less distinct clusters , 48 vs . 72 respectively ( Table 2 ) . Moreover , the three largest clusters in the munc18aS306D , S313D simulation encompass about 27% of the total structures population compared with only 16% as determined in the wild-type munc18a cluster analysis . Analyzing the size of the syntaxin-binding cavity in the three largest clusters of the munc18aS306D , S313D compared to the wild-type shows that the three largest clusters in the phosphomimetic munc18aS306D , S313D simulation demonstrated a smaller variance in the values of the distances between the centers of mass of domains 3a and 1 exemplifying that the size of the cavity is relatively unchanged compared to the size of the cavity in the wild-type three main clusters . As detailed in Table 2 , the mean distance between the centers of mass of the two domains is significantly shorter for the phosphomimetic munc18aS306D , S313D illustrating that a closed-cavity conformation predominates in the phosphomimetic munc18aS306D , S313D three largest clusters ( Table 2 ) . In summary , the cluster analysis shows that munc18aS306D , S313D samples less distinct conformations during the simulations . The mutant is more rigid in the cavity's region than in the wild-type simulation and a closed-cavity conformation predominates . In order to identify the driving force for the cavity closure process , we examined the energetic components ( Lennard-Jones [LJ] and electrostatic ) of munc18aS306D , S313D during the simulation time . Inspection of the change in the energetic components of munc18aS306D , S313D shows that the closing motion of the protein was correlated with a decrement in the sum of the electrostatic and LJ energy components of the system indicating stabilization of the structure ( Figure 6 , A and B ) , and the formation of extra electrostatic and hydrophobic interactions . Specifically in the cavity area , we monitored the time-dependent pattern of hydrogen bonds and found that three to five additional hydrogen bonds were formed during the simulations between residues located on both sides of the cavity ( Figure 6D ) . The later is in contrast to the wild-type simulation , that during the same simulation time , the number of hydrogen bonds between domains 3a and 1 fluctuated between 0–2 ( Figure 6C ) . The interactions between residues located on both sides of the cavity kept them in proximity and stabilized the closed-cavity conformation . In addition to the Hydrogen bonds that were formed during the munc18aS306D , S313D simulation , LJ interactions between residues in domains 3a and 1 further stabilized the closed state ( Figure 6B and Figure 7 , A and B , residues in green and yellow ) . Table 3 summarizes the interactions observed during the simulation between residues located on both sides of the cavity . It should be noted that Arg39 [9] and other munc18a residues that are essential for its interaction with syntaxin , were found to be involved in the inter-domain interactions , bringing both sides of the cavity together . Similarly , Lys46 ( domain 1 ) that is involved in the interaction with syntaxin forms an electrostatic interaction during the free munc18aS306D , S313D simulation ( M2 ) with Asp262 located in domains 3a . Figure 7C depicts the decrease in the distance between Lys46 and Asp262 to ∼2 . 5 nm , as they approach each other during the simulation , forming a stable electrostatic interaction already after 5 ns of the simulation ( Figure 7 , C–E ) . The analysis of the energetic components of the munc18aS306D , S313D system during the simulation shows that the mutant protein ( munc18aS306D , S313D ) is energetically-stabilized in the closed-cavity conformation in which residues on both sides of the cavity interact with each other . Therefore , the binding of syntaxin to munc18aS306D , S313D requires breaking several intra-molecular electrostatic bonds and as a result might become energetically unfavorable . We next investigated whether the closed-cavity conformation is reversible , whether it is induced directly by the insertion of the phosphomimetic mutations and whether the protein can regain its flexibility in the area of the cavity . We removed the phosphomimetic mutations from the structure of the protein in the last frame of the munc18aS306D , S313D simulation and performed another simulation of 36 ns of this structure mutated back to the wild-type ( D306S , D313S ) . This simulation showed that the back-mutated wild-type protein gradually regains its dynamic nature in the cavity area and the cavity starts to reopen ( Figure 8 , A and B ) . The distance between the centers of mass of two regions adjacent to the cavity: residues 35–70 ( domain 1 ) and residues 260–280 ( domain 3a ) increased from 1 . 8 nm to 2 . 2 nm during the 36-ns back-mutation simulation ( munc18aD306S , D313S , Figure 8A ) . Next , we extended this analysis by looking at the relative motion of larger sections of the protein; measurement of the distance between the centers of mass of domains 3a and 1 during the munc18aD306S , D313S simulation indicated that the distance gradually increases from 3 . 4 nm up to 3 . 8 nm , reflecting reopening of the cavity . The opening movement of the cavity was also observed by ED analysis; in Figure 8C , we present a porcupine plot of the fourth eigenvector of the dynamics demonstrating by the direction and length of the ‘needles’ a clear expansion of the cavity . Finally , a straightforward superposition of domains 3a and 1 from the last frames in the simulations of munc18aS306D , S313D ( t = 35 ns , red ) and munc18aD306S , D313S ( t = 36 ns , blue ) indicates that the positions of domains 3a and 1 are further away from each other in the back-mutated munc18aD306S , D313S compared to the phosphomimetic munc18aS306D , S313D and similarly to the wild-type ( Figure 8D ) . The results suggest that the phosphorylation/phosphomimetic mutations induce a closed-cavity conformation that can be reversed upon dephosphorylation/back-mutations of the protein . Mutations of Ser 306 and Ser 313 to Ala in munc18a were shown to turn the protein to be non-phosphorylated and had no affect on syntaxin binding [14] , [25] . To check the specificity of the closed-cavity conformation to the phosphomimetic mutations in these positions , another simulation ( 36ns long ) was performed , following the dynamics of the non-phosphorylated mutant munc18aS306A , S313A under the same conditions as in the previous simulations . Measurement of the distance between the centers of mass of domains 3a and 1 during the munc18aS306A , S313A simulation shows that the distance was fluctuating between 3 . 5 to 4 . 2 nm ( Figure 8E ) , similarly to the fluctuations that were observed in the wild-type simulation ( Figure 1C ) . We did not track any substantial movement of domains 3a and 1 towards each other , thus , no closure of the cavity was observed as was depicted in the phosphomimetic munc18aS306D , S313D simulations . Analysis of the time-dependent change in the number of hydrogen bonds between domains 3a and 1 shows that the number of hydrogen interactions remained 0 or 1 during most frames of the simulation indicating that no new hydrogen bonds were formed during the simulation between residues located in domains 3a and 1 , in contrast to the phosphomimetic munc18aS306D , S313D simulations ( Figure 8F ) . In summary , the mutant munc18aS306A , S313A did not adopt a closed-cavity conformation ( Figure 8G ) and the dynamics resembled that of the wild-type state ( Figure 8 , E–G ) indicating specificity of the closing phenomenon to the phosphomimetic mutations in these positions . The current study reveals , for the first time , new conformations that munc18a can adopt when it is unbounded to syntaxin . Based on a rigorous analysis of a comprehensive set of molecular dynamics simulations we were able to monitor the dynamics of the wild-type free munc18a in comparison to its mutant forms ( phosphomimetic , back-mutated and non-phosphorylated mutants ) , focusing on the structural changes that occur during the trajectories in the main syntaxin-binding site , the cavity between domains 3a and 1 . We show that munc18a , in its syntaxin-unbounded form , is in a dynamic equilibrium between conformations varying in the size of its syntaxin-binding cavity located between domains 3a and 1 . Specifically , we found that munc18a can adopt a stable conformation where its cavity , serving as the main syntaxin-binding site , is mostly blocked by inter-domain interactions . This conformation is induced following in silico insertion of phosphomimetic mutations in positions 306 and 313 ( S306D , S313D ) . We propose that the observed reduction in affinity of munc18a to syntaxin following phosphorylation or insertion of phosphomiemtic mutations as shown experimentally is a result of preferential stabilization of a conformation of munc18a where the syntaxin-binding site is less accessible for syntaxin . This conformation of munc18a makes the binding of syntaxin less probable , and energetically and sterically unfavorable . Our proposed mechanistic explanation is supported by a few studies carried out in the past that already speculated that munc18a might have additional distinct conformations different from the one that was resolved ( bound to syntaxin ) in the published crystal structure [9] , [12] . Previous studies suggested that the munc18a conformations could be induced by interactions with other proteins , such as Rab , Rab effector or munc13 [9] . However , to the best of our knowledge , there is no other available resolved conformation ( i . e crystal structure ) of munc18a . Another indication for the existence of several munc18a conformations is the putative binding site of the protein cyclin-dependent kinase 5 ( CDK5 ) in munc18a . CDK5 has been shown to phosphorylate munc18a and to mediate the disassembly of the munc18a-syntaxin-1a complex , with the assistance of other proteins . The site of CDK5-mediated phosphorylation in munc18a is located between domains 2 and 3 ( residue Thr574 ) . In the crystal structure of the munc18a-syntaxin complex , this region of munc18a is buried in the protein and therefore inaccessible , indicating that CDK5 probably interacts with a different conformation of munc18a that was not determined yet [9] , [21] . The closed-cavity conformation of munc18a is specifically induced by the phosphomiemetic mutations; however it is not exclusively present in this mutated form of the protein . Molecular dynamics simulation of another mutated form of munc18a - munc18aF115E [13] showed that the introduction of this mutation induced closure of the cavity as well ( Figure 9 , A–B ) . The closure was initiated by a dominant movement of domain 1 towards domain 3a . These results suggest that the closed-cavity conformation can be driven by several types of mutations . The tendency of the protein to form this conformation might be a general mechanism explaining the impaired binding of several mutated forms of munc18a to syntaxin . The established hypothesis attributes the reduced affinity of the phosphorylated munc18a ( or the phosphomimetic mutant munc18aS306D , S313D ) to syntaxin to the local repulsion of syntaxin by the negative charges of the phosphates ( or glutamates ) in this region of munc18 . This repulsion was suggested to reduce the compatibility and the overall affinity of the complex [14] . The hypothesis presented in the current study , based on extensive molecular dynamics simulation and analyses , challenges this paradigm and suggests that the reduced affinity results from closure of the cavity of munc18a , making it inaccessible for syntaxin binding in this area . Many key biological processes such as the synaptic processes [21] , [26] are regulated by protein phosphorylation . In order to understand the effects of this process , it is essential to characterize specifically the structural changes induced by phosphorylation , leading to a change in the affinity of proteins to target proteins . In this study , we followed structural changes that phosphorylation might induce and we were able to provide a novel mechanism for explaining experimental results showing reduced affinity between proteins . As the potential of phosphorylation to induce substantial conformational changes in proteins was already shown in previous studies [27] , [28] , [29] , [30] , we suggest that the present conventional paradigm , explaining the reduced affinity of the phosphorylated munc18a to syntaxin as merely a local repulsive phenomenon , is rather simplified . Efforts should be aimed at tracking the global dynamic conformational changes that occur in the phosphorylated munc18a or in other mutated forms of munc18a in attempt to resolve munc18a conformations and in particular the closed-cavity conformation . All simulations performed were using the coordinates of munc18a crystal structure that were taken from the recently refined crystal structure of the syntaxin-1a-munc18a complex , determined by x-ray crystallography at a resolution of 2 . 6 Å [9] , [12] . The crystal structure coordinates , taken from the Protein Data Bank ( PDB file: 3C98 . pdb ) , include 556 residues out of the 594 residues of the full sequence of munc18a: 6 residues ( 317–323 ) in domain 3a and 25 residues ( 506–531 ) in domain 2 have not been structurally resolved . In addition , at both terminals; the first three residues of the N-terminal ( amino-acid residues 1–3 ) and residues 593–594 of the C terminal were not resolved as well . Three simulations of wild-type munc18a were performed differing in the tools used for completion and structural prediction of the missing regions . In simulations 1 and 2 , the missing residues were added to the structure and modeled using the Swiss-PDB program ( [16] , http://www . expasy . org/spdbv/ ) and in simulation 3 , the Rosetta software [17] was used for the completion and modeling as detailed below . In the Swiss-PDB , an energy-minimizing computation was performed by the Swiss-PDB tool using Gromos96 implementation of the Swiss-PDBViewer following the addition of the residues . All MD simulations presented in this article were performed using the GROMACS 4 . 0 suite of software [31] , using the GROMACS 53a6 force field [32] . The protein was embedded in a dodecahedron box containing the SPC water molecules [35 , 097 molecules for the Swiss-PDB based structures ( 1 and 2 ) and 34 , 972 molecules for the Rosetta-based structure ( simulation 3 ) that was extended to at least 15 Å between the protein's structure and the edge of the box . Assuming normal charge states of ionizeable groups corresponding to pH 7 , the net charge of munc18a structure is −4e . Hence , 74 sodium and 70 chloride ions were added to the Swiss-PDB structure trajectory box at random positions , to neutralize the system at a physiological salt concentration of 100 mM . Similarly , 73 sodium ions and 69 chloride ions were added to the Rosetta structure trajectory box ( simulation 3 ) . The difference in ion numbers is a result of the difference in the number of water molecules . Prior to the dynamics trajectory , internal constraints were relaxed by energy minimization . Following this step , an MD equilibration run was performed under position restraints for 40 ps . Then , unrestrained MD runs were initiated . Two runs of 35 ns each were performed for the Swiss-PDB structure ( simulations 1 and 2 ) and a single run of ∼35 ns for the selected Rosetta structure ( simulation 3 ) . During the MD runs , the LINCS algorithm [33] was used in order to constrain the lengths of all bonds; the water molecules were restrained using the SETTLE algorithm . The time step for the simulation was 2 fs . The simulation was run under NPT conditions , using the Berendsen coupling algorithm to keep the temperature and pressure constant ( P = 1 bar; τP = 0 . 5 ps; τT = 0 . 1 ps; T = 300 K ) . Van der Waals ( VDW ) forces were treated using a cut-off of 12 Å . Long-range electrostatic forces were treated using the PME method . The coordinates were saved every 1 ps . Low-pass frequency filtering was performed on the simulations using the g_filter tool of GROMACS . The amino acid sequence of the protein Sly1p was fed into I-TASSER ( iterative threading assembly refinement algorithm ) , a 3D protein structure prediction tool [34] , [35] , [36] in order to predict the full length structure of the protein ( 671 residues ) . A partial structure of Sly1p-Sed5p complex crystal structure is available as well ( 1mqs . pdb , [23] ) . One of the best-scored Sly1p model structure obtained by the I-TASSER was chosen as the starting coordinates for the Sly1p MD simulation . The simulation was run for 15 ns under the same conditions and procedure as described for the munc18a . Simulations of the phosphomimetic double-mutant munc18aS306D , S313D were carried out using the same procedure as described for the wild-type simulations . The Swiss-PDB software was used for in silico replacement of Ser 306 and Ser 313 with glutamates . The positions of the mutated residues were optimized and the overall structure was subjected to energy minimization performed by the Swiss-PDB software and then by the GROMACS suite . More details regarding the simulations can be found in Table 1 . The simulations of the back-mutated munc18a ( munc18aD306S , D313S ) , the non-phosphorylated munc18aS306A , S313A and of munc18aF115E were performed in the same procedure described above . The Rosetta program [17] was used to model the missing regions in the crystal structure of munc18a using the loop modeling option as described in details in several studies [18] , [37] , [38] . Repeated runs of the full-length structure were performed generating a total of 1050 plausible structures . The structures were all scored according to their energy and the structure with the lowest score , representing the most probable structure , was chosen for the MD simulation , termed 3 . The required covariance matrix and eigenvectors for the ED analysis were obtained by applying the g_covar program of the GROMACS 4 . 0 package . The analysis was performed on the backbone of the protein . The trajectory was filtered using the GROMACS g_filter program . Movies of 1000 frames representing the pathway between the minimum and maximum points of the movement in the main eigenvectors of each trajectory were formed using the g_anaeig command . ED analysis was also performed using the Dynatraj tool which is a part of the Dynamite server ( http://dynamite . biop . ox . ac . uk/dynamite , [39] ) . Porcupine plots , to visualize the modes of motion taken from the simulations were produced using the Dynatraj tool [40] . Cluster analysis was performed for the simulations of wild-type and phosphomimetic mutant munc18aS306D , S313D by the command g_cluster of the GROMACS 4 . 0 package . The cluster analysis was performed using the Gromos algorithm with an RMSD cut-off value of 0 . 2 nm [24] .
Protein phosphorylation plays a significant regulatory role in multi-component systems engaged in signal transduction or coordination of cellular processes , by activating or deactivating proteins . The potential of phosphorylation to induce substantial conformational changes in proteins , thereby changing their affinity to target proteins , has already been shown but the dynamics of the process is not fully elucidated . In the present study , we investigated , by molecular dynamics simulations , the dynamic conformational changes in munc18a , a protein that is crucial for neurotransmitter release and interacts tightly with the SNARE syntaxin-1 . We further investigated the conformational changes that occur in munc18a when it is phosphorylated , reducing its affinity to syntaxin-1a . The results of the simulations suggest that there is a conformational flexibility of the syntaxin-unbounded munc18a that allows changes in the shape of the syntaxin-1a binding cavity . In silico insertion of phosphomimetic mutations into munc18a led to a reduction in the flexibility and closure of the syntaxin-binding site . We suggest that the reduced affinity of phosphorylated munc18a to syntaxin-1a stems from the difficulty of syntaxin-1a to bind to the munc18a closed-cavity conformation , induced by the PKC phosphorylation of munc18a .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[ "neuroscience/neuronal", "signaling", "mechanisms", "neuroscience/theoretical", "neuroscience", "biophysics/theory", "and", "simulation", "biochemistry/theory", "and", "simulation" ]
2011
Dynamic Conformational Changes in MUNC18 Prevent Syntaxin Binding
Type III secretion systems ( T3SSs ) are specialized macromolecular machines critical for bacterial virulence , and allowing the injection of bacterial effectors into host cells . The T3SS-dependent injection process requires the prior insertion of a protein complex , the translocon , into host cell membranes consisting of two-T3SS hydrophobic proteins , associated with pore-forming activity . In all described T3SS to date , a hydrophilic protein connects one hydrophobic component to the T3SS needle , presumably insuring the continuum between the hollow needle and the translocon . In the case of Enteropathogenic Escherichia coli ( EPEC ) , the hydrophilic component EspA polymerizes into a filament connecting the T3SS needle to the translocon composed of the EspB and EspD hydrophobic proteins . Here , we identify EspA and EspD as targets of EspC , a serine protease autotransporter of Enterobacteriaceae ( SPATE ) . We found that in vitro , EspC preferentially targets EspA associated with EspD , but was less efficient at proteolyzing EspA alone . Consistently , we found that EspC did not regulate EspA filaments at the surface of primed bacteria that was devoid of EspD , but controlled the levels of EspD and EspA secreted in vitro or upon cell contact . While still proficient for T3SS-mediated injection of bacterial effectors and cytoskeletal reorganization , an espC mutant showed increased levels of cell-associated EspA and EspD , as well as increased pore formation activity associated with cytotoxicity . EspP from enterohaemorrhagic E . coli ( EHEC ) also targeted translocator components and its activity was interchangeable with that of EspC , suggesting a common and important function of these SPATEs . These findings reveal a novel regulatory mechanism of T3SS-mediated pore formation and cytotoxicity control during EPEC/EHEC infection . EPEC and EHEC are related pathogens causing severe diarrhoeal diseases . EPEC and EHEC form Attaching and Effacing ( A/E ) lesions on the mucosal intestinal surface , corresponding to the destruction of enterocyte microvilli and the intimate bacterial adherence to the host cell plasma membrane onto an actin-rich pedestal structure [1] . A/E pathogens carry the Locus of Enterocyte Effacement ( LEE ) encoding a type III secretion apparatus ( T3SA ) that allows the delivery of bacterial effector proteins directly from the bacterial cytoplasm into the cytoplasm of eukaryotic cells [2] . The translocator proteins EspA , B and D are required for the injection of type III effectors . Upon cell contact , EspB and EspD insert into the host cell plasma membrane and associate into a pore-forming “translocon” complex . The hydrophilic translocator protein EspA polymerizes into a hollow filamentous structure connecting the T3SA needle to the translocon [3] . SPATEs are serine protease autotransporters that are widely spread among Enterobacteriaceae . SPATEs have been reported to cleave host proteins implicated in diverse functions [4 , 5] . EspC has been described to cleave focal adhesion proteins , following cellular internalization , as well as other eukaryotic proteins such as haemoglobin , pepsin and human coagulation factor V [5 , 6 , 7] . Epidemiological studies indicated that EspC was predominantly found in t-EPEC strains and that EPEC strains carrying EspC and the OI-122 pathogenicity island were associated with high virulence [8 , 9] . These observations suggest that EspC could contribute to bacterial virulence by regulating the action of virulence factors . Although secretion of EspC occurs through a T3SS-independent mechanism , several intriguing features link EspC and the T3SS . The expression of EspC is coupled to that of the T3SS , and secretion of EspC is also activated upon cell contact [10 , 11 , 12] . In addition , although the underlying mechanism is unclear , efficient uptake of EspC by host cells requires the T3SS [13] . These observations prompted us to investigate functional links between EspC and the T3SS . Here , we provide evidence that EspC cleaves the translocon components , EspD and EspA . EspC proteolytic activity regulates pore formation mediated by the T3SS to prevent cytotoxicity during bacterial infection . To study T3SS targets of EspC , bacterial secretion profiles were first analyzed . Bacteria were grown in DMEM medium to induce secretion of T3 substrates and bacterial supernatants were analyzed by Coomassie staining ( Experimental Procedures ) . Comparison of the profiles observed for wild-type ( WT ) E2348/69 EPEC and the escN mutant supernatants showed three major bands identified by mass spectrometry as type III secreted substrates EspB and EspD ( forming a single band of 35 kDa ) and EspA ( 22 kDa ) ( Fig 1A , arrows; Experimental Procedures ) , as well as EspC ( 110 kDa ) , an autotransported serine protease belonging to the SPATEs family . Interestingly , the EspC presence correlated with lower levels of EspB/D and EspA . Western-blot analysis showed increased levels of secreted EspA and EspD in the espC mutant , while the levels of EspB were similar to those observed in WT EPEC ( Fig 1B ) . Complementation of the espC mutant with pEspC resulted in a reduction of EspA and EspD levels in the culture supernatant ( Fig 1B ) . In contrast , the levels of EspB were not affected by espC expression . Similar results were obtained in Western blot analysis of bacterial cell fractions , showing that the levels of EspA and EspD were enhanced in the espC mutant and reduced in the ΔespC / pEspC+ complemented strain relative to WT ( S1A Fig ) . In contrast to supernatants , however , the levels of EspB were also affected by EspC , consistent with association of EspB with EspA / EspD [14 , 15] . As expected , proteinase K treatment indicated that EspA , EspB and EspD were secreted ( S1B Fig ) . To test whether EspC directly regulated the levels of EspA and EspD through proteolytic degradation , supernatants from ΔespC mutant were incubated with recombinant EspC . Samples were analyzed by SDS-PAGE followed by Coomassie staining ( Fig 1C ) or by Western-blot ( Fig 1D and 1E ) . EspA and EspD were degraded in an EspC dose-dependent manner ( Fig 1C and 1D ) . In contrast , and consistent with results observed in the supernatant of the espC mutant , the EspB levels were only affected at high concentrations of EspC ( Fig 1D and 1E ) . As shown in Fig 1F and 1G , proteolytic degradation of EspA and EspD was not observed in the presence of the serine protease inhibitor PMSF , or when EspA and EspD were incubated with EspC-S256I , a catalytic site mutant of EspC [16] . Together , these data indicate that EspC promotes the proteolysis of EspA and EspD . To characterize the activity of EspC on EspA and EspD structures , secreted proteins were fractionated by chromatography ( Experimental Procedures ) . When subjected to anion-exchange chromatography , EspA eluted as two distinct peaks at 175 mM ( peak A/D ) and 290 mM NaCl ( peak A ) ( Fig 2A and 2B ) . EspB did not associate with the anion exchange column under the conditions used . Interestingly , EspD co-eluted with EspA in peak A/D . Coomassie staining following SDS-PAGE indicated that EspA and EspD represented the main protein species in peak A/D , while EspA was predominant in peak A ( Fig 2A and 2C ) . We then tested the sensitivity of proteins contained in peaks A/D and A to EspC . To this end , EspC was purified and used at a concentration of 40 nM , which led to significant proteolysis of EspA and EspD , and which was comparable to estimated concentrations of EspC in bacterial culture supernatants ( Experimental Procedures; S2A and S2B Fig ) . As shown in Fig 2D , EspC induced the proteolysis of EspA , with the appearance of degradation products detected by Western Blot analysis ( Fig 2D arrows ) . While these degradation products were observed for both A/D and A fractions , proteolysis of EspA occurred steadily as a function of the incubation period with EspC in the A/D fraction ( Fig 2D ) . In contrast , in the A fraction , EspA appeared less sensitive to EspC proteolysis during the first hour of incubation , and only showed degradation after longer incubation periods ( Fig 2D ) . When quantified in independent experiments , EspA in the A/D fraction showed a rapid and linear decrease , while EspA decrease in the A fraction was observed at similar rate in the A fraction , but only after prolonged incubation ( Fig 2E ) . In the A/D fraction , as for EspA , EspD was also efficiently proteolyzed by purified EspC ( Fig 2D and 2E ) . Kinetic data analysis indicated that proteolysis of EspD in fraction A/D and EspA in fraction A occurred at a similar rate constant , but that proteolysis of EspA in fraction A/D occurred with a rate constant that was 1 . 8-fold higher ( S3 Fig ) . Taken together , these data show that EspC preferentially targets EspA in EspA / EspD-containing fraction associated with punctiform structures . In contrast , EspD in fraction A/D and EspA in the fraction containing EspA alone was less efficiently proteolyzed by EspC . These results suggest that EspC preferentially proteolyzes EspA in complex with EspD . More than 20 different SPATEs have been described in pathogenic Enterobacteriaceae [17] . Phylogenetic analysis of the SPATEs revealed that the closest SPATEs from EspC are the uropathogenic E . coli Sat , the Shigella flexneri SepA and the EHEC EspP [5] . Sequence alignments between EspC and these related SPATEs indicated identity ranging from 29 to 49% , and from 44 to 64% for the protease and passenger domains , respectively ( S4A Fig ) . To test proteolytic specificity of SPATEs towards EPEC translocon components , EspA , B and D , recovered from the supernatant from DMEM-cultured ΔespC mutant were incubated with the SPATEs EspC , Sat , SepA or EspP , secreted from recombinant DH5-α ( Experimental Procedures ) . As shown in S4B Fig , EspP , but not Sat or SepA , induced the proteolysis of EspA and EspD from EPEC . To investigate the role of EspP in the regulation of the levels of EHEC translocator components , an espP deletion mutant isogenic to the WT 85–170 EHEC strain was generated and secretion assays were conducted . Consistent with the results observed for EspC from EPEC , the espP mutant strain showed increased levels of EspA and EspD in the culture supernatant compared to WT EHEC , while EspB levels were not affected by espP expression ( S4C Fig . Also , EHEC EspA and EspD were degraded upon incubation with either EspP or EspC secreted from recombinant DH5-α S4C Fig . These results indicate that EspP and EspC are interchangeable and have a proteolytic activity towards T3SS translocator components . The absence of proteolytic activity of other SPATEs towards translocator components suggest that this activity of EspC / EspP is specific for EPEC / EHEC . Priming in DMEM leads to the combined induction of EPEC genes , the T3SA assembly , as well as to the activation of secretion of translocon components and type III effectors in the medium . Initial events of T3S induction are expected to involve translocator components in association with the T3SA needle at the bacterial surface , either as part of a tip complex or following induction of secretion . In other well-studied systems , such as the Shigella T3SS , one hydrophobic translocon component associates with the hydrophilic component to form a tip complex that negatively regulates T3S in the absence of host cells [18] . Following cell contact-dependent induction of secretion , the membrane-inserted translocon is thought to remain connected to the T3SS-needle through interaction between these hydrophobic and hydrophilic components . Although such tip complex has not been visualized for EPEC , the ortholog prediction indicates that the EPEC tip complex consist of EspA and EspD that we identified as the main EspC substrates . To determine if EspC regulates EspA/D in association with the T3SS , bacterial strains were primed for various time periods in cell culture medium , and protein levels were analyzed by immunofluorescent staining of bacteria immobilized onto coverslips . To unambiguously analyse structures on individual bacteria , we constructed and analyzed bfp mutant strains that do not form bacterial clusters ( Experimental Procedures ) . After 30 min priming , EspA punctiform structures were occasionally detected on the bacterial surface of WT and ΔespC mutant strains ( Fig 3A ) . After priming for 5 hours , EspA filaments could be readily detected at the surface of bacterial strains , suggesting that filaments evolved from punctiform structures . No difference , however was observed between WT and the ΔespC mutant strains , with 52 ± 1 . 4 and 51 . 5 ± 1 . 6% of bacteria associated with EspA filaments , respectively , with no striking difference in length and number of EspA filaments ( Fig 3A and 3B ) . As expected , EspA labelling was not observed for an isogenic T3SS-deficient mutant ΔescN ( Fig 3A and 3B ) . Western-blot analysis did not reveal EspD expression after 30 min incubation , and while EspD was detected after 5 hours incubation , no significant difference in the levels of EspA , EspB or EspD could be observed in supernatants ( Fig 3C and 3D ) or in bacterial pellets ( Fig 3E ) of WT and ΔespC mutant strain . Altogether , these results indicate that EspC proteolytic activity does not regulate EspA/D structures at the bacterial surface during the 5 hours-priming in cell culture medium . Rather , EspC-mediated proteolysis appears to target EspA/D released after prolonged incubation in cell culture medium , which induces the release of translocator components normally occurring upon cell contact . While priming in culture medium triggers the assembly of the EPEC T3SA and the up-regulation of T3S substrates , it does not recapitulate the sequence of regulatory events occurring upon cell contact . For instance , cell contact has been reported to stimulate further up-regulation of EspD and EspA as well as their secretion [19 , 20] . To analyze the role of EspC in the regulation of EspD and EspA levels upon cell contact , immunofluorescence staining was performed on HeLa cells challenged for 45 min with primed EPEC strains ( Experimental procedures ) . We observed EspA staining in association with all EPEC micro-colonies , bound or not to epithelial cells ( Fig 4A ) . As described previously , EspD staining was not observed for all bacteria but whenever detected , was visualized as punctiform structures on the bacterial surface ( S5B Fig; [20 , 21 , 22 , 23] ) . Interestingly , EspD staining was not observed for micro-colonies that were not associated with cells , and was observed at 45 min , but not at 15 min post-infection , consistent with cell contact triggering EspD secretion ( S5A Fig ) . In WT EPEC , these EspD punctiform structures rarely co-localized with EspA ( Fig 4A ) . Strikingly , the isogenic espC mutant showed a massive increase in EspD staining ( Fig 4B ) , with virtually all micro-colonies showing EspD staining ( Fig 4A and 4B ) . EspA also showed increased staining in the ΔespC mutant ( Fig 4A and 4B ) , and structures showing EspD and EspA co-staining were readily observed ( Fig 4A ) . As expected , such staining was not observed in cells infected with the T3SS-deficient ΔescN strain ( Fig 4A ) . To confirm these results , Western blot analysis was performed on lysates of HeLa cells infected with bacteria ( Experimental Procedures ) . Consistent with EspD fluorescence staining , cells challenged with the ΔespC strain displayed higher amounts of secreted EspD compared to cells challenged with the WT strain ( Fig 4C ) . As expected , EspD was not detected in lysates of cells infected with the T3SS-deficient ΔescN mutant strain ( Fig 4C ) . When detergent solubilization procedures used to isolate translocon components associated with host cell membranes were used , only a minute fraction of secreted EspD was detected ( S5C Fig ) . In these fractions , samples challenged with the ΔespC strain also showed higher EspD levels than those challenged with WT EPEC ( S5C Fig ) . When cell challenge was performed with WT EPEC in presence of the serine protease inhibitor PMSF , or with the ΔespC mutant complemented with the protease deficient EspC-S256I , the levels of secreted EspD were comparable to those observed for the ΔespC mutant , consistent with the down-regulation of translocator components by EspC proteolysis ( S6A and S6B Fig ) . As shown in S7A and S7B Fig , when mixed infection experiments were performed using WT/GFP and ΔespC strains at a 1:1 ratio , heterogenous microcolonies containing variable relative amounts of these bacteria could be detected in association with cells . When quantification was performed by scoring the percentage of microcolonies associated with EspD staining , irrelevant of the EspD staining intensity , the percentage of EspD-positive microcolonies in mixed infections was similar to that observed with the ΔespC strain , suggesting that EspC complementation was inefficient . Consistently , immunofluorescence analysis revealed that within these mixed microcolonies , EspD staining was predominantly associated with ΔespC bacteria while virtually not detected with WT bacteria . These observations suggested that EspC had a more efficient cis-proteolytic activity on EspD and a poorly diffusible trans-complementation activity ( S7A and S7B Fig ) . These results indicate that EspC negatively regulates the levels of extracellular EspA and EspD released upon host cell contact . We next investigated the role of EspC in T3SS-mediated cellular responses during bacterial infection . As shown in S8A Fig , the ΔespC mutant strain did not show detectable alteration in its ability to polymerize actin at the sites of bacterial cell contact compared to WT . Also , translocation of the Tir effector was not altered in the ΔespC mutant , consistent with the absence of a role for EspC in the formation of A/E lesions ( [24]; S8A , S8B and S8C Fig ) . In addition , dispersion of micro-colonies adhering to cells was observed in the ΔespC mutant to a similar extent to that observed for WT bacteria ( S8D Fig ) . These observations suggested that EspC-mediated proteolysis of translocator components did not negatively regulate type III effector translocation into host cells or EPEC-mediated actin pedestal formation . To analyze if EspC regulates T3SS-mediated pore formation during cell infection , we monitored the incorporation of the Lucifer Yellow ( LY ) fluorescent dye in cells challenged by bacteria [25 , 26] ( Experimental Procedures , Fig 5A ) . As shown in Fig 5B and 5C , WT EPEC induced dye uptake in HeLa and Caco-2/TC7 cells , with 26 . 3 ± 5 . 7% and 20 . 9 ± 2% of cells showing LY incorporation , respectively . In contrast , no LY incorporation could be detected when cells were challenged with the T3SS-deficient mutant ΔescN , indicating that dye uptake was dependent on the T3SS activity . When cells were challenged with the ΔespC strain , however , dye uptake increased significantly , with 43 . 6 ± 6% and 43 . 5 ± 3 . 8% of fluorescent cells in HeLa or Caco-2 / TC7 cells , respectively . To rule out the possibility that increased uptake of LY mediated by ΔespC mutant strain was due to fluid-phase uptake , HeLa cells were first loaded with calcein . The decrease of calcein fluorescence mediated by pore formation in the cell plasma membrane and dye leakage was monitored following bacterial challenge for 45 min ( Experimental Procedures ) . As shown in Fig 5D and 5E , the WT strain induced a 9 . 7 ± 0 . 9% decrease in calcein fluorescence relative to the T3SS-deficient mutant ΔescN . Consistent with the LY loading experiments indicative of higher pore forming activity , however , dye leakage induced by the ΔespC strain was significantly more pronounced relative to WT , with a 29 . 6 ± 1 . 4% decrease in calcein fluorescence ( Fig 5D and 5E ) . The ΔespC / pEspC+ strain induced a modest 3 . 7 ± 1 . 3% decrease of dye leakage , indicative of complementation . These results are consistent with EspC negatively regulating the formation of T3SS-dependent pores in the plasma membranes of cells during bacterial infection . Host cell death can result from membrane injury induced by pore forming toxins [27 , 28] . Since our evidence indicated that EspC regulates the T3SS-induced pore formation in host cell membranes , we investigated whether increased pore formation associated with the espC mutant correlated with an increase in bacterial-induced cell death . Cells were challenged with bacteria and further incubated for 17h in the presence of gentamicin to prevent bacterial growth in the extracellular medium ( Experimental Procedures ) . Under these conditions , bacteria-induced cytotoxicity was observed for WT EPEC , with 47 . 1 ± 7 . 3% of remaining adherent cells relative to uninfected samples ( Fig 6A and 6B ) . Cytotoxicity was specific for the T3SS since it was not observed for the escN mutant ( Fig 6A and 6B ) . Strikingly , the espC mutant was more cytotoxic than WT EPEC , with only 27 ± 7 . 7% of surviving cells recovered ( Fig 6A and 6B ) . Also , nuclei from espC-infected cells appeared more condensed than nuclei from uninfected , ΔescN or WT-infected cells ( Fig 6A ) . Consistently , cells infected by the ΔespC mutant strain displayed an average reduction in nuclei area of 26% compared to uninfected cells ( Experimental Procedures; Fig 6C and 6D ) . Interestingly , no DNA fragmentation was observed associated with this phenotype , suggesting that cytotoxicity was not related to caspase-dependent apoptosis . Accordingly , although bacterial infection induced detectable levels of caspase 3 cleavage that were inhibited by zVAD , this caspase inhibitor did not protect from the ΔespC-induced cytotoxicity ( Fig 6E and S9 Fig ) . Numerous eukaryotic targets with unrelated functions have been described for SPATEs , probably reflecting the diversity of adaptation requirements of Enterobacteriaceae to their niche [4] . Here , we describe a novel function for EspC , as a regulator of pore formation induced by the EPEC T3SS . We show that in vitro and during cell challenge , EspC preferentially targets EspA/EspD containing structures , and regulates the T3SS-dependent pore formation during cell challenge . EspA was shown to interact with EspD in vitro [29] . By analogy with findings in other T3SS , EspA association with EspD may occur in a complex involved in host cell membrane sensing at the tip of the T3SA , which in the case of EPEC would correspond to the tip of EspA filaments . The association between EspD and EspA is also believed to establish the physical connection between the EspB/D translocon inserted into host cell membranes and the bacterial T3SA . Our experiments did not reveal a role for EspC in the regulation of EspA/D structures at the surface of primed bacteria . Indeed , after 5-hour priming , corresponding to the priming conditions for cell infection , the ΔespC mutant and WT showed a similar percentage of bacteria associated with EspA filaments , with no detectable differences in EspA filament number or length . Also , similar levels of EspA and EspD were observed in the ΔespC mutant and WT strains . We could also detect EspA punctiform structures on the surface of bacteria primed for 30 min that are reminiscent of those observed by others [30 , 31] . Our time course priming studies suggest that these EspA punctiform structures were precursors of EspA filaments . In other studies , an espD mutant was shown to form EspA punctiform structures but failed to form EspA filaments , indicating that EspD is required for transition from EspA-containing punctiform structures to filaments [21] . The mechanism implicating EspD in this EspA-filament transition , however , remains unclear . Our preliminary results suggest that the percentage of bacteria associated with EspA punctiform structures is higher in the ΔespC mutant after priming for 30 min . However , it is unlikely that EspA punctiform structures are regulated through the proteolytic activity of EspC on EspA/D that we report here , since we could not detect EspD at early priming time points . While we could not detect EspD at the bacterial surface in the absence of contact with host cells , immunostaining experiments identified EspC-sensitive EspA/D structures that were released in the supernatant of bacteria upon prolonged priming conditions . These observations suggest that if an EPEC EspA/D tip complex exists , its association with EspA filaments at the bacterial surface is too transient to allow its detection , possibly because growth in priming conditions also triggers the secretion of translocon components [32] . Alternatively , EspD may be in a configuration that is not recognized by our antibody when part of a tip complex at the bacterial surface . Our results are consistent with EspC targeting EspA/D structures corresponding to translocator components normally released following cell contact . Mixed infection experiments indicated that within the same microcolony , EspC controls the levels of EspD associated with EspC-expressing bacteria , but not those of bacteria deficient for EspC . These results indicate that during cell infection , the action of EspC on translocator components is poorly diffusible and most efficient at the close contact of EspC-secreting bacteria . Along the same lines , the activity of EspC on EspA and EspD secreted upon cell contact appears to be more efficient than that characterized on soluble EspA and EspD released from bacteria grown in vitro . Indeed , we observed a drastic increase in EspA/D levels in the ΔespC mutant compared to WT after 45 min following cell challenge . The T3SS-dependent pore forming activity was also increased in the ΔespC mutant , suggesting that EspC-mediated proteolysis of EspA/D translocator components negatively regulate the formation of translocons inserted in the host plasma membrane . Contact with host cell membranes and insertion of the T3S translocon is expected to involve a change of configuration of EspA and EspD . Following translocation , EspD is exposed on the external surface of the plasma membrane as shown by its sensitivity to degradation by proteases [20 , 33] . This indicates that under their membrane-inserted configuration , translocon components could also serve as targets for EspC . The timely control of EspC-mediated degradation of translocator components suggests that EspC interacts with T3SA components prior to cell contact . Interestingly , while EspC is not a T3SS substrate , it is recovered with a high SILAC ratio in the T3SS secretome [34 , 35] . Furthermore , in vitro evidence indicates that EspC interacts with EspA [13] . Such interactions may be relevant for the control of pore formation mediated by the T3SS described in our studies . In turn , these interactions could also permit a cross-regulation of EspC activity . Indeed , EspC was reported to induce the cleavage of fodrin [7 , 12 , 36 , 37] . However , this intracellular fodrin cleavage activity requires the endocytosis of EspC , a process that is inefficient in the absence of a functional T3SS [37] . Once translocated , EspC may cleave fodrin , an activity linked to EPEC-induced cell detachment and cytotoxicity [7] . Promoting epithelial cell death is a dead-end for a bacterial enteropathogen , which may serves a purpose at late stages of infection and tissue colonization into a given host . The release of bacteria-containing killed cells in the intestinal lumen has been proposed to promote bacterial shedding in the environment to allow bacterial dissemination to other hosts [38] . During the initial stages of infection , along with EPEC injected type-3 effectors that down-regulate inflammation , the control of pore formation and cytotoxicity by EspC likely favours bacterial colonization of the epithelium . Consistent with a role for EspC in the negative regulation of cytotoxicity that we report here , a mutant defective for an EspC ortholog in Citrobacter rhodentium shows increased virulence and damaging potential observed during in vivo infection of mice [39] . The formation of a pore mediated by the insertion of the T3SS translocon in host cell plasma membranes is thought to be a requisite for the injection of type III effectors , and therefore probably represents a conserved feature for all T3SS . While pore-formation can be assessed in red blood cells , it is not observed in epithelial cells that are proficient for membrane-damage repair . Perhaps emphasizing the importance of their regulation , translocon-associated pores are controlled by injected type III effectors [40 , 41] . Such regulation may not be critical for invasive bacteria for which active translocons are expected to be removed from the plasma membrane during their inclusion in the phagocytic vacuole , but T3SS-mediated pore-formation likely represents an acute issue for extracellular pathogens such as EPEC / EHEC growing at the surface of epithelial cells . Of interest , among SPATEs expressed by T3SS-proficient enterobacteriaceae , EspP from EHEC is phylogenetically the most closely related to EspC and shares a similar protease activity towards translocon components , consistent with an adaptation of these SPATEs to extracellular growth on epithelial cells . The targeting of bacterial virulence factors by SPATEs is an emerging concept [42] . In addition to the cleavage of T3SS translocator components described in this study , EHEC EspP was reported to inactivate hemolysin ( Hly ) to regulate Hly-induced pore formation [43] . Obviously , a critical feature of SPATE-mediated proteolysis of pore-forming proteins is the timely control of this activity , so as not to inhibit pore formation . The study of the regulation of pore-forming activity linked to the EspA/D translocator components by EspC will likely provide insights into important steps occurring during host cell membrane's recognition by T3SSs . Bacterial strains and plasmid constructs used in this study are listed in S1 Table . Bacterial culture conditions are described in S1 File . Human epithelial HeLa cells were grown at 37°C in a 5% CO2 incubator in RPMI ( Life technology ) containing Glutamax and supplemented with 10% heat-inactivated fetal calf serum ( FCS ) ( Life Technology ) . TC7 cells ( Caco-2/TC7 ) [44] were grown in a 10% CO2 incubator in DMEM containing 4 . 5 g/L glucose , supplemented with 15% heat-inactivated FCS and 1% non-essential amino acids ( Life technology ) . Deletions of the entire bfpA and espC genes in EPEC or espP gene in EHEC were created by allelic exchange using the λ Red recombinase method [45] ( S1 File ) . A non-proteolytic form of EspC ( EspC-S256I ) was generated by site directed mutagenesis on pJLM174 [35] using the Quick-change method ( Stratagen , Netherlands ) and the following oligonucleotides 5’- CTA CCG GTG GAG ACA TTG GTT CCG GTT TCT ATC-3’ and 5’- GAT AGA AAC CGG AAC CAA TGT CTC CAC CGG TAG-3’ . The construct was verified by DNA sequencing . HeLa cells were plated on glass cover slips the day before the experiment at a density of 4 x 105 cells for cytotoxicity assays , or 5 x 104 cells for all other assays . TC7 cells were plated at a density of 7 x 104 cells and allowed to polarize and differentiate for 14 days before bacterial challenge . Bacterial cultures were primed for 5 h in DMEM , centrifuged and resuspended in fresh DMEM to remove secreted proteins . Bacteria were used at a final OD600nm of 0 . 8 for dye loading or translocation assays , or 0 . 05 for immunofluorescence analysis . For dye loading experiments , samples were centrifuged for 3 min at 2000 g using a plate-holder in a swinging bucket centrifuge 5810 ( Eppendorf ) prior to incubation at 37°C to synchronize the infection . Immunostaining was performed as described in the Supplemental information . Samples were mounted in Dako mounting medium ( DAKO ) and analysed using an Eclipse Ti microscope ( Nikon ) equipped with a 100 x objective , a CSU-X1 spinning disk confocal head ( Yokogawa ) , and a Coolsnap HQ2 camera ( Roper Scientific Instruments ) , controlled by the Metamorph 7 . 7 software . Analysis by epifluorescence microscopy was performed using a DMRIBe microscope ( LEICA microsystems ) using 380 nm , 470 nm , or 546 nm LED source excitation , equipped with a Cascade 512 camera ( Roper Scientific ) driven by the Metamorph 7 . 7 software . Images were analyzed using the Metamorph software . The levels of EspA and EspD secreted at bacterial-cell contact sites were quantified by delimiting an area corresponding to the bacterial microcolony according to DAPI staining . The integrated fluorescence intensity of relative EspA/EspD staining in the corresponding area was measured and expressed as a ratio to that obtained for DAPI staining . Bacteria were grown for 5h or 16 h in DMEM . The equivalent of 10 mls of bacterial culture were centrifuged at 1500 g for 15 min at 4°C . Supernatants were filtered-sterilized using a 0 . 22 μpore-sized filter , and proteins were subjected to precipitation using trichloroacetic acid ( TCA ) at a 5% final concentration and analyzed by SDS-PAGE followed by Coomassie blue staining and / or by Western Blot using the ECL-Plus detection system ( Western Lighting Plus-ECL , Perkin-Elmer ) . Protein loading was normalized using anti-OmpA Western blotting . Quantification of the protein band integrated density was performed using the ImageJ software and expressed as the average of at least 3 independent experiments . To purify EspC , strain DH5-α ( pJLM174 ) was grown for 16 h in LB medium plus arabinose ( 0 . 2% , wt/vol ) and ampicillin ( 100 μg/ml ) at 37°C with shaking . Following centrifugation at 6 , 000 x g for 15 min to remove bacteria , supernatants were filtered through 0 . 22-μm-pore-size filters ( Stericup , Millipore ) . Proteins from sterilized supernatant were precipitated using 40% ( wt/vol ) ammonium sulfate at 4°C . Pellets were resuspended in 25 mM Tris-HCl ( pH 7 . 4 ) , 25 mM NaCl , and 1 mM β-mercaptoethanol ( βME ) in a volume equal to 1/50 of the supernatant and dialyzed 3 times against 100 volumes of the same buffer . EspC was purified by FPLC using an anion exchange column ( Mono Q , GE Healthcare ) using a 25 mM–1000 mM NaCl 20 mls—linear gradient . Fractions were analyzed by SDS-PAGE and Coomassie staining . EspC eluted as a single peak at 265 mM NaCl . EspC-containing fraction were pooled , dialyzed against 25 mM Tris-HCl ( pH 7 . 4 ) , 50 mM NaCl , 1 mM βME . Protein concentration was estimated using a spectrophotometer ( Nanodrop , ND-1000 ) . In typical experiments , about 3 mg of purified EspC could be obtained per millilitre of bacterial culture . EspA and EspD were purified from the supernatant of a DMEM-cultured ΔespC strain , using anion exchange chromatography in a procedure similar to that of EspC , except that proteins were resuspended in 50 mM Tris-HCl ( pH 7 . 4 ) , 50 mM NaCl , 1 mM βME in a volume equal to 1/200 of the supernatant and dialyzed against same buffer . Under these conditions , EspB did not bind to anion exchange column . EspA and EspD co-eluted at 174 mM NaCl ( peak A/D ) , another EspA fraction free of EspD was eluted at 290 mM NaCl ( peak A ) . Proteins were dialysed against 15 mM Tris-HCl ( pH 7 . 4 ) , 115 mM NaCl and 1 mM βME . The concentration of the various protein species was estimated by densitometry following SDS-PAGE and Coomassie blue staining , using purified BSA to perform standard curves . The EspC , EspC-S256I , Sat , SepA and EspP were expressed and recovered in the supernatant of DH5-α recombinant strains as the major protein species detected by Coomassie staining . For each SPATE , the protein concentration was adjusted to 25 nM . The SPATE-containing supernatants were incubated with the ΔespC strain supernatant prepared as described above , for 16 hours at 37°C . When specified , PMSF was added during the incubation procedure at a final concentration of 2 mM . Samples were TCA-precipitated and analyzed by Western Blot . For controlled proteolysis , purified EspA , EspD and EspC were used at a final concentration of 65 nM 250 nM and 40 nM , respectively , in 15 mM Tris-HCl ( pH 7 . 4 ) , 115 mM NaCl , 2 mM MgCl2 and 1 mM βME . Reactions were carried out at 37°C for different time periods ( 30 , 60 , 120 , 240 , 360 min or 16 hours ) , and stopped by the addition of Laemmli sample loading buffer and boiling for 10 min . Samples were analysed by Western Blot . Cells grown onto glass coverslips were challenged with primed EPEC strains . Samples were incubated in DMEM containing 1 mM of Lucifer yellow in a wet chamber at 37°C for 15 or 45 min . In control experiments , samples were incubated with 1 mM 70 kDa dextran coupled to rhodamine . Samples were washed with PBS and fixed in 3 . 7% PFA . Samples were labelled with DAPI and analyzed using an inverted epifluorescence microscope . Images corresponding to LY fluorescence were thresholded to generate binary images . LY positive cells were scored from binary images and results were expressed as the average percentage ± SEM of LY positives cells relative to total cells . Different samples within the same experiment were treated in the same conditions during image acquisition and analysis . Cells grown on coverslips were loaded with 3 μM calcein-AM ( Life Technology ) for 30 min in EM buffer ( HEPES 25 mM pH 7 , 3; NaCl 120 mM; CaCl2 1 . 8 mM; MgCl2 0 . 8 mM; KCl 7 mM; glucose 5 mM ) , at 21°C . After three washes in EM buffer , cells were incubated for another 30 min at 21°C . Cells were challenged with primed EPEC strains for 45 min . Samples were mounted in an observation chamber on a plate heated at 37 °C on an inverted epifluorescence microscope . For each experiment , image acquisition was performed using identical conditions for each sample . The percentage of calcein leakage was calculated as 100 x ( 1 - I / IΔescN ) , with I corresponding to the averaged integrated fluorescence intensities of individual cells for the sample , and IΔescN corresponding to that of cells infected with the T3SS-deficient ΔescN mutant strain . Bacterial strains were primed in DMEM for 5 h , and HeLa cells were infected with primed bacteria for 45 min at 37°C . The medium was removed and samples were further incubated for one hour in fresh medium . Gentamicin was added to kill bacteria and samples were incubated for 17h . Samples were fixed with 3 . 7% PFA and analyzed by transmission light microscopy following hematoxylin / eosin staining , or epifluorescence microscopy for DAPI staining . Alternatively , samples were washed 3 times with PBS , trypsinized and resuspended cells were scored using a Malassez cell counting chamber . To analyze nuclear shrinkage , the nuclei area was delimited by thresholding images acquired from DAPI stained samples to generate binary images . The nuclei size was analyzed using the Metamorph software . Samples within the same experiment were treated under the same conditions during image acquisition and analysis . When specified , the caspase inhibitor zVAD was added at a 50 μM final concentration for 30 minutes prior to bacterial challenge and maintained throughout the experiment . Staurosporine was used at 1 μM as a positive control for caspase-activation . Statistical analysis was performed using an unpaired Student’s t-test with unequal variance .
Enteropathogenic Escherichia coli ( EPEC ) is an important diarrheal pathogen responsible for infant diarrhoea associated with significant morbidity and mortality rates in developing countries . Upon ingestion EPEC colonizes the intestinal mucosa , causing characteristic lesions on enterocytes . Using a type III secretion system ( T3SS ) acting as a molecular syringe , EPEC injects numerous bacterial proteins into host cells that disrupt the intestinal epithelium homeostasis . Injection of T3SS proteins requires the insertion into the host cell plasma membrane of bacterial protein complex , called the "translocon" , associated with pore-forming activity . In addition to the T3SS , EPEC also secretes other bacterial toxins involved in virulence . Among these , the EspC is a protease reported to degrade various host proteins . In this paper , we have characterized an "unsuspected role" for EspC . We show that EspC degrades the T3SS translocon components following cell contact and regulates T3SS-dependent pore formation in epithelial cells . The EspC control of pore formation limits cytotoxicity and thus , is expected to limit the emission of danger signals , which would otherwise favour bacterial clearance at the onset of infection . This work describes a novel regulatory mechanism of pore formation mediated by the T3SS , that are likely to be relevant for other extracellular pathogens .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[]
2015
The Serine Protease EspC from Enteropathogenic Escherichia coli Regulates Pore Formation and Cytotoxicity Mediated by the Type III Secretion System
In the United States , the introduction of the heptavalent pneumococcal conjugate vaccine ( PCV ) largely eliminated vaccine serotypes ( VT ) ; non-vaccine serotypes ( NVT ) subsequently increased in carriage and disease . Vaccination also disrupts the composition of the pneumococcal pangenome , which includes mobile genetic elements and polymorphic non-capsular antigens important for virulence , transmission , and pneumococcal ecology . Antigenic proteins are of interest for future vaccines; yet , little is known about how the they are affected by PCV use . To investigate the evolutionary impact of vaccination , we assessed recombination , evolution , and pathogen demographic history of 937 pneumococci collected from 1998–2012 among Navajo and White Mountain Apache Native American communities . We analyzed changes in the pneumococcal pangenome , focusing on metabolic loci and 19 polymorphic protein antigens . We found the impact of PCV on the pneumococcal population could be observed in reduced diversity , a smaller pangenome , and changing frequencies of accessory clusters of orthologous groups ( COGs ) . Post-PCV7 , diversity rebounded through clonal expansion of NVT lineages and inferred in-migration of two previously unobserved lineages . Accessory COGs frequencies trended toward pre-PCV7 values with increasing time since vaccine introduction . Contemporary frequencies of protein antigen variants are better predicted by pre-PCV7 values ( 1998–2000 ) than the preceding period ( 2006–2008 ) , suggesting balancing selection may have acted in maintaining variant frequencies in this population . Overall , we present the largest genomic analysis of pneumococcal carriage in the United States to date , which includes a snapshot of a true vaccine-naïve community prior to the introduction of PCV7 . These data improve our understanding of pneumococcal evolution and emphasize the need to consider pangenome composition when inferring the impact of vaccination and developing future protein-based pneumococcal vaccines . Pneumococcal conjugate vaccines ( PCV ) target capsular serotype-specific polysaccharides of the respiratory pathogen Streptococcus pneumoniae , which causes substantial morbidity and mortality [1 , 2] . Since the heptavalent PCV and 13-valent PCV were introduced in the United States ( US ) in 2000 and 2010 , respectively , their effectiveness in reducing pneumococcal carriage and invasive disease has been well documented [3–6] . In communities where PCV has been introduced , the prevalence of vaccine serotypes ( VT ) in carriage and invasive disease consistently decreases , resulting in an overall reduction in pneumococcal disease . However , in a process called “serotype replacement , ” non-vaccine serotypes ( NVT ) subsequently increase in carriage after vaccine introduction , leading to slight increases in NVT-associated disease in almost all populations where the vaccine is introduced [7 , 8] . Because polysaccharide serotypes change rarely during pneumococcal evolution , common pneumococcal lineages typically contain only one or a few serotypes . As a result , PCV implementation removes lineages containing only VT from the population , while lineages including both VT and NVT experience genetic bottlenecks [9–11] . Forecasting which serotypes , and more generally which pneumococcal lineages , will increase in frequency in carriage and disease is an active area of research with significant public health importance . For S . pneumoniae , the most commonly used vaccines globally target a fraction of the more than 93 recognized capsular serotypes [12] . The bacteria’s capsule ( CPS ) is the most important determinant of virulence and the strongest predictor of prevalence [13] , as well as the target of PCVs; thus , changes in CPS serotype frequency have been the focus of many analyses of vaccine effect . However , selection acts on genes outside the operon determining CPS serotype . Whole-genome sequencing data has enabled investigation of variation in multiple genomic loci and genome content among pneumococci , focused on loci involved in host immunity and niche adaption . We focus here on two categories of proteins . The first is antigens ( hereafter , when we use the generic term antigen we refer to proteins that elicit an immune response , not to the polysaccharide capsule ) . Antigens such as pneumococcal surface proteins A and C ( pspA and pspC ) and pilus are of specific interest as possible targets for non-capsular polysaccharide based vaccines [14] . Together with other components of the pneumococcal genome , the capsule and non-capsular antigens comprise the overall antigenic profile of a pneumococcus [15–17] . Moreover , evolution among metabolic genes gives rise to distinct metabolic-profiles among pneumococcal lineages , which may be adapted for specific metabolic niches [18 , 19] . Thus , multiple loci may interface with the host , affecting the overall evolutionary success of a lineage at a population level . Gene content varies tremendously among pneumococcal lineages [20 , 21] . The pneumococcal pangenome consists of “core genes” shared by ≥99% of strains and “accessory genes” present at frequencies ≤99% . Accessory genes may include polymorphic antigens , phage and plasmid-related chromosomal islands , and integrative and conjugative elements ( ICE ) harboring antimicrobial resistance genes . The latter are mobile genetic elements ( MGE ) , which are often acquired through horizontal gene transfer ( HGT ) and may remain stable in pneumococcal lineages [21] . Variations in gene content among lineages of a bacterial species are associated with ecological niche specialization and are important for adaptation to changing environments , including selection by vaccine-induced and natural host immunity [21–23] . For the pneumococcus , MGE affect the bacteria’s ability to recombine ( i . e . , competence ) [24] , antimicrobial susceptibility [25] , and carriage duration [26] . Accessory loci may also be acted upon by negative frequency dependent selection ( NFDS ) , hinting at their underlying role in non-serotype-specific immunity and S . pneumoniae ecology [27] . Taken together , gene variation beyond the capsular polysaccharide loci may significantly impact virulence , fitness , transmission , and , in turn , the overall epidemiology and ecology of pneumococcal strains . Before PCV introduction , Navajo and White Mountain Apache ( N/WMA ) Native American communities in the Southwestern US experienced rates of invasive pneumococcal disease ( IPD ) 2–5 times higher than the general US population [28 , 29] . Pneumococcal carriage prevalence among N/WMA pre-PCV7 was 50% among all ages and 75% among children <2 years of age , significantly higher than the general population [30] . Thirty-eight percent of all pneumococcal carriage isolates were PCV7 serotypes [30] . After introduction of PCV7 , carriage prevalence of PCV7 VT declined , and the rate of IPD among N/WMA caused by VT decreased by 89% [31] . However , carriage prevalence of NVT strains increased , resulting in no overall change in pneumococcal carriage prevalence among children or adults [5] . Also , despite increased NVT carriage there was no corresponding increase in the rate of IPD caused by NVT . After introduction of PCV13 in 2010 , carriage of PCV13-specific serotypes declined by 60% among children <5 years of age within the first two years [6] . Yet , overall IPD rates among N/WMA still remain higher than those in the general US population [32] . Here , we analyze a sample of 937 pneumococci collected over 14 years and spanning before , during , and after the introduction of PCV7 and PCV13 vaccines among N/WMA . To understand the evolutionary impact of vaccination and characterize the shift from VT to NVT , we assessed the recombination , evolution , and pneumococcal population history , classified by serotype and by whole-genome sequencing data , across vaccine introduction periods . Furthermore , we investigated metabolic loci variation and pangenome composition over time , with a focus on pneumococcal antigens . This study included pneumococci isolated from a subset of participants of three prospective , observational cohort studies of pneumococcal carriage among N/WMA families described elsewhere ( hereafter , “parent” studies ) [1 , 6 , 30] . Briefly , participants living on reservations in the southwest USA were enrolled during three periods: 1998–2001 , 2006–2008 and 2010–2012 . Nasopharyngeal ( NP ) swab specimens were obtained during visits to Indian Health Services ( IHS ) facilities or the participants’ home to determine pneumococcal carriage status ( S1 Fig ) [30] . A random subsample of isolates was selected from each time period , with an oversampling of isolates post-PCV7 ( S2 Fig ) . A single isolate was chosen from each participant; however , previous pneumococcal carriage history was not assessed . With the exception of a subset of isolates collected from 2006–2008 , all isolates were obtained from children ≤5 years of age . Genomic DNA from S . pneumoniae isolates were sequenced on the Illumina HiSeq , yielding ≥30-fold coverage per isolate . Paired-end 100 bp reads were filtered by quality and length . Serotypes were determined by mapping reads to concatenated CPS locus sequences of 93 pneumococcal serotypes using SRST2 [12 , 33] . Serotypes for isolates identified as serogroup 6 were further resolved using PneumoCaT [34] . Multilocus sequence type ( MLST ) was determined through a similar approach using SRST2 . De novo genome assemblies were generated with Velvet [35] and annotated using Prokka v1 . 11 [36] . After annotation , the pangenome was analyzed with Roary , and a concatenated alignment of clusters of orthologous genes ( COGs ) shared among ≥99% of all strains ( i . e . , core genome ) was abstracted [37] . Pneumococcal population structure was assessed using core genome SNPs with hierBAPS , which was run three times using maximum clustering sizes of 20 , 40 , and 60 [38] . A maximum likelihood ( ML ) phylogeny was estimated using RAxML v8 . 1 . 5 with GTR+Γ nucleotide substitution model and 100 bootstrap replicates [39] . Sequence clusters ( SCs ) ( i . e . , lineages ) identified using hierBAPS were annotated on the core genome phylogeny . For the study period during which pediatric and adult isolates were collected ( 2006–2008 ) , the proportion of isolates by SC was compared between age groups to 10 , 000 random deviates of a Dirichlet distribution [40] . A subsample of isolates from each SC and 25 publicly available reference genomes were aligned using Parsnp and visualized using Gingr to identify the most appropriate genome for reference-based mapping [41] . The phylogenetically closest genome was selected for reference-based mapping of isolates belonging to that SC . For four out of 27 SCs , a monophyletic match was not available; therefore , we generated references by refining , ordering , and concatenating the best draft assembly in the SC . A second de novo assembly was generated with SPAdes and assemblies were then merged using Zorro [42] . After this , SSPACE and GAPFILLER were used to scaffold the assembly and remove Ns [43 , 44] . Final contigs were ordered using Mauve , manually curated using ACT , and concatenated [45] . Filtered Illumina reads from isolates comprising each SC were mapped to the selected reference using SMALT v0 . 7 . 6 and SNPs were identified using SAMtools v1 . 3 . 1 [46] . SNPs were filtered requiring a depth of coverage of five and a minimum alternate allele frequency of 0 . 75 . The output was analyzed as previously described to generate whole-genome multiple sequence alignments for each SC [9 , 47] . Next , we identified recombination among SCs using Gubbins [48] . Gubbins identifies SNPs introduced through recombination and allows censoring for downstream phylogenetic analysis . Results of Gubbins analyses were visualized using Phandango [49] . For SCs in which over 50% of the genome was censored due ancestral recombination events , we either sub-clustered SCs clearly delineated monophyletic clades ( e . g . , SC19 which was comprised of serotypes 15A and 17F ) or removed divergent isolates that were significantly affected by recombination . Sub-clustered SCs were annotated on the ML phylogeny and then reanalyzed with Gubbins . For comparison between vaccine periods , isolates were subdivided into three epochs and six sub-epochs by year of collection: pre-PCV7 sub-epochs 1A ( 1998 ) and 1B ( 1999–2001 ) ; post-PCV7 sub-epochs 2A ( 2006 ) and 2B ( 2007–2008 ) ; PCV13 sub-epochs 3A ( 2010 ) and 3B ( 2011–2012 ) ( S2 Fig ) . Collection years were grouped to balance sample sizes among sub-epochs . To determine the representativeness of the genomic sample to the parent studies from which the sample was drawn , we compared the serotype distribution and serotype diversity ( Simpson’s D ) of unique carriage isolates from the three parent studies of pneumococcal carriage [1 , 6 , 30] , by epoch , to that of the sample . Core genome alignments were generated for isolates in each sub-epoch using Roary , and population genomic statistics including Tajima’s D [50] , Watterson’s estimator ( Θw ) [51] , and nucleotide diversity were calculated for each period using 0-fold and 4-fold degenerate sites . The ratio of diversity at non-synonymous sites to synonymous sites ( πN/πS ) was also calculated as a measure of selection . The same statistics were calculated for each SC . Code for calculating population genetic statistics using Roary output is available at http://github . com/c2-d2/Projects/NWMA_Pneumo/ . ML phylogenies of SCs , inferred from recombination-censored alignments , were used to test temporal signal by assessing correlation between strain isolation date and root-to-tip distance . SCs with poor root-to-tip correlation were assessed for residual recombination and phylogenetic signal . SCs determined to have sufficient temporal signal were analyzed with BEAST v1 . 8 . 2 [52] . For each SC or sub-SC a combination of strict and relaxed molecular clock models and constant and Gaussian Markov random field ( SkyGrid ) demographic models [53] were tested using recombination-free SNP alignments , ascertainment bias correction [54 , 55] , and HKY nucleotide substitution model . For SCs in which the coefficient of variation for relaxed molecular clock models was high ( i . e . , significant rate heterogeneity across the tree ) , a random local clock ( RLC ) model was also tested [56] . Markov chain Monte Carlo lengths for each model run ranged from 150 million to 1 billion depending on the size of the SC and length of the SNP alignment . MCMC chains were sampled to obtain 10 , 000 trees and 10 , 000 parameter estimates in the posterior distribution . Effective sampling size ( ESS ) values were assessed to determine sufficient mixing using Tracer v1 . 6 . 0 , and runs with ESS values of 200 for all parameters were accepted . Marginal likelihood estimates ( MLE ) were obtained for each model using path-sampling and stepping-stone analysis , and models were compared using Bayes Factors [57 , 58] . Parameter estimates for the evolutionary rate , root height ( i . e . , TMRCA ) , and Ne were obtained from the best-fit model . For SCs in which SkyGrid demographic models were fit , the slope of the Ne change over time was calculated to determine directionality , and the 95% highest posterior density ( HPD ) was used to determine significance . To assess the impact of PCV7 on the pneumococcal pangenome we compared frequencies of polymorphisms in core genes and accessory genome COGs among sub-epochs , focusing on antigens and metabolic loci for the core genome and on antigens for the accessory genome analysis . We identified metabolic genes using coding sequences found in S . pneumoniae reference strain D39 ( RefSeq: NC_008533 . 1 ) that were annotated as “Metabolism” according to KEGG Orthology ( KO ) groupings of the KEGG database ( http://www . genome . jp/kegg/ ) and were assigned to a known metabolic pathway ( KEGG pathway spd01100 ) . Pangenome analysis using Roary was repeated including D39 , and COGs found in the core genome ( i . e . , present among all 937 taxa ) with ≥90 BLAST identity to metabolic genes were abstracted . A concatenated alignment of core metabolic COGs was then constructed , and biallelic SNP sites were identified . To assess changes to the accessory genome , we obtained the binary presence-absence matrix of accessory COGs present in frequencies ranging from 5–95% among all taxa . This frequency range was conservatively selected to mitigate the effect of genome assembly and annotation errors in COG identification . Last , we used a previously described method to identify the variants of 19 polymorphic antigens [15] . These antigens have measurable interactions with the host immune system , and therefore are thought to be under the greatest population level host immune pressure . Ten additional antigens were evaluated ( lysM , lytB/C , pcpA , pcsB , phtE , piaA , piuA , psaA , SP2027 , pce ) but were excluded because they were deemed nearly monomorphic due to their low nucleotide diversity . Using the concatenated nucleotide alignment of metabolic loci and a binary presence absence alignment accessory COGs and antigen variants , ML phylogenies were inferred using RAxML with GTRGAMMA ( nucleotide ) or GTRCAT ( binary ) substitution model and 100 bootstrap replicates . The cophenetic ( patristic ) distances of each phylogeny were read into R , and the meandist function in the package vegan was used to calculate within-group distances for three population groupings: serogroup , serotype , and SC . Within-group distances for population stratifications were then compared . For each set of genomic loci ( metabolic , accessory COGs , and antigens ) , frequencies were computed for each of the six sub-epochs . Mean squared errors ( MSEs ) were then calculated to assess changes in frequencies from Epoch 1A . This was done by subsampling 75 individuals with replacement from each sub-epoch and performing 1000 bootstrap replicates of each comparison ( e . g . , Epoch 1A vs . 1B , 1A vs . 2A , 1A vs . 2B , and so on ) . The significance of changes in antigen distributions among epochs was additionally tested by comparing the proportion of antigen variants between Epochs 1–3 to 10 , 000 random deviates of a Dirichlet distribution . The Navajo Nation , White Mountain Apache tribe and the IRBs of the Johns Hopkins Bloomberg School of Public Health , the Navajo Nation and the Phoenix Area IHS approved this study . During the original pneumococcal carriage studies from which these isolates were obtained , written informed consent was obtained from adult participants and from caregivers of child participants . Assent was obtained from children 7–17 years . Isolates were obtained from NP swabs , as previously described , and de-identified for analysis . We analyzed genomic data from a total of 937 pneumococcal carriage isolates collected from N/WMA Native Americans in Southwestern US between 1998 and 2012 . All isolates were obtained from children ≤5 years of age with the exception of 125 isolates ( 13 . 3% of total ) collected from individuals 6–76 years of age during 2006–2008 . Isolates collected from 1998–2001 ( n = 274 ) were obtained from communities that served as the control for cluster-randomized PCV7 trials and therefore represent a vaccine naïve population . Isolates collected during 2006–2008 ( n = 398 ) represent the post-PCV7 pneumococcal population , and isolates from 2010–2012 ( n = 265 ) were sampled during the implementation of PCV13 ( S2 Fig ) . Whole-genome sequencing data has been deposited in NCBI sequence read archive ( SRA ) under accession number ERP009399 , BioProject PRJEB8327 . Individual accession numbers are provided in supplementary file 1 . Pangenome analysis of de novo genome assemblies identified 8 , 674 COGs , of which 1 , 111 were present in ≥ 99% of strains ( i . e . , the core genome ) . Analysis of population structure using hierBAPS identified 27 SCs , two of which ( SC27 and SC4 ) were polyphyletic in the ML phylogeny ( Fig 1 ) . SC27 was comprised of low frequency genotypes whereas SC4 contained three distinct monophyletic clades that were bifurcated by branches with low bootstrap support . Based on recombination analysis using Gubbins and assessment of temporal signal ( i . e . , molecular clock ) , SC4 as well as 10 other SCs were further subdivided , as it was evident that substantial ancestral recombination events occurred on branches separating dominant monophyletic clades . This subdivision is consistent with the biological definitions of lineages or sub-populations [59 , 60] . Subsequent analysis focused on 33 SCs or sub-SCs that varied in size from 10 to 71 isolates ( Table 1 ) . The proportion of isolates belonging to each SC differed between age groups for only four of 27 SCs , among isolates collected from 2006–2008 . SC07 ( serotype 35A ) and SC15 ( serotype 15A ) were more common among children ≤5 years of age , 0 . 8% and 1 . 6% adults compared to 3 . 3% and 4 . 4% children , respectively ( p = 0 . 03 and 0 . 05 ) . SC08 ( serotype 35B ) and SC26 ( serotypes 19A/15C ) were more common among adults , 5 . 6% and 14 . 4% adults compared to 2 . 2% and 8 . 4% children , respectively ( p = 0 . 05 and 0 . 04 ) . For temporal comparison , we divided study periods into three epochs and six sub-epochs ( 1A/B , 2A/B , 3A/B ) ( S2 Fig ) . To verify representativeness of isolates used for genome sequencing in this study , we obtained prevalence data on 3 , 868 carriage events from children ≤5 years of age in the parent N/WMA carriage studies from which the genomic sample was drawn . This included 1227 events from Epoch 1 , 1038 from Epoch 2 , and 1603 from Epoch 3 . For the major epochs , the proportions of NVT , PCV7 , and PCV13 serotypes in our sample were comparable with the serotype dynamics characterized by the three N/WMA parent studies ( Fig 2A ) . The exception was the proportion of NVT and PCV7 VT in Epoch 1 , which was due to differences between serological and genomic assignment of serogroup 6 isolates . In Epoch 1 , serotypes 6B and 6C were both assigned to serotype 6B by the Quellung reaction used in the parent carriage study . This was subsequently resolved in the current study using a genomic approach to determine serotype , and later carriage studies were able to distinguish 6B from 6C . In pre-PCV Epoch 1 , 26 . 3% of the sample was comprised of PCV7 VT , mostly serotypes 23F , 9V , 14 , and 19F . Post-PCV7 , the proportion of PCV7 VT in Epoch 2 fell to 1 . 8% . The prevalence of PCV13 VTs declined steadily from 17 . 5% in Epoch 1 to 11 . 3% in Epoch 3 . The reduction in PCV13-specific VT after the introduction of PCV7 was likely due to the cross-reactivity of the 6B component of PCV7 with serotype 6A [61] , which can be inferred from the elimination of SC17 ( serotype 6A ) after Epoch 1 ( Fig 3 ) . Fluctuations in serotype distribution were reflected in measures of serotype diversity . Simpson’s D , which summarizes diversity as the probability that two isolates chosen at random are different , increased from Epoch 1 to 2 , reflecting an increase in previously low-frequency NVT serotypes as well as the introduction of previously unobserved serotypes ( Fig 2B ) . Fig 3 illustrates how the composition of the 27 main SCs changed during each of the three epochs . Of two lineages containing PCV7 VT only in Epoch 1 , one ( SC12 ) disappeared after vaccination , and another remained , with only PCV7 NVT isolates in Epochs 2 and 3 . In SCs containing both PCV7 VT and NVT , the VT lineages were largely eliminated . After Epoch 1 , the composition of the pneumococcal population in our sample and parent carriage studies shifted to a predominance of NVT and PCV13 VT , with the largest increases in serotypes 23B and 15C . While in most cases the NVT increases arose from serotypes previously observed in Epoch 1 , serotypes belonging to SC10 , SC22 , and SC24 were not detected until Epoch 2 . PCV13 VTs in our sample were not significantly impacted between Epoch 2 and 3 . Further comparison of PCV13 implementation data from N/WMA communities during Epoch 3 sampling demonstrated incomplete vaccine coverage and persistence of PCV13 vaccine serotypes ( S3 Fig ) . This finding is consistent with the previous observation that the impact of PCV13 on carriage among underimmunized children was not detected until vaccine coverage in the community reached 58% [6] . This coverage level was not attained until February 2011 , at which point 52% of the Epoch 3 sample had been collected . As a result , our assessment of the impact of PCV13 on the overall pneumococcal population was limited . We used Watterson’s theta ( ΘW ) –proportional to the number of polymorphic sites—and Tajima’s D to assess the impact of vaccine on population level genetic diversity and population size . Under neutrality and constant population size , ΘW = 2Neμ , where Ne is the effective population size and μ is the mutation rate [51] . Selective removal of several clusters of related strains , such as lineages or sub-lineages associated with VT , should lead to a reduction in ΘW . A related measure , Tajima’s D , tests for evidence of population growth , with negative values suggesting population expansion ( due to the presence of rare variants at high frequencies ) and positive values suggesting balancing selection or population contraction [50] . Consistent with our expectations , ΘW decreased from Epoch 1B to 2A , illustrating an overall decrease in pneumococcal genomic diversity , while the average number of pairwise differences ( π ) was unaffected ( Fig 2C ) . Tajima’s D values computed for the polymorphic nucleotide sites in the core genome increased from -0 . 59 in Epoch 1B to 0 . 07 in 2A , signifying a removal of rare variants consistent with a species-wide population bottleneck ( Fig 2D ) . By Epoch 3B both ΘW and Tajima’s D returned to pre-PCV7 levels while π increased . No discernible changes in either measure were associated with PCV13 introduction . After the population genetic bottleneck induced by PCV7’s removal of VT , genetic diversity ( i . e . , ΘW ) may have been augmented by 1 ) clonal expansion of NVT lineages due to selection or genetic drift ( to increase ΘW such lineages would have to have been so rare post-bottleneck that they were not sampled ) , 2 ) introduction of new lineages , or 3 ) recombination . We hence examined evidence for each of these among individual SCs . Recombination rates ( r/m ) varied among SCs , ranging from 0 to 15 . 0 , averaging 4 . 25 ( Table 1 and S4 Fig ) . While coalescent analysis found SCs varied in mutation rates ( S5 Fig ) , there was no significant difference between the median evolutionary rates of NVT and VT SCs ( 95% CI: -1 . 06e-06–8 . 54e-06 , F ( 1 , 29 ) = 2 . 55 , p = 0 . 12 ) . Therefore , high evolutionary rates among NVT lineages were not solely responsible for recovering the diversity lost due to the removal of PCV7 VT . To investigate the contribution of introduction of new lineages or expansion of previously unsampled ones , we estimated the TMRCAs ( i . e . , lineage age ) of SCs . Overall , the median TMRCA was 1955 and ranged from 1839 ( SC21: 6A/C ST473 ) to as recent as 2000 ( SC10: 19A ST320 ) ( S6 Fig ) . Two SCs that were not identified during Epoch 1 sampling emerged following vaccination: SC10 ( S7 Fig ) , which is all type 19A and ST320 , and SC24 , largely comprised of serotype 23A ( S8 Fig ) related to PMEN clone Colombia23F-26 . Estimated TMRCA for SC10 was 2000 [95% HPD: 1996–2004] . The lineage age , taken together with its low level of genetic diversity ( Θw = 0 . 0006 ) and negative Tajima’s D value ( -2 . 15 ) , suggests that this SC was introduced after the implementation of PCV7 among southwest Native Americans and is currently experiencing population expansion . SC24 was first identified in 2006 during Epoch 2 , but its most recent common ancestor was estimated at 1958 [95% HPD: 1928–1980] , near the median TMRCA among all SCs . Considering its prevalence in Epoch 2 and moderate level of diversity ( Θw = 0 . 003 ) , it is likely that SC24 was not recently introduced and that its was present in the population before PCV7 but at a sufficiently low frequency not to be sampled until 2006 , by which time its frequency may have increased . Furthermore , SC24’s low Tajima’s D value ( -1 . 63 ) is consistent with population expansion . We hypothesized that post-PCV7 changes in pneumococcal populations would be visible as decreases in the effective population size ( Ne ) of predominantly VT lineages and increases in those of predominantly NVT lineages . The effective population size can be interpreted as the number of genomes contributing offspring to the next generation , and changes in Ne can be used to measure population growth or contraction . Inferring demography among SCs identified that over half ( 56% ) fit constant population size models based on MLEs ( Table 1 ) . Furthermore , while the remainder of SCs best fit a fluctuating Ne model ( i . e . , Skygrid ) , assessment of Ne trajectories identified only three that were significantly different from a constant size based on HPDs . These three SCs ( SC11 , SC17 , and SC26-A ) were found to be decreasing throughout the study period; one was PCV13 VT ( SC 17 ) and two were NVT ( SC11 and SC 26-A ) . To assess bias potentially introduced by removing recombination , we tested the association between recombination rates and inferred demography , which we found to not be significant ( F ( 1 , 30 ) = 0 . 44 , p = 0 . 51 ) [62] . Overall , these findings show that the relatively subtle increases in sample frequencies of individual SCs containing NVT are not visible as departures from a constant Ne . To test the hypothesis that selective removal of PCV7 VT disrupted accessory genome content , we compared accessory size and frequencies of 2370 COGs and 53 variants of 19 antigens between pre-PCV7 Epoch 1 to post-PCV7 epochs . Further , we tested the concurrent effect on metabolic loci by assessing frequencies of 22 , 434 biallelic SNPs found among 256 metabolic genes present in the core genome . For metabolic loci , accessory COGs , and antigen variants , within-group diversity was minimized when SC population groupings were assigned , compared to serogroup and serotype ( S9 Fig ) . The introduction of PCV7 resulted in an overall reduction in pangenome size , illustrated by the difference in logarithmic pangenome curves for Epochs 2A and 3B ( S10 Fig ) . A comparison of pre-PCV7 Epochs 1A and 1B provided a baseline estimate of stochastic , temporal fluctuations in frequencies in the absence of an effect of vaccine . Plotting COG frequencies in subsequent epochs demonstrated perturbation in pneumococcal accessory COGs frequencies following introduction of PCV7 ( S11 Fig ) . This perturbation is characterized by the dispersion of frequency scatterplots comparing Epochs 1A vs . 2A [R2 = 0 . 96 , MSE = 8 . 26x10-3 ( 95% CI: 8 . 32x10-3–8 . 40x10-3 ) ] and 2B [R2 = 0 . 98 , MSE = 6 . 65 x10-3 ( 95% CI: 6 . 60x10-3–6 . 70x10-3 ) ] ( Figs 4 and S11 ) . This effect was also observed when comparing the frequencies of polymorphic antigens and metabolic loci between epochs ( Figs 5 , S12 and S13 ) . For all sets of genomic loci , MSE in comparison to Epoch 1A are smaller for 1B than for any of the subsequent epochs , illustrating the disruption caused by PCV7 . While this observation alone could be explained by drift leading to increasing divergence in frequencies over time , a further observation cannot: in each example , MSEs decreased from Epoch 2 to 3 , indicating metabolic loci , accessory COGs , and antigen frequencies were trending back toward pre-PCV7 values ( Fig 5 ) . This trend was observed when isolates collected from individuals >5 years of age were removed from Epoch 2 and the analysis repeated . This led us to compare Epoch 3A ( post-PCV7/pre-PCV13 ) to previous sub-epochs to determine whether the pre-PCV7 Epochs 1A/B or the immediately preceding Epoch 2B were better predictors of COG/antigen frequencies . For accessory genome COG frequencies and metabolic loci , Epoch 2B was a better predictor of 3A frequencies; however , for antigens , pre-PCV7 Epoch 1B was the best predictor of Epoch 3A frequencies ( S14 Fig ) . Taken together , we found that antigen variant frequencies largely returned to pre-PCV7 values; however , some perturbations were not resolved ( Fig 6 ) . This was due largely to pspC groups 1/5 ( p = 0 . 01 ) and srtH Var-I ( p = 0 . 004 ) , which remained at higher frequencies at Epoch 3 , and rrgA Var-I ( p<0 . 001 ) , which was completely removed from the population . The impact of PCV7 introduction on pneumococcal serotype distributions has been well-characterized in the N/WMA and other communities , but the pneumococcal genome-wide impact has been investigated in fewer populations [3 , 63] . We studied genomes from a sample spanning the introduction of PCV7 and PCV13 , which , based on serotype distribution , were representative of the full set of data from which the sample was drawn . Beyond the expected impact on serotypes , we find the effect of vaccine on the pneumococcal population could be observed as changes in population level diversity , metabolic loci , size of the pneumococcal pangenome , and frequencies of accessory genes including polymorphic antigens . We further illustrate how pneumococcal genomic diversity and frequencies of accessory genome COGs rebounded after the population bottleneck induced by the selective removal of VT lineages by PCV7 . These findings help explain how the frequency distribution of polymorphic antigens , for example , largely return to baseline frequencies after being disrupted by vaccine . The post-PCV7 pneumococcal population in N/WMA saw the complete removal of two SCs and a significant reduction in prevalence of three . The population bottleneck was characterized by changes in levels and patterns of genomic diversity , decreasing ΘW and increasing Tajima’s D ( Fig 2 ) . Subsequently , the removal of VT pneumococci was counterbalanced by the expansion of SC9 and the emergence of two previously unobserved SCs , SC10 and SC24 . In Epoch 2 , we identified minor variations in the distribution of SCs by age group for four SCs . As none of the SCs contained PCV7 VT , differences likely resulted from variation in acquired serotype-specific immunity among children and adults [64] . Overall , population structure of SCs was comparable , consistent with pneumococcal transmission dynamics and the wide-ranging impact of the PCV7 vaccine on carriage in children and adults [5] . Despite the changes in the prevalence of SCs over time , no consistent pattern of change in the Ne of these SCs was detectable through coalescent analysis of individual SCs ( Table 1 ) . This lack of signal may be due to a number of factors . It may be that where vaccine pressure was strong enough to drastically change the population size of an SC , it was eliminated ( e . g . , SC12 ) , so the temporal signal was lost; where changes were more modest , e . g . in SC including both VT and NVT , the method may have been too insensitive to detect a change . While assessment of Ne did not clearly identify consistent changes , we did detect the post-PCV7 emergence of two SCs . By comparing TMRCA and core genome diversity , we infer that that the first , SC10 , appears to have been recently introduced among N/WMA , while the second , SC24 , appears to have become detectable due to the vaccine [8 , 65] . It is worth noting that assessing Ne and other population genetic parameters of pneumococcal lineages makes implicit assumptions about defining SCs as populations and a collection of SCs as a metapopulation , which , to varying degrees , may compete or interact with one another through recombination . Indeed , this definition is more complex and requires consideration of competition , gene flow , and niche overlap among lineages [60 , 66 , 67] . Here , we statistically define SCs and find that these populations are often good predictors of serogroup , metabolic profile , and gene content , thus generally demonstrate genomic coherence consistent with the concept of a bacterial population . Pneumococcal genomic data from carriage studies in the US are limited [9] . The N/WMA sample provides an opportunity to assess post-vaccine changes in the pneumococcal populations across demographically and geographically varied regions and , at large , the generalizability of bacterial pathogen population dynamics . Comparable analysis of population structure of 616 carriage isolates from Massachusetts collected between 2001 and 2007 found less structure ( 15 monophyletic SCs ( n = 616 ) ) compared to the N/WMA sample ( 25 monophyletic SCs ( n = 937 ) ) [9] , and unlike Massachusetts , where the post-PCV7 population emerged largely from the pre-existing serotype diversity , in the N/WMA sample we observed seven previously unidentified serotypes and two entire SCs post-PCV7 . Considering carriage data from the larger parent studies , 13 previously unidentified serotypes , excluding 6C , were observed post-PCV7 . This difference aside , SC composition and pneumococcal population dynamics were consistent between N/WMA and Massachusetts . For example , SC9 ( also SC9 in the Massachusetts study [9] ) experienced a near identical population shift post-PCV7 ( S15 Fig ) . This SC , which is comprised of VT 23F and NVTs 23A and 23B , is thought to have arisen through multiple serotype-switching events . In the N/WMA sample , it was one of the most successful in terms of overall prevalence in Epoch 1 . As observed in Massachusetts , PCV7 effectively removed 23F isolates from the SC; however , SC9 NVTs subsequently increased 3 . 5% from Epoch 1 to 3 . This shows that these changes were not restricted to the Massachusetts population , but were replicated in a very different setting , and may suggest that SC9 occupies a specific niche . Consistent with this hypothesis , we find that the antigen profiles for VT 23F and the NVT 23B population that replaced it , to be largely consistent with the exception of zmpA ( S16 Fig ) . Taken together , we observe similar pneumococcal population dynamics in two geographically and demographically distinct populations that share common vaccine histories , suggesting that response to population shaping processes are relatively consistent . We find that each SC is defined by a unique profile of metabolic loci , accessory COGs , and antigen variants . These profiles are most resolved at the SC level rather than serotype or serogroup , as the same serotype can be found in multiple SCs due to switching events . Moreover , within an SC , these genomic loci show significant linkage disequilibrium despite appreciable recombination among pneumococci [68] . Consistent with this linkage , we observed a coincident impact of PCV7 on genetic diversity , accessory COG frequencies , polymorphic antigens , and metabolic loci . The population genomic perturbation that resulted from the removal of PCV7 VT was significantly mitigated by Epoch 3 , with frequencies of antigen variants , in particular , returning to pre-vaccine values . A recently proposed model of NFDS provides one putative mechanism for the maintenance of antigen variants and accessory COGs at optimal frequencies [27] , and variant-specific host immunity provides a biologically plausible mechanism for NFDS on antigens . Early evidence of balancing selection among pneumococci was the reemergence of strains possessing a type 1 pilus after PCV7 significantly reduced piliated serotypes [69] . In the current study , we also observe the reemergence of type 1 pilus driven by serotype 19A ST320 ( SC10 ) . And while the observation with the pilus involved a change in presence-absence frequency , we now see the same dynamic extending to frequencies of antigen variants . Yet , due to linkage it is difficult to untangle which loci are being acted upon by selection and which reflect hitchhiking . Alternatively , balancing selection could be acting upon metabolic loci which are important to niche adaption and have been implicated in post-vaccine metabolic shifts [18] . In an effort to identify which loci may be driving post-vaccine success of SCs , we considered the frequencies of metabolic loci , accessory COG , and antigen variants separately . We find PCV13-era ( Epoch 3 ) frequencies of polymorphic antigens are better predicted by pre-PCV7 ( Epoch 1 ) frequencies than the immediately preceding period . In addition , we observe that overall COG frequencies seemed to trend toward pre-PCV7 norms with increasing time since vaccine introduction , while frequencies of metabolic loci remained disrupted . This does not rule out variation in metabolic loci or other core genes such as GroEL as driving forces for pneumococcal population structure [70]; however , it remains difficult to assign fitness differences based on observed genetic variation . For example , two SCs may be divergent in metabolic loci but capable of exploiting the same metabolic niche . Previous models have proposed that recombination is the mechanism underlying the post-vaccine shift in metabolic , virulence , and antigenic loci [18] . However , we argue that in our sample , recombination has likely not had enough time to shuffle antigen variants or other COGs into different genomic backgrounds . For example , if we again consider the replacement of VT 23F by NVT 23B belonging to SC9 , we observe that both populations possess similar antigenic profiles ( S16 Fig ) . Yet , the TMRCA of the 23B population , and all associated recombination events , predate the introduction of PCV7 ( S15 and S16 Figs ) . This illustrates that at least in this case , an existing population possessing a near identical antigenic profile contributed to the rebalancing of the distribution of antigen variants in the overall pneumococcal population . Overall , the pneumococcal accessory genome is comprised of varying types of MGE ( e . g . , phages and antigens ) , and it is likely that their distribution is controlled by many different , yet interconnected , processes [17] . As such , the underlying dynamics maintaining antigenic variant and accessory COG frequencies require further investigation . Through comprehensive analysis of serotype distribution and population dynamics of S . pneumoniae spanning the introduction of PCV7 and PCV13 among N/WMA communities , we gain a broad understanding of the impact of vaccine on population structure , serotype distribution , and pangenome composition . After the introduction of PCV7 , we observe clonal replacement of VT by NVT as well as clonal expansion of vaccine-associated serotypes during a period when carriage prevalence remained unchanged . Further , we show PCV7 significantly disrupted accessory COG frequencies , including frequencies of polymorphic antigens important to host-pathogen interactions . This post-PCV7 period of ‘flux’ in serotype diversity and accessory COG distribution was normalized by Epoch 3 , demonstrating rapid adaption to the post-vaccine landscape . Moving forward , continued genomic surveillance will be required to monitor the emergence of new lineages and to investigate the impact on post-PCV13 pneumococcal populations . Last , as balancing selection appears to be an integral component of pneumococcal adaption and considerable serotype-lineage-accessory genome linkage exists , the joint effect of removal of vaccine serotypes and linked antigens on host-susceptibility to extant lineages merits further study , as it has significant implications for the future of protein-based pneumococcal vaccines . For example , protein-based vaccines should consider the prevalence of polymorphic variants across host populations and either include multiple variants of the same antigen or target those in greatest frequency .
Pneumococcal disease caused by the bacteria Streptococcus pneumoniae remains a significant cause of morbidity and mortality despite the existence of an effective vaccine . This is because the vaccines only target a small proportion of the total pneumococcal population . Introduction of vaccine in the United States removed vaccine serotypes leaving an open niche that was rapidly filled by non-vaccine serotypes . Forecasting which serotypes , and more generally which pneumococcal lineages , will increase in frequency in carriage and disease is an active area of research with significant public health importance . Here , we investigate the evolutionary impact of vaccination on the pneumococcal population using genomic data from a collection of 937 pneumococcal isolates collected from 1998–2012 among Native American communities . We find the impact of vaccine on the pneumococcal population could be observed in reduced diversity and changing frequencies of genes . Diversity subsequently rebounded through expansion and in-migration of non-vaccine lineages . Further , frequencies of genes coding for protein antigens important to host-pathogen interaction were initially disrupted but later returned to pre-vaccine values , suggesting selection may have acted in maintaining frequencies . These data improve our understanding of pneumococcal evolution and emphasize the need to consider genome composition when inferring the impact of vaccination .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[ "taxonomy", "medicine", "and", "health", "sciences", "ecology", "and", "environmental", "sciences", "immune", "physiology", "immunology", "population", "genetics", "antigenic", "variation", "vaccines", "phylogenetics", "data", "management", "population", "biology", "infectious", "disease", "control", "immune", "system", "proteins", "infectious", "diseases", "computer", "and", "information", "sciences", "ecological", "metrics", "genomics", "proteins", "antigens", "comparative", "genomics", "evolutionary", "systematics", "species", "diversity", "genetic", "loci", "biochemistry", "ecology", "physiology", "genetics", "biology", "and", "life", "sciences", "conjugate", "vaccines", "computational", "biology", "evolutionary", "biology" ]
2018
The impact of serotype-specific vaccination on phylodynamic parameters of Streptococcus pneumoniae and the pneumococcal pan-genome
A long-standing goal in artificial intelligence is creating agents that can learn a variety of different skills for different problems . In the artificial intelligence subfield of neural networks , a barrier to that goal is that when agents learn a new skill they typically do so by losing previously acquired skills , a problem called catastrophic forgetting . That occurs because , to learn the new task , neural learning algorithms change connections that encode previously acquired skills . How networks are organized critically affects their learning dynamics . In this paper , we test whether catastrophic forgetting can be reduced by evolving modular neural networks . Modularity intuitively should reduce learning interference between tasks by separating functionality into physically distinct modules in which learning can be selectively turned on or off . Modularity can further improve learning by having a reinforcement learning module separate from sensory processing modules , allowing learning to happen only in response to a positive or negative reward . In this paper , learning takes place via neuromodulation , which allows agents to selectively change the rate of learning for each neural connection based on environmental stimuli ( e . g . to alter learning in specific locations based on the task at hand ) . To produce modularity , we evolve neural networks with a cost for neural connections . We show that this connection cost technique causes modularity , confirming a previous result , and that such sparsely connected , modular networks have higher overall performance because they learn new skills faster while retaining old skills more and because they have a separate reinforcement learning module . Our results suggest ( 1 ) that encouraging modularity in neural networks may help us overcome the long-standing barrier of networks that cannot learn new skills without forgetting old ones , and ( 2 ) that one benefit of the modularity ubiquitous in the brains of natural animals might be to alleviate the problem of catastrophic forgetting . To test our hypotheses , we set up an environment in which there is a potential for catastrophic forgetting and where individuals able to avoid this forgetting receive a higher evolutionary fitness , meaning they are more likely to reproduce . The environment is an abstraction of a world in which an organism performs a daily routine of trying to eat nutritious food while avoiding eating poisonous food . Every day the organism observes every food item one time: half of the food items are nutritious and half are poisonous . To achieve maximum fitness , the individual needs to eat all the nutritious items and avoid eating the poisonous ones . After a number of days , the season changes abruptly from a summer season to a winter season . In the new season , there is a new set of food sources , half of them nutritious and half poisonous , and the organism has to learn which is which . After this winter season , the environment changes back to the summer season and the food items and their nutritious/poisonous statuses are the same as in the previous summer . The environment switches back and forth between these two seasons multiple times in the organism’s lifetime . Individuals that remember each season’s food associations perform better by avoiding poisonous items without having to try them first . We consider each pair of a summer and winter season a year . Every season lasts for five days , and in each day an individual encounters all four food items for that season in a random order . A lifetime is three years ( Fig . 2 ) . To ensure that individuals must learn associations within their lifetimes instead of having genetically hardcoded associations [47 , 62] , in each lifetime two food items are randomly assigned as nutritious and the other two food items are assigned as poisonous ( Fig . 3 ) . To select for general learners rather than individuals that by chance do well in a specific environment , performance is averaged over four random environments ( lifetimes ) for each individual during evolution , and over 80 random environments ( lifetimes ) when assessing the performance of final , end-of-experiment individuals ( Methods ) . This environment selects for agents that can avoid forgetting old information as they learn new , unrelated information . For instance , if an agent is able to avoid forgetting the summer associations during the winter season , it will immediately perform well when summer returns , thus outcompeting agents that have to relearn summer associations . Agents that forget , especially catastrophically , are therefore at a selective disadvantage . Our main results were found to be robust to variations in several of our experimental parameters , including changes to the number of years in the organism’s lifetime , the number of different seasons per year , the number of different edible items , and different representations of the inputs ( the presence of items being represented either by a single input or distributed across all inputs for a season ) . We also observed that our results are robust to lengthening the number of days per season: networks in the experimental treatment ( called “P&CC” for reasons described below ) significantly outperform the networks in the control ( “PA” ) treatment ( p < 0 . 05 ) even when doubling or quadrupling the number of days per season , although the size of the difference diminished in longer seasons . The addition of a cost for connections ( the P&CC treatment ) leads to a rapid , sustained , and statistically significant fitness advantage versus not having a connection cost ( the PA treatment ) ( Fig . 4 ) . In addition to overall performance across generations , we looked at the day-to-day performance of final , evolved individuals ( Fig . 5 ) . P&CC networks learn associations faster in their first summer and winter , and maintain higher performance over multiple years ( pairs of seasons ) . The presence of a connection cost also significantly increases network modularity ( Fig . 4 ) , confirming the finding of Clune et al . [23] in this different context of networks with within-life learning . Networks evolved in the P&CC treatment tend to create a separate reinforcement learning module that contains the reward and punishment inputs and most or all neuromodulatory neurons ( Fig . 6 ) . One of our hypotheses ( Fig . 1 , bottom ) suggested that such a separation could improve the efficiency of learning , by regulating learning ( via neuromodulatory neurons ) in response to whether the network performed a correct or incorrect action , and applying that learning to downstream neurons that determine which action should be taken in response to input stimuli . To quantify whether learning is separated into its own module , we adopted a technique from [23] , which splits a network into the most modular decomposition according to the modularity Q score [65] . We then measured the frequency with which the reinforcement inputs ( reward/punishment signals ) were placed into a different module from the remaining food-item inputs . This measure reveals that P&CC networks have a separate module for learning in 31% of evolutionary trials , whereas only 4% of the PA trials do , which is a significant difference ( p = 2 . 71 × 10−7 ) , in agreement with our hypothesis ( Fig . 1 , bottom ) . Analyses also reveal that the networks from both treatments that have a separate module for learning perform significantly better than networks without this decomposition ( median performance of modular networks in 80 randomly generated environments ( Methods ) : 0 . 87 [95% CI: 0 . 83 , 0 . 88] vs . non-modular networks: 0 . 80 [0 . 71 , 0 . 84] , p = 0 . 02 ) . Even though only 31% of the P&CC networks are deemed modular in this particular way , the remaining P&CC networks are still significantly more modular on average than PA networks ( median Q scores are 0 . 25 [0 . 23 , 0 . 28] and 0 . 2 [0 . 19 , 0 . 22] respectively , p = 4 . 37 × 10−6 ) , suggesting additional ways in which modularity improves the performance of P&CC networks . After observing that a connection cost significantly improves performance and modularity , we analyzed whether this increased performance can be explained by the increased modularity , or whether it may better correlate with network sparsity , since P&CC networks also have fewer connections ( P&CC median number of connections is 35 . 5 [95% CI: 31 . 0 , 40 . 0] vs . PA 82 . 0 [74 . 0 , 97 . 1] , p = 7 . 97 × 10−19 ) . Both sparsity and modularity are correlated with the performance of networks ( Fig . 7 ) . Sparsity also correlates with modularity ( p = 5 . 15 × 10−40 as calculated by a t-test of the hypothesis that the correlation is zero ) , as previously shown [23 , 66] . Our interpretation of the data is that the pressure for both functionality and sparsity causes modularity , which in turn helps evolve learners that are more resistant to catastrophic forgetting . However , it cannot be ruled out that sparsity itself mitigates catastrophic forgetting [1] , or that the general learning abilities of the network have been improved due to the separation into a skill module and a learning module . Either way , the data support our hypothesis that a connection cost promotes the evolution of sparsity , modularity , and increased performance on learning tasks . We next investigated whether the improved performance of P&CC individuals is because they forget less . Measuring the percent of information a network retains can be misleading , because networks that never learn anything are reported as never forgetting anything . In many PA experiments , networks did not learn in one or both seasons , which looks like perfect retention , but for the wrong reason: they do not forget anything because they never knew anything to begin with . To prevent such pathological , non-learning networks from clouding this analysis , we compared only the 50 highest-performing experiments from each treatment , instead of all 100 experiments . For both treatments , we then measured retention and forgetting in the highest-performing network from each of these 50 experiments . To illuminate how old associations are forgotten and new ones are formed , we performed an experiment from studies of association forgetting in humans [11]: already evolved individuals learned one task and then began training on a new task , during which we measured how their performance on the original task degraded . Specifically , we allowed individuals to learn for 50 winter days—to allow even poor learners time to learn the winter associations—before exposing them to 20 summer days , during which we measured how rapidly they forgot winter associations and learned summer associations ( Methods ) . Notice that individuals were evolved in seasons lasting only 5 days , but we measure learning and forgetting for 20 days in this analysis to study the longer-term consequences of the evolved learning architectures . Thus , the key result relevant to catastrophic forgetting is what occurs during the first five days . We included the remaining 15 days to show that the differences in performance persist if the seasons are extended . P&CC networks retain higher performance on the original task when learning a new task ( Fig . 8 , left ) . They also learn the new task better ( Fig . 8 , center ) . The combined effect significantly improves performance ( Fig . 8 , right ) , meaning P&CC networks are significantly better at learning associations in a new season while retaining associations from a previous one . To further understand whether the increased performance of the P&CC individuals is because they learn more , retain more , or both , we counted the number of retained and learned associations for individuals in 80 randomly generated environments ( lifetimes ) . If we regard performance in each season as a skill , this experiment measures whether the individuals can retain a previously-learned skill ( perfect summer performance ) after learning a new skill ( perfect winter performance ) . We tested the knowledge of the individuals in the following way: at the end of each season , we counted the number of sets of associations ( summer or winter ) that individuals knew perfectly , which required them knowing the correct response for each food item in that season . We formulated four metrics that quantify how well individuals knew and retained associations . The first metric ( “Perfect” ) measures the number of seasons an individual knew both sets of associations ( summer and winter ) . Doing well on this metric indicates reduced catastrophic forgetting because it requires retaining an old skill even after a new one is learned . P&CC individuals learned significantly more Perfect associations ( Fig . 9 , Perfect ) . The second metric ( “Known” ) is the sum of the number of seasons that summer associations were known and the number of seasons that winter associations were known . In other words , it counts knowing either season in a year and doubly counts knowing both . P&CC individuals learned significantly more of these Known associations ( Fig . 9 , Known ) . The third metric counts the number of seasons in which an association was “Forgotten” , meaning an association was completely known in one season , but was not in the following season . There is no significant difference between treatments on this metric when measured in absolute numbers ( Fig . 9 , Forgotten ) . However , measured as a percentage of Known items , P&CC individuals forgot significantly fewer associations ( Fig . 9 , % Forgotten ) . The modular P&CC networks thus learned more and forgot less—leading to a significantly lower percentage of forgotten associations . The final metric counts the number of seasons in which an association was “Retained” , meaning an association was completely known in one season and the following season . P&CC individuals retained significantly more than PA individuals , both in absolute numbers ( Fig . 9 , Retained ) and as a percentage of the total number of known items ( Fig . 9 , % Retained ) . In each season , an agent can know two associations ( summer and winter ) , leading to a maximum score of 6 × 80 × 2 = 960 for the known metric ( 6 seasons per lifetime ( Fig . 2 ) , 80 random environments ) . The agent can retain or forget two associations each season except the first , making the maximum score for these metrics 5 × 80 × 2 = 800 . However , the agent can only score one perfect association ( meaning both summer and winter is known ) each season , leading to a maximum score of 6 × 80 = 480 for that metric . In summary , this analysis reveals that a connection cost caused evolution to find individuals that are better at gaining new knowledge without forgetting old knowledge . In other words , adding a connection cost mitigated catastrophic forgetting . That , in turn , enabled an increase in the total number of associations P&CC individuals learned in their lifetimes . To further test whether the improved performance in the P&CC treatment results from it mitigating catastrophic forgetting , we conducted experiments in a regime where retaining skills between tasks is impossible . Under such a regime , if the P&CC treatment does not outperform the PA treatment , that is evidence for our hypothesis that the ability of P&CC networks to outperform PA networks in the normal regime is because P&CC networks retain previously learned skills more when learning new skills . To create a regime similar to the original problem , but without the potential to improve performance by minimizing catastrophic forgetting , we forced individuals to forget everything they learned at the end of every season . This forced forgetting was implemented by resetting all neuromodulated weights in the network to random values between each season change . The experimental setup was otherwise identical to the main experiment . In this treatment , evolution cannot evolve individuals to handle forgetting better , and can focus only on evolving good learning abilities for each season . With forced forgetting , the P&CC treatment no longer significantly outperforms the PA treatment ( Fig . 10 ) . This result indicates that the connection cost specifically helps evolution in optimizing the parts of learning related to resistance against forgetting old associations while learning new ones . Interestingly , without the connection cost ( the PA treatment ) , forced forgetting significantly improves performance ( Fig . 10 , p = 2 . 5 × 10−5 via bootstrap sampling with randomization [68] ) . Forcing forgetting likely removes some of the interference between learning the two separate tasks . With the connection cost , however , forced forgetting leads to worse results , indicating that the modular networks in the P&CC treatment have found solutions that benefit from remembering what they have learned in the past , and thus are worse off when not allowed to remember that information . We hypothesized that a key factor that causes modularity to help minimize catastrophic forgetting is neuromodulation , which is the ability for learning to be selectively turned on and off in specific neural connections in specific situations . To test whether neuromodulation is essential to evolving a resistance to forgetting in our experiments , we evolved neural networks with and without neuromodulation . When we evolve without neuromodulation , the Hebbian learning dynamics of each connection are constant throughout the lifetime of the organism: this is accomplished by disallowing neuromodulatory neurons from being included in the networks ( Methods ) . Comparing the performance of networks evolved with and without neuromodulation demonstrates that with purely Hebbian learning ( i . e . without neuromodulation ) evolution never produces a network that performs even moderately well ( Fig . 11 ) . This finding is in line with previous work demonstrating that neuromodulation allows evolution to solve more complex reinforcement learning problems than purely Hebbian learning [25] . While the non-modulatory P&CC networks perform slightly better than non-modulatory PA networks , the differences , while significant ( P&CC performance 0 . 72 [95% CI: 0 . 71 , 0 . 72] vs . PA 0 . 70 [0 . 69 , 0 . 71] , p = 0 . 003 ) , are small . Because networks in neither treatment learn much , studying whether they suffer from catastrophic forgetting is uninformative . These results reveal that neuromodulation is essential to perform well in these environments , and its presence is effectively a prerequisite for testing the hypothesis that modularity mitigates catastrophic forgetting . Moreover , neuromodulation is ubiquitous in animal brains , justifying its inclusion in our default model . One can think of neuromodulation , like the presence of neurons , as a necessary , but not sufficient , ingredient for learning without forgetting . Including it in the experimental backdrop allows us to isolate whether modularity further improves learning and helps mitigate catastrophic forgetting . In the experiments we performed , we found evidence that adding a connection cost when evolving neural networks significantly increases modularity and the ability of networks to learn new skills while retaining previously learned skills . The resultant networks have a separate learning module and exhibit significantly higher performance , learning , and retention . We further found three lines of evidence that modularity improves performance and helps prevent catastrophic forgetting: ( 1 ) networks with a separate learning module performed significantly better , ( 2 ) modularity and performance are significantly correlated , and ( 3 ) the performance increase disappeared when the ability to retain skills was artificially eliminated . These findings support the idea that neural modularity can improve learning performance both for tasks with the potential for catastrophic forgetting , by reducing the overlap in how separate skills are stored ( Fig . 1 , top ) , and in general , by modularly separating learned skills from reward signals ( Fig . 1 , bottom ) . We also found evidence supporting the hypothesis that the ability to selectively regulate per-connection learning in specific situations , called neuromodulation , is critical for the benefits of a connection cost to be realized . In the presence of neuromodulatory learning dynamics , which occur in the brains of natural animals [24 , 54] , a connection cost could thus significantly mitigate catastrophic forgetting . This work thus provides a new candidate technique for improving learning and reducing catastrophic forgetting , which is essential for advancing our goal of making sophisticated robots and intelligent software based on neural networks . It also suggests that one benefit of the modularity ubiquitous in natural networks may be improved learning via reduced catastrophic forgetting . While we found these results hold in the experiments we conducted , much work remains to be done on the interesting question of how catastrophic forgetting is avoided in animal brains . Future work in different types of problems and experimental setups are needed to confirm or deny the hypotheses suggested in this paper . Specific studies that can investigate the generality of our hypothesis include studying whether the connection cost technique still reduces interference when inputs cannot be as easily disentangled ( for instance , if certain inputs are shared between several skills ) , investigating the effect of more complex learning tasks that may not be learned at all if the agent forgets between training episodes , and further exploring the effect of experimental parameters , such as the length of training episodes , number of tasks , and different neural network sizes and architectures . Additionally , while we focused primarily on evolution specifying modular architectures , those architectures could also emerge via intra-life learning rules that lead to modular neural architectures . In fact , there may have been evolutionary pressure to create learning dynamics that result in neural modularity: whether such “modular plasticity” rules exist , how they mechanistically cause modularity , and the role of evolution in producing them , is a ripe area for future study . More generally , exploring the degree to which evolution encodes learning rules that lead to modular architectures , as opposed to hard coding modular architectures , is an interesting area for future research . The experiments in this paper are meant to invigorate the conversation about how evolution and learning produce brains that avoid catastrophic forgetting . While the results of these experiments shed light on that question , the importance , magnitude , and complexity of the question will yield fascinating research for decades , if not centuries , to come . We utilize a standard network model common in previous studies of the evolution of modularity [23 , 57] , extended with neuromodulatory neurons to add reinforcement learning dynamics [25 , 69] . The network has five layers ( Supp . S1 Fig ) and is feed-forward , meaning each node receives inputs only from nodes in the previous layer and sends outputs only to nodes in the next layer . The number of neurons is 10/4/2 for the three hidden layers . The weights ( connection strengths ) and biases ( activation thresholds ) in the network take values in the range [-1 , 1] . Following the paper that introduced the connection cost technique [23] , networks are directly encoded [70 , 71] . Information flows through the network from the input layer towards the output layer , with one layer per time step . The output of each node is a function of its inputs , as described in the next section . The neuromodulated ANN model in this paper was introduced by Soltoggio et al . [25] , and adapted for the Sferes software package by Tonelli and Mouret [69] . It differs from standard ANN models by employing two types of neurons: non-modulatory neurons , which are regular , activity-propagating neurons , and modulatory neurons . Inputs into each neuron consist of two types of connections: modulatory connections Cm and non-modulatory connections Cn ( normal neural network connections ) . The output of a neuron is decided by the weighted sum of its non-modulatory input connections , as follows: a i = φ ∑ j ∈ C n w i j a j + b i ( 1 ) where i and j are neurons , aj is the output of neuron j , bi is the bias of neuron i , wij is the weight of the connection between neuron i and j , and φ is a sigmoid function that maps its input to a value in the range [−1 , 1] , allowing both positive and negative outputs . Only non-modulatory connections ( outgoing connections from non-modulatory neurons ) are plastic . Their weight modification depends on the sum of modulatory inputs to the downstream neurons they connect to and a constant learning rate η . Their weight change is calculated by the following two equations: m i = φ ∑ j ∈ C m w i j a j ( 2 ) ∀ j ∈ C n : Δ w i j = η · m i · a i · a j ( 3 ) Equation 2 describes how the modulatory input to each neuron is calculated . φ is a sigmoid function that maps its input to the interval [−1 , 1] ( thus allowing both positive and negative modulation ) . The sum includes weighted contributions from all modulatory connections . Equation 3 describes how this modulatory input determines the learning rate of all incoming , non-modulatory connections to neuron i . η is a constant learning rate that is set to 0 . 04 in our experiments . The ai⋅aj component is a regular Hebbian learning term that is high when the activity of the pre- and post-synaptic neurons of a connection are correlated [45] . The result is a Hebbian learning rule that is regulated by the inputs from neuromodulatory neurons , allowing the learning rate of specific connections to be increased or decreased in specific circumstances . In control experiments without the potential for neuromodulation , all neurons were non-modulatory . Updates to the weights of their incoming connections were calculated via Equation 3 with mi set to a constant value of 1 . Our experiments feature a multi-objective evolutionary algorithm , which optimizes multiple objectives simultaneously . Specifically , it is a modification of the widely used Non-dominated Sorting Genetic Algorithm ( NSGA-II ) [72] . However , NSGA-II does not take into account that one objective may be more important than others . In our case , network performance is essential to survival , and minimizing the sum of connection costs is a secondary priority . To capture this difference , we follow [23] in having a stochastic version of Pareto dominance , in which the secondary objective ( connection cost ) only factors into selection for an individual with a given probability p . In the experiments reported here , the value of p was 0 . 75 , but preliminary runs demonstrated that values of p of 0 . 25 and 0 . 5 led to qualitatively similar results , indicating that the results are robust to substantial changes to this value . However , a p value of 1 was found to overemphasize connection costs at the expense of performance , leading to pathological solutions that perform worse than the PA networks . Evolutionary algorithms frequently get stuck in local optima [5] and , due to computational costs , are limited to small population sizes compared to biological evolution . To better capture the power of larger populations , which contain more diversity and thus are less likely to get trapped on local optima , we adopted the common technique of encouraging phenotypic diversity in the population [5 , 73 , 74] . Diversity was encouraged by adding a diversity objective to the multi-objective algorithm that selected for organisms whose network outputs were different than others in the population . As with performance , the diversity objective factors into selection 100% of the time ( i . e . the probability p for PNSGA was 1 ) . Technically , we register every choice ( to eat or not ) each individual makes and determine how different its sequence of choices is from the choices of other individuals: differences are calculated via a normalized bitwise XOR of the binary choice vectors of two individuals . For each individual , this difference is measured with regards to all other individuals , summed and normalized , resulting in a value between 0 and 1 , which measures how different the behavior of this individual is from that of all other individuals . Preliminary experiments demonstrated that , for the problems in this paper , this diversity-promoting technique is necessary to reliably obtain functional networks in either treatment , and is thus a necessary prerequisite to conduct our study . This finding is in line with previous experiments that have showed that diversity is especially necessary for problems that involve learning , because learning problems are especially laden with local optima [74] . All experiments were implemented in the Sferes evolutionary algorithm software package [75] . The exact source code and experimental configuration files used in our experiments , along with data from all our experiments , are freely available in the online Dryad scientific archive at http://dx . doi . org/10 . 5061/dryad . s38n5 . The variation necessary to drive evolution is supplied via random mutation . In each generation , every new offspring network is a copy of its parent that is randomly mutated . Mutations can add a connection , remove a connection , change the strength of connections , move connections and change the type of neurons ( switching between modulatory and non-modulatory ) . Probabilities and details for each mutational event are given in Supp . S1 Table . We chose these evolutionary parameters , including keeping things simple by not adding crossover , to maintain similarity with related experiments on evolving modularity [23] and neuromodulated learning [76] . The fitness function simulates an organism learning associations in a world that fluctuates periodically between a summer and a winter season . During evolution , each individual is tested in four randomly generated environments ( i . e . for four “lifetimes” , Fig . 2 ) that vary in which items are designated as food and poison , and in which order individuals encounter the items . Because there is variance in the difficulty of these random worlds , we test in 4 environments ( lifetimes ) , instead of 1 , to increase the sample size . We further increase the sample size to 80 environments ( lifetimes ) when measuring the performance of final , evolved , end-of-experiment individuals ( e . g . Figs . 8 and 9 ) . Individuals within the same generation are all subjected to the same four environments , but across generations the environments are randomized to select for learning , rather than genetically hard-coded solutions ( Fig . 3 ) . To start each environment ( note: not season ) from a clean slate , before being inserted in an environment the modulated weights of individuals are randomly initialized , which follows previous work with this neuromodulatory learning model [76] . Modulatory connections never change , and thus do not need to be altered between environments . In the runs without neuromodulation , all connections are reset to their genetically specified weights . Throughout its life , an individual encounters different edible items several times ( Fig . 2 ) . Fitness is proportional to the number of food items consumed minus the number of poison items consumed across all environments ( Supp . S7 Fig ) . Individuals that can successfully learn which items to eat and which to avoid are thus rewarded , and the best fitness scores are obtained by individuals that are able to retain this information across the fluctuating seasons ( i . e . individuals that do not exhibit catastrophic forgetting ) . Our modularity calculations follow those developed by Leicht and Newman for directed networks [67] , which is an extension of the most well-established modularity optimization method [65] . That modularity optimization method relies on the maximization of a benefit function Q , which measures the difference between the number of connections within each module and the expected fraction of such connections given a “null model” , that is , a statistical model of random networks . High values of Q indicate an “unexpectedly modular” network . For undirected networks , the null model traditionally corresponds to random networks constrained to have the same degree sequence as the network whose modularity is measured . Leicht and Newman extend this model to directed networks by distinguishing between the in-degree and out-degree of each node in the degree sequence [67] . The probability that the analyzed network has a connection between node i and j is therefore k i i n k j o u t / m , where k i i n and k j o u t are the in- and out-degrees of node i and j , respectively , m is the total number of edges in the network , and the modularity of a given decomposition for directed networks is as follows: Q = 1 m ∑ i j A i j - k i i n k j o u t m δ c i , c j ( 4 ) Aij is the connectivity matrix ( 1 if there is an edge from node i to node j , and 0 otherwise ) , m is the total number of edges in the network , and δci , cj is a function that is 1 if i and j belong to the same module , and 0 otherwise . Our results are qualitatively unchanged when using layered , feed-forward networks as “null model” to compute and optimize Q ( Supp . S2 Table ) . Maximizing Q is an NP-hard problem [77] , meaning it is necessary to rely on an approximate optimization algorithm instead of an exact one . Here we applied the spectral optimization method , which gives good results in practice at a low computational cost [67 , 78] . As suggested by Leicht and Newman [67] , each module is split in two until the next split stops increasing the modularity score . Each experimental treatment was repeated 100 times with different stochastic events ( accomplished by initiating experiments with a different numeric seed to a random number generator ) . Analyses are based on the highest-performing network from each trial . The experiments lasted 20 , 000 generations and had a population size of 400 . The environment had 2 different seasons ( “summer” and “winter” ) . Each season lasted 5 days , and cycled through 3 years ( Fig . 2 ) . In each season , 2 poisonous items and 2 nutritious items were available , each item encoded by a separate input neuron ( i . e . a “one-hot encoding” [64] ) . Considering the fact that visiting objects in a different order may affect learning , the total number of possible different environments is 25 , 920 . Each day we randomize the order in which food items are presented , yielding 4 ! = 24 different possibilities per day . There are in total 5 days per season , and an individual lives for 6 seasons , resulting in 5 × 6 = 30 days per lifetime ( Fig . 2 ) , and thus 24 × 30 = 720 different ways to visit the items in a single lifetime . In addition to randomizing the order items are visited in , the edibility associations agents are supposed to learn are randomized between environments . We randomly designate 2 of the 4 items as nutritious food , giving 4 2 = 6 different possibilities for summer and 6 different possibilities for winter . There are thus a total of 6 × 6 = 36 different ways to organize edibility associations across both seasons . In total , we have 720 × 36 = 25 , 920 unique environments , reflecting the 720 different ways food items can be presented and the 36 possible edibility associations . As mentioned in the previous section , four of these environments were seen by each individual during evolution , and 80 of them were seen in the final performance tests . In both cases they were selected at random from the set of 25 , 920 . Unless otherwise stated , the test of statistical significance is the Mann-Whitney U test . 95% bootstrapped confidence intervals of the median are calculated by re-sampling the data 5 , 000 times . In Fig . 4 , we smooth the plotted values with a median filter to remove sampling noise . The median filter has a window size of 11 , and we plot each 10 generations , meaning the median spans a total of 110 generations . While measuring the forgetting and retention of evolved individuals ( e . g . Figs . 8 and 9 ) , further learning was disabled . The process is thus ( 1 ) learn food associations , ( 2 ) measure what was learned and forgotten without further learning , and ( 3 ) repeat . Disabling learning allows measurements of what has been learned without the evaluation changing that learned information .
A long-standing goal in artificial intelligence ( AI ) is creating computational brain models ( neural networks ) that learn what to do in new situations . An obstacle is that agents typically learn new skills only by losing previously acquired skills . Here we test whether such forgetting is reduced by evolving modular neural networks , meaning networks with many distinct subgroups of neurons . Modularity intuitively should help because learning can be selectively turned on only in the module learning the new task . We confirm this hypothesis: modular networks have higher overall performance because they learn new skills faster while retaining old skills more . Our results suggest that one benefit of modularity in natural animal brains may be allowing learning without forgetting .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods" ]
[]
2015
Neural Modularity Helps Organisms Evolve to Learn New Skills without Forgetting Old Skills
Interactions between commensal fungi and gut immune system are critical for establishing colonic homeostasis . Here we found that mice deficient in Dectin-3 ( Clec4d-/- ) , a C-type lectin receptor that senses fungal infection , were more susceptible to dextran sodium sulfate ( DSS ) -induced colitis compared with wild-type mice . The specific fungal burden of Candida ( C . ) tropicalis was markedly increased in the gut after DSS treatment in Clec4d-/- mice , and supplementation with C . tropicalis aggravated colitis only in Clec4d-/- mice , but not in wild-type controls . Mechanistically , Dectin-3 deficiency impairs phagocytic and fungicidal abilities of macrophages , and C . tropicalis-induced NF-κB activation and cytokine production . The conditioned media derived from Dectin-3-deficient macrophages were defective in promoting tissue repairing in colonic epithelial cells . Finally , anti-fungal therapy was effective in treating colitis in Clec4d-/- mice . These studies identified the role of Dectin-3 and its functional interaction with commensal fungi in intestinal immune system and regulation of colonic homeostasis . Inflammatory bowel disease ( IBD ) , mainly Crohn disease and ulcerative colitis , is a chronic inflammatory disorder of the gut . Extensive studies have suggested that the etiology of IBD involves environmental and genetic factors that lead to dysfunction of the epithelial barrier , with consequent deregulation of the mucosal immune system and responses to microbiota [1 , 2] . Therefore , interactions between the commensal microbiota and gut immune system are critical for establishing colonic epithelial homeostasis . Mammalian gastrointestinal tract is colonized with multiple microbial communities , including bacteria , fungi and viruses . Although the vast majority of studies on commensal microbiota have focused on bacteria , commensal fungi were reported [3] and were linked with a number of gastrointestinal disease including IBD [4] , irritable bowel syndrome [5] , gastric ulcers[6] and chemotherapy-induced enteric disorders [7] . Gastrointestinal tract of healthy individuals contains 66 fungal genera and 184 fungal species , with Candida as the dominant fungal genera [8] . These fungi could become pathogenic as the result of a change in the environment , for example the loss or reduction of bacterial or suppression of immune system . Recently , Iliev et al . showed that dextran sodium sulfate ( DSS ) treatment allows pathogenic fungi to invade the intestinal wall and that Dectin-1 plays an important role in protecting the host from colitis [9] . Therefore , deficiencies in genes involved in innate and adaptive immune pathways may lead to disorders characterized by intestinal manifestations and loss of microbial diversity [10] . However , a recent study by Tang et al . showed that suppression of Dectin-1 signaling protects mice from experimental colitis by decreasing S100A8 and S100A9 antimicrobial peptide production [11] , which allows the overgrowth of L . murinus that trigger T regulatory cell expansion in the gut . These findings suggest that Dectin-1 can play opposite roles in DSS-induced colitis dependent on the microbial community in the gut . Several mammalian C-type lectin receptors ( CLRs ) , including Dectin-1 , Dectin-2 , Dectin-3 , and Mincle , function as pattern recognition receptors sensing fungal infections and inducing multiple signaling cascades , which lead to expression of various pro-inflammatory cytokines and antimicrobial proteins [12–18] . Specifically , Dectin-1 recognizes β-glucans on the surface of fungal yeast cells , whereas Dectin-2 recognizes α-mannan on the surface of fungal hyphae . Dectin-3 , a CLR also known as MCL/CLECSF8/Clec4d , functions as a pattern recognition receptor for sensing fungal infections by recognizing α-mannans [19 , 20] . Our previous data indicated that Dectin-3 forms a heterodimeric complex with Dectin-2 , which recognizes α-mannans and has greater sensitivity in sensing Candida albicans infections than either the Dectin-2 or Dectin-3 homodimer , leading to potent activation of NF-κB–dependent antifungal immune responses [12] . Dectin-3 is expressed by peripheral blood neutrophils , monocytes , and various subsets of dendritic cells [20] . Recognition of these cell wall components by CLRs induces the Syk/Caspase recruitment domain 9 ( CARD9 ) /NF-κB–dependent signaling pathway [21] , leading to production of inflammatory cytokines in innate immune cells and participating in antifungal responses . Furthermore , the induced pro-inflammatory cytokines regulate Th17 and Th1 cell differentiation . Subsequently , cytokines produced by Th17 and Th1 cells activate neutrophils and macrophages that mediate the clearance of infected fungi in vivo . Although investigators have explored the role of Dectin-3 in systemic immunity , its function in the gastrointestinal immune system has yet to be investigated . In this study , we have identified the role of Dectin-3 and its functional interaction with commensal fungi in intestinal immune responses and regulation of colonic homeostasis . We found that Dectin-3-deficient ( Clec4d-/- ) mice were more susceptible to DSS-induced colitis compared with wild-type mice . The specific fungal burden of C . tropicalis was markedly increased in the gut after DSS treatment in Clec4d-/- mice . Mechanistically , absence of Dectin-3 impairs the phagocytic and fungicidal abilities of macrophages . Dectin-3 is also required for C . tropicalis-induced CARD9/Bcl10 complex formation and NF-κB activation , which in turn induce tissue-repairing program in colonic epithelial tissues . Although recent studies indicate that Dectin-3 plays important roles in innate immune responses against fungal and bacterial infections [12 , 22–24] , the role of Dectin-3 in mucosal immunity has not been examined . To determine the role of Dectin-3 in mucosal immune responses , we first check the body weight and colons for spontaneous colitis . Both female and male Clec4d-/- mice have normal increase of body weight compared with wild-type mice ( S1A and S1B Fig ) . On histological examination , intestinal epithelial cell appeared normal and no observed spontaneous colitis was found in Clec4d-/- mice ( S1C Fig ) . We therefore induce intestinal injury and inflammation using DSS colitis model . After treatment with 2 . 5% DSS for 7 days and being given water for an additional 4 days , Clec4d-/- mice had greater weight loss , shorter colon lengths , and higher clinical scores than wild-type mice ( Fig 1A–1D ) . Similar results were found in another independent experiment ( S1D Fig ) . Clec4d-/- mice also exhibited increased mucosal erosion , inflammatory cell infiltration , crypt destruction and loss of goblet cells in the colon , compared with wild-type mice ( Fig 1E and 1F ) . Similar results were obtained comparing co-housed animals ( S1E Fig ) . These data suggest that Clec4d-/- mice were more susceptible to DSS-induced colitis . We then detected innate and adaptive immune cells in the mesenteric lymph nodes ( MLN ) and colonic lamina propria ( LP ) of mice given DSS . In a comparison with wild-type mice , we found that more macrophages were recruited to MLN and colonic LP in Clec4d-/- mice , whereas fewer Th17 cells were located in the MLN and LP of Clec4d-/- mice ( Fig 2A–2C ) . We also measured colonic expression levels of both cytokines and chemokines using qRT-PCR . No significant differences were observed between wild-type and Clec4d-/- mice in colitis induction stage ( S2A Fig ) . After water recovery stage , the expression of IL-6 , IL-17a , TNF-α , and macrophage inflammatory protein-2 ( MIP-2 ) was lower in Clec4d-/- mice than in wild-type mice ( Fig 2D ) . We also found increased expression of chemokine CXCL1 in the colons of Clec4d-/- mice than in wild-type . Similar results were found in the protein production levels of IL-6 , TNF-α , IL-17a , and MIP-2 in colonic LP cells supernatant ( S2B Fig ) . The colonic expression of IL-10 , IL-22 , IL-1β and IFN-γ were increased similarly in wild-type and Clec4d-/- mice ( S2C Fig ) . Systemic IL-6 , TNF-α , and IL-17a levels in serum were similar in wild-type and Clec4d-/- mice after DSS treatment ( S2D Fig ) , suggesting that these responses were specific for the intestinal environment . Furthermore , we found decreased expression of IL-6 and IL-17a and increased expression of CXCL1 in MLNs of Clec4d-/- compared with wild-type ( S2E Fig ) . IL-6 protects the intestinal epithelium injury by regulating Stat3 signaling [25] . Therefore , the defect in IL-6 expression observed in Clec4d-/- mice also may play a role in impaired epithelial restitution . Since Dectin-3 is involved in anti-fungal immunity [12] , we then examined the fungal burden in the colon . The basal level of total fungal burden in colons has no difference between Clec4d-/- and wild-type mice ( S3A Fig ) . After DSS treatment , the total fungal burden in the colon was markedly higher in Clec4d-/- mice than in wild-type mice ( Fig 3A and 3B ) . A previous study proved that Candida is the major intestinal fungal genus in mice [9] . Therefore , we quantified the relative levels of C . tropicalis , C . albicans , and C . glabrata in colon using quantitative PCR . We found that only C . tropicalis increased in Clec4d-/- mice during colitis ( Fig 3C ) . The proportion of C . tropicalis in total fungi was increased after DSS treatment ( Fig 3D and S3B Fig ) . Since Dectin-3 also involved in innate immune responses against bacterial infections , which may affect colitis progression , we then detected the bacteria burden before and after DSS treatment . As shown in S3C Fig , no difference of total bacteria burden was found between Clec4d-/- and wild-type mice . Given that C . tropicalis is an opportunistic pathogen , we further analyzed its role in the development of colitis in Clec4d-/- mice . We supplemented mice with C . tropicalis and gave them DSS as outlined in Fig 3E . This supplement could increase the fungal burden in both Clec4d-/- and wild-type mice ( S3D Fig ) . As shown in Fig 3F , the body weights of mice given C . tropicalis and DSS decreased by 20% on day 15 . We had to sacrifice the mice on day 16 according to our protocol . We found shorter colon length and more severe clinical scores in Clec4d-/- mice supplemented with C . tropicalis than those in Clec4d-/- mice ( Fig 3F–3I ) . However , the pathological changes of distal colon and cecum were similar in Clec4d-/- mice supplemented with or without C . tropicalis ( S3E Fig ) . In contrast , supplementation with C . tropicalis did not aggravate colitis in WT mice . To exclude the effect of fungal supplementation alone on colitis , we supplemented mice with C . tropicalis alone without DSS , and found that C . tropicalis could not induce colitis in both wild-type and Clec4d-/- mice without DSS treatment ( S4A–S4C Fig ) . The above C . tropicalis supplement experiment suggests an impaired fungal killing ability of Dectin-3 deficient mice . To determine the role of Dectin-3 in antifungal immunity , we examined phagocytic and fungicidal abilities in primary macrophages obtained from bone marrow ( BMDMs ) following challenge with C . tropicalis . We found that wild-type BMDMs were able to limit the intracellular replication of C . tropicalis , whereas Clec4d-/- BMDMs had a much larger fungal load ( Fig 4A ) . CFU assays demonstrated an increased number of viable yeasts recovered from Clec4d-/- macrophages ( Fig 4B ) . The difference between wild-type and Clec4d-/- macrophages in killing the phagocytosed C . tropicalis , was not due to a difference in phagocytosis by these macrophages , as wild-type BMDMs showed a higher phagocytosis ability ( Fig 4C and 4D ) than Clec4d-/- BMDMs . Therefore , Dectin-3 may have an important role in the initial fungal killing process and may prevent the phagocytosed C . tropicalis from escaping from phagosomes . Our previous data has proved that trehalose 6 , 6'-dimycolate ( TDM ) -induced Mincle expression is dependent on Dectin-3-mediated NF-κB activation via CARD9-BCL10- MALT1 complex [23] . This result prompted us to investigate whether Dectin-3 contributes to the NF-κB activation following challenge with C . tropicalis . Indeed , both hyphae and yeast form of C . tropicalis stimulation could effectively induce NF-κB activation in wild-type BMDMs ( S5 Fig ) , and it was significantly defective in Clec4d-/- BMDMs ( Fig 5A ) . Consistently , IκBα phosphorylation and degradation were partly defective in Clec4d-/- macrophages upon C . tropicalis stimulation ( Fig 5B ) . Previous studies have shown that CLRs recognize fungi and induce inflammatory responses through the adaptor protein CARD9 [21 , 26] , and CARD9 forms a complex with Bcl10 following C . albicans stimulation [27] . We then examined the inducible CARD9/Bcl10 complex formation upon C . tropicalis stimulation , and found that the formation of this complex was defective in Clec4d-/- BMDMs following the stimulation of C . tropicalis ( Fig 5C ) . The expression levels of some NF-κB regulated genes , including IL-6 , TNF-α , and IL-10 , were markedly lower in Clec4d-/- BMDMs than in wild-type BMDMs ( Fig 5D ) . Together , these data indicated that Dectin-3 plays a crucial role in NF-κB mediated inflammatory response to C . tropicalis infection . In our DSS-induced colitis model , we found that body weight loss did not differ among wild-type and Clec4d-/- mice during DSS treatment stage but did during recovery from treatment ( water supplementation stage ) . Moreover , the defect in IL-6 expression observed in Clec4d-/- mice also may play a role in impaired epithelial restitution . These data suggested that Dectin-3 plays a role in intestinal healing . Therefore , we performed histological analysis of colon in the different stages of our fungal supplementation experiments . We found that both wild-type and Clec4d-/- mice had colon epithelial cell damage with DSS treatment . After recovery from treatment , most of the epithelial cells in the colons of wild-type mice were repaired while Clec4d-/- mice had impaired tissue repair ( Fig 6A ) . Examination of colons of these mice revealed that more fungi invaded damaged tissue in DSS-treated Clec4d-/- mice than in wild-type mice ( Fig 6B ) . Furthermore , we performed a wound-healing assay to examine the tissue repair function in a normal colon epithelial cell NCM460 in vitro . We obtained BMDMs from wild-type and Clec4d-/- mice and stimulated the cells with C . tropicalis for 12 hours . We then collected BMDMs supernatants and added them to NCM460 cells . After 12 h , NCM460 cells cultured with Clec4d-/- BMDMs supernatant had decreased cell migration as determined using the wound-healing assay ( Fig 6C and 6D ) and decreased expression of p-STAT3 ( S6A Fig ) . We also found markedly lower p-STAT3 and Mcl-1 and higher cleaved caspase-3 and Noxa expressions in colon tissues obtained from Clec4d-/- mice than in that obtained from wild-type mice ( Fig 6E ) . To determine the nature of the cytokine responsible for tissue repair , IL-6 antibody was added into BMDMs supernatant . As shown in S6A and S6B Fig , IL-6 blocking in WT BMDMs supernatant can inhibit cell migration and down-regulated the expression of p-STAT3 in NCM460 cells . This data suggest that defective activation of NF-κB and IL-6 production lead to an impaired tissue repair of colon tissues and a more severe colitis in Clec4d-/- mice . To determine whether an altered fungal burden contributes to colitis severity in the absence of Dectin-3 expression , we suppressed fungal growth in mice via treatment with fluconazole as outlined ( Fig 7A ) . Fluconazole treatment could significantly inhibit the proliferation of fungus in Clec4d-/- colitis mice ( S7A Fig ) . Although fluconazole treatment could slightly increase bacteria burden in both wild-type and Clec4d-/- mice , no significant difference of bacteria burden was found between wild-type and Clec4d-/- mice ( S7B Fig ) . The treatment led to reduced weight loss ( Fig 7B ) and lower clinical and histological scores in Clec4d-/- mice ( Fig 7C–7E ) . We also checked the IL-6 expression level and found no difference between wild-type and Clec4d-/- mice upon fluconazole treatment ( Fig 7F ) . Taken together , these results further support the conclusion that an inability to control fungi in the gut leads to more severe colitis in Dectin-3-deficient mice . In summary , we proposed our working model as follows ( Fig 8 ) : C . tropicalis is a type of commensal fungus that does not induce colitis under normal situations . After tissue damage induced by DSS , C . tropicalis translocated to the LP and activated NF-κB signaling via Dectin-3 and CARD9 , triggering anti-fungal innate immune responses . In Clec4d-/- mice , Dectin-3-dependent NF-κB activation was defective , and IL-6 production was decreased . Loss of these innate immune effector molecules impaired tissue repair , leading to increased microbial translocation and chronic stimulation of mononuclear cells , which exacerbated the vicious cycle of colitis . Genetic variants that confer susceptibility to IBD in humans highlight the importance of innate immune interactions with intestinal microbiota in both initiating and controlling inflammation . Commensal and pathogenic microorganisms are recognized according to conservation of molecular patterns by pattern-recognition receptors . Herein we describe for the first time that Dectin-3 can recognize C . tropicalis and is involved in the pathogenesis of colitis . We observed several important findings . First , C . tropicalis is an opportunistic pathogen , and its burden is specifically increased in Clec4d-/- mice during induction of colitis . Second , C . tropicalis can induce NF-κB activation and cytokine production via Dectin-3 signaling . Third , Clec4d-/- mice is more susceptible to DSS-induced colitis than wild-type mice , and C . tropicalis aggravates the development of colitis . In mammals , the gastrointestinal tract is colonized by a wide range of microorganisms . Colonization by some commensal or pathogenic microorganisms can be detrimental , leading to infectious diseases . Different commensal microorganisms do not necessarily share the same mechanisms of IBD induction . For example , 129S/SvEv IL10-/- mice associated with either Escherichia coli or Enterococcus faecalis have different clinical signs of IBD [28] . In our Dectin-3 knockout mice , increased burden of C . tropicalis was the main cause of severe colitis . In the data by Iliev et al . , it was also shown that C . tropicalis was the dominating species in Dectin-1-deficient mice upon DSS treatment [9] . Transplantation of feces from wild-type to Dectin-1 deficient mice did not reduce symptom severity implying that disease severity was host-mediated rather than owing to microbe dysbiosis . In our study , the basal level of total fungal burden in feces was similar between Clec4d-/- and wild-type mice . But after DSS-treatment , the total fungal burden in the colon was markedly higher in Clec4d-/- mice than that in wild-type mice . These results suggest that the disease phenotype in Clec4d-/- mice is affected by the genotype of the mouse , not by initial differences in microbe . However , the role of Dectin-3 in fungal defense is not specific to C . tropicalis . Our previous study has proved that Dectin-3 can form a heterodimer with Dectin-2 and recognize C . albicans hyphae . Here we focus on C . tropicalis due to the dominate role of C . tropicalis in gut . We did not try gavaging C . albicans or C . glabrata in our study . Together , these data suggest that CLR ( both Dectin-1 and Dectin-3 ) deficiency leads to altered immunity to commensal fungi in the gut . The unique microbial environment of the intestines makes the innate immune system central to intestinal homeostasis . This system is not simply a host-defense mechanism against invading pathogens , as it also modulates microbial killing and affects IEC proliferation , differentiation , and survival . A balance between cell death and survival is important for the maintenance of intestinal homeostasis . NF-κB , a master transcriptional regulator that is activated by various cytokine and pattern-recognition receptors , controls the expression of pro-inflammatory mediators and enhances the survival of cells by inducing the expression of anti-apoptotic genes during colonic inflammation [29] . In mice deficient in the NF-κB p50 subunit , the colonic inflammation becomes persistent [30] . Similarly , mice with IEC-specific deletion of the NF-κB component RELA exhibit increased susceptibility to chemically induced colitis [31] . In the present study , we found that after C . tropicalis stimulation , induction of the expression of pro-inflammatory cytokines , such as IL-6 and TNF-α , was defective in Clec4d-/- mice but not in WT mice . IL-6 produced by immune cells is a key cytokine in antifungal immunity and tissue repair [25] , leading to induction of Th17 cell differentiation . Th17 cells are T helper cells that were characterized relatively recently and play major roles in host defense against fungal infections [32] . In consistent , we also found decreased numbers of Th17 cells in the colons and mesenteric lymph nodes of mice . After epithelial damage , several pathways function in a coordinated manner to restore homeostasis . Cytokines and chemokines are secreted by epithelial and immune cells , which recruit more immune cells to the site of injury and induce cellular proliferation . Specifically , the inflammasome/caspase 1/IL-18/IL-18 receptor/Myd88 axis mediates tissue repair in the intestines [33] . IL-18 binds to the IL-18 receptor , which is expressed by myeloid cells in the lamina propria , and signals through the adaptor Myd88 . If this innate immune signaling pathway is impaired ( as observed in mice deficient in caspase 1 , NLRP3 , IL-18 , IL-18 receptor , or Myd88 ) , persistent tissue damage leads to translocation of commensal microorganisms to the sub-mucosa , where they stimulate immune cells . Secretion of cytokines by activated immune cells results in IEC apoptosis and chronic intestinal inflammation . Studies by Grivennikov et al have proved that IL-6 protects the intestinal epithelium from injury by regulating intestinal trefoil factor and/or AMP secretion [25] . Therefore , the defect in IL-6 expression observed in Dectin-3 deficient mice also may play a role in impaired epithelial restitution . In the present study , we found that Dectin-3–deficient mice had impaired healing of epithelial wounds , which is due to the defective activation of NF-κb and less production of IL-6 , indicating an inherent defect in restitution of colon epithelial barrier . Dectin-1 and Dectin-3 , both belongs to CLRs group , have different structures and ligand spectrum . Dectin-1 contains immunoreceptor tyrosine-based activation motif ( ITAM ) -like motif in the cytoplasmic portion . Dectin-1 was proved to be a β-glucan receptor . It can sense C . albicans , P . carinii , Leishmania infantum , Coccidioides posadasii , Histoplasma capsulatum and Mycobacterium spp . Dectin-3 does not have any signaling motif in their cytoplasmic domains and , instead , recruit the ITAM-containing adaptor molecule FcRg to transduce signals . The CRD domain of Dectin-3 is also atypical , because it lacks conserved triple motif essential for Ca2+-dependent carbohydrate recognition . Dectin-3 has a similar ligand spectrum as Mincle in pathogen recognition and recognizes pathogens with high-mannose type and TDM . Dectin-3-deficient mice are more sensitive to systemic C . albicans infection than wild-type mice and develop milder inflammation upon immunization with TDM . Dectin-3-deficient mice are also highly susceptible to Klebsiella pneumonia infection and die from septic shock . Both Dectin-1 and Dectin-3 mediated carbohydrate recognition induces phagocytosis of pathogens , NF-kB activation , and proinflammatory cytokine production in macrophages . As to IBD , both Dectin-1 and Dectin-3 plays an important role in the maintenance of the intestinal microbe . A polymorphism in the gene of Dectin-1 was identified to be strongly linked with a severe form of ulcerative colitis in humans . However , there is no association between human Dectin-3 and IBD based on currently available GWAS databases . Therefore , a further analysis and identification of Dectin-3 mutations in IBD patients would provide a molecular basis to apply anti-fungal treatment as a potential new therapy for some IBD patients . Actually , we are collecting human sample from colitis patients , we will confirm whether genetic variation in Dectin-3 influences susceptibility of IBD in our future experiment . Animal care and experimental protocols were in accordance with the NIH “Guide for the Care and Use of the laboratory Animals” . All animal experiments and procedures were conducted under the protocol and were approved by the Institutional Animal Care and Use Committee at The University of Texas MD Anderson Cancer Center ( Protocol Number 00000911-RN00 ) . Clec4d+/- mice were obtained from the NIH-supported Mutant Mouse Regional Resource Centers ( http://www . mmrrc . org ) . Clec4d-/- mice generated as previously described [12] were crossed 5 generations onto C57BL/6J background ( 96 . 88% ) . Progeny homozygous for Clec4d-/- and Clec4d+/- ( wild-type ) mice with the same genetic background were bred separately for the experiments , and 8–10 weeks female mice were used . All animals were housed in modified barrier facility at the University of Texas MD Anderson Cancer Center . The C . tropicalis strain ( W4162870 ) is recovered from a patient with candidemia , and kindly provided by Dr . Sarah L . Gaffen ( University of Pittsburgh , PA ) . A single C . tropicalis colony was grown overnight at 30°C in yeast peptone dextrose medium . For preparation of the fungal hyphal form , C . tropicalis was washed , resuspended in complete RPMI 1640 medium , and grown for 3 h . The GFP- C . tropicalis strain is kindly provided by Dr . Richard Bennett ( Brown University , RI ) . For DSS-induced colitis model , wild-type and Clec4d-/- mice were given drinking water supplemented with 2 . 5% DSS ( MP Biomedicals ) for 7 days and then given water for an additional 4 days . For our fungal supplementation experiment , before and upon colitis induction , mice were given four doses of C . tropicalis ( 1 × 108 yeast/mouse/dose ) every other day . For our fungal ablation experiments , mice were given fluconazole ( 0 . 5 mg/ml; Sigma-Aldrich ) in drinking water 4 days prior and throughout the DSS and water stages for a total of 17 days . Body weight , stool consistency , and gross blood were checked daily . After mice were sacrificed , their colon lengths were measured . Paraffin-embedded colon tissue samples were sectioned and stained with hematoxylin and eosin ( H&E ) at the Research Histology Facility at MD Anderson . Colitis severity was assessed by a blinded pathologist using clinical and pathological scores as described previously [34] . Briefly , clinical scores were calculated based on weight loss , stool consistency and occult blood as follows: weight loss: 0 ( 0–5% ) , 1 ( 5–10% ) , 2 ( 10–20% ) , and 3 ( >20% ) ; stool consistency: 0 ( normal ) , 1 ( soft but still formed ) , 2 ( very soft ) , and 3 ( diarrhea ) ; occult blood: 0 ( negative hemoccult ) , 1 ( positive hemoccult ) , 2 ( blood traces in stool visible ) , and 3 ( rectal bleeding ) . Scoring system for inflammation-associated histological changes in the colon is: 0 ( no evidence of inflammation ) , 1 ( low level of inflammation with scattered infiltrating mononuclear cells , 1–2 foci ) , 2 ( moderate inflammation with multiple foci ) , 3 ( high level of inflammation with increased vascular density and marked wall thickening ) , 4 ( maximal severity of inflammation with transmural leukocyte infiltration and loss of goblet cells ) . Colonic LP cells were isolated from the study mice as described previously [35] with some modifications . Briefly , colons were isolated , resected , opened longitudinally , washed , and cut into pieces . Intestinal pieces were incubated in a digestion medium consisting of RPMI 1640 , 5% FBS , 1 . 5 mg/ml collagenase type IV ( Sigma-Aldrich ) , 5 U/ml DNase ( Roche Diagnostics ) , and 1% penicillin-streptomycin for 30 min at 37°C with gentle shaking . The cell suspensions were filtered through a mesh and then centrifuged at 1300 rpm . LP cells were used for flow cytometry , western blot , and cytokine analysis . For surface staining , cells were washed and stained with fluorescent-conjugated antibodies for 20 minutes at 4°C . For intracellular cytokine staining , cells were incubated for 5 hours at 37°C with 50 ng/mL phorbol myristate acetate ( Sigma-Aldrich ) , 1 mmol/L ionomycin ( Sigma-Aldrich ) , and 1 mL/ mL GolgiPlug ( BD Biosciences ) . Surface staining was performed followed by intracellular staining using the BD Cytofix/Cytoperm Kit ( BD Biosciences ) . The following antibodies were used for our analysis ( BD Pharmingen ) : F4/80 ( #552958 ) , CD11c ( #550261 ) , CD4 ( #553729 ) , IL-17A ( #560438 ) , IFN-γ ( #561040 ) CD25 ( #561038 ) and FoxP-3 ( #560408 ) . Fluorescently labeled cells were acquired on a FACS Calibur flow cytometer ( BD Biosciences ) and analyzed using FlowJo Analysis Software ( Tree Star , Inc , Ashland , OR ) . Feces was collected from study mice and suspended in 50 mM Tris buffer ( pH 7 . 5 ) containing 1 mM EDTA , 0 . 2% β-mercaptoethanol ( Sigma-Aldrich ) , and 1000 U/ml lyticase ( Sigma-Aldrich ) . The mixture was incubated at 37°C for 30 min , and fungal genomic DNA was isolated from feces and colons using a QIAamp DNA Stool Mini Kit ( QIAGEN ) according to the manufacturer's instructions . For evaluation of fungal rDNA in feces , 100 ng of total fecal DNA was used as a template and fungal 18S rDNA was evaluated using quantitative PCR analysis . For detection of specific fungi , qPCR was performed in genomic DNA using fungal-specific primers . Quantitative real-time PCR was performed using SYBR Green with an ABI StepOnePlus system ( Life Technologies ) with the following primers: 18S rDNA ( ATTGGAGGGCAAGTCTGGTG; CCGATCCCTAGTCGGCATAG ) , C . albicans ( CTGTTTGAGCGTCGTTTC; ATGCTTAAGTTCAGCGGGTAG ) , C . tropicalis ( TTTGGTGGCGGGAGCAATCCT; CGATGCGAGAACCAAGAGATCCGT ) , and C . glabrata ( CTGCGCTTAACTGCGCGGTT; TGCGAGAACCAAGAGATCCGTT GC ) . The total fungal burden was calculated by the ΔCt method and normalized to the weight of the fecal samples and the amount of total DNA used . Relative quantity of the specific fungal burden was also calculated by the ΔCt method and normalized to the weight of the fecal samples . The proportion of specific fungi was the ratio of specific fungal burden to total fungal burden . For in vivo fungal staining , embedded intestinal specimens were sectioned , mounted on microscope slides , and incubated for 40 min in PBS containing 2% FCS . Intestinal sections were stained with an anti-GFP antibody ( ab13970; Abcam ) . Slides were rinsed with PBS and stained for 5 min with 0 . 1 μg/ml DAPI ( Invitrogen ) and overlaid with a mounting medium ( VECTASHIELD; Vector Laboratories ) . Slides were examined using a Zeiss Axio Observer fluorescence microscope . All compared images were collected and processed identically . Primary cultures of BMDMs obtained from mice were prepared and purity of macrophages was confirmed using flow cytometry as described previously[27] . Briefly , bone marrow cells were harvested from the femurs and tibias of mice . Erythrocytes were then removed from the cells using a hypotonic solution . Cells were cultured for 7 days in DMEM containing 30% conditioned medium from L929 cells . GFP-C . tropicalis ( 5×106 ) was reuspended in RPMI 1640 supplemented with 5% fetal bovine serum ( FBS ) and added onto 1×106 wild-type and Clec4d−/− BMDMs , and incubated at 37°C in a 5% CO2 incubator for 1 hour . Wells were washed and fresh media containing fluconazole ( 300 μg/ml ) was added . At 6 hours and 18 hours , BMDMs were washed three times with PBS , lysed in water , and C . tropicalis CFU were photographed and calculated by plating on YPD agar . For phagocytic ability of BMDM experiment , flow cytometry of the GFP fluorescence of wild-type and Clec4d−/− BMDMs infected with GFP–C . tropicalis ( MOI , 5 ) for 0 , 0 . 5h , and 1h , respectively , then washed extensively with cold PBS and fixed with 2% paraformaldehyde . GFP+ BMDMs were calculated . BMDMs were stimulated with C . tropicalis or LPS , and nuclear extracts of BMDMs were prepared . Five micrograms of the resulting nuclear protein was incubated with a 32P-labeled NF-κB ( E3291 ) or Oct-1 ( E3241 ) probe ( Promega ) for 15 min at room temperature and then subjected to PAGE and exposed to x-ray film . Cell lysates from BMDMs were immunoprecipitated with indicated antibody-conjugated agarose . The resulting immunoprecipitates and lysates were subjected to SDS-PAGE and then blotted using indicated antibodies . Phosphorylated IκB kinase α/β ( #2697 ) , extracellular signal-regulated kinase 1/2 ( #9101 ) , P38 ( #4631 ) , IκBα ( #9246 ) , and STAT3 ( ##9145 ) ; cleaved caspase 3 ( #9664 ) ; and P38 ( #9212 ) were purchased from Cell Signaling Technology . Antibodies against IκB kinase α ( sc-7218 ) , IκBα ( sc-371 ) , Noxa ( sc-30209 ) , and actin ( sc-8432 ) were purchased from Santa Cruz Biotechnology . An anti-CARD9 antibody was purchased from Sigma ( C7862 ) . BMDMs were stimulated with C . tropicalis for 12 h , and BMDM supernatants were collected after stimulation . The ELISA kits for TNF , IL-6 , IL-10 , MIP-2 , IL-1β , IL-17a and CXCL1 were purchased from eBioscience . All the supernatant samples were measured in triplicate according to the ebioscience manufacturer’s protocol . Wounded-monolayer NCM460 cells were washed two or three times to remove detached cells . The initial size of the wound on the monolayer was determined using inverted microscopy immediately after the cells were washed . After 6 and 12 h of incubation in the BMDMs supernatants stimulated with C . tropicalis , wound closure was calculated as the percentage of the remaining initial wound area . An unpaired Student t-test was used to evaluate differences between experimental groups . Statistical analysis was performed using the Prism software program ( version 5 . 0; GraphPad Software ) . P<0 . 05 was defined as statistical significance .
C-type lectin receptors ( CLRs ) comprise a diverse family of soluble and trans-membrane proteins that function as pattern recognition receptors ( PRRs ) . Dectin-3 ( also known as MCL/CLECSF8/Clec4d ) , a myeloid cell-specific CLR family member , could recognize bacterial and fungal components and induce intracellular signaling pathways that regulate the immune response . Although investigators have explored the role of Dectin-3 in systemic immunity , its function in the gastrointestinal immune system is not clear . Using a dextran sodium sulfate ( DSS ) -induced colitis mice model , we show here Dectin-3-deficient mice were more susceptible to DSS-induced colitis compared with wild-type mice . The specific fungal burden of a commensal fungi C . tropicalis was markedly increased in the gut after DSS treatment in Dectin-3-deficient mice , and antifungal therapy could effectively protect these mice from colitis . Taken together , we demonstrate the important function of Dectin-3 and its functional interaction with commensal fungi in intestinal immune responses and regulation of colonic homeostasis .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "blood", "cells", "innate", "immune", "system", "medicine", "and", "health", "sciences", "immune", "physiology", "cytokines", "immune", "cells", "pathology", "and", "laboratory", "medicine", "immunology", "colitis", "developmental", "biology", "fungi", "signs", "and", "symptoms", "gastroenterology", "and", "hepatology", "molecular", "development", "inflammatory", "bowel", "disease", "digestive", "system", "white", "blood", "cells", "inflammation", "animal", "cells", "immune", "response", "immune", "system", "gastrointestinal", "tract", "diagnostic", "medicine", "anatomy", "cell", "biology", "physiology", "biology", "and", "life", "sciences", "cellular", "types", "macrophages", "colon", "organisms" ]
2016
Dectin-3 Deficiency Promotes Colitis Development due to Impaired Antifungal Innate Immune Responses in the Gut
Visceral leishmaniasis belongs to the list of neglected tropical diseases and is considered a public health problem worldwide . Spatial correlation between the occurrence of the disease in humans and high rates of canine infection suggests that in the presence of the vector , canine visceral leishmaniasis is the key factor for triggering transmission to humans . Despite the control strategies implemented , such as the sacrifice of infected dogs being put down , the incidence of American visceral leishmaniasis remains high in many Latin American countries . Mathematical models were developed to describe the transmission dynamics of canine leishmaniasis and its control by culling . Using these models , imperfect control scenarios were implemented to verify the possible factors which alter the effectiveness of controlling this disease in practice . A long-term continuous program targeting both asymptomatic and symptomatic dogs should be effective in controlling canine leishmaniasis in areas of low to moderate transmission ( R0 up to 1 . 4 ) . However , the indiscriminate sacrifice of asymptomatic dogs with positive diagnosis may jeopardize the effectiveness of the control program , if tests with low specificity are used , increasing the chance of generating outrage in the population , and leading to lower adherence to the program . Therefore , culling must be planned accurately and implemented responsibly and never as a mechanical measure in large scale . In areas with higher transmission , culling alone is not an effective control strategy . Visceral leishmaniasis or kala-azar is the most severe clinical form of leishmaniasis , a serious public health problem worldwide [1] , [2] . In Latin America , the agent of the so-called American visceral leishmaniasis ( AVL ) is Leishmania ( Leishmania ) chagasi transmitted , in Brazil , mainly by sandfly Lutzomyia longipalpis [Lutz & Neiva , 1912] . So far , the findings related to the epidemiology of AVL point to a spatial correlation between the occurrence of disease in humans and high rates of infection in dogs , suggesting that , in the presence of the vector , canine visceral leishmaniasis is a key factor for triggering transmission to humans [3] . Overall , the incidence of AVL remains high despite the large-scale control strategies that have been implemented . These strategies focus on early diagnosis and treatment of human cases , vector control to reduce sandfly population , as well as the removal of infected dogs and health education [4] . Despite the lack of solid evidence in literature , culling dogs with canine visceral leishmaniasis ( CVL ) has been the major strategy for controlling this disease in Brazil . Many authors argue that this strategy has low cost-benefit and many are against it , often encouraging the non-delivery of animals to slaughter [5] , [6] , [7] , [8] , [9] . Other professionals , however , admit that this strategy can produce positive results [6] , [10] , [11] . Two possible factors associated with the low effectiveness of culling programs are: ( 1 ) the discontinuity of these programs , which may occur for several reasons , including the lack of a structured surveillance system , budget issues and lack of adequately trained professionals; ( 2 ) Problems related to the logistics in delivering control measures , for example , low infected dog screening rates and lack of a reliable and valid diagnostic test to detect dogs in the early stages of infection , leaving out asymptomatic infectious dogs that are capable of conveying the parasite to the vectors , thus , cooperating with the continuity of the transmission cycle [12] , [13] . Mathematical modeling has been applied in studies of visceral leishmaniasis in order to understand the transmission dynamics of this infection [8] , [14] , [15] , [16] , [17] , [18] , [19] and the impact of control strategies . Hasibeder et al . ( 1992 ) and Dye et al . ( 1992 ) proposed and implemented models to estimate the basic reproduction number ( R0 ) of CVL , which was estimated between 1 . 44 and 11 , this large uncertainty being attributed to the poor performance of the available diagnostic tests . Their model predicts that in areas where R0 is at the upper limit of the R0 range , culling would be successful only if intensively implemented . In real settings , however , R0 estimation is highly uncertain as it depends on how the seropositivity is measured , and on the many assumptions of the underlying model , such as the homogeneous exposure of dogs to sandflies and homogeneous response to infection . Later , Dye ( 1996 ) [20] alerted that culled dogs tend to be rapidly substituted by younger and susceptible ones , reducing the effectiveness of this strategy , compared to alternatives such as vector control , drugs and vaccines . Other studies have explored the effectiveness of imperfect control programs , assessing the effect of imperfect diagnostic tests 24 , and discontinued dog culling programs [15] . They found that a high sensitivity test , together with the immediate sacrifice , was sufficient to control the disease . On the other hand , with a low sensitivity test , the effectiveness of the program was lost , whether or not the dogs were sacrificed immediately . This paper seeks to revisit this problem , focusing on the relevance of asymptomatic infections in a scenario of imperfect control characterized by sub-optimal screening , diagnosis and slaughter rates . We further investigated an unexplored component that is the impact of the low specificity of the diagnostic test . We hope to contribute to the understanding of CVL transmission dynamics and the factors that modulate the control effectiveness . An expression for the basic reproduction number of CVL was derived from the SEI2D model , without control , using the next generation matrix method [24] . The mathematical derivation is found in the appendix ( Text S1 ) . The Basic Reproduction Number is: ( 9 ) For modeling purposes , the intervention program was divided into three components: screening , diagnosis and sacrifice . Screening measures the monthly capture rate and application of the diagnostic test to dogs in the population . ( 10 ) The parameter d represents the probability of a dog to be positively diagnosed given it has been subjected to a diagnostic test and is infected ( test sensitivity ) . ( 11 ) Once positively diagnosed , the dog has a chance f of being put down . The delay between the screening and the sacrifice is 1/u . The product r x dz measures the rate of misclassification of uninfected dogs . This rate depends on r ( screening rate ) and dz which corresponds to the test's probability of false positive ( 1- specificity ) . By varying these parameters , r , d , dz , u e f , one can investigate the impact of a variety of imperfect control programs . Here , we considered variations of two hypothetical programs , both of which have been continuously implemented for 40 years . The first one was based on data from the CVL control program implemented in Belo Horizonte , Brazil , considered to be of good quality , within the possibilities of the country ( data provided by the Subcoordenation of Vector Transmitted Zoonosis and Rabies/SVS/MS ) . In this program , the screening rate is 6% per month followed by the immediate sacrifice of 85% of the screened dogs with positive diagnosis . We implemented this scenario , assuming a diagnostic test with 90% sensitivity and 100% specificity . A second scenario was built representing a worse situation , in which the screening rate is 4% per month and time to culling is 4 months as in Courtenay et al . ( 2002 ) . In this scenario , diagnostic tests were applied with sensitivity and specificity ranging from 80% to 100% . In both programs , we investigate two protocols: one targeting exclusively symptomatic dogs and screening all dogs , regardless of the presence or absence of symptoms . To investigate the impact of diagnostic tests with low specificity , we compared the number of erroneously culled dogs by programs using tests with specificity of 80 , 90 and 100% . By quantifying the number of dogs that were needlessly put down , we have a measure of the negative impact of the control strategy . The effectiveness of the control programs was assessed by comparing the prevalence before and after 40 years of the establishment of the Control Program . The control program was considered successful if it were capable of reducing CVL prevalence below 1% . Considering that prevalence is measured by imperfect diagnostic methods , we further distinguished between real success and perceived success . Real success is achieved when the true prevalence decreases below 1% , while perceived success is achieved when the measured prevalence decreases below this threshold . At last , to investigate the impact of uncertainties in the specification of model parameters in the success of the control programs , we performed a multivariate uncertainty and sensitivity analysis . The procedure was as follows: First , uniform probability density functions were defined for each life-history parameter ( i , qa , qb , p , λ , a , μ ) with intervals equal to 0 . 75 and 1 . 25 times the default parameter value . Secondly , one thousand values were draw from each of these distributions , producing 1000 sets of parameters . To maintain the transmission constant , for each new set of parameters , β was calculated from the R0 equation so that the R0 of all the simulations was kept at the same level . After running the model SEI2D with each set of parameters , we recorded the success of the control program after 40 years as positive if final prevalence was less than 1% and failure otherwise . Using the expression of R0 derived from the SEI2D model and the parameter values presented in Table 1 , we obtained R0 = 1 . 09 for the low endemicity scenario and R0 = 1 . 29 for the high endemicity scenario . These values are low compared with those reported by other authors but were based on prevalence observed in the field [21] , [22] . In the sensitivity analysis section , we discuss scenarios with higher R0 . According to the SEI2D model , a CVL control program with 6% monthly screening rate , a diagnostic test with 90% sensitivity , and no delay between screening and culling should be effective in controlling CVL if implemented continuously for 40 years , that is , prevalence is reduced below 1% . In the low endemicity area , the success is reached by only targeting symptomatic dogs . Under slightly higher transmission , however , successful control requires the sacrifice of symptomatic and asymptomatic dogs . That is , limiting the intervention to clinically positive dogs was not sufficient to control the disease below the 1% prevalence level , leaving it at 2% instead . When screening is reduced to 4% per month , a less sensitive test is used ( 80% ) and elimination time increases to an average of four months , the good performance of the control program targeting symptomatic dogs only is still preserved in the low endemicity area , with final prevalence reaching values below 1% . As transmission increases , targeting just symptomatic dogs becomes no longer effective , resulting in final prevalence of 6% . To ensure prevalence below 1% , at least 30% of the asymptomatic dogs should be put down ( results not shown ) . Figure 2 shows that , if control targets both symptomatic and asymptomatic dogs , the impact of improving the sensitivity from 80 to 90% is negligible . On the other hand , in a program targeting symptomatic dogs only , improving the test's sensitivity to 90% is very advantageous to improve its effectiveness . To further investigate the relevance of asymptomatic dogs on control , we parameterized the model once again , assuming that all asymptomatic dogs were non-infectious , but still positive for the diagnostic tests . These individuals are the dogs considered cured according to Lanotte et al . ( 1979 ) [25] . In this case , their elimination has no effect on the success of the control program . The low and moderate endemicity scenarios simulated here are in the low range of the estimated values for R0 . The performance of imperfect culling programs in an area with extremely high transmission rate , corresponding to R0 = 9 , was evaluated and , in this case , none of the culling strategies were effective ( Table 2 ) . Actually , the R0 threshold under which CVL is controlled is R0 = 1 . 41 for programs targeting any seropositive dog . A maximum R0 = 1 . 106 is required for the success of programs targeting clinically positive dogs only . One of the main arguments against culling programs is the unnecessary sacrifice of healthy dogs that are erroneously diagnosed , leading to speeches against culling , which reduces the number of animals delivered to zoonosis centers , increasing the ethical and social costs of this strategy . Here , we calculated the number of unnecessarily sacrificed dogs in a program using a diagnostic test with 80% sensitivity and either 80% and 90% specificity , during the five years of the control application ( years 35 to 40 after control implementation ) . In the high endemicity area , a program using a test with 80% specificity , this number was 5821 animals , which corresponds to 38% of all dogs put down . Increasing specificity to 90% , only a slight reduction was obtained . However , restricting culling to symptomatic dogs only is not sufficient to control the disease below the 1% prevalence level . This result poses a dilemma to control programs in high endemicity areas as the success of culling is only achieved if asymptomatic dogs are included and this is done at the expense of putting down non-infected dogs . In areas with low endemicity , on the other hand , restricting culling to symptomatic dogs can both control the disease and reduce the risk of putting down healthy animals . Figure 3 shows the proportion of the parameter space – corresponding to a variety of natural history situations – that were controlled by culling programs with screening rates equal to 4 , 6 or 8% , and test's sensitivity equal to 80 or 90% . It is clear that the success of the culling programs is highly dependent on the transmission rate and that increasing screening effort is required in areas with high transmission . Moreover , it is clear that increasing screening effort is more effective than increasing the sensitivity of the diagnostic tests from 80 to 90% . However , one must consider the costs associated with such effort for a routine program . Figure 4 shows the life-history parameters associated with the success or failure of the control program that targeted asymptomatic and symptomatic dogs , with 0 . 04% screening effort in an area with R0 = 1 . 41 . The most important parameters refer to the asymptomatic population . In summary , the higher the proportion of dogs becoming or remaining asymptomatic , the most effective the program is . The reason is the lower transmissibility of these asymptomatic dogs . This study aimed at assessing the effectiveness of culling dogs in the control of canine visceral leishmaniasis in scenarios where implementation occurs imperfectly . This investigation was based on a mathematical model for CVL that introduces a class of infectious asymptomatic dogs which contributes , at a lower rate , to the transmission cycle [15] . This model differs from previous models , in which asymptomatic dogs are assumed to be uninfectious [7] , which may be true for European dogs that are well nourished [26] but not necessarily for all dog populations . The infectiousness and proportion of asymptomatic dogs had strong impact on the success of control strategies . As a matter of comparison , we also simulated a simple SI model as parameterized for CVL . This is in line with part of CVL modeling literature using SIR-like models [16] , [27] . Overall , when compared with the SEI2D model , SID generates more optimistic expectations , with successful control being reached at faster rates . Most researchers agree that the sensitivity , specificity and reproducibility of the available serological tests are substandard [6] , [11] , [15] , [16] , [25] , [28] , [29] , [30] . Sensitivity depends on the methodology used , and the specificity varies with the choice of the antigen . Low sensitivity increases the chance of permanence of false-negative animals in the environment [12] . An aggravating issue in the permanence of asymptomatic dogs is the difficulty of tracking these dogs , turning them into a silent reservoir [12] . The main result of our simulations is that , in areas with very low transmission ( baseline prevalence of 3% ) , culling of symptomatic dogs by a realistic program with 4% screening and testing per month and a mean delay to culling of four months , is sufficient to maintain prevalence under 1% , which we considered a successful endpoint . The advantage of this program is the focus on symptomatic dogs only , what reduces the burden of killing apparently healthy dogs , providing a better grip of the program by the population . However , the control program was successful in interrupting transmission of CVL in areas of low transmission , possibly because the endemic equilibrium in these simulations was fragile , and a simple disturbance in the system lead to R0<1 . In areas with slightly higher endemicity ( R0 = 1 . 29 , prevalence = 15% ) , on the other hand , removing clinically diagnosed dogs is not sufficient as a control strategy because the asymptomatic population is large enough to maintain transmission . This is in accordance with many studies [12] , [13] , [15] , [31] . In this case , a program would have to be capable of including at least 30% of the asymptomatic but infectious dog population in order to maintain infection prevalence below 1% . A further complication of targeting asymptomatic dogs is the increased chance of putting down healthy dogs as the diagnostic tests available have low specificity . This is a serious problem in areas with lower endemicity , where the positive predictive values of the tests tend to be low . The models studied here suggest that in the simulated area , 79% of dogs would be wrongly eliminated by tests with 80% specificity . The unnecessary sacrifice of non-infected dogs burdens the program and feeds the discourse against dog culling and increases society's aversion to the control program [12] . The emotional onus and social cost of euthanizing dogs , whether they are ill or not , must be considered in evaluating of culling dogs as a control strategy against AVL . To avoid the erroneous sacrifice of false-positive dogs , it must be ensured that the tests have high specificity reducing the social cost of this strategy . The transmission rate of CVL in real settings can be much higher than the ones simulated here [17] , [20] . As the transmission rate increases , the effectiveness of the culling program rapidly declines unless investment in screening is enhanced ( Figure 3 ) . In high transmission areas , the required effort may become too high to be feasible , and combined strategies , such as vector control , may become necessary . In any scenario , control effectiveness requires continuity , that is , no interruptions in the application of control measures . In practical terms , the inclusion of asymptomatic dogs in a control program stumbles in several difficulties: the difficulty of screening and testing these dogs , as well as their diagnosis , and convincing the delivery of apparently healthy dogs for culling [12] , [31] . An excellent program would be the one which is more efficient and less costly . A control program aimed only at symptomatic dogs has apparently lower cost than one targeting all infected dogs . However , such program by itself will not control transmission . These results overall , suggest that strategies should differ in areas with high and low transmission , with more integrated approaches being the choice in the former and culling of symptomatic dogs being a choice in the latter . We did not investigate the relative effectiveness of other strategies in the same scenario . Dye ( 1996 ) suggested that insecticide application can be more effective than culling , but this is based on the assumption that the impact of insecticides on the sandfly population is high , and resistance is absent or low . Palatnik-de-Sousa et al . ( 2004 ) argues that using diagnostic tests with greater sensitivity collaborates for the greater effectiveness of a culling program , by minimizing the percentage of false-negative dogs . However , Dye et al . ( 1993 ) assures that even if a highly efficient serological test was used , about 20% of the cases would remain undetected , especially in animals which , at the time of the test , were in the incubation or seroconversion phases . Using tests with greater sensitivity and lower specificity may incur in greater social cost and reduced efficiency due to low social acceptance . In summary , the analysis of the models suggests that besides investments on the improvement of diagnostic tests , further effort is required to improve the control program itself , considering the logistics and resources required for implementation of control for longer periods . The results of this study are limited to cases in which the model is valid . The model assumed a canine population of constant size , but it is possible that , in some contexts , these populations are actually increasing or decreasing . Another limitation of the model is not explicitly considering the dynamics of the vector . It is known from the study of other diseases such as dengue and malaria that the vectorial capacity can be affected by climate and environmental conditions , including variation from one year to another . The model also does not consider other potential hosts such as wild animals . The impact of control would be lower if these animals were present . Furthermore , the model assumes that all dogs are homogeneously exposed to the risk of vector contact . In real situations , the risk is expected to vary spatially and future studies should consider this dimension . Finally , control is implemented continuously , but in real situations this is rare . Future studies should investigate the impact of strategies applied in pulses at different times of the year . With all this , we must evaluate the results of this study with caution and by a realistic point of view , noting that the canine sacrifice was effective in controlling the CVL only in a scenario in which the control was implemented monthly and with the same effort for 40 years .
Visceral leishmaniasis is listed as a neglected tropical disease and is considered a public health problem worldwide . The disease has been documented since 1885 , the first case being reported in India . After over 120 years , the incidence of the disease remains high despite control strategies implemented . In areas where the disease is zoonotic , such as in Brazil , identification as well as removal of infected dogs is recommended in highly endemic areas for they are considered to be the reservoir of the Leishmania chagasi parasite . The theoretical basis that supports the culling of infected dogs is the assumption that the incidence of human infection is directly related to the number of infectious dogs . However , there is no consensus among researchers on the effectiveness of this strategy for controlling either human or canine visceral leishmaniasis . In this context , mathematical models can provide a basis for determining the strategies with the greatest potential for success . This paper aims to contribute to this discussion by introducing further complexities into the problem , in particular , the imperfect diagnosis of this infection and the time gap between laboratory diagnosis and culling and the presence of asymptomatic infections .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results", "Discussion" ]
[ "population", "modeling", "epidemiology", "infectious", "disease", "epidemiology", "disease", "dynamics", "population", "dynamics", "population", "biology", "infectious", "disease", "modeling", "biology", "computational", "biology" ]
2013
Culling Dogs in Scenarios of Imperfect Control: Realistic Impact on the Prevalence of Canine Visceral Leishmaniasis
Sensory neuron diversity is required for organisms to decipher complex environmental cues . In Drosophila , the olfactory environment is detected by 50 different olfactory receptor neuron ( ORN ) classes that are clustered in combinations within distinct sensilla subtypes . Each sensilla subtype houses stereotypically clustered 1–4 ORN identities that arise through asymmetric divisions from a single multipotent sensory organ precursor ( SOP ) . How each class of SOPs acquires a unique differentiation potential that accounts for ORN diversity is unknown . Previously , we reported a critical component of SOP diversification program , Rotund ( Rn ) , increases ORN diversity by generating novel developmental trajectories from existing precursors within each independent sensilla type lineages . Here , we show that Rn , along with BarH1/H2 ( Bar ) , Bric-à-brac ( Bab ) , Apterous ( Ap ) and Dachshund ( Dac ) , constitutes a transcription factor ( TF ) network that patterns the developing olfactory tissue . This network was previously shown to pattern the segmentation of the leg , which suggests that this network is functionally conserved . In antennal imaginal discs , precursors with diverse ORN differentiation potentials are selected from concentric rings defined by unique combinations of these TFs along the proximodistal axis of the developing antennal disc . The combinatorial code that demarcates each precursor field is set up by cross-regulatory interactions among different factors within the network . Modifications of this network lead to predictable changes in the diversity of sensilla subtypes and ORN pools . In light of our data , we propose a molecular map that defines each unique SOP fate . Our results highlight the importance of the early prepatterning gene regulatory network as a modulator of SOP and terminally differentiated ORN diversity . Finally , our model illustrates how conserved developmental strategies are used to generate neuronal diversity . Making sense of a complex environment requires a high level of functional diversity in neuronal classes that comprise both the peripheral and central nervous system . Little is known about how limited genetic resources are utilized to reproducibly spawn a large number of neuronal classes . Sensory systems , especially the olfactory system , are prime examples of both this neuronal diversity and how it enables organisms to survive in a complex world . The olfactory system drives behaviors fundamental to organisms’ survival , like foraging , toxin and predetor avoidance , as well as social behaviors such as courtship , aggression and parenting [1] . To detect and decifer the chemical cues shaping these behaviors , animals are equipped with a diverse array of olfactory receptors ( ORs ) that evolve rapidly [2–6] . The Drosophila olfactory system is a great model to study neuronal diversification because: ( 1 ) the organizational principle of the olfactory system is conserved across species; ( 2 ) it is a complex system with sufficient diversity that calls for sophisticated mechanisms of differentiation; yet , ( 3 ) its numerical complexity is much reduced as compared to mammals , which makes systems-level investigation possible . Adult flies have two pairs of olfactory sensory appendages: the third segment of antenna ( funiculus ) and the maxillary palp [7] . The surfaces of these olfactory organs are covered by multiporous sensory hairs , called “sensilla” . Each antenna and maxillary palp contains about 410 and 60 sensilla , respectively , that house clusters of 1–4 olfactory receptor neurons ( ORNs ) [8 , 9] . There are approximately 1300 ORNs per antenna and 130 per maxillary palp [8 , 10] . Each ORN typically expresses a single receptor gene from a repertoire of 80 genes , creating a total of 50 adult ORN classes that are clustered into stereotypical combinations within 22 individual sensilla subtypes [11] . Antennal sensilla have three major morphological types: club-shaped basiconica ( ab: antennal basiconic ) , spine-shaped trichoidea ( at ) , and cone-shaped coeloconica ( ac ) , in addition to the rare intermediate type ( ai ) [10] . Basiconic sensilla are subdivided into large , thin and small types . Each morphologically distinct sensilla type is further segmented into generally 4 or 3 sensilla subtypes , which are defined by the unique subsets of ORN classes that express invariable combinations of olfactory receptors [7 , 9 , 12] . Basiconic and trichoid sensilla contain ORNs that express conventional insect OR genes , except for two ORN classes ( Gr21a/Gr63a- and Or10a/Gr10a-expressing neurons ) in the large basiconic subtype ab1 that ( co- ) express gustatory receptors ( GRs ) [13 , 14] . Coeloconic sensilla generally contain ionotropic receptor ( IR ) -expressing ORNs [15–17] . Because of the zonal localization of sensilla types/subtypes and their defined relationships to olfactory receptor genes , the expression of a given receptor is accordingly restricted to a specific zone , and thus all ORNs collectively form a sensory map on the antenna [7 , 18–20] . Interestingly , despite the evolutionary separation between Drosophila and mammals , the principle of zonal restriction of OR expression seems to be conserved [21–23] . It is unclear , however , how different zones are generated and how they regulate the distribution and diversity of different ORN classes . In flies , the olfactory appendages develop from the antennal discs , which are specified by Distal-less ( Dll ) , Homothorax ( Hth ) and Extradenticle ( Exd ) [24 , 25] . hth is an anterior-posterior ( A/P ) homeotic selector gene that is sufficient to confer antennal identity in other tissues . Likewise , the homeotic gene antennapedia ( antp ) induces leg fate and a gustatory appendage [25–29] . The legs contain gustatory receptor neurons ( GRNs ) that sense non-volatile chemicals , and GRNs also display neuronal diversity with distinct receptor profiles [30–32] . Both the antennae and legs are ventral , segmented appendages and parallels between them have been drawn for years [26–29] . Indeed forced expression of antp can transform antennae into legs [26–28] . In either case , the proximodistal ( PD ) axis of the 3D adult tissue is constructed by the extension of the 2D sheet-like imaginal disc from the center . True joints are formed along this axis in both appendages , although the legs are more segmented [11 , 33 , 34] . In addition , the alignment between segments appears to be more linear in the leg , reflecting the “telescope-out” motion of the disc during the morphogenic event as opposed to the “fanning” motion in the antenna . Both sensilla-covered chemosensory organs ( funiculus and tarsi ) develop from the distal regions of the corresponding discs . The tarsi are further segmented , which sets natural boundaries for the position of a given GRN class . In contrast , the funiculus possesses a contiguous anatomy allowing the flow of ORN precursors within a certain range [11 , 33 , 35] . It is believed that fly antennae and legs are evolutionarily related , and some common molecules have been discovered to account for the segmental features of their tissue-level analogy along the PD axis [27 , 28 , 36 , 37] . However , how the differentiation of the cellular components , especially the complex array of chemosensory neurons housed in the antennae/legs , is coordinated with or by these morphogenic events remains a mystery . We recently reported that the rotund ( rn ) gene locus , known to control tarsal segmentation , has a critical function in diversifying ORN classes during the antennal disc development [38 , 39] . Rn is required in a subset of sensory organ precursors ( SOPs ) to confer novel sensilla subtype differentiation potentials from some default potentials within each sensilla type lineage . In rn mutants , ORNs in rn-positive sensilla subtype SOPs are converted to lineage-specific default rn-negative fates , resulting in only half of the normal ORN diversity . Through a developmental transcriptome analysis and in light of the knowledge about leg development , we found that Rn , together with BarH1/H2 ( B-H1/2 , Bar or B ) , Apterous ( Ap ) , Dachshund ( Dac ) , and Bric-à-brac ( Bab ) , is part of the conserved PD gene regulatory network module that plays a crucial role in patterning the antennal precursor field prior to proneural gene-mediated SOP selection . Interactions among these PD genes separate the developing antennal disc into seven concentric domains . Each ring is represented by a unique combination of the aforementioned transcription factors , and encodes the differentiation potentials for a limited number of sensilla subtypes . Genetic perturbations of the network lead to predictable changes in the ratios of different sensilla subtypes and corresponding ORN classes . In addition , using endogenously tagged Rn protein in vivo , we show direct binding of Rn to Bar and bab regulatory regions in the antennal disc . This same network module was previously shown to control the segmentation of tarsi in the developing leg and we show that Rn controls neuronal development of the GRNs in the leg as well . We propose a three-step mechanism to explain ORN diversification , beginning with the prepatterning of the precursor field by a gene regulatory network , followed by SOP selection by proneural genes , and Notch-mediated neurogenesis leading to terminal differentiation . The final precursor potentials are largely determined by the prepatterning phase . In our model , each step operates in a context-dependent manner: in a different context , the same transcription factor network with the same logic steps can result in completely different neuronal identity outputs . This combinatorial approach enables the same small , conserved set of genes to specify , in parallel , a broad range of chemosensory neurons . Previously , we demonstrated that the Krüppel-like transcription factor Rn cell-autonomously diversifies ORN classes by branching off novel sensilla subtype lineages from parallel default ones . In rn mutants , ORN diversity is reduced almost by half . Neurons from at4 sensilla in the trichoid zone , ac2 in the coeloconic zone , and ab1 and ab9 in the basiconic zone are all expanded at the expense of specific ORNs in rn-positive sensilla subtypes [38] . To reveal the molecular mechanism by which Rn modulates ORN precursor identities , we compared transcript abundances from a time-course RNAseq analysis in wild type ( w1118 ) , heterozygous and homozygous rn mutant flies ( see Materials and Methods ) at four temporal landmarks during antennal development . We surveyed the prepatterning ( larval ) , SOP selection ( 8hr pupal ) , neurogenesis ( 40hr pupal ) and terminally differentiated adult stages [38 , 40–44] . For the adult stage , changes in OR expression in rn mutants were consistent with the overall trend described from our previous report ( S2 Fig ) , suggesting that our experiment effectively identifies genes whose expression is affected by rn . To find key developmental genes likely acting downstream of Rn , we focused on the three early stages . As Rn is only expressed during larval and early pupal periods , we reasoned that the genes under direct Rn control would show differential expression in one or more of these early time points . A Venn diagram generated from the final lists for all early stages reveals that some genes may be misregulated only in one particular stage , while others show misregulation—both up and down—across multiple stages ( S1A and S1B Fig , also see Materials and Methods ) . GO term analysis showed an excess of misregulated genes with potential functions in development , such as transcription factors and signaling molecules ( an in-depth analysis of the dataset is beyond the scope of this study , and will be published elsewhere ) . In addition , functional clustering analysis using the online tool , DAVID [45 , 46] , for each category in the Venn diagram , uncovered a functional group including homeodomain ( -like ) proteins BarH1/2 ( B-H1/2 , Bar or B ) and Bric-à-brac1 ( Bab1 ) as being modified in rn mutants . Interestingly , both B-H1/2 and bab1 showed changes in transcription levels only during early developmental stages ( Fig 1A ) . It is important to note that B-H1 and B-H2 , as well as Bab1 and Bab2 , are functionally redundant , and have extensively overlapping expression patterns ( S3B Fig ) [47 , 48] . Because only bab1 but not bab2 was included in the initial functional clustering analysis , we re-examined the RNAseq datasets for bab2 . We found that bab2 had an overall higher level of expression than bab1 , and similar trend of misregulation exists for bab2 ( Fig 1A ) . While the p values were still above the arbitrary cutoffs in several cases—likely due to the cellular heterogeneity of antennal disc samples used in transcriptome analysis—the interrelatedness of these genes and rn suggested that they together might have an important role in ORN diversity . Thus , we focused on B-H1/2 and Bab1/2 in this study and explored their roles in ORN diversification further . Rn was previously reported to function in a gene regulatory network together with B-H1/2 , Dachshund ( Dac ) , Apterous ( Ap ) and Bab1/2 to pattern the segmentation of the Drosophila leg disc in the proximodistal ( PD ) axis [49 , 50] . In the leg , the temporally dynamic PD gene regulatory network , under the influence of morphogen gradients , defines a number of concentric domains on the leg disc via cross-regulation , which in turn determines individual segment identities . These data led us to hypothesize that the neuronal diversity phenotypes observed in rn mutants arise due to the changes of expression domains for the PD network components during antennal disc patterning . To test this hypothesis , we first systematically examined the spatial patterns of these factors in the developing antennae of wild type animals , and found that each factor is expressed in a concentric ring along the PD axis of the discs ( Fig 1B ) . The gene expression patterns are remarkably similar between the legs and antennae , suggesting that these two organs share the same molecular tool kits that pattern their respective discs [39 , 51–53] . Rn was previously shown to be a positive regulator of Bab1/2 and a negative regulator of B-H1/2 in the developing leg disc [39 , 51] . Given the evolutionary relationship between the leg and the antennae , we thought a similar regulatory network may exist in the antennal disc [54] . Indeed , the regulatory relationships of PD genes from the leg-patterning network can explain the misregulation of B-H1/2 and bab1/2 in the antennal disc from our RNAseq data . This idea was then confirmed by examining their in vivo expression patterns ( Fig 1C–1E and S3A Fig ) . B-H1 is normally expressed in the center of the disc , bounded by the central fold ( Fig 1C ) . In rn mutants , B-H1 is expanded outside of the central cells into cells that are normally rn-positive and B-H1/2-negative , but the expansion is confined within the distal boundary of Dac ( Fig 1C ) . The ectopic cells that are labeled with B-H1 antibody in rn mutants are positive for the rn promoter reporter ( Fig 1C ) , suggesting that this rn-positive precursor domain may have switched fates as a result of the loss of Rn and the expansion of B-H1/2 . On the other hand , Bab2 expression is significantly reduced in rn mutants ( Fig 1D and 1E ) . Consistent with the RNAseq results , we did not detect obvious changes in ap expression in the third instar larval stage ( S3A Fig ) . Taken together , these results suggest that a common PD gene regulatory network module operates in parallel during leg and antennal disc development . In rn mutant antennae , the number of ORNs in some rn-negative sensilla ( e . g . Or47b ORNs in at4 ) are increased at the expense of ORNs in rn-positive sensilla [38] , and this occurs in parallel to the expansion of Bar in the antennal disc . To test if the expansion of Bar leads to an increase in at4 ORNs in rn mutants , we analyzed Bar/rn double mutants . Normally , approximately 60 Or47b ORNs are found in wild type flies , and this number is increased to ~90 in rn mutants [38] . We first generated eyFLP-induced MARCM clones , which induced small clones that are either wild type or Bar mutant in approximately 20% of all ORNs ( Fig 2C and 2D ) . These analysis showed that the number of total Or47b ORNs in Bar mutant clones was not significantly different than that of the wild type clones ( Fig 2C and 2D ) . However , when generating similar Bar mutant antennal clones in rn mutant animals , we detected a statistically significant suppression of Or47b ORN expansion down to ~80 cells using ANOVA and post-hoc Student’s T-test ( Fig 2B , S1 Table , p<0 . 001 , Fcrit = 3 . 25 , df = 39 ) . We specifically detected a loss of Or47b ORNs from the ectopic antennal zone seen in rn mutants ( Fig 2A ) . These data suggest that the expansion of Bar is causal for the increase of at4 ORN fates in rn mutants . As at4 sensilla are rn and dac-negative [35 , 38] , they are likely developed from the Bar-positive inner circle of the disc ( Fig 1B ) . Consistently , at4 ORNs express the Bar promoter reporter ( see below for fate mapping and genetic analyses ) . Remarkably , Bar seems to be dispensable for the endogenous at4 fate ( Fig 2C and 2D ) , presumably due to the robustness of this fate to genetic perturbations . Next we wanted to know if the genes in the network directly regulate each other . We focused on the function of Rn , as this may help explain the misregulation of B-H1/2 and bab1/2 in rn mutant . Previous in vitro assays have shown that Rn binds to a T-rich motif ( T13 ) in the LAE ( leg and antennal enhancer ) sequence upstream of bab2 to activate its expression during leg and antennal development . However , in vivo evidence for Rn binding targets has been missing due to the lack of a high-quality antibody . We generated a fly line that carries an EGFP endogenous tag for Rn ( Rn-EGFP ) , which was confirmed and validated for functionality [55] . We then used EGFP antibodies to do chromatin immunoprecipitation ( ChIP ) followed by qPCR to test binding of Rn to bab or Bar regulatory elements in the antennal discs . qPCR primers were designed in the first 2kb upstream of the transcription start site ( TSS ) in the Bar loci ( Materials and Methods ) . A primer set covering the T13 motif in the bab2 enhancer was used as a positive control , while the M1 motif region from Or82a promoter was used as a negative control [38 , 39] . ChIP on antennal disc tissues was able to detect direct binding of Rn to the published bab2 enhancer and the promoter region of B-H2 using the Rn-EGFP line , and further confirms that Rn does not bind to OR promoters ( Fig 2E ) . We noticed that the binding of Rn to bab2 enhancer is more robust compared to B-H2 sites , which might arise due to the differences in the genomic organization of these binding sites . Since both B-H2 and Or82a contain T13-like motifs in their upstream regions , the binding of Rn seems to require some special chromatin environment and/or the facilitation of binding by other factors in addition to the presence of a T13 consensus sequence . While our analysis cannot distinguish whether Rn binds to a different motif in the B-H2 promoter , these results suggest that the concentric TF domains may be formed by cross-regulatory relationships , and that Rn regulates components of the network through directly binding to their regulatory elements . We noticed that the expression domains of several PD factors overlap in the third instar antennal disc , and therefore we wanted to dissect the spatial relationships between these factors more carefully . To simplify the descriptions , we use the central fold ( CF ) as a landmark , which is usually observed as an unstained dark circle in a superficial section of confocal images , to separate the disc into inner and outer regions ( Fig 3B and S3C Fig ) . In the outer region , Dac , Rn and Bab are expressed from more proximal to more distal area in the disc ( Fig 3B ) . Due to the substantial overlap in their expression patterns , these three factors divide the region into four concentric rings . We number the rings starting with the outermost one being R ( 1 ) , and therefore R ( 1 ) to R ( 4 ) are assigned to this region ( Fig 3A and 3B and S3D Fig ) . Bab here is expressed in a gradient , similar to its previously reported expression in the leg discs [56] . Our results show that the highest level of Bab is found near the central fold , and its expression decreases toward both outermost and innermost areas of the disc ( Fig 3B and S3C Fig ) . Three more rings can be found inside the central fold . R ( 5 ) is the only ring that shows quadruple labeling by 4 factors examined ( Rn , Bab , Ap , and Bar ) ( Fig 3C and 3D ) . This ring also corresponds to the only region that expresses Rn inside of the central fold ( Fig 3C–3E ) . Bar expression cannot be detected in the centermost region ( Fig 3C–3E ) . Taken together , the partial overlapping patterns of Dac , Rn , Bab1/2 , B-H1/2 and Ap expression demarcate seven concentric ring domains in the third instar antennal disc , and each ring is marked by a unique combination of prepatterning factors ( Fig 3A and S3D Fig ) . Next we asked which precursor identities are generated from each of these seven domains . As all of the components within a sensillum arise from a single SOP , we wanted to know the sensilla subtype identities of SOPs from each concentric domain . To do this , we used promoter-driven reporter lines for each individual gene to label ORN axons . Because ORN sensory identities are closely linked with the glomerular identities in the brain , we can infer which ORN classes express the given factor from the glomerular labeling pattern in this analysis ( Fig 4 and Table 1 and S4 Fig ) . Bar- and bab-GAL4s were analyzed at both adult and pupal stages , whereas ap-GAL4 was analyzed only at mid-pupal stages due to the lack of adult expression . Our model makes predictions as to how manipulations of the patterning network would lead to changes in ORN diversity . As previously reported , rn mutation effectively halves the amount of ORN diversity in the antenna [38] . We constructed a scheme to depict the spatial relationships of the PD transcription factors in rn mutants ( Fig 5B ) . In this model , the TF combinations in Rn-positive domains , namely R ( 2 ) to R ( 5 ) , are altered due to the changes to the expression of Bar , Bab and Rn . As a result , R ( 1 ) would be expanded into R ( 2 ) and the proximal portion of R ( 3 ) . Similarly , R ( 6 ) would be expanded into R ( 4 ) and R ( 5 ) in rn mutants ( Fig 5B and S1 Text ) . Because we observed an expansion of Bar in rn mutants and this expansion is required for their ORN phenotypes , we wanted to test the effects of ectopic Bar expression in the rn expression domain on ORN populations . Ap was previously shown to protect Bar from being repressed by Rn during leg development [51] . Therefore , either overexpressing Bar directly or indirectly by overexpressing Ap should at least partially recapitulate the adult ORN phenotypes in rn mutants . We analyzed OR expression as readouts of ORN classes using quantitative RT-PCR ( qRT-PCR ) for a panel of 20 olfactory receptor genes representing each of the antennal sensilla subtypes in these genetic backgrounds . We confirmed that this assay provides a reliable readout of ORN fates , by showing that the predicted OR expression profiles in rn mutants were readily reproduced ( Fig 5C ) [38] . As predicted , when Bar or Ap was overexpressed , the changes in the expression of the majority of ORs trended towards changes observed in rn mutants ( Figs 6D and 7D ) . One exception was Or47b in at4 that was downregulated in Bar overexpression lines . We have already shown , however , that the expansion of Bar expression accounts for the increase of at4 sensilla in rn mutants . To reconcile this discrepancy , we re-examined the expression of Or47b using a reporter line in Bar overexpressing flies . Although we observed an overall decrease in the number of Or47b neurons consistent with the qPCR result , the domain of expression was expanded to the medial region similar to the manner observed in rn mutants ( S5 Fig ) . In agreement with the expansion of at4 sensilla , glomerular sizes appeared larger for all at4 ORNs ( Or47b , Or88a , and Or65a ) compared to wild type ( S6A , S6B and S6D Fig ) , a similar phenomenon observed in rn mutants . ( S6A and S6C Fig ) [38] . In contrast , the target glomeruli are lost for ab5 ORNs , which show dramatic reduction based on OR expression in qRT-PCR ( Figs 6D and 7D and S6A–S6D Fig ) . We further validated the expression of a subset of ORs in the antenna using reporter lines , we could recapitulate the same changes in OR expression uncovered by the qPCR analysis ( S6E–S6M Fig ) . Of particular note is that ORs in the same sensillum ( ab10: Or49a , ab7: Or67c ) changed in a similar manner to their partner OR genes ( ab10: Or67a , ab7: Or98a ) ( S6E–S6M Fig and Figs 6D and 7D ) . These results suggest that manipulating the PD gene network causes switches of SOP fates and ORN populations . During our examination of Bar-overexpressing larval antennal discs , we found that the central fold ( CF ) disappeared ( Fig 6B and 6C and S7 Fig ) . In contrast , ap and Bab expression patterns were unaffected ( S7 Fig ) . Similarly , Dac showed normal expression , despite the reported function of Bar to repress Dac in the distal area , which suggests that the repression may be time-sensitive and/or context-dependent [47 , 52] . Next we examined the effects of Ap overexpression on the expression patterns of the network genes in the antennal disc . As expected , Ap overexpression resulted in the ectopic expression of Bar protein outside of its normal boundaries in the antennal disc ( Fig 7B and 7C and S8A and S8B Fig ) . Similar to rn mutants , an expanded Bar zone is bounded by the distal limit of Dac in this background ( Fig 7B and 7C ) . However , unlike in rn mutants , Bar does not fully extend to the boundary , and hence , these proximal cells in R ( 4 ) are positive for Rn but negative for Dac and Bar ( S8D Fig ) . They also express Bab and Ap , making them a separate subpopulation within R ( 4 ) . In addition , we saw a loss of Rn expression in R ( 5 ) , leaving the domains within the central fold devoid of Rn expression ( S8A–S8C Fig ) . The simplest interpretation of this data is that increased levels of Ap repress Rn expression in a context-dependent manner . Moreover , we found that Dac expression is decreased in R ( 2 ) and R ( 3 ) that also express Rn ( Fig 7B and 7C ) , suggesting that rn promoter-mediated Ap expression represses Dac in this overlapping domain . Because Dac represses Bab [57] , the reduction in Dac expression should theoretically cause an increase in Bab expression , although we cannot detect any obvious changes for Bab . This may be due to its overall low concentration in this region by repression from other factors [57] . Based on these analyses , we drew similar illustrations for precursor domains in the Ap and Bar overexpression backgrounds ( Figs 6A and 7A and S1 Text ) . They reveal different patterns of gene expression for a number rings compared to the rn mutants , which may account for their differences in adult ORN classes as shown by the qPCR results ( Figs 6D and 7D ) . We conclude that the PD gene regulatory network function in combinations to diversify precursor and ORN fates . We next examined the requirement of Bar in producing the four fates that arise from the Bar-positive region ( Fig 5A ) . To do this , we created Bar mutant clones that delete both BarH1 and BarH2 . However , our analysis did not reveal any significant changes in adult OR expression ( S9A Fig and Fig 2C and 2D ) . The most likely explanation for this observation is that Ap and Bar may have partially redundant functions for some sensilla subtypes , such as at4 and ac2 in R ( 6 ) . Consistent with this , transheterozygous ap mutant alleles using apmd544GAL4 and the nap1 deficiency did not affect ORNs from at4 and ac2 precursors , either ( S9B Fig ) . In fact , only three sensilla subtypes ( ab2 , ab6 , and ac1 ) from R ( 7 ) and R ( 5 ) showed modest decreases in OR expression in ap mutants ( Fig 5A and S9B Fig ) . These results suggest that the developmental refinement of SOP fates in the three inner rings are robust , which makes their dependence on factors like Bar and Ap limited . It has been shown in the leg that this network of PD genes functions under the control of an EGF signaling gradient , which is highest at the center of the disc and decreases outward [33] . There , EGFR signaling represses Rn and activates Bar expression [33 , 49] . We next tested the hypothesis whether perturbations in EGFR signaling can cause modifications to ORN fates . To do this , we expressed a constitutively active EGFR [33] using rn89GAL4 and performed qRT-PCR on ORs ( S10A Fig ) . As expected , these experiments showed that ectopic activation of EGFR function is associated with an expansion of Bar and reduction of Rn expression in the antennal disc ( S10B and S10C Fig ) . In addition , the ORN classes originated from R ( 1 ) , R ( 2 ) , R ( 3 ) , and R ( 5 ) precursor domains were affected in the adult . These results suggest that EGFR signaling may indirectly regulate ORN diversity by modulating the PD gene network . Bab is partially activated by Rn , and it is significantly downregulated in rn mutants ( Fig 1D and 1E ) . It is plausible to think that Bab functions downstream of Rn to specify rn-positive precursor fates . If this is the case , we should see reduced expression of the receptors from the eight rn-positive sensilla subtypes in a bab mutant . To our surprise , only two of the eight receptors tested showed reductions , and another two were even increased in the babPR72 hypomorphic allele ( Fig 8B ) . We noticed a range of changes for sensilla subtypes from the same ring ( Fig 8B ) . For example , among sensilla specified in R ( 7 ) , ab2 and ab6 are reduced , whereas at2 is increased in the bab mutant . The simplest interpretation is that different levels of Bab are required to distinguish these fates in the same ring . When the overall level of Bab is decreased , some sensilla requiring higher Bab may be converted to the ones that require lower Bab . Similarly , only ac1 from R ( 5 ) is reduced , and one explanation is that the lowered Bab expression is still above the threshold for specifying ac4 , but not for ac1 . Alternatively , ac1 ( requiring higher Bab ) may be converted to ac4 ( requiring lower Bab ) , and compensates for the loss of the endogenous ac4 , which may die due to the reduction of Bab . The same reasoning can be applied to ai1 versus ab7 from R ( 3 ) . For R ( 4 ) , we saw increases in at1 and ab5 , and a trend towards downregulation for at3 in the mutant , albeit the latter was not significant ( Fig 8B ) . We then counted the number of Or19b neurons housed in at3 sensilla , and found that it is significantly reduced ( Fig 8C and 8D ) . This discrepancy in the qPCR result may be due to the random fluctuation of gene expression levels , especially when the changes in the numbers of cells are small . This result suggests that similar conversions may occur in R ( 4 ) among the three fates when Bab is reduced . In contrast , the Bab-positive at4 and ac2 from R ( 6 ) appear to be normal in this hypomorphic allele . This could either be because these two sensilla are specified with wider ranges of Bab levels or some other factors are needed to differentiate the two fates . Taken together , these data suggest that Bab could be an essential factor to distinguish alternate SOP fates within a ring using its concentration gradient . The distal portions of both the legs and antennae are chemosensory organs covered by sensilla . The legs , being part of the gustatory system , display neuronal and molecular diversity that is characterized by a huge variety of gustatory receptors expressed on the legs . Unlike ORNs , individual gustatory receptor neurons ( GRNs ) express multiple receptors , and a given GRN class can be found in different sensilla within different GRN clusters [32] . Because the legs and antennae use the same molecular network to pattern these chemosensory appendages , we asked if a similar genetic program operates to pattern the adult GRN fates . We tested this hypothesis in rn mutants as rn is thought to be required for the development of tarsal segment 3 ( ta3 ) , and this segment is lost in rn mutants [58] . However , we do not have a reporter line that uniquely labels ta3 . On the other hand , Gr5a and Gr61a are expressed in the mid and hind legs , where they are restricted to the GRNs in ta4 and ta5 . In both cases , we could reproducibly detect an extra neuron in the mid or hind legs of rn mutant ( Fig 9 and S11 Fig ) . To confirm this result , we used a reporter to label the bitter receptor Gr58c that is expressed by a partner neuron in the same sensilla . We observed ectopic Gr58c neurons in rn mutants ( Fig 9 ) . In contrast , the Gr43a-expressing neurons , which coexpress Gr61a but are housed in another sensillum appeared be unchanged in the mutant ( S11 Fig ) . These results suggest that the sensilla , 5b and 4s , that house the Gr5a/Gr61-expressing sugar neurons and the Gr58c-expressing bitter neurons are expanded towards the proximal segment of the legs in rn mutants ( Fig 9 ) . Taken together , we speculate that the same molecular network is used in parallel to diversify chemosensory neurons in the antennae and legs . We propose a conserved stepwise strategy to explain the overall ORN diversity . First , the prepatterning phase generates distinct pools of epithelial cells with unique differentiation potentials . This is followed by sensory organ precursor selection by proneural genes . Finally , these precursors undergo neurogenic divisions that allocate alternate fates into daughter cells through Notch signaling and terminal selector transcription factors . One salient aspect of this cellular diversification strategy is its modularity . Each step is driven by context-independent rules , yet produces vastly different neuronal outcomes across systems in a developmentally context-dependent manner . For example , Rn is used to generate distinct precursor differentiation potentials in both the antennal and leg discs to increase the complexity of the patterned precursor fields , which give rise to ORNs and GRNs , respectively . Similarly , Notch is reiteratively used during SOP divisions to generate each sensillum . Its function of segregating binary cell fates is context-independent , although the exact fates being segregated are quite different for each sensillum [61] . Therefore , this stepwise mechanism simplifies the overall difficulty of creating neuronal diversity all at once by logically deconstructing similar differentiation processes into single-purpose steps with shared control elements . Even though our findings are in the PNS , there are similar examples of stepwise patterning and diversification in the fly CNS . For example , different neuroblast lineages in the Drosophila embryonic CNS are first specified by spatially restricted factors within specific positions of an orthogonal grid in the embryo [62] . Anterior-posterior axis specification is controlled by Hox-segment polarity genes , which determine the overall fate , just as in the PNS ( leg vs . antenna ) . Dorsoventral patterning is controlled by cross-regulatory transcription factors , which are turned on in response to different concentrations of morphogens such as Hedgehog and Dpp . Similar to olfactory SOPs , patterning of the neuroepithelium is followed by the expression of proneural genes and selection of neuroblasts , which undergo asymmetric divisions and neurogenesis . The division patterns and factors that are asymmetrically segregated into each daughter cell are remarkably similar regardless of the neuroblast lineages to which they belong [62 , 63] . There are also parallels between our findings in the fly olfactory system and the more complex vertebrate olfactory system . Even though stochastic selection has been proposed as a mechanism for the expression of specific OR genes by different ORNs , the restriction of mammalian OR expression into distinct zones suggests that a deterministic mechanism may also be at play [21] . Consistent with this hypothesis , some OR classes that are restricted to specific domains in the mammalian olfactory epithelium were shown to contain known TF binding sites [64 , 65] . Interestingly , some TFs , such as the mammalian orthologue of Apterous , Lhx2 , have evolutionarily conserved developmental functions in olfactory neurons [66] . We suspect that some of the mechanisms used in diversifying fly ORNs may also be used in the mammalian system during the step of OR zonal separation . Examples of similar neuronal diversification cascades utilizing gene regulatory networks under morhogen gradient control are also seen in the vertebrate CNS and PNS . In the classic example of spinal cord neuron diversification , morphogen gradients ( BMP/Shh ) along the D/V axis of the neural tube activate different sets of transcription factors in the precursors to set up a number of domains prior to neurogenesis , thereby diversifying both progenitor and sensory neuronal subtypes they generate [67 , 68] . Recently , the radial glia that give rise to neurons of the cortex were also found to be heterogeneous [69] . Such combinatorial TF network modules confer positional and temporal information to each neural stem cell in order to create a diverse progenitor population in the mammalian cerebral cortex [69] . Segmental patterning of these neural stem cells contributes to neuronal and glial diversity [70] . Similarly , cortical projection neuron fates can be switched among lineages when the corresponding gene network is modified , changing the zonal partitioning of the neocortex [71] . These results , in light of our findings , point to a common strategy composed of modular and simple commands functioning in a nested manner to increase neuronal diversity in multiple developmental contexts . We associate terminal differentiation of ORNs with the specific selection of an OR gene for expression . At least in flies , it is possible that OR expression and ORN diversity are regulated by a set of “terminal selector genes , ” similar to those proposed by Oliver Hobert [72] . Here , a TF , or combination of TFs , directly regulates the expression of genes required for terminal differentiation and function [72] . So far , a list of postmitotic TFs have been shown to directly regulate OR expression by associating with OR promoters [73–77] . However , the loss of these TFs only affects the expression of specific subsets of OR genes , and yet most OR genes have binding sites for these factors . The functional specificity of each TF and their expression patterns in ORN classes have not been well defined . It is possible that chromatin states around OR promoters in different ORN classes govern how these TFs function within each class . Epigenetic modulations of chromatin status have been shown to play an important role in numerous developmental processes , including the development of the olfactory system [78–82] . The prepatterning TFs could recruit epigenetic modifying factors to change the open and closed states of the chromatin around genes critical for different fates . These modifications can be inherited during cell divisions , and affect the genomic accessibility of later factors . Regardless of the exact mechanisms in between , it is likely that the expression/function of terminal selector genes are regulated , at least in part , by the developmental context established by the prepatterning network that we have described . Establishing a clear link between the early patterning networks and the late terminal selector TF network will be critical to resolving these paradoxical results . This patterning network is remarkable in its functional and structural similarity in driving neuronal diversity in two related chemosensory organs , the antennae ( olfactory ) and the legs ( gustatory ) . At the top of the hierarchy of the cascade , Hox genes determine the overall neuronal identities within the discs during embryogenesis . Olfactory lineages , for instance , are determined by the gene homothorax in the antennal disc , whereas homothorax is inhibited by Antennapedia in the leg discs conferring gustatory identities [83] . Strikingly , regardless of the particular Hox gene , the PD gene network module seems to perform similar diversification commands in both chemosensory systems . It will be interesting to ask how different sets of identity genes are regulated in different tissues on the molecular level , and how the cellular memory is passed down through the cascade . Understanding the diversification process from homogeneous fields of precursors to diverse , terminally differentiated neuronal populations will provide key insights into how cascades of master regulatory transcription factor networks can generate and modify the cellular complexity seen in multicellular organisms . Understanding this diversification process can also help us understand the origins of this complexity . At an evolutionary scale , clear analogy exists between ORN precursor diversification process and the segment diversification during early embryogenesis along the myriapods-insect lineage [84] . The addition or elimination of TFs governing either process might reflect , or likely instruct the generation of new fates . Based on a modern version of “the law of development” postulated almost two centuries ago [85] , the acquisition of increased complexity of a tissue and the concomitant genetic changes over evolutionary time is recapitulated by the temporal role and developmental order of the genes that establish the complexity . Under this assumption , a primordial state in antennal development might be the expression domains of Bar and Dac in the antennal disc . Rn was then added to this network later in development and evolution . This would explain the dramatic decreases in ORN diversity and the expansion of specific ORN populations in default sensilla subtypes in rn mutants as well as Bar/Ap overexpression . Indeed , the onset of Rn expression is later in third instar discs compared to those of Dac/Bar/Ap [51] , and Rn seems to be unique to Arthropods , especially insects . Thus , it is plausible that Rn is a newer addition to the network . Conceivably , Rn evolved to generate novel olfactory neurons in order to help the ancestral Arthropods exploit novel olfactory niches . babA128 , Df ( 3L ) babPR72 , were from Frank Laski . Df ( 1 ) B263-20 FRT19A , UAS-BarH1M13 were from Tetsuya Kojima . UAS-Egfr . λtop4 . 4 was from Amanda Simcox . rntot , rntod was previously described [38] . OR-CD8 GFP , OR-GAL4 , IR-GAL4 , GR-GAL4 lines were from Leslie Vosshall , Barry Dickson , Richard Benton and John Carlson , respectively [7 , 20] . Or67dGAL4 knock-in stock was a published line showing faithful expression of Or67d [86] . rn89 , bab1Agal4-5 ( #6802 ) , bab1Pgal4-2 ( #6803 ) , apmd544 , aprK568 , UAS-ap , Df ( 2R ) nap1 , tubP-GAL80 ey-FLP FRT19A , FRT19A , UAS-CD8 GFP , UAS-Syt GFP , UAS-FLP were all from Bloomington Stock Center . NP4099 ( BarGAL4 ) was from Drosophila Genetic Resource Center . Genotypes for fly genetics: Fig 1B . w1118 . aprk568 . Rn-EGFP Fig 1C–1E . rn+/-: UAS-CD8 GFP/+; rn89GAL4/TM6b . rn-/-: UAS-CD8 GFP/+; rn89GAL4/rntod S3A Fig aprK568/UAS-CD8 GFP; rn89GAL4/rntod S3B Fig bab1A128/+ S3C Fig Rn-EGFP Fig 2A . Bar+/- rn+/-: eyFLP FRT19A TubGAL80/+; Or47b::mCD8-GFP/+; rntot/+ . Bar+/- rn-/-: eyFLP FRT19A TubGAL80/+; Or47b::mCD8-GFP/+; rntot/FRT rntot . Bar-/- rn-/-: eyFLP FRT19A TubGAL80/ Df ( 1 ) B263-20 FRT19A; Or47b::mCD8-GFP/+; rntot/FRT rntot Fig 2C . Control: eyFLP FRT19A TubGAL80/FRT19A; Or47b GAL4 UAS-GFP/+ . Bar MARCM: eyFLP FRT19A TubGAL80/ Df ( 1 ) B263-20 FRT19A; Or47b GAL4 UAS-GFP/+ Fig 2E . w1118 . Rn-EGFP Fig 3B and 3C . Rn-EGFP Fig 3D and 3E . aprK568/+ , Rn-EGFP Fig 3F . aprK568/+ Fig 4A–4C and S4A and S4B Fig . NP4099 ( BarGAL4 ) ; UAS-CD8GFP . apmd544; UAS-CD8 GFP . bab1Pgal4-2/UAS-CD8 GFP . NP4099 ( BarGAL4 ) ; UAS-Syt GFP . bab1Pgal4-2/UAS-Syt GFP . Fig 5C . w1118 . rntot/FRT rntot Fig 6B . UAS-CD8GFP/+; rn89GAL4/TM6b Fig 6C . UAS-CD8GFP/+; rn89GAL4/UAS-BarH1M13 Fig 6D . UAS-CD8GFP/+; rn89GAL4/TM6b . UAS-CD8GFP/+; rn89GAL4/UAS-BarH1M13 S5 Fig Or47b::mCD8GFP/+; rn89GAL4/TM6b: Or47b::mCD8GFP/+; rn89GAL4/UAS-BarH1M13 S6C Fig rn89GAL4/rntod . S6E–S6M Fig OR::mCD8GFP . OR::mCD8GFP; rn89GAL4/UAS-BarH1M13 . OR::mCD8GFP; rn89GAL4/UAS-ap S7A Fig Rn-EGFP UAS-BarH1M13/TM6B S7B Fig Rn-EGFP UAS-BarH1M13/rn89GAL4 S7C Fig UAS-BarH1M13/TM6B S7D Fig UAS-BarH1M13/rn89GAL4 Fig 7B and S8A and S8C Fig . UAS-ap/+; Rn-EGFP/TM6b Fig 7C and S8B–S8D Fig UAS-ap/+; Rn-EGFP/rn89GAL4 Fig 7D . UAS-CD8GFP/+; rn89GAL4/TM6b . UAS-CD8GFP/+; rn89GAL4/UAS-ap S9A Fig eyFLP FRT19A TubGAL80/FM6 . eyFLP FRT19A TubGAL80/ Df ( 1 ) B263-20 FRT19A S9B Fig w1118 . Df ( 2R ) nap1/ apmd544 Fig 8B . w1118: babPR72 Fig 8C . Or19b::mCD8GFP/+; babPR72/TM6b: Or19b::mCD8GFP/+; babPR72 S10 Fig UAS-CD8GFP/+; rn89GAL4/TM6b . UAS-CD8GFP/+; rn89GAL4/UAS-Egfr . λtop4 . 4 Fig 9 and S11 Fig rn+/-: GR-GAL4/UAS-CD8 GFP; rntot/TM6b . rn-/-: GR-GAL4/UAS-CD8 GFP; rntot/FRT rntot . For the RNAseq analysis , wandering third instar larval antennal discs ( ~70 for each genotype ) , 8hr APF pupal antennae ( ~50 for each genotype ) , 40hr APF pupal antennae ( ~50 for each genotype ) , and adult antennae ( 150 males and 150 females ) from w1118 , rntot/TM6b , and rntot/rntot flies were dissected . RNA was extracted with RNeasy kit ( Qiagen ) following manufacturer's instructions , and was treated with on-column DNase digestion ( Qiagen ) . We extracted RNA only from the antennal portion of the larval eye-antennal discs in order to remove contamination by transcripts from the developing eye . All samples were diluted to 20ng/ul in 55ul volume with H2O , out of which 3 . 5ul was used for quality control using Bioanalyzer ( Duke Microarray Core Facility ) . The concentrations were measured again with Qubit 2 . 0 ( Life Technologies ) , and 700ng RNA was diluted to 50ul total volume with H2O for each sample . RNA sequencing libraries were prepared with TruSeq Stranded mRNA Sample Prep Kit ( Illumina ) following the manufacturer's instructions . For the RNA fragmentation step , 94°C , 2min was used with the intention to obtain a median size ~185bp . PCR amplification was done with 15 cycles . A total of 24 multiplexed libraries ( barcoded ) were accessed for quality and mixed altogether before separating to two identical pooled libraries , which are subject to cluster generation followed by Illumina 50bp paired-end sequencing by UNC High-Throughput Sequencing Facility ( HTSF ) . Drosophila melanogaster transcriptome ( r5 . 57 ) was downloaded from flybase and bwa indexed was created with bwa-0 . 7 . 8 . Each sequencing file was aligned to the transcriptome , and . sam files for each sample were generated by putting two alignments from both reads together . At least over 80% of the total reads were able to align to the reference . After that , count tables were made for each sample with a customized python script , and further consolidated into a matrix containing transcript ID and read counts from all genotypes for each stage with a Ruby script . These matrices were used as inputs for differential expression analysis using a customized DESeq2 R script . For each stage , we first filtered out ORs/IRs/GRs from the RNAseq datasets , and excluded the genes with low expression levels in all three genotypes ( normalized count < 20 ) . We then narrowed the analysis down to genes that show the same trend of differential expression when comparing homozygous vs . w1118 and homozygous vs . heterozygous datasets . Because the heterozygous background may have some dominant effects due to the presence of the balancer chromosome and the heterozygous flies do not show any OR phenotypes [38] , saving genes that pass both comparisons would help remove irrelevant genes modified by the balancer chromosome and meanwhile enhance the discovery confidence . Because of these stringent filtering steps , we could maximize our gene lists with a more relaxed cutoff ( unadjusted p < 0 . 1 ) for the gene ontology ( GO ) and functional clustering analyses . Samples were fixed with 4% paraformaldehyde , washed with phosphate buffer with 0 . 2% Triton X-100 , and staining as previously described [87] . Primary antibodies were used in the following dilutions: rabbit α-GFP 1:1000 ( Invitrogen ) , chicken α-GFP 1:700 ( Aves Labs ) , rat α-Ncad 1:20 ( Developmental Studies Hybridoma Bank ) , mouse α-Bruchpilot 1:20 ( Developmental Studies Hybridoma Bank ) , mouse α-CD2 1:1000 ( Serotec ) , mouse α-Dac 2–3 1:20 ( Developmental Studies Hybridoma Bank ) , rabbit α-β galactosidase 1:800 ( Invitrogen ) , mouse α-β galactosidase 1:800 ( Promega ) , rat α-Bab2 1:1500 ( Frank Laski ) , rabbit α-Bar-H1 1:100 ( Tetsuya Kojima ) . The following secondary antibodies were used: Alexa 488 goat α-rabbit 1:1000 , Alexa 488 goat α-chicken 1:1000 , goat α-mouse-Cy3 1:100 , goat α-rat-Cy3 1:200 , goat α-rabbit-Cy3 1:200 , Alexa 568 goat α-mouse IgG highly cross-adsorbed 1:300 , Alexa 647 goat α-rat 1:200 , Alexa 633 goat α-mouse 1:200 , Alexa 647 goat α-mouse 1:200 . Confocal images were taken by an Olympus Fluoview FV1000 or Zeiss LSM 510 ( Light Microscopy Core Facility ) . Antennae from approximately 50 flies were dissected for each genotype and at least three biological replicates were analyzed for each genotype . RNA was extracted with an RNeasy kit ( Qiagen ) , treated with on-column DNase digestion ( Qiagen ) , and then reverse transcribed into cDNA using the SuperScript First-Strand Synthesis System for RT-PCR ( Invitrogen ) . qPCR was performed using the FastStart Universal SYBR Green Master Mix ( Roche ) or the FastStart Essential DNA Green Master Mix using standard protocol . Expression for each gene was analyzed in triplicate . Ct values were used to calculate dilution factors for each gene based upon standard curves created for each gene . Dilution factors were then normalized to the average factor of all ORs tested . See Table 2 for Primers used . This procedure is modified based on a published protocol [88] . For each genotype , approximately 800 eye-antennal discs were dissected . The samples were cross-linked with 1% formaldehyde in dissection buffer for 10min at room temperature . To quench cross-linking , glycine was added to 125mM final concentration , and the samples were incubated for 5min . The discs were homogenized and sonicated in a Bioruptor machine for 13min ( high frequency; 30 sec ON/30 sec OFF ) . The chromatin was pre-cleared with pre-washed Dynabeads Protein G ( Life Technologies ) for 1hr at 4°C on a nutator . The pre-cleared chromatin was split into 2 tubes ( 1ml/tube ) , and another 20ul ( 2% ) was saved as input and stored at -20°C . 5ug Anti-GFP antibody ( Ab290 ) or an equal amount of normal rabbit IgG were added to either tube , followed by overnight incubation at 4°C . Beads were added to both tubes , and the samples were incubated for 2 hours at 4°C on a nutator . Beads was briefly rinsed with wash buffer I ( 50mM K-HEPES , pH7 . 8 , 140mM NaCl , 1mM EDTA , 1mM EGTA , 1% Triton X-100 , 0 . 1% Na-deoxycholate , 0 . 1% SDS ) , and washed 1X with wash buffer I , 1X with wash buffer II ( the same as buffer I , except that NaCl is 500mM ) , 1X with wash buffer III ( 250mM LiCl , 0 . 5% Igepal CA-630 , 0 . 5% Na-deoxycholate , 1XTE ) , 2X with the TE buffer , at 4°C , 5min/each wash . The chromatin was eluted 2X with pre-warmed elution buffer ( 1% SDS , 100mM NaHCO3 ) . For each elution , beads were incubated in 100ul solution for 10min at 65°C , with gentle vortexing every 2–3 min . To reverse cross-link , 5M NaCl was added to each tube , followed by overnight incubation at 65°C . The ChIP-ed DNA was treated with RNase and proteinase K , and extracted by PCR purification columns ( Qiagen ) . The purified DNA was tested for enrichment of DNA fragments by qPCR . For each target gene , up to 150bp amplicons were selected every ~300bp in the first 2kb region upstream of the coding region ( Table 3 ) . A primer pair covering T13 motif within the bab2 LAE ( leg and antennal enhancer ) was used as a positive control for ChIP-qPCR analysis [39] . The M1 motif upstream of the rn-positive Or82 promoter was used as a negative control . Statistical analysis of Bab2 expression levels , qPCR results , and neuron counts was by unpaired , two-tailed Student’s t test . Single factor ANOVA was used to analyze the number of Or47b neurons in rn/Bar double mutant analysis . Post-hoc , unpaired , two-tailed Student’s t tests were calculated after ANOVA . For all tests , * p< 0 . 1 , ** p<0 . 01 , *** p<0 . 001 .
Drosophila uses 50 different olfactory receptor neuron ( ORN ) classes that are clustered in combinations within distinct sensilla subtypes to decipher a complex chemical environment . Each sensilla subtype houses 1–4 ORN identities that arise through asymmetric divisions from a single multipotent sensory organ precursor ( SOP ) . How each class of SOPs acquires a unique differentiation potential that accounts for ORN diversity is unknown . Here , we show that Rn , along with BarH1/H2 ( Bar ) , Bric-à-brac ( Bab ) , Apterous ( Ap ) and Dachshund ( Dac ) , is part of a conserved proximodistal ( PD ) gene regulatory network module that patterns the antennal disc into seven concentric rings and diversifies SOP identities . Each ring expresses a unique combination of the aforementioned transcription factors , and encodes the differentiation potentials for a limited number of sensilla subtypes . Genetic perturbations of the network lead to predictable changes in ORN diversity . These data suggest that the diversification of precursor fields by the prepatterning network is the first step to neuronal diversification , followed by SOP selection by proneural genes , and Notch-mediated neurogenesis . As each step operates in a context-dependent manner , deployment of the same transcription factor network module may regulate neuronal diversity in parallel systems with completely different fate outputs .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[]
2016
A Functionally Conserved Gene Regulatory Network Module Governing Olfactory Neuron Diversity
Talin serves an essential function during integrin-mediated adhesion in linking integrins to actin via the intracellular adhesion complex . In addition , the N-terminal head domain of talin regulates the affinity of integrins for their ECM-ligands , a process known as inside-out activation . We previously showed that in Drosophila , mutating the integrin binding site in the talin head domain resulted in weakened adhesion to the ECM . Intriguingly , subsequent studies showed that canonical inside-out activation of integrin might not take place in flies . Consistent with this , a mutation in talin that specifically blocks its ability to activate mammalian integrins does not significantly impinge on talin function during fly development . Here , we describe results suggesting that the talin head domain reinforces and stabilizes the integrin adhesion complex by promoting integrin clustering distinct from its ability to support inside-out activation . Specifically , we show that an allele of talin containing a mutation that disrupts intramolecular interactions within the talin head attenuates the assembly and reinforcement of the integrin adhesion complex . Importantly , we provide evidence that this mutation blocks integrin clustering in vivo . We propose that the talin head domain is essential for regulating integrin avidity in Drosophila and that this is crucial for integrin-mediated adhesion during animal development . The formation and maintenance of three-dimensional tissue architecture requires fine-tuning of adhesion between cells and the extracellular matrix ( ECM ) . Integrins are the principal family of cell-ECM adhesion receptors in metazoans and are comprised of an alpha and beta subunit [1] . The extracellular domain of integrins binds to the ECM and their cytoplasmic tail domains mediate linkage to the actin cytoskeleton via adapter proteins . The strength and stability of cell-ECM attachment varies in response to the cellular context: stable , long-lasting adhesion is used to preserve tissue architecture while short-term matrix attachment is used for dynamic processes such as cell migration during embryonic morphogenesis [2] . Thus , determining the strength and duration of adhesion to the ECM has important consequences for animal development and tissue maintenance . The strength and duration of integrin binding to the ECM is controlled by two different mechanisms: by changing the conformation of integrins , and by regulating their clustering . Changes to the conformation of integrins , a process known as integrin activation , modulates the affinity of integrins for their ECM ligands . During activation , the heterodimer switches from a bent low-affinity state to an extended high-affinity state . In comparison , clustering of integrin receptors increases the avidity or accumulated strength of multiple integrin interactions with ECM ligands . Essential roles for both integrin activation and integrin clustering have been demonstrated in various systems and cell types . It is not known whether all cells where integrins are known to function use these two regulatory mechanisms . There are examples of cells and tissues that use regulation by: activation ( for example platelets; [3] , [4] ) , clustering ( such as skin; [5]–[7] ) or both ( for example , several types of leukocytes; [4] , [8] ) . An intriguing possibility is that there are tissue-specific contexts that require a particular mode of regulating integrin , either clustering or activation . However , the identification of such tissue-specific contexts requires the ability to abrogate clustering and/or activation , in vivo , across tissues and compare the observed phenotypes . The large cytoplasmic protein talin is a central mediator of both integrin clustering and activation making it a perfect target for studies aiming to understand both processes . In particular , the N-terminal region of talin , known as the head domain , has been implicated in binding to and activating integrins , a process known as “inside-out” activation [9] . The talin head is composed of an atypical FERM ( Band 4 . 1/Ezrin/Radixin/Moesin ) domain that is made up of four lobular subdomains ( F0 , F1 , F2 , and F3;[10] ) . The F3 subdomain mediates direct interactions with the β−integrin cytoplasmic tail and is required to induce integrin activation [11] , [12] but the other subdomains also contribute to integrin activation [9] , [13] . The mechanisms by which the talin head domain mediates integrin activation have been studied extensively [14] . The model emerging from these studies is that plasma membrane interactions mediated via the F1 , F2 and F3 subdomains , together with F3-dependent β-tail interactions , induce a change in angle of the β-integrin transmembrane domain relative to the membrane [15]–[18] . It is this movement that promotes the conformational changes that drive integrin activation . In comparison , although talin has an established role in inducing integrin clustering [19] , [20] the mechanism that mediates this function has yet to be elucidated . Nonetheless , Saltel and co-workers [21] suggest based on their studies that clustering involves similar conformational changes in integrin to those that take place during activation . Studies in Drosophila melanogaster have generated useful insight into the regulation of integrin function in vivo [1] , [22]–[26] . This is because flies are particularly amenable to transgenic and mutational structure/function analysis . Also , there is an array of developmental processes during fly development that are integrin-dependent allowing for integrin function to be analyzed in diverse contexts . Previous studies in the fly have addressed the role of talin head mediated integrin activation in vivo . When introduced into fly talin , a mutation that abolishes the ability of the talin head to bind integrin via Integrin Binding Site 1 ( IBS-1 ) in the F2-F3 domain ( R367A in fly talin ) , resulted in a phenotype that was mild , consistent with slight weakening of the attachment between integrins and the ECM [25] . Furthermore , phenotypes were only in the muscle at prominent , well-characterized sites of integrin-mediated adhesion known as myotendinous junctions ( MTJs ) . MTJs are sites where integrins mediate stable attachment between muscles and the overlying epidermis . Subsequent studies showed that a second , more C-terminal , integrin binding site ( IBS-2 ) in talin is the main linker between integrins and the Intracellular Adhesion Complex ( IAC ) and that this interaction is of particular importance during morphogenetic events that require more dynamic adhesion [24] . Thus , it was posited that that the head domain is predominantly required for stabilizing rather than establishing Cell-ECM adhesion . These studies leave unresolved questions with respect to the role of talin-head dependent integrin activation , as the IBS-1 mutation used did not specifically abrogate activation , but rather completely disrupted binding between talin head and integrin . Intriguingly , work using insect cell-culture argues that the canonical talin-head induced integrin activation does not occur in Drosophila [27] . Overall , the existing body of data suggests the talin head contributes in novel and as of yet undetermined ways to regulating integrin-mediated adhesion in Drosophila . Here , we utilize a structure/function approach to investigate the role of the talin head domain in the context of the developing fly embryo . We use targeted mutations in talin that abolish talin head-mediated integrin activation while leaving all other functions of the talin head intact . We are thus able to confirm that canonical talin head-mediated inside-out activation is indeed not essential for fly development . Importantly , we identify a point mutation in the talin head that phenocopies complete deletion of the talin head , and interferes with the reinforcement of cell-ECM adhesion . A key feature of this mutation is that it not only disrupts talin head domain-mediated integrin activation but also impinges on integrin clustering . Biochemical analysis of provides a mechanistic basis for the phenotype underlying the mutation , identifying an essential intramolecular interaction between the F2 and F3 subdomains of talin . Our results suggest that a major function of the talin head is to induce integrin receptor clustering , and to promote adhesion maturation . Moreover , we provide genetic evidence demonstrating that clustering is the primary mechanism by which integrin function is regulated in developing fly embryos . We sought to introduce a mutation into the talin head that disrupted its ability to activate integrins but not other aspects of its function . We relied on the extensive knowledge of talin structure generated by previous NMR and crystallographic analysis of the talin-head interaction with integrin in order to do this . Previous studies identified a mutation ( Fig . 1a–b; L325R in talin1 , L331R in talin2 ) in mammalian talin that specifically abrogates the integrin-activating function of talin , but does not substantially affect the ability of the talin head to bind to integrin [28] . When this mutation is introduced into the talin head , it blocks the conformational change in integrin that drives activation . The residue identified specifically attenuates the interaction between the talin head and integrin at the membrane proximal region of the β-integrin cytoplasmic tail , while maintaining the interaction between the talin head and the distal regions of the β-integrin cytoplasmic tail [28] . We introduced this mutation into fly talin ( L334R ) to study its effects . Our lab has previously developed and extensively utilized a protocol to assess the effect of point mutations in fly talin . This protocol relies on the dominant-female sterile germline clone technique [29]; ( see Materials and Methods ) to remove all the endogenous talin from fly embryos . To replace the endogenous talin we used ubiquitously expressed full-length talin rescue constructs , either wild-type ( WT ) talinGFP , or talin point mutants . The WT talinGFP construct rescued the embryonic lethality associated with loss of talin . Surprisingly , we found that talinGFP*L334R-rescued embryos were in some cases indistinguishable from WT talinGFP-rescued embryos; we observed many L334R-rescued embryos hatching to the larval stages ( Fig . 1c–g ) . Talin-dependent morphogenetic movements germband retraction ( GBR ) and dorsal closure ( DC ) were not affected by the L334R mutation ( Fig . 1c , d , g ) . However , , we observed that approximately 20% of late stage 17 embryos possessed a form of muscle detachment defect ( Fig . 1e; Supplemental Fig . S1 ) . Previous work on talin*R367A , a mutation that abrogates talin binding to integrin via its IBS-1 domain and consequently blocks activation , also revealed late muscle detachment defects ( Supplemental Fig . S1; [24] , [25] ) . However , the late muscle defects in talin*R367A mutant were stronger and more penetrant than those of the talinGFP*L334R-rescued embryos ( Supplemental Fig . S1 ) . A hallmark of the talin*R367A phenotype is detachment of integrins from the ECM , marked by staining for the protein Tiggrin ( Fig . 1i; [24] , [25] ) . In contrast , in talinGFP*L334R-rescued embryos , we could not detect detachment of integrins from the ECM ( Fig . 1j ) . Similar to WT talinGFP-rescued embryos ( Fig . 1h ) , they exhibited complete overlap of integrin and ECM signal at MTJs . A possible explanation for the difference between the R367A and L334R mutation would be that the L334R mutation might not block integrin activation in flies . To directly test this possibility , we used a cell culture based activation assay to confirm that the L334R mutation indeed blocked the ability of the talin head to activate integrins . GFP-tagged fly talin head constructs , either WT or L334R , were transiently expressed in CHO cells that stably express human αIIbβ3 integrins . Activation was assessed using an established flow cytometry-based assay to quantify activation of αIIbβ3 integrins . As was previously shown [27] , fly talin robustly activates human integrins when expressed in cell culture ( Fig . 1k ) . In comparison , we found that the L334R mutation in the talin head abrogated integrin activation ( Fig . 1k ) . Consistent with the mild phenotype observed in talinGFP*L334R-rescues , we found a number of subtle differences in sub-cellular localization and dynamics between WT and L334R mutant talin . First , recruitment of talinGFP*L334R to sites of integrin-mediated adhesion was slightly reduced compared to WT talinGFP ( Fig . 1l ) . This effect was more pronounced in the presence of endogenous talin suggesting that the mutant protein is outcompeted by the untagged WT protein for integrin binding ( Fig . 1m ) . Second , we found that the turnover of talinGFP*L334R was elevated compared to that of WT talinGFP ( Fig . 1n ) , when we employed a Fluorescence Recovery After Photobleaching ( FRAP ) protocol that we developed to analyze the stability of components of the integrin adhesion complex in sites of Cell-ECM attachment in the fly muscle [30] , [31] . Taken together , these data suggest that talin head-mediated integrin-activation , or at least L334-dependent activation , is dispensable for most of fly embryogenesis but does play a small role late in development in stabilizing adhesion to maintain tissue architecture . We sought to identify roles of the talin head beyond integrin binding and activation . To this end , using the same approach described above , we replaced endogenous talin with a ubiquitously expressed construct that deletes the talin head ( residues 1–448 ) but leaves the rest of talin intact: headless-talinGFP . Importantly , head deletion resulted in severe phenotypes resembling complete loss of talin ( Fig . 2a–c ) . GBR and DC were severely disrupted ( Fig . 2d–e ) , as was stable muscle attachment to the ECM ( Fig . 2f ) . While the headless-talinGFP localized poorly in the presence of endogenous talin it exhibited robust localization in talin null embryos ( Fig . 2g , h ) even though its overall expression appeared somewhat lower compared to wild type ( Supplemental Fig . S2 ) . Therefore headless-talinGFP was able to retain functional interactions that supported recruitment to sites of adhesion . Nonetheless , FRAP analysis showed that headless-talinGFP was substantially less stable at sites of adhesion ( Fig . 2i ) . In addition , the adhesion complex that is normally recruited to sites of adhesion by talin ( [25] , [32]; Fig . 2j , l ) , was absent or severely reduced in headless-talinGFP rescue embryos ( Fig . 2k , m ) . These results show that deletion of the head results in severe defects in recruitment and stabilization of the adhesion complex; consequently , loss of talin head function blocks talin-dependent morphogenetic events . Thus far , we have shown that talin-head mediated integrin activation plays only a minor role in talin function during fly develeopment and that despite this , the head domain has other essential functions in mediating integrin-based Cell-ECM adhesion . To uncover the mechanism underlying additional roles for talin we turned to a previously isolated allele of talin , rhea17 . This allele encodes a talin protein containing a missense mutation in the talin head and importantly , produces a phenotype that is similar to that observed when the talin head is deleted in full ( Fig . 2b , Fig . 3f ) . The rhea17 allele was originally uncovered in a screen for dominant enhancers of a hypomorphic integrin allele [32] . We sequenced the rhea17 allele ( see materials and methods and Supplemental Fig . S3 ) and found that it contains a mutation that replaces a conserved glycine , G340 ( G331 in mammalian talin1 , G334 in mammalian talin2 ) , to a glutamate ( G340E; Fig . 3a–b ) . Phenotypic analysis of embryos homozygous for this mutation ( Fig . 3f–g ) showed that integrin-dependent morphogenetic processes GBR and DC as well as stable muscle attachment to the ECM were severely disrupted compared to heterozygous controls ( Fig . 3f–g , c–e ) . Interestingly , embryos that had a GBR phenotype were more likely to have a DC phenotype . For example , while 38% of the total population of rhea17 mutant embryos analyzed displayed DC defects , amongst those that also have GBR defects , the proportion of embryos with DC defects increased to about 63% . This connection seems likely , because the morphology of the amnioserosa , an extra-embryonic tissue required for both GBR and DC [33] , [34] , is defective as a result of GBR failure . To ensure that the phenotypes we observed in embryos homozygous for the rhea17 mutation were not due to the accumulation of background mutations , we analyzed the phenotype of embryos trans-heterozygous for the rhea17 allele and a talin null allele ( rhea17/Df ) . This revealed an even stronger phenotype , suggesting that the phenotype observed in the rhea17 homozygous mutants is not due to background mutations . Furthermore , this implied that the rhea17 allele is a hypomorphic mutation that retains some functionality in comparison to complete loss of talin ( Fig . 3c–e ) . Of note , the GBR phenotype of rhea17/Df embryos was stronger than that of rhea79 talin null mutant embryos ( Fig . 3c ) . It is thus possible that the rhea17 might be acting in a dominant negative fashion in this process . This is consistent with what we have previously observed for some mutations in integrin that give rise to stronger GBR phenotypes than the loss of function mutants [35] . Another possibility is that the rhea79 allele may have accumulated background mutations that slightly suppress the talin null phenotype . A possible explanation for the strong phenotype observed in rhea17 mutant embryos was that the mutation compromises the stability of talin protein such that functional defects observed could arise from insufficient levels of talin protein . However , analysis of rhea17 embryos revealed that the G340E mutation did not affect either the localization to or levels of , both integrin ( Fig . 3h–j ) and full-length talin at sites of adhesion at MTJs ( Fig . 3k–m ) . Furthermore , Western blot analysis did not reveal any detectable degradation products associated with the rhea17 mutation ( Fig . 3o ) . Finally , side-by-side analysis of embryos heterozygous for either the rhea17 allele or the rhea79 talin null allele demonstrated that there was about twice as much talin protein in the rhea17/+ embryos compared to rhea79/+ . This result indicated that full-length talin protein product from the rhea17 allele was expressed at levels comparable to the wild type allele , demonstrating that the talin protein containing the G340E mutation is stable and sufficiently expressed ( Fig . 3o ) . Since the full-length G340E talin encoded by the rhea17 allele is able to localize to sites of adhesion we asked whether this mutation blocked the ability of the talin head to bind to and activate integrins . To this end , we again employed a flow cytometry-based αIIbβ3 integrin activation assay . This showed that that the G340E point mutation , much like the L334R mutation , blocked the ability of the talin head to activate integrins ( Fig . 3n ) . The phenotype observed in rhea17 embryos cannot be explained by a defect in integrin activation alone since our data demonstrates that blocking activation by itself does not give rise to a severe phenotype ( Fig . 1 ) . Therefore , we hypothesized that the underlying cause of the rhea17 phenotype is due to a defect in another function associated with talin: integrin clustering [19] , [21] . Integrin clustering in the fly can be assessed using a well-established in vivo assay in the context of the fly imaginal wing disc epithelium [25] , [32] , [36] . In the imaginal wing disc integrins mediate adhesion between the epithelial layers and form discrete puncta that colocalize with other adhesion complex components including talin , on the basal surface of the epithelium [32] . In the absence of talin these clusters fail to form , indicating a role for talin in integrin clustering ( Fig . 4a; [32] ) . Interestingly , clonal patches of homozygous rhea17 mutant cells also failed to form integrin clusters ( Fig . 4b ) . In comparison , neither the R367A mutation , nor the L334R mutation , disrupted integrin clustering ( Fig . 4c–d; [25] ) . These results are in line with the hypothesis that the G340E mutation in rhea17 directly impinges on the ability of talin to cluster integrins . If the G340E mutation impacts integrin clustering , we predicted that this would affect the recruitment of integrins to sites of adhesion . To test this idea we analyzed integrin recruitment to MTJs . Consistent with previous reports [37] , integrin recruitment to MTJs exhibited a substantial increase between embryonic stages 16 and 17 in control WT embryos ( Fig . 4e , g , i ) , but in rhea17 embryos , this increase did not occur ( Fig . 4f , , h , j ) . While this result hints at a defect in integrin clustering , it only provides indirect support for this hypothesis . Furthermore , to more directly test the role of clustering we utilized neomycin , a reported inhibitor of integrin clustering that works by sequestration of PI ( 4 , 5 ) P2 membrane phospholipids [19] , [38] , [39] . It was previously shown that neomycin treatment results in increased turnover of integrins at focal adhesions [19] . FRAP analysis of MTJs revealed that the mobile fraction of integrin-YFP increased by over 45% at sites of adhesion in embryos treated with neomycin compared to control vehicle-treated embryos ( Fig . 4k ) . The mobile fraction of integrin-YFP also increased over heterozygous controls in rhea17 mutant embryos ( Fig . 4l ) . Overall , we present three lines of evidence which taken together , support the idea that clustering is disrupted by the presence of the G340E mutation in talin . Both the fact that the talin head was required for adhesion complex assembly ( Fig . 2 ) , and that the G340E mutation interfered with integrin clustering ( Fig . 4 ) , led us to hypothesize that the rhea17 mutation might give rise to defective adhesion complex assembly and maintenance . Since MTJs grow and mature over several hours of embryonic development ( stages 16–17 ) , they serve as a useful system to study the maturation of integrin-mediated adhesions . In WT embryos talin is localized at MTJs as they form during stage 15 and then undergoes substantial enrichment between stages 16 and 17 as Cell-ECM adhesions are consolidated and reinforced ( Fig . 5a , c , e ) . Recruitment of other adhesion complex components including PINCH ( Fig . 5f , h , j ) and pFAK ( Fig . 5k , m , o ) also occurred at stage 16 and was maintained through stage 17 . In rhea17 embryos , although talin is well recruited to MTJs by stage 16 , its recruitment is not reinforced in stage 17 ( Fig . 5b , d , e ) . Strikingly , other adhesion complex components such as PINCH ( Fig . 5g , i , j ) and pFAK ( Fig . 5l , n , o ) were also initially recruited to MTJs at stage 16 at levels comparable to WT , but by stage 17 , the levels were drastically reduced . Intriguingly , we found that MTJs are longer in rhea17 embryos compared to WT controls , further suggesting a failure in adhesion maturation and consolidation ( Fig . 5p–r ) . Altogether , these data demonstrate that the G340E mutation in the talin head impinges on the ability to reinforce integrin-mediated adhesions , consistent with a defect in adhesion maturation , leading to the breakdown of the Cell-ECM adhesion . We have shown that talin protein in rhea17 mutants fails to cluster integrins and reinforce integrin-mediated adhesions . However , the data presented so far does not explain the mechanism by which this effect is mediated . When put into the context of the large body of knowledge that exists about the structure of the talin head , the nature of molecular lesion in rhea17 provides some intriguing hints about this mechanism . Specifically , the G340E mutation is expected to disrupt the coordinated movement of the F2 and F3 domains which is essential for activation and clustering . Structural modeling of the talin head , using the solved crystal structure of the mouse talin head in complex with β1-integrin ( [15]; PDB number 3G9W ) , predicted that residue equivalent to G340 ( G331 ) is located on the surface of F3 at its interface with F2 ( Fig . 6a , inset ) . In the WT talin head this glycine allows the close packing of these two sub-domains . However , substitution for a glutamate inserts a charged carboxyl group into this close gap disrupting the fixed orientation of the F2 and F3 domains , which should allow them to move independently of one another . Since it has been proposed that a tri-partite interaction between integrin , talin head , and the phospholipid bilayer is required to facilitate stable adhesion and to promote inside-out signaling , it is quite possible that disrupting the coordination of F2 and F3 would destabilize these interactions . Consistent with such an effect the G340E mutation rendered the talin head domain proteolytically sensitive to cleavage of the F3 subdomain from the F0-F2 subdomains ( Fig . 6b ) . Furthermore , we used MALDI-TOF mass spectrometry and peptide mass fingerprinting to confirm that the cleaved fragment we observed was indeed the F0-F2 domain , indicating that the F3 had been lost ( Fig . 6c ) . This in vitro evidence is in line with the assertion that F2 and F3 become structurally uncoupled from one another when G340 is mutated . Structural analysis of the mammalian talin head has shown that the orientation of the F2-F3 domains is fixed and that this has important implications for the talin function , for example during inside-out activation [15] . It was shown previously that a cluster of basically charged residues in the F2 domain , the membrane orientation patch ( MOP; Fig . 6a ) , play a key role in integrin activation via electrostatic interactions with the plasma membrane . In order to achieve optimal interaction of the MOP with the phospholipid headgroups , the F2-F3 module needs to reorient relative to the membrane . Multiscale molecular dynamics simulations suggest that this reorientation results in a ∼20° change in the tilt of the β-integrin transmembrane domain [16] leading to integrin activation via dissociation of the α-integrin and β-integrin transmembrane domains . Therefore , if the effect of the G340E mutation on clustering is due to disruption in the process of tilting and dissociation of the α-integrin and β-integrin transmembrane domains than mutations that specifically disrupt this process should also impinge on clustering . We therefore tested three mutations that have been proposed to have such an effect: first the L334R mutation in talin which was suggested to block the change in tilt angle of the β-integrin transmembrane domain [28] . Second , we tested two mutations in integrin previously suggested to promote dissociation of the α-integrin and β-integrin transmembrane domains , D807R ( D723R in β3 integrin; [35] , [40] , [41] ) and G792N ( G708N in β3 integrin;[35] , [42] ) . We found that clustering was similar to wildtype for all three mutations when we replaced endogenous talin or integrin with the mutant proteins ( Fig . 4d and Supplemental Fig . S4 ) . We therefore did not find evidence that supports the idea that the G340E mutation is defective in clustering due to interference with tilting and dissociation of the α-integrin and β-integrin transmembrane domains . Another possibility is that the G340E mutation disrupts the ability of the F2-F3 domains to bind the plasma membrane , which effectively reduces the affinity of the interaction between the talin head and integrin [15] . To test if specific membrane interactions might play a role in talin recruitment to sites of integrin-mediated adhesion , we again used the drug neomycin . Neomycin sequesters PIP2 , which is the predominant phospholipid type that the talin head interacts with at the plasma membrane [15] . Consistent with this idea , we found that with neomycin treatment the recruitment of talinGFP to MTJs in embryos was significantly reduced compared to controls ( Fig . 6d ) . In comparison , we failed to see a reduction in recruitment with the G340E mutation suggesting that the phenotype arises via a different mechanism than simple disruption of talin head's interaction with the plasma membrane ( Fig . 3k–m; Fig . 6e ) . This result implies that F2-F3 coordination could be important for integrin clustering through a mechanism other than by ensuring membrane embedding of the talin head domain . Here we define a function for the talin head domain in regulating integrin adhesion by modulating clustering and therefore avidity for the ECM in Drosophila . Our analyses support several notable conclusions: ( 1 ) canonical talin-dependent integrin activation is largely dispensable for integrin function in flies; ( 2 ) a major function of the talin head is to promote integrin clustering and adhesion maturation; ( 3 ) long-term maintenance of the integrin adhesion complex may depend on a novel , previously uncharacterized mechanism that coordinates the spatial arrangement of F2 and F3; ( 4 ) disruption of clustering and/or adhesion reinforcement leads to severe defects in integrin-mediated processes during fly embryogenesis . It is established that modulating integrin affinity by the talin head is a key mechanism of integrin regulation in mammalian systems [3] , [11] , [12] , [43] . However , previous work suggested that this mechanism was not a major player in the fly [20] , [25] , [27] . Our study provides a possible explanation reconciling these results by suggesting that the major regulatory function of the talin head in flies is in modulating avidity rather than affinity . Furthermore , we show that the regulation of integrin clustering underlies reinforcement of Cell-ECM adhesions to drive their maturation during a key developmental transition . Multiple studies in a variety of cell types have shown that the modulation of avidity is important for integrin function in different contexts [4] , [7] , [8] , [44]–[46] . For example , in leukocytes , it has been shown that the formation of integrin microclusters precedes ligand binding [45] . Interestingly , also in leukocytes , activated integrins are unable to mediate stable adhesion to ligand if the lateral mobility of integrins is restricted implicating a key role for integrin clustering in stable adhesions [44] . Furthermore , in both platelets and leukocytes , defective regulation of integrin clustering rather than integrin affinity has been shown to underlie the severe phenotypes caused by loss of kindlin , thus exemplifying the critical role of integrin clustering [47] . We would therefore argue based on our data , and the work of others that avidity regulation can in some instances be the main method of regulating integrin function . The work we present here fits very well with the conclusions of two cell-culture based studies from Wehrle-Haller and co-workers that explore the role of talin in regulating integrin avidity [19] , [21] . First , Cluzel et al used quantitative live imaging approaches to demonstrate that the formation of integrin clusters required the talin head and integrin [19] . Similar to what we observed , adhesion maturation ( defined by recruitment of actin and other integrin adhesion complex proteins ) only occurred following talin-head dependent clustering of integrins [19] . Furthermore , consistent with our data , their results also indicated that integrin clustering leads to the stabilization of integrin-mediated adhesions [19] . Second , to elucidate the molecular mechanisms by which the talin head supports integrin clustering , Saltel et al used quantitative live imaging , mutational analysis , and computational structural modeling approaches [21] . Based on their findings , they propose that the ability of the talin head domain to promote clustering relies on two separate interactions: between the F2-F3 subdomains of the head and the plasma membrane , and between F3 and the β-integrin cytoplasmic tail [21] . As a result both the F2 and F3 subdomains were equally required for integrin clustering; mutations that disrupted either the F2-F3/membrane interface or the F3/integrin interaction abrogated not only clustering but also focal adhesion maturation [21] . Our results support these findings in a physiological context as our analysis suggests that coordination between the F2 and F3 domains of the talin head is important for integrin clustering and adhesion reinforcement . Furthermore , we identify a specific conserved residue that may mediate this function . In general , the striking similarity between our data , which is derived from very different systems than those used by Cluzel , Saltel and co-workers , supports the notion that the role of talin and the talin head domain in regulating clustering is a basic conserved mechanism that may be important in various cell types . What is the possible effect of the G340E mutation we describe ? We envision two possibilities . First it could be that it is required to coordinate conformational changes within the F2 and F3 domain . Such coordination is required to ensure that concomitant with F2 binding to the plasma membrane through its membrane orientation patch , the F3 is able to bind to the integrin cytoplasmic tail to induce tilting ( [15] , [16] , [18] , [48] ) . However , we were unable to obtain results that support a role for tilt in regulating clustering in the fly , although we cannot discount this possibility . Secondly , the G340E mutation is in the F3 domain of talin in an area of the protein required to bind to the plasma membrane in such way so as to maximize its affinity for integrin [16] , [48] . Therefore it is plausible that the G340E mutation might affect the affinity of talin binding to the integrin by weakening the interaction between the talin head and the plasma membrane . Such reduced affinity would lessen the capacity of talin to experience and transduce the forces necessary for clustering and maturation , since increased force would disrupt the talin-integrin bond . However , our results suggest that the G340E mutation , and disrupted F2-F3 coordination , might have a different consequence than weakening of the membrane-integrin-talin interaction . Specifically , in rhea17 embryos , we did not observe the reduction in talin recruitment to sites of adhesion that might be expected if the G340E mutation affects talin binding to the membrane . Therefore , elucidating how proper F2-F3 coordination might contribute to integrin clustering remains an intriguing question for future study . An intriguing difference between our results and previous studies of integrin clustering is that in the fly , this process appears to occur independently of activation . In comparison , a number of mammalian integrin studies have found a mechanistic interdependence between clustering and activation [44] . This is because both processes were hypothesized to depend on separation of the salt bridge that forms between the α and β-integrin subunits [49] . However , only a subset of mutations that disrupt integrin activation also abrogate integrin clustering [21] . Based on contradictory results from studies in keratinocytes and platelets , we speculate that the relationship between activation and clustering might depend on the type of ECM ligand involved . It is known that blood cells such as platelets encounter soluble ligands and depend on talin-dependent control of integrin affinity as well as integrin clustering , to stably bind ligand and initiate a clotting response [3] , [4] . In contrast , keratinocytes in skin face insoluble , high-density ligands [50]–[52] and mutations in β-integrin that strongly disrupt integrin activation and ligand binding have no effect on integrin function [6] . Importantly , in keratinocytes , there is a known requirement for integrin clustering [5] , [7] . We propose that , in the context of fly development , integrins typically encounter a high density of insoluble ECM ligands and thus primarily regulate their ligand-binding function by receptor clustering rather than through activation . The rhea17 mutant allele affects clustering as well as reinforcement and maturation of cell-ECM adhesions . This introduces the possibility that these two events are linked . Such linkage might occur because integrin clustering can create a concentrated platform for adhesion complex formation and maintenance . Subsequently , it is reasonable to postulate that integrin clustering promotes robust adhesion complex assembly and maturation . In line with this prediction , in rhea17 mutant embryos , we observed progressive loss of adhesion complex components from MTJs between during an important developmental transition ( stages 16–17 ) . During this period of growth , sarcomeres form and muscle contraction begins and in wild-type embryos the recruitment of talin and other adhesion complex components increases dramatically to provide resistance to the growing tensile forces ( [37] , this study ) . We anticipate that as increased force is exerted on integrin-mediated adhesions at MTJs , the talin head is required to facilitate adhesion reinforcement . If the stability of the talin head at adhesions is compromised , as in rhea17 , this reinforcement cannot occur leading to the disintegration of MTJs and subsequent muscle detachment in stage 17 , which is precisely what we observed . In further support of this hypothesis , we discovered that MTJs were longer in rhea17 mutant embryos , suggesting that in the absence of effective integrin clustering , adhesions are not able to consolidate into tight , compact MTJs able to support the forces of muscle contraction . Both in Drosophila and in mammalian cell culture , mutations have been identified that prevent integrin binding but do not prevent integrin clustering [21] , [25] . This observation raises the question of whether the talin head must bind to integrin in order to induce clustering and adhesion reinforcement . One possibility is that the talin head binds to integrins via its C-terminal IBS-2 domain , freeing up the head to act as a scaffold for other adhesion complex components . In many ways , the phenotype caused by the rhea17 mutation resembles the effect of disrupting of the IBS-2 domain [24]; in both cases , attachment between integrins and the actin cytoskeleton is severely compromised . Perhaps F2-F3 interactions are required not only for coordinated integrin and membrane binding , but also for talin head binding to other adhesion complex proteins . This hypothesis constitutes an intriguing avenue of future study . In summary , our results provide insights into how integrins are regulated under physiological conditions to give rise to stable tissue architecture . Our work suggests that the canonical model of talin head function as an integrin activator should be modified to include an additional essential role as an orchestrator of integrin clustering and adhesion complex reinforcement . We furthermore illustrate how specific inter-domain interactions in the talin head contribute to the regulation of integrin function . Based on our data we propose the following model for talin function in the fly embryo: talin is recruited to integrin initially through its IBS2 domain ( see [24] ) , which helps assemble an adhesion complex that links to the cytoskeleton . During embryogenesis , there is an increasing need to generate stronger adhesion as the growth of the embryo generates proportionally greater mechanical strain upon the tissues . It is at this point that clustering becomes essential . Talin is then recruited to integrins via its head domain and is stabilized within growing adhesive contacts by coordinated interactions between F2-F3 and the plasma membrane . These stabilized talin-head-integrin complexes form clusters and act as a scaffold for adhesion complex assembly and cytoskeletal attachment that is maintained and reinforced throughout tissue growth and development . Failure to cluster integrins results in severe defects in reinforcement of cell adhesion , and subsequently Cell-ECM adhesions breakdown in the face of increasing mechanical force [31] . Thus , our work sheds novel light on the molecular mechanisms that act through talin to promote adhesion receptor clustering and adhesion complex stability , crucial aspects underlying tissue morphogenesis and homeostasis . The generation of talinGFP is previously described [30] . To make pUbi-talinEGFP*L334R mutant construct , pBS-talinGFP was mutated using the QuikChange Lightning mutagenesis kit ( Stratagene ) . The talinGFP*L334R cassette was sub-cloned into the pUbi63E vector using a strategy similar to that used to generate the WT talinGFP construct [30] . The making of pUASp-GFP-TalinHead was described previously [25] . This construct was directly mutated to contain the L334R point mutation using the QuikChange mutagenesis kit ( Stratagene ) . All rescue experiments were performed in mutant background such that both maternal and zygotic contributions of talin were eliminated , using the rhea79 allele and the Dominant Female Sterile technique [29] . The rhea79 allele was generated by a P-element excision that covers the entire rhea locus . See [32] for a complete characterization . Females of the genotype yw , hs-Flp/+; pUbi-talinGFP , talinGFP*L334R , or headless-TalinGFP/+; rhea79a , FRT2A/OvoD1 , FRT2A were subjected to a heatshock-regime during the larval stages to generate a mosaic germline in order to give rise to rhea mutant oocytes with maternally supplied rescued transgenes . Virgins were then crossed to rhea79a/TM6b , dfd-GMR-nvYFP males . Embryos without the fluorescent balancer were selected for analyses . Using this approach we find that WT talinGFP rescued embryos resemble WT embryos and that over-expression of transgenic talin does not cause any deleterious effects or ectopic signaling integrin ( [24] , [25]; Supplemental Fig . S5 ) . The rhea17 allele was sequenced according to conventional protocols by sequencing of the entire rhea coding sequence ( see Supplemental Table S1 for a list of primer pairs ) . The G340E mutation was identified in exon 5 through comparison of genomic DNA from homozygous wild type OR flies and heterozygous rhea17 flies ( Supplemental Fig . S3 ) . Maternal-zygotic rhea17 mutants were generated via the Dominant Female Sterile germline-clone technique and crossed to rhea17/TM3 , dfd-GMR-nvYFP males or , to assess the phenotype of rhea17 over a deficiency , to rhea79/TM3 , dfd-GMR-nvYFP males . For all talin FRAP experiments , talinGFP constructs were heterozygous and expressed in a w1118 background . For integrin-YFP FRAP experiments , the transgene was either expressed in a heterozygous w1118 background ( neomycin experiments ) , or expressed in a rhea mutant background ( either rhea17 or rhea79 ) . UAS-driven transgenes were expressed in the muscle using the muscle-specific mef2-Gal4 driver . For analysis of integrin clustering with talin transgenes , yw , hsFLP;;GFP-FRT2A virgins were crossed to males of the genotype of either rhea79 , FRT2A/TM3 , dfd-GMR-nvYFP , headlessTalin-GFP/Y;;rhea79/TM3 , dfd-GMR-nvYFP , L334R; rhea79 FRT2A/TM3 , dfd-GMR-nvYFP or rhea17 , FRT2A/TM3 , dfd-GMR-nvYFP . For analysis of integrin clustering with integrin transgenes ubi-GFP , FRT101; hsFLP males were crossed to virgin progeny from mysXG43 , FRT101/FM7 , Kr>GFP crossed to males of the genotype Ubi-integrinYFP*D807R or UBi-integrinYFP*G792N . For both integrin and talin transgenes larval progeny were subject to a heat-shock regime in order to induce clones . Wandering third instar larvae were selected for dissection and analysis . Embryos and third instar imaginal wing discs were fixed and stained according to standard protocols . The following antibodies were used in our analysis: rabbit anti-talin ( 1∶500 ) , mouse anti talin ( 1∶50; DSHB ) mouse monoclonal anti- βPS-integrin ( 1∶50; DSHB ) , rat anti-αPS2-integrin ( 1∶200 , 7A10 ) , mouse anti-tiggrin ( 1∶1000; Liselotte Fessler , UCLA ) , mouse anti-Myosin Heavy Chain ( 1∶200; Dan Kiehart , Duke University ) , rabbit anti-PINCH ( 1∶1000; Mary Beckerle , University of Utah ) , rabbit anti-phospho-FAK ( 1∶200; Invitrogen ) and rabbit anti-paxillin ( 1∶1000; [53] ) . Rhodamine-conjugated phalloidin ( Invitrogen ) was used to stain actin filaments ( 1∶400 ) . Fluorescently- conjugated Alexa-Fluor-488 , Cy3 and Cy5 secondary antibodies were used at 1∶400 dilution ( Molecular Probes ) . Images were collected using an Olympus FV1000 inverted confocal microscope and an UplanFL N 40×1 . 30 NA oil objective or a UplanSApo 60×1 . 35 NA objective . For all micrographs of whole embryos , or of MTJs , z-stacks were assembled from 8–12 1 . 0–2 . 0 µm confocal sections . Embryos were staged as described in [23] . For the line scan analyses in Fig . 1 , fluorescence intensity profiles across the boxed areas were obtained using ImageJ ( NIH , Bethesda , MD ) and then normalized to the maximum fluorescence intensity in each channel . See also [24] . Recruitment of integrin , talin , or talinGFP to MTJs was calculated according to our previously established method ( [23]–[26] , [35] , [36] ) . Briefly , the mean fluorescence intensity of the signal of interest was measured at the MTJ and the cytoplasm and a ratio of MTJ∶cytoplasmic signal was determined . This value was averaged based on measurements of 5 MTJs ( all ventral-lateral attachments from hemi-segments A2–A6 ) from at least 3 embryos . Localization of IAC components to MTJs was quantified as follows , adapted from the method described in [37]: for each IAC component ( talin , PINCH , and pFAK ) , recruitment was determined by measuring the mean fluorescence intensity at MTJs , which was then expressed as a ratio over the mean fluorescence of αPS2-integrin staining at each MTJ . For each genotype , at 5 MTJs were measured from at least 3 embryos . MTJ length was assessed by measuring the length∶width ratio of the D-V length of ventral-lateral MTJ over the A-P width of adjacent VL1 muscles in the hemisegment posterior to each MTJ that was measured . All images were taken using the same gain and offset settings . This value was averaged based on measurements of 5 MTJs ( hemi-segments A2–A6 ) from at least 3 embryos . All quantitative analyses were obtained using ImageJ ( NIH , Bethesda , MD ) and two-sided Student's t-tests were performed using Prism5 ( GraphPad Software Inc . , La Jolla , CA ) . Stage 17 embryos were collected and prepared for FRAP as described previously [30] . Briefly , embryos were collected from apple juice plates , dechorinated in 50% bleach for 4 minutes , washed with PBS and mounted onto glass slides in PBS . FRAP analysis was performed at room temperature . Photo-bleaching was performed using a 473 nm laser at 5% power with the Tornado scanning tool ( Olympus ) for 2 seconds at 100 mseconds per pixel . Fluorescence recovery was recorded over 5 minutes at 1 frame every 4 seconds . To control muscle twitching in and out of focus , multiple regions of interest ( ROIs ) were selected in non-photobleached regions; only samples for which intensities within control ROIs remained steady throughout the FRAP experiment were used . The mobile fraction and statistical tests were performed using Prism 5 software . Neomycin treatment , which was used to inhibit integrin clustering through sequestration of PI ( 4 , 5 ) P2 phospholids [19] , [38] , [39] , was carried out according to the embryonic drug delivery protocol described in [54] , in w1118 embryos expressing ubi-integrinYFP [30] . The control for integrin FRAP in rhea17 zygotic mutant embryos was rhea79/+ embryos expressing ubi-integrinYFP . We have previously observed that genetic background influences the baseline turnover of integrin adhesion complex components and therefore each experiment requires its own control [31] . In particular , we find that the drug delivery protocol lowers turnover in vehicle-only treated embryos , and thus explains the differences in the mobile fractions of the controls shown in Fig . 4k versus Fig . 4l . Westerns were carried out as previously described [17] . Animals assayed either were heterozygous for the rhea79 talin null allele or the rhea17 allele . Antibodies used were rabbit anti-talin antibody ( 1∶2000; [26] ) and mouse monoclonal anti-β-actin ( 1∶5000 , AbCam 8224 ) . All constructs were expressed in E . coli BL21 Star ( DE3 ) cultured in 2YT media . GST-tagged talin head recombinant proteins ( residues 1–409 ) and mutants therein were purified using glutathione sepharose resin ( GE Healthcare ) and eluted by TEV cleavage . Protein concentrations were determined using extinction coefficients at 280 nm . Proteomics were carried out by the University of Leicester Proteomics Facility ( PNACL , University of Leicester ) essentially as described previously [55] . The TEV eluted proteins were run on an SDS-PAGE gel and the corresponding bands were excised . The gel stab was digested with trypsin and subjected to LC-MS/MS mass spectrometry using an RSLCnano HPLC system ( Dionex , UK ) and an LTQ-Orbitrap Velos mass spectrometer ( Thermo Scientific ) . The data were analysed using Mascot ( Matrix Science Ltd . ) and Scaffold ( Proteome Software ) . The identified peptide fragments were visualized on a structural model of fly talin head produced using the mouse talin1 head construct . All structure images were generated with PyMol ( PyMOL Molecular Graphics System , Version 1 . 5 . 0 . 4 Schrödinger , LLC . ) . The modeling of the potential consequence of the G340E mutation was carried out using the atomic structure of the mouse talin2 F2F3 domain in complex with the beta1d cytoplasmic tail ( PDB: 3G9W [15] ) . The residues that define the extensive interface between the F2 and F3 interface are conserved between mouse and fly with G340 crucial to enable close packing of the two domains . Visualization of the G340E mutant was made using the mutagenesis function in PyMol . The activation state of αIIbβ3 integrins was assessed by measuring the binding of the ligand mimetic anti-αIIbβ3 monoclonal antibody PAC1 in flow cytometric assays as described previously [56] . A CHO cell line stably expressing αIIbβ3 [57] , [58] was transfected with the indicated GFP tagged fly talin head cDNA using polyethylenimine ( PEI ) and 18 h later cells were suspended and stained with αIIbβ3 integrin activation-specific PAC1 IgM ( BD Biosciences ) in the presence and absence of the ligand binding inhibitor EDTA ( Sigma ) . αIIbβ3 integrin expression was assessed separately by staining with monoclonal antibody D57 [57] , a gift from M . Ginsberg ( UCSD ) . Cells were washed and PAC1 binding to live , transfected ( GFP-positive ) cells was assessed with Alexa647-conjugated goat anti-mouse IgM ( Invitrogen ) . In parallel , bound D57 to live expressing cells with similar GFP fluorescence intensity was detected using Alexa 647 fluorophore-conjugated goat anti-mouse IgG ( Invitrogen ) . Activation was quantified and an activation index was calculated as defined by the formula AI = ( F – F0 ) / ( Fintg ) , where F is the geometric mean fluorescence intensity ( MFI ) of PAC1 binding , F0 is the MFI of PAC1 binding in the presence of EDTA , and Fintg is the standardized ratio of D57 binding to transfected cells . The Fintg expression ratio was defined as follows: Fintg = ( Ftrans ) / ( Funtrans ) , where Ftrans is the geometric MFI of D57 binding to GFP-positive cells and Funtrans is the MFI of D57 binding to untransfected cells . FACS data analysis was carried out using FlowJo FACS analysis software and statistical analysis using GraphPad Prism software .
Cells are the building blocks of our bodies . How do cells rearrange to form three-dimensional body plans and maintain specific tissue structures ? Specialized adhesion molecules on the cell surface mediate attachment between cells and their surrounding environment to hold tissues together . Our work uses the developing fruit fly embryo to demonstrate how such connections are regulated during tissue growth . Since the genes and molecules involved in this process are highly similar between flies and humans , we can also apply our findings to our understanding of how human tissues form and are maintained . We observe that , in late developing muscles , clusters of cell adhesion molecules concentrate together to create stronger attachments between muscle cells and tendon cells . This strengthening mechanism allows the fruit fly to accommodate increasing amounts of force imposed by larger , more active muscles . We identify specific genetic mutations that disrupt these strengthening mechanisms and lead to severe developmental defects during fly development . Our results illustrate how subtle fine-tuning of the connections between cells and their surrounding environment is important to form and maintain normal tissue structure across the animal kingdom .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "missense", "mutation", "developmental", "biology", "mutation", "point", "mutation", "integrins", "animal", "genetics", "organism", "development", "invertebrate", "genetics", "cell", "biology", "cell", "adhesion", "embryogenesis", "genetics", "muscle", "development", "biology", "and", "life", "sciences", "organogenesis", "molecular", "cell", "biology", "morphogenesis", "molecular", "biology" ]
2014
The Talin Head Domain Reinforces Integrin-Mediated Adhesion by Promoting Adhesion Complex Stability and Clustering
Short-term synaptic plasticity ( STP ) critically affects the processing of information in neuronal circuits by reversibly changing the effective strength of connections between neurons on time scales from milliseconds to a few seconds . STP is traditionally studied using intracellular recordings of postsynaptic potentials or currents evoked by presynaptic spikes . However , STP also affects the statistics of postsynaptic spikes . Here we present two model-based approaches for estimating synaptic weights and short-term plasticity from pre- and postsynaptic spike observations alone . We extend a generalized linear model ( GLM ) that predicts postsynaptic spiking as a function of the observed pre- and postsynaptic spikes and allow the connection strength ( coupling term in the GLM ) to vary as a function of time based on the history of presynaptic spikes . Our first model assumes that STP follows a Tsodyks-Markram description of vesicle depletion and recovery . In a second model , we introduce a functional description of STP where we estimate the coupling term as a biophysically unrestrained function of the presynaptic inter-spike intervals . To validate the models , we test the accuracy of STP estimation using the spiking of pre- and postsynaptic neurons with known synaptic dynamics . We first test our models using the responses of layer 2/3 pyramidal neurons to simulated presynaptic input with different types of STP , and then use simulated spike trains to examine the effects of spike-frequency adaptation , stochastic vesicle release , spike sorting errors , and common input . We find that , using only spike observations , both model-based methods can accurately reconstruct the time-varying synaptic weights of presynaptic inputs for different types of STP . Our models also capture the differences in postsynaptic spike responses to presynaptic spikes following short vs long inter-spike intervals , similar to results reported for thalamocortical connections . These models may thus be useful tools for characterizing short-term plasticity from multi-electrode spike recordings in vivo . Short-term synaptic plasticity ( STP ) refers to fast and reversible changes of synaptic strength caused by the recent history of presynaptic spiking activity [1] . STP occurs on timescales from milliseconds to few seconds , and includes mechanisms for both facilitation of transmitter release , where synaptic strength increases with consecutive presynaptic spikes , and depression , where synaptic strength decreases . Facilitation and depression are mediated by the dynamics of presynaptic calcium and the depletion and replenishment of vesicles in the presynaptic terminals [1] . The relative contribution of facilitation and depression varies across synapses , cell types , and brain regions [2 , 3] with facilitation dominating at some synapses and depression at others . By shaping postsynaptic responses evoked by trains of presynaptic action potentials , STP alters neuronal information processing [4–6] . In vitro studies have shown that STP has profound effects on temporal filtering [7] , network stability [7] , and working memory [8] . Moreover , there is bidirectional interaction between STP and long-term synaptic changes: STP can determine the magnitude of long-term plasticity [9–12] , and long-term synaptic changes also modify STP [9–13] . This results in an interplay between STP and long-term plasticity on multiple timescales [13 , 14] . Therefore , characterization of short-term plasticity in different systems is crucial for understanding neural computations . Traditionally , short-term plasticity is studied using intracellular recordings where responses of the postsynaptic neuron to presynaptic stimulation are directly measured as evoked postsynaptic potentials or currents . Based on results of intracellular recordings Tsodyks , Markram , and colleagues developed a computational model that describes STP in terms of dynamics of resources and their utilization [15 , 16] . The Tsodyks-Markram ( TM ) model provides a phenomenological description of the short-term dynamics of synaptic responses in terms of 1 ) changes in the probability of transmitter release ( utilization ) , related to the dynamics of presynaptic calcium and , 2 ) the use and replenishment of synaptic vesicles ( resources ) . The TM model accurately captures the dynamics of synaptic responses caused by STP , links the observed diversity in synaptic dynamics to differences in the model parameters ( utilization , recovery of resources , and their time constants ) , and allows prediction of postsynaptic responses to an arbitrary sequence of presynaptic stimuli [17] . Although several alternative models of STP have been proposed [18 , 19] , the TM model is the most broadly used because it provides a compact description of STP with biophysically relevant parameters . The TM model had been successfully used in a number of intracellular studies to assess synaptic dynamics in different connections [17 , 20 , 21] and changes of synaptic dynamics induced by long-term plasticity [13] , adaptation [22] or injury [23] . Traditionally TM model parameters are estimated from responses to presynaptic stimuli applied in bursts of different frequencies [13 , 16 , 22 , 23] . A recent study presented a Bayesian approach that estimates TM model parameters by fitting postsynaptic responses induced by stochastic trains of presynaptic spikes [17] . Thus , STP parameters can be extracted from responses to in vivo-like presynaptic activity . Here we ask whether it is possible to estimate STP parameters using only the spike trains of pre- and postsynaptic neurons without access to postsynaptic potentials or currents . If available such a method would greatly expand the possibilities for studying STP in vivo . Although multiple intracellular recordings or simultaneous extra and intracellular recordings in vivo are possible [10 , 24–26] , they are technically prohibitive for large-scale studies . Techniques for large-scale extracellular recordings , on the other hand , allow simultaneous recording of spiking from hundreds of neurons [27–29] . Prior studies compared cross-correlograms calculated using presynaptic spikes occurring after short or long inter-spike intervals , and found evidence for both short-term facilitation [30] and depression [31 , 32] of synaptic transmission in vivo . This split-correlograms approach , however , does not allow for a detailed reconstruction of synaptic weight for each presynaptic spike or estimation of underlying release probability and vesicular resources . Here we develop two statistical methods that use pre- and postsynaptic spike trains to estimate the dynamics of short-term plasticity . Both approaches are based on a generalized linear model ( GLM ) that predicts postsynaptic spiking as a function of the observed pre- and postsynaptic spikes [33–37] . In these GLM-based methods we allow the effect of the presynaptic spikes to vary on short timescales as a function of the presynaptic spike timing . In a first model , the effect of presynaptic spikes is determined by the nonlinear dynamical equations of the TM model ( TM-GLM ) . In a second model , we introduce a functional description of short-term plasticity based on a generalized bilinear model ( GBLM ) . Although the parameters in the second approach are no longer linked to biophysical properties the GBLM allows us to capture a wide range of neuronal interactions and synaptic dynamics . To validate our models , we recorded spike responses of pyramidal neurons in vitro ( cortical slices , layer 2/3 pyramids ) to intracellularly injected currents composed of synaptic inputs with the known pre-defined short-term plasticity . We show that , using only pre- and postsynaptic spike trains , the TM-GLM can recover the underlying parameters of STP , and the GBLM is able to reconstruct synaptic dynamics using a descriptive plasticity “rule” . Estimates provided by each of the two models were in good correspondence to ground truth values for a wide range of synaptic weights and time scales of facilitation and depression . Additionally , using simulated neurons we show that estimation of STP by these models is robust to several potential confounds: spike frequency adaptation , noise from probabilistic vesicle release , and spike sorting errors . The methods developed here , thus , have the potential to serve as powerful tools for large-scale studies of short-term synaptic plasticity in vivo , including alterations of short-term plasticity during different behaviors , during learning , or as a result of pathology . To mimic recordings from pairs of neurons with known connectivity and short-term plasticity we made intracellular recordings from layer 2/3 pyramidal neurons in slices of rat visual cortex , and recorded spiking responses of neurons to injection of fully-defined fluctuating current . The injected current was designed to mimic the postsynaptic effect of synaptic inputs which have different strength and express unique synaptic dynamics ( Fig 3 ) . To synthesize the current , we used a population of 96 presynaptic neurons where the spike times of each neuron were generated using an inhomogeneous Poisson process with a mean rate of 5 Hz . Six pools of 16 neurons ( 8 excitatory and 8 inhibitory in each pool ) expressed five distinct types of STP , each defined by a unique set of parameters and ranging from strong depression to strong facilitation , along with a sixth pool of neurons which did not express STP ( Table 1 ) . STP of synaptic responses was implemented according to the TM model . Average synaptic weights for the 16 inputs in each pool ranged from strongly excitatory to strongly inhibitory , with excitatory and inhibitory inputs having the same amplitudes but opposite signs . This resulted in a balanced fluctuating current . Using the membrane potential responses to the injected current we detected postsynaptic spikes as positive-slope zero crossings . Thus , in this dataset we knew the timing of presynaptic spikes of each simulated presynaptic neuron , the time-varying synaptic weight , and the timing of the postsynaptic spikes . To illustrate how STP at a single synapse affects postsynaptic firing in the presence of many other inputs , we performed a separate recording where the injected current had additional structure . One out of 96 presynaptic neurons repeatedly discharged with a pattern typically used for testing STP in slice experiments ( 9 regularly spaced spikes + 1 after a delay ) , while the spiking of the remaining 95 presynaptic neurons followed uncorrelated inhomogeneous Poisson processes as described above . This resulted in a repeating test pattern at one synapse embedded in fluctuating noise produced by the activity of the remaining presynaptic neurons . The strength of this synapse was increased to increase signal-to-noise ratio . The average postsynaptic current , membrane potential , and peristimulus time histogram of spiking ( PSTH ) in response to the test stimulation patterns demonstrate that the effect of a single strong input ( >100pA ) is clearly observable [Fig 4] . Moreover , synapses with different short-term synaptic dynamics: depression , facilitation and no plasticity produce distinct postsynaptic responses at all levels . In recordings with in vivo-like activity , the effects of short-term synaptic plasticity will be more subtle , since presynaptic spike times do not occur in such regular , repeating patterns under natural conditions and synaptic weights in neuronal connections are much weaker . The remaining analysis focuses on the recordings without the test patterns , where the strongest synaptic weights were ~30pA . In previous studies a split-correlogram approach had been used to reveal the effects of short-term plasticity on postsynaptic spike statistics in vivo [30 , 31] . By calculating cross-correlograms separately for presynaptic spikes following short ISIs ( or in bursts ) and for spikes following long ISIs ( isolated spikes ) , evidence was found for both short-term facilitation [30] and depression [31 , 32] of synaptic transmission in vivo . To determine if this method of analysis could reveal effects of STP in our data obtained with inhomogeneous Poisson presynaptic spiking , we split presynaptic spike trains into spikes following inter-spike intervals shorter than the 10th percentile and longer than the 90th percentile of ISI distribution ( Fig 5A ) . Separate analysis of the postsynaptic effects of presynaptic spikes from these two groups revealed clear differences between synaptic inputs with distinct types of plasticity [Fig 5B] . In connections with depressing synapses the PSCs , PSPs , and , most importantly , peak spike counts in the correlograms were much reduced for short intervals . In connections with facilitating synapses , the postsynaptic effects were slightly increased following short presynaptic intervals [Fig 5B] . In synapses with intermediate forms of plasticity the effect of ISI on postsynaptic responses was less pronounced and was between the two extremes . Note that because of temporal summation after short ISIs the increase of the postsynaptic responses ( PSCs , PSP , and spike count ) is evident in the short interval correlograms even shortly before 0ms , similar to the results from in vivo study [30] . Thus , the effects of STP on spike responses of neurons to injection of a fully-defined current were clearly expressed in the difference between split-correlograms , consistent with results reported for in vivo recordings [30–32] . Our results show that the effects of ISI on split-correlograms were more pronounced for depressing than for facilitating inputs . One possible reason for such asymmetry may be that the presynaptic spike statistics used here does not fully elicit the effects of facilitation . To address this issue , we examined the distribution of PSP amplitudes as a function of inter-spike intervals for synapses with the different types of STP used in our model [Fig 5C] . While in depressing synapses the PSP amplitudes monotonically increase as ISIs increase , the response amplitudes in facilitating synapses depend on the ISIs in a non-monotonic way . At facilitating synapses , there is an ISI range in which PSP amplitudes are elevated , but for both shorter and longer ISIs the amplitudes are reduced ( Fig 5C ) . This pattern makes it difficult to distinguish facilitation in split cross-correlograms , since short and long ISIs can produce similar PSP amplitudes . Moreover , facilitating responses also have higher variability than depressing responses for any given ISI , likely since stronger facilitation enhances the variability of utilization ( release probability ) compared with depressing synapses ( Eq 3 ) . These factors appear to hinder detection of short-term facilitation with split-correlogram analyses . The examples considered above show results for the strongest , excitatory simulated inputs ( ~30 pA ) . Weaker excitatory synapses and inhibitory synapses express similar dynamics in their PSC and PSP amplitudes , however , the postsynaptic effects are less pronounced and show greater variability . For weak facilitating synapses there is often no detectable difference between the postsynaptic responses to short and long intervals . This analysis exposes a fundamental drawback of the split-correlogram approach: its low sensitivity to transient effects . By explicitly modeling how synapses vary in response to the history of presynaptic spiking , rather than modeling the average responses to only a single previous ISI , model-based approaches can more accurately reconstruct synaptic dynamics and distinguish between different types of STP . We extend the GLM framework to include short-term synaptic plasticity implemented according to the Tsodyks-Markram model ( see Methods ) . The TM model describes the dynamics of synaptic transmission using two coupled differential equations for resources R and their utilization ( release probability ) u with a set of four parameters θ = {D , F , U , f} ( Eq 3 in the Methods ) . To fit the TM-GLM to the observed spike trains we use an alternating coordinate ascent to maximize the ( penalized ) likelihood of observed postsynaptic spiking . Namely , we update the plasticity parameters with fixed GLM parameters and then update the GLM parameters with fixed plasticity parameters , alternating between the two optimization problems until the maximum is achieved . The TM formalism assigns a weight to each spike of the presynaptic neuron , while the GLM parameters characterize the influence of prior postsynaptic spiking and coupling between pre- and postsynaptic activity ( as scaled by the TM weights ) . To facilitate convergence of the TM and the GLM parameters we impose prior constraints on both these parts of the model ( see Methods ) . Using pre- and postsynaptic spike trains , we thus obtain estimates of both traditional GLM parameters and a complete set of parameters θ = {D , F , U , f} describing short-term plasticity in the TM model . We fit the TM-GLM separately for each simulated connection in our in vitro recording . The 96 simulated presynaptic inputs had different weights and different types of STP , and our goal is to compare how these synaptic properties affect estimation of STP . Specifically , we have six sets of parameters corresponding to strong depression , depression , depression/facilitation , facilitation , strong facilitation , and a control set with no plasticity ( Table 1 ) . Although the optimization of the TM parameters is not convex , we find that , after adding informative priors ( see Methods ) the global optimum can be quickly found using random restarts . TM-GLM estimates of the time constant for depression D and the release probability U are closer to underlying true values than the estimates of the facilitation time constant F and its magnitude f . Fig 6A shows results of bootstrapping to estimate the parameter uncertainty for the different types of plasticity . Note that high variability in the estimation of facilitation parameters is not a specific drawback of our model , but represents a more general problem . Indeed , previous work showed that estimates of facilitation parameters were non-precise even when direct measurements of postsynaptic responses , PSPs or PSCs ( and not postsynaptic spikes as used in our model ) were fitted [17 , 22] . Particularly for depressing synapses ( where U is large and F is small ) , the estimation of f is not well-posed . In this case , it may make more sense to use a more restricted TM model with fewer parameters [15 , 16] or to use a fully Bayesian approach where the posterior can be more completely assessed . More generally , the difficulty of estimating facilitation parameters might be a consequence of a relatively weaker effect of facilitation on postsynaptic activity as compared to depression . This interpretation is supported by the observation that despite the deviation of estimated parameters of facilitation from the true value , the model with the estimated parameters accurately predicts the steady-state filtering properties of dynamic synapses ( Fig 6C ) . Note that some of the bias in parameter estimation may be due to the choice of priors . Here we chose our priors to avoid local minima in the posterior that occur near the edges of parameter space , where F or f are close to zero . However , as the number of observations increases these biases will be reduced , since likelihood will have a larger impact on the posterior than the prior . In general , the accuracy and confidence of the estimates will be affected by many factors , such as , the number and pattern of presynaptic spikes , number of postsynaptic spikes , the synaptic weight , and the type of STP . For large-scale analysis of STP in neuronal networks it might be important to distinguish between different types of plasticity at a synapse ( e . g . facilitating vs depressing ) and attribute certain types of plasticity to different classes of synaptic connections , rather than to extract the exact parameter values for each synapse . Again , although the problem is not convex , we find that the different types of plasticity can be distinguished based on spiking observations alone . For the 5 strongest excitatory inputs with each type of plasticity we compare the likelihood under the different settings of the TM parameters used in the recording [Fig 6B] . This analysis treats the problem of STP-identification as a classification problem . If the data do not provide a clear indication of the type of STP , e . g . for very weak synaptic inputs which have little effect on postsynaptic spiking , then the likelihood should be similar under all models–both facilitating and depressing . However , here we find that the true parameters do have the highest likelihoods , with depressing inputs having high likelihoods under the depressing model and facilitating inputs have high likelihoods under the facilitating model . Additionally , even though the estimated parameters may differ from their true values , the ( steady-state ) synaptic dynamics of the estimated models typically matches the dynamics of the true models [Fig 6C] . Depressing synapses show characteristic low-pass filtering , while facilitating synapses have band-pass filtering with cutoff frequencies depending on the exact TM parameters . The TM-GLM estimates the short-term dynamics of a synapse described with biophysically realistic parameters that are related to the vesicle and calcium dynamics . In many cases , however , it might be useful to detach the description of the coupling between pre- and postsynaptic spiking from the biophysics of synaptic dynamics at an individual synapse . To describe neuronal interactions in terms of ISI-dependent modifications , we introduce a generalized bilinear model ( GBLM , Fig 2 ) that captures functional changes in the synaptic efficacy for different presynaptic intervals . In this model , the coupling term changes as function of presynaptic spiking , e . g . at facilitating synapses it increases for short ISIs , and at depressing synapses it decreases for short ISIs . We use basis splines to fit a smooth modification function ( see Methods ) that describes how the coupling term has been adjusted following different presynaptic intervals . We further assume that the effect of the modification is transient , decaying exponentially [Fig 2] . Compared to the TM-GLM , the GBLM has simplified description of the dynamics of coupling but provides a more explicit characterization of the effects of different ISIs on the modification of the coupling term . The GBLM provides clearly distinct estimates of the modification functions for synaptic connections with different types of short-term plasticity [Fig 7] . For simulated inputs expressing the same type of STP , but having different weights ( among strongest 3 ) or different signs ( excitatory and inhibitory ) , the estimates of the modification functions were similar . These modification functions were estimated by maximizing the regularized log-likelihood . For stability , the spline basis was designed to have no effect on very short or very long ISIs where there is typically little data . However , for depressing synapses the modification function decreases the relative synaptic strength for ISIs between 0 and 1s , and for facilitating synapses the modification function increases the relative synaptic strengths . Both the TM-GLM and the GBLM accurately describe split cross-correlograms for all examined types of STP , and for both excitatory and inhibitory inputs for the in vitro experiment [Fig 8A] . However , in addition to the spike statistics we can also compare how well the models reconstruct the time-varying individual PSC amplitudes . After estimating the plasticity dynamics for each simulated input using the TM-GLM ( npre* ) and the GBLM ( w ( t ) ⊙ npre ) we then calculate correlations between the true PSC amplitudes and the estimated amplitudes under the two models [Fig 8B] . We find that the weights of the simulated inputs have a substantial effect on the reconstruction of PSC amplitudes . The estimated amplitudes at strong synapses ( both excitatory and inhibitory ) are reconstructed much more accurately than amplitudes at the weak synapses . Additionally , we find that the PSCs of depressing synapses are much more reliably reconstructed than PSCs of facilitating synapses ( r = 0 . 95±0 . 01 for synapses with strong depression vs . r = 0 . 34±0 . 06 for synapses with strong facilitation ) . This is consistent with our observation that the PSCs of depressing synapses are more reliably related to ISIs compared to facilitating synapses [Fig 5] . Finally , the TM-GLM model appears to consistently out-perform the GBLM ( average correlation for the TM-GLM across all types of plasticity and weights is r = 0 . 70±0 . 03 compared to r = 0 . 52±0 . 03 for the GBLM ) . In vitro recordings of responses to simulated presynaptic spikes have the advantage that the postsynaptic spikes are generated by the biophysics of a real neuron . However , estimation of STP from spike trains recorded in the intact brain in vivo may be compromised by several additional factors , not considered in this controlled experimental setting . Below we will analyze possible effects of four such factors on STP estimation: spike frequency adaptation , stochastic release of transmitter , uncertainty of spike sorting , and correlated common input . To examine how these sources of variability may affect the estimation of short-term synaptic plasticity from spikes we simulated postsynaptic spike trains using leaky integrate-and-fire model neurons receiving synaptic inputs with defined STP in the presence of noise . For simplicity , we focus on model synapses with strong depression , strong facilitation , and no plasticity ( Table 1 ) . One factor that affects postsynaptic firing is spike frequency adaptation . In particular , an after-hyperpolarization ( AHP ) current mediating fast spike frequency adaptation can change the pattern of postsynaptic firing and may act to mask the influence of presynaptic STP on generation of postsynaptic spikes . To test if our models can differentiate the effects of AHP currents ( IAHP ) , which alter the dynamics of the postsynaptic neuron , from the effects of short-term synaptic plasticity , we simulated two leaky integrate-and-fire ( LIF ) neurons with and without an IAHP [39] ( see Methods ) . In response to a long depolarizing pulse , the LIF neuron without an IAHP fires at a stationary rate . The LIF neuron with the IAHP , on the other hand , rapidly adapts–with a firing rate peaking immediately after the depolarization onset and gradually decreasing to a lower steady-state . After stimulus offset the firing rate of the adapting LIF decreases below the pre-stimulus level [Fig 9A] . These effects are not due to synaptic dynamics but reflect the dynamics of the postsynaptic neuron itself . We simulated pre- and postsynaptic spike trains using the LIF model neurons ( with and without an IAHP ) receiving inhomogeneous Poisson input with short-term synaptic dynamics governed by the TM model and applied our models to estimate STP from these spike trains . Results from the TM-GLM and GBLM for the two leaky IF neurons show that the adaptation properties mediated by IAHP current are mostly captured in the post-spike history filters [Fig 9B and 9C] . For connections with depression , facilitation , or no STP , the estimated TM parameters and the modification functions estimated with the GBLM are similar with and without the IAHP . Although frequency adaptation occurs on a similar timescale to short-term synaptic plasticity , the methods here thus seem to be able to distinguish purely postsynaptic dynamics from the time-varying effect of the presynaptic neuron on the postsynaptic neuron . One further potential source of noise that is not included in the Tsodyks-Markram model , and that was not accounted for in our experiments in slices , is stochastic vesicle release . Although the TM model and the GBLM treat the synaptic transmission as deterministic and the PSC/PSP amplitudes can take any value , in real synapses PSC/PSP amplitudes are fundamentally stochastic with vesicles being probabilistically released from a limited number of sites . Compared to our in vitro experiments using the deterministic release , it may be more difficult to estimate STP parameters from the spiking of real neurons with stochastic release . To study how stochastic release impacts the estimation of STP parameters , we use a quantal model of synaptic plasticity [40 , 41] . In this model , the resources of the TM model are discretized based on the number of release sites and are then released according to a Binomial distribution with a time-varying probability given by the utilization variable of the TM model ( see Methods ) . We simulated pre- and postsynaptic spike trains from LIF model neurons driven by inhomogeneous Poisson input with synaptic dynamics governed by the quantal TM model . The amplitudes of the postsynaptic currents are now noisy rather than deterministic functions of the presynaptic spike timing . In our simulations , increasing the number of release sites decreases the variance of the PSC amplitudes . For depressing synapses , stochastic release leads to a systematic bias in the estimates of the TM model parameters compared to their values under deterministic release [Fig 10A] . For facilitating synapses , on the other hand , the TM parameter estimation was not substantially affected . Similarly , the modification functions estimated with GBLM for depressing synapses were changed as the number of release sites is varied , while the modification functions for facilitation are more stable . Both the TM-GLM and GBLM can still distinguish between depression and facilitation , but considering stochastic release may be necessary for accurate parameter estimates in vivo . Another potential source of uncertainty , that may affect the estimation of synaptic dynamics from spikes , is imperfect spike sorting . In practice spike sorting from in vivo recordings is not a perfect process , and inaccuracies in spike sorting can lead to biased estimates of neural response properties [42] . Here , we simulated presynaptic and postsynaptic spike trains using LIF model neurons with strongly depressing or facilitating dynamics on inhomogeneous Poisson input ( as above , See Table 1 for parameters ) . We then simulated the effects of imperfect spike sorting by randomly deleting and inserting spikes into both the pre- and postsynaptic spike trains before estimating STP . For insertion , we randomly selected spikes from two other inhomogeneous Poisson neurons ( same baseline firing rates ) and assigned the spikes to pre- and postsynaptic neurons . For both the TM-GLM and GBLM we find that the imperfect assignment of spikes ( both addition and deletion ) results in only small biases in the estimation of STP parameters for connections with strong facilitation and depression [Fig 10B] . Despite these small biases , we were able to distinguish between facilitation and depression even as the proportion of spike sorting errors becomes large ( 20–40% insertion/deletion ) . In vivo , neurons often have common synaptic input from unobserved sources . Common input introduces correlations in pre- and postsynaptic spiking that are not due to synaptic connections between the recorded neurons . In cortical networks , for instance , the strength of the common input can vary according to the level of synchronous activity with weak common input in asynchronous , irregular network states and strong common input in synchronous states [43] . To study how such correlations would affect STP estimation we simulated a microcircuit with different levels of synchrony . In this simulation , two presynaptic neurons receive input from three sources: 1 ) a private , slowly fluctuating current , 2 ) a shared/common , slowly fluctuating current , and 3 ) an independent white noise current . The postsynaptic neuron receives the common input , an independent white noise current , and inputs from each presynaptic neuron–one with a depressing synapse and one with a facilitating synapse [Fig 11A] . We then vary the strength of the common input using a weight parameter w , which determines how much of each neuron’s input is originating from the shared/common source and how much of the input comes from the private current . As the weight of the common input increases there is a short-term synchronization between the spiking of all neurons [Fig 11B] . At low ( w = 0 . 25 ) and medium ( w = 0 . 5 ) common input both the TM-GLM and GBLM were able to discriminate between depressing and facilitating inputs , but at w = 0 . 75 neither model was able to distinguish between the depressing and facilitating input . This simulation demonstrates that , at least in some situations , strong common input can cause both models to fail to estimate underlying short-term synaptic plasticity . Intracellular recordings in brain slices have revealed a diversity of STP across cell types and anatomical connections [2 , 3] . Moreover , the details of STP at a given type of synapse may change depending on a multitude of factors , such as changes during development [44] , neuromodulation [45] , or induction of long-term plasticity [13] . Because STP critically affects information processing , understanding operation of neuronal networks during natural behavior requires large-scale analysis of STP in vivo . However , since large-scale intracellular recordings are not feasible in vivo , alternative methods are necessary for such studies . Large-scale extracellular recordings , on the other hand , are feasible in vivo . Existing techniques allow simultaneous recording of spiking of hundreds of neurons , and this number appears to be growing exponentially [27] . Characterizing short-term plasticity using spike observations is more difficult than using intracellular ( PSC/PSP ) signals , but short-term synaptic plasticity does have observable effects on spike statistics . Prior evidence for STP in vivo obtained from spike trains alone employed a split cross-correlogram approach , in which the postsynaptic response to presynaptic spikes following short ISIs was compared to that following long ISIs . Several studies using this approach analyzed strong thalamocortical connections and found evidence for both short-term facilitation and depression [30–32] . To the best of our knowledge , however , the split cross-correlogram approach has not revealed evidence of short-term plasticity in weaker synapses , such as corticocortical connections . Here we introduce two new model-based methods to characterize short-term synaptic plasticity from pre- and postsynaptic spiking . By explicitly modeling synaptic dynamics these models are able to recover a detailed description of short-term plasticity . These models reproduce the results from split cross-correlograms ( Fig 8 ) , but also provide an explicit characterization of the dynamics of STP and allow reconstruction of PSP amplitudes for each presynaptic spike . To validate our methods , we used spiking of layer 2/3 pyramidal neurons in vitro induced by injection of a current composed of PSCs from an artificial population of presynaptic neurons , whose spiking and plasticity parameters are known . Even though each presynaptic input represents only a small fraction of the total injected current , we can accurately estimate the synaptic dynamics from pre- and postsynaptic spiking . In this setting , both model-based methods , the TM-GLM and GBLM , can robustly distinguish between different types of STP , and can reconstruct PSP amplitudes for a wide range of synaptic weights for both excitatory and inhibitory connections . The TM-GLM provides a compact description of STP with four parameters related to the vesicular release and calcium dynamics in the presynaptic terminal . The GBLM provides a functional description of how the synaptic weight changes as a function of presynaptic ISIs . An advantage of the GBLM approach is that the synaptic modification rule is not constrained by the biophysics of single synapses , but has the potential to capture more complex dependences , including polysynaptic effects . One further advantage of the GBLM over the TM-GLM model is that the synaptic dynamics are assumed to be linear , which increases both the speed and robustness of the optimization process . Depending on whether a functional or a biophysical description is required , the two methods may thus both be useful tools for large-scale characterization of short-term synaptic plasticity from spiking activity . Estimating synaptic plasticity from in vivo multi-electrode recordings of spiking activity will introduce several additional challenges . One challenge is simply detecting the connections between neurons . Strong monosynaptic connections are typically expressed in cross-correlograms as clear peaks ( or troughs , for inhibition ) with short latency and sharp onset , but weak connections or connections between neurons with low firing rates are difficult to detect in cross-correlograms . In previous work , we showed that model-based approaches can increase the sensitivity of detection for weak connections compared to traditional cross-correlation approaches [46] , and the GLM-based approaches here are likely to have similar advantages . A second challenge is that short-term synaptic plasticity isn’t the only source of variation in the observed postsynaptic responses to presynaptic spikes . Changes in the excitability of the postsynaptic neuron , stochasticity of vesicle release , and spike sorting errors can alter the statistics of the response and could potentially bias our estimates of short-term synaptic plasticity . To study how these sources of variability affect estimation of STP parameters we simulated spike trains of connected leaky integrate-and-fire model neurons , and introduced each of these confounding variables individually . We found that adding an after-hyperpolarization current ( IAHP ) to the postsynaptic neuron impacts only the post spike-history filters in both the TM-GLM and GBLM , and does not substantially change STP estimation . Stochastic vesicle release and spike sorting errors , on the other hand , lead to biases in the estimation of short-term synaptic plasticity for our models . However , even with these additional noise sources , both the TM-GLM and GBLM are still able to reliably distinguish between connections expressing short-term facilitation and depression . A third challenge is that correlations between the spiking of two neurons may be produced by common input rather than , or in addition to , the synaptic connection between the neurons . In our experiments with current injection in neurons in slices , inputs were generated as independent inhomogeneous Poisson processes , without the correlations that are present in vivo . To understand how correlated spiking can affect STP estimation , we simulated a small , feed-forward network of neurons with common input . We found that as the common input becomes stronger , the synchronization between pre- and postsynaptic spikes can interfere with the estimation of STP . The TM-GLM and GBLM were able to estimate synaptic dynamics only when common input was weak , but failed to accurately estimate the underlying synaptic dynamics for neurons with strong common input . While our in vitro experiment and simulations allowed us to compare STP estimation under controlled conditions with known synaptic dynamics , more work may thus be needed to account for all the dependencies that occur between pre- and postsynaptic neurons in vivo . Finally , a fourth challenge is that the assumptions of the TM model itself do not necessarily describe the dynamics of all interactions between the pre- and postsynaptic neurons . The TM model only aims to describe presynaptic mechanisms of STP . However , postsynaptic factors such as desensitization or saturation of postsynaptic receptors may play a role in STP at some synapses , and the synaptic weight may vary on other timescales ( e . g . due to LTP/LTD ) . Replacing the TM model used here with alternative models of plasticity may be a tractable approach to address these challenges [18 , 47 , 48] . For instance [17] shows that in many cases the original three-parameter Tsodyks-Markram model is sufficient to describe STP . Alternatively , since the GBLM is not constrained by single-synapse biophysics , it may , in some cases , provide a more flexible first-order description of short-term dynamics , including those that are not well described by the TM model . Rather than describing anatomical connectivity , the two model-based methods introduced here describe the plasticity of functional interactions between neurons . Many of the techniques that have been used to improve models of functional connectivity without plasticity can be used to improve the TM-GLM and the GBLM presented here . For instance , it may useful to model multiple inputs simultaneously or to include latent common input in the model [49–51] . More structured regularization techniques may allow more accurate reconstruction with smaller sets of data [52 , 53] . To improve models of synaptic dynamics it may be useful to consider additional timescales [19] , a higher-order expansion of the ISI dependencies , or other types of plasticity occurring on longer time scales , such as spike-timing dependent plasticity [54–56] . Applying these methods in vivo may then allow us to characterize short-term plasticity during natural behavior and in larger populations than previously possible . All animal use procedures conform to the principles outlined in the Guide for the Care and Use of Laboratory Animals ( National Institutes of Health publication no . 86–23 , revised 1985 ) and were approved by the Institutional Animal Care and Use Committee at the University of Connecticut . Approaches using generalized linear models ( GLMs ) have proved to be effective tools for estimation neuronal connections from spike train data [57–59] . The standard GLM assumes that the spike train is a binary sequence of observations , m ( t ) , generated from a Poisson process . For a single pair of neurons , we model the conditional intensity , λ ( t ) , of this process as a linear combination of a baseline firing rate μ , a contribution from the presynaptic neuron rxt and weighted contribution from the postsynaptic spike-history syt passed through an exponential nonlinearity ( Fig 1 ) . λ ( t|μ , r , s ) =exp⁡ ( μ+rxt+syt ) m ( t ) ∼Poisson ( λ ( t|μ , r , s ) ) xt=[x1 ( t ) , x2 ( t ) , … , xL ( t ) ] , xj ( t ) =n ( t ) *bj ( t ) yt=[y1 ( t ) , y2 ( t ) , … , yL ( t ) ] , yj ( t ) =m ( t ) *bj ( t ) ( 1 ) where n ( t ) and m ( t ) are the pre- and postsynaptic spike trains , respectively . Our goal is to estimate the set of model parameters r=[βc ( 1 ) , βc ( 2 ) , … , βc ( L ) ] , s=[βh ( 1 ) , βh ( 2 ) , … , βh ( L ) ] and μ , describing the coupling , k ( t ) , and post-spike filter , h ( t ) , which best predicts the postsynaptic firing m ( t ) . k ( t ) =∑j=1Lβc ( j ) bj ( t ) ;h ( t ) =∑j=1Lβh ( j ) bj ( t ) ;bj ( t ) =12cos⁡ ( log⁡ ( t+Cj ) +π ) +12 ( 2 ) where bj ( t ) are raised-cosine basis functions which reduce dimensionality and allow a smooth representation of the two filters [34] . This stochastic model of a Poisson spiking neuron has a guaranteed convex log-likelihood which gives a unique set of parameters for its global maximum [37] . In order to model plasticity , we modified the GLM , allowing the contribution of coupling to vary over time . A conventional GLM treats all presynaptic spikes n ( t ) , equally , with each presynaptic spike having the same “weight” when influencing conditional intensity , λ ( t ) . To account for short-term facilitation and depression we modify the weights of each spike according to the phenomenological Tsodyks-Markram ( TM ) model [16] . This four parameter model , sometimes referred to as the extended Tsodyks-Markram model , describes the dynamics of resources R and their utilization u by the following system of differential equations: dR ( t ) dt=1−R ( t ) D−u ( t− ) R ( t− ) δ ( t−ts ) du ( t ) dt=U−u ( t ) F+f[1−u ( t− ) ]δ ( t−ts ) ( 3 ) where resources , R ( t ) , represent the portion of available vesicles which instantly decreases after each spike at ts and gradually recovers with depression time constant D . The second equation describes release probability ( utilization of resources ) , which instantly increases after each spike by f[1−u ( t− ) ] , where f is the magnitude of facilitation and decays back to the baseline value , U , with facilitation time constant F . The amplitude of the postsynaptic current Isyn ( ts ) evoked by presynaptic spike at ts is described by Isyn ( ts ) =AR ( ts ) u ( ts ) ( 4 ) where A is the maximal current that can be evoked at that synapse if all resources are recovered ( R = 1 ) and are released at once . With different sets of parameters θ = {D , F , U , f} this model can reproduce diverse types of short-term plasticity ( depression , facilitation or a mixture of both ) observed experimentally [17] . Using these dynamics , we create a “marked” point-process n* ( t ) =AR ( t ) u ( t ) n ( t ) ( 5 ) where n* ( t ) captures the amplitudes of PSCs at the time-points of presynaptic spikes and is zero otherwise . By using n* ( t ) instead of n ( t ) in the modified GLM ( TM-GLM ) , we account for STP in the coupling term . Note that when n* ( t ) is constant ( R ( t ) and u ( t ) constant ) the TM-GLM will describe a steady-state synapse with no short-term plasticity , and , in this case , TM-GLM is identical to the original GLM . With the modified coupling term the original observation model is rewritten as λ ( t|μ , r , s ) =exp⁡ ( μ+rxt*+syt ) m ( t ) ∼Poisson ( λ ( t|μ , r , s ) ) ( 6 ) Using the TM-GLM our goal is to estimate the static parameters of the synapse ϕ = {μ , r , s} , as well as the plasticity parameters θ = {D , F , U , f} , given the pre- and postsynaptic spike trains . Specifically , we aim to find maximum a posteriori ( MAP ) estimates of θ and ϕ that optimize p ( θ , ϕ|npost ( t ) , npre ( t ) ) ∝p ( npost ( t ) |θ , ϕ , npre ( t ) ) p ( θ , ϕ|npre ( t ) ) =p ( npost ( t ) |θ , ϕ , npre ( t ) ) p ( θ ) p ( ϕ ) ( 7 ) To prevent over-fitting and assure nonnegative values , we introduce weakly informative priors on the plasticity parameters p ( θ ) to span the parameters space only over meaningful intervals and prevent the optimization from getting stuck at local minima . We then use coordinate ascent , maximizing the log posterior by alternating between optimizing the plasticity parameters given the GLM parameters and fitting the GLM given fixed plasticity parameters . Although this posterior is not guaranteed to be convex , in many cases , the non-convexity of GLM-like models does not lead to optimization problems [60–62] . Previous work estimating STP parameters from intracellular recordings suggests that , rather than point estimates , a fully Bayesian approach may provide a more accurate understanding of the parameters [17 , 40] . Although it is possible to use MCMC to sample from the posterior , the large number of function evaluations ( compared to optimization ) makes it less attractive for our model with spike observations . When optimizing plasticity parameters ( GLM parameters fixed ) we randomly restart over the θ-space and use priors {0 < D , F < 2} ∼ gamma ( α = 1 . 2 , β = 2 ) and {0 < U , f < 1} ∼ beta ( 1 . 01 , 1 . 01 ) . We optimize the plasticity parameters in the log-domain using two-metric projection and numerical differentiation of the posterior [63] . Additionally , we found that when optimizing the plasticity parameters , convergence is improved by normalizing the static coupling term k ( t ) and optimizing an amplitude A ( with prior p ( A ) = cauchy ( 0 , 50 ) ) alongside the parameters θ . These prior distributions and parameters were chosen to prevent the model from reaching the boundaries ( e . g . U or f at 0 or 1 ) , but they do introduce bias into the parameter estimates and may not necessarily work well for all sets of data . When optimizing the static GLM parameters ϕ ( plasticity parameters fixed ) we would typically assume p ( ϕ ) to be flat . However , we found that in some cases the coupling term k ( t ) interacts with the plasticity parameters . For instance , an excitatory depressing synapse will show a biphasic coupling term where a negative component can partially account for the reduced impact of a burst . To prevent this type of ambiguity we introduced a quadratic penalty on negative coupling coefficients r with the improper prior logp ( r ) ∝ −ηr2 1r<0 . In practice , we use LBFGS optimization of the penalized log-likelihood and this ensures that the estimated coupling term is approximately positive for excitatory inputs and negative for inhibitory inputs . With limited data or when extending the model to multiple inputs additional types of regularization may be useful [58] . The phenomenological model , described above , gives a clear view of the synaptic dynamics by searching over the θ-space of STP parameters . However , in cases were TM assumptions on synaptic dynamics such as vesicle release and changes of the calcium changes in presynaptic terminal doesn’t hold , it may be preferable to have a model of STP that is not constrained to the TM dynamics . In a second type of model–the generalized bilinear model—instead of searching over the space of STP parameters we directly infer a short-term synaptic modification “rule” . This generalized bilinear model ( GBLM ) compartmentalizes the coupling term into a stationary and a short-term plastic modification ( Fig 2 ) . Here the modification term , w ( t ) , weights the static coupling term depending on the history of presynaptic spiking . For a synapse with no plasticity , w ( t ) equals to one and the coupling term , rxt is static and does not depend on previous presynaptic spiking . For a synapse with plasticity , w ( t ) >1 would increase the static coupling term rxt to account for facilitation and w ( t ) <1 would decrease the rxt to account for depression . In both cases , the effect of w ( t ) on the coupling term decays with time and the coupling term recovers to its static form . We defined the modification function as: w ( t ) =1+∑kqk∑l=0Tδ ( tk−l ) exp ( −lτ ) qk=αBm ( Δtk ) ( 9 ) where q ( ∙ ) determines the amplitude of exponentially decaying effects from previous spikes on the synaptic weight . Here k indexes the presynaptic spikes with times tk and previous inter-spike intervals Δtk . Although we could attempt to fit the decay function ( instead of using single exponential ) and its time-constant ( τ = . 2 ) we fixed them to increase the robustness and speed of the maximum likelihood parameter fitting . Spikes are then convolved with the exponential kernel weighted by the modification terms q ( ∙ ) . To ensure q ( ∙ ) is a smooth function we represent it using the B-spline bases , Bm ( Δt ) , with log-spaced sampling knots in Δtk . The final model is linear in both the stationary parameters , {μ , r , s} , and STP parameters , q . To estimate the parameters , we alternate between two GLMs: fitting {μ , r , s} with fixed q and fitting q with fixed {μ , r , s} ( both using iterative reweighted least squares—IRLS ) . Although the two GLMs are log-concave in this problem , the joint likelihood of {μ , r , s , q} is not guaranteed to be concave . However , we find that in practice convergence is fast using the alternating method and random restarts results in the same final solution . Slices of visual cortex were prepared from male Wistar rats ( P21-P23 ) as described in detail in our prior work [38] . Extracellular solution used during preparation of slices and for perfusion of recording chamber contained ( in mM ) : 125 NaCl , 2 . 5 KCl , 2 CaCl2 , 1 MgCl2 , 1 . 25 NaH2PO4 , 25 NaHCO3 , 25 D-glucose and was bubbled with 95% O2 and 5% CO2 . Patch clamp electrodes for whole cell recordings were filled with K-gluconate based solution ( in mM: 130 K-Gluconate , 20 KCl , 4 Mg-ATP , 0 . 3 Na2-GTP , 10 Na-Phosphocreatine , 10 HEPES ) and had a resistance of 4–6 MΩ . Whole-cell recordings were made from layer 2/3 pyramidal neurons of rat visual cortex . Membrane potential responses to injection of fully-defined fluctuating current ( [38]; see below ) were recorded using the bridge mode of a Dagan BVC-700A amplifier ( Dagan Corporation , USA ) . Data were digitized at 20 kHz ( Digidata 1440A , Molecular Devices , USA ) and stored for further processing . Timings of postsynaptic spikes were determined as positive-slope zero crossings of the membrane potential signal . An artificial current for injection was designed to mimic the postsynaptic effect of a realistic cortical circuitry with inputs of different strength and unique short-term synaptic characteristics ( Fig 3 ) . Current for injection was synthesized using a population of 96 presynaptic neurons ( 6 pools of 16 neurons , 8 excitatory and 8 inhibitory ) . Five sets of STP parameters were chosen to cover the whole spectrum of the plasticity from strong depression to strong facilitation . The sixth set of synapses did not express short-term plasticity . For each neuron , we generated an inhomogeneous Poisson spiking series with the log rate generated using a cubic spline function with 1 knot/s and standard normally distributed amplitudes . The rate is then scaled to generate an average spike rate of 5Hz and the spikes are weighted to generate the postsynaptic current amplitudes of the TM model . The weighted series of postsynaptic current amplitudes was then convolved with a synaptic integration kernel to generate the artificial postsynaptic current traces . We generated the kernel as a difference of two exponentials with time constants of 1ms and 10ms . Eight different synaptic weights with a normal inverse cumulative distribution function ( μ = . 7 & σ = . 93 ) were used to create a pool of excitatory synapses . Same synaptic weights , but with a negative sign were used to generate currents produced by inhibitory neurons . Because the number and weight distributions for excitatory and inhibitory presynaptic neurons were the same , the total input current was balanced . We used 20 different realizations of the current for injection . The duration of each current trace was 46s . Injections of fluctuating currents were separated by intervals of 60–100s . The amplitude of the injected current was adjusted to produce membrane potential fluctuations of 10–15 mV . DC current was added to achieve the average postsynaptic firing rate of ~5Hz . Thus , we knew the timing of presynaptic spikes for each simulated presynaptic neuron contributing a synaptic connection as well as its amplitude and the parameters governing its short-term plasticity . We used individual pairs of pre- and postsynaptic spike trains to compare the parameters of short-term plasticity , estimated by the models , to the ground truth values . To examine the limitations of our models more thoroughly , we simulated a leaky integrate-and-fire model neuron receiving presynaptic input with short-term synaptic plasticity . In particular , to examine the effect of spike frequency adaptation we simulate a postsynaptic neuron with and without an after-hyperpolarization current [39]: τmdVmdt=− ( Vm−Erest ) −rmgsra ( t ) ( Vm−Ek ) +rmI ( t ) τsradgsra ( t ) dt=−gsra ( t ) ifVm=Vththen{Vm→Vresetgsra ( t ) →gsra ( t ) +Δgsra ( 10 ) where Ek is the reversal potential due to K+ , gsra ( t ) is the spike rate adaptation conductance , which changes with rate Δgsra = 200nS , and Ek = 80 mV is the reversal potential . The other parameters were set to Erest = 80mV , Ek = 80mV , Vreset = 80mV , Vth = 54mV , τm = 10ms , τsra = 100ms , and rm = 10MΩ . Similar to the in vitro experiment above , I ( t ) is synthesized by simulating a presynaptic input with short-term synaptic plasticity ( inhomogeneous Poisson spiking with Tsodyks-Markram PSC amplitudes ) . We then adjust the DC current , noise , and synaptic strength to achieve the desired postsynaptic spike rate ( 5Hz ) along with a cross-correlogram similar to those obtained by the strongest synapses in the in vitro experiment . Although the TM model treats short-term synaptic plasticity as a deterministic process , synaptic transmission is a discrete , stochastic process where a discrete number of vesicles are present and probabilistically released following presynaptic spikes . To model this additional variability , we use LIF simulations , as above , where rather than having PSC amplitudes be synthesized from the TM model we use a quantal , stochastic extension of the TM model . First , to make the TM model discrete , we consider an integer number of release sites , nmax , where , at any point in time , only a fraction of resources are available to be released , am = ⌊nmax Rm⌋ . Following each presynaptic spike , a discrete number of vesicles is released km∼Binomial ( am , um ) ( 11 ) giving the PSC amplitudes Im=kmnmax ( 12 ) Following a spike , the resources and utilization at the following spike ( after interval Δt ) are given by Rm+1=1− ( 1–am−kmnmax ) e−ΔtDum+1=U+ ( um+f ( 1–um ) −U ) e−ΔtF ( 13 ) The codes for both TM-GLM and GBLM are available at https://github . com/abedghanbari2/stsp .
Information processing in the nervous system critically depends on dynamic changes in the strength of connections between neurons . Short-term synaptic plasticity ( STP ) , changes that occur on timescales from milliseconds to a few seconds , is thought to play a role in tasks such as speech recognition , motion detection , and working memory . Although intracellular recordings in slices of neural tissue have identified synaptic mechanisms of STP and have demonstrated its potential role in information processing , studying STP in intact animals , especially during behavior , is experimentally difficult . Unlike intracellular recordings , extracellular spiking of hundreds of neurons simultaneously can be recorded even in behaving animals . Here we developed two models that allow estimation of STP from extracellular spike recordings . We validate these models using results from in vitro experiments which simulate a realistic synaptic input from a population of presynaptic neurons with defined STP rules . Our results show that both new models can accurately recover the synaptic dynamics underlying spiking . These new methods will allow us to study STP using extracellular recordings , and therefore on a much larger scale than previously possible in behaving animals .
[ "Abstract", "Introduction", "Results", "Discussion", "Methods", "and", "models" ]
[ "action", "potentials", "medicine", "and", "health", "sciences", "vesicles", "nervous", "system", "membrane", "potential", "electrophysiology", "neuroscience", "synaptic", "plasticity", "cellular", "structures", "and", "organelles", "neuronal", "plasticity", "mood", "disorders", "developmental", "neuroscience", "animal", "cells", "postsynaptic", "potentials", "mental", "health", "and", "psychiatry", "cellular", "neuroscience", "cell", "biology", "anatomy", "synapses", "physiology", "neurons", "biology", "and", "life", "sciences", "cellular", "types", "depression", "neurophysiology" ]
2017
Estimating short-term synaptic plasticity from pre- and postsynaptic spiking
Cytomegalovirus ( CMV ) infection elicits a very strong and sustained intravascular T cell immune response which may contribute towards development of accelerated immune senescence and vascular disease in older people . Virus-specific CD8+ T cell responses have been investigated extensively through the use of HLA-peptide tetramers but much less is known regarding CMV-specific CD4+ T cells . We used a range of HLA class II-peptide tetramers to investigate the phenotypic and transcriptional profile of CMV-specific CD4+ T cells within healthy donors . We show that such cells comprise an average of 0 . 45% of the CD4+ T cell pool and can reach up to 24% in some individuals ( range 0 . 01–24% ) . CMV-specific CD4+ T cells display a highly differentiated effector memory phenotype and express a range of cytokines , dominated by dual TNF-α and IFN-γ expression , although substantial populations which express IL-4 were seen in some donors . Microarray analysis and phenotypic expression revealed a profile of unique features . These include the expression of CX3CR1 , which would direct cells towards fractalkine on activated endothelium , and the β2-adrenergic receptor , which could permit rapid response to stress . CMV-specific CD4+ T cells display an intense cytotoxic profile with high level expression of granzyme B and perforin , a pattern which increases further during aging . In addition CMV-specific CD4+ T cells demonstrate strong cytotoxic activity against antigen-loaded target cells when isolated directly ex vivo . PD-1 expression is present on 47% of cells but both the intensity and distribution of the inhibitory receptor is reduced in older people . These findings reveal the marked accumulation and unique phenotype of CMV-specific CD4+ T cells and indicate how such T cells may contribute to the vascular complications associated with CMV in older people . The effective control of infectious agents requires a range of different arms of the immune system . CD4+ T cells play a pivotal role in orchestrating these events , including support for both antibody production and the expansion and effector function of CD8+ T cells . However it is now well established that CD4+ T cells can also exert crucial effector functions which may be mediated by cytokine production or direct cytotoxicity [1–4] . In chronic viral infections such as cytomegalovirus ( CMV ) these effector functions are important for control of lytic replication and suppression of viral reactivation . Human leukocyte antigen ( HLA ) class I tetramers have made a huge contribution to the study of antigen-specific CD8+ T cell immune responses through their ability to allow the visualisation and phenotypic analysis of cells isolated directly from blood and tissue [5] . In contrast , the study of antigen-specific CD4+ T cells has been limited by the relative lack of HLA class II tetramers . Although virus-specific CD4+ T cells can be detected relatively easily by their functional response following exposure to antigen , this alters their phenotype and transcriptome and does not permit analysis of the resting cellular profile . As such , much less is known about the profile and function of antigen-specific CD4+ T cells . CMV is a β-herpesvirus that establishes a lifelong chronic infection and which is well controlled in healthy people . Initial infection leads to the expansion of very large numbers of virus-specific T cells within peripheral blood which are maintained throughout life and increase even further with age [6–8] . The virus has evolved multiple mechanisms to evade HLA class I and class II-restricted T cell immune responses and a state of functional latency is established after infection , which is associated with intermittent episodes of viral replication ( reviewed in [9 , 10] ) . HLA class I tetramers have revolutionized the study of the CMV-specific CD8+ immune response and have been pivotal in defining the unique immunodominance of the virus , the phenotype of virus-specific cells and unique aspects of their transcriptome [11] . CMV-specific CD4+ T cells are also critical effector populations in the control of CMV infection where they maintain function of virus-specific CD8+ T cells and suppress viral replication at specific tissue sites [12–17] . Indeed , delayed reconstitution of CMV-specific CD4+ T cells correlates with viral reactivation and CMV disease in organ transplant recipients and is associated with prolonged urinary viral shedding in children undergoing primary infection [18] . In relation to murine CMV , mice lacking CD4+ T cells are impaired in their ability to clear virus from the salivary gland which is an important site of viral latency [19 , 20] . CMV-specific CD4+ T cells have been identified in vitro using cell culture and epitope screening technology . Indeed , the use of peptide pools spanning the whole viral proteome has shown a very broad and strong CD4+ T cell response against many viral proteins of which the most immunodominant are glycoprotein B ( gB ) and the major tegument component phosphoprotein 65 ( pp65 ) [21] . These studies have shown that the CMV-specific CD4+ T cell response is of unusually strong magnitude and increases further during ageing [15 , 22–24] . However , such analyses have relied on the interrogation of cells that have been stimulated with antigen for several hours prior to analysis and the almost complete absence of HLA class II tetramers has greatly limited the ability to determine the profile of virus-specific CD4+ T cells directly ex vivo . HLA class II tetramers have recently been used to identify T cells recognising influenza [25 , 26] , hepatitis C virus [27 , 28] , HIV [29] and Epstein-Barr virus [30] . Here we have used three HLA class II tetramers to carry out the first comprehensive analysis of the phenotypic and transcriptional profile of unmanipulated CMV-specific CD4+ T cells . We show that CMV-specific CD4+ T cells are found at very high frequencies within peripheral blood , exhibit a highly differentiated and cytotoxic phenotype which would target them to activated endothelium through CX3CR1-fractalkine binding . These features reveal the extraordinary magnitude of the CMV-specific CD4+ T cell pool that must be maintained to suppress viral reactivation and indicate potential mechanisms that may underlie the development of vascular disease during chronic CMV infection . Glycoprotein B and pp65 are the two most immunodominant target proteins for CMV-specific CD4+ T cells . We obtained HLA-peptide tetramers that contained three epitopes derived from gB or pp65 , the gB-derived DYSNTHSTRYV peptide restricted by HLA-DRB1*07:01 ( DR7 ) as well as two pp65-derived epitopes , AGILARNLVPMVATV and LLQTGIHVRVSQPSL , which are restricted by HLA-DRB3*02:02 ( DR52b ) and HLA-DQB1*06:02 ( DQ6 ) respectively ( Table 1 ) . These epitopes are subsequently named by the first three amino acids of their respective peptide sequence throughout this paper . To confirm the specificity of all three tetramers we initially used the reagents to stain CD4+ T cell clones specific for the cognate HLA class II-peptide complex . This confirmed strong and specific binding whilst very little background was observed following staining of peripheral blood molecular cells ( PBMCs ) from CMV-seronegative individuals who expressed the appropriate HLA allele contained within each tetramer ( Fig 1 ) . The sensitivity of detection of virus-specific T cells through use of tetramer staining was defined by mixing aliquots of peptide-specific T cell clone ( 5% , 1% , 0 . 5% , 0 . 25% and 0 . 1% ) with PBMCs taken from a CMV-seronegative individual . This approach showed that CMV-specific T cells could be identified reliably at frequencies as low as 0 . 01–0 . 05% of the total CD4+ T cell population ( Fig 1A ) We next went on to use the HLA class II tetramers to enumerate CMV-specific CD4+ T cells within the peripheral blood of healthy donors . PBMCs were isolated from 73 CMV-seropositive individuals between the age of 24 and 88 years who all expressed the appropriate HLA class II allele contained within the tetramer ( Table 1 ) . These were then stained directly with the HLA class II tetramer and analysed by flow cytometry ( Fig 1B and 1C ) . CMV-specific CD4+ T cells were observed in 74% of the donors that were tested . The median frequency of CMV-specific CD4+ T cells was 0 . 45% of the total CD4+ subset , and this varied between 0 . 75% , 0 . 21% and 0 . 66% for T cells specific for the LLQ , AGI and DYS epitopes respectively . The proportion of CD4+ epitope-specific T cells ranged from 0 . 01% up to a remarkable value of 24% of all CD4+ T cells within one individual . Interestingly , 26% of the donors ( 19/73 ) carried peptide-specific T cell populations representing over 1% of the total CD4+ T cell pool . An increase in the number of CMV-specific CD8+ T cells in association with aging , sometimes termed ‘memory inflation’ , has been demonstrated through the use of HLA-peptide tetramer staining . We therefore analysed the frequency of virus-specific CD4+ T cells in relation to age ( Fig 1D ) . Although the very largest tetramer-staining populations were indeed identified in the older donors in our study , no clear increase in CMV-specific CD4+ T cells was observed with age as many younger donors also carried substantial frequencies of CMV-specific CD4+ T cells . CMV-specific T cells have been detected previously by Interferon ( IFN ) -γ production following antigen stimulation [23 , 31] and we were interested to compare the relative number of CD4+ T cells identified by HLA class II tetramers compared to this functional response . Analysis of intracellular cytokine staining ( ICS ) for IFN-γ production after stimulation with CMV peptide was therefore performed within a panel of donors . A strong correlation was observed between these two values ( rSpearman = 0 . 83; p = 0 . 008; Fig 1E ) although it was of interest that the number of cells detected by tetramer staining was greater than the value obtained by cytokine detection . This indicates that both the number of virus-specific cells has been underestimated in previous studies using cytokine detection and that peptide-specific T cells display other functional responses in addition to production of IFN-γ . To further investigate the functional properties of CMV-specific CD4+ T cells following activation , we next went on to stimulate PBMCs with peptide prior to assessment of the profile of cytokine production using ICS . The predominant expression pattern was of combined IFN-γ , TNF-α and MIP-1β ( CCL4 ) production , with a further subset which failed to generate MIP-1β ( Fig 2 ) . Of note , the proportion of TNF-α+ cells in most donors exceeded that of IFN-γ+ cells . Interestingly , in three individuals the proportion of IFN-γ+ cells was only between 36–65% and large populations of cells were observed which produced IL-4 , usually in isolation and sometimes in combination with other cytokines . Indeed , in one donor these comprised up to 60% of peptide-specific cells . Virtually no CMV-specific CD4+ T cells produced IL-17A or IL-10 in response to antigen stimulation and no significant differences were observed between T cells recognising the gB or pp65-derived epitopes . We next undertook a more detailed analysis of the phenotype of CMV-specific CD4+ T cells . Expression of CCR7 and CD45RA was used to define naïve , central memory ( CM ) , effector memory ( EM ) and revertant CD45RA+ effector memory ( EMRA ) cells . A median of 88% of CMV-specific CD4+ T cells displayed an EM phenotype ( CCR7-CD45RA- ) with only 3 . 3% of cells expressing a CCR7+CD45RA- profile typical of CM cells ( Fig 3A ) . In addition , re-expression of CD45RA was found on only a minor subset ( 1 . 8% ) of effector memory cells , in marked contrast to the profile observed commonly on CMV-specific CD8+ T cells [6] . The surface expression of the additional differentiation markers CD27 , CD28 , CD57 and CD45RO was then assessed in order to undertake a more detailed phenotypic analysis . A Boolean gating strategy was used to investigate expression patterns of these markers and determine the differentiation hierarchy of the CMV-specific CD4+ T cells ( Fig 3B ) . As anticipated , the majority of the global CD4+ T cell population in each donor expressed a naïve phenotype but CM and EM populations were also observed in comparable proportions ( S1 Fig ) . The differentiation pattern of the CMV-specific T cells was very different ( Fig 3B ) and this , when combined with current understanding of T cell biology [32 , 33] , allowed us to model the profile of CMV-specific CD4+ T cell differentiation ( Fig 3D ) . This revealed that CMV-specific CD4+ T cells can be detected in several stages of differentiation and reveals progression through a dual CD45RA+CD45RO+ stage prior to loss of CD45RA expression and attainment of CCR7+CD45RO+ central memory phenotype . Further differentiation led to downregulation of CCR7 followed by a sequential loss of CD27 and CD28 expression . Indeed , 64% of cells exhibited a predominant CD27-CD28- profile whereas 22% ( 12/55 ) displayed a largely CD27-CD28+ profile . CD57 expression is a predominant feature of CMV-specific T cells and was observed almost exclusively on CD27-CD28- cells , with minor expression on the CD27-CD28+ population . A final stage of differentiation , in a minority of cells , was the re-expression of CD45RA which coincided with complete loss of CD45RO expression . Further examination of virus-specific populations within individual donors revealed a moderate degree of heterogeneity in relation to differentiation status . Indeed it was noteworthy that 9% ( 5/55 ) of responses were characterised by a dominant central memory phenotype whilst only 3 responses exhibited late stage CD45RA+ effector memory ( EMRA ) differentiation ( Fig 3B ) . As such we next went on to examine potential factors that might be related to the differentiation profile of virus-specific CD4+ T cells . Interestingly , this was not correlated with the magnitude of the immune response , a pattern that is different to the profile for virus-specific CD8+ T cells where clonal expansion is associated with a greater degree of differentiation [6] . In contrast , antigenic specificity may be an important factor as CD4+ T cells specific for DYS ( glycoprotein B ) displayed a more differentiated phenotype compared to pp65-specific T cells ( Fig 3C ) . As such , loss of CD28 or gain of CD57 expression was seen on 62% and 41% of gB-specific T cells respectively , compared to only 30% and 18% of CD4+ T cells specific for the epitopes from pp65 . Moreover , an EMRA phenotype was observed only on the CD4+ T cells which were specific for DYS . The availability of HLA-peptide tetramers allows the direct analysis of antigen-specific T cells without prior stimulation and this was felt to be particularly valuable in the assessment of CD4+ T regulatory function as FoxP3 expression can be induced following activation through the TCR [34] . In order to investigate whether CMV-specific CD4+ T cells contain natural T regulatory cells we stained cells directly ex vivo with HLA class II tetramer , anti-CD4 , anti-CD25 , anti-CD127 and intracellular anti-FoxP3 . However , virtually no CMV-specific T cells were found to exhibit a CD4+CD25+CD127low/-FoxP3+ T regulatory phenotype ( S2 Fig ) . The availability of HLA class II-peptide tetramers allowed us to undertake direct purification and transcriptional analysis of CMV-specific CD4+ T cells , an approach that has been important in relation to determining novel features of the equivalent CD8+ T cell subset [11] . CMV-specific CD4+ T cells were isolated from the blood of five CMV-seropositive donors by staining with tetramer followed by high purity cell sorting . Two of these populations were specific for epitope DYS and three recognised the peptide LLQ . Effector memory T cells isolated from CMV seronegative individuals were used as a comparator group . The pattern of normalised gene expression was compared initially between the combined transcriptome of the CMV-specific T cell samples and the effector memory population from CMV-negative donors . Global expression patterns were broadly similar between the two groups , reflecting the shared effector memory phenotype . However 55 mRNA transcripts differed by at least two-fold expression between the two groups , of which 35 were upregulated in CMV-specific T cells and 20 genes were lower within this group ( Fig 4A and S1 Table ) . We also compared the individual transcriptional profiles of DYS- and LLQ-specific T cell populations and here 12 of the 55 genes that exhibited differential expression between the combined profile of CMV-specific and control EM cells were also differentially expressed in both the DYS- and LLQ-specific T cells . 36 genes were altered only within the DYS-specific populations and 7 genes exhibited differential regulation within LLQ-specific T cells alone ( Fig 4B , S2 and S3 Tables ) , probably reflecting the more marked differentiation profile observed for the DYS-specific population . Relative expression levels ( aquantile normalised expression ) for selected transcripts are depicted in Fig 4C comparing DYS and LLQ-specific CD4+ T cells , as well as CD4+ EM cells from CMV seronegative donors . An increase in relative transcription levels was often observed for LLQ-specific T cells which was then further enhanced in DYS-specific T cells explaining why more significant differences in gene expression were observed in comparisons between DYS-specific T cells only and EM T cells . The function of many proteins encoded from the genes upregulated in CMV-specific T cells is related to cytotoxic function , such as granzymes B , H and A , granulysin and perforin . Expression of the chemokines CCL3 ( MIP-1α ) and CCL4 ( MIP-1β ) was strongly increased and indicates an important role for CMV-specific CD4+ T cells in attracting cells of the innate immune system to the site of viral recognition . The increased pattern of transcription of CX3CR1 in DYS-specific T cells is of particular note as this chemokine receptor has been shown to be a discriminative marker for CMV-specific CD8+ T cells and is thought to attract cells to areas of stressed endothelium which express the membrane-bound ligand fractalkine [11] . In addition we observed marked overexpression of ADRB2 , the gene encoding the β2-adrenergic receptor , on these cells which forms an important link between the sympathetic nervous system and the immune system . Additional upregulated genes of interest in CMV-specific T cells included the G protein coupled-receptor GPR56 and fibroblast growth factor-binding protein 2 ( FGFBP2 ) , both of which have been previously associated with cytotoxic activity , and the secreted extracellular matrix protein SPON2 . As changes in level of transcription do not always translate into the same changes at protein level , further analysis would be needed to confirm some of these observations . Several genes were downregulated in CMV-specific T cells of which the most striking pattern was seen for ADAMTS6 , a member of the ADAMTS family ( a disintegrin and metalloproteinases with thrombospondin ) . These secreted proteins have roles in mediating cell adhesion and proteolytic shedding and it is of interest that ADAMTS6 expression is increased by TNF-α [35] . The physiological importance of this will require further investigation as the substrate for ADAMTS6 is currently unknown . CD27 expression was also reduced , reflecting its marked reduction at the cell surface , and levels of TNFRSF4 ( OX40 ) transcript , which is induced following cell activation , was also low suggesting that CMV-specific CD4+ T cells are largely resting in the steady state . To further validate results from the microarray analysis , and investigate differences observed between LLQ- and DYS-specific CD4+ T cells , we performed qPCR analysis for genes that were found to exhibit differential expression on microarray . These include the chemokines CCL3 , CCL4 and CCL5 , GZMB and perforin , as well as CX3CR1 and the ADRB2 gene . For all seven genes we confirmed increased levels of transcription in CMV-specific T cells compared to the CD4+EM population ( Fig 4D ) . The pattern of expression was broadly reflective of that seen in the microarray analysis . Particularly high transcript levels in CMV-specific T cells were observed for GZMB ( 30 to 50-fold increase ) , CCL4 ( 6 to 19-fold increase ) and CX3CR1 . Of note , in DYS-specific T cells transcription of CX3CR1 was found to be 1000-fold higher than in EM cells and levels were also 12-fold higher in LLQ-specific T cells . Gene expression of ADRB2 was also increased in both LLQ- and DYS-specific T cells . We next went on to investigate the protein expression of four genes whose transcription had been revealed to be strongly upregulated by microarray analysis . As such , tetramer staining was combined with antibodies to granzyme B , perforin , FasL and CX3CR1 . CMV-specific CD4+ T cells were found to possess a very strong cytotoxic phenotype with up to 96% of cells staining positive for granzyme B on direct ex vivo analysis ( Fig 5A ) . This pattern was particularly strong for glycoprotein B-derived DYS-specific T cells which exhibited a median expression level of 78% , compared to 61% and 45% of the pp65-derived LLQ- and AGI-specific T cells respectively . Perforin expression was also found on 57% , 39% and 27% of these three populations respectively and all T cells that expressed perforin showed co-expression of granzyme B . A very strong correlation was observed in relation to the expression of CX3CR1 with cells of a cytotoxic phenotype ( Fig 5A ) . In CMV seronegative individuals the proportion of CD4+ EM T cells expressing markers of cytotoxicity was only around 1% . Given the clinical importance of CMV infection in older people we further analysed the expression of granzyme B , perforin and CX3CR1 in relation to the age of the donor ( Fig 5B ) . Interestingly , the substantial cytotoxic potential of CMV-specific CD4+ T cells was found to increase even further with aging and this was particularly the case for pp65-specific T cells , within which perforin expression increased from 18% within donors aged 20–35 years compared to 43% in those aged over 60 years . The cytotoxic profile of DYS-specific T cells also tends to increase with age but this effect was less marked as the cytotoxic phenotype was already strongly established in these cells at an early age . The expression of CX3CR1 was consistently high on CMV-specific T cells from donors at all age groups with 80–90% of both gB and pp65-specific cells carrying this receptor . Expression of FasL was detected on only a very small proportion of CD4+ T cells , with a median of 0 . 7% of CMV-specific cells and 0 . 48% of the global effector memory CD4+ pool expressing this marker ( S3 Fig ) . To investigate whether CMV-specific CD4+ T cells are indeed capable of killing target cells directly ex vivo , we isolated virus-specific cells from peripheral blood using tetramers and determined lysis of autologous or HLA-matched lymphoblastoid cell lines ( LCLs ) loaded with cognate peptide . CD4+ effector memory T cells ( CCR7-CD45RA- ) from two CMV-seronegative individuals were selected as controls . CMV-specific CD4+ T cells displayed remarkable cytotoxic capacity and this was seen for cells specific for all three peptide targets ( Fig 5C ) . At an effector:target ( E:T ) ratio of 1:1 LLQ- and AGI-specific T cells lysed around 77% of target cells within 20 hours . DYS-specific T cells were able to kill 56% of target cells even at an E:T ratio of 0 . 5:1 . No killing was observed by CD4+EM cells isolated from CMV seronegative individuals . Previous studies have shown that CMV-specific CD4+T cells can express the co-stimulatory molecule NKG2D [36] which is almost always present on cytotoxic NK cells and CD8+ T cells . However these analyses relied on functional activation with viral antigen and we therefore examined NKG2D expression on unmanipulated cells through the use of HLA class II tetramers . Interestingly , NKG2D expression was negligible on the CD4+ EM T cell population in CMV seronegative individuals , with a median expression of only 0 . 66% , but was observed on 23% of CMV-specific CD4+ T cells ( Fig 5D ) . Importantly , this proportion did not show any increase in relation to aging . Relatively little is known about the pattern of expression of inhibitory markers such as PD-1 or Tim-3 on antigen-specific CD4+ T cells . Expression of PD-1 on virus-specific CD8+ T cells has been associated with functional impairment of immune responses against HIV or HCV , both chronic infections in which antigen load may remain high for prolonged periods of time [37 , 38] . However in healthy individuals PD-1hi CD8+ T cells usually demonstrate an effector memory phenotype and do not necessarily exhibit functional ‘exhaustion’ [39] . To get a better understanding of PD-1 expression on CD4+ T cell populations we initially analysed the memory phenotype of PD-1+ cells within the global CD4+ T cell subset and the CMV-specific T cell population . This showed that within the CD4+ T cell population 62% of PD-1+ cells carried an EM phenotype and 22% were CM cells ( Fig 6A ) . Of the PD-1+ CMV-specific CD4+ T cells 94% displayed a CCR7-CD45RA- EM phenotype and virtually no cells had a CCR7+CD45RA-CM phenotype ( Fig 6B ) . We therefore examined the pattern of expression of PD-1 on CMV-specific CD4+ T cells and contrasted this with the pattern of staining on the CD4+EM T cell population . PD-1 expression was observed on a median 47% of CMV-specific CD4+ T cells and exhibited a remarkable range of expression , with cells from some donors showing hardly any evidence of PD-1 expression whereas virtually 100% of cells were positive in other individuals ( Fig 6C ) . When PD-1 expression was evaluated in relation to peptide specificity it was clear that PD-1 expression was markedly reduced on DYS-specific CD4+ T cells , where it was observed on 29% of virus-specific cells , compared to 51% and 47% of AGI- or LLQ-specific T cells , respectively . Interestingly , PD-1 expression was very low on the CD4+EM T cell population and was observed on less than 10% of cells ( Fig 6C ) . When analysing the median fluorescence intensity ( MFI ) of PD-1 on these cell populations we found the same trends within virus-specific T cells , and much lower MFI values were detected on CD4+EM T cells ( Fig 6D ) . Of interest , Tim-3 expression was not detected on a significant proportion of any CD4+ T cells ( S4 Fig ) . It might be anticipated that the proportion of T cells expressing an inhibitory marker would increase with age . Therefore we examined both the percentage of PD-1+ cells and the MFI of PD-1 expression on CMV-specific T cells , and the CD4+EM T cell subset , in relation to age . The proportion of PD-1+ cells did not change in relation to the age of the donor ( Fig 6E ) however a non-significant trend was observed towards lower levels of PD-1 protein expression ( MFI ) on the surface of PD-1+ CMV-specific CD4+ T cells in older adults ( Fig 6F ) . In this study we have used HLA class II-peptide tetramers to identify and characterise CMV-specific CD4+ T cells , without the need for functional identification , for the first time . This allowed a comprehensive analysis of the resting phenotype and transcriptional profile of antigen-specific CD4+ T cells recognising glycoprotein B and pp65 , the two viral proteins recognised most frequently by CMV-specific CD4+ T cells [21] . The CD4+ T cell response against cytomegalovirus is of interest for several reasons , related primarily to the unusual magnitude and phenotype of virus-specific cells , and also their potential role in the vascular damage that is reported in association with CMV infection in older people [40 , 41] . Here we identified CMV epitope-specific responses ranging from 0 . 01% up to a remarkable 24% of the total CD4+ repertoire . The median peptide-specific response was 0 . 45% of the CD4+ T cell repertoire which is considerably higher than CD4+ T cell responses against viruses such as influenza , hepatitis C Virus and Epstein-Barr Virus [30 , 42 , 43] . Interestingly , the frequency of cells identified by tetrameric staining was greater than that detected by expression of IFN-γ following peptide stimulation . As such , although this confirms the strong Th1 profile of CMV-specific CD4+ T cells , it also reveals that a considerable proportion of peptide-specific CD4+ T cells demonstrate an alternative cytokine profile . Indeed , TNF-α was the cytokine most frequently produced by CMV-specific CD4+ T cells . In addition , a proportion of virus-specific T cells was found to secrete IL-4 in some donors although the great majority of CMV-specific CD4+ T cells display a polarized Th1 phenotype [44] . IL-4 production by CMV-specific CD4+ T cells has previously been reported through ELISPOT analysis [45] . Of interest , although IL-4 production within pp65-specific T cells was usually observed with co-expression with IFN-γ , IL-4 production by gB-specific cells was seen in the absence of other cytokines . As such these observations indicate novel functional roles for IL-4 in the setting of chronic infection which warrant further investigation [46] . Functional assays using stimulation with whole viral lysate have demonstrated an increase in virus-specific CD4+ T cells with age [22] and although a similar trend was observed in our analysis it is likely that demonstration of such T cell ‘memory inflation’ will require analysis of responses directed against a wide range of different epitopes . Many younger donors already had high frequencies of epitope specific T cells in the periphery . This may be a reflection of the duration of viral carriage as some individuals acquire the virus very early in life whereas others only do so at a later time . Almost all CMV-specific CD4+ T cells expressed an effector memory phenotype , although in contrast to CMV-specific CD8+ T cells , very few had reverted to re-expression of the CD45RA isoform rather than CD45RO . In vitro studies suggest that re-expression of CD45RA occurs during prolonged absence of antigenic stimulation [47] which may indicate that CMV-specific CD4+ T cells undergo antigen recognition in vivo more frequently than CD8+ subsets . Latent CMV is maintained within cells of the monocytic lineage and viral reactivation occurs during maturation into dendritic cells [48] . As such , CMV-specific CD4+ T cells are likely to undergo regular antigenic stimulation and this may serve to retain expression of the CD45RO isoform . Tetramer staining also allowed a detailed analysis of the membrane phenotype of CMV-specific CD4+ T cells and indicated that CD27 expression is lost early during differentiation and is then followed by loss of CD28 in the majority of the population . Loss of CD28 expression on CD4+ T cells is very unusual and , indeed , the CD4+CD28- phenotype is virtually unique to CMV-specific T cells [49] and indicates that alternative mechanisms of T cell co-stimulation become important in the CMV-specific immune response , potentially through molecules such as 4-1BB [50] . CD57 is a poorly characterised molecule , but is again highly specific for CMV-specific T cells and was found to be expressed reciprocally with CD28 , suggesting that it may itself have a potential role in co-stimulation . NKG2D is one such potential alternative costimulatory molecule and can synergise with TCR-dependent activation of CMV-specific CD4+ T cells to enhance a range of effector functions [36 , 51 , 52] . Indeed , we found that NKG2D was expressed on 23% of CMV-specific CD4+ T cells , a remarkably high level for a molecule typically associated with expression on NK and CD8+ T cells . The high degree of differentiation seen for CMV-specific CD4+ T cells was largely independent of the frequency of epitope-specific T cells , indicating that alternative factors , such as the environment in which T cell priming occurs or the availability of antigen , may influence their phenotypic profile ( reviewed in [53] ) . However the antigenic specificity of virus-specific CD4+ T cells did have a marked influence on the differentiation status such that CD4+ T cells specific for glycoprotein B-derived epitope DYS consistently displayed a more differentiated phenotype than pp65-specific cells . Glycoprotein B is highly unusual in that it has direct access to the HLA class II antigen-processing pathway of infected cells . This mechanism is likely to mediate frequent reactivation of gB-specific T cells and could explain the more ‘driven’ differentiation phenotype of CD4+ T cells specific for peptides from this protein [54 , 55] . A unique aspect of our study was the direct isolation of virus-specific CD4+ T cells without the need for antigenic stimulation , such that we were able to analyse their resting transcriptional profile . Even though differences to CD4+EM T cells of CMV seronegative individuals were generally modest , the most striking observations were seen in the expression of genes related to cytotoxic function and chemotaxis . Remarkably , the profile of cytolytic genes upregulated in CMV-specific CD4+ T cells closely corresponds to the pattern seen in CD8+ cytotoxic T cells ( CTL ) . These included granzymes A , B and H , as well as perforin , granulysin and NKG7 . Importantly , we were also able to determine that Fas ligand is not expressed by virus-specific CD4+ T cells , indicating that the delivery of mediators from cytotoxic granules is the dominant mechanism of target cell lysis . Remarkably , the use of tetrameric staining reveals that this profound cytotoxic potential of CMV-specific T cells is observed within resting cells in the bloodstream . Furthermore isolated virus-specific CD4+ T cells very efficiently kill antigen-loaded target cells directly ex vivo suggesting that they are primed for rapid target cell lysis in the event of an episode of viral reactivation . CMV-specific CD4+ T cells also expressed a distinctive profile of chemokines and additional proteins that indicate an important role in chemotaxis and tissue migration . Indeed , one of the most fascinating features of CMV-specific CD4+ T cells is their high level of expression of CX3CR1 , a chemokine receptor that binds CX3CL1 ( fractalkine ) and has already been identified as a specific marker for CMV-specific CD8+ T cells [11] . The CX3CR1/CX3CL1 axis plays an important role in both the adhesion and transmigration of lymphocytes to endothelial cells during inflammation [56 , 57] . Interestingly , endothelial cells are a principal target tissue for CMV infection and the expression of CX3CR1 on CMV-specific T cells may therefore act to localise adaptive immunity to sites of viral reactivation . This mechanism of endothelial-targeting may be highly relevant to the potential development of endothelial immunopathology mediated by CMV-specific cytotoxic T cells [58 , 59] . Indeed a close link has been observed between augmented CMV-specific immune responses and a range of inflammatory conditions ( reviewed in [60] ) and the proportion of CD4+CD28- T cells has been shown to correlate directly with cardiovascular mortality in some studies [61] . CMV-specific CD4+ T cells also exhibit a very similar chemokine secretion profile to that of virus-specific CD8+ cells , with production of the inflammatory mediators CCL3 ( MIP-1α ) , CCL4 ( MIP-1β ) and CCL5 ( RANTES ) , all of which would support recruitment of innate immune cells such as macrophages and NK cells . The pattern is associated with high level expression of IFN-γ and TNF-α but very little production of IL-2 [23 , 62] and is again typical of a differentiated Th1 profile . An additional interesting observation from the microarray data was the detection of high levels of ADRB2 mRNA in glycoprotein B-specific T cells . ADRB2 encodes the β2-adrenergic receptor which allows cells to respond to systemic production of epinephrine and forms an important link between the sympathetic nervous system and the adaptive immune response [63 , 64] . The CMV genome contains promoter elements that can bind epinephrine leading to increased viral gene transcription [65] and physiological stress is established as a risk factor for reactivation of many herpesviruses . Interestingly , CX3CR1+ T cells have been shown to be the major T cell subset released into the circulation following administration of epinephrine , in a mechanism partially mediated by reversal of their resting adherence to endothelium [66] . This may provide a mechanism of rapid mobilisation for CMV-specific T cells in response to stress induced viral reactivation , but this will need further investigation . T cell activation and effector function are finely tuned events and CD4+ T cells show an extremely high sensitivity for their cognate antigen [67] . Moreover lytic synapse formation has a very low threshold for activation [68 , 69] . A proportion of CD27-CD28- CMV-specific CD4+ T cells has previously been described to exhibit regulatory function in vitro [70] but FoxP3+ cells were not detectable in this study and antigen-specific T cells were not shown to produce IL-10 . This suggests that regulatory mechanisms are likely to be focussed at the level of the effector T cell and here PD-1/PD-L1 interactions play a key role as a negative feedback mechanism controlling TCR-dependent effector function of T cells [71] . High levels of PD-1 have been detected on CD8+ effector memory T cells [39] and similarly we find that PD-1 expression on CD4+ T cells is also largely confined to this subset . PD-1 expression was seen on nearly half of all CMV-specific CD4+ T cells and could serve to limit T cell activation . As such it is noteworthy that the level of membrane PD-1 expression ( MFI ) on virus-specific CD4+ cells tends to decrease during aging and may therefore lower the activation threshold of T cells at a time when their cytotoxic potential is actually increasing . Together these factors may influence virus-host balance across the life course and may contribute towards immunopathology . In summary our data reveal that people who are chronically infected with cytomegalovirus , which represents the great majority of the global population , harbour substantial populations of virus-specific CD4+ T cells within their bloodstream . These highly differentiated cells display a strongly cytotoxic phenotype , may be targeted to activated endothelium and have the potential to respond to physiological ‘stress’ through detection of epinephrine ( Fig 7 ) . In addition , this cytotoxic profile increases further with age whilst the level of inhibitory PD-1 on the surface declines . These findings reveal the exceptional evolutionary adaptation of the CD4+ T cell response towards the control of CMV . In addition they shed light on the potential mechanisms whereby CMV infection may serve to mediate tissue damage , most particularly vascular disease , and indicate a range of potential novel immunotherapeutic targets . The study was approved by the West Midlands ( Black Country ) Research Ethics Committee ( REC 07/Q2702/24 ) and all donors gave written informed consent before participation . The donor cohort included samples from laboratory personnel , the blood transfusion service and healthy older adults recruited from the ‘Birmingham 1000 Elders group’ ( REC 2002/073 ) . A total number of 73 CMV-seropositive donors , aged between 24–88 years , with appropriate HLA-genotype were included in the study . PBMCs were isolated from 50ml heparinized blood by density gradient centrifugation using Lympholyte-H cell separation media ( Cedarlane ) and aliquots of 10x106 cells were cryopreserved in RPMI1640 ( Sigma-Aldrich ) containing 20% fetal calf serum ( FCS ) and 10% DMSO and stored in liquid nitrogen until use . To identify donors with the appropriate HLA-genotype , genomic DNA was isolated from PBMCs using the GenElute Blood Genomic DNA Kit ( Sigma-Aldrich ) according to manufacturer’s instructions . Typing for HLA class II alleles was performed by PCR technique as described previously [72] . Phycoerythrin ( PE ) -conjugated custom-made HLA class II tetrameric complexes were purchased from the Benaroya Research Institute at Virginia Mason ( Seattle , Washington ) . Three tetramer complexes were used in this study: They were comprised of the CMV gB-derived epitope DYSNTHSTRYV in the context of HLA-DRB1*07:01 ( DR7 ) [73] , or pp65-derived epitopes AGILARNNLVPMVATV within HLA-DRB3*02:02 ( DR52b ) [74] and LLQTGIHVRVSQPSL within HLA-DQB1*06:02 ( DQ6 ) [75] . Initially , the specificity of each tetramer was confirmed by screening against CD4+ T cell clones recognizing the tetramer’s cognate HLA class II-peptide complex or against PBMCs from a CMV-seronegative donor expressing the appropriate HLA-allele . Optimal tetramer concentration and staining times were distinguished at the outset and constant conditions used throughout the study . To identify cytokine-producing T cells following activation , 1x106 PBMC were resuspended in RPMI1640 ( Sigma-Aldrich ) supplemented with 10% FCS and 1% Penicillin/Streptomycin and stimulated with peptide ( 5μg/ml final concentration ) overnight at 37°C with 5% CO2 in the presence of BrefeldinA ( 10μg/ml final concentration; Sigma-Aldrich ) . Unstimulated cells and cells stimulated with Staphylococcus Enterotoxin B ( final concentration 0 . 2μg/ml; Sigma-Aldrich ) served as controls . Following overnight incubation the cells were stained with LIVE/Dead fixable violet or aqua stain ( Invitrogen ) as described above , T-cells were then identified by staining with anti-CD4-PE ( RPA-T4 , BD Bioscience ) or anti-CD3 AF700 ( SK7 , Biolegend ) and anti-CD4 APC-Cy7 and B cells excluded by staining with anti-CD19 pacific blue ( H1B19 , eBioscience; dump channel ) . Fixing was carried out with 4% paraformaldehyde ( in PBS; Sigma-Aldrich ) for 15 min at RT before permeabilising with 0 . 5% Saponin ( in PBS; Sigma-Aldrich ) for 5 min . Intracellular cytokines were then stained with the following antibodies: anti-IFN-γ FITC antibody ( 25723 . 11 , BD Bioscience ) or anti-IFN-γ PE-Dazzle ( 4S . B3 ) , anti-TNF-α PE-Cy7 ( Mab11 ) , anti-IL-10 PE ( JES3-9D7 ) , anti-IL-4 BV421 ( MP4-25D2; all Biolegend ) , anti-IL17A ( eBio64DEC17 , eBioscience ) and anti-MIP-1β ( 24006 , R&D ) and followed by a final wash in staining buffer . Acquisition was carried out on an LSR II flow cytometer and DIVA software ( BD Bioscience ) collecting 300 , 000 live lymphocytes and data analysed using FlowJo software version 7 . 6 . 5 ( Tree Star ) . For the analysis single , viable , CD19- lymphocytes were gated before identification of cytokine-producing CD4+ T cells . For the multi-cytokine panel Boolean gating was used to determine all possible combinations and further analysis was carried out using SPICE version 5 . 3 . To analyse the transcriptional profile of CMV-specific T cells we sorted DYS- and LLQ-specific CD4+ T cells from CMV-seropositive healthy donors and for comparison CD4+ T cell subsets from CMV-seronegative healthy individuals . For this PBMCs were isolated from 120 mL of heparinised blood by density gradient centrifugation and the CD4+ T cell population was enriched by negative selection ( StemCell Technologies ) according to manufacturer’s instructions . CD4+ T cells from CMV-seropositive donors were then stained with LIVE/Dead fixable far red stain ( Invitrogen ) , washed and re-suspended in 600 μL of human serum prior to incubation with PE-conjugated HLA class II tetramer ( HLA-DR7:DYS or HLA-DQ6:LLQ ) for 1 h at 37°C and 5% CO2 . After washing , cells were stained with anti-CD4 PE-CF594 ( RPA-T4 , BD ) and re-suspended in RPMI + 10% FCS . From the single , viable lymphocyte population CD4+tetramer+ cells were then sorted on a MoFlow Cell Sorter ( Beckman Coulter ) consistently reaching a purity of 98–99% . Following CD4-enrichment cells of CMV-seronegative donors were stained with LIVE/Dead fixable far red stain ( Invitrogen ) , anti-CD4 PE-CF594 ( RPA-T4 , BD Biosciences ) , anti-CCR7 FITC ( 150503 , R&D ) and anti-CD45RA PE ( HI100 , eBioscience ) . Based on their expression profile effector memory cells ( CCR7-CD45RA- ) were then sorted to high purity . Total RNA of the sorted cells was extracted using an RNeasy Plus Mini kit ( Qiagen ) according to manufacturer’s instruction . The RNA integrity was checked and it was subsequently labelled before hybridisation to Agilent human gene expression 8x60k microarrays ( G4858A ) according to manufacturer’s instructions following the standard Agilent Low Input Quick Amp labelling protocol . Due to low mRNA yield , CD4+CD45RO+ T cells sorted from a CMV-seronegative donor served as the reference sample on the two-colour microarray slide . These were not taken into account for the analysis . The Microarrays were carried out by the Functional Genomics , Proteomics and Metabolomics Facility at the School of Biosciences , University of Birmingham . Microarray data was analysed with the R Limma Package ( Bioconductor ) [78–80] . Normalisation was performed with the Loess ( intra-array ) and Aquantile ( inter-array ) methods . An adjusted p-value ( Benjamini and Hochberg's method ) of 0 . 05 and below was taken as significant for differences in gene expression or otherwise a 2-fold change in expression levels . Further analysis of output data was completed in Excel ( Microsoft Corp ) . Hierarchical clustering was performed on the MultiExperiment viewer version 4 . 9 ( MeV , [77] . Functional enrichment analysis was completed using DAVID version 6 . 7 [81 , 82] . Microarray data are available in the ArrayExpress database ( www . ebi . ac . uk/arrayexpress ) under accession number E-MTAB-4510 . cDNA generated from the same samples used for the microarray analysis was used to analyse transcription of a selected number of genes that were differentially expressed between CD4+EM cells and CMV-specific CD4+ T cells . TaqMan Gene Expression Assays for CCL3 ( Hs04194942_s1 ) , CCL4 ( Hs04421399_gH ) , CCL5 ( Hs00982282_m1 ) , GZMB ( Hs00188051_m1 ) , PRF1 ( Hs00169473_m1 ) , ADRB2 ( Hs00240532_s1 ) , CX3CR1 ( Hs04187059_m1 ) and GAPDH ( 4310884E ) were bought from Thermo Fisher Scientific . Specific target amplification was carried out on the cDNA using 2× TaqMan PreAmp Master Mix ( Life Technologies ) and 0 . 2× primer mix ( 20× TaqMan assays diluted in water ) . Reactions were subjected to 95°C for 10 min , followed by 12 cycles of 95°C for 15 s and 60°C for 4 min . These pre-amplified samples were then diluted 1:5 with water prior to Q-PCR analysis using the TaqMan Gene Expression Assays . The relative transcription was calculated comparing with average transcription level of the three control CD4+EM T cells and GAPDH assays served for normalization . Assays were performed in duplicate for two to three donors each ( EM , LLQ and DYS ) . Transcription levels in CD4+EM T cells were set to 1 . To analyse the cytotoxic capacity of CMV-specific CD4+ T cells directly ex vivo , CD4+TM+ cells were separated as described above and then co-cultured over night with CFSE labelled ( 0 . 5μM; Invitrogen ) autologous or HLA-matched LCLs which were loaded with the cognate peptide . CFSE-labelled LCL alone served as control . In addition CD4+ CCR7-CD45RA- ( EM ) cells of two CMV seronegative individuals were sorted and served as effector cells . Killing of target cells was assessed on a BD Accuri flow cytometer ( BD Biosciences ) by quantifying live CFSE-labelled LCL using counterstaining with Propidium Iodide ( PI ) to identify dead cells . All conditions were carried out in triplicate . For the analysis initially the T cell population and the LCL population were gated on , using FSC and SSC scatter plots , to determine the true ratio of effector to target cells . Then the number of live LCLs was determined using CFSE and PI and percent killing of target cells was calculated . Statistical analysis of flow cytometry data was performed using GraphPad Prism5 . The non-parametric Mann-Whitney U-test was applied for comparison of two groups , and the Kruskal-Wallis test ( with Dunn’s multiple comparison ) for comparison of more than two groups of continuous measurements . To analyse the strength of associations between variables Spearman’s rank test was used . A p-value of less than 0 . 05 was considered statistically significant .
Cytomegalovirus ( CMV ) is a member of the herpesvirus family and most humans carry chronic CMV infection . This drives the development of large expansions of CD8+ CMV-specific T cells , which increase further during ageing . CMV infection is associated with vascular disease and increased risk of mortality in older people , which may be related to damage from this CMV-specific immune response . Here we used a set of novel reagents called HLA class II tetramers to make a detailed study of CMV-specific CD4+ T cells . We show that CMV-specific CD4+ T cells are found at remarkably high frequencies within blood , representing up to a quarter of all such white cells . In addition they demonstrate a range of unique features . Firstly they carry a chemokine receptor that directs the cells to activated endothelial cells within blood vessels . Secondly , they express epinephrine receptors which would allow them to respond rapidly to stress . Finally , these CD4+ T cells are unique as they are strongly cytotoxic and equipped with the ability to directly kill virally-infected cells . HLA class II tetramers therefore reveal a profile of unique features which provide insight into how CMV-specific CD4+ T cells may be involved in vascular immunopathology .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "blood", "cells", "innate", "immune", "system", "medicine", "and", "health", "sciences", "immune", "cells", "immune", "physiology", "cytokines", "immunology", "cell", "differentiation", "developmental", "biology", "molecular", "development", "bioassays", "and", "physiological", "analysis", "cytotoxic", "t", "cells", "research", "and", "analysis", "methods", "specimen", "preparation", "and", "treatment", "staining", "white", "blood", "cells", "memory", "t", "cells", "animal", "cells", "t", "cells", "immune", "system", "microarrays", "cell", "staining", "cell", "biology", "phenotypes", "physiology", "genetics", "biology", "and", "life", "sciences", "cellular", "types" ]
2016
Cytomegalovirus Infection Leads to Development of High Frequencies of Cytotoxic Virus-Specific CD4+ T Cells Targeted to Vascular Endothelium
Revealing QTLs with a minor effect in complex traits remains difficult . Initial strategies had limited success because of interference by major QTLs and epistasis . New strategies focused on eliminating major QTLs in subsequent mapping experiments . Since genetic analysis of superior segregants from natural diploid strains usually also reveals QTLs linked to the inferior parent , we have extended this strategy for minor QTL identification by eliminating QTLs in both parent strains and repeating the QTL mapping with pooled-segregant whole-genome sequence analysis . We first mapped multiple QTLs responsible for high thermotolerance in a natural yeast strain , MUCL28177 , compared to the laboratory strain , BY4742 . Using single and bulk reciprocal hemizygosity analysis we identified MKT1 and PRP42 as causative genes in QTLs linked to the superior and inferior parent , respectively . We subsequently downgraded both parents by replacing their superior allele with the inferior allele of the other parent . QTL mapping using pooled-segregant whole-genome sequence analysis with the segregants from the cross of the downgraded parents , revealed several new QTLs . We validated the two most-strongly linked new QTLs by identifying NCS2 and SMD2 as causative genes linked to the superior downgraded parent and we found an allele-specific epistatic interaction between PRP42 and SMD2 . Interestingly , the related function of PRP42 and SMD2 suggests an important role for RNA processing in high thermotolerance and underscores the relevance of analyzing minor QTLs . Our results show that identification of minor QTLs involved in complex traits can be successfully accomplished by crossing parent strains that have both been downgraded for a single QTL . This novel approach has the advantage of maintaining all relevant genetic diversity as well as enough phenotypic difference between the parent strains for the trait-of-interest and thus maximizes the chances of successfully identifying additional minor QTLs that are relevant for the phenotypic difference between the original parents . Many genetic traits are quantitative and show complex inheritance . Because these traits are so prevalent in nature , understanding the underlying factors is important for various biological fields and for applications like industrial biotechnology and agricultural practice [1] . Recently , baker's yeast Saccharomyces cerevisiae has become an important subject for studies in quantitative genetics [2] , [3] . In particular the availability of high-density genetic markers , the ease of performing experimental crosses and the powerful technologies for precise genetic modification [4] , [5] , do not only allow efficient QTL mapping but also rapid identification of causative genes and their experimental validation and interaction analysis . S . cerevisiae displays many quantitative traits that are also important in other cell types , including industrial microorganisms and cells of higher , multicellular organisms . Such properties include thermotolerance [6] and oxidative stress tolerance [7] , the capacity to produce small molecules , such as acetic acid [8] and ethanol tolerance [9] , [10] . Other quantitative traits that have been studied in yeast include transcriptional regulation [11] , sporulation efficiency [12] , telomere length [13] , cell morphology traits [14] , mitochondrial genome instability [15] , global gene expression [16] , evolution of biochemical pathways [17] and resistance to chemicals [18] . A major remaining challenge in quantitative trait studies is the efficient mapping of minor quantitative trait loci ( QTLs ) and identification of their causative genes . Minor QTLs have a subtle influence on the phenotype , which is easily masked by epistasis [19] , gene-environment interactions [20] , low association to the phenotype because of limited sample size and complex interactions with other QTLs . Minor QTLs are important because together they can produce in an additive or synergistic manner equally dramatic effects on the phenotype as major QTLs . Actually , the work of Bloom et al . [21] , in which a large panel of individually genotyped and phenotyped yeast segregants was used , has shown that for 46 quantitative traits , the assembly of all detected loci could explain nearly the entire additive contribution to the heritable variation . The minor QTLs identified should be truly relevant for the trait of interest in the original parent strains and not generated in some unrelated way during the mapping procedure . Several methods have been reported to identify minor QTLs . Sinha et al . [22] used a targeted backcross strategy to first eliminate a major QTL . Subsequent mapping revealed a novel allele that had an epistatic interaction with the first major QTL . A disadvantage of backcrossing is the reduction of genetic diversity , which likely leads to loss of minor QTLs . In a different approach , Lorenz and Cohen [23] fixed major QTLs either in the superior parent or in the inferior parent and successfully identified minor QTLs by linkage analysis by repeating the QTL mapping with the new parent strains . A potential problem caused by elimination of major QTLs in one parent is that the phenotypic difference between the two parent strains is reduced . This may make it more difficult to evaluate the phenotype of the extreme segregants in comparison with the superior parent . Parts et al . [24] used many millions of segregants and multiple inbreeding steps to facilitate the detection of statistically significant minor QTLs . The use of such a high number of segregants , however , is only feasible for selectable phenotypes . Swinnen et al . [10] made use of more stringent phenotyping , i . e . tolerance to higher ethanol levels , which revealed several additional minor QTLs . The disadvantage of this approach is that higher stringency of phenotyping requires higher numbers of segregants to be phenotyped to obtain enough segregants with the superior phenotype . In the study of Bloom et al . [21] , aimed at identifying the source of missing heritability , linkage analysis was performed with a large panel of individually genotyped and phenotyped yeast segregants , which enabled detection of many QTLs with a small effect . In this work we have extended previous approaches to identify minor QTLs to QTL mapping by pooled-segregant whole-genome sequence analysis and we eliminated the effect of major QTLs in both parents . Our method is based on the observation that superior haploid segregants of natural or industrial diploid strains usually contain mutations that to some extent compromise rather than promote the trait of interest . As a result genetic mapping with such segregants usually reveals QTLs , which are linked to the inferior parent rather than to the superior parent . This allows the construction of two new parent strains , which are both downgraded for the trait of interest by replacement of a superior allele with an inferior allele from the other parent . This maintains a large phenotypic difference between the new parent strains . They also retain all genetic diversity , in particular all remaining minor QTLs . We show the effectiveness of this approach by first mapping QTLs involved in high thermotolerance of a selected yeast strain compared to a control strain , identifying causative genes linked to the superior and inferior parent , constructing two downgraded parent strains and repeating the genetic mapping with the new parents . This revealed several new minor QTLs , which we validated by identifying the causative gene in two QTLs . Interestingly , the two novel causative genes identified in this study are both involved in pre-mRNA splicing , which suggests an important role for RNA processing in conferring high thermotolerance . We have screened a total of 305 natural and industrial isolates of S . cerevisiae for their ability to grow at high temperature , i . e . 40–41°C , on solid YPD plates . Not a single yeast strain was able to grow with a reasonable rate at 42°C . The strain MUCL28177 showed very good growth at 41°C and was chosen for further analysis . After sporulation , we selected a haploid segregant MUCL28177-21A , further referred to as 21A , which also showed excellent growth at 41°C compared to the control strain BY4742 . Strain 21A was crossed with the laboratory strain BY4742 , that is unable to grow at 41°C . The hybrid 21A/BY4742 diploid strain grew at least as well as the 21A strain at 41°C , indicating that the high thermotolerance of 21A is a dominant characteristic . Phenotyping of 950 segregants of the 21A/BY4742 diploid strain revealed a range of thermotolerance . It resulted in 58 segregants with similar growth at high temperature as 21A . The growth of the original strain MUCL28177 , the parent strains 21A and BY4742 , the hybrid diploid strain 21A/BY4742 and ten representative segregants with varying thermotolerance , is shown in Figure 1 . The 58 thermotolerant segregants were pooled based on dry weight and genomic DNA isolated from the pool . Genomic DNA samples from the pooled segregants and from parent strain 21A were sequenced . The sequence reads obtained were aligned with the sequence of the reference S288c genome , which is essentially the same as that of the inferior parent strain BY4742 . A set of quality-filtered SNPs to be used as genetic markers , was acquired essentially as described before [10] . For each chromosome , the SNP variant frequency was modeled using an additive logistic regression model [10] , [25] . The results are shown in Figure 2 . In the top panel , the raw SNP frequencies are plotted against the chromosomal position along with the modeled frequency ( smoothed lines ) . The middle panel shows contrasts between selected pools and an unselected pool along with 95% simultaneous confidence bands . Upward and downward deviations from 0 indicate putative QTLs containing causative alleles from the superior and inferior parent , respectively . Normally , only linkage with the superior parent strain is expected . However , since the original MUCL28177 diploid strain is a natural isolate , it is likely heterozygous . Hence , the 21A segregant may contain recessive mutations that compromise to some extent thermotolerance in spite of the fact that its overall thermotolerance was only slightly lower than that of the MUCL28177 parent strain . We calculated 2-sided p-value profiles along the chromosome that were adjusted for multiple testing ( Text S1 online: Supplementary Methods ) and five regions show significant p-values ( 0 . 05 significance level , Figure 2 ) . We chose four regions with the smallest p-values for further analysis ( Table S1 online ) . For these loci , selected SNPs were scored in individual thermotolerant segregants ( up to 62 after additional segregant isolation and phenotyping ) and a binomial exact test with FDR adjusted p-values was used for assessing statistical significance [10] , [26] . Three QTLs ( QTL1 , QTL2 and QTL3 ) were confirmed to exhibit statistically significant linkage to the high thermotolerance phenotype ( 0 . 05 FDR level , Table 1 ) . QTL1 and QTL2 showed linkage with the genome of the superior 21A parent strain , while QTL3 showed linkage with the genome of the inferior BY4742 parent strain . We concentrated our work first on QTL1 and QTL3 , because they showed the strongest linkage to the superior and inferior parent , respectively . The subtelomeric regions often show deviations from the 50% value of the SNP variant frequency , but this is also observed in the unselected pool . It may be caused by complications with the mapping of repetitive sequences , which are known to be commonly present in subtelomeric regions . We have analysed for instance the right subtelomeric region of chromosome X , in the mapping with the original parents , using SNP detection in the individual segregants and found a p-value that failed to indicate significant linkage ( results not shown ) . We first fine-mapped QTL1 by scoring eight selected SNPs in individual thermotolerant segregants , which reduced the size of the locus to about 60 , 000 bp ( Figure 3A ) . Detailed analysis of the 21A sequence of this region showed that 22 out of the 33 genes and putative ORFs present contained at least one non-synonymous mutation in the ORF compared to the BY4742 sequence ( Figure 3A ) . Next we applied reciprocal hemizygosity analysis ( RHA ) [6] to identify causative gene ( s ) in QTL1 . RHA is used to test for a possible contribution to the phenotype of each allele of the candidate gene in a hybrid genetic background . For each of the 22 genes with non-synonymous mutations , we constructed two 21A/BY4742 hybrid strains in which either the 21A or the BY4742 allele was deleted , so that each strain only contained one specific allele of the candidate gene . Comparison of the growth at high temperature ( 41°C ) of the two hybrid strains did not show any difference for the 22 candidate genes , except for MKT1 ( Figure 3B , Figure S1 online and data not shown ) . The hybrid strain with the MKT121A allele showed better growth than the strain with the MKT1BY4742 allele . We further confirmed the relevance of MKT1 by demonstrating that MKT1 deletion reduced thermotolerance in the 21A strain background ( Figure S2 online ) . Since 21A with either mkt1Δ or MKT1BY4742 showed the same growth at 40 . 7°C and since BY4742 showed the same growth at 40 . 7°C as BY4742 mkt1Δ , the MKT1BY4742 allele behaves as a loss of function allele for thermotolerance when assayed under our conditions and in our haploid strain backgrounds ( Figure S2 online ) . In a previous QTL mapping study of thermotolerance with a clinical isolate of S . cerevisiae and the lab strain S288c , the MKT1 allele of the clinical isolate was also identified as a causative gene [6] and in a follow-up study , out of two polymorphisms in Mkt1 , D30G and the conservative substitution K453R , the D30G mutation was identified as the causative mutation [27] . Sanger sequencing of MKT121A confirmed that Mkt1-21A has the same mutations . END3 and RHO2 , which are located close to MKT1 in the same QTL , were also reported to have an allele-specific contribution to thermotolerance [6] . However , in the current experimental setup , the RHO2 alleles from our two genetic backgrounds did not produce a difference in thermotolerance , while for END3 there may be a slight difference ( Figure 3B ) . Sequence alignment using the Illumina sequencing data shows that END321A lacks the causative SNP ( C773T ) found in END3YJM145 [27] . In the case of RHO2 it is known that SNPs in the 3′UTR of RHO2YJM145 are responsible for the phenotypic effect on thermotolerance . RHO221A contains the same SNPs in its 3′UTR except for insertion of an A six base pairs downstream of the ORF . Hence , this insertion in RHO2YJM145 may cause the growth advantage at high temperature . QTL3 is linked to the genome of the inferior parent strain , indicating that BY4742 contains a superior genetic element for thermotolerance in this region . We fine-mapped QTL3 by scoring seven selected SNPs in 62 thermotolerant segregants individually . This reduced the locus to 40 , 000 bp ( Figure 4A ) . Detailed analysis of the 21A sequence in this region revealed 13 genes and putative ORFs with at least one non-synonymous mutation ( Figure 4A ) . To accelerate identification of the causative genes in this region , we first performed ‘bulk RHA’ . Instead of comparing alleles for each single gene , we first made a reciprocal deletion in the hybrid strain of a fragment with multiple genes . We divided the 40 , 000 bp region of QTL3 into two fragments , with the first containing 11 genes and the second 14 genes ( Figure 4A , 4B ) . For each fragment , we constructed two hybrid strains with one strain containing only the fragment from the 21A background and the other only the fragment from the BY4742 background ( Figure 4B ) . Comparison of growth at high temperature ( 41°C ) showed that FRAGMENT1BY4742 conferred better growth at high temperature than FRAGMENT121A , while there was a much smaller difference between the strains with FRAGMENT2BY4742 or FRAGMENT221A ( Figure 4C ) . We then applied RHA for the six individual genes of FRAGMENT1 that had at least one non-synonymous mutation ( Figure 4C ) . This identified PRP42BY4742 as a superior allele for thermotolerance compared to PRP4221A , whereas for the other genes there was no allele-specific difference in thermotolerance ( Figure S3 online ) . We also tested growth at high temperature of strains containing a heterozygous deletion of either the complete FRAGMENT1 or only the PRP42 gene together on the same plate . We found the growth at 41°C to be similar whether the complete FRAGMENT1 or only the PRP42 gene from either BY4742 or from 21A was deleted ( Figure S4 online ) . This suggests that PRP42 was likely the only causative gene in FRAGMENT1 and thus also seems to exclude the other genes without non-synonymous mutation in their ORF as possible causative gene . As an additional control , we also performed RHA with the seven genes of Fragment 2 with a non-synonymous mutation in their ORF and we did not find any difference between the alleles from the two parent strains in conferring thermotolerance ( data not shown ) . PRP4221A has eleven mutations compared to PRP42BY4742 , with three of them being non-synonymous and the other eight synonymous ( Table S2 online ) . The three polymorphisms in Prp42 , H296Y , F467S , and E526Q , are non-conservative substitutions , but it is difficult to predict a possible effect on the function or structure of the protein . They are located in domains without strong conservation ( data not shown ) . Since no mutation was present in the promoter and terminator region , the difference in thermotolerance conferred by the two PRP42 alleles is likely due to the change in protein sequence and thus in functionality . We have investigated the presence of these mutations in 22 other yeast strains , isolated from various sources , and of which the whole genome has been sequenced ( Table S2 online ) , and found that among the three non-synonymous mutations , C886T is unique to 21A , whereas the other two mutations ( C1400T and C1576G ) are present in all other strains except in the lab strains S288c , CEN . PK113-7D and W303 . If we assume that the inferior PRP42 allele is rare ( like the inferior MKT1 allele in S288c ) , then C886T is the best candidate for the causative mutation . On the other hand , we cannot exclude that C886T is only one of the causative mutations , that it requires interaction with one or more of the other mutations or that a combination of the other SNPs is causative for the phenotype . We next constructed two downgraded parent strains each with their own superior allele replaced by the inferior allele of the other parent: 21ADG: 21A mkt1Δ:: MKT1BY4742 and BY4742DG: BY4742 prp42Δ::PRP4221A . Growth at 41°C of 21ADG was reduced compared to 21A , confirming the importance of MKT121A for high thermotolerance in 21A ( Figure 5A ) . At 41°C , BY4742 and also BY4742DG are not able to grow ( Figure 5A ) . Hence , we reduced the temperature to 40 . 7°C , which allowed to demonstrate reduced growth of BY4742DG compared to BY4742 ( Figure 5B ) . Also at 41°C , we could demonstrate the beneficial effect of PRP42BY4742 compared to PRP4221A by comparing growth of the 21ADG/BY4742 and 21ADG/BY4742DG hybrid strains ( Figure 5A ) . The availability of the four hybrid diploid strains also allowed us to demonstrate that in this background the effect of the MKT1 and PRP42 genes on thermotolerance is independent . The hybrid diploids , 21ADG/BY4742 and 21A/BY4742DG , each with replacement of one superior allele , both showed reduced growth at 41°C compared to the original hybrid of the parent strains , 21A/BY4742 , while the hybrid of the two downgraded parent strains , 21ADG/BY4742DG , in which both superior alleles are replaced , showed further reduced growth ( Figure 5A ) . ( In this figure all strain pairs were put on the same plate . ) Figure 5 shows that both at 41°C and 40 . 7°C , the two downgraded parent strains , 21ADG and BY4742DG , still show a strong difference in thermotolerance . We sporulated the 21ADG/BY4742DG diploid strain and phenotyped 2464 segregants for thermotolerance . Examples are shown in Figure 5B . The segregants showed a range of thermotolerance and also transgressive segregation [28] , since some of the segregants showed poorer thermotolerance than the inferior BY4742DG parent ( e . g . segregant 9 in Figure 5B ) while others showed better thermotolerance than the superior 21ADG parent ( e . g . segregant 8 in Figure 5B ) . This suggests the presence of additional QTLs and causative genes influencing thermotolerance . From the 2464 segregants derived from the diploid 21ADG/BY4742DG we selected 58 thermotolerant segregants that grew at 40 . 7°C at least as well as the 21ADG superior parent strain , and repeated the pooled-segregant whole-genome analysis . We have used the same set of SNPs as generated in the previous sequencing of the 21A parent strain compared to S288c , for the mapping of QTLs linked to thermotolerance . A total of ten regions have a 2-sided p-value low enough for significance ( Figure 2 ) . Interestingly , two regions can be discerned with a clear difference between the original and downgraded pool ( Figure 2 , Table S1 online ) . The previous peak indicating linkage of one or more causative elements in the region between about 400 , 000 bp and 600 , 000 bp on chromosome XIV with the superior parent 21A ( QTL1 ) has shifted to a more upstream position in the mapping with the 21ADG downgraded superior parent ( QTL4 ) . In the region between 600 , 000 bp and 800 , 000 bp on chromosome XII , there is a new conspicuous peak , indicating linkage with the 21ADG superior parent ( QTL5 ) . We confirmed the statistical significance of these two new QTLs by scoring selected SNPs in the individual thermotolerant segregants and performing a binomial exact test ( Table 2 ) . For the remaining seven regions , the SNPs showed about 50% variant frequency in the unselected pool ( Figure S5 online ) . This suggests that the putative weak linkage from these regions is not caused by allelic incompatibilities . In addition , the significant association of the causative element ( s ) in QTL3 with the inferior parent ( 71% of the segregants had the genotype of the inferior parent , as determined by individual segregant genotyping ) observed in the first mapping was completely abolished in the second mapping ( 52% of the segregants had the genotype of the inferior parent ) , which reaffirms that PRP42 is the only causative gene in this locus . In a previous QTL mapping study of thermotolerance [22] , the authors identified the NCS2 allele of a clinical isolate as a superior allele compared to the inferior allele from the S288c control strain . Since NCS2 is located in the central region of QTL4 and since the NCS221A allele contains the same mutation ( A212T ) as identified in the previous study , we have tested whether NCS221A is also a causative allele in our genetic background . For that purpose , we performed RHA for NCS2 using a hybrid diploid strain constructed from the two downgraded parent strains . We found that the NCS221A allele supported higher thermotolerance compared to the NCS2BY4742 allele , indicating that also in our genetic background the NCS2 allele from the superior strain acted as a causative gene ( which does not preclude the presence of other causative genes ) . Deletion of the inferior NCS2BY4742 allele in the hybrid diploid strain also caused a conspicuous drop in thermotolerance ( Figure S6 online ) . Fine-mapping of QTL5 by scoring six selected SNPs individually in all 58 thermotolerant segregants enabled us to reduce the size of the QTL from 150 , 000 bp to 40 , 000 bp ( Figure 6A ) . We then divided this region into three fragments and performed bulk RHA with each fragment in the 21ADG/BY4742DG diploid strain ( Figure 6A ) . ( The fragments had an overlap of one gene . ) Evaluation of thermotolerance with the pairs of reciprocally deleted hemizygous strains revealed that FRAGMENT121A and FRAGMENT221A conferred higher thermotolerance than the corresponding fragments from the inferior BY4742DG parent . For FRAGMENT3 there was no difference ( Figure 6B ) . We then performed RHA with all individual genes of Fragments 1 and 2 containing non-synonymous mutations in their ORF ( as indicated in Figure 6A ) . However , for none of the genes tested there was a different effect on thermotolerance of the two alleles ( data not shown ) . We then applied RHA to the remaining genes in FRAGMENT2 and found that the SMD221A allele conferred higher thermotolerance compared to the SMD2BY4742 allele ( Figure 6B ) . Hence , it apparently acted as a causative allele in both FRAGMENT1 and FRAGMENT2 , since it was the only gene present in the overlap between the two fragments . The observation that replacement of FRAGMENT121A with FRAGMENT1BY4742 caused a similar reduction in thermotolerance compared to the replacement of FRAGMENT221A with FRAGMENT2BY4742 is consistent with SMD2 being the only causative gene in QTL5 . We confirmed by Sanger sequencing that SMD221A only displayed SNPs in the promoter and terminator region as compared to SMD2BY4742 ( data not shown ) . Hence , a difference in expression level may be responsible for the difference in thermotolerance . We have compared SMD2 transcription levels in different strains and with incubation at different temperatures . We found a higher level of SMD2 expression for 21A compared to BY4742 in cells growing exponentially in liquid cultures ( YPD at 30°C ) and also 21ADG showed a higher level of SMD2 expression under these conditions than BY4742DG ( Figure 6C ) . The difference in SMD2 expression level is also clear for the 21A/BY4742 RHA pairs , but there is no significant difference for the 21ADG/BY4742DG RHA pairs ( Table S3 online ) . This indicates that the mechanism of SMD2 in influencing thermotolerance cannot be solely due to differences in its transcript level , and other mechanisms such as post-transcriptional regulation may play a role . In the cross with the original parents , the QTL5 region did not show any indication of linkage to the genome of the superior parent strain 21A , with 37 out of 58 thermotolerant segregants of 21A/BY4742 having the SMD221A allele ( confirmed by genotyping the individual segregants , data not shown ) . We have also applied RHA for SMD2 in the original 21A/BY4742 hybrid . Interestingly , we could not detect any difference in thermotolerance at the two temperatures tested ( 40 . 7°C and 41°C ) ( Figure 7A ) . Knowing that 21ADG/BY4742DG lack only two superior alleles as compared to 21A/BY4742 and both PRP42 and SMD2 encode proteins forming the same spliceosomal complex , we constructed double hetero-allelic mutations for PRP42 and SMD2 in the 21A/BY4742 background , and evaluated thermotolerance of the strains . In the hybrid with the inferior PRP42 allele , the superior SMD2 allele caused higher thermotolerance compared to the inferior SMD2 allele , whereas in the hybrids containing the superior PRP42 allele , the two SMD2 alleles did not influence thermotolerance differently ( Figure 7B ) . The identification of SMD2 as a causative gene for thermotolerance indicates that our new approach of mapping with the downgraded parent strains is able to reveal minor loci and causative genes that escape detection in QTL mapping with the original parents , in this specific case because of epistatic interaction . We expressed the two PRP42 alleles from a centromeric plasmid in the parent 21A strain ( Figure S7A online ) and in the 21A prp42Δ strain ( Figure S7B online ) . In both cases , there was no difference in thermotolerance between the strains . On the other hand , comparison of the thermotolerance of strain 21A and that of the two heterozygous RHA strains showed that the RHA strain expressing the 21A allele had clearly lower thermotolerance than the other two strains ( Figure S7C online ) . The thermotolerance of the heterozygous RHA strain expressing the superior PRP42 allele from BY4742 was not higher than that of the 21A strain . These results show that the BY4742 allele of PRP42 is not able to enhance the thermotolerance level of the 21A strain further , apparently indicating that other factors become limiting for thermotolerance . One such other factor may be SMD2 . In the 21A strain it is present for 100% in the superior form , while in the heterozygous RHA strain , it is only present for 50% in the superior form . Hence , a dosage effect of SMD2 may possibly be limiting for thermotolerance in the heterozygous RHA strain expressing the superior PRP42 allele from BY4742 . The difference in ploidy or in the genetic constitution between the haploid 21A strain and the diploid RHA hybrid strains may also play a role , although this seems to be contradicted by the fact that we mapped the superior PRP42 allele using haploid segregants of the superior and inferior parents . Also in the study of Sinha et al . [27] , replacement of the inferior allele of MKT1 with the superior allele in the S288c strain did not cause the expected improvement in thermotolerance . In this paper we have shown that identification of new minor QTLs involved in complex traits can be successfully accomplished by crossing parent strains that have both been downgraded for a single QTL . Using this approach we have identified new QTLs and new causative genes , revealing an important role for RNA processing in high thermotolerance . This method has the advantage of maintaining all relevant genetic diversity and enough phenotypic difference between the two parent strains and thus significantly increases the chances of identifying minor QTLs . In principle , successive rounds of minor QTL mapping could be performed in this way by sequentially downgrading the two parent strains further , making use each time of a causative gene identified in a QTL linked to the superior parent and in a QTL linked to the inferior parent . The following yeast strains were used: prototrophic and heterothallic diploid strain MUCL28177 , which was isolated from orange juice in the region of Strombeek-Bever , Belgium , its haploid segregant MUCL28177-21A , referred to as 21A , and BY4742 ( Matα his3Δ1 leu2Δ0 ura3Δ0 lys2Δ0 ) [40] . Yeast cells were grown in YPD medium containing 1% ( w/v ) yeast extract , 2% ( w/v ) bacteriological peptone , and 2% ( w/v ) glucose . 1 . 5% ( w/v ) Bacto agar was used to make solid nutrient plates . Transformants were grown on YPD agar plates containing 200 µg/ml geneticin . Mating , sporulation and isolation of haploid segregants were done using standard protocols [41] . Strains were inoculated in liquid YPD and grown in a shaking incubator at 30°C overnight . The next day the cells were transferred to fresh liquid YPD at an OD600 of 1 and grown for 2 to 4 h to enter exponential phase . The cell cultures were then diluted to an OD600 of 0 . 5 and 5 µl of a fourfold dilution range was spotted on YPD agar plates , which were incubated at different temperatures . Growth was scored after two days incubation for all conditions . All spot tests were repeated at least once , starting with freshly inoculated cultures . Repetitions of the thermotolerance assays may show slight differences in growth intensity . Hence , the strains to be tested were always spotted together with the relevant controls on the same plate . SNPs were scored in individual segregants by PCR . At a given chromosomal location , two SNPs spacing between 500 and 1 , 500 bp were chosen for the design of specific primers . For a given SNP , two primers either in the forward or reverse direction , were designed with one mismatch at their 3′ ends . First , a gradient PCR was applied using genomic samples of 21A and BY4742 as templates , with each template tested with two primer combinations ( primer pair based on the sequence of BY4742 and primer pair based on the sequence of 21A ) . The annealing temperature at which the best distinguishing power was obtained with the two parents was used for scoring of the SNPs in the individual segregants . All the ORFs of non-essential genes in the centre of the QTL were deleted separately in both 21A and BY4742 . PCR-mediated gene disruption was used [46] . Plasmid pFA6a was used as a template to amplify a linear DNA fragment containing the kanMX4 cassette [47] , with 50 bp homologous sequences for the target regions at both ends . Transformants growing on YPD geneticin plates were verified by PCR with several combinations of internal and external primers . The verified haploid deletion strains were subsequently crossed with the matching wild type haploid to generate the hybrid diploids . For RHA with essential genes and fragments containing multiple genes , transformation was performed directly in the hybrid diploid . External SNPs primer pairs together with internal primers within the kanMX4 cassette were used in different combinations to determine in which parent the allele or the fragment had been deleted . For each heterozygous deletion hybrid , at least two isogenic strains were made and evaluated for thermotolerance . The growth of strains in the RHA test should always be compared within the strain pairs and not between the strain pairs , since the loss of one copy of a gene can cause an effect on the growth of the strains under non-restrictive conditions or even under restrictive conditions if the gene is important for the phenotype and because of the variability between different thermotolerance assays . . The replacement of MKT121A with MKT1BY4742 in 21A was performed by a two step transformation . For the first transformation , a linear DNA fragment with the AMD1 gene from Zygosaccharomyces rouxii flanked by 50 bp sequences that are homologous to the two sides of the MKT1 ORF was amplified from plasmid pFA6a-AMD1-MX6 [48] by PCR , and transformed into 21A . Transformants were grown on YCB ( Yeast Carbon Base 1 . 17% , phosphate buffer 3% , Bacto agar 2% ) plates containing 10 mM acetamide . Single colonies were checked for the correct replacement with the use of external and internal primers . For the second transformation , colonies were transformed with a linear DNA fragment containing the MKT1BY4742 ORF , together with ∼100 bp downstream and upstream . Transformants were grown on YNB galactose ( 0 . 17 Yeast Nitrogen Base w/o amino acids and ammonium sulfate , 1 . 5% Difco agar , 0 . 01% galactose , pH 6 . 5 ) containing 100 mM fluoroacetamide . Colonies were first checked for the presence of MKT1 by PCR , and then confirmed by DNA sequencing . The replacement of PRP42BY4742 with PRP4221A in BY4742 was performed in a two step transformation . For the first transformation , a URA3 gene was inserted ∼50 bp downstream of the PRP42 ORF in BY4742 . Colonies growing on –URA plates were confirmed to have a correct insertion by PCR . For the second transformation , a linear DNA fragment containing the ORF of PRP4221A together with ∼400 bp downstream and upstream was transformed into the previous colonies , and the transformants were grown on 5-FOA plates . Colonies were first checked for the right DNA polymorphism by SNP primer pairs , and then confirmed by DNA sequencing . All sequence data have been deposited in the Sequence Read Archive ( SRA ) at the National Center for Biotechnology Information ( NCBI ) and can be accessed with account number SRA058979 .
Most traits of organisms are determined by an interplay of different genes interacting in a complex way . For instance , nearly all industrially-important traits of the yeast Saccharomyces cerevisiae are complex traits . We have analyzed high thermotolerance , which is important for industrial fermentations , reducing cooling costs and sustaining higher productivity . Whereas genetic analysis of complex traits has been cumbersome for many years , the development of pooled-segregant whole-genome sequence analysis now allows successful identification of underlying genetic loci with a major effect . On the other hand , identification of loci with a minor contribution remains a challenge . We now present a methodology for identifying minor loci , which is based on the finding that the inferior parent usually also harbours superior alleles . This allowed construction for the trait of high thermotolerance of two ‘downgraded parent strains’ by replacing in each parent a superior allele by the inferior allele from the other parent . Subsequent mapping with the downgraded parents revealed new minor loci , which we validated by identifying the causative genes . Hence , our results illustrate the power of this methodology for successfully identifying minor loci determining complex traits and with a high chance of being co-responsible for the phenotypic difference between the original parents .
[ "Abstract", "Introduction", "Results", "Discussion", "Materials", "and", "Methods" ]
[ "genetic", "networks", "quantitative", "traits", "genetic", "maps", "gene", "function", "genome", "sequencing", "genome", "analysis", "tools", "model", "organisms", "linkage", "maps", "trait", "locus", "analysis", "molecular", "genetics", "linkage", "(genetics)", "genome", "complexity", "biology", "trait", "locus", "phenotypes", "genetic", "screens", "heredity", "gene", "identification", "and", "analysis", "genetics", "yeast", "and", "fungal", "models", "saccharomyces", "cerevisiae", "genomics", "gene", "networks", "complex", "traits" ]
2013
QTL Analysis of High Thermotolerance with Superior and Downgraded Parental Yeast Strains Reveals New Minor QTLs and Converges on Novel Causative Alleles Involved in RNA Processing
Many strategies to control opisthorchiasis have been employed in Thailand , but not in the other neighbouring countries . Specific control methods include mass drug administration ( MDA ) and health education to reduce raw fish consumption . These control efforts have greatly shifted the epidemiology of Opisthorchis viverrini ( OV ) infection over the last decade from presenting as densely concentrated "heavy" infections in single villages to widespread "light" OV infections distributed over wide geographical areas . Currently , the "gold standard" detection method for OV infection is formalin ethyl-acetate concentration technique ( FECT ) , which has limited diagnostic sensitivity and diagnostic specificity for light OV infections , with OV eggs often confused with eggs of minute intestinal flukes ( MIFs ) in feces . In this study , we developed and evaluated the diagnostic performance of a monoclonal antibody-based enzyme-linked immunosorbent assay for the measurement of OV excretory-secretory ( ES ) antigens in urine ( urine OV-ES assay ) for the diagnosis of opisthorchiasis compared to the gold standard detection FECT method . We tested several methods for pre-treating urine samples prior to testing the diagnostic performance of the urine OV-ES assay . Using trichloroacetic acid ( TCA ) pre-treated urine , we compared detection and quantification of OV infection using the urine OV-ES assay versus FECT in OV-endemic areas in Northeastern Thailand . Receiver operating characteristic ( ROC ) curves were used to determine the diagnostic sensitivity and specificity of the urine OV-ES assay using TCA pre-treated urine , and to establish diagnostic positivity thresholds . The Positive Predictive Value as well as the likelihood of obtaining a positive test result ( LR+ ) or a negative test result ( LR- ) were calculated for the established diagnostic positivity threshold . Diagnostic risks ( Odds Ratios ) were estimated using logistic regression . When urine samples were pre-treated with TCA prior to use in the urine OV-ES assay , the analytical sensitivity was significantly improved . Using TCA pre-treatment of urine , the urine OV-ES assay had a limit of detection ( LoD ) of 39 ng/ml compared to the LoD of 52 ng/mL reported for coprological antigen detection methods . Similarly , the urine OV-ES assay correlated significantly with intensity of OV infection as measured by FECT . The urine OV-ES assay was also able to detect 28 individuals as positive from the 63 ( 44 . 4% ) individuals previously determined to be negative using FECT . The likelihood of a positive diagnosis of OV infection by urine OV-ES assay increased significantly with the intensity of OV infection as determined by FECT . With reference to FECT , the sensitivity and specificity of the urine OV-ES assay was 81% and 70% , respectively . The detection of OV-infection by the urine OV-ES assay showed much greater diagnostic sensitivity and diagnostic specificity than the current "gold standard" FECT method for the detection and quantification of OV infection . Due to its ease-of-use , and noninvasive sample collection ( urine ) , the urine OV-ES assay offers the potential to revolutionize the diagnosis of liver fluke infection and provide an effective tool for control and elimination of these tumorigenic parasites . Opisthorchis viverrini ( OV ) infection is a major public health problem in the Mekong River Basin region of Southeast Asia , especially in Thailand , the Lao People’s Democratic Republic ( Lao PDR ) , Cambodia , and Vietnam [1 , 2] . The clinical sequelae of chronic opisthorchiasis are several advanced hepatobiliary pathologies [3] , the most concerning being advanced periductal fibrosis and intrahepatic cholangiocarcinoma ( CCA ) [4 , 5] . Based on its strong association with CCA , OV has been classified as a Group I biological carcinogen by the World Health Organization’s International Agency on Research in Cancer [6] . As chronic OV infection has such a fundamental role in the induction of bile duct fibrosis and CCA , a comprehensive strategy to control and eliminate these neglected tropical diseases ( NTDs ) has been undertaken in Thailand over the last several decades . After a decade of mass drug administration ( MDA ) and public health educational efforts to control consumption of the raw fish intermediate host in which OV metacercariae encyst , the epidemiology of OV infection has changed dramatically . Initially presenting as densely concentrated heavy infections , OV infections today appear "light" and are spread across extensive geographical regions , especially in Northeastern Thailand , where these control program have been most effective . Hence , a detection method which is analytically sensitive yet also rapid and easy to apply is required to monitor this “shifting landscape” of OV infection [7] . Currently , the “gold standard” diagnostic for OV infection is the formalin ethyl-acetate concentration technique ( FECT ) , which quantifies OV eggs in feces . The FECT method has several important drawbacks , including limited analytical sensitivity ( A-Sn ) : i . e . , light intensity infections can go undetected , requiring extensive fecal sampling over the course of days , which can be logistically onerous ( if not impossible ) in the resource limited settings where OV transmission is currently occurring . In addition , FECT has been shown in several studies [8] to have a limited analytical specificity ( A-Sp ) , with OV eggs often confused with the eggs from minute intestinal flukes infection ( MIFs ) and accurate distinction , requires the presence of an experienced microscopist [9 , 10] . Finally , advanced hepatobiliary pathologies from chronic opisthorchiasis such as biliary tract obstruction from bile duct fibrosis or primary biliary sclerosis , can obstruct the flow of eggs into the lumen and hence into feces , making the detection of light OV infection by coprological method nearly impossible [11] . Together , these limitations decrease the utility of FECT where OV transmission occurs in the Mekong Basin Region [8] . To address these limitations , several immunological and molecular assays have been developed to detect the presence of OV in feces , with varying diagnostic accuracy [3 , 10 , 12] . The molecular methods tend to be highly specific , but they often lack analytical sensitivity because of the presence of numerous PCR inhibitors in the feces [13] . Antigen detection in feces using a capture ELISA has also been developed and is a promising approach , as it is highly sensitive for light infections [14–16] . Monoclonal antibody ( mAb ) -based systems have been shown to be sufficient to detect OV adult secretory products from light infections , i . e . , when only a few adult worms are present and when eggs cannot be detected in the feces in an animal model [16 , 17] . However , these mAb-based antigen detection methods remain coprological assays; i . e . , they require numerous fecal samples for analysis . Immunodiagnostic methods utilizing human serum or plasma are also useful for detecting OV-infection and the concomitant risk of the hepatobiliary pathologies [18 , 19] . However , serodiagnostic methods require a blood draw , blood processing , cold chain refrigeration of sera or plasma , and trained phlebotomists , and usually prove logistically onerous and even unfeasible given the limited infrastructure of many of the laboratories in these research poor setting . Less invasive and easier to handle sample matrices such as saliva or urine are the ideal specimens to be collected for detecting OV in resource poor OV endemic setting in Southeast Asia [12] . Recently , our group has shown that parasite-specific IgG can be readily detected in urine of individuals with chronic opisthorchiasis [20] which has implications for the study of opisthorchiasis-induced hepatobiliary and renal abnormalities as well as the detection of people at risk of developing OV-induced CCA . The utility of urine samples , particularly ease and non-invasiveness of the collecting technique , has not been examined to date . In the current study , we developed and optimized an ELISA protocol to quantify the level of crude excretory-secretory ( ES ) OV antigen extract in urine samples and then assessed its diagnostic accuracy for the detection of OV-infection in field setting , with urine from individuals from OV endemic areas at variable infection intensities . We then compared the relationship between the detection of OV infection by our novel urine OV-ES assay and OV infection determined by the current “gold standard method” of FECT . This study is the first report on the performance of a method for urinary antigen detection for the diagnosis of opisthorchiasis . The study began in December 2013 and ended in December 2014 . Table 1 shows the sample sets used in the current manuscript . Sample Set 1 used to develop and Sample Set 2 to field-test the performance of the mAb-ELISA for OV-ES with TCA treated urine ( urine OV-ES assay ) compared to the current “gold standard” coprological technique FECT . Sample Set 3 was used to determine cross reactivity of the mAb-ELISA for OV-ES with individuals mono-infected with helminths . Negative urine samples from Sample Set 1 ( n = 10 ) were “spiked” with varying concentration of crude OV adult ES antigen extract starting with 5000 ng concentration of OV-ES , two-fold serially diluted to produce a standard calibration curve ( S2 Fig ) . A best-fit linear regression line was obtained from the serially diluted spiked urine . The regression line was utilized in each assay and functioned as the assay standard calibration curve from which antigen concentrations in urine samples were predicted by interpolation . To elucidate the limit of detection ( LoD ) of the assay , the same set of negative TCA treated urine samples were spiked with serially decreasing concentrations of OV-ES antigen . The highest concentration of spiked urine that interpolated below the value resulting from interpolating unspiked negative urine onto the standard calibration curve was considered the LoD of the assay . Urine from Sample Set 1 was used to evaluate the performance of urine treatment methods in the assay as follows: initially , the OD values obtained from the urine OV-ES assay were transformed to a ratio between the OD of the samples and the OD of reference urine samples . Wilcoxon rank-sum test was performed to compare the distributions of OD values obtained from various treatment methods ( freezing only ( i . e . , no treatment ) , heating , alkaline treatment , or TCA treatment ) ; Bonferroni correction was used to adjust for multiple comparisons ( Table 3 ) . A Receiver Operating Characteristic ( ROC ) curve [26] comparing different urine treatment methods in the urine OV-ES assay was constructed by plotting the sensitivity against 1-specificity for the OD values detected using each urine treatment method on FECT confirmed positives and negatives . The Area under the Curve ( AUC ) was a measure used to determine the probability of correctly identifying an OV positive individual ( as determined by FECT ) as a True Positive and an OV negative individual ( as determine by FECT ) as a True Negative . The urine treatment methods that generated the ROC curve with the highest AUC was considered the optimal urine treatment method for use in the urine OV-ES assay . Diagnostic sensitivity and diagnostic specificity , established from ROC curve analysis , were utilized to evaluate the performance of the urine OV-ES assay in an OV endemic area ( Sample Set 2 ) . To further characterize assay performance , the Positive Predictive Value ( PPV ) was calculated using a OV prevalence rate obtained from the field study which is comparable to the previous data from the OV endemic areas in the region [20 , 27] . These values were also used to compare the diagnostic accuracy of the OV-ES assay , using TCA treated urine to the gold standard FECT . The threshold for diagnostic positivity was obtained by maximizing the sensitivity and specificity of each data point from the ROC curve . The relationship between OV infection status in this sample and urinary antigen concentrations , determined by the urine OV-ES assay , was investigated with a logistic regression model . The model was used to describe the association between elevated antigen levels and positive OV infection status; odds ratios ( ORs ) , with corresponding 95% confidence intervals ( CIs ) , are reported in the results and describe this relationship . A 0 . 05 significance level ( alpha = 0 . 05 ) was utilized to determine meaningful predictors in the model . In an attempt to identify potential sources of false positivity , urine from OV-negative individuals with other helminths infections ( Sample Set 3 ) was treated with TCA and infection status was evaluated with the OV-ES urine antigen assay . The other helminths infections in this study sample included Strongyloides stercoralis ( n = 56 ) , minute intestinal flukes ( MIFs ) ( n = 13 ) ; hookworms ( n = 10 ) ; Taenia sp . ( n = 6 ) ; Echinostoma sp . ( n = 4 ) ; and Trichuris trichiura ( n = 3 ) . A false positive diagnosis resulted when a values from the urine of an individual infected with one of these helminths was above the cutoff value established from sample set 2 . This human subject protocol was approved by the Human Ethics Committee of Khon Kaen University , Thailand ( reference number HE561478 ) and written informed consent was obtained from individual subjects and those with parasite-positive examination by FECT were treated with anthelmintic drugs . The experimental protocol for monoclonal antibody production was approved by the Institutional Animal Ethical Committee , Khon Kaen University ( AEKKU-NELAC 26/2557 ) and was performed in strict accordance with the guideline for the Care and Use of Laboratory Animals of the National Research Council of Thailand . Statistical analyses were performed using SPSS 21 ( IBM , Chicago , IL , USA ) and SAS 9 . 3 ( Cary Institute , NC , USA ) . Kendall’s tau-b correlation test was used to determine the correlation between urinary antigen concentration and EPG . The usefulness of urinary antigen , detected by mAb-ELISA for diagnosis of opisthorchiasis , was evaluated in a multiple logistic regression model . Diagnostic accuracy of the mAb-ELISA for OV ES using TCA treated urine in terms of clinical sensitivity , clinical specificity , and the positive predictive values and negative predictive values were estimated by Receiver Operation Curve ( ROC ) curve analysis using MedCal ( MedCalc Software , Ostend , Belgium ) . Statistically significant level was set as p<0 . 05 . As shown in Fig 1 , the TCA method produced an ROC curve with the highest AUC of 0 . 9925 . The sensitivity and specificity for each urine treatment method are reported in Table 3 . Samples treated by the TCA method increased the OD range of the assay as shown in Table 3 , i . e . , treating urine with TCA produced the highest OD levels in the urine OV-ES assay compared to the other treatment methods . Fig 1 and Table 3 show the ROC curve analysis using individuals from Sample Set 1 ( n = 50 ) to compare the diagnostic performance of antigen detection when urine is pretreated using TCA , freezing ( -20°C ) , heating at 70°C , or alkaline before use in urine OV-ES ELISA . The sensitivity against 1-specificity for the antigen levels detected by each ELISA method from confirmed positives and negatives by the gold standard method ( FECT ) . Table 4 shows the threshold to obtain the diagnostic cutoff for positivity ( i . e . , OV infection ) using the urine OV-ES assay as determined by the point on the ROC curve ( Fig 2 ) where the diagnostic sensitivity ( D-SN ) and diagnostic specificity ( D-SP ) were concurrently maximized . The positive and negative predictive values ( PPV , NPV ) and positive and negative likelihood ratios ( LR+ , LR- ) were estimated based upon a prevalence of OV of 50% are also presented in Table 4 . It should be noted that despite the fact that the D-SN and the D-SP decreases when applied in the field setting decreases , the assay’s diagnostic performance remains robust with an AUC of 0 . 846 , a PPV of 75% , NPV of 76% , and a LR+ of 2 . 69 and LR- of 0 . 27 . Moreover , despite being developed in a small and targeted sample set ( Sample Set 1 ) , the application to a larger , randomly selected group ( Sample Set 2 ) verified the diagnostic performance of the urine OV-ES assay ( S1 Table ) . A logistic regression model was applied to this data set to determine the odds of having a positive diagnosis based on increasing OV-ES urine antigen levels as measured by the urine OV-ES assay and are presented in Table 4 . It should be noted that a one unit increase in urine OV as detected by the urine OV-ES assay has an increased odds of having OV of 9% while 10 unit increase urine OV as detected by the urine OV-ES assay has an increased odds of having OV of 230% . Based on urine samples from participants in Sample Set 2 ( Table 1 ) , tests for cross reactivity of the urine OV-ES assay for other helminth infections endemic to the area were used to evaluate the analytical specificity of the assay as shown in Table 3 . Most yielded negative tests except for 2 of 56 ( 3 . 57% ) subjects with S . stercoralis infections , and 1 of 10 subjects with hookworm infections ( 10% ) as seen in Fig 3 . The gold standard method for detecting OV infection is the coprological technique FECT , which is relatively sensitive for medium to heavy OV infection , but lacks analytical sensitivity for light OV infections . As shown in recent several studies , even light OV infection remains strongly associated with significant hepatobiliary pathology , such as advanced periductal fibrosis and CCA , for which this food borne trematode is best known [28] . Moreover , the FECT also lacks analytical and diagnostic specificity ( especially against the MIFs which are endemic in Mekong Basin Region ) , and can be logistically challenging , especially in the resource-poor settings of Southeast Asia , where OV is the most prevalent and its concomitant morbidity and mortality from intrahepatic bile duct cancer or CCA is highest in the world [29] . The coprological techniques for OV are especially onerous given the changing distribution of OV infection due to mass drug administration ( MDA ) , with the distribution of OV infection going from densely clustered medium to heavy infections to widely dispersed light OV infections , but with the rates of hepatobiliary pathology , especially CCA , remaining the highest in the world . Mass Drug Administration using praziquantel for opisthorchiasis in Thailand commenced in 1984 , when the prevalence of OV was on average 63% , with MDA continuing until 2001 which brought the average prevalence of OV down to 9 . 3% [1] . As such , the current coprological methods for the detection of OV infection are insufficient for this new “landscape” of OV-infection . Herein , we propose a urine-based detection method by the urine OV-ES assay , which utilizes advances already made in the serological testing for OV , but applied to pretreated urine samples . Currently , the serological diagnosis of OV is problematic due to the invasiveness of blood draw , cold chain requirements , a lack of analytical and diagnostic specificity , and the persistence of antibodies even after treatment for OV infection , making the determination of an active OV infection from a former OV infection impossible . However , the monoclonal antibodies used in the serological assays can be modified for use in alternative approaches to diagnose active OV infection: i . e . , urine OV-ES assay . Here , we adopted antibody approach using the non-invasive sample matrix ( urine ) in an easily applied assay that could be used to determine OV infection in these new epidemiological circumstances . In our study , the effects of four treatment protocols on urine specimens were compared . The TCA-treated urine yielded higher OD values in the urine OV-ES assay , with a greater discriminatory power between “known OV positive” and “known OV negative” samples than when urine was treated by freezing , heating , or alkaline prior to use in the urine OV-ES assay . These data confirm findings in a recent copro-antigen analysis that also utilized anti-OV-ES antibodies for the detection of opisthorchiasis [17] . The pre-treatment with TCA has been also successfully added to serum and urine samples to precipitate out the interfering proteins for successful antigen detection in schistosomiasis [24 , 30 , 31] . In the present study , LoD of the urine OV-ES assay was 39 ng/mL , which is better than the copro-antigen detection using this same monoclonal antibody of 52 ng/mL as previously reported [17] . The greater analytical sensitivity of the urine OV-ES assay is not surprising since the same monoclonal antibody clone ( clone KKU505 ) and only a slightly different urine OV-ES assay protocol were used in the current study from that of Watwiengkam and colleagues [17] . In regards to the greater specificity of the urine OV-ES assay , cross reactivity with urine from OV-negative individuals with other parasitic infections endemic to the region was observed in only 2 of the 56 subjects infected with S . stercoralis ( 3 . 57% ) and in only 1 from 10 subjects infected with hookworm ( 10% ) . It is likely that these cases may represent mixed infections of S . stercoralis or hookworm with low intensity O . viverrini infection , or that O . viverrini was a recent infection in the pre-patent period where eggs are yet to be detected . Furthermore , the results confirmed that there was no cross reactivity with other helminth parasites endemic to the area since the levels of antigen in subjects with O . viverrini infection alone or with mixed infection with other parasites ( i . e . S . stercoralis , echinostomes , hookworm , minute intestinal fluke and Taenia ) were not different . However , further tests of urine samples from non-endemic areas of opisthorchiasis are required to ensure the lack of this observed cross reactivity . In this study , analyses of quantitative urinary-antigen levels revealed that the intensity of OV infection was significantly correlated with the concentration OV-ES antigen in urine . This finding was similar to copro-antigen levels and intensity of infections with both OV [17] and experimental clonorchiasis [32] . Furthermore , a significant correlation has also been found between egg counts and levels of S . haematobium antigen in urine samples [33] . Therefore , the measurements of antigen level in feces and urine samples provide an advantage not only for diagnosis of light infection but also for the diagnosis of the intensity of OV infection . Whether the urinary antigen associates with severity of disease or not as observed in other diseases [34–38] is the subject of future investigations by our group . In comparison with copro-antigen detection in opisthorchiasis reported previously by Watwiengkam et al [17] , urinary antigen level was approximately 25 times lower than copro-antigen at infection intensities of 1–100 EPG . The explanation for this discrepancy is probably due to the fact that OV adult worms secrete antigen directly into the bile duct before being passively swept into the gastrointestinal tract and voided in the feces as copro-antigens . By contrast , using our urinary antigen detection method , the parasite antigen may be sequestrated or trapped in the liver or in other tissue , and diffused into the circulation before being excreted via the kidneys with urine . Therefore , unknown quantities of OV-ES antigen may be trapped in tissue with residual amounts are excreted into urine . Whether levels of urinary antigen detected have any relationship to kidney pathology as observed previously by our group with antibody to O . viverrini in urine is not known [20] . It worthy to note that , although very little underlying information is available , the fact that we are able to detect OV-ES antigen in urine specimens forms the basis for further studies to elucidate the mechanisms involved in the antigen secretion pathway of infected individuals , as well as its contribution in hepatobiliary pathology and formation of CCA . With reference to the gold standard diagnosis of FECT , performance of the urine OV-ES assay on field samples yielded a high diagnostic sensitivity and diagnostic specificity of 81% and 70% , respectively . The prevalence of OV by antigen detection ( 64% ) was higher than that determined by FECT ( 53% ) . The main difference was due to the finding of a considerable number of antigen positive cases ( 28 out of 63 cases , 44% ) that were deemed the egg-negative using FECT . Autopsy data have shown that a considerable number of egg-negative cases ( 70% ) can occur in individuals with worm loads <20 [39] , which suggests that the current gold standard FECT diagnosis may not be ideal and caution is needed when interpreting the data generated by this method . Another possibility is that the result of examinations using only a single fecal sample may underestimate the presence of light infection in endemic populations multiple fecal samples may help to increase the sensitivity of fecal examination [40] . Nevertheless , without a reliable gold standard method , a repeat of a specific method or combination of methods is required; for example , a combination of fecal examination and serological tests has been suggested for more accurate diagnosis of clonorchiasis [41] . Moreover , it is not feasible to detect eggs in cases of opisthorchiasis in the presence of biliary obstruction due to periductal fibrosis or in the case of low infection intensity [42] . Each infected person in this study was likely at a different stage of infection , particularly when worms are still immature and therefore eggs were not detectable by FECT . In this situation , metabolic products may be secreted in considerable quantity from the parasites such that it is possible to be readily detected by the urine OV-ES assay . Interestingly , we found that among the 97 cases from egg-positive individuals , antigens were detected only in 89 cases while 8 cases were found to be negative following urine OV-ES assay . These subjects had low EPG ( EPG 7–19 ) , which may be a problem with false-negative results . In addition to worm burden , the concentration of OV antigens is influenced by many other factors , such as duration and degree of biliary fibrosis , or the formation of circulating immune complexes that are difficult to detect [43] . In order to improve the diagnostic accuracy even more when the OV-ES assay is used in low transmission scenarios , one possibility is to modify the current protocol to increase volume of the urine sample in the assay . In the current protocol , we used 100 μL of sample diluted in 4% TCA , with only 50 μL of urine sample was used . A similar strategy has been successfully investigated in detecting circulating antigen in urine in Asian schistosomiasis ( Schistosoma japonicum ) in which the results showed that by increasing volume of urine sample from 250 μL to 2 mL of urine , the diagnostic sensitivity has increased 5 folds [44 , 45] . It is known that the gold standard diagnostic method by conventional fecal examination i . e . FECT has limitation in diagnostic accuracy [8] . However , the OV-ES assay in the current format also has a potential drawback in that it requires sophisticated instrumentation ( i . e . , an microplate reader ) , reagents , and well-trained technical staff . However , these drawbacks can be mitigated if the OV-ES assay is further developed into a simplified strip test kit for point of care ( POC ) use similar to that for schistosomiasis [46] . More studies on larger sample population , more tests in communities with different endemic settings and more tests with concurrent trematode infections i . e . schistosomiasis are needed . In conclusion , our results show that the urine OV-ES assay can be an effective tool for urinary-antigen detection in opisthorchiasis by the incorporation of a TCA urine treatment protocol , and strongly suggest that TCA-treated urine is a good alternative for detection of ES antigens of O . viverrini . Urine antigen and intensity of O . viverrini infection levels were strongly correlated . The advantage of using urine samples is the non-invasive ease of collection as well as high acceptability by individuals and the community . Moreover , this protocol has offered high sensitivity and specificity , which is essential for the effective surveillance and control of opisthorchiasis . Further studies are required to evaluate the performance and utility of urinary antigen detection for diagnosis of O . viverrini and to assess the effect of chemotherapy in opisthorchiasis .
Improved diagnostic methods for the detection of Opisthorchis viverrini ( OV ) infection in humans is required for effective surveillance and control of this food borne parasite and the prevention of OV-induced bile duct cancer ( cholangiocarcinoma or CCA ) . In this study , a novel urinary antigen detection method was established for quantitative diagnosis of opisthorchiasis by a monoclonal antibody-based enzyme-linked immunosorbent assay ( urine OV-ES assay ) . Analysis of paired feces and urine samples from 235 subjects in Don Chang sub-district in Khon Kaen Province , Northeast Thailand revealed 81% sensitivity and 70% specificity of the urine OV-ES assay when compared to the current gold standard diagnostic method . Moreover , levels of antigen detected by the urine OV-ES assay significantly correlated with intensity of OV infection ( P< 0001 ) , with and the proportion of antigen positive diagnosis associated with increasing intensity of infection . Forty four percent of individuals determined to be egg negative subjects by the gold standard method formalin ethyl-acetate concentration technique were positive by the urine OV-ES assay . The ease and noninvasiveness of urine sample collection and the high diagnostic accuracy of the urine OV-ES assay provide an alternative means for the diagnosis of human opisthorchiasis and facilitate the prevention and control of opisthorchiasis in resource limited setting of Southeast Asia .
[ "Abstract", "Introduction", "Materials", "and", "Methods", "Results", "Discussion" ]
[]
2015
Advances in the Diagnosis of Human Opisthorchiasis: Development of Opisthorchis viverrini Antigen Detection in Urine
The pathogenesis of rabies is associated with the inability to deliver immune effectors across the blood-brain barrier and to clear virulent rabies virus from CNS tissues . However , the mechanisms that facilitate immune effector entry into CNS tissues are induced by infection with attenuated rabies virus . Infection of normal mice with attenuated rabies virus but not immunization with killed virus can promote the clearance of pathogenic rabies virus from the CNS . T cell activity in B cell–deficient mice can control the replication of attenuated virus in the CNS , but viral mRNA persists . Low levels of passively administered rabies virus–neutralizing antibody reach infected cells in the cerebellum of B cell–deficient mice but are not sufficient to mediate virus clearance . Production of rabies virus-specific antibody by B cells invading CNS tissues is required for this process , and a substantial proportion of the B cells that accumulate in the CNS of mice infected with attenuated rabies virus produce virus-specific antibodies . The mechanisms required for immune effectors to enter rabies virus-infected tissues are induced by infection with attenuated rabies virus but not by infection with pathogenic rabies viruses or immunization with killed virus . T cell activities can inhibit rabies virus replication , but the production of rabies virus–specific antibodies by infiltrating B cells , as opposed to the leakage of circulating antibody across the BBB , is critical to elimination of the virus . These findings suggest that a pathogenic rabies virus infection may be treatable after the virus has reached the CNS tissues , providing that the appropriate immune effectors can be targeted to the infected tissues . Rabies viruses spread from peripheral sites of entry to the central nervous system ( CNS ) tissues via axonal transport thereby bypassing the specialized features of the neurovasculature known as the blood-brain barrier ( BBB ) . Once the virus reaches CNS tissues three alternative outcomes are likely: ( 1 ) the BBB remains intact and the infection is lethal due to the absence of an antiviral CNS immune response ( 2 ) immune effectors cross the BBB and mediate a CNS antiviral immune response with extensive immunopathology that contributes to the disease , or ( 3 ) immune effectors cross the BBB and clear the virus from the CNS without significant pathological consequences . It is well known that in humans naturally infected with rabies virus the latter outcome is exceedingly rare . In addition , CNS inflammation is generally limited in individuals who succumb to rabies [1] . Consequently , it is probable that the BBB remains intact through much of the course of rabies infection in humans . In the absence of a mechanism to compromise the barrier function of the neurovasculature , circulating rabies virus-specific immune effectors , whether raised by the infection or by active or passive immunization , would be unable to mediate an antiviral response in CNS tissues . This may explain why conventional post-exposure treatment of human rabies , consisting of active and passive immunization , is unsuccessful if begun after the appearance of signs of the disease [2]–[4] . At this stage of the infection the virus has likely begun to replicate in the CNS . Thus , the primary function of current post-exposure regimens may be limited to preventing the virus from reaching CNS tissues . Unlike humans where rabies viruses may take weeks to reach the CNS from the site of exposure [5] , the spread of most rabies virus strains to the CNS in mice is rapid with virus generally being detectable in CNS tissues within 48 hours of infection [6] . Nevertheless , normal mice survive infection with laboratory-attenuated strains of rabies virus [7] . While certain of these viruses may be deficient in the capacity to spread from the periphery to the CNS , most of the attenuated rabies virus variants that we have tested spread to and replicate in the CNS but are cleared by immune effectors that cross the BBB and infiltrate neural tissues [7] . In contrast , BBB integrity is maintained and immune effectors do not accumulate in the CNS tissues during infection of mice with common pathogenic rabies virus strains , despite the development of virus-specific immunity in peripheral lymphoid organs and innate immunity in the infected CNS tissues [7] , [8] . These observations have led us to speculate that the lethal outcome of infection with wildlife and pathological strains of rabies virus is at least in part due to the evasion of immune clearance as a consequence of the maintenance of BBB integrity [8] . Perhaps the best evidence that this may be the case is that disruption of the BBB in mice infected with a highly lethal silver-haired bat-associated rabies virus ( SHBRV ) , by triggering autoimmune CNS inflammation , promotes the clearance of the virus from the CNS tissues and survival [9] . Due to the associated pathology , the approach of using an autoimmune response to induce elevated BBB permeability and permit rabies virus-specific immune effectors to infiltrate CNS tissues is clearly inappropriate for use in human rabies . On the other hand , the neuroimmune response induced by infection with attenuated rabies virus , which also has the appropriate specificity , is not associated with significant pathology . In this study , we show that the functional changes in the BBB required to deliver immune effectors to the CNS tissues can be induced in mice infected with a lethal rabies virus strain by immunization with a live-attenuated virus vaccine strain but not by administration of killed virus vaccine . Furthermore , our data suggests that clearance of rabies virus from CNS tissues is dependent upon the production of virus-specific antibodies by infiltrating B cells . Eight to 12-week old wild-type control 129/SvEv and C57BL6 mice and JHD−/− mice on a C57BL6 background were obtained from the in-house breeding colony at Thomas Jefferson University . RAG-2−/− mice on a 129/SvEv background were obtained from Taconic ( Germantown , NY ) . Mice were infected or immunized intranasally ( i . n . ) with 105 focus forming units of CVS-F3 , CVS-N2c or UV-inactivated CVS-F3 in PBS as previously described [10] . In some experiments , mice were infected and immunized with a combination of the viruses . Where indicated , CVS-F3-infected JHD−/− mice were treated intraperitoneally with 1 mg of the monoclonal , rabies virus glycoprotein-specific , virus-neutralizing antibody 1112 in 500 µl of saline or with the vehicle alone at the time points noted in the figure legends . All procedures were carried out according to the protocols approved by the Institutional Animal Care and Use Committee of Thomas Jefferson University . BBB integrity was assessed by quantifying the leakage of a low molecular weight fluorescent marker ( Na-fluorescein , 376 kDa ) from the circulation into CNS tissues as previously described [10] . Briefly , 100 µl of 10% solution of Na-fluorescein was injected intraperitoneally and after 10 minutes mice were anesthetized and cardiac blood was collected followed by transcardial perfusion . Serum samples as well as supernatants of homogenized and centrifuged tissues were clarified by precipitating proteins with 15% TCA and the level of fluorescence measured with a CytoFluor™II fluorimeter . The amount of Na-fluorescein in the CNS tissue is normalized to its level in serum by ( µg of Na-fluorescein in CNS tissue/mg of tissue ) / ( µg of Na-fluorescein in serum/µl of serum ) and is expressed as a fold increase in fluorescence uptake by comparison with the results obtained from naïve controls . For immunohistochemical analysis , brains from perfused mice were snap frozen in Tissue-Tek O . C . T . Compound ( Sakura Finetex , Torrance , CA ) , sectioned using a Thermo Shandon cryostat ( Pittsburgh , PA ) , and fixed in either 80% acetone or 95% ethanol . Immunoglobulin ( Ig ) was detected using either biotinylated monoclonal rat anti-mouse kappa light chain ( 1 hour at 1∶50 ) ( BD Pharmingen , San Jose , CA ) followed by Alexa Fluor 568 streptavidin ( 1 hour at 1∶1000 ) ( Invitrogen , Eugene , OR ) or the VECTASTAIN ABC-AP KIT with polyclonal rabbit anti-mouse biotinylated IgG ( 1∶200 ) developed using the peroxidase antiperoxidase method and 3′3-diaminobenzidine as substrate ( Vector Laboratories , Burlingame , CA ) according to the manufacturer's protocol . For the additional staining shown in one of the figure panels ( #3D ) , a 1 hour incubation with 1 mg/ml 1112 was performed prior to detection of Ig . To assess virus infection sections were stained for 1 hour with FITC-conjugated anti-rabies virus nucleoprotein monoclonal antibody ( 1∶50 ) ( Centocor , Malvern , PA ) . Photographs were taken with a Nikon digital camera on an Olympus BX-60 microscope . Total RNA was isolated from CNS tissue samples and mRNA expression levels of rabies virus nucleoprotein , CD4 , CD8 , IFN-γ and L13 in CNS tissues were measured by quantitative reverse-transcriptase ( RT ) -PCR as previously described [10] . Real-time quantitative RT-PCR was carried out on cDNA using specific primer and probe sets and a Bio-Rad iCycler iQ Real Time Detection System ( Hercules , CA ) . The number of copies of specific mRNAs in each sample was determined as previously described [10] and normalized to the mRNA copy number of the housekeeping gene L13 in that sample . Data are expressed as the number of copies of mRNA for a particular gene in a sample per copy of mRNA for the housekeeping gene L13 in that sample . Mononuclear cells were prepared from peripheral blood collected by retro-orbital bleeding in heparinized capillary tubes by centrifugation at 300 g for 20 minutes . The white cell layer was washed in PBS twice before analysis . Mononuclear cells were isolated from CNS tissues as described elsewhere by isolation at the interface of a 30/70 Percoll ( Sigma ) step gradient centrifuged at 800 g for 25 minutes [11] . For flow cytometry , mononuclear cells were suspended in staining buffer ( PBS with 2% FBS and 0 . 1% NaN2 ) and incubated with anti-CD16/32 ( 1 ug/106 cells ) ( 2 . 4G2 BD Pharmingen , San Jose , CA ) antibody to prevent non-specific binding . Cells were washed in PBS and incubated with anti-mouse CD19 ( 1∶1000 ) ( 1D3 , BD Pharmingen , San Jose , CA ) and MHC class-II ( 1∶1000 ) ( 120 . 1 , BD Pharmingen , San Jose , CA ) antibodies . Phenotypic characterization of antibody-labeled cells was performed on a BD-FacsCaliber Flowcytometer . CD19-MHC class II double-positive cells were defined as B cells . Numbers of rabies virus-antigen specific antibody secreting B cells were assessed using Millipore Multiscreen HA® ELISPOT plates coated with 5 ug/mL of UV-inactivated whole rabies virus . Peripheral blood or brain derived mononuclear cells were suspended in RPMI media supplemented with 25 mM HEPES and 10% FBS and 200 , 000 cells were incubated in each well for 18 hours . Plates were washed and bound rabies virus-specific antibodies were detected by addition of alkaline-phosphatase conjugated anti-mouse IgG antibody ( 1∶500 ) ( Sigma , St . Louis , MO ) followed by BCIP/NBT substrate . Spots were counted using a dissecting microscope . Results are expressed as the mean±standard error mean ( S . E . M . ) . Statistical significance of the differences between groups was tested using the Mann-Whitney test and the symbol * indicates a p value<0 . 05 . Mouse models are not considered to be particularly suited to studies of post-exposure prophylaxis ( PEP ) with rabies due to the rapid spread of the viruses to the CNS . However , our prior studies suggest that the lethal outcome of rabies in mice is more a consequence of the inability to deliver immune effectors into CNS tissues than its spread [8] . In our view , the key feature is that BBB integrity is maintained during infection with lethal rabies viruses , while infection with attenuated rabies virus variants causes enhanced BBB permeability and a virus-clearing CNS immune response [7] , [10] . The reason for this difference could be that infection with highly pathological rabies virus strains causes the inhibition of immune mechanisms that mediate the changes in BBB function necessary for rabies-specific immune effectors to cross . Alternatively , these may not be triggered due to a subtle difference in rabies virus immune mechanisms induced by pathogenic and attenuated viruses . To distinguish between these two possibilities , mice were infected with the attenuated CVS-F3 variant or immunized with killed CVS-F3 and then 3 or 5 days later were super-infected with the pathogenic CVS-N2c rabies virus . CVS-F3 was administered first because the virus spreads to , and replicates in the CNS more slowly than CVS-N2c ( data not shown ) . The delays were limited to 3 and 5 days so that the CVS-N2c infection would have the 48 hours required to spread to the CNS before the appearance of serum rabies virus-specific antibodies approximately 8 days following CVS-F3 infection [10] . To control for unanticipated effects caused by the administration of either immunogen , groups of mice were also given both live and inactivated CVS-F3 . As shown in Table 1 , administration of an inactivated CVS-F3 vaccine preparation that is effective when given several weeks before a CVS-N2c challenge does not protect when given 3 or 5 days prior to challenge . On the other hand , the majority of mice infected with CVS-F3 as recently as 3 days previously survive CVS-N2c infection regardless of whether or not inactivated virus is also administered . These results suggest that the processes required to clear pathogenic rabies virus from CNS tissues are induced by infection but not immunization with CVS-F3 . JHD−/− mice lack B cells but have functional T cells and , unlike RAG-2−/− mice , which lack both T and B cells , are capable of elevating fluid-phase BBB permeability in response to the infection ( Fig . 1A ) . These mice are therefore suitable for analyzing the effects of antibody administration on CVS-F3 infection . As a preface to such studies we compared the course of CVS-F3 infection in JHD−/− and RAG-2−/− mice . As is the case for wild-type mice infected with CVS-F3 [10] , both JHD−/− and RAG-2−/− mice lose weight as the infection progresses ( Fig . 1A ) . However , while RAG-2−/− mice continue to lose weight and die approximately 20 days following infection , up to 70% of JHD−/− mice survive past this time-point , most showing a modest weight gain ( Fig . 1B ) . At the same time virus replication , which continues to increase in RAG-2−/− mice , becomes reduced in the JHD−/− mice ( Fig . 1C ) . These JHD−/− mice exhibit signs of rabies infection including ataxia and partial paralysis but survive the infection for extended periods of time ( >40 days ) . This raises the possibility that T cell activities may be able to partly control the virus infection independently of antibody . To gain insight into the contributing T cell subsets , we compared the levels of CD4 , CD8 and IFN-γ mRNAs in CNS tissues from wild-type conventional mice 24 days after CVS-F3 infection when there is little virus replication remaining ( see below ) and JHD−/− mice 40 days after infection . As can be seen in Fig . 2 , the levels of CD4 and CD8 mRNA are somewhat lower in the JHD−/− mice but IFN-γ mRNA levels have remained relatively high and may therefore be contributing to the control of virus replication . The inability of JHD−/− mice to clear CVS-F3 from the CNS reaffirms the importance of rabies virus-specific antibodies in this process . However , little is known with respect to how these antibodies may be delivered to infected CNS tissues . Our studies of mice clearing CVS-F3 suggest that the leakage of naturally developing antibodies from the circulation into the CNS tissues may be minimal since elevated BBB permeability occurs before serum antibody titers peak [10] . In addition , over the short term , extensive fluid phase exchange across the BBB is seen but little accumulation of markers of the molecular mass of antibody is detectable [12] . Nevertheless , it may be expected that some antibody would cross the BBB in conjunction with infiltrating immune cells and , over time , sufficient levels may accumulate to impact virus replication . To examine this possibility , JHD−/− mice were treated with 1 mg of the mouse IgG1 monoclonal rabies virus-neutralizing , glycoprotein-specific antibody 1112 , which is highly effective in post-exposure treatment models [13] , on each of days 7 and 9 post-infection when BBB permeability is at a peak . Several hours later , CNS tissues were obtained and stained with antibodies specific for rabies nucleoprotein and for mouse IgG to determine if there was any antibody associated with infected cells . While extensive infection of Purkinje cells can be readily detected with nucleoprotein-specific antibodies in sections from the cerebellum of JHD−/− mice ( Fig . 3A ) , as expected , there is no evidence of IgG in sections from animals that had not received antibody ( Fig . 3B ) . IgG-specific staining of Purkinje cells in sections from mice receiving 1112 antibody could be detected ( Fig . 3C ) but the treatment of these sections with additional 1112 antibody in vitro prior to IgG detection resulted in more extensive staining ( Fig . 3D ) . When cells stained for both nucleoprotein ( Fig . 3E , G green ) and antibody ( Fig . 3F , G red ) were examined more closely , distinct inclusions of nucleoprotein and antibody/glycoprotein can be seen . These findings suggest that low levels of 1112 antibody can leak from the circulation to interact with rabies virus-infected cells in the CNS tissues provided that their application coincides with elevated BBB permeability . To determine whether 1112 antibody administration to CVS-F3-infected JHD−/− mice leads to the clearance of the virus from CNS tissues , we administered saline or 1 mg of the antibody 5 times at two day intervals between days 7 and 15 post-infection . This antibody dose regimen achieved a half-maximal serum rabies-specific antibody titer of approximately 1/240 , which is roughly equivalent to the serum titer found in normal mice 8 days post-infection with CVS-F3 , during the period of time when BBB permeability is maximal . Viral nucleoprotein mRNA levels in the CNS tissues of surviving animals , both saline and antibody treated , were substantial several weeks later ( Fig . 4 ) at a time when they are virtually undetectable in wild-type mice [10] . Moreover , no impact on the health or survival of the mice was noted . CVS-F3 clearance from the CNS tissues of wild-type mice occurs prior to the development of high titers of circulating virus-neutralizing antibodies ( VNA ) and after BBB permeability has peaked [10] . However , B cells that have infiltrated the CNS tissues express high levels of κ-light chain mRNA during this time period indicating that there is likely to be substantial antibody production in the CNS tissues [10] . To assess this possibility more directly , we used antibodies specific for mouse IgG to stain CNS tissues from wild-type mice infected 12 days previously with CVS-F3 . Extensive foci of antibody are seen throughout the cerebellum ( Fig . 5 ) . At higher magnification the antibodies appear to be diffusing in stellate patterns from the foci ( Fig . 5 ) . To determine if B cells may be the source of these antibodies and whether or not they are likely to be rabies virus-specific , we assessed rabies virus-specific antibody production by B cells from the peripheral blood and CNS tissues of CVS-F3-infected mice . While the proportion of CD19+ B cells in mononuclear cells recovered from the CNS tissues of CVS-F3-infected mice is lower than in peripheral blood from the same animals , the fraction of the cells that produce rabies virus-specific antibodies is considerably higher ( Fig . 6 ) . This suggests that B cells producing rabies virus-specific antibodies either selectively invade or expand in the CNS tissues in response to CVS-F3 infection . Prompt administration of PEP is the recommended course for an individual who has come in contact with a rabid animal . Since this prevents the development of clinical rabies it is impossible to be certain how many of the tens of thousands of people who receive PEP on an annual basis have actually been infected with the virus . It is also impossible to know how far the rabies virus may have spread before being cleared by the immune effectors provided or induced by PEP and the infection . The commonly held view that pathogenic wildlife rabies virus that has spread to the CNS cannot be cleared by immune mechanisms is supported by the absence of significant immune cell infiltration into the CNS tissues of individuals who die from rabies [1] and the failure of PEP in individuals that have developed signs of rabies [2] , [4] . Our studies in animal models of rabies suggest that this is a consequence of the inability of virus-specific immune effectors to cross the BBB and enter CNS tissues infected with pathogenic rabies viruses [7] , [8] . The rabies virus-specific immune effectors that are raised in lethally infected mice are able to clear rabies virus from the CNS if provided access across the BBB . For instance , when the BBB is compromised by the induction of autoimmune CNS inflammation , rabies-specific immune effectors infiltrate CNS tissues and can clear the highly pathogenic SHRBV [9] . In addition , the adoptive transfer of immune effectors recovered from mice lethally infected with SHBRV results in clearance of the attenuated CVS-F3 virus from the CNS tissues of mice lacking T and B lymphocytes [8] . In contrast , the transfer of cells from mice clearing CVS-F3 has no impact on the outcome of SHBRV infection [8] . Regardless of the infecting virus strain , elements of the innate immune response that are important for the early control of virus replication and for attracting immune cells into infected tissues are induced [7] , [8] , [10] . These findings led us to speculate that functional changes at the BBB required to provide immune effectors access to the CNS tissues are induced during infection with attenuated rabies virus strains but not during pathological rabies virus infection [7]–[9] . A key issue examined in this study is whether or not this is due to an inhibitory process triggered by infection with pathogenic rabies virus . If so , it may be expected that BBB integrity would be maintained during infection with both pathogenic and attenuated rabies viruses and the outcome would be lethal , but it is not . Infection with an attenuated rabies virus induces BBB integrity changes and immune effector entry into CNS tissues regardless of whether or not there is also an ongoing infection with pathogenic rabies virus . However , protection is not provided by immunization with killed virus . We therefore conclude that the generation of a rabies virus-specific immune response in the periphery is not sufficient to clear pathogenic rabies viruses from the CNS tissues . A mechanism selectively induced by infection with attenuated rabies virus , likely manifested at the BBB , is necessary to provide immune effectors access to CNS tissues . To gain further insight into the mechanism of rabies virus clearance from the CNS tissues , we have used gene-deleted mice to study the role ( s ) of different antiviral immune effectors in the CNS tissues of mice clearing the attenuated rabies virus CVS-F3 . Mice lacking T and B cells cannot clear this virus and die from the infection [8] , [14] . CD8 T cells contribute to , but are not required for the clearance of CVS-F3 as clearance is merely delayed in mice without this cell population [14] , [15] . On the other hand , JHD−/− mice , which lack B cells but have functional CD4 and CD8 T cells , often survive CVS-F3 infection over extended periods despite being unable to clear virus from CNS tissues and exhibiting neurological symptoms . This leads us to conclude that elements of the T cell response , likely including IFN-γ production by CD4 and CD8 T cells , can control certain features of the infection that make significant contributions to its lethality but that antibody is required for virus clearance . To examine the contribution of circulating antibody to virus clearance from CNS tissues , we administered high levels of the rabies virus neutralizing mouse monoclonal 1112 antibody to CVS-F3-infected JHD−/− mice during the stage of infection when BBB permeability is maximal . While leakage of a 150 kDa molecular weight marker from the circulation into the CNS tissues of CVS-F3-infected mice is minimal over a 4-hour period [12] , antibodies present in the circulation over a more extensive period of time can evidently leak into the CNS tissues of the infected mice . 1112 antibody was found associated with the Purkinje cells in the cerebellum that express high levels of rabies virus antigen . The antibody was primarily localized in inclusion bodies which is consistent with previous in vitro studies showing that 1112 antibody is rapidly internalized by rabies virus-infected neuroblastoma cells where it accumulates in intracellular vesicles [13] . Of note in our studies is that the intracellular inclusions of glycoprotein-specific 1112 are generally distinct from inclusions of the virus nucleoprotein . The amounts of antibody reaching rabies virus-infected cells in vivo appears to be relatively low as considerably greater amounts of the antibody can bind to the cells when applied to tissue sections in vitro . While it is possible that even low levels of virus-neutralizing antibody may impact the replication and spread of the virus while BBB permeability is enhanced , treatment of CVS-F3-infected JHD−/− mice with 1112 antibody failed to clear the virus . It should be noted with respect to the origin of the antibodies that participate in rabies virus clearance that serum rabies virus-specific antibody titers peak some time after BBB integrity has been restored [10] . The presence of cells expressing the B cell phenotypic marker CD19 and mRNAs specific for κ- light chain in the CNS tissues of mice clearing CVS-F3 [8] , [10] led us to examine the possibility that rabies virus-specific antibodies are produced by infiltrating B cells . The current findings indicate that this is the case . Focal concentrations of antibodies can be readily detected in the CNS tissues of mice clearing CVS-F3 and a high proportion of B cells recovered from the tissues produce rabies virus-specific antibodies in vitro . This leads us to conclude that the high levels of antibodies required for rabies virus clearance from the CNS tissues are produced at the site of infection rather than diffusing in from the circulation . In this case , passively administered antibody during PEP would primarily impact virus in the periphery and an active immune response leading to elevated BBB permeability and immune effector delivery to the CNS tissues would likely be required to clear virus from the CNS . As certain of the aspects of BBB function required for immune cell infiltration are unchanged by CNS infection with pathogenic rabies viruses [8] , once the virus has reached the CNS a PEP protocol capable of altering the BBB , so that virus-specific immune effectors can reach the infected tissues , is required . Inactivated CVS-F3 can induce rabies virus-specific T and B cells , but fails to promote recovery from CVS-N2c infection over a time frame during which the administration of live CVS-F3 is therapeutic . We consider that this is a consequence of the inability of the inactivated virus to induce the functional changes in the BBB that are required for antiviral immune effectors to enter CNS tissues . In our view , administration of a live-attenuated rabies virus vaccine is the most reasonable , currently available , approach to providing the appropriate immune effectors access to the CNS tissues . The results of our experiments with a new , highly attenuated recombinant rabies virus vaccine which expresses three copies of a mutated glycoprotein gene , strongly support this hypothesis [16] . In these studies , the triple G vaccine was shown to promote immune effector delivery into CNS tissues and normal mice were found to survive the intracranial injection of a mixture of the vaccine virus and a highly pathogenic dog strain which was nearly 100% lethal when administered alone [16] . The triple G vaccine also proved effective in the post-exposure treatment of mice infected with a highly pathogenic dog rabies virus several days previously [16] . However , when UV-inactivated and given peripherally to mimic conventional post-exposure vaccination , there was little protective effect [16] . In addition to boosting the antiviral response , attenuated rabies virus vaccines spread to the CNS where they trigger the mechanisms required for T cells and B cells to enter the tissues and clear , not only the attenuated , but also pathogenic rabies viruses . It is clear from the commonly lethal outcome of rabies that these mechanisms are not induced in a timely fashion in the context of the spread of a wildlife rabies virus to the human CNS .
Every year over 50 , 000 people die from rabies worldwide , primarily due to the poor availability of rabies vaccine in developing countries . However , even when vaccines are available , human deaths from rabies occur if exposure to the causative virus is not recognized and vaccination is not sought in time . This is because rabies virus immunity induced by the natural infection or current vaccines is generally not effective at removing disease-causing rabies virus from brain tissues . Our studies provide insight into why this is the case and how vaccination can be changed so that the immune response can clear the virus from brain tissues . We show that the type of immune response induced by a live-attenuated rabies virus vaccine may be the key . In animal models , live-attenuated rabies virus vaccines are effective at delivering the immune cells capable of clearing the virus into CNS tissues and promote recovery from a rabies virus infection that has spread to the brain while conventional vaccines based on killed rabies virus do not . The production of rabies-specific antibody by B cells that invade the CNS tissues is important for complete elimination of the virus . We hypothesize that similar mechanisms may promote rabies virus clearance from individuals who are diagnosed after the virus has reached , but not extensively spread , through the CNS .
[ "Abstract", "Introduction", "Methods", "Results", "Discussion" ]
[ "virology/vaccines", "neurological", "disorders/infectious", "diseases", "of", "the", "nervous", "system", "immunology/immune", "response", "virology", "immunology", "virology/immune", "evasion", "immunology/immunity", "to", "infections", "virology/host", "antiviral", "responses" ]
2009
The Production of Antibody by Invading B Cells Is Required for the Clearance of Rabies Virus from the Central Nervous System
Imputation-based association methods provide a powerful framework for testing untyped variants for association with phenotypes and for combining results from multiple studies that use different genotyping platforms . Here , we consider several issues that arise when applying these methods in practice , including: ( i ) factors affecting imputation accuracy , including choice of reference panel; ( ii ) the effects of imputation accuracy on power to detect associations; ( iii ) the relative merits of Bayesian and frequentist approaches to testing imputed genotypes for association with phenotype; and ( iv ) how to quickly and accurately compute Bayes factors for testing imputed SNPs . We find that imputation-based methods can be robust to imputation accuracy and can improve power to detect associations , even when average imputation accuracy is poor . We explain how ranking SNPs for association by a standard likelihood ratio test gives the same results as a Bayesian procedure that uses an unnatural prior assumption—specifically , that difficult-to-impute SNPs tend to have larger effects—and assess the power gained from using a Bayesian approach that does not make this assumption . Within the Bayesian framework , we find that good approximations to a full analysis can be achieved by simply replacing unknown genotypes with a point estimate—their posterior mean . This approximation considerably reduces computational expense compared with published sampling-based approaches , and the methods we present are practical on a genome-wide scale with very modest computational resources ( e . g . , a single desktop computer ) . The approximation also facilitates combining information across studies , using only summary data for each SNP . Methods discussed here are implemented in the software package BIMBAM , which is available from http://stephenslab . uchicago . edu/software . html . We now describe more formally the imputation-based approaches to association mapping that we take here , and introduce notation for use in later sections . Readers whose primary interest is in our practical findings may wish to skip the Results section , and refer back to this section for reference . The imputation-based approach we take here is based on using a “prospective” model relating phenotypes to genotypes , and so is appropriate for analyzing phenotypes and genotypes on individuals sampled randomly from a population . We consider its applicability to other designs , such as case-control studies , in the Discussion . ( See [4] for related work based on a retrospective likelihood . ) Let Y denote measured phenotype values for n randomly-sampled individuals on whom genome-scan genotype data G have been collected . We use yi to denote the phenotype of individual i , and gij for the genotype of individual i at SNP j ( coded as 0 , 1 or 2 copies of the minor allele ) , with g·j denoting the vector of genotypes at SNP j . We also assume the availability of denser genotype ( or haplotype ) data , H , on a “panel” of unrelated , unphenotyped , individuals . In our examples this panel will consist of subsets of individuals from the International HapMap Project . For notational simplicity we assume that the columns of H and G are augmented , as necessary , so that they refer to the same set of SNPs ( i . e . so they have the same number of columns , with the jth column of each referring to the same SNP j ) . Thus , if a SNP j is typed in H but not in G then g·j will consist of all “missing” observations . The aim of our imputation-based mapping approach is to assess whether a SNP , j , that is genotyped in H but not in G , is associated with the phenotype Y . Let β denote the parameters in a model that relates phenotypes Y with genotypes g·j at SNP j . To obtain a likelihood for β , given the observed data Y , G , H , we take the standard regression-based approach of conditioning on the genotypes G , H , and using the conditional distribution p ( Y|G , H , β ) as the likelihood . This approach implicitly assumes that the genotypes alone contain no information about β ( or , at least , it ignores any such information ) . [Although in principal the genotypes alone could provide information about whether a SNP has an effect on phenotype ( e . g . through signatures of selection ) , such information seems likely to be both limited , and difficult to extract . ] Under this assumption the likelihood for β can be written as ( 1 ) where the last line invokes the assumption that the genotype distribution does not depend on β . From Expression ( 1 ) we see that , for an untyped SNP , the likelihood is a weighted average of the complete data likelihood ( i . e . the likelihood if the genotypes at SNP j were observed to be g·j ) over all possible values of g·j . The weights , Pr ( g·j = g|G , H ) , are determined by the distribution of the unobserved genotypes given the observed genotypes ( which can be obtained in several ways; here we use methods from [7] ) . This sum has too many terms to be computed directly . However , if the two distributions that occur in each term of this sum are independent across individuals , as they will be in settings considered here , then the likelihood simplifies to a computationally-tractable form: ( 2 ) We will compare two different approaches to using this likelihood to test the null hypothesis H0 that SNP j is unassociated with phenotype , versus the alternative hypothesis H1 that SNP j is associated with phenotype . The first approach is to use the generalized likelihood ratio test statistic , ( 3 ) where βˆ0 and βˆ1 denote the maximum likelihood estimates of β under H0 and H1 respectively . Under standard theory , 2log ( Λ ) has an asymptotic χ2 distribution under H0 . The second approach we consider is the Bayesian approach from [2] . In Bayesian statistics , the strength of the evidence for H1 vs H0 is given by the Bayes factor for H1 vs H0 , defined as ( 4 ) where p0 ( · ) and p1 ( · ) denote prior distributions on β under H0 and H1 . Note that this expression for the Bayes factor bears some resemblance to the likelihood ratio Λ , but whereas in Λ the numerator and denominator are maximised with respect to unknown parameters β , in the Bayes factor they are integrated with respect to these parameters , weighted by the prior ( which has the effect of averaging over plausible values of the parameters ) . Large values of the Bayes factor indicate evidence for H1 over H0 , whereas small values indicate evidence for H0 over H1 . Bayes factors have a number of general advantages over p values as measures of evidence [2] , [8] , [9] , and , as we show below , a particular advantage , at least in principle , when testing untyped SNPs . For a typed SNP at which genotypes are observed to be g , the Bayes factor , BF ( g ) , can sometimes be computed analytically ( see below ) . For an untyped SNP the Bayes factor is the weighted average of BF ( g ) over all possible values for g . Indeed , substituting the likelihood ( 1 ) into the numerator of ( 4 ) gives ( 5 ) Since this sum typically has too many terms to be computed directly , it must be approximated by computational methods; we compare three different methods later in this paper . We now give explicit expressions for the likelihood and Bayes factor in the case of a continuous ( quantitative ) phenotype Y , which will be our primary focus for the rest of this paper . We assume a normal linear model relating Y to the genotypes g·j at a SNP j of interest: ( 6 ) where μ is the phenotype mean for individuals with gij = 0; a and d represent , respectively , an additive and dominance effect for SNP j; 1 ( A ) is the indicator function , taking value 1 if A is true , and 0 otherwise; and ϵi are independent and identically distributed N ( 0 , 1/τ ) error terms , where 1/τ represents the variance of Y within each genotype class . Thus the parameters in this model are β = ( μ , a , d , τ ) , and the null hypothesis H0 is a = d = 0 . Under this model , the likelihood ( 2 ) becomes , for each yi , a mixture of three normal distributions: ( 7 ) where N ( ·;α , τ ) denotes the normal density with mean α and variance 1/τ , and α0 = μ , α1 = μ+a+d , α2 = μ+2a . To compute Bayes factors , we use a prior based on prior D2 from [2] . Among other assumptions , this prior assumes that a and d are independent and normally distributed , with mean 0 , and respective variances . As noted by [2] , the assumption that a and d are a priori independent is not ideal , since one might expect a and d to be dependent . However , dependence can easily be introduced by averaging results from prior D2 over several values of σa and σd , as we do here . Furthermore , in simulations , [2] found that the results from this prior generally agreed well with results from another prior , D1 , based on more realistic assumptions . Under this prior , the Bayes factor for a SNP with genotypes g , BF ( g ) , can be computed analytically ( [2] , Protocol S1 ) : ( 8 ) where Y̅ is the mean of the phenotypes Y; ν−1 is the 3×3 diagonal matrix with diagonal elements ; and X is an n×3 design matrix , which is a function of the genotypes g . Specifically , the first column of X is a vector of 1s , the second column is the vector of genotypes g , and the third column is a vector of indicators for whether the genotypes are heterozygotes ( Xi3 = 1 ( gi = 1 ) ) . To compute the likelihood ratio ( 3 ) and Bayes factor ( 5 ) we now need two things . First , we need expressions for Pr ( g·j = g|G , H ) ; that is , we need to have ways to estimate the ( distribution of ) genotypes at SNP j from the observed genotype data G , H . Second , because the sum ( 5 ) is over all possible genotypes at SNP j , and therefore typically contains a very large number of terms , we need efficient computational methods for approximating this sum . ( In contrast , Λ can be obtained relatively easily by numerical optimisation of ( 2 ) over β . ) The first two sections of the results , Genotype Imputation , and Bayes Factor Calculation , deal with each of these two issues in turn . Subsequent sections deal with comparisons of the use of the Bayes factor and Λ to detect associations with phenotypes , and the effects of imputation accuracy on power to detect associations . While there are many possible approaches to predicting unknown genotypes from patterns of LD [e . g . 10] , [11] , [12] , both [2] , [3] use similar methods based on Hidden Markov models: the PAC model from [13] implemented in software PHASE and IMPUTE , and the cluster-based model from [7] implemented in software fastPHASE and BIMBAM . In comparisons in [7] the two models produced very similar accuracy for imputed genotypes , and were more accurate than other methods considered . Both models also produced approximately calibrated predictions ( e . g . genotypes assigned a probability of 90% by these models were correct in approximately 90% of cases ) . Here we focus on the cluster-based model from [7] because it has certain computational and practical advantages over the PAC model ( e . g . it can deal easily with unphased panel data ) . However , we expect many of our conclusions to apply more generally ( e . g . see Discussion ) . In brief , the model from [7] assumes that each SNP along each sampled haplotype has an ( unknown ) cluster membership that changes , in a Markovian way , along the chromosome . Conditional on cluster memberships , alleles are sampled independently from cluster-specific and SNP-specific allele frequencies , θ . Other parameters of the model include the jump rates for cluster memberships , r; and the probabilities of jumping to each cluster , α . In the most general version of the model , which we consider here , both r and α are allowed to vary along the genome; see [7] for full details . [7] give an EM algorithm for estimating the parameters ( θ , α , r ) from either phased or unphased genotype data , and describe how the model can be used to impute missing genotypes and estimate haplotypic phase . For example , given parameter estimates ( θ̂ , α̂ , r̂ ) , the distribution of a missing genotype in individual i at SNP j can be approximated as Pr ( gij = g|gi· , θ̂ , α̂ , r̂ ) where gi· is the vector of observed genotypes in individual i . [7] compared several approaches to genotype imputation using data with genotypes missing at random . In their comparisons the cluster-based model provided more accurate imputed genotypes than other methods provided that i ) the number of clusters was sufficiently large , and ii ) predictions of genotype probabilities were averaged across multiple applications of the EM algorithm . Here we examine factors affecting the accuracy of genotypes imputed using this model in the context of imputation-based association mapping , where patterns of missingness are highly structured , not random . We took a genotype data , G , on chromosome 22 from 675 Caucasian individuals enrolled in the PRINCE study [14] who have been genotyped on the Illumina 317k chip as part of an ongoing genome-wide association study to study statin response and lipid-related phenotypes . To provide a set of test SNPs on which to assess imputation accuracy we masked all 675 genotypes at every 25th SNP ( masking 220 SNPs in each trial , this procedure being repeated 5 times by shifting the starting SNP , so masking 1100 SNPs in total ) . We then assessed various strategies for using the cluster-based model to impute the masked genotypes , using subsets of the HapMap data as a panel , H . Except where noted , results here are based on using the 60 HapMap CEU ( European ) parents as the panel . We assessed accuracy of imputed genotypes by comparing the true genotypes with the “best guess” imputed genotypes ( i . e . the genotype assigned the highest probability ) and computing the genotype error rate , being the proportion of best guess genotypes that are incorrect . The Bayes factor for an untyped SNP ( Equation ( 5 ) ) involves a sum over a very large number of terms , and it is computationally impractical to compute this sum directly . In practice then we must use methods to approximate this Bayes factor . Here we compare three different approaches to making this approximation . The first appproach is the “naive” Monte Carlo estimator used by both [2] , [3] , given by: ( 9 ) where g ( 1 ) , … , g ( M ) are independent and identically-distributed samples from Pr ( g·j|G , H ) . For sufficiently large M this estimator will give an accurate approximation to the BF ( 5 ) . More precisely , it converges to the true value as M→∞ . However , in practice we have found that for moderate values of M ( = 1 , 000 say ) this estimator can have a large standard deviation , producing unreliable estimates for some SNPs ( see below ) . The second approach is based on an importance sampling estimator [e . g . 15]: ( 10 ) where g ( 1 ) , … , g ( M ) are independent samples from an “importance sampling distribution” Q ( · ) . With judicious choice of Q the standard deviation of this estimator , and hence the accuracy of the approximation , can be much improved compared with BFnaive . Our importance sampling function is described in Text S1 . The third approach we consider is motivated by the simple idea of replacing unobserved genotypes at SNP j with their posterior mean , and computing a Bayes factor based on these posterior mean genotypes . To describe this approach more precisely , note that the Bayes factor for a typed SNP can be written as a function of a design matrix X ( equation ( 8 ) ) . The approximation we consider is to replace X with its expected value , X̅ ( so in the second column of X each element gij is replaced with E ( gij|G , H ) , and in the third column I ( gij = 1 ) is replaced with Pr ( gij = 1|G , H ) ) and to compute an approximate Bayes factor , BFmean , based on this expected design matrix: ( 11 ) Since the elements of X̅ can be computed analytically ( using either the PAC or cluster-based model ) the approximation BFmean is very quick to compute , reducing the number of Bayes factor evaluations by a factor of M ( and thus , typically , reducing computation for each SNP by orders of magnitude , once imputation has been performed ) compared with BFnaive and BFIS . We compared the three approximations by applying them to the genome-scan genotype data on chromosome 22 , described above , with phenotypes simulated under both null and alternative hypotheses ( see next section for details ) . Bayes factors were computed by averaging over σa = 0 . 05 , 0 . 1 , 0 . 2 , 0 . 4 and σd = σa/4 . For the sampling-based estimators BFnaive and BFIS we used M = 1 , 000 . Values of BFmean were generally similar to both BFIS , and BFnaive ( Figure 2 ) , with the agreement with BFIS being better ( presumably because BFIS has generally smaller standard error than BFnaive; see red vertical bars on Figure 2 ) . Furthermore , at SNPs where the values disagreed most strongly , the standard error of the Monte-Carlo estimators tended to be large ( see red vertical bars on Figure 2 ) , suggesting that at these SNPs the Monte-Carlo estimates may be unreliable . In summary , BFmean appears to provide an adequate approximation to the Bayes factor , more accurate ( for M = 1 , 000 ) than the naive Monte Carlo estimates used by [2] , [3] , and orders of magnitude quicker to compute . To explain why , and under what circumstances , BFmean provides an accurate approximation to the Bayes factor for H1 vs H0 , note that BFmean is , in fact , the Bayes factor for a different alternative hypothesis , ( 12 ) where ϵi are independent and identically distributed N ( 0 , 1/τ ) . The likelihood for this alternative hypothesis is very similar to the likelihood for H1 ( Equation 7 ) . In particular Yi has the same mean under and H1 . The differences between the two are that i ) under , Yi is normal , whereas under H1 it is a mixture of three normals; and ii ) under the variance of Yi is the same for each i , whereas under H1 the variance of Yi is larger for those i whose genotypes are less certain . These differences will be subtle unless a and d are large compared with 1/τ ( since with small a and d the three components of the mixture will have similar means , making the mixture of normals very similar in shape to a normal , and ensuring that differences among individuals in variance of Yi are small ) . In particular , our empirical results suggest that , at least for priors with σa<0 . 4 , BFmean generally provides an accurate approximation to the full Bayes factor . For different priors , using substantially larger values of σa , we have seen examples where the accuracy of the approximation is much poorer ( data not shown ) . However , in genome-wide association studies , where effect sizes of single SNPs tend to be small , the use of such large values of σa will not generally be desirable or appropriate ( although see next section ) . We note that it would be relatively straightforward to develop an improved approximation to the Bayes factor for H1 by modifying to account for the different variances across i . We do not pursue this here since BFmean appears adequate for our purposes . In subsequent sections we use BFmean to approximate the Bayes factor . Motivated by the fact that BFmean is the Bayes factor for vs H0 , we will also consider an analogous likelihood ratio statistic , Λmean , defined to be the likelihood ratio statistic for vs H0 . The statistic Λmean has several practical advantages over Λ: it can be computed analytically ( Text S2 ) , which results in a moderate computation saving ( a factor of around 10–20 in our implementation , although the exact saving will depend on the details of the numerical optimisation scheme used to obtain Λ ) ; and Λmean can be easily obtained from standard regression software since is a standard linear regression . ( In terms of ranking SNPs in order of significance , Λmean is equivalent to the standard F statistic for this regression . ) In this section we use both theoretical arguments and simulation experiments to compare and contrast the use of Bayes factors vs p values from likelihood ratio statistics for testing untyped SNPs for association . Besides comparing the Bayesian and frequentist approaches to inference , the results in Figure 4 also illustrate that , at least when using the Bayes factor to test imputed SNPs , the imputation-based approach is relatively robust to very poor imputation accuracy: even with very high average imputation error rate ( 25% error rate with YRI panel ) the imputation-based method using the Bayes factor as a test statistic performs better than testing typed markers only . This is reassuring for the use of imputation in studies where no well-matched panel is available . We also examined the effect of smaller changes in imputation accuracy ( imputation error rates in range 6 . 2% to 7 . 3% ) . As might be expected , such small changes in imputation error rate produce correspondingly small changes in performance ( comparable in magnitude to the difference between BFmean and Λ for the CEU panel in Figure 4; data not shown ) . While we have focussed here on testing quantitative phenotypes for association with genotypes , all of the key results should be expected to also apply to binary ( 0/1 ) phenotypes . Indeed , although the natural way to analyse a binary trait is via a logistic , rather than a linear , regression , for the small effect sizes that are typical in genetic studies the two approaches to analysis might be expected to produce similar results ( e . g . see [19] , p . 18 ) . To examine this further , we simulated binary phenotypes Y′ , by taking the quantitative phenotypes Y simulated above , and setting . We then compared the Bayes factors for these binary phenotypes under the linear model ( using the mean genotype approximation based on ( 8 ) ) , and the logistic model [using the Laplace method to approximate necessary integrals , as was done in 9; see Text S3] . The results ( Figure 5 ) show a strong correspondence between Bayes factors based on the two different models , supporting the idea that results obtained here for quantitative traits will apply also to binary traits . In summary , we have addressed a number of practical issues that arise in implementing imputation-based association mapping for a quantitative trait . Key findings include: i ) when using the model of [7] to impute untyped SNPs , fitting the model to only the panel produces substantially improved imputation accuracy compared with fitting the model to all data; ii ) although imputation accuracy is affected by mismatches in genetic background between panel and study samples , imputed genotypes can be accurate provided the panel contain at least some samples that are representative of the study sample; iii ) when computing Bayes factors based on imputed genotypes , simply replacing the imputed genotypes with their posterior mean produces a good approximation to a full analysis; iv ) when ranking imputed SNPs for association , in our simulations Bayes factors produced better rankings than Likelihood ratio statistics; our explanation for this , based on theoretical arguments , is that the Bayes factors take better account of the different amounts of information in different imputed SNPs; v ) The power of Bayesian imputation-based association mapping is relatively robust with respect to imputation accuracy: even when average imputation accuracy is low , imputation-based analysis can increase power compared with testing typed SNPs only . More generally , since in our simulations small differences in imputation accuracy had only a small effect on power , we conclude that choice of imputation method may ultimately matter less than what is done with the imputed genotypes . Thus , when selecting software to use in performing imputation-based analyses , it seems important to consider the range of analyses that can be performed , as well as the imputation method used . For example , our software , BIMBAM , can perform not only the single-SNP tests based on Bayes factors described here , but also multi-SNP analyses of individual genes or regions; and it can easily exploit any ( unphased ) resequencing data that may be available on regions of interest . All these features can increase power to detect , and ability to explain , associations [2] . While our findings are based on use of a particular imputation method , in the particular context of association mapping of a quantitative phenotype , all of them ( except perhaps i ) above ) are likely to apply more generally . For example , the PAC model used for imputation in [3] has many similarities with the one we use here , including the fact that imputation is essentially performed by modelling each sampled haplotype as a mosaic of template haplotypes ( in the PAC model these template haplotypes are the panel haplotypes , where in the cluster-based model these template haplotypes are estimated from the panel haplotypes , and are in some sense a summary of the panel haplotypes ) . As such , finding ( ii ) regarding the impact of mismatches in genetic background between panel and cohort individuals are also likely to apply to this method . Similarly , findings ( iii–vi ) seem likely to apply not only to quantitative traits , but also binary traits , or case-control studies , particularly given the correspondence between the Bayes factors based on linear and logistic models , in Figure 5 . Although imputation-based analyses involve considerably more computation than simply testing typed SNPs , these analyses are nevertheless now practical with very modest computational resources . For example , using the methods we describe here ( and particularly findings i ) and iii ) above ) , implemented in the software package BIMBAM , analysing the whole of chromosome 22 , in 675 individuals , with E = 5 and K = 10 takes just under an hour on a single Mac PRO desktop computer ( with 3 . 0 GHz CPU and 8G memory , although memory used for simulations here was less than 1G ) . Besides the gain in computational convenience , the effectiveness of Bayes factors based on posterior mean genotypes also has important implications for sharing and combining data across studies . Specifically , we have in mind situations where , for political , ethical , or other reasons , sharing individual-level genotype and phenotype data among investigators working on similar studies is more difficult than sharing summary-level data on each SNP . In these cases , a simple approach is to share and compare Bayes factors ( or p values ) for each SNP . However , better summaries of the data can be easily shared , to allow more powerful subsequent analyses . Specifically , for a quantitative trait , for a typed SNP the Bayes factor ( 8 ) depends on the phenotype and genotype data only through the matrices XtX , XtY and the number YtY . Similarly , for an imputed SNP , BFmean depends only on X̅tX̅ , X̅tY and YtY . These three quantities represent only summary level data ( for example , in the case of the observed SNP they are essentially equivalent to knowing the number of individuals in each genotype class , and the within-genotype-class means and standard deviations of the phenotype ) . Further , if these three quantities are known for multiple studies , they can be computed for the combined study . Specifically , if Xs and Ys denote , respectively , the design matrix and phenotype vector for study s ( s = 1 , … , S ) , and X and Y denote the combined design matrix and phenotype vector for the combination of data across all S studies , then it is straightforward to show that ( 13 ) Since XtX , XtY and YtY suffice to compute the Bayes factor for the combined study , this demonstrates that , from appropriate summary-level data from each study , one can compute the Bayes factor BFmean for each SNP as if one had possession of the combined data across all studies . ( Note that the same summary data also suffice to compute Λmean , allowing frequentist inference for the joint data also to be performed without sharing individual-level data . ) Of course , combining results across studies can present many challenges , including differential inclusion criteria or phenotype definitions; differential genotyping biases , e . g . due to differential DNA quality [20]; and different phenotype distributions and/or genetic backgrounds within different studies . Some of these problems may be easier to solve than others ( e . g . to minimize this last problem we advocate quantile normalizing the phenotype values within each study , to an N ( 0 , 1 ) distribution , prior to computing summary statistics . ) However , our results do provide a framework for combining results across studies when these other challenges can be surmounted . Our software BIMBAM includes an option to output and input summary statistics of this form , allowing multiple investigators to easily perform a combined imputation-based analysis of multiple data sets , without sharing individual level data among groups . The methods described here , like those from [3] , are based on a prospective likelihood . However , they have already been applied to other designs , such as case-control studies [9] , where use of a retrospective likelihood would be more appropriate . We now consider the validity of this approach . For case-control studies , with observed covariates , the correspondance between Bayesian analyses based on prospective and retrospective likelihoods is derived by [21] . In brief , they show that the Bayesian analysis based on a prospective likelihood is equivalent to a particular Bayesian analysis using the retrospective likelihood , provided that the prospective analysis uses an ( improper ) uniform prior on the baseline log-odds of the disease ( e . g . log-odds of disease for genotype 0 ) . [This is the prior implemented in our software , and it can also be shown ( MS , unpublished data ) that the approximate BF from [17] can be derived using this prior . This was not the prior used in [9] , although it is unclear how much this matters in practice . ] This result justifies the use of Bayesian methods based on a prospective model , with an appropriate prior on the baseline log-odds , for analysing typed SNPs in case-control studies . The result does not apply directly to unobserved covariates , and hence does not apply directly to untyped SNPs . However , we note that i ) if a SNP is easy to impute , with high confidence genotypes , then the result seems likely to hold approximately , since it is almost as if the SNP is observed; ii ) if a SNP is difficult to impute , with very low confidence genotypes , the Bayes factors from prospective and retrospective analyses will both be close to 1 , since there will be little information regarding association at such SNPs . Thus , for SNPs that are very easy or very difficult to impute the prospective analysis will approximately agree with a retrospective analysis . This gives grounds for optimism that results from applying these prospective methods to case-control data will not be generally misleading . The situation is slightly less clear for data obtained by genotyping individuals that are at the extremes of a quantitative phenotype distribution ( e . g . [22] ) , since the [21] result applies to binary phenotypes . Thus , although we speculate that direct prospective analyses of the observed quantitative phenotypes will be generally satisfactory here also , it might be prudent to also analyse such data treating the two extremes as binary ( case-control ) phenotypes .
Genotype imputation is becoming a popular approach to comparing and combining results of multiple association studies that used different SNP genotyping platforms . The basic idea is to exploit the fact that , due to correlation among untyped and typed SNPs , genotypes of untyped SNPs in each study can be inferred ( “imputed” ) from the genotypes at typed SNPs , often with high accuracy . In this paper , we consider several issues that arise when applying these methods in practice , including factors affecting imputation accuracy , the importance of taking account of imputation uncertainty when testing for association between imputed SNPs and phenotype , how imputation accuracy affects power , and how to combine results across studies when only single-SNP summary data can be shared among research groups .
[ "Abstract", "Introduction", "Results", "Discussion" ]
[ "mathematics/statistics", "genetics", "and", "genomics" ]
2008
Practical Issues in Imputation-Based Association Mapping